Beruflich Dokumente
Kultur Dokumente
AFT Impulse
User’s Guide
Information in this document is subject to change without notice. No part of this User’s
Guide may be reproduced or transmitted in any form or by any means, electronic or
mechanical, for any purpose, without the express written permission of Applied Flow
Technology.
"AFT Impulse", "AFT Fathom", "AFT Mercury", "Applied Flow Technology", and the
AFT logo are trademarks of Applied Flow Technology Corporation
Chempak is a trademark of Madison Technical Software, Inc.
Microsoft, "Visual Basic", Excel and Windows are trademarks or registered trademarks
of Microsoft Corporation. “CAESAR II” is a registered trademark of COADE, Inc.
1. Introduction .................................................................... 1
3. Fundamental Concepts................................................ 59
Detailed Contents
Summary..................................................................................................iii
Detailed Contents ..................................................................................... v
Nomenclature....................................................................................... xxix
1. Introduction .................................................................... 1
Terminology ............................................................................................. 1
Modeling capabilities ............................................................................... 2
Interface features ...................................................................................... 3
Thermophysical property data .................................................................. 4
Who can use AFT Impulse ....................................................................... 4
Engineering assumptions in AFT Impulse ............................................... 4
Installing AFT Impulse............................................................................. 5
Check hardware and system requirements......................................... 5
Read the README.TXT file............................................................. 5
Run the setup program ....................................................................... 6
Getting started with AFT Impulse ............................................................ 6
Example models ................................................................................. 7
Using online help...................................................................................... 7
Verification models .................................................................................. 7
AFT Impulse overview ............................................................................. 7
Input windows.................................................................................... 8
Output windows ................................................................................. 8
Converting models and databases from previous versions....................... 9
Pipe friction........................................................................................ 9
Infinite pipe junctions ........................................................................ 9
Transient junction data....................................................................... 9
Junction loss models .......................................................................... 9
Reservoir transients............................................................................ 9
Databases ......................................................................................... 10
3. Fundamental Concepts................................................ 59
Pipes and junctions ................................................................................. 59
Convention for flow entering and exiting............................................... 60
Features for modeling irrecoverable losses ............................................ 60
Aligning objects......................................................................... 81
Creating and Managing Saved Groups ............................................ 81
Special group usage.......................................................................... 82
Bookmarks on the Workspace.......................................................... 82
Last View ......................................................................................... 82
Panning the Workspace.................................................................... 82
Customizing features........................................................................ 83
Workspace Preferences ............................................................. 83
Adjusting the Workspace size ................................................... 83
Selective display of pipes and junctions.................................... 83
Scale/flip workspace.................................................................. 84
Renumbering the Workspace objects ........................................ 85
Using the Renumber Wizard ..................................................... 86
Incrementing Object Numbers................................................... 86
Background Graphic ........................................................................ 86
Specifications windows.................................................................... 87
Printing and exporting the Workspace............................................. 88
The Model Data window ........................................................................ 88
Display and printing features ........................................................... 90
The Model Data Control window .................................................... 90
Pipe data display........................................................................ 90
Junction data display ................................................................. 90
Show selected pipes and junctions ............................................ 91
Scenario Format......................................................................... 91
Database connections ................................................................ 92
The Output window ................................................................................ 93
Transient output files ....................................................................... 94
The Output Control window ............................................................ 95
Pipe and Junction parameters .................................................... 95
Pipe Transient parameters ......................................................... 97
General output ........................................................................... 97
Summaries ................................................................................. 97
Format and action ...................................................................... 98
ω pump speed
ρ density
μ dynamic viscosity
Subscripts
f fluid
i junction or pipe station at which solution is sought
j junctions with pipes connecting to junction I
m motor
R rated conditions for a pump, usually BEP
o stagnation
∞ infinity, far away, ambient
new current time for computation
old most recent time for computation
Introduction
Terminology
The term “waterhammer” has been used for more than one-hundred
years, and describes a transient phenomenon in piping systems caused by
Modeling capabilities
AFT Impulse can be used to model a wide variety of pipe system
transients, including:
• Transients in open and closed (recirculating) systems
• Network systems that branch or loop, with no limit on the number of
loops
• Systems with valve transients
Interface features
AFT Impulse's graphical interface is based on drag-and-drop operations,
which make it simple to build a model of a generalized pipe system. You
control the arrangement, and you benefit from the direct visual feedback
regarding the layout of your model.
Data is entered for the components in Specifications windows, which are
opened by double-clicking the component of interest. Additional global
editing features simplify making large-scale changes to the model.
AFT Impulse handles both traditional English and SI systems of units.
You assign units to all input parameters by choosing from lists. This
highly flexible approach removes the burden of hand calculated unit
conversions.
Note: You must use the setup program to install AFT Impulse. You
cannot simply copy the files from the installation disks to your hard
drive.
Before you run the setup program, make sure your computer platform
meets the minimum requirements.
Feature Required
Processor Pentium II or higher
Hard disk space 40 MB
CD-ROM drive Yes
Monitor resolution SVGA (800x600)
RAM 128 MB
MS-Windows Windows 98, NT 4.0,
2000, ME, XP or later
versions
Example models
An auxiliary help file (called ImpulseExamples.hlp) is installed with
AFT Impulse and leads the user through modeling a number of real
world systems. This Help file can be accessed by choosing "Show
Examples" from the help menu. The example models discussed in
ImpulseExamples.hlp are installed in the EXAMPLES folder.
Verification models
A large number of verification models have been built and compared to
published results from the open literature. These are included in the
Verification sub-folder below AFT Impulse. Along with the models you
will find documented comparisons in the Verification help file also
installed in the Verification folder.
Model Data
Visual Report
Workspace Output
Graph Results
Input windows
The two windows that function exclusively as input windows are the
Workspace window and the Model Data window. These two windows,
one graphical and the other text-based, work together to process model
input data with immense flexibility. The tools provided in these two
windows allow you to model a large variety of pipe networks.
The Visual Report window can function in support of both input and
output data. As an input window, it allows you to see the input data
superimposed on the pipe system schematic created on the Workspace.
Output windows
The two windows that function exclusively as output windows are the
Output window and the Graph Results window. The Output window is
text-based, while the Graph Results window is graphical. These two
windows offer a powerful and diverse range of features for reviewing
analysis results for modeling errors, gaining a deeper understanding of
the pipe system's flow behavior, and preparing the results for
documentation.
As an output window, Visual Report allows you to see the output results
superimposed on the pipe system schematic created on the Workspace.
The five primary windows form a tightly integrated, highly efficient
system for entering, processing, analyzing, and documenting
waterhammer analyses of pipe networks.
Pipe friction
AFT Impulse 1.0 only worked with constant friction factor models.
When these models are imported into version 4.0, the friction model for
the pipe is set to Explicit Friction Factor or Frictionless, depending on
whether the version 1.0 pipe had a zero or non-zero friction factor.
Reservoir transients
AFT Impulse 1.0 allowed reservoir surface level transients as a delta
from the steady-state. These will all get converted automatically to the
version 4.0 convention of specifying the actual liquid surface level.
Databases
AFT Impulse 1.0 databases can be converted using the AFT Convert
utility provided with the software.
Combining pipes
AFT Impulse 4.0 will offer to combine pipes for you when you import
AFT Fathom models. This will allow models to run much more
efficiently. When combined, AFT Impulse will create intermediate pipe
elevations at the points where junctions are deleted. All pipe resistance
and junction losses will be maintained in the model.
Databases
AFT Fathom and AFT Mercury databases can be converted using the
AFT Convert utility.
General
• Users can create transient force files for use in CAESAR II pipe
stress software
• Select Special on pipe material
• Significantly enhanced global editing of pipes and junctions (data
now shown in categories)
• Batch runs for scenarios
• All printing can go directly to a PDF file
• In grids, cells that have drop-down capability have a persistent drop
down symbol
• Initialization and data files now saved to user folders
Pipes
• Enhanced fittings & losses entry for pipes
• New pipe friction model uses resistance in dH/Q2
Junctions
• Valves now support Cv vs. percent open (like Control Valves)
• More advanced control valve logic
Transient Control
• Users can define “force sets” for purposes of plotting force data
• Runtime estimates based on users previous
Solver
• Power Law and Bingham Plastic now cover both laminar and
turbulent flow
Workspace
• Renumber Increment Added
• Shortcut button provides fast access to explanations of shortcut
functions
• Selections can be rotated right or left
• New window added to show all undefined pipes and junctions and
missing properties
Model Data
• Can see data for all scenario direct ancestors on Model Data
Output
• Users can print Model Data with output
Graph Results
• Can graph forces on pipes
This chapter is designed to give you the big picture of AFT Impulse's
layout and structure. Some of the more basic concepts will be used to
build an eight-pipe, eight-junction model to solve a waterhammer
problem published in the literature (Karney, et al., 1992). This system
has reservoirs, a surge tank, a relief valve and an exit valve.
This chapter is not intended to replace the more in-depth discussions
given in later chapters. To acquire a more detailed understanding of AFT
Impulse's menus and functionality before creating a model, skip over this
chapter.
A number of other example model discussions are included in a Help file
distributed with AFT Impulse called ImpulseExamples.hlp. It can be
opened from the Help menu by choosing "Show Examples".
Toolbars
Workspace
Toolbox
Minimized
primary
windows
Status Bar
¾ Next, drag a branch junction from the Toolbox and drop it to the
lower right of the reservoir J1. This becomes junction J2 (Figure 2.2b).
Figure 2.2b Walk Through Model with one reservoir and one
branch placed
¾ Now, drag a surge tank junction from the Toolbox and drop it to the
right of the reservoir J1. This junction takes on the number J3 (Figure
2.2c).
Figure 2.2c Walk Through Model after placing the Surge Tank
¾ To add a branch junction, select a Branch from the Toolbox and place
it on the Workspace as shown in Figure 2.2e. The Branch will take on
the default number J5. Similarly, add another branch at junction J6.
¾ Add valve junctions by selecting the junction icons from the Toolbox
and placing them on the Workspace as shown in Figure 2.2e. These
junctions will become J7 and J8, respectively.
¾ Before continuing, save the work you have done so far. Choose Save
As from the File menu and enter a file name ("WalkThru", perhaps) and
AFT Impulse will append the “.IMP” extension to the file name.
¾ To create a pipe, click on the Pipe Drawing tool icon. The pointer will
change to a cross-hair when you move it over the Workspace. Draw a
pipe below the junctions, similar to that shown in Figure 2.2f.
The pipe object on the Workspace has an ID number (P1) shown near
the center of the pipe.
¾ To place the pipe between J1 and J2, grab the pipe in the center with
the mouse and drag it so that its left endpoint falls within the J1
Reservoir icon and drop it there. Next, grab the right endpoint of the
pipe and stretch the pipe, dragging it until the endpoint terminates within
the J2 Branch icon (see Figure 2.2g).
¾ Activate the pipe drawing tool again and press the mouse button down
on the J2 Branch. Stretch the pipe up to the J3 Surge Tank and release
the mouse button. Draw a third pipe from the J3 Surge Tank to the J4
Reservoir. Draw a fourth pipe from the J3 Surge Tank to the J5 Branch.
Draw a fifth pipe from the J6 Branch to the J5 Branch. Draw a sixth pipe
from the J2 Branch to the J6 Branch. Draw a seventh pipe from the J6
Branch to the J7 Valve.. Finally, draw an eight pipe from the J6 Branch
to the J8 Valve. Your model should now look similar to Figure 2.2h.
Figure 2.2h Walk Through Model with all pipes and junctions
placed
¾ Open the Parameter and Units Preferences window from the Options
menu. This Walk Through model is in metric units. The default in AFT
¾ Select the Unit Preferences tab and then select the Default Unit System
as SI. Click the Apply Default Units button. See Figure 2.3. Finally click
OK to close the window.
¾ Next, click the checkmark on the Toolbar that runs across the top of
the AFT Impulse window. This opens the Checklist window (see Figure
2.4). The checklist contains six items. Each item needs to be completed
before AFT Impulse allows you to run the Solver.
Figure 2.5 The Output Control window lets you customize the
output
Figure 2.6 The System Properties window lets you enter physical
properties of the fluid
¾ Select System Properties from the Analysis menu to open the System
Properties window. For this example, select the AFT Standard fluid
option, then choose “Water at 1 atm” from the list and click the Add to
Model button. The properties for AFT Standard water are given only as
a function of temperature. Enter 20° C in the temperature box, click the
Calculate Properties button and click OK.
¾ Open the checklist once more or observe the Status Bar and you should
now see the third item checked off.
Object status
Every pipe and junction has an object status. The object status tells you
whether the object is defined according to AFT Impulse's requirements.
To see the status of the objects in your model, click the floodlight on the
Toolbar (alternatively, you could choose Show Object Status from the
View menu). Each time you click the floodlight, Show Object Status is
toggled on or off.
When Show Object Status is on, the ID numbers for all undefined pipes
and junctions are displayed in red on the Workspace. Objects that are
completely defined have their ID numbers displayed in black. (These
A. Define Reservoir J1
¾ Click OK. If Show Object Status is turned on, you should see the J1 ID
number turn black again, telling you that J1 is now completely defined.
Inspection
window
B. Define Branch J2
¾ Click the Optional tab and enter an imposed flow rate of -2 m3/sec (the
negative sign means that the flow is out of the junction – a flow sink).
¾ Click the check box to Model Short Connector Pipe. This allows you
to model a short pipe which is lumped together with the surge tank for
each solution. The data for the connector pipe is as follows: Friction
factor is 0.02, Pipe Diameter is 0.5 meters, Pipe Area is 0.1963 square
meters (it is cylindrical), Pipe Length is 30 meters, and Elevation
Change is 30 meters (i.e., the pipe is vertical because the elevation
change is the same as the length).
You also can model a flow restrictor (e.g., orifice) if one exists.
E. Define Valve J7
¾ Choose the loss model as Cv. This valve will initiate the transient for
this system. Until the valve changes position, the entire pipe system is in
a steady-state condition. Enter a Cv of 2392.6. This is the steady-state
value.
¾ Select the Transient Data tab. In this transient the valve partially
closes over a period of 10 seconds, then holds steady at the new position
for the remainder of the simulation. Enter the following Cv vs. time data
in the table (see Figure 2.10):
Time Cv
0 2392.6
10 797.4
100 797.4
F. Define Valve J8
¾ Click the Transient Data tab and select the Single Event option for the
Initiation of Transient. Select Pressure Stagnation at Pipe as the Event
Type, Greater Than or Equal To as the Condition, and a Value of 1.688
MPa at the outlet of Pipe 8. This Event represents the cracking pressure
required initiate the opening transient of the relief valve.
Enter the valve transient Cv data below in the Transient Data table (see
Figure 2.11a):
Time Cv
0 0
3 651.2
63 0
100 0
The data represents the valve initially closed at time zero. Time zero is
the moment of cracking. The valve then opens and gradually closes
again over a period of 60 seconds.
Because the valve is closed initially, set the valve Special Condition on
the Optional Data tab to Closed (see Figure 2.11b).
G. Define Pipe P1
The next step is to specify all the pipes. To open the Pipe Specifications
window, double-click the pipe object on the Workspace.
¾ First open the Pipe Specifications window for Pipe P1 (Figure 2.12).
Choose the Pipe Material as Unspecified, choose the User Specified
Wavespeed option, and select the Friction Model Data Set as
Unspecified with the Friction Model as "Explicit Friction Factor". Enter
a length of 1001.2 meters, a diameter of 1.5 meters, and a friction factor
of 0.012.
Note: This example from Karney assumes all friction factors are
known ahead of time. Normally this will not be the case. In most cases
you will access the roughness for the pipe from the pipe material
database supplied with AFT Impulse, or input your own roughness
value. Then AFT Impulse will calculate the friction factor using standard
methods (see Chapter 8).
¾ Open the Specifications window for each of the other pipes and enter
the following data (pipe 1 data was just entered so ignore it):
Pipe Length Diameter Friction Wavespeed
(meters) (meters) Factor (meters/sec)
1 1001.2 1.5 0.012 996.3
2 2000 1 0.013 995.3
3 2000 0.75 0.014 995
4 502.5 0.5 0.015 1000
5 502.5 0.5 0.015 1000
6 1001.2 1 0.014 996.3
7 2000.2 0.75 0.013 995.1
8 100.12 1 0.014 996.3
After entering the data for all the pipes, the fourth checklist item should
be completed. If it is not, see if the Show Object Status is on. If not,
select Show Object Status from the View menu or toolbar. If the fourth
checklist item is not completed at this point, see if any of the pipes or
junctions have their number displayed in red. If so, you did not enter all
the data for that item.
¾ Before continuing the model, save it to file one more time. It is also a
good idea to review the input using the Model Data window.
Customize view
General data
Pipe data
Junction data
Figure 2.13 The Model Data window shows the input data in text
form
Figure 2.14 The Transient Data tab in the Model Data Junction
data area shows all transient data entered
¾ Select Section Pipes on the Analysis menu to display the Section Pipes
Window (Figure 2.15). For this model the controlling pipe is P8. This is
the pipe with the shortest end-to-end communication time (i.e., L/a – the
length divided by the wavespeed). To satisfy the MOC, the following
equation must be applied:
Li
ni =
ai Δt
where n is the number of sections in pipe i, L is the length, and a is the
wavespeed. The Δt is the time step. Since all pipes in the network must
be solved together, the same time step must be used for each pipe. With
a given length and wavespeed for each pipe, it can be seen from the
above equation that it is unlikely that the number of required sections, n,
for each pipe will be a whole number.
¾ For Min. and Max. Sections in Controlling Pipe enter 1 section for
each. Limit the Max Error to 1% for the search (usually 5 or 10% is
Note: The error in the Section Pipes window relates only to sectioning
roundoff and not to overall model accuracy.
¾ Click OK to accept the current settings. The last checklist item should
be completed. The model is ready to be solved.
¾ Choose Run from the Analysis menu or click the arrow icon on the
toolbar. During execution, the Solution Progress window displays
(Figure 2.17). You can use this window to pause or cancel the Solver's
activity.
Customize view
General results
Pipe results
Junction results
¾ Select Output Control from the Analysis menu one more time. Select
the Pipe tab. The Reorder scroll bar on the far right allows you to reorder
parameters in the list.
¾ Select the Velocity parameter and use the Reorder scroll bar to move it
up to the top of the parameter list.
¾ Click OK to display the changes to the current results. You will see
in the Pipes table that the first column now contains velocity and the
third column contains the volumetric flow rate. The Output Control
window allows you to obtain the parameters, units and order you prefer
in your output. This flexibility will help you work with AFT Impulse in
the way that is most meaningful to you, reducing the possibility of
errors.
¾ Choose Select Graph Data from the View menu or the Toolbar to
open the Select Graph Data window (see Figure 2.20). From the All Pipe
Stations… list choose pipe 7. Select the Outlet of pipe 7. This is the
station connected to the exit valve. A graph of this station shows the
pressure and flow histories at the valve.
Figure 2.20 The Select Graph Data window controls the Graph
Results content
¾ Click the Add >> button to add this station to the list on the right.
Figure 2.24 The Visual Report Control window selects content for
the Visual Report window
¾ Click the Visual Report Control button on the Toolbar (or View
menu) and open the Visual Report Control window, shown in Figure
2.24. Default parameters are already selected, but you can modify these
as desired. For now, select Max Pressure Stagnation and Min Pressure
Stagnation in the Pipe Results area. Click the Show button. The Visual
Report window graphic is generated (see Figure 2.25).
It is common for the text in the Visual Report window to overlap when
first generated. You can change this by selecting smaller fonts or by
dragging the text to a new area to increase clarity (this has already been
done in Figure 2.25, as has the selection to show units in a legend). This
window can be printed, copied to the clipboard for import into other
Windows graphics programs, saved to file, or printed to an Adobe PDF
file.
Figure 2.25 The Visual Report window integrates results with the
model layout
Conclusion
You have now used AFT Impulse's five primary windows to build a
simple model. Review the rest of this User’s Guide for more detailed
information on each of the windows and functions.
Fundamental Concepts
Specifying losses
Local losses can be specified in two ways. The most flexible way of
including a loss factor in an AFT Impulse model is to associate it with a
junction specifically defined for that type of loss.
Secondly, you can include loss factors with pipes. When you specify
Fittings & Losses in the Pipe Specifications window, the loss is assumed
to be distributed evenly along the pipe length, much like friction losses.
For this reason a pipe-associated loss factor is referred to as a distributed
loss. The losses at junctions, on the other hand, are point losses.
When modeling for waterhammer, it is preferred that losses for all static
components be included in pipes as Pipe Fittings & Losses. This reduces
the number of pipes and model run time.
Variable resistance
When experiencing waterhammer transients, the flowrate in a pipe will
vary for each pipe computing station at an instant in time, and all pipe
stations will vary over time. As the flowrate varies, so does the Reynolds
number. This means that, strictly speaking, the pipe friction factor will
change as the flowrate changes, and will also change from station to
station.
This change in friction factor is typically neglected for two reasons.
First, it complicates the computation method. And second, it is usually
not important.
AFT Impulse allows the user to model either constant friction factor or
variable. The variable friction factor model results in model runtimes
Description of waterhammer
Waterhammer is a phenomenon that occurs in all liquid piping systems
whenever some event disturbs the steady state. Rapid disturbances can
cause large transient pressures that in extreme cases can cause a
catastrophic failure of the piping system and/or damage to associated
equipment.
For more slowly occurring disturbances, the transient pressures may be
small or even unnoticeable. In such cases it is sometimes said that
waterhammer doesn’t exist. What is meant by such statements is that
large, potentially destructive waterhammer doesn’t exist. Regardless of
the magnitude, waterhammer always exists when the liquid velocity in a
piping system is changed. In cases of small velocity changes
waterhammer transients are small and are typically and justifiably
ignored by engineers.
The disturbances consist of coupled pressure and velocity waves that
propagate throughout a pipe system. Across these waves a discontinuity
in magnitude exists. An example is given later in this section to clarify.
To picture what happens during a waterhammer event, consider a 1220
meter long pipeline flowing water in a steady state condition. Any
disturbance in the fluid will propagate near the acoustic speed of the
liquid in both the upstream and downstream directions.
It turns out that the propagation speed is influenced by the piping
material, wall thickness, diameter and structural support method. These
other factors combine into a parameter called the wavespeed, which is
traditionally referred to as a. The wavespeed is thus analogous to the
acoustic speed, but combines the effect of pipe interaction. Although the
wavespeed varies from case to case, a nominal value of 1220 m/sec is
reasonable for water.
Now picture a transient event in this pipe, specifically that the valve is
suddenly closed at the downstream end (see Wylie, et al., 1993, pp. 7).
Instantaneous waterhammer
With few exceptions, it is safe to calculate the maximum possible
waterhammer pressure surge by using the instantaneous waterhammer
equation. The instantaneous waterhammer equation assumes that the
transient event occurs either instantaneously or rapidly enough such that
it is in effect instantaneous.
In such a case, it can be shown (Wylie, et al., 1993, pp. 4) by use of the
mass and momentum equation that the pressure transient is given by the
following equation:
ΔP = − ρ a ΔV (3.1)
where:
ΔP = pressure surge
ρ = density
a = wavespeed
ΔV = velocity change
By adding the pressure surge to the existing static pressure, one can
obtain the maximum theoretical pressure in the pipe. However, in some
cases the pressure can exceed the instantaneous prediction. These are
discussed in Chapter 9.
Wavespeed
When a transient event is initiated in a pipe system, the remainder of the
system must adjust to the new conditions. In order to adjust, the
existence of the event must be communicated to the rest of the system.
This communication takes place at the wavespeed of the fluid. The
wavespeed is somewhat analogous to the sonic speed of the liquid.
However, the wavespeed is affected by the pipe structure.
See Chapter 9 for a mathematical description.
Conceptual example
Let's consider a simple example of a valve closure. A frictionless pipe is
connected to an upstream reservoir and a downstream valve. Fluid is
V = Vsteady V = Vsteady
V=0 V=0
a a
P P
Psteady ΔPinstantaneous Psteady ΔPinstantaneous
x x
V V
Vsteady
x x
-Vsteady
V = Vsteady V = Vsteady
V=0 V=0
a a
P P
Psteady Psteady
ΔPinstantaneous ΔPinstantaneous
x x
V V
Vsteady
x x
-Vsteady
This chapter explores the features found in AFT Impulse's five primary
windows and explains the role each plays in the analysis process.
Overview
AFT Impulse has five primary windows. The primary windows are
subordinate to the AFT Impulse window and can be maximized or
minimized within the boundaries of the AFT Impulse window. The
primary windows are permanent in that you work in one of these
windows at all times.
The primary windows work together to provide tools for entering model
input, analyzing results for accuracy, and preparing results for
documentation. Figure 4.1 summarizes the workflow using the primary
windows.
• The Workspace window allows you to build the model visually and
see the model layout.
• The Model Data window is the text-based complement to the
Workspace window. The Model Data window shows input data in
text form. This window works hand-in-hand with the Workspace
window to provide exceptional flexibility in manipulating data.
• The Output window displays the results of the analysis in text form
and lets you produce attractive, effectively organized printed output.
• The Visual Report window merges the output data with the pipe
system layout from the Workspace to present a unique perspective of
Model Data
Visual Report
Workspace Output
Graph Results
Toolbars
Workspace
Toolbox
Minimized
primary
windows
Status Bar
Branch Reservoir
Tee/Wye Valve
General Component
If you double-click the Pipe Drawing Tool it remains active until you
click it again a single time. This allows you to draw a series of pipes
without returning to the Toolbox each time.
This feature is useful in graphically bending the pipe around another part
of the model or graphically showing that there are embedded elbows as
Fittings & Losses.
You can add a segment to a pipe in two ways. The easiest way to add a
single segment is to select Pipe Segments and Add from the Arrange
menu. If there are no segments, then a single segment can also be added
by selecting the Add/Remove Segment from the Toolbar. If the pipe
Junction icons
Below the four drawing tools are nineteen junction icons. The junctions
allow you to model a large variety of pipe system components important
in waterhammer. To add a junction to a model, drag the desired icon
from the Toolbox and drop it anywhere on the Workspace.
As you move the mouse pointer over the Toolbox, a Tool Tip identifies
the type of junction under the mouse pointer.
The types and order of junction icons on the Toolbox can be modified
through the Toolbox Preferences window. For more details on how to
customize the Toolbox, refer to Chapter 7.
Editing features
Junction and pipe objects on the Workspace can be manipulated as
individual items or as groups. This applies to cutting, copying, pasting,
deleting, duplicating, dragging, locking, and undoing. Pipe objects on the
Workspace can be stretched by selecting the pipe at one end and
dragging it to a new location.
Aligning objects
Tools to align components in your model are located on the Arrange
menu.
Last View
The last view of the Workspace is always saved in memory and can be
displayed by selecting Last View from the View Menu or Toolbar.
Customizing features
Workspace Preferences
The Workspace Preferences on the Options menu lets you customize the
appearance and behavior of the Workspace. New defaults can be saved
so AFT Impulse will always appear and behave according to your
preferences.
See Chapter 7 for more information on Workspace Preferences.
Scale/flip workspace
The Workspace can be scaled horizontally and/or vertically by choosing
Scale/Flip Workspace on the Arrange menu (see Figure 4.9). Scaling
will move the junctions and will stretch or shrink the pipes so that the
area the model covers is changed by the percentage specified.
Flipping the Workspace horizontally and/or vertically will move the
pipes and junctions to their mirror image locations. This is useful when
Background Graphic
You can load a graphic into the background of the Workspace and build
your model on top of that graphic. The graphic could be, for instance, a
topographic map or a facility drawing. Such graphics are loaded from the
File menu by choosing Load Background Picture. At your option the
background graphic can be included in Workspace printouts.
Specifications windows
Each object on the Workspace has an associated Specifications window.
The Specifications window lists all the input data for the selected object.
There are three ways to open an object’s Specifications window from the
Workspace:
1. Double-click the object
2. Select the object and click the Open Pipe/Jct Window icon on the
Toolbar
3. Select the object and press ENTER
The Specifications windows can also be accessed through the Model
Data window.
General data
section
Customize
View
Pipe data
section
Junction data
section
The other display tab in the pipe area is the Pipe Detail Summary, which
assembles a list of details about each pipe in another format useful to
some engineers.
The junction data area is below the pipe data area. Here the junction data
is separated into tables for each junction type. Click the folder tabs to
see the data for any type of junction.
The transient data for each junction is shown on the Transient Data tab
(see Figure 4.12).
To edit information in the Model Data window, you can open the
Specifications window for the pipe or junction. This can be done by
double-clicking the far left column where the pipe or junction ID
numbers are shown. Alternatively, you can use the Global Edit windows
opened from the Edit menu.
Scenario Format
A powerful Model Data feature when using scenarios is to show all
direct ancestor data in the Model Data along with data from the current
scenario. The Scenario Format features allow you to do this, and to
highlight where data changes occur. See Figure 4.14. Note that the
scenario name is appended to the pipe and junction number in the left
columns.
Database connections
In the lower left of the Model Data Control window there is a Database
checkbox. If this box is checked, your Model Data Control parameters
are set up as determined by the database to which you are connected.
This is referred to as an active database. To make it inactive, uncheck
the box or change one of the Model Data Control settings controlled by
the database.
If the checkbox is unchecked, but enabled, you are connected to a
database but the settings are not being passed to the Model Data Control
window. The database is thus inactive. To make it active, check the box
then click the OK button.
If the checkbox is disabled, there is no connected database.
Customize view
General section
Description of
output terms
Database connection
Figure 4.16 The Output Control window selects parameters for the
Output window. Here the Pipe parameters (for steady-
flow results) are shown.
General output
The General section is the text area located at the top of the Output
window. The project title, reference information and other settings
affecting the General Output are defined here.
You can enter a descriptive title which will be used in the Model Data,
Output and Visual Report windows. A title is required and can have up
to 100 characters.
You can keep a lengthy explanation or any other documentation about
your model in the Reference Information section. Names of projects,
individuals, and assumptions can all be kept with the model. This
information can then be included in your reports to improve tracking.
The checkboxes allow you to enable certain special reports to be shown
in the General Section.
Summaries
Several Special Summary Reports can be displayed in the Output
window. These include a Pump Summary and Valve Summary. The
contents of these summary reports and the engineering units can be
specified in the Summaries list at the right.
Database connections
In the lower left of the Output Control window there is a Database
checkbox. If this box is checked, your Output Control parameters are set
up as determined by the database to which you are connected. This is
referred to as an active database. To make it inactive, uncheck the box or
change one of the Output Control settings controlled by the database.
If the checkbox is unchecked, but enabled, you are connected to a
database but the settings are not being passed to the Output Control
window. The database is thus inactive. To make it active, check the box
then click the OK button.
If the checkbox is disabled, there is no connected database.
See Chapter 7 for an in-depth discussion of how databases are
configured and administered through your local or wide area network.
Command buttons
There are eight buttons at the bottom of the Output Control window.
Impulse has built-in default parameters, units and settings which you can
choose by clicking the Impulse Default button. You can also develop
your own settings, tailored to your project or industry, and have these
used by default (instead of Impulse’s defaults). To make your own
default, first select the output parameters, units and settings you would
like to use then click the Set As Default button. Your settings will be
saved and will be used each time any new project is initiated. If you
make changes to the settings, and want to get back to your defaults, click
the User Default button. The saved default settings are updated only
when you click Set As Default.
You can save the output settings (except for the selected pipes/junctions
and junction deltas) to a file by pressing the Save Control Format button
and entering a file name. These settings are loaded again by pressing the
Load Control Format button and choosing the file name. For example,
you may have a final report format that is always desired. You can load
this format before generating final results. If you have another format
you use for reviewing model accuracy, you may want a larger number of
parameters in the output.
Figure 4.17 The Output window shows steady flow results in the
Pipes table and the various junction tables. Transient
data is shown on the transient tabs.
Buttons to
increment or
decrement
currently
displayed time
Time step
selection tool
slides up and
down
Figure 4.19 The Transient Max/Min table details the maximum and
minimum values for the entire simulation.
Figure 4.20 Detailed and Summary versions are available for the
Transient Max/Min table
Figure 4.23 The Select Graph Data window controls the Graph
Results window
Profile Along a Flow Path and EGL, HGL and Elevation Profile
These two types of graphs will plot the selected parameter along a flow
path. The pipes that comprise the path must form a single, continuous
path. The independent variable is the length along the pipe(s).
A path of pipes can be specified in one of two ways. First, the user can
select the pipes in the displayed list. This permits only a single path to be
specified.
Second, one or more flow paths can be specified using groups (see
Figure 4.24). Only groups which consist of a pipe sequence will be
displayed. Groups are set up on the Workspace using the Groups tool on
the Edit menu.
The results can be the values at a particular point in time, or the overall
maximum and/or minimum values. The Profile Along a Flow Path graph
type also allows cross-plotting of Design Alerts such as max/min
allowed pressures.
Both of these graph types support animation, discussed in the next
section.
Another advantage of the Animate Using Output File option is that it can
be made to run much faster. When the animation speed control is set to
increase speed, intermediate time steps are skipped and not plotted. If
the speed is increased too much, the quality of the animation deteriorates
because too much missing time data makes it difficult to follow what is
happening. When the speed is increased, the Animate Using Output File
option just skips time steps and reads in the data for each required time
step.
On the other hand, the Animate Using Solver option cannot skip
calculation steps because it is running the Transient Solver. To obtain
the results for a future time step, it must calculate all of the intermediate
time steps.
The biggest advantage of the Animate Using Solver option is that it does
not require one to save all of the required pipe station data to disk in
order run an animation. Because it is obtaining data from the Transient
Solver, it can always run any animation desired. If one uses the Animate
Using Output File option and has not saved all of the required data, the
Figure 4.26 Transient Force data can be plotted using force sets
defined in the Transient Control window
General display
The General Display tab allows you to configure several other items.
The title, defined in the Output Control window, and filename can be
shown or not shown depending on your preference. The scenario name
can also be shown.
Figure 4.29 The Visual Report with the units and Color Map in the
Legend
The units for all parameters can be shown in a legend instead of next to
the parameter value (see Figure 4.27 or 4.28). This will reduce the
amount of text and crowding shown in the Visual Report window. The
legend can be moved to any location in the Visual Report window. If this
The order of entry is not important. Impulse will sort the entries first by
value and then by the operator, with the > being above the <. The pipes
Figure 4.31 The Visual Report Control Color Map allows you to
specify the display colors for pipes according to their
output values.
Command buttons
You can save the output settings (except for the selected pipes/junctions)
to a file by clicking the Save Options button and entering a file name.
These setting are loaded again by pressing the Load Options and
choosing the file name. This file may be shared among engineers or
incorporated into a company-wide database.
Impulse will initially try to place the text for pipes and junctions based
on their positions on the Workspace. You can drag the text to new
locations which will give the best visual presentation. These positions
are saved with the model. You can reset the locations back to where
Impulse initially placed the text by clicking the Reset Locations button
and then confirming the operation.
If you have made changes which you don’t want to keep, click the
Cancel button. Click Show to apply your settings.
The Toolbars
The Toolbars offer quick access to the features used most frequently in
AFT Impulse. Figure 4.32 shows the Toolbars. There are actually six
toolbars in all. One is the Common Toolbar, and it is always shown. Its
features are common to all five Primary windows. In addition, there is
one Toolbar for each Primary window, offering features appropriate for
work in that window. Each of the functions on the Toolbar is also
available through the menu system.
Creating objects
The pipe flow model is assembled on the Workspace by arranging
graphical objects (pipes and junctions) that represent parts of a physical
pipe flow system. In addition to pipe and junction objects, annotation
objects can also be added.
Pipes
To create a pipe, select the Pipe Drawing Tool from the Toolbox and
draw a line on the Workspace to represent the pipe. To draw multiple
pipes more quickly, hold down the CTRL key while drawing each pipe.
This keeps the Pipe Drawing Tool active, so you don't have to select it
for each pipe. Alternatively, you can double-click the Pipe Drawing Tool
to “lock” it down. You can then continue to draw pipes until you click
the Pipe Drawing Tool a second time.
To draw a vertical or horizontal pipe, hold down the SHIFT key while
drawing the pipe.
This feature is useful in graphically bending the pipe around another part
of the model or graphically showing that there are embedded elbows as
Fittings & Losses.
You can add a segment to a pipe in two ways. The easiest way to add a
single segment is to select Pipe Segments and Add from the Arrange
menu. If there are no segments, then a single segment can also be added
by selecting the Add/Remove Segment from the Toolbar. If the pipe
already has one or more segments, this toolbar selection will remove all
segments.
The other way to segment a pipe is to select the pipe on the Workspace
then choose the Pipe Segments and Manager from the Arrange menu. A
window will display where you can add, remove and merge segments. A
“handle” will appear in the middle of the pipe for each new segment. As
the cursor passes over the handle, the cursor will change to a crosshair.
Press and hold down the left mouse button and drag the middle of the
pipe to the desired location. By holding the SHIFT key the pipe segment
will move only at right angles.
Junctions
To place a junction, drag the icon from the Toolbox onto the Workspace.
The nineteen available junctions represent various pipe system
components.
When you duplicate a junction using the Duplicate or Copy and Paste
features on the Edit menu, the new junction retains the engineering
information associated with the original junction.
Junctions are shown on the Workspace with default ID numbers, which
you can change in the Junction Specifications window. The letter “J”
precedes the ID number to signify that it is for a junction. From the
junction’s Specifications window you can change the ID number to any
desired value greater than zero and up to 30,000. Identical ID numbers
can be assigned to both a pipe and a junction because the “P” or “J” will
distinguish the two.
Unlike pipes, junctions do not have reference positive flow directions.
Junctions for which the flow direction is important (such as pumps and
control valves) derive the flow direction from the flow direction of the
connecting pipes.
Morphing junctions
A junction can be “morphed” from one type to another. To morph a
junction, hold down the CTRL key and select a junction from the
Toolbox and drop it onto an existing Workspace junction. The junction
type will change to the new junction type, and data that can be kept will
be copied into the new junction.
Splitting pipes
If you hold down the SHIFT key while selecting a junction from the
Toolbox and then drop the junction onto an existing pipe, the pipe will
split into two pipes. The physical length of the original pipe will be
automatically halved, and the new pipe will be assigned the balance.
Thus the sum of the two pipe lengths will equal the length of the original
pipe. Any fittings/losses in the original pipe will be left in the original
pipe, and the new pipe will have no fittings/losses.
Annotations
To create an annotation, select the annotation tool on the Toolbox and
draw the outline of the annotation. When you release the mouse button,
the Workspace Annotation window will display. Here you can enter text
for the annotation, draw lines and arrows, change colors, and draw
outlines. You also can change the colors of the annotation.
Annotations can be cut, copied, pasted, duplicated, and moved like pipe
and junction objects. When you select an existing annotation, resizing
handles appear.
The Annotation Manager assists you in specifying which annotations
will display in which scenarios.
Moving objects
Objects on the Workspace can be moved individually or as groups.
Clicking a Workspace object selects it. Pipe objects change color when
selected. Junction and annotation objects are outlined in red when
selected. This color is configurable in the Workspace Preferences.
To move an object, select it, drag it within the Workspace, and drop it in
the desired location. When an object is dragged beyond the existing
Workspace area, the Workspace is expanded accordingly.
A pipe object can be stretched by grabbing the handles (small black
squares) at the pipe endpoints and moving an endpoint to a new location.
To prevent accidental movement, lock the objects on the Workspace.
The Lock feature is accessed from the Edit menu or the lock button on
the Toolbar.
To group multiple objects for movement or other operations, hold down
the SHIFT key when selecting the objects. Objects can also be grouped
by drawing a box around them using the Selection Drawing Tool on the
Toolbox. You can use the Select Special on the Edit menu to group
objects based on specified criteria.
You can also select all objects in a flow path as follows:
• Select one object in the flow path
• Choose Select Flow Path from the Edit menu or Toolbar
Tip: You can undo all pipe and junction movements using Undo on the
Edit menu or Toolbar. You also can press the escape key during the
movement to cancel the move.
Keyboard modifiers
When you drag a junction, the endpoints of any connected pipes are
moved with the junction, thus maintaining the graphical connection.
This functionality may be changed through the Workspace Preferences
window.
Editing objects
The objects you place on the Workspace can be edited with the editing
commands from the Edit menu or the Toolbar. Objects can be cut,
copied, pasted, duplicated, and deleted. These operations can be
performed on individual objects or on groups. The Edit menu provides
one level of undo for each editing operation. Alternately, you can use the
Undo button on the Toolbar.
The Copy Graphics feature on the Edit menu will copy the specified
Workspace objects to the Windows clipboard. This image can then be
pasted into other Windows applications.
Connecting objects
During construction of a new model, objects can be placed anywhere on
the Workspace. In order to assemble a model that is ready to submit to
the Solver, you must connect the objects properly.
Remember that connectivity only exists between junctions and pipes.
There are no junctions that connect to junctions, and no pipes that
connect to pipes.
Defining objects
Each Workspace object in a model must be defined before AFT Impulse
can obtain a solution. AFT Impulse examines your model for proper
definition before it gives you access to the Solver.
To define an object, you must specify all the required property data for
the object and satisfy its connectivity requirements as described in the
following sections.
The Show Object Status feature identifies pipe and junction objects in
the model that are not completely defined.
Inspecting objects
The Inspection feature is a time saving feature for reviewing the data
associated with a Workspace pipe or junction object. Inspection displays
the data for an object in read-only format (see Figure 5.3).
Figure 5.4 The six Checklist items help you track the status of
your model
Status bar
Status light
Figure 5.5 The Status Bar shows the status of each Checklist
item. The Model Status Light turns green when the
model is complete.
Analysis Type
The first item on the Analysis menu is Analysis Type. This drops down
to two menu choices: Transient and Steady Only. By default the
Transient option is selected when a new model is opened. A transient
model always performs a steady-state analysis first, and then proceeds to
the transient solution. The Transient menu selection requires all six
Checklist items to be completed. The Steady Only option requires only
the first four.
Section Pipes
The fifth Checklist item, Section Pipes, is only used when the Analysis
Type is Transient. The details of pipe sectioning are related to the
transient solution method, the Method of Characteristics (MOC)
discussed in Chapter 9. In summary, the pipe must be broken down into
computing sections so the MOC can be used. There are strict criteria on
how this is done, and the Section Pipes window automates this process.
The Section Pipes window can only be opened after required pipe input
and fluid properties has been entered, as these data impact how the
sectioning is performed.
Transient Control
The sixth Checklist item, Transient Control, is only used when the
Analysis Type is Transient. Here the model start and stop times are
defined, as well as what data to save to the output file and estimates of
model run time and output file size.
Output Control
Most engineering analyses involve processing a very large amount of
information, some of which is critical but most of which is of lesser
Description of
output terms
Database connection
System Properties
The System Properties window gives you control over the fluid and
environmental properties that influence your pipe system analysis.
Figure 5.8 shows the System Properties window, which is accessed
through the Analysis menu.
The three fluid properties that are required for all analyses are density,
dynamic viscosity and liquid bulk modulus of elasticity. If only a steady-
state analysis is being performed, the liquid bulk modulus of elasticity
will not be used although input is still required. Fluid vapor pressure is
an optional specification.
These properties can be entered by hand when the fluid is selected as
Unspecified, or they can be obtained from the database list of AFT
Standard or Chempak fluids. When selected from the list, density,
The basis of all mixtures is either mass or mole, and is specified on the
System Properties window itself. There are no limits to the number of
components in a particular mixture.
The mixtures you create here are referred to as pre-mixtures because you
are specifying the composition before the model is run.
Accuracy option
There is one accuracy option for Chempak fluids and mixtures. The
State Property Accuracy option influences the number of elements used
in curve fits for calculating property data. Typically you should use the
high accuracy option.
Figure 5.10 The Create Mixture window (opened from the System
Properties window) allows you to create predefined
mixtures for the model.
Viscosity models
The default viscosity model is Newtonian, which applies to a wide
variety of important liquids including water. Some liquids exhibit a
dependence of viscosity on the fluid dynamics, and these fluids are
called non-Newtonian. When one of these model is selected, additional
input parameters are required (see Figure 5.11).
AFT Impulse offers four non-Newtonian viscosity models. Two of the
models apply to pulp and paper. These are the Duffy model and the
Brecht & Heller model. Also available are models for Power Law fluids
and Bingham Plastics. See Tilton, 1997 for more information on these
viscosity models.
Also available are models for Power Law fluids and Bingham Plastics.
See Darby, 2001 for more information on these viscosity models.
More extensive discussion of these viscosity models is given in Chapter
8.
Atmospheric pressure
AFT Impulse uses a default atmospheric pressure of 1 standard earth
atmosphere, or 101.325 kPa. You are free to change this value. This
pressure is also used to convert absolute pressures to gauge pressures.
Keep in mind that atmospheric pressure varies with altitude. If you are
doing one project at Sea Level and another at 2500 meters you should
adjust the Atmospheric Pressure. See Figure 5.12 for the System Data
tab.
Gravitational acceleration
The default for gravitational acceleration is 1 standard earth
gravitational acceleration, or 9.81 m/s2. You can change this to a
multiple of standard earth accelerations (that is, number of g's) according
to the design environment for your system. In principle, the body force
on a fluid system does not have to be due to gravity. See Figure 5.12 for
the System Data tab.
Section Pipes
In a transient simulation, a common time step must be used for all pipes.
The Method of Characteristics (MOC) requires that each pipe section
satisfies the following:
Δx
a= (5.1)
Δt
where a is the wavespeed in the pipe, Δx is the length of the section and
Δt is the time step. The length of a pipe section is just the length of the
pipe divided by the number of sections in that pipe:
L
Δx = (5.2)
n
The time step can be obtained by combining Equations 5.1 and 5.2:
L
Δt = (5.3)
na
The maximum time is obtained from the pipe with the smallest value of
L/a. This pipe will have one section (n = 1) and thus the time step will
be:
L
Δt max = (5.4)
a
min
In any pipe system there will be one pipe that is the controlling pipe.
The controlling pipe is that pipe which has the least number of sections,
frequently only one. Once the controlling pipe is identified, the time step
is determined by solving Equation 5.4 for Δt. Then the number of
sections in the remaining pipes is obtained from:
Li
ni = (5.5)
ai Δt
The Section Pipes window automatically determines which pipe is the
controlling pipe. It also provides searching tools that allow the user to
identify a sectioning strategy that has acceptable error levels.
Error is introduced at this point because of an approximation that is
usually required in multi-pipe systems. Because the number of sections
in pipes is derived from Equation 5.5, typically the number of sections,
n, will not be a whole number. Since partial sections in a pipe cannot be
modeled, an alternative must be found.
The wavespeed for each pipe is the least certain parameter. Therefore,
the wavespeed in each pipe is allowed to depart slightly from its original
calculated value in order to cause n to be a whole number. Introducing
this error term modifies Equation 5.5 to the following:
Li
ni =
(1 ± ψ )ai Δt
where ψ is the accepted uncertainty in wavespeed up to ±15% (Wylie,
et al., 1993, pp. 54).
Figure 5.14 shows the Section Pipes window. The controlling pipe (#8)
is shown at the right. The sectioning criteria are shown at the upper left,
where the average and maximum sectioning errors are shown. The table
at the bottom shows the selected sectioning strategy and resulting
modified wavespeed. The resulting time step size is shown at the right.
Numbering convention
Once a pipe is broken into sections, computation takes place where the
sections join together. These are called pipe stations. The total number
of stations in a pipe is equal to the number of sections plus 1. The station
at the beginning of the pipe is numbered zero. See Figure 5.15.
Computing
stations 0 1 2 3 4 5 6
Pipe sections 1 2 3 4 5 6
J1 J2 J3
P1 P2
P1 J2 P2
Figure 5.16 The pressure drop across the valve at J2 in the AFT
Impulse model at the top is given by the difference
between the solutions at Pipe 1, station 4 and Pipe 2,
station 0.
For junctions with more than one connecting pipe, the principle is the
same (see Figure 5.17).
J1 J3
P1 P2
J2
P3
J4
P2 2,0
1,0 P1 2,1
1,1 J2 2,2
1,2 2,3
2,4
1,3
3,0 P3
3,1
3,2
3,3
3,4
Figure 5.17 Pipe station convention for AFT Impulse model. Pipes
1, 2 and 3 have 3, 4 and 4 pipe sections, respectively.
Because there is no loss at the J2 branch, the
pressure solution for Pipe 1, computing station 3 (i.e.,
1,3) is the same as both Pipe 2, computing station 4
(i.e., 2,4) and Pipe 3 computing station 0 (i.e., 3,0). The
flow solutions at these three station will sum to zero
at all times (in the absence of vapor cavitation).
Transient Control
Figure 5.18 shows the Transient Control window. The first tab is an area
called the Transient Solver Control area. This area offers numerous input
fields detailed in the following descriptions.
Figure 5.19 Force set data is defined on the Forces tab in the
Transient Control window
Flow
y
PA
Pressure
PB
Distance
Area changes
In cases where force sets span across two pipes, and the pipes have a
different diameter, the area change itself cannot physically create an
unbalanced force even though the static pressure changes. AFT Impulse
will automatically account for this in the force balance calculation.
See the Examples Help for additional explanation of these issues.
Figure 5.21 This window exports data to file for direct use with
CAESAR II. It is accessed from the File menu in any of
the output windows.
Current
Workspace
scenario
Scenario
Tree
Rename,
delete, clone,
promote &
save
scenarios by
clicking here
Create a new
scenario by
clicking here
‘3 in PVC Pipe’, ‘3.5 in PVC Pipe’ and ‘4 in PVC Pipe’ were created by
first selecting the Base Scenario then clicking on Create Child. ‘North
hydrant open’ and ‘South hydrant open’ were, in turn, created by
selecting each of the above children scenarios for the three different pipe
run sizes, then clicking on Create Child (Note: alternatively, one could
have made the Base Scenario current, created the ‘3 in PVC Pipe’ child,
then selected it and created the ‘North hydrant open’ and ‘South hydrant
open’ scenarios below it, then, with ‘3 in PVC Pipe’ still selected, click
on Other Actions / Clone With Children twice, then rename the new
clones ‘3.5 in PVC Pipe’ and ‘4 in PVC Pipe’). The ‘North hydrant’ and
‘South hydrant’ scenarios may be thought of as grand children of the
Base Scenario. This concept of scenario ancestry is helpful in keeping
track of the differences and similarities between scenarios discussed
later.
A box with a plus sign next to a scenario indicates it may be expanded to
reveal children scenarios while a box with a minus sign indicates the
level is fully expanded with all children scenarios shown, if any. No box
indicates there are no children. Levels may be expanded by clicking on
the plus sign and condensed by clicking on the minus sign.
Diameter Length
Base 3 25
Child #1 __ __
GrandChild #1 __ __
Figure 5.24a Scenario Manager logic – Blank fields for Child #1 and
GrandChild #1 mean that the data is to come from the
parent. If the Base scenario data is changed, all
descendants are changed.
If a child scenario does not have a blank field, then data for that property
is initiated at that scenario level (see Figure 5.24b).
Diameter Length
Base 3 25
Child #1 2 __
GrandChild #1 __ __
If a scenario is changed, and its child has different data, then the change
will not pass downwards (see Figure 5.24c).
Diameter Length
Base 3 25
Child #1 2 __
GrandChild #1 __ __
Diameter Length
Base 6 40
Child #1 2 __
GrandChild #1 __ __
If a child scenario has data that is different than the parent, its children
cannot relink to the parent (see Figure 5.24d).
If a child scenario’s data, which was previously changed and is thus
different from the parent, is changed back to the same vale as the parent,
the inheritance link is re-established (see Figure 5.24e). Its descendant’s
link is also re-established.
Diameter Length
Base 3 25
Child #1 2 __
GrandChild #1 3 __
Diameter Length
Base 3 25
Child #1 3 __
GrandChild #1 3 __
Diameter Length
Base 3 25
Child #1 __ __
GrandChild #1 __ __
Diameter Length
Base 3 25
Child #1 2 __
GrandChild #1 __ 15
Waterhammer Assistant
The Waterhammer Assistant can be opened from the Help menu or
Toolbar. The Waterhammer Assistant reviews your model and offers
helpful advice (see Figure 5.25).
By default, whenever a model is run AFT Impulse checks with the
Waterhammer Assistant to see if it has any advice. If it does, the
Waterhammer Assistant is displayed. The user then has the option of
canceling the run or ignoring the advice and continuing. If you do not
want the Waterhammer Assistant to be consulted before each run, you
can specify this choice in the General Preferences window opened from
the Options menu.
Workspace Find
The Find window utility shown in Figure 5.26 helps you find a particular
pipe or junction by its ID number and name. The Find utility becomes
more useful for models that expand over multiple Workspace pages.
You can access the Find utility by selecting Find on the View menu or
by clicking the button on the Toolbar. The Workspace is moved so that
the pipe or junction is located near the center of the screen.
Figure 5.26 The Find feature for the Workspace is helpful for
locating a specific pipe or junction. It is opened from
the View menu or Toolbar.
Reverse Direction
Each pipe you create has a reference positive flow direction as indicated
by the arrow on the pipe. If you want to reverse the reference positive
flow direction, you can do so by manually moving the pipe endpoints.
However, this procedure is tedious. The Reverse Direction utility toggles
the reference positive flow direction for the selected pipe or pipes.
The Reverse Direction utility is accessed from the Arrange menu or from
the reverse direction button on the Toolbar.
Select Special
Select Special is a tool for selecting or deselecting objects based on
certain criteria (see Figure 5.27) and allows you to quickly select a group
of objects.
When first opened, Select Special will select all the objects currently
selected on the Workspace. You may alternatively select objects based
on several other criteria available from the Selection Type drop-down
menu. Objects in the Pipe Selections or Junction Selections list may be
selected or deselected based on the criteria by clicking the Select Pipes
or Select Junctions button or the Deselect Pipes or Deselect Junctions
button.
Invert will deselect all selected objects and select all unselected objects.
Clicking on the Select All or None buttons will result in all of the
objects being selected or deselected respectively.
Several select/deselect operations may be performed in succession.
When finished, click on OK button.
Tip: To quickly show only a part of the model in the Output window,
first select the objects on the Workspace for which you want to see
output. Then, on the Show Selected Pipes/Junctions tab in the Output
Control window, select the Workspace buttons in the pipe and junction
areas. Finally, click OK in the Output Control window and only the
pipes and junctions you selected on the Workspace will be shown in the
Output.
Special Conditions
Some junctions have Special Conditions that alter the normal state of the
junction. For example, a valve can be closed, a pump turned off, or a
relief valve opened.
The special conditions are set in one of three ways:
1. By selecting the junction(s) on the Workspace and choosing Special
Conditions from the Edit Menu
2. By selecting the junction(s) on the Workspace and clicking the
Special Condition icon on the Toolbar
3. By opening the junction’s Specifications window and clicking the
appropriate condition on the Optional tab
The junctions that have special conditions set are shown using a special
symbol before the ID number (an “X” by default). This symbol can be
customized in the Workspace Preferences window.
As shown in Figure 5.28, when the special condition results in a section
of the model being closed or “turned off”, the pipes which will have zero
steady-state flow are displayed as dotted lines. This indicates visually
that they are cut off from flow; the junctions are outlined with a dotted
line. The default dotted line can be changed in the Workspace
Preferences window.
Dashed lines
for pipes and
around
junctions
indicate closed
model sections
Merging models
Two models can be merged together using the Merge command on the
File menu. Here’s how:
• Open the first model
• Choose Merge on the File menu
• From the file list, select the model to be merged
If there are duplicate numbers, the objects in the second model will be
changed. The second model will be selected (highlighted) so that you
can move it into position (be careful not to click off the selected portions
or the second model may be deselected).
Print Preview/Special
The Print Preview/Special window shown in Figure 5.29 allows you to
access special print features such as Print Preview and Fit to One Page.
It is accessed from the File menu.
The Print Preview/Special window also offers you a preview of the print
content and format for all primary windows. By clicking the Print
Preview button you can review your printed material at several zoom
states. Individual pages can be printed.
Batch runs
AFT Impulse scenarios or models can be run in batch mode, a feature
that is especially useful for running a group of scenarios or models
during lunch or overnight.
Print Content
With either the Model Data or Output window active, you can open Print
Content from the View menu (see Figure 5.30), allowing you to specify
the content and format of Model Data and Output reports.
You can include Model Reference Information (entered in the Output
Control window) to document your model or make important comments
you want kept with the model.
The various tables in the General section, Pipe section, and Junction
section can be selectively printed. All Pipe and Junction Notes are
shown in the Model Data General section and can be printed at your
option.
Math calculator
For convenience, AFT Impulse offers you quick access to the Windows
Calculator; simply select Calculator from the View menu or press
CTRL+K in any specifications window.
Steady-state solver
The Steady-State Solver obtains a balanced steady-state solution to the
system being modeled. It employs a Newton-Raphson matrix method
that is iterative nature. An initial guess to the steady-state solution is
made, and the Newton-Raphson method progresses the guess towards a
final solution that agrees with the governing equations within some
tolerance. Details are given in Chapter 8.
Transient solver
The Transient Solver models the surge transients in a pipe system using
the Method of Characteristics (MOC). Unlike the iterative Newton-
Raphson matrix method used for steady-state solutions, the MOC is a
direct solution method that marches in time. Details are given in Chapter
9.
To function properly, the MOC requires an accurate steady-state solution
to provide the initial conditions. Thus when one runs a transient
simulation in AFT Impulse 4.0, the Steady-State Solver is always run
first. The steady-state solutions are then automatically passed on to the
Transient Solver, which uses them as the initial conditions for the MOC.
This will reduce artificial transient problems to a bare minimum.
Highlight feature
The highlight feature identifies required input data in the Specifications
window. You can toggle the highlight feature on and off in any of the
following ways:
• Double-click within the specifications window (outside of the
tabbed area)
• Press the F2 function key
• Choose Highlight in Pipe and junction Windows from the Options
menu
The highlighting feature may be especially useful when you are first
learning to use AFT Impulse or when you are having difficulty obtaining
a defined object status.
Jump feature
By clicking the Jump button in a Specifications window, you can move
directly to the Specifications window of a selected pipe or junction.
Pipe number
Every pipe must have an ID number. When you create a new pipe, AFT
Impulse assigns a default pipe number. The pipe number can be
changed; however, duplicate pipe ID numbers are not accepted, and ID
numbers must be greater than zero and up to 30,000. The pipe numbers
you choose have no bearing on the model connectivity, direction, or
layout. They are merely convenient identifiers.
Pipe name
A name can be assigned to each pipe for reference purposes. The default
name is simply “Pipe.” Names do not need to be unique. The name can
be shown on the Workspace, in the Visual Report and in the Output
window. By using names, attention can be called to important or critical
pipes.
Connected junctions
The connected junctions area shows you AFT Impulse's interpretation of
the junctions that are connected to the particular pipe. Every pipe must
be connected to two junctions in order to be completely defined. When a
junction does not yet exist at one of the pipe endpoints, “None” appears
in the area where the junction ID number would normally be displayed.
All pipes on the Workspace have a reference positive flow direction. The
reference positive direction is indicated on the Workspace by an arrow
on the pipe. Based on the reference direction, AFT Impulse identifies an
upstream junction and a downstream junction.
To determine the properties of the connected junctions, you can use the
Inspection utility. To inspect, position the mouse pointer over the
connected junction ID number and hold down the right mouse button.
Pipe Model
The Pipe Model tab (Figure 6.1a) allows you to input geometric data for
the pipe.
Pipe material
The Pipe Specifications window allows you to leave the pipe material
unspecified or to choose a pipe material from a list.
AFT Impulse provides default pipe material data for several different
kinds of pipe. In addition to these pipe materials, you may add your own.
Access to the Pipe Material Database is available on the Database menu.
By entering your own custom pipe materials, you can build a pipe
material database that is saved to disk and read in during startup. Once
you enter materials data, AFT Impulse treats the pipe materials as if they
were native to the program. More detail on building a custom pipe
material database is given in Chapter 7.
When you select a pipe material from the materials list, two things
happen:
1. In the Size area, the dropdown lists for nominal pipe Size and Type
are enabled
2. In the Friction Model area, a default friction data set and
accompanying handbook pipe roughness value is entered
The Type entry allows you to select the specific type, class, or schedule
of material for the nominal pipe size. Once the type is selected, the pipe
diameter is shown in the diameter field. This diameter cannot be edited.
In addition, if transient modeling is enabled and the user has selected
Calculated Wavespeed, the pipe wall thickness, modulus of elasticity,
and Poisson ratio are shown. All of these parameters come from the pipe
material database, and cannot be edited. These three parameters are
needed for wavespeed calculations.
Pipe diameter
The inner diameter of the pipe is entered in the Inner Diameter field in
the Pipe Specifications window. For pipes whose material is unspecified,
the diameter and units may be entered directly.
If the material is specified from the Pipe Material list, the diameter
cannot be edited; the value is derived from the database for the selected
material, size, and type.
The reduction in diameter due to scaling can be accounted for by
entering a percent reduction in the ID Reduction - Scaling field. Zero
percent would represent no reduction.
Friction Model
A common problem in specifying pipe roughness is that long term
operation of the pipe often results in deposits forming on the inside of
the pipe. This is commonly referred to as fouling. Fouling increases
roughness and, in severe instances, decreases the area available for fluid
transport (also causing greater pressure loss in the pipe). You are
encouraged to seek out applicable pipe design data when specifying pipe
roughness in your model.
The Friction Model area of the Pipe Specifications window offers eight
methods for specifying frictional models. Methods 1, 2, 4 and 5 all rely
on the Darcy-Weisbach friction factor method of calculating pipe
pressure drop.
1. Absolute roughness – AFT Impulse's default method is to specify
the roughness as an absolute average roughness height. Values of
pipe roughness can be found in many pipe handbooks or from
manufacturer's data.
Pipe Support
The Pipe Support selection is made from the provided drop down list to
obtain the constant c1. This constant is required for pipes in which the
wavespeed is calculated. AFT Impulse provides seven common pipe
support types.
The meaning of the c1 constant and its relationship to the seven support
types is given in Chapter 9.
Pipe Wavespeed
The wavespeed is the speed at which transient events propagate through
the pipe. Typical values for the wavespeed are 700-1600 m/sec (2000-
5000 ft/sec).
This parameter depends on the liquid acoustic velocity and the pipe
material and support. AFT Impulse will calculate the wavespeed for you,
or you can provide your own user specified wavespeed.
Figure 6.2 Fittings & Losses window opened from the Pipe
Specifications window
The Area Changes loss does not change the size of the pipe, it only puts
in a loss factor based on the pipe diameter and the area ratio of the loss.
When this option is selected, notice the illustration of how the loss is
defined, where Apipe is the area of the pipe and As is the area of the loss.
These area changes are useful to model fittings between a pipe and
another junction like a pump or control valve.
Design Alerts
A design alert (Figure 6.1c) allows you to specify certain maximum
and/or minimum parameters for the pipe. After running the Steady-State
or Transient Solvers, any exceeded value will be identified in the
Warnings section of the Output window and the specific cell containing
the value in the Pipes section will be highlighted. Design Alerts can be
cross-plotted in the Graph Results window vs. the actual results. One
good example where this is useful is for cross-plots of maximum and
minimum allowed operating pressures.
Optional input
There are several optional input parameters (Figure 6.1d). These are
described in the following sections.
Workspace display
You have the option of specifying what to display on the Workspace.
You can choose to display the pipe ID number, the pipe name, the pipe
nominal size, and the pipe schedule. This is helpful in large models to
Design Factor
You can specify two design factors for each pipe – one for the pipe
friction and one for fittings & losses that you add to the pipe.
Design factors allow you to specify multipliers on the pipe friction
calculations and fittings & losses that are applied by the Steady-State
and Transient Solvers during the solution process. Design factors are
helpful for adding safety margins to your design calculations.
Parallel Pipes
Here you can specify that the pipe is one of a number of parallel pipes.
The pipe properties thus represent a single pipe. This is useful, for
example, to represent a bundle of tubes in a heat exchanger.
Fluid Properties
This tab is only visible when variable properties are being modeled
(Figure 6.1e). Variable property modeling is enabled in the System
Properties window.
For models with variable fluid properties, you can assign different fluid
properties to each pipe in the model. This is convenient for modeling
systems that are subject to temperature variations and systems that have
multiple fluids. The pipe fluid properties are displayed in the Pipe
Specifications window, and can be modified by selecting the Fluid
Properties button. Besides varying the fluid properties, the viscosity
model can varied from pipe to pipe. Thus some pipes in the model can
be modeled as Newtonian, while others as non-Newtonian.
Notes
Each pipe can have notes associated with it. This is useful for listing
assumptions, reference documents, drawings, measurements, etc.
Status
The Status tab shows which required input data has not been entered.
Note: AFT Impulse does not prevent negative flow through junctions.
Should this occur, AFT Impulse will use the same loss factor or pressure
drop data referenced to the defined upstream pipe, just as in forward
flow. Because it is possible that the loss factor or pressure drop data
would not be the same for both flow directions, the flow solution could
be misleading. For this reason it is important that you either properly
define the pipe flow directions for junctions or pay close attention to the
flow directions that result. Should negative flow occur in any junction,
AFT Impulse will give a warning in the output.
Format #1:
Junctions with one or two connecting pipes
The first basic window format is for junction types that allow only one
or two connecting pipes (see Figure 6.3). While junctions do not have an
explicitly defined flow direction like pipes do, typically those with two
pipes or less adopt a direction from the connecting pipes. For these
junction types, upstream and downstream pipes are recognized by AFT
Impulse based on the reference positive flow direction of the connecting
pipes. The upstream and downstream pipes are displayed separately near
the top of the window.
In this first basic window type, it is generally important to have the
pipe’s reference positive flow directions specified in the physically
correct directions. A good example of this is a Pump junction which
will add pressure to the system in the direction of positive flow through
the connecting pipes. The Pump junction interprets where to add the
pressure based on the directions of the connecting pipes.
To inspect the connected pipe information in this window format,
position the mouse pointer over the connected pipe ID and hold down
the right mouse button.
Fixed display of
connecting pipes
Format #2:
Junctions with more than two connecting pipes
The second basic window format for junction Specifications windows is
for junctions that allow more than two connecting pipes (see Figure 6.4).
These junctions typically allow up to twenty-five connecting pipes; the
exception is the Tee/Wye junction, which allows only three.
Because the number of pipes connected to a junction may vary, the
second basic format uses a table sized according to the number of
connecting pipes.
An example of this second window format is the Branch Specifications
window, shown in Figure 6.4, which has three connecting pipes. To
determine AFT Impulse's interpretation of the model connectivity for
this second window format, you can review the contents of the table
areas. There are separate loss factors for flow going into the pipe from
the junction and flow going out of the pipe into the junction. The
Junction Number
Each junction has an ID number (see Figures 6.3 and 6.4) that you are
free to change. However, no two junction numbers can be the same, and
all junction numbers must be greater than zero and up to 30,000.
Duplicate junction numbers are not accepted. The junction numbers you
choose are arbitrary and have no bearing on model connectivity,
direction, or layout. They are merely convenient identifiers.
Junction Name
Each junction can be assigned a name for reference purposes (see
Figures 6.3 and 6.4). The default name is the junction type. The name
does not need to be unique. The name can be displayed on the
Workspace, Visual Report, Model Data and Output windows.
Junction Elevation
Each junction has an entry for the elevation of the junction (see Figures
6.3 and 6.4). Next to each elevation entry is a set of units. For junctions
that use Format #1, the default inlet elevation is entered by the user, and
the outlet elevation is assumed to the same. A different outlet elevation
can be entered by clearing the “Same as Inlet” checkbox. Different inlet
and outlet elevations will have no effect on the flow solution, but will
offset the local static pressure at the connecting pipe by the hydrostatic
pressure difference.
Junctions that follow Format #2 have a single elevation and offsets from
that elevation (if they exist) for each individual connecting pipe entered
in the connecting pipe table.
Junction elevations are used to account for pressure changes due to
gravity and thus to allow calculation of absolute pressures in pipes. In a
sense, junction elevations are arbitrarily defined. The magnitude of the
elevation is not important from a solution standpoint, but the relative
elevation difference between junctions is of critical importance because
of its effect on the pressure results.
Elevation is required input for every junction. AFT Impulse uses this
input to determine the relative height changes between one end of a pipe
and the other.
In the Parameter and Unit Preferences window you can set a default
elevation for your system. When a default elevation is specified, each
time you open a new junction Specifications window the default
Database List
The Database List (see Figures 6.3 and 6.4) is a list of all custom
equipment you have entered into the Component database. It displays
custom equipment from both the local and network database sources.
Selection of an item automatically retrieves the data and enters it into the
junction. Changes to the custom input data are not accepted. To stop
using a database selection, choose (None) from the database list.
Pipe connectivity
In each junction Specifications window (see Figures 6.3 and 6.4), you
can see AFT Impulse's interpretation of your system's connectivity. All
junctions must have the proper number of connecting pipes in order to
be completely defined (refer to Table 6.1).
Transient Data
Many junction types have an area where active transient data can be
specified. Table 6.2 lists the junction types.
The active transient data is entered on the Transient Data tab (Figure
6.5). Table 6.2 lists the type of data that can be entered for each junction
type. If a junction does not have any active transient behavior, there is no
need to enter any data on the Transient Data tab.
In most cases, transient data can be entered as absolute values or as a
percentage of steady-state. The distinction is made when choosing
between Absolute Values and Relative to Steady-State Value (Figure
6.5).
The first data point always needs to match the steady-state value which
is usually entered on the tab at the far left. If, for example in Figure 6.5,
the steady-state data for Cv of 1000 is changed to 750, the transient data
will also need to be changed. However, if in Figure 6.5 the Relative
option is specified, the first data point would be 100%, and when the Cv
is steady-state is changed from 1000 to 750 there is no need to change
the transient data since the first data point is still 100%. But it is now
relative to a different steady-state value.
Repeat Transient
If the transient data is periodic, you can enter the data for one cycle of
the period and then tick the check box for Repeat Transient. This will
cause the one cycle of transient data to be repeated once it has reached
the end. The repetition will continue until the end of the simulation.
Figure 6.6 The Optional tab offers several optional data inputs
and is very similar for each junction type
Special Conditions
Some junctions allow you to set Special Conditions (see example in
Figure 6.6). This feature allows you to override the default behavior of
the junction. For example, the special condition for a Valve junction is
to close the valve. The special condition for a Relief Valve is for it to be
open no matter what the pressure. Not all junction types support special
condition settings.
The junction for which the Special Condition is set will have an ‘X’
placed before the junction ID number. If the special condition is to close
the junction, by default it will be outlined with a dashed line. These
settings may be changed in the Workspace Preferences window.
Chapter 5 discusses how to apply transient data to junctions with steady-
state Special Conditions.
The Vacuum Breaker Valve, Gas Accumulator, Liquid Accumulator,
Surge Tank, and Relief Valve junctions have a unique kind of Special
Condition called Ignore. This makes the steady-state and transient
solvers ignore the presence of the junction. Thus not only is the data
ignored, the junction itself is ignored. This is convenient when locating
one of these devices by trial and error. Rather than having to delete the
junction, it can be ignored.
Design Factor
You can specify a design factor for each junction (see example in Figure
6.6). Design factors allow you to specify multipliers on the junction
pressure loss. These multipliers are applied by the Steady-State and
Transient Solvers during the solution process. In the case of a pump, the
design factor multiplies the pressure rise. Design factors are helpful for
adding safety margins to your design calculations.
Note: If you are designing for some minimum condition, you may want
to use the design factors differently for pumps than for other junction
types. If you use 1.1 for all junctions with a pressure loss, you will get an
extra 10% pressure drop. Using the same 1.1 on your pump will give an
extra 10% pressure rise. This will work against the 10% margin on your
pressure losses. If you want to use a pump design factor in this case, it
might be best to use 0.9.
Notes
Each junction can have associated notes. This is useful for listing
assumptions, reference documents, drawings, measurements, etc. Notes
are entered in the provided area on the Notes tab.
Status
The Status tab shows which required input data has not been entered.
Tip: If you see that the required pipe in the Specifications window states
“None” but a pipe is connected to the junction graphically, then check
the Inflow/Outflow setting on the junction or the reference positive flow
direction of the pipe.
The Assigned Flow junction type allows you to specify positive flow
rates as mass flow rates. Negative flow rates are not accepted for steady
flow, although are for transient flow. If, for example, you have an
outflow type junction and your system is physically flowing in, you
cannot assign a negative flow rate to the junction. You must reverse the
connecting pipe flow direction to be consistent with the actual direction
and change from an outflow type to an inflow type.
Transient Data
On the transient data tab a flowrate transient can be entered. A special
feature is the ability to model a sinusoidal flowrate transient. Input for
the amplitude and frequency is required. This transient is summed to that
specified in the transient data table. A chopped sine wave takes the
absolute value of the sine function.
More information on transient data, including event transients, is given
earlier in this chapter and in Chapter 10.
Special Conditions
You can set a Special Condition for an Assigned Flow junction, which
will turn the flow off and make it act like a Dead End during steady-
state.
The Assigned Pressure junction type has much in common with the
Reservoir junction. In each case you specify parameters in order to
achieve a known pressure. However, the Assigned Pressure junction
allows you to specify either stagnation or static conditions. This is useful
if, for example, you are modeling a system where the pressure in a pipe
is known and the location is being used as a boundary in the model. If
the measured conditions are for a location with a velocity, then they
represent static conditions and the choice of static properties should be
selected. In this case, the choice of a Reservoir junction would be
inappropriate. Similar to the Reservoir, an Assigned Pressure junction
causes the rest of the system to distribute the flow in a manner consistent
with the defined pressure. You can specify the units for pressure by
selecting from the adjacent drop-down list.
The distance of each connecting pipe (if any difference exists) from the
elevation can be entered in the pipe table.
Transient Data
On the transient data tab a pressure transient can be entered. A special
feature is the ability to model a sinusoidal pressure transient. Input for
the amplitude and frequency is required. This transient is summed to that
specified in the transient data table. A chopped sine wave takes the
absolute value of the sine function.
More information on transient data, including event transients, is given
earlier in this chapter and in Chapter 10.
For each connecting pipe, the pipe table lists the reference flow direction
and up to two loss factors. The first loss factor is for physical flow out of
the pipe and into the branch. The second is for physical flow into the
pipe and out of the branch. The loss factors can be specified
independently or left as zero. To edit within the pipe table, simply click
in the appropriate column and row.
The distance of each connecting pipe (if any difference exists) from the
elevation can be entered in the pipe table.
The Branch junction also offers you the option of specifying a flow
source or sink at the junction. According to AFT Impulse's convention, a
source of flow into the system is defined as positive, while a
consumption of flow with flow out of the system (a sink) is defined as
negative.
Transient Data
On the transient data tab a transient source or sink flowrate can be
entered. More information on transient data, including event transients,
is given earlier in this chapter and in Chapter 10.
A check valve is a device that allows flow in only one direction. AFT
Impulse assumes that the Check Valve is initially open. If the flow
solution indicates that forward flow will not occur, AFT Impulse closes
the valve. The valve can close during the steady-state or transient portion
of the simulation.
A check valve Special Condition can be either open or closed.
Specifying losses
You specify the losses for the Check Valve on the Valve Model tab in
the Specifications window. For convenience, you can specify the
constant loss characteristics of a Check Valve as a valve coefficient (CV)
Transient Data
The closing of a check valve is a complicated transient process because
the mechanical movement is coupled to the hydraulics of the pipe.
Sophisticated models exist in the literature which attempt to model this
complicated process, but these require extensive data about the valve
and typically more simplified models are employed.
AFT Impulse offers a more simplified model, whereby the user can
specify a closing profile transient of Cv or K vs. time. Remember that
when using the K loss model, a value of –1 (negative one) means the
valve is closed. The closing will occur as soon as the closing velocity
criteria is met (discussed above). If no closing transient profile is
entered, the valve will close instantaneously.
You can also specify an opening profile of Cv or K vs. time, which is the
profile the valve will follow if it reopens. If no opening transient profile
is entered, the valve will open instantaneously.
The zero time for the closing transient is relative to when the valve starts
to close because of backflow. For the opening transient, it is relative to
when the valve starts to open.
Special Conditions
Control valves have two special conditions. To open fully and not
control, or to close.
Transient Data
The Control Valve junction is the only one with two transient tabs. The
first is the valve reaction to inability to achieve control, and the second
is to the control point. More information on transient data, including
event transients, is given earlier in this chapter and in Chapter 10.
Failure Transient
Control valves have what might be called inherent event logic. That is,
rather than respond to events at some other location, the failure
transients respond to events of the valve itself. Specifically, the events
are the valve's loss of control and possibly regaining of control.
The failure transient is the valve action when it loses control. If the valve
is set to close upon failure, the closure transient will be used. This is a K
or Cv table vs. time. If the valve were to reopen, the opening transient
would be used. The time scale for the closing or opening of a valve is
relative to when the failure occurs. That is called time zero in a relative
sense.
You can elect to not enter data for the closing or opening transient. If the
valve loses control, the valve action is assumed to be instantaneous. If
the failure action is to fully open, then the closing transient data will
have no meaning and not be used. In a sense the opening transient will
have meaning, but in reality when the valve loses control it is already
fully open by that point, so no transient will occur. The valve will just
stay fully open.
The closing and opening transients are thus intended for use when the
valve is set to close upon failure to control.
In steady flow, dead ends have no flow and thus play no role in the
hydraulic solution to the system. However, in transient flow pipes that
terminate in dead ends can have flow. Thus dead ended lines can and do
PA
Orifice
Connector
m& A Pipe
LC
Pump Model
The basic definition of the pump is performed on the Pump Model tab.
You can model a pump with a pump curve or an assigned flow. When
Submerged pump
If a pump is submerged, there is no upstream pipe. Thus only one
connecting pipe is needed for the discharge. This is modeled by ticking
the check box and entering the pump suction conditions.
Pump configurations
Pump data can be entered for multiple configurations. This is available
only for pumps modeled as pump curves. The default is a single
configuration.
A pump configuration is a pump with a specific impeller trim and
operating speed. Multiple impeller trims and operating speeds can be
specified as part of the pump, and then a particular combination can be
chosen.
These parameters are set in the Auxiliary Input tab.
Multiple configuration
selections displayed
Transient Data
As discussed previously, pumps can be modeled as a pump curve or as
an assigned flow. Using assigned flow pumps in steady flow modeling is
useful for sizing a centrifugal pump. In AFT Impulse, however, that is
not appropriate during transient modeling. When modeling a centrifugal
pump transient, a pump with a pump curve is essential. Positive
displacement pumps act as constant flow devices, and the assigned
flowrate feature allows one to adequately model such pumps.
Figure 6.24 Pump with four quadrant data for pump trip modeling
Controller transient
If the pump has a variable speed controller (see Figure 6.25), a control
flow or pressure transient can be specified. Again, this transient can be
time-based or event-based (see Figure 6.26).
Figure 6.26 transient essentially increases the speed of the pump to
achieve 300 m3/hr of flow.
Note that the resulting pump speed transient with time can be saved to
the output file (see Transient Control window in Chapter 5). If saved, the
speed can be plotted vs. time.
Special Conditions
Pump junctions in AFT Impulse have two types of Special Conditions.
The first is to turn the pump off and have no flow through it. This is
called "Pump Off No Flow". The second type of Special Condition turns
the pump off but allows flow to go through the pump. This is called
"Pump Off With Flow Through".
Why would one want to use one Special Condition rather than the other?
The simple answer is that the first Special Condition is more appropriate
for positive displacement type pumps and the second for centrifugal
pumps.
Here's why. When a positive displacement pump is turned off and has a
pressure difference across it, it will usually act like a closed valve and
not allow flow to go through it. Thus the first Special Condition would
be most appropriate. For instance, assume one wants to model the
transient that occurs during the startup of a positive displacement pump.
The best way to do this would be to use the first Special Condition with
Viscosity corrections
Viscosity corrections to standard pump curves have been published in
the Hydraulic Institute Standards. The method, based on the 14th
edition, 1983, is employed. The method itself consists of a semi-
logarithmic diagram relating flow rate, head rise and viscosity. The
details of this diagram have been digitized into numerical relationships
inside AFT Impulse 4.0. The corrections to head rise, flow rate and
efficiency, referred to as CH, CQ and CE, respectively, are given in the
Pump Summary output.
The viscosity corrections are assumed to be valid during transients.
Specifying losses
You specify the losses for the Relief Valve on the Valve Model tab in
the Specifications window. For convenience, you can specify the
constant loss characteristics of a Relief Valve as a valve coefficient (CV),
as a loss factor (K), or as a variable Cv where Cv depends on pressure.
Chapter 8 details the relationship between K and CV.
If a pipe returns to the reservoir above the liquid surface, the depth is
entered as a negative number, or the elevation is specified higher than
that of the liquid surface. Pipes that discharge above the liquid surface
are assumed to have liquid free fall to the liquid surface. AFT Impulse
applies the proper boundary condition for pipes with negative depths. If
the pipe is above the liquid surface, the only appropriate condition is for
the fluid to be flowing from the pipe into the reservoir. Fluid cannot flow
from the reservoir into the pipe as the pipe is above the liquid surface. If
this occurs, AFT Impulse will assume the fluid flowing into the pipe is
the same as the reservoir fluid, solve the system, and then give a warning
in the output.
Transient Data
On the transient data tab the surface elevation can be changed with time.
More information on transient data, including event transients, is given
earlier in this chapter and in Chapter 10.
Loss Model
The discharge coefficient for the junction is entered in the appropriate
box. The physical area of the exit flow can also be entered. These two
parameters are used to calculate the discharge flow rate, which depends
on the difference between the internal pressure in the system and the
specified exit pressure.
The K Fire Sprinkler model uses the K value as defined by the fire
sprinkler industry. Whereas the standard K loss value is dimensionless,
the K Fire Sprinkler has units associated with it. See Chapter 8 for more
details.
Transient Data
On the transient data tab the discharge flow area or K data can be varied
with time to simulate the opening or closing of the junction. More
information on transient data, including event transients, is given earlier
in this chapter and in Chapter 10.
Tank geometry
The Tank Height is the physical top of the tank in relation to the
elevation. If there is a Connector Pipe, then the Tank Height is relative
to the top of the Connector Pipe. This height is used in the case of
overflow. If the surge tank overflows, spillage will occur and the liquid
will exit the pipe system not to be recovered. If you do not want to
model the overflow, specify a very tall tank or do not enter a Tank
Height.
If the Tank Cross-Sectional Area is specified as Constant, the surge tank
is assumed to be a vertical tank with constant cross-sectional area. If the
tank cross-sectional area changes with height, this can be entered by
selecting the variable option and entering the cross-sectional variation in
the table.
The Tank Cross-Sectional Area is used for the purpose of determining
the liquid volume. It will determine how high the liquid will rise or drop,
since a given volume of liquid will flow into the tank during a time step.
LL Orifice
Connector
m& T Pipe
LC
Transient Data
On the transient data tab the surface pressure can be varied with time to
simulate a tank pressurization. More information on transient data,
including event transients, is given earlier in this chapter and in Chapter
10.
LA
LL Orifice
Connector
m& T Pipe
LC
Loss factors
There are two Loss Models available for Tee/Wye junctions: Simple and
Detailed. If you choose the Simple type of Tee/Wye, no loss factors will
be calculated. This is equivalent to a Branch with no loss factors entered
for the connecting pipes. In some systems the losses associated with a
Tee/Wye are small compared to the overall loss in the system. To avoid
unnecessary calculations you should choose the Simple type of Tee/Wye
unless you need the actual losses.
The Detailed type of Tee/Wye will calculate the loss factors as they
depend on the flow split as well as the geometry. Because of the
dependence on the flow split, loss factors usually cannot be calculated
before performing an analysis. AFT Impulse incorporates the best
available models to calculate tee and wye loss factors during the actual
solution. These calculations account for the percentage of the flow split
Transient Data
On the transient data tab a transient Cv or K table can be entered to
simulate the opening or closing of the valve. For a closed valve, the K
factor is infinite. To account for this, enter a K factor of –1 (negative
one) in the transient data table. More information on transient data,
including event transients, is given earlier in this chapter and in Chapter
10.
Special Conditions
The Special Condition for a valve always closes it.
Figure 6.38 The Global Pipe Edit window allows you to change
data for many pipes at the same time. This is how the
window appears when first opened.
Figure 6.39 The Global Pipe Edit window after selecting the
“Select Pipe Data” button. Data can now be applied.
Choice of editing
common or
specific data
Figure 6.40 Global Junction Edit window. This is how the window
appears when first opened to edit Common data.
Figure 6.41 Common Data Edit window for global junction editing.
Figure 6.42 The Global Junction Edit window after selecting the
“Select Common Junction Data” button. Data can now
be applied.
Option for
specific types of
junctions.
Junction type
selected in list.
6. Enter the junction data you want to change and click the OK button
in the Spray Discharge Specifications window (or whatever window
type is displayed).
7. The Global Junction Edit window now lists the parameters to change
in a list on the right (see Figure 6.44). Select the parameters you
want to change.
8. Click the Apply Selections button.
Parameter Preferences
The first parameter you can set is the Use Most Recent Pipe Size Data
feature. When this is selected, AFT Impulse remembers your most recent
pipe material size setting and sets new or unspecified pipes to this pipe
material and size.
You can also set a default pipe material and default friction model. You
can change the material and loss model of each individual pipe at any
time, but this will be the first choice.
Unit Preferences
You can set the units you prefer to use in AFT Impulse on the Unit
Preferences tab (see Figure 7.2). These units are selected as a first choice
in the unit selection boxes. You can choose any of the units available. To
set a new default, choose the unit type from the list on the right and click
the Set As Preferred Unit button.
You also can customize the units used by AFT Impulse. For each
parameter, AFT Impulse can use many different units. However, if only
a few are ever typically used, then the others can be removed. This will
decrease the number of items in the unit selection boxes, making it easier
to choose the one desired. If a preferred unit is removed from those
available to Impulse, a new one will have to be chosen before leaving
this area. You also must have one set of units available for each
parameter.
To customize the units, first choose the parameter to modify on the list
on the left. Then, among the unit types on the right, check or uncheck the
boxes for the units you want or do not want to see.
You can select or unselect all displayed units by clicking the appropriate
button. Also, English and SI units can be selected as a group for
parameters by clicking the appropriate button.
Figure 7.2 Using the Unit Preference tab in the Parameter and
Unit Preferences window, you can set the default units
you want to work with and choose to show only the
units you want to work with.
Tip: To choose only English or SI units for all parameters, go to the Unit
Preferences tab, choose the Default Unit System then click the Apply
Default Units button. This will use AFT Impulse’s default units. This
will also change the preferred units.
Database connections
In the lower left of the Parameter and Unit Preferences window there are
two Database checkboxes. If these boxes are checked, your Parameter
and Unit Preferences parameters are set up as determined by the
database to which you are connected. This is referred to as an active
database. To make them inactive, uncheck the boxes or change one of
the Parameter and Unit Preferences settings controlled by the database.
You can specify database connections separately for parameters and
units.
If the checkboxes are unchecked, but enabled, you are connected to a
database but the settings are not being passed to the Parameter and Unit
Preferences window. The databases are thus inactive. To make them
active, check the boxes then click the OK button.
If the checkboxes are disabled, there are no connected databases.
Later in this chapter we’ll discuss how databases are configured and
administered through your local or wide area network.
Command buttons
There are eight buttons at the bottom of the Parameter and Unit
Preferences window. Impulse has built-in default parameters, units and
settings which you can choose by clicking the Impulse Default button.
You can also develop your own settings, tailored to your project or
industry, and have these used by default (instead of Impulse’s defaults).
To make your own default, first select the parameters, units and settings
you would like to use, then click the Set As Default button. Your settings
will be saved and will be used each time any new project is initiated. If
you make changes to the settings, and want to get back to your defaults,
click the User Default button. The default settings are updated only
when you click Set As Default.
You can save the parameter and unit settings to a file by pressing the
Save Preferences button and entering a file name. These setting are
loaded again by pressing the Load Preferences button and choosing the
file name. The format files you create can be placed on a network for
Workspace Preferences
The Workspace Preferences window is accessed from the Options menu.
The features offered in this window allow you to customize the
appearance and behavior of the Workspace.
By varying the pipe thicknesses you can easily see the feeding pipes, the
large flow area pipes, or the small supporting flow pipes.
Each individual pipe allows you to modify its thickness, which overrides
the default size set here.
Display Options
From the Display Options tab you can change the symbols used, text
content, and location of the text on the Workspace (see Figure 7.4).
Workspace grid
A grid can be displayed on the background for modeling convenience.
The grid can be fine, medium or coarse. The grid lines can be solid or
dashed. And finally, the pipes and junctions can be snapped to the grid
you specify.
The grid can be included in printouts by specifying its inclusion in the
Print Preview/Special window.
Workspace symbols
Special symbols can be displayed on the Workspace adjacent to pipes
and junctions to denote special behavior. For example, an X (or symbol
of your choice) can represent a Special Condition. If you want to use
Popup menu
The popup menu can be disabled (it is enabled by default) here. The
popup menu is a context sensitive menu available on the Workspace.
Sample Workspace
The tab for Sample Workspace shows you a sample of how the
Workspace pipes and junctions will appear if you click the OK button.
Database connections
In the lower left of the Workspace Preferences window there is a
Database checkbox. If this box is checked, your Workspace Preferences
parameters are set up as determined by the database to which you are
connected. This is referred to as an active database. To make it inactive,
uncheck the box or change one of the Workspace Preferences settings
controlled by the database.
Command buttons
There are eight buttons at the bottom of the Workspace Preferences
window. Impulse has built-in default parameters, units and settings
which you can choose by clicking the Impulse Default button. You also
can develop your own settings, tailored to your project or industry, and
Toolbox Preferences
The Toolbox Preferences window is accessed through the Options menu.
Figure 7.6 shows the layout of the Toolbox Preferences window. The
features in this window allow you to customize the Toolbox in the
manner that is most productive for you.
Once you have changed the appearance and functionality of the Toolbox,
you can set your changes as the user default values by clicking the Set
As Default button. From then on, AFT Impulse will always load with
your default Toolbox settings.
You can also reorder the icons on the Toolbox by dragging and dropping
them to the desired locations on the Current Toolbox.
Toolbox shortcuts
You can show or hide Toolbox shortcut information.
Database connections
In the lower left of the Toolbox Preferences window there is a Database
checkbox. If this box is checked, your Toolbox Preferences parameters
are set up as determined by the database to which you are connected.
This is referred to as an active database. To make it inactive, uncheck
the box or change one of the Toolbox Preferences settings controlled by
the database.
If the checkbox is unchecked, but enabled, you are connected to a
database but the settings are not being passed to the Toolbox Preferences
window. The database is thus inactive. To make it active, check the box
then click the OK button.
If the checkbox is disabled, there is no connected database.
Later in Chapter 7 a lengthy discussion is given of how databases are
configured and administered through your local or wide area network.
General Preferences
The General Preferences window (Figure 7.7) allows you to:
• Specify automatic model saving
• Options for opening models with scenarios
• What content to display on the Status Bar
• Require a check of the Waterhammer Assistant before running
models
• Enable Tip of the Day feature
• Number of data points to show in input tables
• Whether or not to confirm before clearing output results
• Setting parameters for “button flashing” for curve fit buttons on
junction Specifications windows
• Default locations for saving models
• Common database locations
Figure 7.8 Fittings & Losses Edit window allows you to and edit
fitting and loss data that is used in pipes.
Once you have entered a custom fluid, it will appear in the Fluid List
area of the System Properties window whenever you use AFT Impulse.
You can add new fluids to the database by clicking the Add New Fluid
button and opening a new window (see Figure 7.10). The constants
provided allow you to describe the variation of density, dynamic
viscosity, bulk modulus of elasticity, and vapor pressure as they depend
on temperature or solids concentration for newtonian fluids. In addition,
there is an option to describe the variation of density, viscosity, rigidity
coefficient and yield stress for a Bingham Plastic non-newtonian fluid as
it depends on solids concentration. This window provides the tools to
perform polynomial curve fits of fluid data; or, if you already have
appropriate constants, you can input them directly. In addition, the raw
data can be interpolated
Density, viscosity and bulk modulus are required for all fluids. Vapor
pressure is required if cavitation modeling is desired.
The Fluid Database window also allows you to open the Change Fluid
Data window, which is essentially the same window as the Add New
Fluid window. For easy editing, data for the fluid you wish to change is
entered for you in the appropriate areas. The original curve fit data you
entered is also displayed.
The Add New Fluid/Change Fluid Data windows displays cross-plotting
of your custom curves against the curve fit data points. You can also
import data from text files into these windows.
After you have entered a custom material, each time you open a Pipe
Specifications window the material will be available to you for quick
selection. This minimizes the need to work with handbooks.
Pipe materials typically have a range of nominal sizes and a specific
type, class, or schedule that describes the actual dimension. AFT
Impulse works with the inner diameter of the pipe, since that is the
parameter that directly affects the fluid mechanics.
You can also add multiple Friction Data Sets for each pipe material.
This allows you to keep, for example, different friction values that may
apply to the same pipe under different conditions. These conditions
could be related to the age of the pipe or the type of fluid being carried.
¾ To add custom pipe materials to the user database, choose the entry
location you desire in the list on the left (Pipe Materials, Sizes and
Database Manager
Accessed from the Database menu, Database Manager allows you to –
• Connect to and disconnect from databases
• Create and delete databases
• Review the content of databases
• Move contents between databases
Editing databases
The Edit Database tab of Database Manager allows you to create and
delete databases, and edit their contents by copying sections or items
from one database to another. The tab is divided into a Source Database
area on the left and a Changing Database area on the right (see Figure
7.15).
Figure 7.15 Edit Database tab allows you to select sections from a
specified database and add them to an existing or new
database.
The file and descriptive name of the selected database are displayed as
well as the sections available in that database, i.e. fluids, components or
pipe materials.
Database
connection
Overview
Users of AFT Impulse on a network can access common databases to
simplify model verification, improve analysis and reporting consistency,
and eliminate redundant work.
To accomplish this there is some initial setup that must be done. First,
the data for components, materials, formats, etc., needs to be created and
saved to various data files. Secondly, these files must be made available
to the users of AFT Impulse by creating a file named DATABASE.LIB
in the IMPULSE folder. Thirdly, the users need to connect to the new
databases listed and select the sections to use. Each of these broad steps
is discussed below. To maintain control and integrity of the data, it
would be advised that one person is given the responsibility to create and
IMP_USER4.DAT
AFT Impulse Local User Database
VailMaterials.dat
Development Output Control Settings
Development Output.dat
My VR Format.dat
My Custom Visual Report Format
This discussion will focus on the first and second methods. Let’s assume
that the person responsible for the network database has already created
a new pipe material. Also, assume that the goal is to create a network-
wide database to contain this material, as well as two Output Control
settings and a Visual Report format.
The information for the material is copied from the Local User Database
(IMP_USER4.DAT) to a new file using the Database Editor in the
Database Manager (see Figure 7.15). For the source database choose
Local User Database and then select Pipe Materials. Then click Create
New Database and enter a file name, VailMaterials.dat, for this example.
Then enter a description of the database. This description is how AFT
Impulse users will refer to the database and should concisely describe
what is in the database (for this example we’ll use “Vail Resort
Materials”). Then click the Copy button to copy the pipe materials from
∑ m&
j =1
ij =0 (8.2)
where ΔPf is the frictional pressure loss. The total pressure change
between junctions is given by the momentum equation in the form of the
Bernoulli equation:
1 1
P1 + ρV12 + ρgz1 = P2 + ρV2 2 + ρgz2 + ΔPf (8.4)
2 2
The static and stagnation pressure are related as follows:
1
Po = P + ρV 2 (8.5)
2
Substituting Equation 8.5 into 8.4
Po,1 + ρgz1 = Po,2 + ρgz2 + ΔPf
Solving for the frictional pressure drop for a constant area pipes yields:
Po,i − Po, j + ρg ( zi − z j ) = ΔPf (8.6)
⎛ f ij Lij ⎞ 1
Rij′ = ⎜ + K ij ⎟ (8.9)
⎜ Dij ⎟ 2 ρA 2
⎝ ⎠ ij
Note that R ′ is not the same as R, which is the resistance used in the
Output Control window and is expressed by the following equation:
ΔH = RQ 2
n 1/ 2
⎛ Po,i − Po, j + ρg ( zi − z j ) ⎞
∑ ⎜
⎜ R′ij
⎟
⎟
=0 (8.10)
j =1 ⎝ ⎠
where n is the number of pipes connected to junction i.
To be completely general, Equation 8.10 should be written for junction i:
n 1/ 2
⎛ Po,i − Po, j + ρg ( zi − z j ) ⎞
∑ ⎜
⎜ R′ij
⎟
⎟
= m& i, Applied (8.11)
j =1 ⎝ ⎠
to allow for application of boundary condition flow rates to a boundary
junction node.
Equation 8.11 as applied to each junction in the network represents the
system of equations that need to be solved to determine the stagnation
pressure at each junction. To solve this system, the Newton-Raphson
method is employed. In the Newton-Raphson method, new values for
each unknown are calculated based on the previous value and a
correction that uses the first derivative of the function.
In this instance the function would be of the form:
n 1/ 2
⎛ Po,i − Po, j + ρg ( zi − z j ) ⎞
Fi = ∑ ⎜
⎜ R′ij
⎟
⎟
− m& i, Applied (8.12)
j =1 ⎝ ⎠
The method involves finding all the junction stagnation pressures, Po,i ,
that cause all of the Fi to go to zero, thus satisfying Equation 8.11 at all
junctions.
When applied to a system of equations, the Jacobian matrix contains all
the required derivative information to employ the Newton-Raphson
technique. The Jacobian, JF, is given by:
2 3 4
Consider the reservoir in Figure 8.1. When the liquid in a reservoir flows
into connecting pipes, the static pressure immediately drops due to the
increase in velocity (see Figure 8.2).
Po,2
P2
ΔPloss
1
ρV2 2
Pressure
Po,4
2 1
ρV42 P4
2
J4 J5
P3
J1 J2 J3
P1 P2
P4
Now consider the systems in Figure 8.4. The first system appears to be
closed, while the second appears to be open. If the same boundary
condition (i.e., surface elevation and pressure) is used for J1, J11 and
J12 in the second system, to AFT Impulse it will appear as an identical
system to the first system. The reason is that AFT Impulse takes the first
system and applies the J1 reservoir pressure as a boundary condition to
pipes P4, P9 and P10. The second system uses three reservoirs to apply
boundary conditions to P4, P9 and P10. But if the reservoirs all have the
same elevation and pressure, the boundary conditions are the same as J1
in the first system. Thus the same boundary condition is used for P4, P9
and P10 in both models, and they appear identical to AFT Impulse.
J8 J9 J6
P8 P7
P9 P6
J1 J5 J3
P4 P3
P10 P11
J4 J2 J10
P1 P2
J11 J8 J9 J6
P9 P8 P7
J12
P6
P4
J1 J5 J3
P3
P10 P11
J4 J2 J10
P1 P2
Figure 8.4 The first system is closed, and the second open. In
both systems the flow into P10 is the sum of P4 and
P9. If the second system has the same conditions at
J1, J11 and J12, the two system will appear identical
to AFT Impulse.
But how is the flow balanced at J1 in the first system? Looking at the
first system, one sees that to obtain a system balance, whatever flows
into P10 must come back through P4 and P9. Because there is overall
system balance by the Steady-State Solver, it will give the appearance of
a balanced flow at the pressure junction J1. If there is only one boundary
(i.e., junction) where flow can enter or leave the pipe system, then no
Roughness-based methods
Roughness-based methods use a pipe roughness value to calculate the
pressure drop. This can be in the form of an absolute roughness (which
has units of length) or a relative roughness that ratios the pipe roughness
to its diameter.
Laminar flow
For laminar flow AFT Impulse uses the standard laminar equation:
64
f = Re < 2300 (default) (8.14)
Re
Note that the laminar friction factor does not depend on the user
roughness.
The System Properties window allows you to change the default laminar
transition Reynolds Number (see Chapter 5).
Turbulent flow
The Colebrook-White iterative friction factor equation is used to obtain
friction factors in the turbulent flow regime.
−2
⎡ ⎛ε 9.35 ⎞⎤
f = ⎢1.14 − 2 log⎜ + ⎟⎥ Re > 4000 (default)
⎜ D Re f ⎟
⎣⎢ ⎝ ⎠⎦⎥
The System Properties window allows you to change the default
turbulent transition Reynolds Number (see Chapter 5).
Transition flow
For Reynolds Numbers in the transition flow regime, a linear
interpolation is used between the laminar and turbulent values.
Hydraulically smooth
The friction factor in the turbulent range is calculated using Colebrook-
White with roughness = 0 while in the laminar range it continues to be
calculated using Equation 8.14 above.
Hazen-Williams method
AFT Impulse also offers the Hazen-Williams method of specifying
irrecoverable loss information. The Steady-State Solver converts the
Hazen-Williams factor to a Darcy-Weisbach friction factor (Walski
1984, 37). This allows a consistent solution approach to be used for all
pipe system models, while retaining the flexibility of the two approaches
to account for losses.
17.25
CHW =
f (VD) 0.081
0.54
Resistance
The pipe resistance relates head loss to volumetric flow rate. In equation
form, the head loss is:
ΔH = RQ 2
Using standard relationships, the friction pressure drop is related to
volumetric and mass flow rate as follows:
Rgm& 2
ΔPf = RρgQ 2 =
ρ
dP 139.212 fQ 2 ( ρ ρ water )
=
L D5
1
f = 2
⎡ ⎛ Re ⎞⎤
4 ⎢0.86859 ln⎜ ⎟⎥
⎣ ⎝ 1964
. ln( Re) − 38215
. ⎠⎦
10,059Qν
Re =
(
d ν 2 − 818.56 )
where:
dP = pressure drop (psid)
L = length (miles)
f = friction factor, MIT
Q = volumetric flow rate (barrels/hour)
ρ = density (lbm/ft3)
ρwater = density of water (62.3 lbm/ft3)
D = diameter (inches)
Re = Reynolds number, MIT
ν = kinematic viscosity (Seconds Saybolt Universal)
d 2.5 ( dP L) ⎡ ⎛ d 3(ρ ρ ⎤
water )( dP L) ⎟
⎞
Q= ⎢log⎜ + 4.35⎥
⎜ ⎟
5.9113( ρ ρ water ) ⎢⎣ ⎝
0.5
μ2 ⎠ ⎥⎦
where:
dP = pressure drop (psid)
L = length (miles)
Q = volumetric flow rate (barrels/hour)
ρ = density (lbm/ft3)
ρwater = density of water (62.3 lbm/ft3)
Frictionless pipes
Frictionless pipes are convenient for connecting junctions that may not
have a physical pipe between them. There are, however, some
limitations where a frictionless pipe may be located in a model. An
example is a frictionless pipe that connects two Assigned Pressure
junctions. Such a pipe would have an infinite flow rate. In such cases
AFT Impulse identifies inappropriate placement of frictionless pipes and
informs the user when a model is run.
Vmax = K ′C σ
Vwater = 122
. C 1.4
If V < Vmax, then
dH
= FKV α C β D γ
L
where:
dH /L = head loss per length (m water / 100 m pipe)
C = pulp consistency (% dryness)
Vmax = max velocity (for Duffy Equation) in above equation
(m/s)
Vwater = water velocity in above equation (limit at which pulp
effects are negligible), (m/s)
D = inner pipe diameter (mm)
D 0.205Vρ
Re′ =
C1.157
3.97
f =
Re′1.636
fV 2 LK
dH =
D
where:
Re’ = pseudo Reynolds Number (dimensionless)
D = inner diameter (feet)
V = average velocity (ft/s)
ρ = stock density (lbm/ft3)
C = pulp consistency (% dryness)
f = pseudo friction factor (dimensionless)
K = constant depending on pulp type (dimensionless)
L = length (feet)
0.2728n −1 2
fT =
Re pl1 (1.87 + 2.39n )
1
α=
1 + 4−Δ
Δ = Re pl − Re plc
8 D nV (2 − n )ρ
Re pl =
K (2(3n + 1) n )n
(
f = f L m + fT m )1 m
40000
m = 1 .7 +
Re
where:
64 ⎛ He He 4 ⎞
fL = ⎜ 1+ − ⎟⎟
Re ⎜⎝ 6 Re 3 f 3 Re 7 ⎠
DVρ
Re =
Rc
He is the Hedstrom number:
D 2 ρS y
He =
Rc 2
10 a
fT =
Re 0.193
( (
a = −5.88 1 + 0.146 exp − 2.9 × 10−5 He ))
Design factors
In each pipe you can specify a Design Factor for the pipe friction. This is
a multiplier that is applied to the friction factor calculated with the
preceding methods.
Pj , new − Pj , old
If (For All Junctions) < TOLrel Then
Pj , new
Convergence = True
Else
Convergence = False
End If
Table 8.1 shows a tolerance example for pressures. Notice how relative
change does not have units, but absolute change has units (of psia).
The default in AFT Impulse is relative tolerance because experience
indicates it is the most robust.
AFT Impulse 4.0 offers options for combining absolute and relative
tolerance. You can tell AFT Impulse to assess convergence based on
whether either an absolute or relative criteria is satisfied, or whether
both absolute and relative criteria are satisfied.
Figure 8.6 shows the tolerance input area in the Steady Solution Control
window. Table 8.2 shows an example of how the four different criteria
would be applied to flow rate iterations.
Relaxation
Another feature of iterative solvers is called relaxation. In short,
relaxation slows the Steady-State Solver's progress toward a solution and
in the process reduces the ability of poorly behaved (non-linear)
components to destabilize the solution. AFT Impulse implements a
proprietary adaptive relaxation scheme that reduces the relaxation
dynamically in response to highly non-linear features of the problem.
However, you are free to override AFT Impulse's adaptive approach.
When applied to the flow solution, the relaxation is used as follows:
There are two relaxation parameters in AFT Impulse 4.0 (See Figure
8.7): Flow rate relaxation and pressure relaxation.
Pressure Relaxation
Because of the nature of the matrix solver, experience indicates that if
pressure relaxation is used, it should always be set to 0.5 or 1.0.
Anything else will usually destabilize the solution. A pressure relaxation
value of 0.5 usually works best.
Maximum Iterations
The Maximum Iterations parameter in the Steady Solution Control
window determines when the Steady-State Solver considers the model
nonconvergent; it does not directly affect the solution. There may be
many reasons why a model does not converge. You want to be sure that
the Steady-State Solver does not continue indefinitely searching for a
solution that cannot be obtained because of an input error or ill-behaved
system.
Most properly defined models will converge in 50,000 iterations. For
larger models, more than 50,000 iterations may occasionally be required,
especially when relaxation is being used. AFT Impulse's default of
50,000 iterations is sufficient for most models.
Matrix Method
AFT Impulse’s network solution method requires solution of matrices
thousands of times. The default method is Gaussian Elimination, which
is very robust and usually provides the fastest convergence.
For larger systems, Gaussian Elimination with Pivoting and the LU
Decomposition methods can be better and are therefore provided.
Figure 8.8 The Solution Progress window shows you the status
of the Steady-State and Transient Solvers. This
includes showing how far out of tolerance the Steady-
State Solver is, and when the steady solution has
converged.
Design factors
Each junction that allows modeling of pressure losses allows input of a
Design Factor. The design factor is multiplied by the K factor
determined by methods discussed in this chapter.
Many junction types allow modeling of K factors as well as other
pressure loss methods. The Design Factor is also applied to these other
methods as a multiplier on the pressure loss.
Area change
Two standard area change loss geometries are available: the conical
transition and the abrupt transition.
The conical expansion correlation (Crane 1988, A-26) is:
2
⎛ μ ⎞⎛ Aup ⎞
K up = 2.6⎜ sin ⎟⎜⎜1 − ⎟
⎟ (μ < 45 degrees)
⎝ 2 ⎠⎝ Adown ⎠
2
⎛ Aup ⎞
K up = ⎜⎜1 − ⎟
⎟ (μ > 45 degrees)
⎝ Adown ⎠
The conical contraction correlation (Crane 1988, A-26) is:
⎛ μ ⎞⎛ A ⎞
0.8⎜ sin ⎟⎜1 − down ⎟
⎝ 2 ⎠⎜⎝ Aup ⎟
⎠
K up = (μ < 45 degrees)
2
⎛ Adown ⎞
⎜ ⎟
⎜ Aup ⎟
⎝ ⎠
⎛ A ⎞ μ
0.5⎜1 − down ⎟ sin
⎜ Aup ⎟⎠ 2
⎝
K up = (μ > 45 degrees)
2
⎛ Adown ⎞
⎜ ⎟
⎜ Aup ⎟
⎝ ⎠
The abrupt expansion correlation (Idelchik 1994, 208) is:
Bend
90 degree bends
The three Bend loss correlations are all for turbulent Reynolds Numbers
(Crane 1988, A-29).
The K values for a smooth, flanged bend are provided in Table 8.5.
r/d K r/d K
1 20 f T 8 24 f T
1.5 14 f T 10 30 f T
2 12 f T 12 34 f T
3 12 f T 14 38 f T
4 14 f T 16 42 f T
6 17 f T 20 50 f T
⎛ r ⎞
K = (n − 1)⎜ 0.25πf T + 0 .5 K ⎟ + K
⎝ D ⎠
where n is the number of 90 degree bends, K is the loss factor for one 90
degree bend (Table 8.5), and fT is given by Table 8.6.
The Mitre bend is given by Table 8.7.
μ (deg.) K
0 2 fT
15 4 fT
30 8 fT
μ
45 15 f T
60 25 f T
75 40 f T
90 60 f T
Valve
Standard valve loss models are used from Idelchik, Miller and Crane.
Orifice
The orifice loss factors are all for turbulent Reynolds Numbers. The
sharp-edged orifice shown in Figure 8.9, is given by the following
(Idelchik 1994, 218):
2⎡ 2
⎛ Aup ⎞ ⎛ ⎞
0.375 ⎤
⎢0.707⎜1 − Aorifice ⎟ ⎛ Aorifice ⎞⎥
K up = ⎜ ⎟ + ⎜⎜1 − ⎟
⎜ Aorifice ⎟
⎝ ⎠
⎢ ⎜
⎝ Aup ⎟⎠ ⎝ Adown ⎟⎠⎥
⎢⎣ ⎥⎦
2⎡ 0.75 2
⎛ Aup ⎞ ⎛ ⎞ ⎛ Aorifice ⎞
K up = ⎜ ⎟ ⎢ K ′⎜1 − Aorifice ⎟ + ⎜⎜1 − ⎟
⎜ Aorifice ⎟
⎝ ⎠
⎢ ⎜ Aup ⎟⎠ ⎝ Adown ⎟⎠
⎣⎢ ⎝
0.375 ⎤
⎛ Aorifice ⎞ ⎛ A ⎞ ⎥
+2⎜⎜1 − ⎟ K ′⎜1 − orifice ⎟
Adown ⎟⎠ ⎜ Aup ⎟⎠ ⎥
⎝ ⎝ ⎥⎦
where:
K ′ = 0.03 + 0.47 • 10
(− 7.7r / Dorifice )
Tee/Wye
The loss factors calculated for tee and wye junctions involve
complicated correlations that depend on the flow split, the ratio of flow
areas, and the angle of the connecting pipes. The models used by AFT
Impulse are taken mostly from chapter 7 of Idelchik's Handbook of
Hydraulic Resistance (1994). A few of the more important equations
that AFT Impulse uses will be presented here, and references will be
given for the others.
Ac Ast Ast Ac
As As
Diverging Case
Side branch (Idelchik 1994, 418):
ΔPs ⎡ ⎛ V ⎞2 ⎛ Vs ⎞ ⎤
K c, s = ⎢ s
= λ 1 + ⎜⎜ ⎟⎟ − 2⎜⎜ ⎟⎟ cosα ⎥
2 ⎢ ⎝ Vc ⎠ ⎥
ρVc / 2 ⎣ ⎝ Vc ⎠ ⎦
As / Ac Qs / Qc λ
≤ 0.35 ≤ 0.4 1.1 - 0.7 Qs / Qc
≤ 0.35 > 0.4 0.85
> 0.35 ≤ 0.6 1.0 - 0.6 Qs / Qc
> 0.35 > 0.6 0.6
As / Ac Qs / Qc τ st
≤ 0.4 0 - 1.0 0.4
> 0.4 ≤ 0.5 2 (2 Qs / Qc - 1)
> 0.4 > 0.5 0.3 (2 Qs / Qc - 1)
Converging Case
Side branch (Idelchik 1994, 417):
ΔPs
K c, s =
ρVc2 / 2
⎛ ⎛ Q A ⎞2 A ⎛ Qs ⎞
2
Ac ⎛ Qs ⎞
2 ⎞
= A′⎜⎜1 + ⎜⎜ s c ⎟⎟ − 2 c ⎜⎜1 − ⎟⎟ − 2 ⎜⎜ ⎟⎟ cosα ⎟
⎟⎟
⎜ ⎝ Qc As ⎠ Ast ⎝ Qc ⎠ As ⎝ Qc ⎠
⎝ ⎠
A2
Kvalve = 1460 2
Cv
Note: The fire sprinkler K is not the same as the dimensionless K value
as given by Equation 8.18 and used extensively throughout AFT
Impulse.
ρ water
CD A = K sprinkler
2
Pumps
Pumps can operate at variable speeds. Affinity laws (also called
homologous laws) allow an estimation of the pump performance at
speeds other than the design speed.
Before discussing the affinity laws, it should be noted that the affinity
laws are an approximation. In many cases pump manufacturers will have
performance data for a number of speeds. These data should be used if
available. When not available, the affinity laws can be used with an
understanding that an approximation is being made.
The affinity laws for pumps are as follows:
2
ΔH1 ⎛ s1 ⎞
=⎜ ⎟
ΔH 2 ⎜⎝ s2 ⎟⎠
(8.19)
Q1 s1
=
Q2 s2
where s is the speed. If the pump data is input as a polynomial, the speed
will affect the curve as follows:
2 3 4
ΔH1 = a + bQ1 + cQ1 + dQ1 + eQ1
2 3 4
ΔH 2 = s 2 ΔH1 = s 2 a + s 2bQ1 + s 2 cQ1 + s 2 dQ1 + s 2eQ1
2 3 4
Q2 ⎛Q ⎞ ⎛Q ⎞ ⎛Q ⎞
ΔH 2 = s 2 a + s 2b + s 2 c⎜ 2 ⎟ + s 2 d ⎜ 2 ⎟ + s 2 e⎜ 3 ⎟
s ⎝ s ⎠ ⎝ s ⎠ ⎝ s ⎠
3 4
Q 2 Q
ΔH 2 = s a + sbQ2 + cQ2 + d 2 + e 23
2
s s
(8.20)
where ΔH1 is the input curve that typically represents 100% speed and
ΔH2 is the head rise at speed s. In this form, the speed s will be a decimal
and not a percent. That is, 50% speed will result in the appropriate s of
0.5. AFT Impulse determines this automatically in the Steady-State and
Transient Solvers.
If you model the pump as a controlled flow or discharge pressure, AFT
Impulse will backsolve Equation 8.20 for the speed s and include that in
the output Pump Summary.
Accuracy options
When performing interpolations for properties, a 2 point or 4 point
interpolation scheme can be used. The 2 point scheme implemented in
AFT Impulse is the Standard Accuracy option. The 4 point scheme is the
high accuracy option.
Viscosity
The dynamic viscosity in AFT Impulse is assumed to be a function of
temperature only. In general this is a good assumption, although systems
at very high pressure usually start to show some pressure dependence.
These properties are modeled as polynomial curve fits of temperature.
Description of waterhammer
Please see Chapter 3 for a summary description of waterhammer in a
simple one-pipe system.
Instantaneous waterhammer
With few exceptions (discussed at the end of the chapter), it is safe to
calculate the maximum possible waterhammer pressure surge by using
the instantaneous waterhammer equation. The instantaneous
waterhammer equation assumes that the transient event occurs either
instantaneously or rapidly enough such that it is in effect instantaneous.
In such a case, it can be shown (Wylie, et al., 1993, pp. 4) by use of the
mass and momentum equation that the pressure transient is given by the
following equation:
ΔP = − ρ aΔV (9.1)
lbm ft ft 1 ft 2 lbf − s 2
Pmax = 62 * 4000 *10 * * + 50 psig
ft 3 s s 144 in 2 32.2 lbm - ft
Pmax = 535 + 50
Pmax = 585 psig
Wavespeed
When a transient event is initiated in a pipe system, the remainder of the
system must adjust to the new conditions. In order to adjust, the
existence of the event must be communicated to the rest of the system.
This communication takes place at the wavespeed of the fluid. The
wavespeed is somewhat analogous to the sonic speed of the liquid.
However, the wavespeed is affected by the pipe structure.
The relationship between these parameters is expressed in the following
equation (Wylie, et al., 1993, pp. 27):
K ρ
a2 =
1 + c1[(K E )(D e )]
where:
a = wavespeed
K = fluid bulk modulus of elasticity
ρ = fluid density
E = pipe modulus of elasticity
D = pipe inner diameter
e = pipe wall thickness
c1 = correction factor
The sonic speed of the liquid is given by the (square root of the)
numerator. The denominator contains the reduction from sonic speed
caused by the pipe properties.
The correction factor, c1, is a function of how the pipe is restrained.
Some examples from Wylie, et al., 1993, page 27-28 are:
1. Thin-walled pipe anchored at upstream end only
c1 = 1 − μ 2
2. Thin-walled pipe anchored throughout
c1 = 1 − μ 2
3. Thin-walled pipe free to expand throughout
c1 = 1
4. Thick-walled pipe anchored at upstream end only
2e
c1 = (1 + μ ) + D (1 − μ 2)
D D+e
5. Thick-walled pipe anchored throughout
c1 =
2e
D
(
(1 + μ ) + D 1 − μ 2
D+e
)
Method of Characteristics
The analysis of waterhammer in liquid pipe systems is based on the
transient mass and momentum equations. For a complete development of
the equations please see one of the many excellent references (Wylie, et
al., 1993, Chaudhry, 1987, and Swaffield et al., 1993). Solving the
equations based on a prescribed relationship between time step and
distance step is the essence of the Method of Characteristics which will
be discussed in depth in the following sections.
Momentum equation
Following from Wylie, et al., 1993, page 22, the momentum equation
can be expressed as
1∂ P ∂V fV V
+ + g sin(α ) + =0 (9.3)
ρ∂ x ∂t 2D
where:
P = pressure
V = velocity
ρ = density
x = distance along pipe
t = time
g = acceleration due to gravity
D = diameter
f = friction factor
α = angle of pipe slope
a2 ∂ V ∂ P
+ =0 (9.4)
g ∂ x ∂t
where:
a = wavespeed
Equations 9.3 and 9.4 are two partial differential equations with the two
unknowns P and V and the two independent variables x and t.
Application of the Method of Characteristics will convert these two
partial differential equations to four ordinary differential equations as
follows. To maintain flexibility with g-level, the development will leave
the parameter P in the equations rather than convert to head, H, as in
Wylie, et al., 1993.
1∂ P ∂V fV V
+ + g sin(α ) + = L1 (9.5)
ρ∂ x ∂t 2D
∂V ∂ P
ρa 2 + = L2 (9.6)
∂ x ∂t
Combine these two equation linearly using a parameter, λ, (Wylie, et al.,
1993, pp. 27)
L1 + λ L2 = 0
or,
⎛ 1 ∂ P ∂ P⎞ ⎛ ∂ V ∂V⎞
λ⎜ + ⎟+⎜ρ + λρ a 2 ⎟
⎝λ ∂ x ∂ t ⎠ ⎝ ∂ t ∂ x⎠
ρ fV V
+ ρ g sin(α ) + =0 (9.7)
2D
If
dx 1
= = λ a2 (9.8)
dt λ
then
1
λ=±
a
and
dx
= ±a
dt
From calculus
dP ∂ P dx ∂ P
= + (9.9)
dt ∂ x dt ∂ t
dV ∂ V dx ∂ V
= + (9.10)
dt ∂ x dt ∂ t
Substituting Equation 9.8 into 9.7
⎛ ∂ P dx ∂ P ⎞ ⎛ ∂ V ∂ V dx ⎞
λ⎜ + ⎟+⎜ρ +ρ ⎟
⎝ ∂ x dt ∂ t ⎠ ⎝ ∂ t ∂ x dt ⎠
ρ fV V
+ ρ g sin(α ) + =0 (9.11)
2D
and substituting Equations 9.9 and 9.10 into 9.11
⎛ dP ⎞ ⎛ dV ⎞ ρ fV V
λ⎜ ⎟+⎜ρ ⎟ + ρ g sin(α ) + =0
⎝ dt ⎠ ⎝ dt ⎠ 2D
or
1 dP dV ρ fV V
± +ρ + ρ g sin(α ) + =0 (9.12)
a dt dt 2D
Now multiply Equation 9.12 by (a dt), which also equals dx
ρ fV V
± dP + ρ adV + ρ g sin(α )dx + dx = 0 (9.13)
2D
The gravity term contains the term α, and the term can be converted as
follows
⎛ dz ⎞
ρ g sin(α )dx = ρ g ⎜ ⎟dx = ρ gdz (9.14)
⎝ dx ⎠
where dz is the vertical change in pipe elevation. Therefore, substituting
9.14 into 9.13
ρ fV V
± dP + ρ adV + ρ gdz + dx = 0
2D
Further substitution of the mass flow rate for velocity yields
a f
± dP + dm& + ρ gdz + dx m& m& = 0 (9.15)
A 2 ρDA2
Integrating Equation 9.15 with the positive sign along the C+
characteristic from point A to P (Figure 9.1) yields
t = 5Δt
t = 4Δt
t = 3Δt
t = 2Δt
P
t = Δt
C+ C-
A B
t=0
x=0 x = i-1 x=i x = i+1 x=L
where:
BM = B + R m& i +1,old
(
+ − m& up, new − m& up, old + m& down, new + m& down, old ) 2Δρt
(9.23)
If no vapor existed previously, the old volume is zero. When the vapor
volume becomes negative, the cavity collapses and the pressure again
rises above vapor pressure. In this case the conventional methods
(Equations 9.19, 9.20) again are applicable.
The DVC method is well understood and documented for interior pipe
points. However, application of the method to general boundary
conditions in multi-pipe systems is limited. The theory for many of the
junctions used in AFT Impulse is developed in Walters, 1991.
Pipe sectioning
The relationship between pipe sectioning and time step is given by
L
a= (9.24)
n Δt
where n is the number of sections and L/n is the length of each pipe
section. If the transient event occurs within some particular time step,
the number of sections, n, must be chosen such that it is consistent with
the time step in Equation 9.24.
Specifying the number of sections for multi-pipe systems is more
complicated. Each pipe has its own wavespeed and length. For each pipe
the following must be satisfied:
Li
ni = (9.25)
ai Δt
In Equation 9.25 as applied to each pipe in the system, the time step Δt
must be the same for all pipes. The length, L, is a given for each pipe and
thus cannot be changed. The wavespeed, a, is also a given for each pipe,
although it is known with less certainty than is the length. The free
parameter is then the number of sections, n.
In any pipe system there will be one pipe that is the controlling pipe.
The controlling pipe is that pipe which has the least number of sections,
sometimes only one. Once the controlling pipe is selected, the time step
is determined by solving Equation 9.24 for Δt. Then the number of
sections in the remaining pipes is obtained from Equation 9.25.
However, this presents a dilemma. Because the number of sections in
remaining pipes is derived from Equation 9.25, typically the number of
sections, n, will not be a whole number. Since partial sections in a pipe
cannot be modeled, an alternative must be found.
Because the wavespeed for each pipe is the least certain parameter, an
uncertainty in wavespeed up to 15% is possible. Therefore, the
wavespeed in each pipe is allowed to depart slightly from its original
calculated value in order to cause n to be a whole number.
Therefore, Equation 9.25 becomes
Li
ni =
(1 ± ψ )ai Δt
where ψ is the accepted uncertainty in wavespeed up to ±0.15 (Wylie,
et al., 1993, pp. 54).
Two other things should be noted about pipe sectioning. Because each
pipe section is explicitly solved along the characteristic lines, breaking
pipes into continuously smaller sections offers no improvement in
accuracy. That is, using twice the number of sections in a pipe will not
yield a more accurate prediction (see Chapter 12). This is in contrast to
other kinds of finite difference methods where more increments offers
improved accuracy.
The other item of note is the effect of breaking the model into additional
increments. According to Equation 9.24, when the number of sections is
Reservoir theory
A reservoir is similar to an assigned pressure. With the known reservoir
height, the pressure can be obtained at the junction and the flow can be
solved directly (Wylie, et al., 1993, pp. 43):
Pknown = Pi, new = H R ρg + Psurface
(
+ − m& up, new − m& up, old + m& down, new + m& down, old ) 2Δρt
The m
& up terms are known because they are specified terms for inflow
junctions. For outflow junctions the m & down terms are known. Similar to
a pipe interior node, when the vapor volume is negative, the cavity
collapses and the fluid pressure then rises above the vapor pressure.
(
Vvapor , new = Vvapor , old + − m& up, new − m& up, old ) 2Δρt
since the m & down terms are always zero. For dead ends at the beginning
of a pipe, the m& up terms are zero. Similar to a pipe interior node, when
the vapor volume is negative, the cavity collapses and the fluid pressure
then rises above the vapor pressure.
Branch theory
At a branch, there can be multiple pipes (Wylie, et al., 1993, pp. 51-53).
An additional relationship is needed, and that relationship is the
conservation of mass. The total mass flow in and out of the junction
must sum to zero.
In addition, the branch has a single pressure solution, Pj, and this
solution is common to all inflowing pipes.
where
C C
SC = ∑ P + ∑ M (9.27)
BP BM
1 1
SB = ∑ +∑ (9.28)
BP BM
In general, there may be a known flow into the branch (a flow sink) or
into the branch (a flow source). Therefore rewrite Equation 9.26 as
⎜
( )
⎛ ∑ − m& in, new − m& in, old ⎞
⎟
(
Vvapor , new = Vvapor , old + ⎜ + ∑ m& out, new + m& out, old ⎟ Δt
)
⎜ ⎟ 2ρ
⎝ ( )
⎜ − m& B, new + m& B, old ⎟
⎠
where m
& B is the branch source/sink term defined as positive flow into
the branch. The m
& in terms are obtained from the positive compatibility
equation, while the m
& out terms are obtained from the negative equation.
Similar to a pipe interior node, when the vapor volume is negative, the
cavity collapses and the fluid pressure then rises above the vapor
pressure.
Tee/Wye theory
The Tee/Wye junction is a static element, and its solution method is
essentially the same as the Branch junction. The only difference is that
the loss factors are updated dynamically similar to steady flow.
Valve theory
A valve junction has a pressure loss across the valve. When the valve
closes this loss becomes infinite (Wylie, et al., 1993, pp. 57). With two
connecting pipes, the upstream and downstream pipe pressures are
obtained from the compatibility equations (Equations 9.19, 9.20):
Pi, new,up = C P,up − BP,up m& i, new,up
The pressure drop is related to the flow rate according to the following:
where
Rvalve = Valve Resistance
Substituting the compatibility equations (Equations 9.19, 9.20) in and
eliminating the pressures
(CP,up − BP,up m& valve )
( )
− C M , down + BM , down m& valve = Rvalvem& valve2
Rearranging terms,
m& valve 2 −
1
Rvalve
(CP,up − BP,up m& valve )
+
1
Rvalve
(CM ,down + BM ,down m& valve ) = 0
Using the quadratic equation, the solution for positive flow through the
valve is
(
m& valve = − 1 Rvalve BP,up + BM , down )
( ) (
+ 1 Rvalve 2 BP,up + BM , down 2 + 4 Rvalve C P,up − C M , down )
For negative flow, the solution is
(
m& valve = 1 Rvalve BP ,up + BM , down )
( ) (
− 1 Rvalve 2 BP ,up + BM , down 2 − 4 Rvalve C P,up − CM , down )
After solving for flow rate, m
& valve , the upstream and downstream
pressures can be obtained from the compatibility equations.
If the valve closes then 1/Rvalve = 0 and
m& valve = 0
Exit valves
If the valve is an exit valve, then the downstream pressure is known and
m& valve = − 1 Rvalve BP ,up
(
+ 1 Rvalve 2 BP,up 2 + 4 Rvalve C P,up − Pexit )
Valve vapor cavitation theory
When the calculated pressure at the upstream or downstream side of a
valve junction drops below vapor pressure, a vapor cavity forms on that
side of the junction.
Vapor cavities can form on either side of the valve or on both sides.
Vvapor ,up, new = Vvapor ,up,old
(
+ − m& up, new − m& up, old + m& valve, new + m& valve, old ) 2Δρt
Vvapor , down, new = Vvapor , down, old
(
+ − m& valve, new − m& valve, old + m& down, new + m& down, old ) 2Δρt
As an example, when a vapor cavity forms downstream of a valve, the
downstream pressure become fixed at the vapor pressure. The
calculation then is similar to an exit valve, which also flows to a fixed
pressure.
m& valve = − 1 Rvalve BP ,up
(
+ 1 Rvalve 2 BP ,up 2 + 4 Rvalve C P,up − Psat )
When a vapor cavity occurs on both sides of the valve, the pressure is at
vapor pressure on both sides and thus there is no pressure drop across
the valve. Accordingly, the flow rate goes to zero.
Note: In steady flow, the infinite pipe junction is usually a known flow,
but can also be a known pressure. Thus the Infinite Pipe capability is
offered through the Assigned Flow or Assigned Pressure junction. It is
specified by setting the transient special condition to No Reflections
(Infinite Pipe).
Assuming that the infinite pipe is at the downstream end, the current
pressure can be solved using the positive characteristic compatibility
equation (Equation 9.19),
Pi, new = C P − BP m& i, new
and the negative characteristic compatibility equation (Equation 9.20)
Pi, new = CM + BM m& i, new (9.30)
m& spray
Pexit
or
2 2 ρC D 2 Aspray 2
sgn( m& )m& + m&
SB
2 ρC D 2 Aspray 2 (S B Pexit − SC )
+ =0
SB
Note: If the spray junction pressure falls below the exit pressure, reverse
flow will occur. If the exit pressure is the atmosphere, the reverse flow
will be air. However, Impulse assumes the inflow is at the pipe system
density. This can lead to erroneous results if ignored. Reverse flow can
be determined by graphing the flow at the spray junction or looking at
the summary of maximum and minimum flows in the Output window.
⎜
(
⎛ ∑ − m& in, new − m& in, old ) ⎞
⎟
(
Vvapor , new = Vvapor , old + ⎜ + ∑ m& out, new + m& out, old ) ⎟ Δt
⎜ ⎟ 2ρ
⎝ (
⎜ + m& spray , new + m& spray , old )
⎟
⎠
where m
& spray is the spray outflow term. The m& in terms are obtained
from the positive compatibility equation, while the m
& out terms are
obtained from the negative equation.
Similar to a pipe interior node, when the vapor volume is negative, the
cavity collapses and the fluid pressure then rises above the vapor
pressure.
Pump theory
The pump junction can model the effects of a pump with a known
change in pump speed (Wylie, et al., 1993, pp. 59-60), or with inertial
data where the pump speed is calculated.
ΔH = a + bQ pump + cQ pump 2
where a, b and c are obtained from curve fits of vendor data. This curve
assumes the pump operates at 100% speed. During the transient, the
speed will change.
The affinity law for pumps given by the following equations
ΔH
= constant1
α2
Q
= constant2
α
where α is the pump speed. Applying these relationships the pump curve
is transformed
ΔH Q pump Q pump 2
= a′ + b′ + c′
α2 α α2
⎛ αb′ c′ ⎞
ΔP = ρg ⎜α 2 a′ + m& pump + m& pump 2 ⎟
⎜ ρ ρ2 ⎟
⎝ ⎠
where the primes on the constants were dropped after they were divided
through by ρg.
By use of the two compatibility equations (Equation 9.19, 9.20),
Pup, new = C P − BP m& pump
This is a quadratic equation for the pump flow rate which can be solved
by the quadratic formula
⎛ BP,up + BM , down − bα ⎞
Q pump = ⎜⎜ ⎟
⎟
⎝ 2c ⎠
⎡ ⎛ 2
(
⎢ ⎜ 4b α a + C P ,up − C M , down
× ⎢1 − 1 −
)⎞⎟1 2 ⎤⎥
⎢⎣ ⎝
⎜ (
BP,up + BM , down − bα 2 ) ⎟
⎠
⎥
⎥⎦
With the solution for the pump flow rate, the pressure upstream and
downstream of the pump can be obtained from Equations 9.19 and 9.20.
⎛ BM , down − bα ⎞
m& pump = ⎜⎜ ⎟⎟
⎝ 2c ⎠
⎡ ⎛
⎢ ⎜
× ⎢1 − 1 −
(
4b α 2 a + PR − CM , down )⎞⎟1 2 ⎤⎥
⎢⎣ ⎝
⎜ (
BM , down − bα 2 ) ⎟
⎠
⎥
⎥⎦
(
+ − m& up, new − m& up, old + m& pump, new + m& pump, old ) 2Δρt
Vvapor , down, new = Vvapor , down, old
(
+ − m& pump, new − m& pump, old + m& down, new + m& down, old ) 2Δρt
As an example, when a vapor cavity forms upstream of a pump, the
upstream pressure becomes fixed at the vapor pressure. The calculation
then is similar to a submerged pump, which also flows from a fixed
pressure.
⎛ BM , down − bα υ ⎞
m& pump = ⎜ ⎟
⎜ 2c ρ 2 ⎟
⎝ ⎠
⎡ ⎛
× ⎢⎢1 − ⎜1 −
(
4b α 2 a + Psat − CM , down )⎞⎟1 2 ⎤⎥
⎢⎣ ⎝
⎜ ( BM , down − bα ρ 2) ⎟
⎠
⎥
⎥⎦
Similar to a pipe interior node, when the vapor volume is negative the
cavity collapses and the fluid pressure then rises above the vapor
pressure.
Figure 9.3 Motor torque vs. speed data is used for startup
modeling
This calculation uses the user provided pump data for head vs. flow and
power vs. flow, and thus only works for positive or zero flows. The mass
flow rate and head are obtained from Equations 9.19 and 9.20 (with the
pressure converted to head) and the pump curve supplied by the user.
The typical shape of a motor torque vs. speed curve is shown in Karassik
(2001, p. 8.25, Figure 26), which is reflected in the Figure 9.4 data.
Figure 9.5 Pump window with pump data entered for specific
speed of 0.46.
v
θ = π + tan −1 (9.39b)
α
Some references choose to define θ slightly differently:
α
θ = tan −1
v
AFT Impulse allows you to select either definition as the basis for the FH
and FB data in Equation 9.39a.
The parameters FH and FB are properties of the pump, and it is typically
assumed that for pumps of similar specific speed they are invariant (see
Brown & Rogers for a discussion questioning this assumption). Figures
9.5 and 9.6 show data for a pump with a specific speed of 0.46 as given
by Wylie, et al., (1993, pp. 147-148).
AFT Impulse provides data of FH and FB vs. θ for pumps of twenty-one
specific speeds. See Table 9.1 for a list of all specific speeds available.
With data for FH and FB, the compatibility equations (9.19, 9.20) are
used as follows:
ΔPpump = C M − C P + (BM + BP )m& pump (9.40)
(
ΔPpump = PR α 2 + v 2 FH ) (9.41)
( )
C P − CM − vm& R (BP + BM ) + PR α 2 + v 2 FH = 0 (9.42)
( )⎝ N
0.9556
⎛ P ⎞
I p = 1.5 107 ⎜⎜ ⎟ (9.46)
3⎟
⎠
(
ΔV = − m& A, new + m& A, old ) 2Δρt
Therefore,
Pj , new = Pj , old −
(
K ′Δt m& A, new + m& A,old )
2Vρ
⎜
( )
⎛ ∑ − m& in, new − m& in, old ⎞
⎟
(
Vvapor , new = Vvapor , old + ⎜ + ∑ m& out, new + m& out, old ⎟ Δt
)
⎜ ⎟ 2ρ
⎝ ( )
⎜ − m& A, new + m& A, old ⎟
⎠
where m & A is the accumulator flow out of the junction. The m& in terms
are obtained from the positive compatibility equation, while the
m& out terms are obtained from the negative equation.
Similar to a pipe interior node, when the vapor volume is negative, the
cavity collapses and the fluid pressure then rises above the vapor
pressure.
1 2K 2 Kρ ⎛ S ⎞
m& exit , new2 + m& exit , new + ⎜⎜ Pexit − C ⎟⎟ = 0
S B Aexit 2 Aexit 2 ⎝ SB ⎠
If using the variable Cv model, the K factor above itself becomes a
function of ΔP, which requires iteration to solve for the mass flow rate.
Variable Cv model
For passive relief valves the valve Cv typically depends on the pressure
difference across the valve, as the pressure difference is what causes the
valve to open. When the pressure difference is below the cracking
pressure, the valve remains closed. When it rises above the cracking
pressure, it opens. As the pressure difference increases, it opens further
until it is fully open. Any further pressure increases do not increase the
Cv, and so the fully open Cv is used for all higher pressures.
When working with the variable Cv model, the cracking pressure is
entered in pressure/head difference (see Figure 9.7). In addition, the first
Cv data must equal zero (see Figure 9.8), and the corresponding
pressure/head to zero Cv must be the same as the cracking pressure/head
difference (shown as 700 kPa in both Figures 9.7 and 9.8).
In Figure 9.8, the first data point represents the valve closed at the
cracking pressure, with any greater pressure difference resulting in a
positive Cv, which means that flow will occur. The maximum pressure
difference in the table (770 kPa in Figure 9.8) represents the pressure
difference to fully open the valve. The Cv of 150 thus represents the Cv
when the valve is fully open. Any pressure differences greater than 770
kPa will result in a Cv of 150, the maximum value.
Figure 9.8 First data point must have Cv of zero with the
pressure/head loss equal to the cracking
pressure/head difference (as shown in Figure 9.7)
C − CM − ΔP
m& i, new = P
BP + BM
LL Orifice
Connector
m& T Pipe
LC
SC ⎛ m& Δt ⎞
− Psurface − ρg ⎜⎜ LL,old − T , old ⎟⎟ + C1
SB ⎝ 2 ρAT ⎠
m& T , new =
⎛ gΔt 1 ⎞
− ⎜⎜ C2 + + ⎟⎟
⎝ 2 AT S B⎠
This flow rate can be back substituted to obtain the junction pressure and
surge tank liquid height.
PA
Orifice
Connector
m& A Pipe
LC
PAV An = C A
where n is the polytropic constant (1.4 for isentropic air, 1.0 for
isothermal) and CA is a constant.
Therefore,
(
V A, new = V A, old − m& A, new + m& A,old ) 2Δρt
By substitution
SC + m& A, new
= PA, new − C1 − C2 m& A, new
SB
or
SC + m& A, new
PA, new = + C1 + C2 m& A, new
SB
Combining equations and simplifying,
( )( )
F = C3 + C4 m& A, new C5 + C6 m& A, new n − C A
When the value of m & A, new has been found that causes F to equal zero,
the equation will be satisfied. To solve by Newton-Raphson the
derivative of F is given by,
dF
dm& A, new
= F′ =
C7
C5 + C6 m& A, new
(
+ C4 C5 + C6 m& A, new n )
where
C7 = − nC AC6
By iteration, m& A, new is calculated. This value can then be back-
substituted to obtain the new junction pressure and new gas volume.
Sonic inflow
(γ +1) (2(1− γ ))
C d Ain Po ⎛ γ − 1⎞
m& = γ ⎜1 + ⎟ (9.48a)
RTo ⎝ 2 ⎠
Subsonic inflow
2γ (γ +1) γ ⎤
C A P 2γ ⎡⎢⎛ Pi ⎞ ⎛ Pi ⎞
⎥
m& = d in o ⎜ ⎟ − ⎜⎜ ⎟⎟ (9.48b)
RTo γ − 1 ⎢⎜⎝ Po ⎟⎠ P
⎝ o⎠ ⎥
⎣ ⎦
Sonic outflow
(γ +1) (2(1−γ ))
C A P ⎛ γ −1⎞
m& = d out i γ ⎜1 + ⎟ (9.48c)
RTL ⎝ 2 ⎠
Subsonic outflow
−2 γ − (γ +1) γ
C d Aout Pi 2γ ⎡⎢⎛ Pi ⎞ ⎛ Pi ⎞ ⎤
⎥ (9.48d)
m& = ⎜ ⎟ − ⎜⎜ ⎟⎟
RTL γ − 1 ⎢⎜⎝ Po ⎟⎠ ⎝ Po ⎠ ⎥
⎣ ⎦
where:
Pi = pressure in the pipe at the junction node
Po = atmospheric pressure
To = atmospheric temperature
TL = liquid temperature
Equation of state
Besides the compatibility equations (9.19 and 9.20), we need another
equation to form a complete set. The ideal gas equation of state is used:
PV = mRT
Substituting terms one obtains
Pi ⎛
RTL ⎝
⎜Vold +
Δt
2
( ⎞
Qdown, new + Qdown,old − Qup, new − Qup, old ⎟ = )
⎠
⎛
⎜ mold +
Δt
2
( ⎞
m& gas, new + m& gas, old ⎟ )
⎝ ⎠
Eliminating the volumetric flowrate using the mass flowrate obtains
Pi ⎛ 2Vold
⎜⎜
RTL ⎝ Δt
+
1
ρL
(
m& down, new + m& down,old − m& up, new − m& up, old )⎞⎟⎟ =
⎠
⎛ 2mold ⎞
⎜ + m& gas, new + m& gas, old ⎟
⎝ Δt ⎠
Pi ⎛ 2Vold 1 ⎛ Pi − C M C −P ⎞⎞
⎜ + ⎜⎜ + & down, old − P i − m& up,old ⎟⎟ ⎟ =
m
RTL ⎜⎝ Δt ρ L ⎝ BM BP ⎠⎠
⎟
⎛ 2mold ⎞
⎜ + m& gas, old + m& gas, new ⎟
⎝ Δt ⎠
(9.49)
Now we can use Equations 9.48 to eliminate m
& gas, new as follows:
⎛ 1 1 ⎞
Pi 2 ⎜⎜ + ⎟
⎝ ρ L BM ρ L BP ⎟⎠
⎛ C CP m& d m& u 2V ⎞
+ Pi ⎜⎜ − M − + old − old + old ⎟
⎟
⎝ ρ L BM ρ L BP ρL ρL Δt ⎠
⎛ 2m ⎞
− RTL ⎜ old + m& gasold + m& sonic ⎟ = 0
⎝ Δt ⎠
where m& sonic is the sonic inflow which can be calculated from Equation
9.49a based completely on input data.
Subsonic Inflow
Equation 9.49 is modified using Equation 9.48b and solved iteratively
using Newton-Raphson. Defining Pr = Pi/Po obtains:
⎛ 1 1 ⎞
Pr 2 Po 2 ⎜⎜ + ⎟⎟
⎝ ρ L BM ρ L BP ⎠
⎛ 2V C C m& d m& u ⎞
+ Pr Po ⎜⎜ old − M − P + old − old ⎟
⎟
⎝ Δt ρ L BM ρ L BP ρL ρL ⎠
⎛ 2m ⎞
− RTL ⎜ old + m& gasold ⎟
⎝ Δt ⎠
⎛ 2γR ⎞ 2 γ
+ Cd Ain PoTL ⎜⎜
(γ − 1)T
⎟⎟ Pr (
− Pr (γ +1) γ = 0 )
⎝ o⎠
(
C1 Pr a − Pr b)0.5 + C2 Pr 2 + C3Pr + C4 = F
0.5C1 (Pr a − Pr b ) (aPr a −1 − bPr b −1 )+ 2C2 Pr + C3 =
−0.5 dF
dP r
2
a=
γ
γ +1
b=
γ
2γR
C1 = Cd Ain PoTL
(γ − 1)To
⎛ 1 1 ⎞
C2 = Po 2 ⎜⎜ + ⎟⎟
ρ
⎝ L MB ρ B
L P⎠
⎛ 2V C C m& d m& u ⎞
C3 = Po ⎜⎜ old − M − P + old − old ⎟
⎟
⎝ Δt ρ L BM ρ L BP ρL ρL ⎠
⎛ 2m ⎞
C4 = − RTL ⎜ old + m& gasold ⎟
⎝ Δt ⎠
where the function F is solved for Pr.
Subsonic Outflow
Equation 9.49 is modified using Equation 9.48d and solved iteratively
using Newton-Raphson. Defining Pr = Pi/Po obtains:
⎛ 1 1 ⎞
Pr 2 Po 2 ⎜⎜ + ⎟⎟
⎝ ρ L BM ρ L BP ⎠
⎛ 2V C CP m& d m& u ⎞
+ Pr Po ⎜⎜ old − M − + old − old ⎟
⎟
⎝ Δt ρ L BM ρ L BP ρL ρL ⎠
⎛ 2m ⎞
− RTL ⎜ old + m& gasold ⎟
⎝ Δt ⎠
+ Pr (C D Aout Po )
γ −1
Pr [
2γRTL − 2 γ
− Pr − (γ +1) γ = 0 ]
Defining terms for Newton-Raphson solution yields
(
C1 Pr − a − Pr − b )0.5 + C2 Pr + C3 + C4 P1 = F
r
(
0.5C1Pr Pr − a − Pr −b )−0.5 (− aPr −a −1 + bPr −b−1 )
1 dF
+ C 2 − C4 =
2 dPr
Pr
2
a=
γ
γ +1
b=
γ
2γRTL
C1 = Cd Aout Po
γ −1
⎛ 1 1 ⎞
C2 = Po 2 ⎜⎜ + ⎟⎟
⎝ ρ L BM ρ L BP ⎠
⎛ 2V C CP m& d m& u ⎞
C3 = Po ⎜⎜ old − M − + old − old ⎟
⎟
⎝ Δt ρ L BM ρ L BP ρL ρL ⎠
⎛ 2m ⎞
C4 = − RTL ⎜ old + m& gasold ⎟
⎝ Δt ⎠
where the function F is solved for Pr.
⎛ 1 1 ⎞
Pi 2 ⎜⎜ + ⎟⎟
⎝ ρ L BM ρ L BP ⎠
⎛ 2V C CP m& d m& u ⎞
+ Pi ⎜⎜ old − M − + old − old ⎟
⎟
⎝ Δt ρ L BM ρ L BP ρL ρL ⎠
⎛ (γ +1) (2(1−γ )) ⎞
⎛ γ −1⎞
+ Pi ⎜ C D Aout γRTL ⎜1 + ⎟ ⎟
⎜ ⎝ 2 ⎠ ⎟
⎝ ⎠
⎛ 2m ⎞
− RTL ⎜ old + m& gasold ⎟ = 0
⎝ Δt ⎠
1
Δz
2
orifice 3
m& 2 2
Forifice = A2
2 ρC D 2 Ao 2
Fg = ρA2 gΔz
Substituting terms and dividing by area A2
fL2 1 1 L dm& 2
P1 − P3 − m& 2 2 − & 2 2 − ρgz = 2
m
D2 2 ρA2 2 2 ρC D 2 Ao 2 A2 dt
Taking average values over time, and denoting current time by a prime,
2 L2
C1 = P3 − P1 + 2 ρgz − m& 2 (9.50)
A2 Δt
fL2 1 1 2 L2
C2 = m& 2 + m& 2 + (9.51)
D2 ρA2 2 2 2 A2 Δt
ρC D Ao
If a connector pipe exists but no orifice then
2 L2
C1 = P3 − P1 + 2 ρgz − m& 2
A2 Δt
fL2 1 2 L2
C2 = m& 2 +
D2 ρA2 2 A2 Δt
Cavitation
When a positive pressure surge occurs, it is always followed by negative
pressure surge that drops the pressure. If the drop in pressure causes
cavitation, then the pressure spike that occurs when the cavity collapses
can cause the local pressure to exceed that predicted by Equation 9.52.
The potential to cavitate can be assessed by comparing the pressure rise
in Equation 9.52 to the steady-state pressure. If the pressure surge is
greater than the steady-state pressure, then cavitation is possible. If it is
less than the steady-state pressure and the fluid has a low vapor pressure
(like water), then it will probably not cavitate. If the fluid has high vapor
pressure, this will need to be subtracted from the steady-state pressure
and then compared to the pressure surge.
Elevation changes
If the pressure surge propagates to an area of the system at a lower
elevation, the peak pressure can be higher because of the higher static
pressure.
Time-based transients
Time-based transients occur in accordance with the time of the transient
simulation. The start and stop times of the simulation are specified in the
Transient Control window. The initiation of time-based transients are
pre-specified before a model is run.
For example, consider a valve closure transient. Assume the valve starts
to close at two seconds into the simulation, and the time it takes for the
valve to close is one second. After that, the valve stays closed. If the
initial valve Cv is 250, the transient would appear as in Figure 10.1.
In the "Initiation of Transient" area there are four options. For time-
based transients, the option is specified as Time. In the Transient Data
area the data is entered. Here the Cv profile of the valve will follow the
profile entered and start to close at 2 seconds.
Event-based transients
Event transients are initiated when some user specified criteria is met.
For example, a valve can start to close when a certain pressure is reached
at a point in the system. If the pressure is never reached, the transient is
never initiated.
Event transients are one of three types. The three types will be discussed
in the following sections.
The data in the Transient Data table has a slightly different meaning than
does a time-based transient. Here time zero is relative to the time at
which the event criteria is first met. For example, if Pipe 2 reaches 1000
kPa at 3.65 seconds, the valve will start its transient at 3.65 seconds. The
valve will close (i.e., Cv = 0) at 1 second after the event initiation, or
4.65 seconds.
Sequential events
Sequential dual events progress from the first event to the second and no
further. For example, these could be two high pressure settings. When
the first setting is reached the valve opens partially, and when the second
setting is reached, which would likely be at a higher pressure than the
first event, the valve opens the rest of the way. Similar to a single event
transient, if the pressure drops and then rises again, the events would not
be initiated again.
Sequential dual events are specified by selecting Dual Event Sequential
in the Initiation of Transient area of the junction window (Figure 10.3).
Similar to Cyclic Dual Events, two event tabs appear where the first and
second event can be specified. The transients that are initiated when the
events occur are specified in the table of the First Transient and Second
Transient tab.
Check valve
The check valve has two built-in events. The first is that it closes when
backflow starts to occur. The second is that it reopens when sufficient
pressure differential occurs. If the user does not enter a closing or
opening transient in the check valve window, these transients are
assumed to be instantaneous.
Relief valve
The relief valve has two built-in events. The first is that it opens when
the cracking pressure is reached. The second is that it closes again when
the pressure falls back below the cracking pressure. If the user does not
enter an opening or closing transient in the relief valve window, these
transients are assumed to be instantaneous.
Repeat transient
For transients that are periodic, you can specify that the transient repeats
itself. This is specified by selecting the "Repeat Transient" checkbox
(see Figure 10.1, 10.2 or 10.3). When Repeat Transient is selected, the
first and last data points must match. After the transient reaches the final
data point, it returns to the first data point and starts over.
Event messages
Events that are initiated during the simulation are displayed in the two
Event Messages area of the Output window (see Figure 10.4). The first
is a list of event messages sorted by junction. The second is a list of
event messages sorted by time. If no events occur, the Event Messages
tab will be hidden.
Figure 10.4 The Event Messages area shows a listing of all events
that occur during the simulation sorted by junction
and time.
Figure 10.6 All junction transient data is shown in the Model Data
window on the Transient Data tab.
General considerations
Surge pressures that are excessively high can burst pipes and damage
equipment. Surge pressures that reduce the local pressure can result in
pipes being crushed (due to atmospheric pressure exceeding the internal
liquid pressure), cavitation and liquid column separation (which can then
result in large pressure spikes when the cavity collapses), and
subatmospheric pressures that are unacceptable for drinking water
pipelines.
The surge pressures result from changing conditions in the system,
which change the fluid velocity. Options include slowing down the
system changes and thus reducing velocity or providing locations which
can deliver extra fluid when the pressure decreases or accept extra fluid
when the pressure rises.
If using a surge suppression device, it is generally desirable to locate the
device as close as possible to the cause of the transient. The practicality
of the different options depend highly on the particular case of interest.
Methods for surge reduction are discussed in detail in Swaffield et al.
(1993, Chapter 6), Wylie et al. (1993, Chapter 10), and Chaudhry (1987,
Chapter 10).
Flexible hose
Flexible hose can sometimes be used in small systems, which reduces
wavespeed and hence pressure surge.
Parallel valves
Rather than use one large valve to stop the flow, perhaps two or more
valves can be used in parallel. The timing of the valve closures is
staggered, thus allowing a portion of the flow to be stopped with each
valve.
Relief valves
Relief valves can be used to avoid high pressure situations. In addition,
inward relief valves can be used for low pressure situations.
Surge tank
Surge tanks can be an effective way to reduce surge pressures if the
steady-state system pressures are relatively low. High steady-state
pressures result in high hydraulic gradelines. Since the liquid level in a
surge tank rises to the local hydraulic gradeline, they would have to be
very tall if located in a high pressure region.
Gas accumulator
Gas accumulators can be effective in spreading out the wave front of a
transient and thus reducing peak pressures. Gas accumulators change the
frequency response of the system and can in fact amplify pressure surge
if not sized and located properly.
Figure 12.2 Model Data window display for the model in Figure
12.1. From this perspective it is much easier to see
that one of the pipes (#4) has length data entered in
miles rather than feet. This is not automatically an
input error, as AFT Impulse accepts pipe input with
mixed engineering units. However, more frequently
than not the user intends to use consistent units, so
this probably is an error.
12
10
Pressure (psid) 6
a
4
0
0 50 100 150 200
Flow (ft3/s ec)
12
10
8
Pressure (psid)
b 6
0
0 50 100 150 200
Flow (ft3/s ec)
Figure 12.3 Top graph shows a third or fourth order pump curve
fit with an inversion at 150 ft3/sec. AFT Impulse
interprets flow rates to the right of this inversion as
acceptable which can cause convergence problems or
erroneous results. All pump curve fits should include
sufficient data such that the curve crosses the zero
pressure rise axis as shown in the bottom curve.
Application example
A real life example is shown in Figure 12.4. The system represents part
of a tank farm for refined petroleum products. Gasoline flows from a
refinery some 100 km away into the tank farm. The supply pipeline is
represented by pipe P1. Pipes P2-P9 are relatively short local pipes at the
tank farm with changing diameters. This model is for gasoline, and the
gasoline comes into junction J10 and flows initially to the tank at J106.
The transient of interest is when the J201 valve is opened and the J104
The transient lasts for about 30 seconds. During the opening and closing
of the valves, the transient waves propagate out to J106 and J212, as
well as to the supply junction J1. Using the wavespeed of gasoline it can
be determined that any waves that propagate to pipe P1 will not reflect
back to the tank farm within 30 seconds, and thus will not impact the
transient. However, during the transient procedure, gasoline still flows
from junction J1 into the system.
Discharge
Tank (J2)
Transfer
Pipeline
Transfer
Supply Pump (J3)
Tank (J1)
Shoreline
Elevation
Lake Michigan
J2
P3
J1 J3 J4
P1 P2
Transfer Discharge
Pipeline Tank (J2)
Transfer
Supply Pump (J3)
Tank (J1)
Shoreline
Elevation
J2
P2
J1 J3
P1
If the goal is to size the pump at J3, the model is Figure 12.8 can be used
with the pump modeled as an assigned flow pump.
Examples
The simplest example is the system shown in Figure 12.9. In AFT
Impulse terms, the system has two assigned flow junctions.
L = 30 m
D = 5 cm
Inlet f = 0.02 Outlet
Q = 0.05 m3/s Q = 0.05 m3/s
El = 0 m El = 0 m
Figure 12.9 Model with two assigned flows. This model does not
have a unique solution.
Obviously, the flow in the pipe is known, but, what is the pressure at the
inlet? At the outlet? It cannot be determined because there is no
reference pressure. The reference pressure is that pressure from which
other pressures in the system are derived. There can be one or more
reference pressures, but there always has to be at least one.
The model in Figure 12.9 can be built with AFT Impulse. If you try to
run this model, AFT Impulse will inform you that it cannot run it
because of the lack of a reference pressure. In AFT Impulse there are
numerous junctions that can act as a reference pressure: the reservoir,
assigned pressure, spray discharge, pump with submerged option, and
valve and relief valve after cracking (with the exit valve option) behave
as reference pressures because that is their nature. Three other junction
types, surge tank, gas accumulator, and assigned flow (when the steady-
state flow is zero) have a special feature that allows them to act as
reference pressures. This feature is found on each junction's Optional
tab, and is discussed at length in Chapter 6.
There are a couple other things worth noting about Figure 12.9. First is
that the model in Figure 12.9 has redundant boundary conditions. When
L = 30 m
D = 5 cm
Inlet Outlet
f = 0.02
b P = 0.1 MPa Q = 0.05 m3/s
El = 1.5 m El = 0 m
L = 30 m
Inlet D = 5 cm Outlet
c P = 0.1 MPa f = 0.02 P = 0 MPa(g)
El = 1.5 m El = 1.5 m
Figure 12.10a-c Top two models with one pressure and one
assigned flow, bottom model with two pressures.
All of these models have a unique solution.
Second, if the user was allowed to specify two flows as in Figure 12.9,
he/she could specify them with different flow rates. Clearly, they must
have the same flow or an inconsistency occurs. The basic reason the
inconsistency would be possible is because Figure 12.9 does not have a
unique solution.
Figure 12.10a-c shows three other model configuration possibilities. It is
not possible to specify inconsistent conditions for any of the Figure
Outlet
Q = 0.025 m3/s
El = 0 m
Inlet
Q = 0.05 m3/s Outlet
El = 0 m Q = 0.025 m3/s
El = 0 m
J1 J2 J3
Inlet Outlet
a P = 0.1 MPa(g) Q = 0.05 m3/s
El = 3 m El = 0 m
FCV
Q = 0.05 m3/s
J1 J2 J3
Inlet Outlet
b P = 0.1 MPa(g) P = 0 MPa(g)
El = 1.5 m El = 1.5 m
FCV
Q = 0.05 m3/s
Figure 12.12a-b The model at the top does not have a reference
pressure after the Flow Control Valve so the
pressure drop across the FCV cannot be
determined. The top model does not have a unique
solution. The bottom model has a pressure
upstream and downstream, and a unique solution
exists.
Figure 12.13 AFT Impulse message when you try to run the model
shown in Figure 12.12a.
Inlet Outlet
P = 0.1 MPa(g) Q = 0.05 m3/s
El = 1.5 m El = 0 m
Pump with Known (fixed) Flow
Q = 0.05 m3/s
Inlet Outlet
P = 0.1 MPa(g) P = 0 MPa(g)
a El = 1.5 m
El = 1.5 m
FCV FCV
Q = 0.05 m3/s Q = 0.05 m3/s
Inlet Outlet
b P = 0.1 MPa(g) P = 0.15 MPa(g)
El = 1.5 m El = 1.5 m
Pump with Fixed FCV
Flow Q = 0.05 m3/s
Q = 0.05 m3/s
Figure 12.15 a-b Neither model above has a unique solution. The
top model has two flow control valves in series.
The bottom model has a pump modeled as an
assigned flow in series with an FCV. In both cases
either a third pressure junction is needed between
the two middle junctions, or one of the junctions
must be changed from a flow controlling device.
Inlet Outlet
a P = 0.1 MPa(g) P = 0.15 MPa(g)
El = 1.5 m El = 1.5 m
Pump with FCV
Fixed Flow Q = 0.05 m3/s
Q = 0.05 m3/s
Inlet Outlet
b P = 0.1 MPa(g) P = 0.15 MPa(g)
El = 1.5 m El = 1.5 m
Pump with PDCV
Fixed Flow dP = 0.05 Mpa
Q = 0.05 m3/s
Inlet Outlet
c P = 0.1 MPa(g) P = 0.15 MPa(g)
El = 1.5 m El = 1.5 m
Pump with Curve FCV
dP = f(Q) Q = 0.05 m3/s
Figure 12.16a-c. The model at the top is the same as Figure 12.15b
and does not have a unique solution. To size the
pump, change the model to the second one shown
above, which uses a pressure drop control valve
(PDCV) rather than an FCV. Once the second
model is run, the pump is sized, a pump curve
exists, and the model at the bottom can be run
using an FCV.
P5^ ^P8
J5 J8
FCV FCV
Q = 100 gpm Q = 100 gpm
P4^ P7^
J1 J2 J3 J4 J7
P1> P2> P3> P6>
This brings up the question: which FCV should you turn into a PDCV?
When FCV’s are put in parallel, frequently the pipe design has one of
the FCV’s further away from the pump than the others. Because of the
additional piping leading to this FCV, it will be the weakest link in the
chain of parallel FCV’s, by virtue of having the lowest pressure drop
across it. The most remote FCV should be chosen as the PDCV. If the
most remote FCV is chosen, when the minimum pressure drop required
is applied to the PDCV all other FCV’s in the parallel system will have
greater than the minimum pressure drop — thus satisfying the pressure
drop requirement for those FCV’s as well.
Figure 12.18a-b. AFT Impulse results for the model in Figure 12.17
after J8 has been changed to a PDCV. Top results
show valve summary and bottom results show
pump size.
Inlet Outlet
Q = 0.05 m3/s P = 0 MPa(g)
El = 3 m El = 3 m
PRV
Pdown = 0.1 MPa(g)
Figure 12.20 shows the possible cases with PRV’s and PSV’s, while
Table 12.2 comments on the six cases. In summary, at least one known
pressure is always needed on the side of a pressure control valve
opposite of the controlled side.
Inlet Outlet
a Q = 0.05 m3/s P = 0 MPa(g)
El = 3 m PRV El = 3 m
Pdown = 0.1 MPa(g)
Inlet Outlet
b P = 0.2 MPa(g) Q = 0.05 m3/s
El = 3 m El = 3 m
PRV
Pdown = 0.1 MPa(g)
Inlet Outlet
c P = 0.2 MPa(g) P = 0 MPa(g)
El = 3 m El = 3 m
PRV
Pdown = 0.1 MPa(g)
Inlet Outlet
d Q = 0.05 m3/s P = 0 MPa(g)
El = 3 m El = 3 m
PSV
Pup = 0.15 MPa(g)
Inlet Outlet
e P = 0.2 MPa(g) Q = 0.05 m3/s
El = 3 m El = 3 m
PSV
Pup = 0.15 MPa(g)
Inlet Outlet
f P = 0.2 MPa(g) P = 0 MPa(g)
El = 3 m El = 3 m
PSV
Pup = 0.15 MPa(g)
a
P1v P4^
J2 J3 J4 J7 J6
P2> P3> P6> P5>
Open
J1 J5
b
P1v P4^
J2 J3 XJ4 J7 J6
P2> P3> P6> P5>
Closed
Figure 12.21 a-b Top model has reference pressures at J1 and J2.
Bottom model has closed valve at J4 which
isolates the J5 and J6 assigned flows. There is no
reference pressure for J5 and J6 and no unique
solution exists for the bottom model.
Artificial transients
An artificial transient occurs when the steady-state results do not match
the transient calculation initial step. During the first time step AFT
Impulse tries to bring the mismatched conditions to what it considers
steady conditions which generates a transient pressure. This transient is
not a true physical transient, but a false “artificial” one based on
mismatched data. If undetected, the artificial transient will interact with
the physical transient being modeled and corrupt the results.
Figure 12.22 The Section Pipes window alerts you to pipes that
could be combined and possibly reduce model
runtime.
Before AFT Impulse initiates the transient simulation, it performs a
“zero time step computation” and compares the predicted results with
the results from the steady-state solution. If the difference exceeds the
artificial transient criteria (0.5% by default), then AFT Impulse will
display a warning in the output. If Stop Run if Artificial Transient
Detected is selected, the simulation will be immediately terminated. This
feature is especially helpful for simulations that take a long time to run.
For example, it is wasteful to run a large model for an hour only to find
at the end that an artificial transient input error was present at the very
beginning, thus nullifying the results.
The artificial transient detection feature is an automatic way for the user
to identify problems of this nature. A manual process can also be used,
and it is instructive to consider the manual process to better understand
artificial transients.
In some cases where pipes are initially not flowing, the first transient
calculation is some very small but non-zero flowrate. This can appear to
Double-clicking tool
If you double-click the Selection Drawing Tool, the Selection Drawing
Tool remains active after you have completed the selection. It remains
active until you click it again a single time
Control key
If you hold down the CTRL key when completing the selection drawing
(just before releasing the mouse button), the Selection Drawing Tool
remains active. The CTRL key thus performs a similar function to
double-clicking the tool.
Shift key
If you first double-click the Selection Drawing Tool or hold down the
CTRL key, you can also hold down the SHIFT key while drawing the
selection box and the enclosed pipes and junctions will be deselected.
Double-clicking tool
If you Double-click the Pipe Drawing Tool it remains active until you
click it again a single time. This allows you to draw a series of pipes
without returning to the Toolbox each time.
Control key
If you hold down the CTRL key when completing the pipe drawing, the
Pipe Drawing Tool remains active, and you can draw a series of pipes
without returning to the Toolbox each time. The CTRL key thus
performs a similar function to double-clicking the tool.
Shift key
If you hold down the SHIFT key down while drawing pipes on the
Workspace, you can draw pipes that are perfectly horizontal or vertical.
This may improve the aesthetics of your model.
Shift key
If you hold down the SHIFT key down while selecting the zoom area,
the zoom state will always be 100%.
Control key
A junction can be “morphed” from one type to another. To morph a
junction, hold down the CTRL key while selecting a junction from the
Toolbox. Drop the junction onto an existing Workspace junction. The
junction type will change to the new junction type, and data that can be
kept will be copied into the new junction.
Shift key
If you hold down the SHIFT key while selecting a junction from the
Toolbox and then drop the junction onto an existing pipe, the pipe will
split into two pipes. The physical length of the original pipe will be
automatically halved, and the new pipe will be assigned the balance.
Thus the sum of the two pipe lengths will equal the length of the original
pipe. Any fittings/losses in the original pipe will be left in the original
pipe, and the new pipe will have no fittings/losses.
Shift key
If you hold down the SHIFT key when you perform an inspection, only
the items not yet defined are displayed.
Control key
If you hold down the CTRL key when you perform an inspection, the
inspection window will show the output data for the pipe or junction.
Control key
By default, when you move an existing junction icon, any connected
pipes will move along with the junction and retain the connection. If you
hold down the CTRL key as you release the junction, only the junction
will move. All pipe connections will be broken. This behavior can be
reversed such that connections will not be retained unless you hold down
the CTRL key. This reversal can be specified on the Workspace
Preferences window.
F2 function key
When in a pipe or junction specifications window you can toggle the
highlight feature by pressing the F2 function key. You can also toggle
this from the Options menu.
F5 function key
When in a pipe specifications window, if you press the F5 function key,
you will jump to the next highest numbered pipe. If you press the CTRL
key while pressing the F5 function key, you will jump to the next lowest
numbered pipe.
Limitations
The size of model you can create with AFT Impulse is limited by the
amount of RAM on your PC. When solving a system, AFT Impulse
creates a square matrix that is double-precision. The size of the matrix is
determined by the total number of tee/wye and branch junctions in your
model.
For example, if you have 100 tee/wye and branch junctions, a 100x100
array is created. With double precision, the amount of RAM needed is
16x100x100, or 160KB. A 1000x1000 array would require 16MB.
In addition to this memory requirements, a relatively small amount of
memory is required for other Steady-State Solver and modeling
activities. AFT Impulse itself requires approximately 15MB of RAM to
run.
Window 95, 98 and ME are limited to displaying a maximum of
approximately 350-400 junction icons as a result of the way in which
these operating systems handle graphics resources. Windows NT, 2000
and XP, on the other hand, has no such limit. One can determine the
amount of free graphics resources by selecting Help from within AFT
Impulse, then clicking on About AFT Impulse and then System Info.
Among the information displayed will be GDI Memory Available
(Graphical Device Interface). As this value approaches zero, Windows
95, 98 and Me will become unstable and, ultimately, crash.
It is important to note that this Windows 95, 98 and ME limitation does
not limit the number of junctions or, as a result, the model size, since
junctions may also be displayed as a solid box, an outline box or not at
all (the junctions are still there even if not displayed, and may be edited
Installation Issues
Customization files
As you customize AFT Impulse, two new files are created in your
application data folder (e.g., C:\Documents and Settings\Your
Account\Application Data\Applied Flow Technology\AFT Impulse\4.0):
IMPULSE4.INI
IMP_USER4.DAT
The IMPULSE4.INI file contains the information to customize the
features in AFT Impulse. The IMP_USER4.DAT contains all custom
database information. If you invest significant time in customizing AFT
Impulse, especially in the area of custom databases, you are strongly
advised to keep backup copies of these files.
Network installations
AFT Impulse 4.0 requires a number of auxiliary files to run properly.
Because of how 32-bit Windows operating systems work (Windows 9x,
NT/2000/XP), certain files must be registered in the Windows registry.
For software installed on a client PC this is accomplished directly during
setup. For software installed on a local or wide area network, however,
the installation becomes two-step process.
The first step is to install AFT Impulse on the network using the
“Network (auxiliary files on client)” in the setup program. This performs
a typical network installation except the auxiliary files are not installed
or registered.
Contacting AFT
Telephone support
Applied Flow Technology can be reached at:
(719) 686-1000 (Voice)
(719) 686-1001 (FAX)
Web site
You can download the latest maintenance releases of Impulse, find out
what new things are happening at AFT, and get the latest information by
visiting our website at:
http://www.aft.com
E-Mail support
AFT can also be reached by E-mail at:
support@aft.com or www.aft.com
Mail support
You can send mail to AFT at:
Applied Flow Technology
P.O. Box 6358
Woodland Park, CO 80866-6358 USA
You can send courier mail to AFT at:
Applied Flow Technology
400 W. Hwy. 24, Suite 201
Woodland Park, CO 80863 USA
for pumps 365, 391, 395 Animate Using Solver See Animation
H theory 387
Initial guesses
Hazen-Williams 202, 338
flow guess 205
H-Equation method 325
generation 218
HGL 108, 326
pressure guess 217
Hide Empty Columns
using for troubleshooting models
Model Data window 90
451
Highlight feature 30, 133, 194
Inspection feature 31, 134, 210, 212
Homologous pumps See Affinity Laws
Showing only undefined items 480
Hydraulic Grade Line See HGL
Installation 485
hydraulic transients
Network installations 486
See waterhammer
S in Output window 93
Visual Report Control window 116, wavespeed 38, 42, 64, 66, 152, 153,
304 203, 368, 370, 377, 378, 424