Beruflich Dokumente
Kultur Dokumente
Chapter sections
1.1 Starting WaveBuild
1.2 Setting General Parameters
1.3 Creating a Simulation Title
From the Start menu select Programs > Ricardo > WAVE > WaveBuild 8.0
Linux/UNIX:
Once the application opens, the WaveBuild GUI should have a blank canvas and appear as in Figure
1.
The title bar is across the top of the WaveBuild GUI window. It lists the name of the currently open
file (NoName.wvm by default, until renamed). If changes have been made to the file and haven't yet
been saved, an asterisk (*) will appear next to the filename. See Figures 2 and 3.
The pull-down menus are listed across the top of the GUI, beneath the title bar (see Figure
4). Clicking these with the left mouse button will open up menus with selection items.
Linux/UNIX users: The Help pull-down menu will appear at the far right side of the window.
Open
Save
Undo
Redo
Find
Select
Place Element
Cut
Copy
Paste
Constants Table
WavePost
WNOISE
Text Editor
Opens the Text Editor specified in the Tools > Options tab.
Favorite Application
The canvas is the main portion of the WaveBuild window and where the flow network will be
assembled and described (see Figure 5).
The canvas is a customizable surface. You can customize the properties for the WaveBuild canvas
by selecting the Canvas Properties... option from the Edit pull-down menu or by right-clicking on the
canvas background and selecting the Edit Canvas Properties... context menu item. Customizable
options include canvas size, annotation display, and grid appearance.
With any new model, the first step should always be to define the general parameters for the
simulation, specifically the units system to initialize all data entry. Open the General Parameters
Panel by selecting General Parameters... under the Simulation pull-down menu.
For this tutorial, all units will be defined in the SI system with mm as the basic unit of length. Select SI
[mm] from the Units option menu. Wherever an input is required in WaveBuild, the units for that
input will be displayed next to the entry field. When a numeric input is provided, if it is in another unit
than the initialized unit, the user can change the units for entry. Changing the selection in this option
menu at a later point will not convert previously-entered values between units systems!
Under the Start Options section of the panel, note the toggle button next to Reinitialize Flowfield
Between Cases. Turning this on will cause WAVE to start from the user-imposed initial conditions for
wall temperatures, gas temperatures, pressures, velocities, and species concentrations in every
subsequent case when multiple cases are defined (Case #1 is unaffected by this setting). This option
is turned off by default because in most instances the final conditions for a converged case are closer
to the final conditions of the following case than the user-imposed initial conditions. Assuming this is
true, each case will converge to the final solution quicker. For this tutorial, the toggle button should
remain off.
In the General Parameters section of the panel, type 30 in the Simulation Duration text field. We
will be simulating an engine and this number will define the number of engine cycles to
simulate. Since we don't know how many cycles our simulation will take to converge, we will set a
typical number for a gasoline engine and note the convergence behavior at run time. If our number is
higher than required, auto-convergence will stop the simulation.
Note that the background of the Simulation Duration field turns yellow. This is because the units are
listed as s for seconds and 30 seconds far exceeds the recommended value for simulation duration in
the intelligent defaults settings. WAVE, by its nature, simulates fluid flow using a timebase of
seconds. When a cyclic process is introduced to the simulation (such as an engine cylinder, an
oscillating flow source, etc.) WaveBuild knows to change to a cyclic timebase. Once we place the
engine cylinders on the canvas, this units field will automatically change to cycles and the background
of the text field will change to white. "Behind the scenes" the WAVE solver will still be solving
everything in a seconds timebase, but user entry fields are changed to cycles for convenience.
Under the Fuel and Air Properties section of the panel, click on the file tag button
to open the
Fuel Property Tag Selector and double-click on INDOLENE (see Figure 7). This will automatically fill
in the word INDOLENE in the text field. Similarly, typing in the word directly into the field indicates
the use of a tag. To specify a particular file not aliased in the tags file, use the file browser
button
to open a file browsing window and select the properties file to use. This will fill the text
field with an absolute file reference, using pointed brackets, <>, to surround the entry, denoting a
filename.
In WAVE, five species are used and transported throughout the model. These are Fresh Air,
Vaporized Fuel, Burned Air, Burned Fuel, and Liquid Fuel; numbered Species 1 to 5 respectively. All
properties for these species are defined in a single properties file (with a .fue extension) that
contains information on how the fuel and air react at different temperatures, pressures, and
concentrations. The appropriate properties file should be selected for the fuel being used in the
model. For simulations using only air, any properties file may be specified since all contain
information on the properties of air. Ricardo ships numerous pre-made properties files and all are
selectable from the default tags list. The user can create their own file if desired by clicking on
the Create Properties File button to use the pre-processor.
Note the End of Cycle Angle text field. This field is used to specify the crank angle that WAVE will
use to denote the start/end of subsequent engine cycles. By default (setting of auto), WAVE uses
the IVC of cylinder #1 as the end of cycle. This is acceptable for almost all engine simulations except
when VVT is employed and continuity between cases is important. In that case, a crank angle can be
specified and should be just after the latest IVC for cylinder #1 in the simulation. For this tutorial, the
default setting of auto is appropriate.
When finished, the General Parameters should appear as in Figure 8.
Click on the Convergence tab. This tab allows the user to activate the auto-convergence
mechanism, which tells WAVE to end the simulation if it converges to within the specified tolerance
value for the specified number of consecutive cycles (default of 1% for 1 cycle) regardless of whether
or not it has reached the specified duration. Turning this option off will cause WAVE to run to the full
specified duration for each case in the simulation.
For all simulations, velocity (U) and pressure (P) are checked for convergence in every computational
cell in the last timestep of each cycle. In simulations where WAVE's structural conduction model is
implemented, the calculated wall temperature (T) is also checked for convergence. In simulations
where WAVE's control systems are used, PID controller tolerances are also included in the
convergence calculation. Finally, the user has the option of adding any cycle-averaged quantities
desired to be included as additional criteria.
For this tutorial, the toggle button should remain on and the default values are acceptable. When
finished, the Convergence tab should appear as in Figure 9.
Press the OK button to close the General Parameters Panel and save the settings.
The first time the file is saved, you will be prompted for a filename and directory where the file will be
written (the file is saved with a .wvm extension, for "WAVE model"). You will also be prompted to add
comments for the file (this option can be deactivated via the File tab on the Options Panel located
under the Tools... pull-down menu). When the save button is subsequently clicked, or if the file is not
a newly created file (opened from a pre-existing file), the currently open file will be overwritten. To
save the file elsewhere or with a different name, select the Save As... option from the File pull-down
menu. Save this file with a descriptive name, such as tut_si1.wvm.
Show
Chapter sections
Move the mouse over the Ambient junction icon listed in the
Elements Tree. Hold down the left mouse button and drag it
onto the WaveBuild Canvas, dropping it anywhere. This will
place one junction of the simple ambient type onto the model
canvas (see the WaveBuild HELP on how to work with the
Elements Tab). To place additional junctions without
continually dragging and dropping, click on the place
element + button in the toolbar across the top of the WaveBuild
GUI. The mouse pointer changes to a "+" sign when the
mouse is moved over the WaveBuild canvas.
The mouse pointer is now in junction placement mode. In
junction placement mode, a junction of the selected type will be placed on the canvas every time you
press any mouse button. The mouse will remain in junction placement mode until you click on the
Select button on the toolbar across the top of the WaveBuild GUI or hit the Esc key to return to
select mode, at which point the mouse pointer will return to appearing as an arrow when moved over
the canvas.
Place two of the Ambient junctions on the canvas approximately 12 grid squares apart
horizontally. Note that when placed on the canvas, junctions "snap" to the nearest grid intersection
point. This behavior can be modified via the Edit > Canvas Properties... menu item.
Drag and drop two Orifice elements from the Elements Tree on the canvas between the ambient
junctions. Finally, drag and drop an Engine Cylinder element
from the Elements Tree between the
Orifice junctions. When finished the WaveBuild canvas should appear as in Figure 2.
This overwrites the .wvm file with the new information and backs up the previous .wvm file to a file
with the extension .wvm_bak1. WaveBuild will create up to 9 backup files with the .bak extension,
all of which can be opened and viewed using WaveBuild.
Show
Chapter sections
3.1 Defining Ambients
3.2 Defining Ducts
3.3 Defining Orifices
Double-click with the left mouse button on the duct labeled duct1 to open the Duct Panel. Across the
top of the Duct Panel is the Duct ID (name for the duct as displayed on the canvas and in the output
files) and, in the Connectivity section, the junctions connected to the duct's Left and Right ends are
listed.
The middle of the Duct Panel is the Schematic section which will update the drawing of the duct to
reflect the entered geometric dimensions (Bend Angle is not drawn in the schematic as it has no
physical meaning in a 1-D model, it is simply used to calculate a pressure loss based on the angle
and bend radius). Fuel Injectors, Spray Impingement Points, and Thermocouples will also appear in
the schematic if present in the duct.
The bottom of the Duct Panel is reserved for two layers of tabs which hold entry fields in which the
user will specify all information relevant to that duct. Of all of those tabs, there are three of primary
importance that we will use in this tutorial. Under the top level tab of Duct Data, the Dimensions tab,
the Coefficients tab, and the Initial Conditions tab are particularly relevant in all ducts and are the
only tabs we will be editing during the course of this tutorial (see Figure 3).
On the Dimensions tab, with the Shape selected as Circular, type the dimensions given
for duct1 in Figure 2 into the appropriate entry fields (Left and Right Diameters and Overall
Length). The engine bore for this example is 78.1 [mm] so, following the general recommendation
for discretizing the model for performance simulation, set the Discretization Length to 35 [mm] (click
here to read a sidebar on Discretization). When completed, theDimensions tab for duct1 should
appear as in Figure 4.
On the Coefficients tab, type 0 (zero) in the Friction and 1.5 in the Heat Transfer coefficient fields
under the Coefficients section of the tab (click here to read a sidebar on Port Flow Testing to
understand why). The default setting of 0.0 for the Pressure Loss Coefficient is suitable for this
simulation. The Left and Right end Discharge Coefficients will be automatically calculated by
WAVE when left as auto based on the diameter of the duct at the relevant end and the diameter of
the neighboring duct/orifice. If desired, this can be overridden by typing in a value for the Discharge
Coefficient from 0.0 to 1.0. For the purpose of this tutorial, the default value of auto is suitable and
should not be changed. When finished, the Coefficients tab should appear as in Figure 5.
Note that the Friction and Heat Transfer Coefficients have default values of 1.0. These values
are multipliers for the standard calculation. Thus values of 0.0 imply that there is no pressure
loss due to friction and no heat transfer occurring along the length of the duct while values of 2.0
imply that twice the standard pressure loss due to friction and twice the standard heat transfer is
occurring. These multipliers may be used as "tuning knobs" to adjust friction and heat transfer
and should be changed according to the surface roughness of the material and flow conditions in
the duct. Keep in mind that surface roughness will affect BOTH of these parameters and that
pressure loss due to increased heat transfer can be much greater (expansion/contraction of the
gas) than pressure loss due to friction!
On the Initial Conditions tab, type the conditions given for duct1 in Figure 2 into the appropriate
entry fields (Pressure, Temperature, and Wall Temperature). In the case of duct1, all of the default
settings are correct for our model. When finished, the Initial Conditions tab should appear as
in Figure 6.
Note that initial conditions of Pressure and Temperature need not be extremely accurate in an
engine simulation as the gas will quickly move through the system and conditions will be recalculated frequently. The initial conditions as set will be flushed out within the first few engine
cycles of the simulation. Default settings of 1.0 [bar] and 300 [K] are appropriate for most engine
Now double-click on the orif1 junction and note the only editable fields for an orifice junction are the
ID of that junction and the junction Diameter (diameter of the orifice plate) if one exists. If an orifice
junction is simply used to join two duct ends together, than the default setting of auto is
sufficient. The auto setting implies that the diameter of the orifice is equal to that of the smaller of the
two attached ducts. If a restriction is to be modeled, a Diameter smaller than the smallest of the two
attached ducts should be used. Type 20 [mm] in the Diameter text field and hit the OK button. Note
the orif1 junction now appears as if an orifice plate is reducing the diameter at the junction.
Click the Undo button in the toolbar
to return the Junction Diameter to auto again. For this
tutorial all orifice junctions will have a Diameter of auto unless otherwise specified.
An orifice diameter can never be larger than the diameter of the smallest attached duct as it would
have no physical meaning, however a value of 0 (zero) makes the junction behave like a closed
seal between the attached ducts.
When completed, the WaveBuild canvas should appear as in Figure 7.
Show
Chapter sections
4.1 The Geometry Tab
4.2 The Operating
Parameters Tab
4.3 The Heat Transfer Tab
There are four primary tabs that are important for every engine. These are Geometry, Operating
Parameters, Heat Transfer, and Combustion.
On
the Geometry tab,
under
the Configuration section enter the
relevant data for this engine as shown
in Table 1.
The No.
of
Cylinders field
does not automatically update when
cylinders are placed on the screen
(this allows for definition of the engine
before the model is built). The user
must enter the correct number of
engine cylinders manually.
The Strokes per Cycle field is used
to define the engine as a 2-stroke or
4-stroke engine (optionally a 6- or 8stroke engine for research purposes).
The Engine Type field is used to
enable different combustion and
emission models when either SI or
Diesel are selected. The Motored
option allows for engine motoring
simulation (engine pumping only, no
combustion via cancelling all injection
events).
On
the Geometry tab,
under
the Friction
Correlation section
enter the relevant data for this engine
as shown in Table 2.
These coefficients are used in the
Chen-Flynn friction correlation model.
This model is used to calculate the
FMEP
(Friction Mean Effective Pressure) for
the engine. When data is collected in
the test cell, it can be plotted and
correlated using the Chen-Flynn
model so that FMEP may be
calculated at non-tested engine
speed/load conditions. The equation
to calculate FMEP in WAVE is:
No. of Cylinders
Engine Type
Spark Ignition
Bore
78.1 [mm]
Stroke
82.0 [mm]
Compression Ratio
150.0 [mm]
0.0 [mm]
10.0
2 [mm]
Clearance Height
ACF
0.35 [bar]
BCF
0.005
CCF
400 [Pa/min*m]
QCF
0.2 [Pa/min2*m2]
Once the SPEED constant has been defined, the background of the Engine Speed text entry field
should turn white to denote an acceptable value. Also note that the simulation title at the top of the
WaveBuild canvas has updated to reflect the value of the SPEED constant as given in the title. Hover
the mouse over the Engine Speed entry in the text field to pop up the tool tip and note that the
constant is now being evaluated correctly as in Figure 4.
The Reference Pressure and Reference Temperature fields can be left as their default values
of 1.0 [bar] and 298 [K] respectively. These values are used in the calculation of volumetric efficiency
for the engine and may or may not correspond to the ambient conditions of the dynamometer cell
when tests are performed (different companies use different practices). When completed,
the Operating Parameters tab should appear as in Figure 4.
Efficiency) and when the intake valves are closed (during compression/combustion/exhaust). Adding
swirl will increase the total heat transfer due to increased charge motion in the cylinder.
Modifications to the standard Woschni model have been added to compensate for varying levels of
IMEP. This can be selected by changing the Woschni Model from Original (1967) to Load
Compensating (1990).
For the purpose of this tutorial, enter the relevant data for this engine as shown in Table 3.
Woschni Model
Original
1.0
1.0
520 [K]
520 [K]
400 [K]
420 [K]
480 [K]
Swirl Ratio
0.0
Show
Chapter sections
.\examples\engine\TUT_si\tut_si1.wvm
Example Output File:
.\examples\engine\TUT_si\tut_si1.wps
The connections from duct2 and duct3 to the cyl1 junction are assumed to be valves (all engine
cylinder connections are assumed to be valves). The blue connection point to duct2 denotes an
intake valve and the red connection point to duct3 denotes an exhaust valve. This is correct for this
model but is purely coincidental. WaveBuild places engine cylinders on the screen with two
connections by default. It assumes that the first connection to be used is an intake valve and the
second connection to be used is an exhaust valve. Any other valves created (by double-clicking on
the cylinder junction and editing the Number of Valves field to be larger than 2) are assumed to be
intake valves. This intake/exhaust status can be changed by highlighting the connection of choice
and changing the Type in the drop-down menu (see Figure 1 for example of a modified Cylinder
Panel). For this tutorial, the two valves created by default are suitable and the number of valves
should not be changed.
Valves are defined globally, for use anywhere in the model. Anywhere a valve is present (on an
engine cylinder, in a valve-specific junction, etc.) the user must tell WAVE which globally defined
valve to use. This, for example, allows the user to define an intake valve once and then use it for all
engine cylinders.
To define a valve, select Valves... from the Model pull-down menu. The Valve List panel will pop
up. As it is blank, it is clear that no valves have yet been defined. Click on the Add button to create a
new valve. This will pop up the Add Valve Panel where the user selects which type of valve to
use. The standard valve on an engine cylinder is a Lift valve, selected by default). Click
the OK button to accept the Lift type and the Lift Valve Editor will automatically appear.
Click on the Edit Lift Profile button to open the Valve Lift Profile Editor.
In the Valve Lift Profile Editor, data must be entered for the behavior of the valve. This behavior is
described as the lift of the valve vs. time (time is entered as cam or crank angle degrees). There are
numerous options for entering this data including:
INTAKE VALVE
Diameter
Lift Profile
35 [mm]
SI1INT tag
Cycle Anchor
330 deg
Profile Anchor
0 [deg]
Duration Multiplier
Lift Multiplier
Lash
Rocker Ratio
Angle Type
Coefficient Profile
1.0
1.414
1.0
Crank
CFTYP tag
Click the OK button to save these setting and close the Valve Lift Profile Editor.
Similar to the Valve Lift Profile Editor, data must be entered for the coefficients. This behavior is
described as values for both the forward and reverse flowing direction (forward implies into the
cylinder, reverse implies out of the cylinder) vs. the lift of the valve (non-dimensionalized by dividing
the lift by the reference diameter). Again, there are numerous options for entering this data including:
For this tutorial the data has been provided already in a pre-formatted external file that is aliased in
the default active.tags file. To select this file, click on the tag button
and select
the CFTYP option. Notice that the array fills automatically by reading the contents of the file aliased
in the active.tags file and a coefficient profile appears in the plot on the right-hand side of the
panel, see Figure 4.
This file contains information for flow coefficients as we can tell by the fact that at zero lift (zero L/D
value) the coefficient has a value of 0 (zero). Flow coefficients must be entered with the first row in
the array being an L/D value of 0 (zero) and forward and reverse coefficient values of 0 (zero).
WAVE can also accept discharge coefficients as input. Discharge coefficients are distinguishable by
the fact that the coefficient value at zero lift (zero L/D value) are non-zero. If the first row of the array
is an L/D value of zero and the forward and reverse coefficients are non-zero, then WAVE assumes
that the array is entered as discharge coefficients. Click on the tag button
and select CDTYP to
display a typical discharge coefficient profile, seeFigure 5. These two profiles are derived from the
same data so either will work and is appropriate for this tutorial.
Click the OK button to save these setting and close the Profile Editor.
Click the OK button on the Lift-Valve Editor panel to save the settings for valve #1 and return to the
Valve List panel.
EXHAUST VALVE
Diameter
Lift Profile
Cycle Anchor
Profile Anchor
Duration
Multiplier
28 [mm]
SI1EXH tag
105 [deg]
0 [deg]
1.0
Rocker Ratio
Angle Type
Coefficient Profile
0
1.0
Crank
CFTYP tag
Chapter sections
6.1 Adding the Injector Element
6.2 Defining the Injector
Click on the Composition tab, where the total composition of the fuel before injection can be
specified. If the aforementioned charge cooling effect is undesirable, then the vapor portion can be
specified here. For this tutorial, the default of 1.0 for Liquid Fuel is suitable and can be left as is
(100% of the injected fuel is in liquid state, 20% of that vaporized when injected).
When completed, the Proportional Injector Panel should appear as in Figures 2-5, below. Click
the OK button to close the Proportional Injector Panel and save the data (when prompted to add
the A_F constant to the Constants Panel, select No).
All that remains is to add the A_F constant to the Constants Panel. Open the Constants Panel
and
type the constant name, A_F, in the Name column, + row. This will automatically add a new row to
the constant panel. There are no units required and the value of the constant in case 1 should be
entered as 14.7 (approximately stoichiometric for the INDOLENE fuel that this simulation is
using). When finished, the Constant Panel should appear as inFigure 6, below.
Phase 2 Steps
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si1.wvm
Example Output File:
3 Requesting Post.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si1.wps
Processing Datasets
4 Running
WAVE
and
Understanding the .out File
5 Introduction to Post
Processing and WavePost
With these five steps, we will run the WAVE solver and create some output data for postprocessing. WavePost will be used to view plots and cycle-averaged results for the model.
Show
Chapter sections
2.1 Duct Time Plots
2.2 Junction Time Plots
2.3 System Time Plots
2.4 Multiple Plots Overlaid
Suppose we are interested in observing the pressure and temperature at the mid-point of the intake
port during the engine cycle. We can request Time Plots at the intake port (duct2) and WAVE will
automatically create these plots at the end of the simulation. Right-click on duct2 and select the Edit
Plots... option from the pop-up menu (see Figure 1). The Duct Plot Panel will open with no plots yet
defined at this location. Click on the Create Plot button to open the Duct Plot List.
The list of plots available at this location is displayed (this list is context-sensitive and generated
based on the canvas item selected). Click on the 201 Pressure plot and then, holding the Shift
button to multiple-select, click on the202 Temperature plot (most valid duct plots are in the 2xx
range). Click on the OK button to close the Duct Plot List and add these plots to the Existing Plots list
in the Duct Plot Panel.
The location along the length of the duct from where plots should take their data is specified under the
Duct Locations region of the panel. Locations are defined by a normalized value from zero to one
(left end of duct is 0, middle of duct is 0.5, right end of duct is 1). Default Locations can be defined for
commonly-used locations to be specified in multiple plots (Default Locations will appear identically for
all ducts when the Duct Plot Panel is opened). Custom Locations are specific locations in the
selected duct where plots should take their data. Clicking on the Use All Locations button will
automatically create Custom Locations for every cell in a duct. To add locations (either Default or
Custom), right-click on the column header and select Insert Column Before or Insert Column After, as
appropriate, from the context-menu.
Pressure and Temperature are scalar values and, in WAVE, are stored at the center of a given cell
(vector values, such as velocity, are stored at cell boundaries). As there is only one cell in duct2, no
location needs to be specified. The time plots will automatically be created at the 0.5 location (halfway along the duct). Specifying any other location along the length of the duct will create identical
plots.
When complete, the Duct Plot Panel should appear as in Figure 2.
Click on the OK button to close the Duct Plot Panel. Note the plot icon now hanging off of duct2. To
edit plots at this location again, simply double-click on this icon and the Duct Plot Panel will open.
Suppose we desired to examine the combustion performance in the power cylinder during the engine
cycle. We may want to create a P-V (Pressure vs. Volume) plot. Right-click on the cyl1 junction and
select the Edit Plots...option to open the Junction Plot Panel. The duct plots created above will be
listed in the Existing Plots list. Click on the Create Plot button to open the Junction Plot List.
Note that there are many more plots in the list than in the Duct Plot List (engine cylinder plots are
typically in the 1xx range). Click on the 111 Linear P-V Diagram and click the OK button to close the
list and add the plot to the Existing Plots list. Also note that an asterisk appears to the right of the plot
type for any plot with the currently selected canvas item as a location.
No location within the junction needs to be specified for the plot as there is only one calculation point
in a junction. When completed, the Junction Plot Panel should appear as in Figure 3, below. Click
the OK button to close the Junction Plot Panel.
Although there is only one cylinder in our engine, we will eventually be creating a 4-cylinder
engine. To observe the behavior of the engine itself (system of cylinders grouped together), click on
the Simulation pull-down menu and select the Time Plot... menu item. This will open the Time Plot
Panel where plots can be created for the engine system, as well as sensors, actuators, and pins (to
be discussed in more advanced tutorials).
The System Location Type should be active by default when the panel opens. Click on the Create
Plot button to open the Time Plot List for the engine system.
A few engine system specific plots are available and even fewer are applicable to our system as
modeled (system plots are typically in the 7xx range). Click on the 701 Engine Torque plot and click
the OK button to close the list.
Again, no location needs to be defined for system plots. When completed, the Time Plot Panel
should appear as in Figure 4. Click the OK button to close the Time Plot Panel. Note, there is no
display of the system plot on the canvas as there is not yet a canvas entity to which it can be
attached.
Perhaps we wish to examine the pressure and temperature in the exhaust port and compare it to the
conditions in the intake port. We could create separate plots for the duct representing the exhaust
port (duct3), but it would be more useful if the data for both ports were on the same plot. Right-click
on duct3 and select the Edit Plots... option from the pop-up menu. Note that the only plots in the
Existing Plots list are the pressure and temperature plots from duct2. This is because these plots are
allowed at duct3 as well (a P-V plot is not sensible in a duct, nor is engine torque).
With plot 201 highlighted, click on the Add Location button to plot the pressure at duct3 on the same
plot as duct2.
Do the same for plot 202 to add duct3 to the plot of temperature in duct2. Click on the Use All
Locations button to request the plots at the center of both cells in duct3 (locations of 0.25 and 0.75).
When completed, the Duct Plot Panel should appear as in Figure 5.
Figure 5: Duct plot panel with duct2 and duct3 plots overlaid
Click the OK button to close the Duct Plot Panel. The model should appear as in Figure 6 with all
time plots added.
Show
Chapter sections
For this simulation, select the VELOCITY and VOLUMETRIC_FLOW datasets. When completed, the
Postprocessing Output Panel should appear as in Figure 1.
Click on the OK button to save the changes and close the panel.
Show
Chapter sections
4.1 Running the Solver
4.2 The .out File
The model has been run through an input check and is all set to create the requested data for postprocessing. To run the model through the actual WAVE solver there are two options:
Run the model in screen mode by clicking on the Run Screen Mode button
in the toolbar
Run the model in batch mode by clicking on the Run Batch Mode button
in the toolbar
Screen mode runs the model at high priority while sending standard output to the screen. A shell will
open and the simulation will pass by, printing output in real-time for the user to examine (output is
also printed to the .out file).
Batch mode runs the model at reduced priority while sending standard output to the .out file. On
PC, a shell will open, noting the version of the solver being used, the window will minimize, and the
simulation will run "quietly" with no output to the screen. When the simulation is finished, the prompt
will return in the opened shell. On UNIX, the process is detached from the current shell (similar to
executing at the shell prompt with the "&") and run in the background. No indication of the process is
given. Status of the simulation can be determined using the "ps -ef" command.
If the model has been altered in any way and not yet saved (noted by the asterisk next to the filename
in the title bar), WaveBuild will prompt you to first save the file before running. Run the solver using
either method and close the shell (if necessary) when the simulation is complete.
DUCT: duct3
DUCT: duct3
Row (2) is a summary for Cylinder #1 during engine cycle 0, the start-up cycle (WAVE simulations, by
default, begin at IVC for cylinder #1 unless otherwise specified in the General Parameters
panel). Only Cylinder #1 will have results for engine cycle 0.
Row (3) is a summary of engine system performance for engine cycle 0.
Rows (4+) will be a summary of every individual cylinder in the system for each engine cycle followed
by a engine system summary for each cycle.
Starting at engine cycle 3, a line (8) denoting auto-convergence conditions is printed to the
output. When auto-convergence conditions are satisfied (14), if the convergence detection was
activated in the General Parameters panel, WAVE will run one more engine cycle and then finish the
case. This happens regardless of whether or not convergence conditions are satisfied in the following
engine cycle.
The column titles as labeled in row (1) relate to the lines summarizing individual cylinder
performance. They are, in order Cylinder Number, Engine Cycle (cumulative from start of case),
Timestep number (cumulative from start of case), Mass Airflow (kg/hr), Volumetric Efficiency, Exhaust
Port Temperature [K], Equivalence Ratio, IMEP [bar], PMEP [bar], Indicated Horsepower [hp],
Indicated Specific Fuel Consumption [g/kW*hr], Cylinder Pressure at IVC [bar], Cylinder Temperature
at IVC [K], and Trapped Fuel/Air Ratio at IVC.
ENG:
ENG
ENG:
ENG
ENG:
ENG
I***
ENG:
ENG
I***
ENG:
ENG
I***
ENG:
ENG
I***
PCYL
1.013
TCYL
301.1
FTR
0.00
1.575
381.0
0.606E
1.569
385.3
0.680E
1.569
385.8
0.680E
1.568
385.9
0.680E
1.569
385.9
0.680E
Next, note the information on elapsed time, number of timesteps, and limiting elements in the system.
The number of timesteps can be used to calculate the average timestep size in CA over the last
cycle (CA per cycle / #timesteps). Our engine is a 4-stroke engine so there are 720 in a single
engine cycle with 776 timesteps in the last engine cycle. This yields an average timestep size of
approximately 0.93 close to the default maximum timestep size of 1 set in the General
Parameters panel. Although there is no hard and fast rule, typical, well-built WAVE models won't
have a timestep below 0.1. If the timestep is smaller than this, it is usually due to poor modeling
practice (extremely small element size within the model).
The limiting element information can be useful in finding an unreasonably small Discretization Length
or Overall Length (for a duct) or Volume (for a y-junction) that may be slowing the simulation down
dramatically. Note in this simulation that duct4 is the limiting element, with sub-volumes 1, 2, and 4
showing up in the list
ELAPSED TIME OUTPUT:
0.28 sec.
0
sec.
LIMITING ELEMENT
% STEPS (DOES NOT HAVE TO ADD TO 100)
------------------------------------------------------------DUCT/VOL: duct4/1
69.1 DUCT/VOL: duct4/2
11.3
DUCT/VOL: duct4/4
18.4
Finally, general result information is printed in tabular format, printable for records or reports. Results
included are Final Output of all elements in the system (ducts and junctions); Engine
Summary; Breathing Quantities; Fuel Burn Progress Summary, Engine Geometry, Operating
Conditions, Predicted Performance, and Engine Out Emissions; and Valve and Emissions specific
output. The output contained from this point forward varies for what is included in the model
(turbine/compressor, emissions modeling, etc.). Examine this portion regularly to grow more familiar
with WAVE output for different models!
-----------------------------------------------------------------------------------------------------------F I N A L
O U T P U T
O F
D U C T S
-----------------------------------------------------------------------------------------------------------Duct
Junction TWALL
TAV
PAV
PMAX
PMIN
UMAX
UMIN
MACH
FLOW
A
[K]
[K]
[bar]
[bar]
[bar]
[m/s]
[m/s] NUMBER
[kg/s] [cm^
duct1
Intake 300.0 299.3
0.989
1.049
0.929
110.0
-31.1
0.3195 0.02209 12.5
orif1 300.0 297.6
0.987
1.105
0.865
114.5
-23.0
0.3423 0.02209
9.6
duct2
orif1
cyl1
310.0
310.0
294.4
294.4
0.988
0.988
1.121
1.121
0.821
0.821
114.5
220.8
-23.0
-634.0
0.3423
1.0000
0.02209
0.02360
9.6
0.0
duct3
cyl1
orif2
400.0 1243.4
400.0 1143.8
0.978
0.935
1.402
1.337
0.706
0.596
808.9
859.4
-437.8
-23.5
1.0000
1.3088
0.02361
0.02361
0.0
6.1
duct4
orif2
Exhaust
400.0 1107.4
400.0 1160.6
0.924
0.995
1.248
1.084
0.507
0.927
859.4
389.1
-23.5
-17.9
1.3088
0.5595
0.02361
0.02362
6.1
12.5
E N G I N E
S U M M A R Y
-----------------------------------------------------------------------------------------------------------NC MASS IN
VOL.EFF. TRAP.RATIO
IMEP
PMEP
IHP
TEXH
RES(%)
EGR(%)
PHI
kg/hr
bar
bar
K
-----------------------------------------------------------------------------------------------------------1 79.56
0.9663
1.000
12.37
-1.210
32.58
1356.
3.460
0.000
0.9903
-----------------------------------------------------------------------------------------------------------NC DPMAX
TH_DPMAX
HTR
EVO
EVC
IVO
IVC
VOL.EFF.
COMBSTART IGNDEL
bar/deg
deg
kW
deg
deg
deg
deg
(PLENUM)
deg
deg
-----------------------------------------------------------------------------------------------------------1 2.044
0.9766E-03 10.34
105.0
405.0
330.0
610.0
0.9747
-24.35
0.000
-----------------------------------------------------------------------------------------------------------B R E A T H I N G
Q U A N T I T I E S
-----------------------------------------------------------------------------------------------------------AMB.VOL.EFF (AIR
IN / AMB. REF.) =
0.966
TRAP.RAT.
(FRESH TR./ FRESH IN ) =
1.000
DEL.EFF.
(FRESH IN / PLEN. REF.) =
1.020
SCAV.RAT.
(GAS
IN / GAS
TR. ) =
0.965
CHARG.EFF. (FRESH TR./ PLEN. REF.) =
1.020
SCAV.EFF.
(FRESH TR./ GAS
TR. ) =
0.965
TOT.DEL.EFF (GAS
IN / PLEN. REF.) =
1.020
RESID.FR.
(RESID TR./ GAS
TR. ) =
0.035
EGR FR.
(RESID IN / GAS
IN ) =
0.000
-----------------------------------------------------------------------------------------------------------F U E L
B U R N
P R O G R E S S
S U M M A R Y
------------------------------------------------------------------------------------------------------------
E N G I N E
G E O M E T R Y
---------------------------------------------------------------------------------------------------------I
DISPL./CYL.
[l] = 0.3928
I
NUMBER OF CYLINDERS =
1.000
I
I
[in^3] =
23.97
I
COMPRESSION RATIO
=
10.00
I
EFFECTIVE CR (VC-TDC
I
BORE
[mm] =
78.10
I
BORE/STROKE
= 0.9524
I
I
[in] =
3.075
I
CON. ROD LENGTH[mm] =
150.0
I
I
STROKE
[mm] =
82.00
I
WRIST PIN OFFSET[mm] =
0.000
I
I
[in] =
3.228
I
CLEARANCE VOL.[m^3] = 0.4365E-04 I
ENGINE TYPE
---------------------------------------------------------------------------------------------------------I
INT. VALVE DIA.[mm] =
35.00
I
EXH. VALVE DIA.[mm] =
28.00
I
#1 EVO [deg]
I
MAX. LIFT [mm] =
12.57
I
MAX. LIFT [mm] =
8.640
I
#1 EVC [deg]
I
I
I
#1 IVO [deg]
I
NO. INTAKE VALVES
=
1.000
I
NO. EXHAUST VALVES =
1.000
I
#1 IVC [deg]
---------------------------------------------------------------------------------------------------------O P E R A T I N G
C O N D I T I O N S
---------------------------------------------------------------------------------------------------------I
RPM
=
6000.
I
#1 INT.PORT PR[bar] = 0.9883
I
#1 IGN DELAY [deg
I
AMB. PRESSURE [bar] =
1.000
I
[inHg] =
29.27
I
#1 COMB. START[deg
I
[inHg] =
29.61
I
#1 INT.PORT TEMP[K] =
291.1
I
INJ.TIMING
[deg
I
AMB. TEMP.
[K] =
298.0
I
[degF] =
64.38
I
MID.INJ.PRESS.[bar
I
[degF] =
76.73
I
#1 EXH.PORT PR[bar] = 0.9780
I
[psi
I
FUEL TYPE (C:H:O) C =
7.300
I
[inHg] =
28.96
I
INJ.DURATION [deg
I
H =
13.90
I
I
FUEL RATE
[kg/hr
I
O =
0.000
I
PISTON VEL.
[m/s] =
16.40
I
(MULTI) [lbm/hr
I
FUEL LHV
[MJ/kg] =
43.18
I
[ft/min] =
3228.
I
#1 FUEL / SHOT [kg
I
[btu/lbm] = 0.1856E+05 I
I
I
(A/F) STOICH
=
14.56
I
I
1% FUEL PWR/CYL [W
I
FUEL MOLEC. WEIGHT =
101.7
I
I
---------------------------------------------------------------------------------------------------------P R E D I C T E D
P E R F O R M A N C E
(I D E A L
G A S E S)
---------------------------------------------------------------------------------------------------------I
INDIC. POWER
[hp] =
32.58
I
BRAKE POWER
[hp] =
27.93
I
PMEP
[bar
I
IND.EFFICIENCY [%] =
37.43
I
BRAKE EFFICIENCY[%] =
32.09
I
[psi
I
IMEP(NET)
[bar] =
12.37
I
BMEP
[bar] =
10.61
I
FMEP
[bar
I
[psi] =
179.4
I
[psi] =
153.8
I
[psi
I
ISFC
[kg/kW/hr] = 0.2228
I
BSFC
[kg/kW/hr] = 0.2598
I
IMEP(GROSS)
[bar
I
[lbm/hp/hr] = 0.3662
I
[lbm/hp/hr] = 0.4272
I
[psi
I
IND. TORQUE
[N*m] =
38.67
I
BRAKE TORQUE [N*m] =
33.15
I
FRICT. TORQUE [N*m
I
[ft*lbf] =
28.52
I
[ft*lbf] =
24.45
I
PUMP. TORQUE [N*m
I
ISAC
[kg/kW/hr] =
3.275
I
AUXILIARY POWER [hp] =
0.000
I
#1 EXHAUST TEMP [K
I
[lbm/hp/hr] =
5.383
I
[kW] =
0.000
I
[degF
I IND. ENERGY BALANCE:(MULTI)
I
FRESH AIR IN[kg/hr] =
79.56
I
#1 PMAX
[bar
I
NET PISTON WORK [%] =
37.43
I
(WET)
[lbm/hr] =
175.4
I
[psi
I
AVAIL.EXH.ENTH. [%] =
9999.
I
TRAPPING RATIO
=
1.000
I
#1 CA AT PMAX [deg
I
DEBIT INTK.ENTH.[%] =
9999.
I
VOL.EFF.(DELIVERED) = 0.9663
I #1 MAX DP/DTH [bar/deg
I
H. TRAN.(IN-CYL)[%] =
15.93
I
#1 VOL.EFF.(PLENUM) = 0.9747
I
[psi/deg
I
BLOWBY AT RING1 [%] =
9999.
I
A/F TRAPPED
=
14.70
I #1 CA AT MAX DP/DTH[de
I
IMBALANCE
[%] =
9999.
I
PHI TRAPPED
= 0.9903
I
#1 MAX AVG.GAS T[K
I
PUMPING WORK
[%] = -3.662
I
RESIDUAL FRAC. [%] =
3.460
I
[degF
I
I
BRAKE POWER
[kW] =
20.83
I
FRIC. (%FUEL ENER.
I
I
I
H.TRAN.(IN-CYL)[kW
---------------------------------------------------------------------------------------------------------E N G I N E
O U T
E M I S S I O N S
---------------------------------------------------------------------------------------------------------I
NOx
[ppm] =
0.000
I HC EMISSIONS(C1) [ppm] =
63.96
I
CO EMISSIONS [ppm
I
NOx AS NO2
[g/hr] =
0.000
I
[g/hr] =
2.622
I
[g/hr
I
(BRAKE SPC.)[g/kW/hr] =
0.000
I
(BRAKE SPC.)[g/kW/hr] = 0.1259
I
(BRAKE SPC.)[g/kW/hr
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------TITLE: SI Tutorial, 4-Cylinder Gasoline Engine at 6000 rpm
------------------------------------------------------------------------------------------------------------
Show
Chapter sections
5.1 The .sum File
5.2 The .wvd File
5.3 The .wps File
5.4 Time Plots in WavePost
5.5 Cycle Average Results
in WavePost
Launch WavePost from WaveBuild by clicking on the WavePost button in the toolbar
. The
WavePost GUI will open and automatically load the .wps created by WAVE (since we requested time
plots in WaveBuild). The network should appear in the main WavePost window identical to its
appearance in WaveBuild (see Figure 1).
The Results Frame in the lower right corner of WavePost shows all results available in the current
session file. Plots are categorized as Time Plots, Sweep Plots, Spatial Plots, or TCMAP
Plots. The Time Plots that we requested in WaveBuild have been automatically created and are
listed in the tree under the Time Plots folder. Double-click on the Pressure plot to see the result
(see Figure 2). Notice that this plot has a data line for the single element in duct2 as well as both of
the elements in duct3, where we decided to "Use All Locations".
Each plot can be individually opened and edited. Elements of the plot, such as the data line, the
axes, the title, etc. are all selectable and changeable. Individual data curves can be hidden through
the Curve Selector Panel (Tools > Curve Selector...). Data can also be Cut, Copied, and Pasted
between plots. A single plot can also be Cloned (File > Clone) to create an exact copy of a plot to
use as a template for new data.
Data can be added to a plot from the existing results (Add > Data...) or imported from an ASCII text
file, Ricardo SDF-formatted file, or, on PC, from an MS-Excel file (File > Import > Excel...,
ASCII..., or SDF...). Data on an existing plot can also be exported to an Excel file (on PC) or to an
ASCII text file by selecting the File > Export... pull-down menu item.
Plots can be printed directly to a printer or to an image file by clicking on the Print button
toolbar.
in the
The plots that were pre-requested in WaveBuild are already created and listed under the Time
Plots folder. But we also requested some Basic and Valve Datasets. This data has been stored in
the .wvd file and we can now create our own time plots using this data. We will create an overlay of
Valve Flow Coefficient (CF) and Valve Discharge Coefficient (CD) vs. non-dimensionalized valve lift
(L/D).
Right-click on the Time Plots folder and select the Add Time Plot... option a blank time plot
window will open. Select the Data... option from the Add pull-down menu to open the Time Data
Panel. In the Output Sets option menu, highlight the single set that is available
(named filename.wvd:Case 1). In the Independent Variable (X) section of the panel, select
the Custom option and then click on the Edit... button to open the X Axis Selector Panel. Highlight
the Junction Cyl1 Intake 1 option in the Elements option menu and select Valve Lift Over
Diameter in the Variables option menu. The X Axis Selector Panel should appear as in Figure 3,
below.
Click the OK button to save the selections and close the panel. Back in the Time Data Panel, under
the Elements option menu, highlight the Junction Cyl1 Intake 1 option and pick Valve Flow
Coefficient in the Variables option menu. When finished, the Time Data Panel should appear as
in Figure 4.
Click the OK button to save the settings and close the panel. The data curve for Flow Coefficient will
appear in the Time Plot and the plot title and axis labels will be automatically generated. Double-click
on the plot title and edit it to read "CF and CD". Double-click on the Y-axis and edit the label to read
"Valve Coefficient" (delete the word Flow). Double-click on the X-axis and edit the label to read
"L/D". Double-click on the plot frame (easiest to do at the top or right-edge of the plot) and click in
the Grid checkbox.
Highlight the data curve and, using the Copy
and Paste
toolbar buttons, paste a second data
curve on the same plot. Double-click on the second data curve in the legend to open the Curve
Panel. Click on the Edit Databutton and then click on the Modify Data Source button. Change the
Variable to VALVE:DISCHARGE_COEFFICIENT and click on the OK button to save the
change. Don't forget to update the Legend in the Curve Panel to reflect the change from flow to
discharge coefficient. Finally, in the "Plot" pull-down menu, change the option from "Cyclic" to
"XY". When finished, the Time Plot should appear as in Figure 5. Close the Time Plot and note that
a fifth plot is in the list under the Time Plots folder named "CF and CD".
Time plots can also be made quickly by right-clicking on an element in the flow network diagram and
selecting a variable to plot. Right-click on duct1 to create a time plot of Velocity at
location 0.0 (Figure 6) and right-click oncyl1 to create a time plot of Pressure (Figure 7).
Save this WavePost session file (.wps) by selecting the Save As... option from the File pull-down
menu. It is important to save a WavePost file under a different name when it has been edited (plots
created, changed, etc.) since WAVE will overwrite the file every time the solver is run. If you would
like to rerun the model file and still have access to the plots and network displays you've just created
save the .wps file to the same directory with a new name.
Show
Phase 3 Steps
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si1.wvm
Example Input File:
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4.wvm
Example Output File:
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4.wps
With these five steps, we will create a parameterized 4-cylinder model that runs un-throttled over a
range of 1000-6000 rpm. We will then observe the behavior of the engine over the speed range by
creating Sweep Plots in WavePost.
Show
Chapter sections
The
single-cylinder
model
currently
has
Ambient
junctions
at
both
ends,
named Intake and Exhaust. We don't need to include these when replicating the engine
network. These ambient junctions can be detached and moved aside for the time being.
Middle-click on the end of duct1 that is attached to the Intake ambient. Holding the middle-mouse
button down, drag the end of duct1 out onto the canvas and release the mouse button. This will
detach the duct from the ambient. Repeat for the end of duct4 attached to the Exhaust ambient.
Using the middle mouse button, drag the Intake ambient further to the left, to move it out of the
way. Repeat for the Exhaust ambient, moving it further to the right. When finished, the model should
appear as in Figure 1.
The engine-cylinder network, with the attached ducts representing the intake and exhaust ports, is
identical for all four cylinders in the engine. We can therefore copy and paste the existing singlecylinder network three times to create the other cylinders in our 4-cylinder engine. Using the left
mouse button, draw a box around the ducts and engine-cylinder network to select the entire system
(be careful not to draw the box around the ambient junctions). All of the selected items within the box
will be highlighted in red. Click on the Copy button in the toolbar . Then click on the Paste
button
in the toolbar and the mouse pointer will become a crosshair icon. Click on the canvas
beneath the cyl1 junction and a duplicate network will be created. Click on the Paste button
and
place the duplicate network two more times to create four identical duct/junction networks to
represent all four engine-cylinders. Note that the ducts and junctions have all been numbered
sequentially.
Plot requests are not duplicated, thus no plots are dangling off of any of the newly created
ducts/cylinders (feel free to request new plots if desired). To hide the plots that currently exist on the
network, right-click anywhere in the white canvas area and select the Edit Canvas
Properties... option from the menu to open the Canvas Properties panel. De-select the Plots toggle
button in the Annotations section of the panel (see Figure 2).
This will simply hide the plot icons and not draw them on the canvas. When finished, the model
should appear as in Figure 3.
Adding an Engine Block icon to the canvas will provide a clickable and selectable object, enabling
direct access to the Engine General Panel (previously accessed through the Model pull-down menu
or an engine cylinder junction) and also allow for right-click selection to add plots and other
attachments (sensors and actuators to be discussed in later tutorials).
Select the Create Engine... option from the Tools pull-down menu to open the Create Engine
Panel. This panel displays currently entered geometric values from the Engine General
Panel. Currently, there is only one cylinder in the engine -- this must be updated to reflect that three
new cylinders have been added. Change the No. of Cylinders text field to 4 and press
the Enter key. This will update the Preview of the engine block on the right as well as the Firing
Order table at the bottom. The Firing Order table will automatically calculate the TDC (top dead
center time) for each cylinder based upon the No. of Cylinders value and the Strokes per
Cycle selection (TDCs are calculated for even firing intervals and are relative to the previous cylinder,
with the first firing cylinder at crank-angle 0). Change the Firing Order to reflect that of a standard 4cylinder engine 1, 3, 4, 2. The default spacing of the cylinder TDCs will be appropriate for this
tutorial. When completed, the Create Engine Panel should appear as in Figure 4.
Click the OK button to close the panel and note the Engine Block icon that is added to the canvas.
When the Engine Block is created, it will have four Engine Cylinder junctions created along with it
by default. Left-click each of these newly-created Engine Cylinders one at a time (they will highlight
in red) and press the Delkey to delete them, leaving an empty Engine Block icon. Move the Engine
Block icon over to the Engine Cylinders that are currently on the canvas by middle-clicking on the
icon and dragging it. The existing Engine Cylinderjunctions can be "dropped" into the icon by
middle-clicking on them, one at a time, and placing them over the cylinder place-holders on the
icon. They will snap into place on the icon and be associated with the icon from that point on. The
bore-spacing of the Engine Block icon can be adjusted by right-clicking on the icon and
selecting Appearance... from the menu. Default WaveBuild grid spacing is 40/square so, if the
cylinder junctions are placed 3 grid squares apart, use a spacing of 120 (see Figure 5).
Show
Chapter sections
Connect the dangling duct ends to the Simple Y-junction by dragging and dropping (use the
middle-mouse button) anywhere on the blue portion of the junction. Any duct end dropped onto the
blue portion of the junction will create its own connection point automatically. Dropping the dangling
duct end on the existing connection point, , will occupy that connection point, leaving no starting
point to draw a duct away from the Y-junction. If there are no connection points on a Yjunction and a duct must start at that junction and be drawn away from it (to follow the Left to Right
convention), simply left-click on the blue portion and drag a duct away from the Y-junction.
Create a new duct between the Intake ambient and the Simple Y-junction on the intake
side.
Enter 50 [mm]
for
both Left and Right
Diameters and 500 [mm]
for Overall
Length. The Discretization Length should be 35 [mm], as used earlier in the single-cylinder
model. The default initial conditions are suitable for this duct.
Create another new duct between the Simple Y-junction on the exhaust side and
the Exhaust ambient (following the Left to Right convention). Enter 50 [mm] for both Left and Right
Diameters and 500 [mm]
for Overall
Length.
Enter 40 [mm]
for
the Discretization
Length.
Appropriate
initial
conditions
for
this
duct
should
be
set
as 1.05 [bar] Pressure, 700 [K] Temperature, and 650 [K] Wall Temperature. When completed, the
model should appear as in Figure 2.
The exhaust-side Simple Y-junction should be set up similarly, with a Diameter of 50 [mm] and
initial conditions similar to the outlet duct Pressure of 1.05 [bar], Temperature of 700 [K], and Wall
Temperature of 650 [K]. The orientation of the ducts should be similar to that of the layout on the
canvas, as in the intake-side Y-junction. When completed, the Simple Y-junction Panel and Openings
panel should appear as in Figure 5 and Figure 6below, respectively.
Show
Chapter sections
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4.wvm
Example Output File:
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4.wps
Make sure to return to Case #1 before continuing by either typing directly into the text field or using
the arrow selection buttons.
The convenience of using multiple cases is that Constant values can be changed from case to case.
For this tutorial, we will step from 6000 rpm down to 1000 rpm, using 1000 rpm increments to
simulate multiple steady-state test points in a speed sweep. Open the Constants Panel
and note
that there are now columns added to the table for every new case created. Also note that the values
from Case #1 are filled in for all of the new cases, using light-gray text. The light-gray text denotes
that the value is carried over from the previous case. If no new value is provided for a constant,
WAVE assumes that the value used in the previous case is suitable.
Select the Case #2 field for the SPEED constant and enter 5000. Enter 4000 - 1000 for Cases #3 #6, respectively. The Constants Panel should appear as in Figure 2. Click OK to close the panel
and save the settings.
When running a speed-sweep simulation, it is recommended to start at the high speed and move
downwards toward the low speed. This is because the WAVE solver is actually working in a timebase
of seconds and the solution tends to converge based on the number of repeated engine cycles. More
cycles can be completed at high RPM in a given amount of time in seconds than at low RPM. This
means a system running a high RPM will tend to finish quicker than at low RPM, thus any problems
with the general setup may be detected earlier when starting at a high RPM.
Typically, with a change in engine speed other parameters change as well, such as combustion
behavior and cylinder temperatures.
Double-click on the Engine Block icon to open the Engine General Panel. Click on
the Heat Transfer tab and enter {PISTON_TEMP}, {HEAD_TEMP}, {LINER_TEMP}, {IV_TEMP},
and {EV_TEMP} in the text fields for the Piston Top, Cylinder Head, Cylinder Liner, Intake Valve,
and Exhaust Valve temperatures, respectively. Click on the Apply button and a message will
appear as show in Figure 3.
WaveBuild has detected that new constants have been used but are not defined in the Constants
Panel. Click on the Yes button to open the Edit Constants panel and edit the profiles for these five
new constants. Enter the values as shown in Figure 4 These profiles describe the temperature of
the combustion chamber cooling slightly with a decrease in engine speed. Click OK to save these
constant profiles and close the Edit Constants panel.
Click on the Combustion tab and enter {CA50} in the Location of 50% Burn Point text field
and {BDUR} in the Combustion Duration (10-90%) text field. Click on the Apply button again to be
queried on adding these constants to the table. Select Yes from the Query window and enter the
profiles as given in Figure 5. These constants help to describe the shorter crank angle duration of
combustion and retarding of spark timing at lower engine speeds. Click OK to save these constant
profiles and close the Edit Constant panel.
Click the OK button to close the Engine General Panel.
Show
Chapter sections
4.1 Changes in the .out File
IHP
0.00
ISFC
0.0000
PCYL
1.013
TCYL
301.2
FTR
0.00
0.00 0.0000
0.00 0.0000
0.00 0.0000
-2.84 -2.2470
1.497
1.614
1.569
1.544
376.9
403.2
403.5
405.2
0.338E
0.597E
0.628E
0.613E
Also note that the same output created for the single-cylinder, run is created for each case in the
multi-case run.
At the beginning of the output for Case #2, note the warning statement:
W*** FLOW FIELD NOT REINITIALIZED AT START OF THIS CASE
This warning tells us that the initial conditions for Case #2 are set by using the final conditions
from Case #1 (controlled by the Reinitialize Flowfield Between Cases toggle button on the General
Parameters Panel). Note that there is no engine cycle 0 for cylinder 1 in any of the cases
following Case #1 since no re-initialization occurs. This practice usually saves the simulation a few
engine-cycles in each case as it is initialized using conditions that are most-likely closer to the
converged results than those specified by the user when setting up the simulation. For the six cases
in this simulation, the reduction in the number of cycles is noted in the table below, showing that a
total of 10 (35 - 25 = 10) fewer engine-cycles are performed if reinitialization is skipped in this
manner:
NUMBER OF CYCLES REQUIRED
Case #
Reinitialized
Flowfield
Non-reinitialized
Flowfield
Total
35
25
#
CASE: 2 ----------------------------------------------------------------------------------------------------TITLE= SI Tutorial, 4-Cylinder Gasoline Engine at 5000 rpm
Show
Chapter sections
5.1 Multi-case Handling in
WavePost
5.2 Creating Sweep Plots
There are two ways to distinguish which plot is from which case:
1. The plots are listed in order of case, from Case #1 upwards. In a run with a small number of
cases, it's easy to pick out the correct plot.
2. If there are too many plots listed to make this practical, then the plots can be filtered by case
number. Check the Show selected case box above the Plots folder tree, this will enable the
filtering by case number. Then, expand the results file list in the Results File folder tree and
select the desired case by left-clicking on the case icon
. This will filter the plots and only
display those with data from the selected case (see Figure 2).
Selecting a case in this way will also change the behavior of displayed results. Manually-created
Time Plots and cycle-averaged results or animations (discussed in Phase 4) will display results from
the selected case.
This will create a new Sweep Plot named Sweep Plot #1 by default. The plot is empty, with no data
displayed. To add data, select the Add pull-down menu and then select the Data... menu item to
open the Sweep Data Panel. This contains a list of all of the keywords in the .sum file that correspond
with a cycle-averaged numeric value. The list defining .sum file keywords and their associated
units is available in the WAVE Help section. Selectbhp for the Dependant Variable (Y) and
select rpm for the Independent Variable (X). The Sweep Data Panel should appear as in Figure
4 when completed. Click the OK button to apply.
WavePost will automatically label the axes (with units) by the .sum variables selected and name
the curve by the filename and keyword selected. Double-click on the axes or the curve to edit these
names. Double-click on the title and edit the name to Power vs. Engine Speed. A different (but
compatible by conversion) units system can be selected for the axis so that users can choose how
data is displayed without doing clumsy numerical conversions manually. When completed, the plot
should appear similar to Figure 5.
WavePost can also easily create Default Sweep Plots, where a summary quantity from an element in
the model can be plotted against a default X-Axis. The default X-Axis used can easily be set by
expanding the Defaults sub-folder under the Sweep Plots folder the first branch of the expanded
tree is the X-Axis Settings. Double-clicking on this branch will open the Default Sweep Plot X-Axis
Panel, as shown below in Figure 6.
If an engine exists in the model, "RPM" will be selected by default. Alternatively, you can select
"Case" or "Cycle" ("Time" is allowed in a non-cyclic run). Otherwise, you can also select any other
summary quantity by selecting the "Custom" option. Click the "OK" button to close the panel.
Right click on the engine block icon and create a sweep plot of torque (Add Sweep Plot >
Performance > Brake Torque vs. > RPM). The sweep plot will open and be added to the plots in
the results tree. Use this method to create a plot of volumetric efficiency as well. The completed
plots should appear as in Figures 7 and 8.
Click on the File pull-down menu and select the Save As... option to save the .wps session file under
a new name, such as Comparison.wps, so that it won't be overwritten by WAVE in the event that
the model file,tut_si4.wvm, is re-run.
Show
Phase 4 Steps
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4.wvm
Example Input File:
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4_intake.wvm
Example Output File:
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4_intake.wps
With these five steps, our model will have a complete air induction system that is representative of a
real engine. Each part of the system will be discussed and modeling practices examined. Animations
will be created in WavePost for reporting and data analysis.
Show
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4_intake.wvm
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4_intake.wps
Before performing any actions make sure the model is in Case #1.
The 4-cylinder model as created has all four intake ports joined together at a single Simple Yjunction. The intake ambient junction will be re-used upstream of the induction system as the
atmospheric conditions before the snorkel, but the Y-junction and duct created for the 4-cylinder
model can be selected and deleted. Highlight both the intake-side Simple Y-junction and
upstream duct (holding down the shift key and left-clicking) and hit the Delete key. This will leave all
four intake port ducts dangling with no junction to attach to upstream (see Figure 1). These will
eventually be attached to the ends of the runners from the intake manifold. For now, they can remain
unattached.
Click and hold the middle mouse button to move the ambient junction named Intake down to the
bottom of the canvas. This will serve as the starting point for the induction system.
The snorkel and zip tube can easily be modeled using only orifice junctions and ducts. A schematic
of the system with required dimensions is shown in Figure 2.
Place orifice junctions and connect with ducts as shown in Figure 3. In place of the air cleaner,
place two Complex Y-junctions and connect with a duct. Remember to follow the Left to Right
convention when creating the ducts.
To make the last duct in the zip tube (duct22 in Figure 3) appear bent on the screen, right-click on
the duct and select Add Control Point from the menu:
A control point can be selected with the middle mouse button and moved to make the duct appear
bent (it will snap to grid points just like junctions do). Multiple control points can be added to a
duct. Using two control points diagonal from each other by one grid square will create the
appearance of a smooth 90 bend, as shown in Figure 4.
Using the schematic from Figure 2, above, edit the ducts representing the snorkel and zip tube and
enter the appropriate Diameters and Lengths (leave the Complex Y-junctions and duct between
them for later). Remember to use 35 [mm] for the Discretization Length for all ducts in the intake
system and set initial conditions to be 1 [bar] Pressure, and 300 [K] for Initial Fluid and Wall
Temperatures.
Note that when setting the Left Diameter of the first duct to 70 [mm] and the Right
Diameter to 50 [mm], with Overall Length of 50 [mm], the Taper Angle field turns yellow, indicating
a warning that the calculated taper angle of11.3099 is outside of the recommended range for this
parameter (see Figure 5). Read this sidebar on modeling tapered ducts and note that we are
modeling upstream of the intake manifold, thus expect constant flow into the engine. Flow will always
be contracting and therefore the slightly high taper angle is not of concern.
Also note that when entering dimensions for the last duct in the zip tube, the 90 Bend Angle should
be included (see Figure 6). The Bend Angle field on the Dimensions tab of the Duct Panel allows
the user to specify how much of a bend occurs across the entire length of the duct. The pressure
drop due to this bend is also then distributed across the entire length of the duct.
Internally, the WAVE solver uses this Bend Angle, the average Diameter of the duct, and the duct
Overall Length to calculate a Cp value based on simple equations (see the WAVE Help section on
ducts for more details). If you would like to specify your own Cp value, simply set the Bend Angle to
0 and set the Pressure Loss coefficient on the Coefficients tab. The Cp value is then divided evenly
amongst the number of elements in the duct and then the pressure loss in each element is calculated
by:
Show
NOTE: This modeling method is only appropriate for performance simulations. Acoustic
simulations should model the air cleaner volumes in smaller sub-sections and the filter with
numerous perforates!
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4_intake.wvm
For our model, the two Complex Y-junctions will represent the two cylindrical volumes and the duct
connecting them will represent the filter. Double-click on yjun1 and enter the appropriate values as
given in Figure 2.
Note
the
auto-calculator
buttons
provided
for
both Volume and Heat
Transfer/Skin Friction Area fields. Clicking on either of these will assume a spherical shape
and
calculate Volume or Heat Transfer/Skin
Friction
Area accordingly
from
the Diameter value (this is identical to using a Simple Y-junction -- allowing the user to enter
their own value is what makes the Complex Y-junction flexible)!
Ducts
are
oriented
similarly
as
on
the
Simple
Y-junction.
The
orientation
of duct19 and duct20 should be used as shown in Figure 1. Note the new fields for each connected
duct: DELX, DIAB, and Thick.
DELX, sometimes referred to as the characteristic length, is the distance from the duct connection
point across the volume. See Figure 3 for a diagram of the DELX values for both Complex Yjunctions.
DIAB, sometimes referred to as the expansion diameter, is the equivalent diameter for the maximum
area that the gas can expand into, perpendicular to the duct entrance. See Figures 4 and 5 for
diagrams for the DIAB values for both Complex Y-junctions.
Thick is the orifice thickness and is used in acoustics simulations to calculate the acoustic end
correction. It is not necessary to set this value in performance simulations as it has no effect
whatsoever on the outcome.
When completed, the orientations for the Complex Y-junctions should appear as shown in Figures
6 and 7.
Show
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4_intake.wvm
Example Output File:
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4_intake.wps
The schematic of the throttle body geometry is shown in Figure 1. In WaveBuild, we will model this
geometry as two ducts connected in the middle using a throttle valve junction. The throttle valve
junction will represent the throttle blade.
Place the required junctions on the canvas and connect with ducts as shown in Figure 2, starting
with the final orifice junction of the zip tube. Edit both ducts to have Overall Lengths of 50 [mm]
and constant Diameters of 60[mm]. Remember to set the Discretization Length to 35 [mm] and
the Initial Conditions to the standard of 1 [bar] Pressure and 300 [K] Initial Fluid and Wall
Temperatures.
The throttle valve junction represents an instance of a butterfly valve type, as defined in the Model >
Valves list. In order to use it, we must first define the butterfly valve. Click on the Valves menu
item in the Model pull-down menu. Click on the Add button in the Valve List and select
the Butterfly valve type, then click on the OK button to open the Butterfly Valve Editor panel. Set the
Bore Diameter to 60 [mm] and the Shaft Diameter to 5 [mm], with a Minimum Plate Angle (angle at
which the valve sits effectively closed) of 5 [deg]. The Calculated Values section of the panel will
allow you to preview the calculated geometric quantities. Use the slider bar to observe the schematic
and values as the butterfly valve angle changes. When completed, the Butterfly Valve Editor panel
should appear as in Figure 3.
With the Forward/Reverse coefficient profile type radio button selected, click on the Edit Flow
Coefficient Profiles button to open the flow coefficient profile editor panel. If you have a defined
profile as a function of angle, it should be entered here. For the purposes of this simulation we will
define a coefficient of 0.5 at 5 deg and 1 at 85.22 deg, allowing a full sweep of realistic coefficient
values from when the valve is closed until it is fully open. When completed, the flow coefficient profile
editor should appear as in Figure 4.
Double-click on the throttle valve junction on the canvas to open the Throttle Valve Junction Panel
and edit the ID of the junction to THROTTLE. Select Valve Number 3 from the pull-down list and set
the Plate Angle to{THROTTLE_ANGLE} (see Figure 5). This parameterizes the throttle plate angle,
allowing us to set the angle in the Constants Panel, so it can change between cases, giving load
control during the simulation.
The background of the Plate Angle field will turn yellow as the constant THROTTLE_ANGLE is not
yet defined. Click the OK button to save the settings and close the panel and, when queried about
adding the constant to the constants table, click the OK button again and set the value 90 for all
cases (setting it in Case #1 only is sufficient, see Figure 6). Modeling the throttle fully open assumes
we are running a full-load speed sweep.
Show
WavePost
Step 4 - The Intake Manifold
In this step, we will model the simple log-plenum intake manifold using more complex yjunctions. The techniques for modeling a large volume with sub-volumes and for modeling a tapered
duct (bellmouth entry to intake runners) will be demonstrated.
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4_intake.wvm
Example Output File:
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4_intake.wps
The schematic of the intake manifold geometry is shown in Figure 1. In WaveBuild, we will model the
inlet pipe as a single duct. The plenum will be simplified and assumed cylindrical in shape (with an
equivalent diameter of 110 mm). Four Complex Y-junctions will be attached together using massless
ducts to represent the plenum volume. Three ducts will be used to represent each intake runner. A
discharge coefficient will be imposed for incoming flow to the first runner duct, to model the bellmouth
entry.
The equivalent network as created in WaveBuild is shown in Figure 2. Drag and drop the
required orifice and y-junction elements onto the canvas from the Elements Tree and connect them
with ducts accordingly. Remember to follow the Left to Right convention. Start the inlet pipe of the
intake manifold at the last orifice junction of the throttle body. Use four Complex Y-junctions to
model the intake manifold plenum. These four Y-junctions will be connected using massless ducts
(zero overall length).
Edit
the duct representing
the
inlet
to
the intake manifold.
Assign Left and Right
Diameters of 60 [mm] and an Overall Length of 150 [mm]. Remember to set the Discretization
Length to 35 [mm] and the Initial Conditions to1.0 [bar] Pressure and 300 [K] InitialFluid and Wall
Temperatures.
The Left and Right Diameters of the massless ducts should equal the equivalent diameter of the Yjunctions (which will also match the DIAB values assigned to the connections) so that there is no
pressure loss due to expansion or contraction. The Y-junctions have an equivalent diameter of 110
mm, so assign Left and Right Diameters of 110 [mm] to the massless ducts, along
with 0 [mm] Overall Lengths (see Figure 3). They will appear gray on the canvas, indicating they
are massless ducts.
The Y-junctions should be edited to have Diameter values of 110 [mm], Volume values
of 0.75e+006 [mm3] (3.0 L divided evenly by four), and Heat Transfer/Skin Friction Area values of
approximately 27300 [mm2] (pi*diameter*length) as in Figure 4.
Orient the duct connections according to the layout on the screen (see Figure 2). The DELX values
for the massless duct connections should be 79 [mm] (the length of each subvolume in the direction
of flow through the massless ducts) and the DIAB values should be 110 [mm] (equal to the massless
duct Diameters so no expansion or contraction occurs).
The DELX values for the runner connections should be 110 [mm] (distance across the subvolume in
the direction of flow into the runners). The DIAB values for the runner connections require some
thought. Should the area used to calculate the DIAB value for each runner connection be the
maximum area the gas can expand into in the Y-junction or the length of the entire plenum?
Technically, the DIAB value should be calculated from the maximum area the gas can expand
into along the length of the single y-junction into which the duct enters. This is because any losses
caused by flow traveling along the length of the plenum will be accounted for by mass transfer from
one y-junction to the next. For our geometry, the maximum area for expansion in the direction of flow
from the runners is equal to 110 mm * 79 mm 4 = 8690 mm 2. Thus DIAB is
approximately 105 [mm]. See Figure 5 for a representative duct orientation.
Note, the pressure and flow after a sudden expansion has a greatly diminished response the larger
the expansion is. Once the DIAB value is approximately twice the diameter of the entering duct, the
effect of the expansion changes very little with further increase in DIAB. Figure 6 illustrates this
effect.
Each runner, as shown in the schematic (see Figure 1), will be represented by three ducts. Edit
the ducts to have the geometry represented in the schematic and remember to set
the Discretization Length to 35 [mm] and theInitial Conditions to 1 [bar] Pressure and 300 [K]
Initial Fluid and Wall Temperatures.
The
middle duct for
each
runner
should
have
a 45 Bend Angle.
To edit multiple ducts at the same time (junctions can also be edited in this manner, in combination
with ducts if desired), multiple-select the items to be edited by holding the shift key and left-clicking on
the items or by drawing a box around the desired items while holding down the left mouse
button. With the desired ducts selected (highlighted in red) right-click on the white background of the
canvas and select the Edit Parameters... menu option. This will open the Duct Template
Panel (since all selected elements are ducts) allowing fields to be set for multiple ducts
simultaneously (see Figure 7 for example of setting duct geometry for all runner ducts). Only the
edited input fields will be set for all selected ducts.
If more than just ducts are selected (e.g. if the orifice elements between the ducts are also selected),
the Edit Parameters... context menu option will open the Parameters Panel, which allows you to set
input values for multiple elements at once, regardless of whether or not they are the same element
type (see Figure 8 for example of setting duct geometry for all runner ducts).
Figure 8: Edit Parameters Panel for setting parameters in all Runner Ducts
The checkbox next to each entry indicates whether the value will be written to each item. Since
multiple items are selected, no one item's given value will show up in the entry fields they will be
zero by default. The zero value will not be written to the items unless the checkbox is selected (or the
number zero is typed into the field).
The first duct is shown in the schematic to have an entry (Left end) Diameter of 70 [mm] and an exit
(Right end) Diameter of 40 [mm]. This creates a Taper Angle of approximately 8.5, higher than the
recommended limit of 7 as discussed in this sidebar on Modeling Tapered Ducts. Moreover, the
actual geometry doesn't have a constant taper it has a bellmouth entry (which has a rapid change in
cross-sectional area) at the beginning and then a gentle taper towards the duct end. This indicates
that flow will most likely separate from the duct wall during reverse flow, from the valve towards the
plenum, and a sudden expansion will occur creating a pressure loss.
To account for this physical phenomenon, both ends (Left and Right) of the first runner duct should
have Diameter values of 40 [mm]. This means that during reverse flow, from the runner duct to the
plenum Y-junction, it will experience a sudden expansion into the junction and the loss will be
accounted for. To accurately represent the effect of the bellmouth entry, the Disch. Coef. from the
plenum junction to the runner duct should be set to a representative high value, in the range of 0.95 1.0 in the Openings panel for each Y-junction (see Figure 9).
Connect the ends of the intake runners to the dangling port duct ends using a
simple orifice junction. If desired, use control points on the middle runner ducts to represent
the Bend Angle. When completed, the entire model should appear similar to Figure 10.
Show
WavePost can use time-based data to animate the network with color contours of basic
datasets. These animations can be viewed in WavePost and then saved as MPEG files for use in
presentations and reports.
In WaveBuild, open the Output and Plotting Panel by selecting the Output and Plotting... item from
the Simulation pull-down menu. Check the Generate Animation Data checkbox and click OK to
save the setting and close the panel (see Figure 1).
Run the WAVE model by clicking on the Run Screen Mode button in the toolbar
(if prompted to
save the model before running WAVE, click the OK button to save and run sequentially).
Launch WavePost from WaveBuild by left-clicking in the Launch WavePost button in the WaveBuild
toolbar . This will open the .wps file named similarly to the .wvm file currently loaded in
WaveBuild. Once WavePost is open, click the Open File button
in the toolbar and open
the .wps file created in Phase 3, Comparison.wps. This WavePost session file is currently only
referencing results from the last model created, without the intake system added on. It should contain
sweep plots created during the previous phase of torque, power, etc.
Click on the Add button at the bottom of the Output Files frame and select the newly
created .wvd file for the WAVE model containing the intake system. The .wvd and .sum files for this
WAVE run will be added to the Output Files list (see Figure 2 below). A Query window will pop-up
prompting whether to add curves to the existing plots using this file (see Figure 3). Click on
the Yes button and every existing plot will add data from the newly added.wvd and .sum files (if
matching data exists in the new files). Open the Sweep Plots to view the comparison of the
performance parameters as in Figure 4.
Note that addition of the intake system has changed the predicted performance results. Power is
decreased near 3000 rpm, but increased above 4000 rpm. The positive tuning effects are more
powerful than the losses due to friction, expansion and contraction, and bends that were added in the
intake system.
Use the buttons on the WavePost toolbar to play the animation. If desired, you can record the
animation to an MPEG file. Click on the record button to open the Movie Recordings Settings
Panel. Enter a name for the movie in the File text field (Velocity.mpg) and set the MPEG
Quality to High (this will use more disk space but create a clearer image) as in Figure 7. Click on
the OK button and a progress bar will pop up to show the progress of the rendering process. This will
take a while to complete.
When finished, a file with the name as given above will be created in the working directory. It should
appear similar to the animation below (right-click on the animation and select Play to repeat).
Proceed to Phase 5 - Adding the Exhaust System and Creating Spatial Plots in
WavePost
Sample Animation
Show
Phase 5 Steps
1 The Exhaust
Manifold
2 The Catalytic
Converter
3 The Resonator
(Silencer)
4 The Complex
Muffler
5 Creating
Spatial Plots in
WavePost
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4_intake.wvm
Example Input File:
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4_exhaust.wvm
ExampleOutput File:
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4_exhaust.wps
With these five steps, our model will have a complete exhaust system that is representative of a real
engine. Each part of the system will be discussed and modeling practices examined. Spatial Plots
will be created in WavePost for reporting and data analysis.
Show
Chapter
sections
Before performing any actions, make sure the model is in Case #1.
The 4-cylinder model with the intake system still has all four exhaust ports joined together at a
single Simple Y-junction. The ambient junction named Exhaust will be re-used downstream of the
exhaust system as the atmospheric conditions after the complex muffler, but the Y-junction and duct
created for the 4-cylinder model can be selected and deleted. Highlight both the exhaustside Simple Y-junction and downstream duct (holding down the Shift key and left-clicking) and hit
the Delete key. This will leave all four exhaust port ducts dangling (see Figure 1). These will
eventually be attached to the ends of the runners from the exhaust manifold.
Middle-click and drag the ambient junction named Exhaust down to the bottom of the canvas. This
will serve as the end point for the exhaust system.
The schematic of the exhaust manifold is shown in Figure 2. It can be easily modeled as
three ducts representing each exhaust runner (connected by orifice junctions) and a Complex Yjunction representing the collector. The collector is approximately spherical in shape, but
a Complex Y-junction should be used as it is more flexible for tuning purposes (Simple Y-junctions
are rarely used in real modeling!). It is extremely simple to make a Complex Y-junction behave
exactly like a Simple Y-junction by using all of the defaults for Volume, Heat Transfer/Skin Friction
Area, and DELX/DIAB settings.
Using the Elements Tree, place all of the required junctions and then draw the required ducts as
shown in Figure 3. Connect the dangling ducts from the exhaust ports to the newly
created orifice junctions. Remember to create a hanging duct leaving the collector Y-junction to
represent the connection with the close-coupled catalytic converter.
Define all the ducts using the values shown in the schematic. Remember to use 40 [mm] for
the Discretization
Length and
set
the Initial
Conditions to
a Pressure of 1.05 [bar], Fluid Temperature of 1100 [K], and Wall Temperature of 750 [K].
The hanging duct that will be connected to the close-coupled catalytic converter can be
assigned Left and Right Diameters of 76 [mm] and an Overall Length of 0 [mm] (i.e. it is
a massless duct). This will approximately model a direct connection between the collector volume
and the entry to the catalytic converter.
Define the Y-junction representing the collector using the values shown in the schematic. Once
the Diameter of 80 [mm] has been entered, click on the auto-calculate buttons
to fill in
the Volume and Heat Transfer/Skin Friction Area fields automatically. The Wall Temperature of
the collector should be higher than the runners. Assign similar initial conditions as the runners, but
use a Wall Temperature of 900 [K]. When completed, the Y-junction Panel for the collector should
appear as in Figure 4.
Orient the ducts as they are displayed in the schematic in Figure 1. Orient the massless
duct leaving the collector to be facing down, out of the plane of the runners, as shown in Figure 5.
Show
Chapter
sections
2.1 The Basic
Geometry
2.2 The Entry
and Exit Cones
2.3 The Brick as
a Catalyst Duct
The schematic of the catalytic converter is shown in Figure 1. In WaveBuild, the entry and exit cones
can be modeled using Complex Y-junctions and the catalyst brick can be modeled using a single
Catalyst Duct connecting the two junctions. Drag and drop the required elements onto the canvas
and connect them. Don't forget to attach the dangling end of the massless duct created with the
exhaust manifold to the Y-junction representing the entry cone. See Figure 2.
The entry and exit cones of the catalytic converter are modeled using complex Y-junctions in order to
account for the pressure loss due to sudden expansion of the gas. The taper angle in these cones
typically far exceeds the recommended maximum of 7, so using a duct to model the entry/exit cones
won't capture the pressure loss due to sudden expansion (see sidebar on Modeling Tapered Ducts).
Double-click on the entry and exit cone Y-junctions and edit their settings to reflect the information
given in the schematic. The Diameters provided are the area-weighted average diameters,
the Volumes and Heat Transfer/Skin Friction Areas are taken from solid-modeling software
(CAD).
Set
the Initial
Conditions to
be 1.05 [bar] Pressure, 1100 [K] Temperature,
and 900 [K] Wall Temperature (see Figure 3).
Orient the duct connections to be directly across from each other for each Y-junction. The DELX
values for both connections should be the length of the cone (50 [mm] for both junctions) while
the DIAB value can be left as the default value for now, identical to the junction diameter
setting: 90.2 [mm] for the entry cone and 76.4 [mm] for the exit cone. Set a Disch. Coeff. of 0.95 for
both the entrance and exit connections of the catalyst to account for the tapered cone geometry
(see Figure 4).
The catalyst brick itself is modeled using a single Catalyst Duct which will actually only represent one
channel through the brick but creates a number of identical channels to represent the entire
brick. The Catalyst Duct is simply an element which allows input of basic geometric parameters of a
the brick in order to describe the single flow channel.
Figure 5 shows the method used to calculate the diameter of a single catalyst brick channel and the
number of identical channels. It assumes that all channels are square in shape. Given the geometry
of the brick, the catalyst duct should be specified to have a Cross Sectional Area of 7854 [mm2],
a Perimeter of 314.59 [mm], a Cell Wall Thickness of 4 [mil], and a Cell Density of 600 [1/in2].
The Overall Length should be 80 [mm] as shown in Figure 1 and remember to set
the Discretization Length to 40 [mm]. Using the formula shown in Figure 5, the Cell Count should
be equal to 7304 and the Cell Diameter(single channel) should be 1.055 [mm], as shown in Figure 6.
Edit the neighboring Y-junctions representing the entry and exit cones and set the DIAB value for
the catalyst duct connection to be 1.17 [mm] (this represents the expansion diameter for a single
channel entering the volume).
A common practice when modeling Catalytic Converters in steady-state, if test data is available, is to
set the brick Wall Temperature to match the temperature of the gas on the downstream side of the
catalyst (from the test data) and then set the Heat Transfer Multiplier to a high value, such
as 5. This ensures that the gas temperature leaving the brick matches the calibration data.
Edit the duct settings and click on the Coefficients tab to set the Heat Transfer value to 5. On the
Initial Conditions tab, set the Pressure to 1.05 [bar], the Fluid Temperature to 1100 [K], and the Wall
Temperature to a constant named {CAT_TEMP}. When prompted to add CAT_TEMP to the
Constants Table, click Yes and enter the values as shown in Figure 7.
Show
Chapter
sections
3.1 The Basic
Geometry
3.2 The Down
Pipe
3.3 The
Resonator
The schematic of the down pipe (from the catalytic converter) and resonator is shown in Figure 1. In
WaveBuild, the down pipe can be modeled as two curved ducts and the resonator can be modeled as
entry and exit ducts with Complex Y-junctions representing the entire length of the inner, perforated
duct and the surrounding can.
Drag and drop two orifice elements onto the canvas as shown in Figure 2, to represent the ends of
the ducts used to model the downpipe. Draw the ducts representing the down pipe starting from the
exit cone of the catalytic converter.
Enter the geometric properties as shown in the schematic in Figure 1. Include a 90 Bend Angle for
each duct and remember to use a Discretization Length of 40 [mm]. Set the Initial Conditions
to 1.05 [bar] Pressure and1000 [K] Fluid Temperature with 800 [K] Wall Temperature. If desired,
use control points to represent the bend angle on the canvas.
With the down pipe modeled, the canvas should appear similar to Figure 3.
The resonator is modeled with a short entry duct (40 mm long as shown in Figure 1), Complex Yjunctions representing the inner perforated duct and the surrounding can, and a short exit duct. The
completed WaveBuild representation of the resonator appears in Figure 4.
The number of consecutive Complex Y-junctions to use can be determined by dividing the total length
of the can (300 mm) by the exhaust discretization size (40 mm). The results in 7 Complex Yjunctions used in series to represent the inner perforated duct and 7 more used to represent the
surrounding can. The easiest method for building this resonator model is to build and define the first
Complex Y-junction for the inner duct and outer can each, and then copy/paste them 6 times.
The first Y-junction representing the inner perforated duct is cylindrical -- 42.857 mm long (300mm/7)
and 50 mm in Diameter. This means it has a Volume of 84149.8 [mm3] The Heat Transfer/Skin
Friction Area can be calculated from the area of the pipe minus the area of the perforates connecting
to the outer can. The area of the perforated section of pipe is 47123.89 mm 2. The total area of the
perforates account for 20057.5 mm 2. Assuming the perforates are evenly distributed along the entire
length of the pipe, we can calculate the surface area of a single Y-junction as 1/7th of the total surface
area. Thus, the total Heat Transfer/Skin Friction Area for the single Y-junction is 3866.63 [mm2]. Set
the Initial
Conditions to 1.05 [bar] Pressure, 900 [K] Fluid
Temperature,
and 700 [K] Wall
Temperature. The Y-junction panel and pipe orientations should appear as in Figures 5 and 6.
The first Y-junction representing the outer can is cylindrical as well, with the volume representing the
inner duct subtracted (don't forget to include the 1 mm thickness of the tube wall). This creates a
torus, or doughnut-shaped volume, with an equivalent Diameter of 108.15 [mm].
The
resulting Volume is 393686.4 [mm3] The Heat Transfer/Skin Friction Area of the entire outer can,
including the end-caps can be averaged over all seven Y-junctions, and is 18781.3 [mm2]. Set
the Initial
Conditions to 1.05 [bar] Pressure, 800 [K] Fluid
Temperature,
and 600 [K] Wall
Temperature. The Y-junction panel and pipe orientations should appear as in Figure 7 and 8.
Copy/paste the two Y-junctions to create seven of each, in series as shown in the figure below. All
of
the Y-Junctions representing
the
inner,
perforated
pipe
should
be
connected
by massless ducts with Diameters of 50 [mm]. All of the Y-junctions representing the outer can
should be connected by massless ducts with Diameters of 108.15 [mm]. And each inner,
perforated pipe Y-junction should be connected to the adjacent outer can Y-junction by
amassless duct which
represents
the
perforates,
with
a Diameter of 11.3 [mm]
and
a Count of 28.57 (=200/7, assuming they are evenly distributed).
Show
Chapter
Example Input File:
sections
4.1 The Basic
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4_exhaust.wvm
Geometry
4.2 The Mid-Pipe Example Output File:
4.3 Modeling the
Muffler in Quasi2D
.\Ricardo\WAVE\8.0\examples\engine\TUT_si\tut_si4_exhaust.wps
The schematic of the mid-pipe and complex muffler is shown in Figure 1. In WaveBuild, the two
segments that represent the mid-pipe can be modeled simply as two ducts, starting from the orifice
junction at the end of the resonator. The complex muffler will be modeled as a series of inter-
connected Complex Y-junctions, massless duct connections, and normal ducts for the muffler
entrance and tailpipe.
Drag and drop two orifice elements onto the canvas as shown in Figure 2, to represent the ends of
the ducts used to model the mid-pipe. Draw the ducts representing the mid-pipe starting from
the orifice junction at the exit of the resonator.
Enter the geometric properties as shown in the schematic in Figure 1. Include a 90 Bend Angle for
the first duct and remember to use a Discretization Length of 40 [mm]. Set the Initial
Conditions to 1.05 [bar] Pressure and800 [K] Fluid Temperature with 600 [K] Wall
Temperature. If desired, use control points to represent the bend angle on the canvas.
With the mid-pipe modeled, the canvas should appear similar to Figure 3.
Because the muffler is such a large and complex volume, the three internal chambers which
physically divide the muffler vertically, as shown in the schematic, should be sub-divided and modeled
with multiple Complex Y-junctions. Each section of perforated pipe will also have to be
represented by Complex Y-junctions attached to the surrounding volume(s) with multi-count
massless ducts to represent the perforates, as in the resonator. The level of refinement is up to
the user and what kind of results they desire: for performance, simple Quasi-2D meshing may be
sufficient but for acoustics, detailed Quasi-3D meshing should be performed.
We refer to meshing in WAVE as "Quasi" meshing as it is not identical to full-scale 3-D CFD but can
approximate the same results. Quasi-2D meshing (Box A in Figure 4) is sub-division of volumes in 2
axial directions (X and Y, in this case) and ignoring the division in the 3rd direction (Z axis). Quasi-3D
meshing (Box B in Figure 4) is sub-dividing volumes in all 3 axial directions. The number of subvolumes increases dramatically between Quasi-2D and Quasi-3D meshing!
The complexity of this connectivity of Y-junctions, massless ducts, and normal ducts has no uniquelycorrect design and different designers might end up with different connectivity matrices for this same
muffler. Also, this modeling work has no real engineering value, it's only routine work that could to be
automated! It is expensive and tedious for you, the engineer, to spend your time on such laborious
tasks. For this reason, Ricardo Software introduced WaveBuild3D as an auto-mesher for large
components in the intake and exhaust systems using a solid-modeling interface to represent the
geometry. The table below shows a comparison for different mesh matrix sizes vs. pre-processing
time of manual meshing (by you the user) and WaveBuild3D (computer program auto-meshing).
* Based on average time an engineer spends to complete one complex Y-junction input = 6 minutes
** Based on meshing using Intel Core 2 Duo T7600
# of Subvolumes
(Matrix Size)
Meshing Type
2D/3D
# of Repeated
Tasks
Engineer
Minutes*
WaveBuild3D
Minutes**
3x3
2D
18
108
0.03
3x3x3
3D
54
324
0.05
9x9x9
3D
1458
8748
0.37
15 x 15 x 15
3D
6750
40500
0.83
nxnxn
3D
2 x n3
2 x n3 x 6
It is left to you, the new WAVE user, to complete modeling the muffler. At this point there is nothing
new to learn. All that needs to be done is repeating the geometric and duct orientation input many
times to model the muffler as a network of ducts, massless ducts, and Complex Y-junctions.
The muffler is shown in WaveBuild3D in Figure 5 and the corresponding WAVE mesh is shown
in Figure 6.
The WaveBuild3D tool creates portions of WAVE models known as Components. Any WaveBuild3D
component can be used in any WAVE model, regardless of whether or not the user has a
WaveBuild3D license. This muffler component is shipped as an example model. To add it to the
model, expand the Tags branch under the Components branch of the Elements tree. Drag and drop
the SI_Muffler component onto the canvas and connect it to the model by attaching the last duct of
the mid pipe (the last orifice element can be deleted, as it is no longer necessary) to the inlet
connection point of the component. Connect the outlet connection point of the muffler component to
the Exhaust ambient using a duct with 50 [mm] diameters and an overall length of 40 [mm] (use initial
conditions similar to the mid-pipe).
When finished, the model should appear similar to Figure 7.
Show
Spatial Plots in
WavePost
5.2 Animating
Spatial Plots
Run the WAVE model by clicking on the Run Screen Mode button in the toolbar
(if prompted to
save the model before running WAVE, click the OK button to save and run sequentially).
Launch WavePost from WaveBuild by clicking in the Launch WavePost button in the WaveBuild
toolbar . This will open the.wps file named similarly to .wvm currently loaded in WaveBuild. Once
WavePost is open, click the Open File button
in the toolbar and open the .wps file created
in Phase 3, Comparison.wps. This WavePost session file is currently referencing results from the
4-cylinder model without intake/exhaust (tut_si4) and from the four-cylinder model with intake only
(tut_si4_intake). It should contain sweep plots created during the in Phase 3 of torque, power,
etc.
Click on the Add button at the bottom of the Output Files frame and select the newly
created .wvd file for the WAVE model containing both the intake and exhaust systems
(tut_si4_exhaust). The .wvd and .sum files for this WAVE run will be added to the Output
Files list. A Query window will pop-up prompting whether to add curves to the existing plots using
this file. Click on the Yes button and every existing plot will add data from the newly
added.wvd and .sum files (if matching data exists in the new files). Open the Sweep Plots to view
the comparison of the performance parameters.
On the Canvas of the WavePost GUI, use the Shift+Left Click to multiple-select
the ducts and junctions along the intake runner to the cyl1 junction (see Figure 1). Right-click on
one of the ducts/junctions in the highlighted group and select the "Add Spatial Plot > Basic >
Pressure" menu item, as shown in Figure 2.
A Spatial Plot of the pressure along this path will be automatically created, showing a scaled view of
the ducts/junctions along the path and the pressure profile in those ducts/junctions on the same
plot (see Figure 3). Therepresentation of the ducts/junctions can be moved using the middle
mouse button or removed from the plot altogether by double-clicking on it and de-selecting
the Display Scale View option, where the crank animation can also be displayed (also accessible
from the Tools > Display pull-down menu item).
Spatial Plots can be animated just like canvas displays of network variables, as demonstrated
in Phase 4. Simply open the Animation Panel by selecting the Animation... menu option from
the Tools pull-down menu (seeFigure 5). The animation can be played or recorded to a MPEG file
for use in reports/presentations. The Time Offset can be set to change the starting point of the
animation, allowing multiple spatial plots to be shown in/out of phase.
The Animated Spatial Plot should appear as the video at the bottom of the page.
CONGRATULATIONS!
The 4-cylinder model is complete! All of the basic principles in WAVE modeling have been introduced
and explained as well as the basic methods of building a model and post-processing the
data. Optional phases follow to elaborate on the basic principles.