Beruflich Dokumente
Kultur Dokumente
INDEX
1.0 Introduction 3
1.0 INTRODUCTION
Pinch Technology provides a basis for the design of process and utility networks. The starting
point is the identification of the minimum practical energy consumption for a given project. Next,
this “target” consumption is achieved using a blend of systematic steps and the designer’s own
decisions. The procedure is fun to use and is quick and easy to apply. It provides a fundamental
and logical basis for design studies.
The systematic aspects of Pinch Technology do not inhibit the designer’s ability to remain “in
control”. Rather, they help to prevent possible options being over looked and a non-optimal design
inadvertently being selected.
Pinch Technology considers in depth both energy and capital costs. Aspects of operability,
materials of construction, plant-layout and safety are taken into account. The procedure is
relevant to the design of new plant as well as to retrofit projects.
(MER Design)
• Threshold Problems
This is the flow sheet of the front end part for a speciality chemical process. The feed is preheated
by process heat recovery (exchangers 1 and 3) and by two steam heaters before it is fed into the
reactor. The reactor product is being cooled down in exchangers 1,2 and 3 and finally with cooling
water before it goes into a separator. The bottom product is preheated before downstream
processing, while the top product is mixed with a recycle stream and the original reactor feed.
First, the problem is defined and second, a design is derived using previous experience.
In this course, we will introduce different way to solve the problem in three stages:
1. Problem Definition
2. Targets (energy, capital)
3. Design
First, the problem is defined. Second, a ‘targeting’ procedure is performed, which predicts, ahead
of any design, the minimum energy and capital costs of the problem. Third, with the information
obtained in the second stage, a design is derived which achieves these targets.
In the ‘problem definition’ stage the process is being stripped-down to its essential unit operations,
namely reactors, separators, etc.
In this way the process ‘stream data’ are extracted. Each process stream is expressed by:
T
supply : initial (starting) stream Temperature
T
target : final stream Temperature
In the targeting stage, the minimum energy and capital costs for a process are identified.
The energy targets give, independently of any design, the minimum heating and cooling
requirements of a process.
Yes, we can predict whether we really need this amount of hot utility.
The capital targets are divided in to “Units targets” and “Area targets”.
The process shown has been designed with six Heat Transfer Units.
The Unit targets give, independently of any design, the minimum number of Heat Transfer Units
required by a process.
The process design shown below requires a total of 692 m2 of heat transfer area.
Can we predict what the minimum required amount of surface area is?
The area targets give, independently of any design, the minimum heat transfer
area required by a process.
This is how the existing process utility consumption, number of units and surface area
requirements compare with the targets. The targets clearly indicate that the existing design can
be improved quite significantly.
As seen before, targets are useful because they give an indication on how efficiently energy
and/or capital are used in a design.
The first relates to how meaningful these targets are, namely is there a design which performs as
good as predicted? The second relates to how compatible the targets are, namely can we have
minimum utility consumption, minimum number of units and minimum surface area at the same
time?
2.5 DESIGN
The targets are meaningful. Here a design is shown which uses the minimum utility requirements,
as predicted by the targets, i.e. 1068 units of steam and no cooling water.
Checked by: HVL Pinch Technology Page 8 of 93
Training Module Rev.0
(No. REL.CHE.066) Date:25th Dec. 2003
Approved by: HVL Author : H.V. Lodhia
Training Reliance Industries Ltd. Jamnagar Refinery Complex
A closer examination of this design also shows that the number of units used and the surface area
required are as predicted by the targets.
There is no Encroachment in :
The design shown before features significant improvements in terms of both energy and capital
costs compared to the existing one. This has not been achieved by altering unit operation
technology, or by using ‘exotic’ materials of construction or by compromising process flexibility or
safety. The design has been improved solely by better process integration.
Track record shows that in some; projects energy and capital savings have been achieved at the
same time. This goes against common belief that if you want to save energy you need to spend
capital. This point is discussed by means of a simple example.
(A) Network for Minimum Capital Cost (but max. energy cost)
(B) Network for Minimum Energy Cost (but max. capital cost)
Here we have three streams requiring heating and three streams requiring cooling.
We would expect that the design (A) which uses steam only for the streams requiring heating and
cooling water only for the streams requiring cooling will be expensive in terms of energy cost but
will have minimum capital cost.
In contrast, the design (B) which has a good integration scheme, will have minimum energy cost
but will be expensive in terms of capital.
Let us put some numbers to our simple example. All six streams have the same enthalpy change
(500 units) and the overall heat transfer coefficient (U) is assumed to be U = 1. The result is not
what was expected. The design using no utilities is better in terms of capital cost as well: less
surface area and half the number of units.
H
Higher T’s (Hence, Heat Transfer is 3000 units)
H
Lower T’s (only 1500 units)
The reason is that the first design features better driving forces, but also transfers more heat.
Q
A=
U T LM
In this example the effect of Q on the area is larger than the effect of DTLM. As a result, the
second design has less energy and less capital cost.
Before we start the working further, we need to introduce some terminology which will be used
throughout the course:
(Reducing in Enthalpy)
They are shown running from right to left in the temperature / enthalpy
(T-H)diagram, i.e. reducing in enthalpy.
T
(Increasing in Enthalpy)
(Left to Right)
They are shown running from left to right in the temperature / enthalpy (T-H) diagram, i.e.
increasing in terms of enthalpy.
T = 20 800
T = 30 800
500 500
T = 10 T = 20
Exchange A Exchange B
The exchangers shown below are now represented on the (T-H) diagram. The hot and
cold streams of each exchanger are placed in the way discussed earlier.
Exchange A
100
T = 20o
O.K.
80
60
T = 10o
Not O.K. 40
20
Exchange B
100
T = 30 o
80
60
T = 20o
40
20
O.K.
Each process stream is shown in terms of initial (Tsupply) and final (Ttarget) temperatures. The
enthalpy change (DH) of all streams are also given, there are four streams requiring heating or
cooling. For operability reasons, reboiler and condenser in column (C1) are not considered in the
problem.
CP of each stream
Heat capacity flow rates of all four streams can be calculates as:
H / (Tsupply – Ttarget )}
CP = {
EXAMPLE PROBLEM 1
Data
Supply Target
Stream Heat Capacity
Temperatur Temperatur
No. Flowrate
e e
& Type (kw / °C)
(°°C) (°°C)
1
180 80 1.0
Hot
2
130 40 2.0
Hot
3
60 100 4.0
Cold
4
30 120 1.8
Cold
T min = 10 °C
Utilities: Steam at 200°C, CW at 25°C
Design a network of steam heaters, water coolers and exchangers for the four process streams.
Where possible, use heat interchange in preference to utilities.
Conventional Design
This is the most popular conventional solution. Are 60 units of steam necessary?
The enthalpy box for the most popular solution (discussed earlier) shows that that enthalpy
imbalance between the hot and cold streams is made up by external utilities, i.e. steam, QH = 60
and CW, QC = 18 units. Whatever the amount of heat interchange between the hot and cold
streams is, the enthalpy imbalance between the hot and cold streams (DH) must always be equal
to the difference of the external utilities (QH-QC).
If the amount of the externally supplied hot utility (QH) is reduced, the amount of the cold utility
(QC) is also reduced by the same amount. The difference (QH –QC) is always equal to the enthalpy
imbalance between hot and cold streams (DH) and is therefore constant.
The question arising now is how far can hot and cold utility be reduced in the problem.
To answer the last question, the first law as applied before, is not sufficient. We need to also use
the second law of thermodynamics, i.e. introduce the temperature dimension to the problem.
This is the temperature/ enthalpy diagram for one of the hot streams from the last Example
problem 1. Since we assumed that CP = constant, the line connecting supply temperature (130°C)
and target temperature (40°C) is a straight line. The enthalpy change of this stream is clearly
shown on the enthalpy axis.
This is the T-H diagram of two hot streams having identical supply and target temperatures, but
different heat capacity flow rates (CP). Hence the slopes of the lines are different. The flatter the
slope the larger the CP.
A simple process is shown. One hot stream and one cold stream are exchanging heat at DTmin =
10°C. The cold stream requires additional heating, provided by steam, while the hot stream
requires further cooling, provided by cooling water.
The process is also shown in the T-H diagram. DTmin = 10°C is observed. The overlap between hot
and cod curves represents the amount of heat transferred in the heat exchanger.
The remaining part of the hot stream is satisfied by cooling water (QC = 30). The remaining part
of cold stream is satisfied by steam (QH = 50).
Note also that for DTmin = 10°C, the amount of hot and cold utilities shown are the minimum
possible, hence the notations QHmin & QCmin are used.
Tt
Ts
In the T-H diagram the curve representing a stream can be shifted horizontally. What matters is
not the absolute value of the enthalpy but the enthalpy change of a stream.
T
Tt
Ts
On the other hand, different streams with the same supply and target temperature will have
different slopes.
Based on what was discussed above (streams can be shifted horizontally in the T-H diagram) we
now move the cold stream curve to the right, up to the point where DTmin = 20°C.
It is clear that the heat transferred between the two streams has been reduced and that steam
and cooling water requirements have increased (the dotted line represents the position of the cold
stream curve for DTmin = 10°C.
It is also clear that the increase in steam consumption is equal to the increase in cooling water.
Note also that for DTmin = 20°C, the amount of hot and cold utilities shown are again the minimum
possible.
If the amount of heating supplied is increased by X units then the amount of cooling required will
be increased by exactly X units.
Stream Data
Stream No.
And Type CP Ts Tr
TS = supply temperature
TT = target temperature
The ‘composite curves’ represent an overview of the stream data for the whole process in the T- H
diagram.
The way by which a hot composite curve is constructed is shown below, using the example
problem 2.
This problem has two hot streams. Their supply and target temperatures are placed in a
descending order (see dotted lines) so that temperature intervals are created (three in this case).
The streams are drawn between their supply and target temperatures with the appropriate slope
(imagine a vertical temperature axis and a horizontal enthalpy axis.)
In the first interval only one stream exists, the one with CP = 3, so the hot composite is a
line with a CP = 3.
In the second temperature interval, both stream exist, so the hot composite is a line with
a CP = 3.0 + 1.5 = 4.5.
The construction continues in the same way for the other intervals.
The cold composite curve is constructed in a similar way as the hot composite curve.
The hot and cold composite curves constitute the ‘composite curves’. Hot and/or cold composite
curve can be moved horizontally. Here they are placed in such a way that the smallest
temperature approach in 10°C.
The overlap between hot and cold composites represent the maximum amount of heat which can
be exchanged between the process streams for DTmin = 10°C.
The overshoot of the cold composite is balanced by hot utility (QHmin = 20).
The overshoot of the hot composite is balanced by cold utility (QCmin = 60).
Note also that for DTmin = 10°C, the amounts of utilities shown are the minimum possible for this
problem. (Example problem 2).
As the composite curves are moved apart (i.e. DTmin increases), the amount of heat recovered is
reduced, so the amount of QHmin and QCmin is increased.
Construct
Composite Curves
Energy Targets
From the stream data, the composite curves can be constructed. For a specific DTmin the energy
targets are then defined. Note that the energy targets are identified prior to any design.
Correct setting of composite curve requires trial and error. May be cumbersome, especially for
problem having several hot and cold streams.
If the composite curves are placed in such a way that there is no overlap between them (i.e. no
integration), then all process streams are satisfied using utilities.
As composite curves are moved closer, they will start to overlap. This overlap represents the
amount of heat given up by the hot streams to the cold streams (process heat recovery). The
remaining of the cold composite is balanced with hot utility, while the remaining of the hot
composite is balanced with cold utility.
As the composite curves are moved closer and closer, the amount of heat recovery increases,
while the amount of hot and cold utilities is reduced.
The minimum amount of hot and cold utilities (QHmin and QCmin) is found when the composite
curves touch. That point is called ‘Pinch’.
At the pinch the driving forces between hot and cold composites are zero. This represents an
infeasible situation, as the heat transfer area required at this point is infinite.
To avoid infinite heat transfer area, the composite curves are moved apart, such that there is a
positive temperature difference between hot and cold streams at the pinch. The heat transfer area
required is reduced, but the minimum hot and cold utilities are increased (higher energy cost.)
The right choice of DTmin is important. It depends on the problem and the underlying economic
factors. There are ways of establishing the optimum DTmin and this will be covered extensively
later in this course.
If we separate the process at the pinch, we can make the following observations:
1. Above the pinch the heat required by the cold streams is provided partly by the hot
streams and partly by hot utility. This part of the process is a net ‘Heat sink’. It is in
enthalpy balance with QHmin units of heat from external utility. No cooling is required.
2. Below the pinch, the heat required by the cold streams is completely provided by the hot
streams. The remaining heat in the hot streams is removed using cold utility. This part of the
process is a net ‘Heat source’. It is in enthalpy balance with QCmin units of external cooling.
No heating is required.
We also observe:
We observe that whilst it is possible to transfer heat from above the pinch to below the pinch the
opposite is not possible. For example, it is feasible to use a hot stream above the pinch in order to
heat a cold stream below the pinch: the driving forces are positive and larger than DTmin and this
happens quite often in existing designs. The opposite however, i.e. using of a hot stream below
the pinch to provide heat to a cold stream above the pinch is infeasible: the driving forces are
either less than DTmin or negative.
As discussed, the pinch separates the process into two independent regions, which are in heat
balance with their utilities if the heat transferred between them is zero: (no heat transfer across
the pinch).
It is possible to transfer heat from above to below the pinch. Let us assume that this amount of
heat is (α).
The part of the process above the pinch, which before was in heat balance with QHmin now loses
(α) units of heat to below the pinch. To restore the heat balance, hot utility needs to be increased
by the same amount, i.e. (α) units. Below the pinch, we are adding (α) units of heat to the part of
the process that had an excess of heat. Therefore the cold utility requirements will also increase
by (α) units.
So, if in any design (XP) amount of heat is transferred from above to below the pinch, then the
actual heat consumption will be (XP) over and above the target consumption.
α 1 : hot utility being used to heat cold streams below the pinch.
α 2 : hot streams above the pinch used to heat cold streams below the pinch.
α 3 : cold utility being used to cool hot streams above the pinch.
When we concentrate on the streams that need hating and cooling only and leave out all unit
operations, the process can be represented in the ‘Grid Diagram’.
The hot streams are running from left to right at the top end of the diagram, while the cold
streams are running from right to left at the bottom end of the diagram.
Heat exchangers between process streams are shown as two circles on the corresponding streams
joined by a line. Heaters and coolers are shown as a single circle with the letter H or C
respectively. The heat load is usually shown below each exchanger.
In the conventional flow sheet, showing the pinch is quite complicated and not so clear, as
streams are scattered all over. The grid diagram features a number of advantages over that of a
conventional flow sheet representation. The first relates to the pinch location.
CONVENTIONAL FLOWSHEET
Pinpointing the pinch on each different stream is inconvenient and will be quite messy, especially
when more streams are involved.
GRID DIAGRAM
In the grid diagram the pinch location is easily indicated. It separating the above and below the
pinch streams. The right part of the grid diagram relates to the below the pinch part of the
process while the left part relates to the above the pinch part of the process.
Correct and incorrect heat transfer is now easily represented in the grid diagram. “Mistakes” in
heat exchanger network design are quickly spotted.
The Implications of the PINCH are easily assessed in the Grid Diagram.
These are the stream data for example problem 2. Reboilers and condensers are not considered
as part of the integration study for operability reasons. The minimum temperature approach
(DTmin) is 10°C.
The composite curves for example problem 2 give the pinch temperatures and the minimum utility
consumption for DTmin = 10°C.
Heat transfer is most constrained at the pinch. We may as well start our design at the pinch while
we still have all our options available for placing heat exchanger.
Given that, at the pinch, we can match any hot stream with any cold stream, the question is :
which streams should we consider first ?
Let us first consider the process above the pinch and find a partner for the stream with CP = 3.0
We have two options :
1. Match with the cold stream having CP = 2.0 . We know that the cold end of this match has
a driving force equal to DTmin = 10°C. Because CP of the hot stream is larger than CP of
the cold stream, the driving forces along this match, as we move away from the pinch, will
decrease, i.e. violation of DTmin’
2. Match with the cold stream having CP = 4.0. The cold end of this match will also have a
driving force equal to DTmin = 10°C. Here, however, CP of the cold stream is larger than
CP of the hot stream so the driving forces will increase along this match as we move away
from the pinch.
Only the second option is possible, and generally we should observer CPcold > CPhot for pinch
matches above the pinch.
Let us now consider the process below the pinch and find a partner for the cold stream with CP =
2.0. We have two options:
1. Match with the hot stream having CP = 1.5. We know that the hot end of this match has a driving
force equal to DTmin = 10°C. Because CP of the cold stream is larger than CP of the hot stream,
the driving forces along this match, as we move away from the pinch, will decrease, i.e. violation
of DTmin.
2. Match with the hot stream having CP = 3.0. The hot end of this match will also have a driving
force equal to DTmin. Here, however, CP of hot stream is larger than CP of cold stream so the
driving forces will increase along this match as we move away from the pinch.
Only the second option is possible, and generally for pinch matches below the pinch we should
observe CPhot > CPcold.
5.4.3 IN GENERAL
The observations made can be generalized and summarized in this easy to remember form where:
OUT: refers to streams “coming out” from the pinch, i.e. cold streams above and hot streams
below the pinch.
IN : refers to streams “coming in” to the pinch, i.e. hot streams above and cold streams below
the pinch.
So, for feasible matches at the pinch, above or below, we should always have
Let us now concentrate on the streams coming out from the pinch, above the pinch (i.e. cold
streams.) We consider five different options at the pinch:
2. Match with a hot stream below the pinch: not possible, because all hot streams below
the pinch have temperatures lower than pinch temperature, i.e. this match will have a DTmin
violation, or negative DT’s.
4. Match with a hot stream above and away from the pinch: possible.
5. Match with a hot stream at the pinch: possible, but the hot stream needs to have a
smaller CP.
Generally, streams “OUT” above the pinch are considered to be “easy” as we have three different
possible options to match them.
Above the pinch, for streams going into pinch (i.e. hot streams). We consider five different options at
the pinch:
2. Use a cooler: possible, but not wanted as this will increase utility requirements (do not use
cold utility above the pinch).
3. Match with a cold stream below the pinch: possible, but not wanted as this will increase
utility requirements (do not transfer heat across the pinch).
4. Match with a cold stream above and away from the pinch : not possible, because these
cold streams have temperatures higher than cold pinch temperature, i.e. this match will have a
DTmin violation or negative DT’s.
5. Match with a cold stream at the pinch: possible, but the cold stream needs to have a
larger CP.
Generally, streams “IN” above the pinch are considered to be “difficult”, as we have only one
possible option to match them.
Below the pinch, for streams coming out from the pinch (i.e. hot streams). We consider five different
options at the pinch:
2. Match with a cold stream above the pinch: not possible, because these cold streams
have temperatures higher than cold pinch temperature, i.e. this match will have a DTmin,
violation or negative DT’s.
4. Match with a cold stream below and away from the pinch: possible.
5. Match with a cold stream at the pinch : possible, but the cold stream needs to have a
smaller CP.
Generally, streams “OUT” below the pinch are considered to be “easy” as we have three different
possible options to match them.
Below the pinch, for streams going into pinch (i.e. cold streams). We consider five different options at
the pinch:
2. Use a heater: possible, but not wanted as this will increase utility requirements (do not use
hot utility below the pinch).
3. Match with a hot stream above the pinch: possible, but not wanted as this will increase
the utility requirements (do not transfer heat across the pinch).
4. Match with a hot stream below and away from the pinch: not possible, because these
hot streams have temperatures lower than hot pinch temperature, i.e. this match will have
DTmin violation or negative DT’s.
5. Match with a hot stream at the pinch: possible, but the hot stream needs to have a larger
CP .
Generally, streams “IN” below the pinch are considered to be “difficult”, as we have only one
possible option to match them.
Only the second option satisfies the CP-inequality. So for stream 2 we have only one option:
match with stream 3.
An explained earlier, we start with the steams going in to the pinch. Out of those, the most
difficult are the ones with the larger CP’s. so, generally, we start with the stream having the
largest CP, then the second largest, etc., etc.
We know that these matches, irrespectively of their heat loads, will always be feasible, i.e. no
DTmin violation. As capital cost depends quite significantly on the number of units used in the
network, we maximize the heat load of each match. This is what we call the “tick-off” heuristic:
the smaller enthalpy content or requirement of the two related streams, determine the heat load
of the match.
This heuristic, ticks-off one stream every time a match is placed, except where the enthalpy
change of the steams involved are equal: see the match between streams 2 and 3, where both
streams are ticked-off.
As there are no other hot streams, the final part of the heat required on stream I, has to come
from hot utility. The heat load of the heater is equal to the hot utility target predicted by the
composite curves.
There is only one option: stream 1 should be matched with stream 2 to satisfy the CP-
inequality.
Follow the “tick-off” heuristic to keep capital cost down. Cold stream I needs additional heat.
The rest of the heat required on stream1 can be supplied by hot stream 4.
The match between streams 1 and 4 has CPin > CPout (does not agree with the CP-inequality). In
this case, however, the hot side driving force DTmin = {90-35} = 55 > DTmin. So even though the
driving forces will be reduced as we move towards the cold side of the match, we do not reach
DTmin violation.
Generally, away from the pinch, the CP-inequality does not matter. However, we must always
check the temperatures, to avoid DTmin violations or temperatures infeasibilities (negative DT’s).
As there are no other cold streams, the rest of the heat of stream 4 is removed with a cooler. The
cooler heat load is 60 units, equal to the minimum cold utility target predicted by the composite
curves.
We now join the above and below the pinch designs. Note that the final design achieves the
energy targets predicted (ahead of any design) by the composite curves, for DTmin = 10°C.
If we take the pinch division out, we have the completed design in the grid diagram
The design from the grid diagram can be easily transformed into the conventional flow sheet
form.
6.1 No Integration
All heat required by the cold streams is supplied by hot utility and all heat available from the hot
streams is rejected to cold utility (i.e. no heat recovery, no overlap between the composite
curves.)
As the composite curves are moved closer together (heat recovery is introduced) the utility
requirements decrease.
Checked by: HVL Pinch Technology Page 59 of 93
Training Module Rev.0
(No. REL.CHE.066) Date:25th Dec. 2003
Approved by: HVL Author : H.V. Lodhia
Training Reliance Industries Ltd. Jamnagar Refinery Complex
We have seen before that if the composite curves are moved even closer (i.e. DTmin < 20°C),
both hot and cold utility requirements are reduced, up to the point where the curves touch each
other.
The same happens in this example. Here, however, at DTmin = 13.7°C, no more hot utility is
enough heat available from the hot streams is enough to provide the duty required by the cold
streams. At this point the process requires cold utility only.
A further reduction in DTmin will shift the composite curves closer, but the process heat recovery
will not increase.
Part of the cold utility can now be used at higher temperatures, but the total amount of cold utility
will remain constant.
6.4 THRESHOLD
The graph of DTmin versus utility consumption shows clearly this “threshold” type of problem.
Namely from DTmin = 0 to DTmin = 13.7, the cold utility is invariant and thereafter the need for the
second utility appears and both utilities increase in parallels.
Contrast this with the behaviour of the “normal’ (pinched) problem. Here, both utilities are always
present (even DTmin = 0 ) and are always a function of DTmin.
6.5 SUMMARY
We notice, however, that not all the hot utility needs to be supplied from high pressure steam
(HP). Part of it can be supplied at lower temperatures, using low pressure steam (HP).
If the problem was already pinched, the introduction of a lower steam level can introduce a
second (utility pinch).
The problem is threshold and in order to reduce energy costs we supply part of the heat
requirement with LP steam. Introduction of the LP steam creates a new pinch and it seems that
we supply heat below the pinch.
Even though we supply heat below the utility pinch, the utility requirement will not increase. This
pinch did not originally exist and was only created because we decided to use LP steam. This is
not the same as using hot utility below the “normal” process pinch.
Generally, in most processes, a number of different utility levels exists. For instance, not all the
process cooling duty needs to be done by refrigeration. If temperatures permit, we might be able
to reject part of that heat into cooling water and so reduce energy costs. By doing so, new utility
pinches may be created. The parts of the process composite curves between the various pinches
are balances by the use of the appropriate utility level. For example, the process streams between
the process pinch and the LP generated pinch, are balanced by the LP steam.
7.4.1Existing Design
7.4.3 SUMMARY
Stream splitting is a powerful technique for better energy recovery but don’t split unless necessary
!!!
NOTE:
Initial inspection would suggest contacting streams 1 and 5 at the cold end of the process. This
would reduce the heat loads on cooler C1 and the heater. Stream 1 is chosen in preference to
stream 2 because of its significantly higher CP and the higher load on the cooler.
However, the integration of a new heat exchanger is not completely straight forward. The new
exchanger would affect the temperatures in “downstream” exchangers 1 and 4 which would lead
to the need for additional area. Then, if additional area were needed in exchanger 4 anyway, we
should once more consider stream 2, with a view to reducing the load on cooler C2. with this type
of reasoning, a network may result as shown above. The overall saving in energy is 2335 units.
But why should we choose the level of energy saving shown in 8.2.2. By installing more exchange
area (i.e. investing more capital), we could have saved more energy. By installing less exchanger
ara, we could save on capital, although we would save less energy an economic analysis for
various energy-recover levels is shown in the graph of energy savings versus capital investment. A
simple calculation shows that the “adopted choice” saves significant energy (about 13%) at a
good payback (2 years).
Both energy and area targets are functions of DTmin and can be related to each other. A typical
plot of area requirement as a function of energy consumption is shown. It is called the
area/energy plot.
This plot shows that for small DTmin, the process requires little energy but large surface area.
Conversely, at a large DTmin, the process requires much energy but less surface area.
An optimum trade-off between energy and capital costs exist and this can be determined from the
detailed cost analysis ahead of design.
The area below the curve is shaded and marked “infeasible”. It is not possible for a design to
perform better than target. But we can do worse than target. The design shown uses too much
energy for the area installed or employs too much area for the energy consumed (it features criss-
crossing).
Generally, there are two types (degrees) of non-optimality. The first degree refers to networks
which achieve the energy and area targets, but have the wrong DTmin (wrong trade-off). The
second degree refers to networks which feature criss-cross heat exchange.
The area/energy plot for a given set of stream data. From the actual energy consumption and
network area the point representing the actual process plant is readily located on the plot. In most
cases, it will be above the target curve.
The existing design shown does not take best advantage of its installed area. It does not recover
as much energy it could. If we were to improve this situation, which way should we move ?
It is often assumed that good retrofits should aim for the optimum new design. This does not
make sense. Who is prepared to throw away area that has already been paid for. Our first
objective must be to use the existing area more effectively.
We now recognize:
Optimum retrofits do NOT lead to today’s optimum grass-root design
Instead we should try to improve on the use of are by reducing criss-crossing whilst shifting the
composite curves closer in order to save energy. The ideal direction to move from the existing
design point would therefore be horizontally to the left. In principle, moving horizontally up to the
targeting curve is possible. However, in practice this is often undesirable. The network structure
there is likely to be different. Hence, it would requires different exchanger sizes and types. If we
are to improve on the use of the exchangers already installed we usually have to invest in new
exchangers.
Usually, many project options are available to the designer. So many lines of possible retrofits
exist.
The most effectiveness of these curves will be different the lower the curve the lower the
investment for a give energy saving.
8.8 Payback
Assume that the best curve is the line shown there. The shape of this curve is typical. Its slope
increased with increasing investment. This implies that the payback period increases with
investment level.
Conservative Assumption:
A good retrofit will at least maintain the existing overall “efficiency” of the network…..
Unfortunately, the “best curve” is difficult to determine. It is a function of plant layout and process
constraints.
But could we predict the curve that at least describes the best we can expect ? The assumption
would be that the network, after retrofit, will use overall area at least as effectively as before. In
other words, the retrofit project will at least not increase the overall level of existing criss-crossing.
We define the area efficiency, as the ratio of minimum area required (target, AT) to that actually used
in the network (Aex) for the existing energy usage.
The value of can be expected to be less than 1 in practical designs. A value of 1 would indicate no
criss-crossing. The lower the value of, the poorer the use of area and the more severe the criss-
crossing.
Conservative Assumption:
α = constant (<1)
If we assume that α is constant for every level of energy saving, we would obtain the “constant
α” line.
8.12 We can predict how much energy for how much capital
From the “constant α” curves, we can determine the energy savings for different levels of
investment.
By using the costs of area and energy, the “constant α” curve can be transformed into a savings/
investment relationship. This curve relates annual payback.
The retrofit target curve has been drawn for our example, based on the conservative assumption
α = constant and on the cost data given. The economic set point for the retrofit by inspection,
discussed earlier has also been included.
For an investment of ₤ 0.29 million, the retrofit by inspection gave an energy saving of
₤150,000/year. But our target curve shows possible savings of ₤ 192,000/- year – an
improvement of 28% This would correspond to a payback of 1.5 years instead of 2 years.
Alternatively, we would expect more than double the savings at 2 years payback (₤ 320.000 as
opposed to ₤150,000).
Having obtained retrofit targets, a design methodology is needed with guarantees that the targets
will be met. Crucial design steps must be conducted correctly.
8.17 The targetd savings determine correct Tmin to initialize the retrofit design
The energy savings, as identified by the targets, define the relative position of the composite
curves, the pinch and the appropriate value of DTmin.
For our example, the value of DTmin for the targeted energy savings is 19 °C.
Savings = 5051 kw
For DTmin = 19°C, we redraw the existing and we identify the exchangers which transfer heat
across the pinch.
This is the outline of the retrofit design procedure that should be followed in order to achieve the
targets.
Exchangers 1,2 and 4 and cooler C2 have been removed from the design. They will leave heating
and cooling requirements for the streams involved.
Position new exchangers and, where possible, reuse exchangers removed in the previous step.
A possible network is shown. Above the pinch, the heater and exchangers 1 and 4 are reused.
Below the pinch, exchanger 2 is reused, but with reduced duty. The remaining enthalpy on stream
4 is taken by exchanger 3. cooler C2 has a reduced duty. Exchanger A is new.
Compatibility with existing network can be improved using loops and paths. The area of the
existing exchangers is to be reused as much as possible.
The final network is shown here. In this design, the surface area of exchanger 3 is fully reused.
Increased area is required in the other three existing matches. There is one new exchanger A.
8.20 Conclusion
Hence,
REAL
Retrofit Target
ANNEXURE – 1
Q 4 : What is the difference between Traditional Design & Design based on Pinch
Technology ?
A : Traditionally, the design problem is tackled in two stages.:
• Problem Definition
• Design
First, the problem is defined and second, a design is derived using previous experience.
Pinch Technology introduces different way to solve the problem in three stages:
• Problem Definition
• Targets (energy, capital)
• Design
First, the problem is defined. Second, a ‘targeting’ procedure is performed, which predicts,
ahead of any design, the minimum energy and capital costs of the problem. Third, with the
information obtained in the second stage, a design is derived which achieves these targets.
Q 7 : Whether both Energy & Capital saving can be achieved at the same time ?
A : Yes, track record shows that in some projects energy and capital savings have been achieved at
the same time. This goes against common belief that if you want to save energy you need to
spend capital.
Checked by: HVL Pinch Technology Page 90 of 93
Training Module Rev.0
(No. REL.CHE.066) Date:25th Dec. 2003
Approved by: HVL Author : H.V. Lodhia
Training Reliance Industries Ltd. Jamnagar Refinery Complex
Q 9 : What is meant by “Heat Sink” & “Heat Source” and why it is called so ?
A : Above the pinch the heat required by the cold streams is provided partly by the hot streams and
partly by hot utility. This part of the process is in enthalpy balance with QHmin units of heat from
external utility. Hence , it is called a net ‘Heat sink’.
Below the pinch, the heat required by the cold streams is completely provided by the hot streams.
The remaining heat in the hot streams is removed using cold utility. This part of the process in
enthalpy balance with QCmin units of external cooling. Hence , it is called is a net ‘Heat source’.
Q 10: Is it possible to transfer heat from below the pinch to above the pinch ?
A : We observe that whilst it is possible to transfer heat from above the pinch to below the pinch the
opposite is not possible. For example, it is feasible to use a hot stream above the pinch in order to
heat a cold stream below the pinch as the driving forces are positive and larger than DTmin and
this happens quite often in existing designs. The opposite however, i.e. using of a hot stream
below the pinch to provide heat to a cold stream above the pinch is infeasible as the driving forces
are either less than DTmin or negative.
So, if in any design (XP) amount of heat is transferred from above to below the pinch, then the
actual energy consumption (both Hot & Cold Utilities) will be (XP) over and above the target
consumption.
Q 12: How many types of cross pinch heat transfer are there ?
A : There are three different types of cross pinch heat transfer:
α 1: Hot utility being used to heat cold streams below the pinch.
α 2: Hot streams above the pinch used to heat cold streams below the pinch.
α 3: Cold utility being used to cool hot streams above the pinch.
Q 14 : What are the Steps for MER (Maximum Energy Recovery) Design ?
A : The Steps for MER (Maximum Energy Recovery) Design are :
Q 15 : What is the difference between the Pinched problem & Threshold problem ?
A : Pinched problem – Both Hot and cold utility required even at DTmin = 0
Threshold problem – One utility disappears at some value for DTmin
Utility pinches may be created. The parts of the process composite curves between the various
pinches are balances by the use of the appropriate utility level. For example, the process streams
between the process pinch and the LP generated pinch, are balanced by the LP steam. If the
problem was already pinched, the introduction of a lower steam level can introduce a second
(utility) pinch.
trade-off between energy and capital costs exist and this can be determined from the detailed
cost analysis ahead of design.
This is the outline of the retrofit design procedure that should be followed in order to achieve the
targets.