Exploring Aspen Energy Analyser to improve Processes

Global Energy Efficiency

Ben Maes

Extended Abstract

Thesis to obtain the Master of Science Degree in

Chemical Engineering

Supervisor: Prof. Henrique Anibal Santos de Matos

Examination Committee
Chairperson: Prof. Sebastião Alves

Supervisor: Prof. Henrique Anibal Santos de Matos

Members of Committee: Prof. Maria Cristina Carvalho Fernandes

June 2018

1
1 INTRODUCTION

1.1 Heat integration [1]

Heat exchange is an important topic nowadays in the industry. With heat exchange there are process
streams that need to be cooled down. They are used to heat up process streams that need to be heated
up and vice versa. With heat integration it’s possible to save a lot of cost because less utility need to be
used. Utilities are streams that only exist in the process to heat up or cool down process streams and
serve no other purpose. Almost all processes have at least some kind of heat exchange present in their
processes but this is often not the best possible design and a more energy efficient design can be
created. There are several possible programs that can be used for heat integration. One of them is AEA
(= Aspen Energy Analyzer). AEA is a program that can work along with, the very commonly used, Aspen
Plus by using the processes designed there. A good method to start with heat integration is pinch
analysis. This is a very popular method that allows the user to gain a design that uses the least amount
of utility possible. AEA isn’t the only program that focusses on heat integration of a process. The
university Instituto Superior Técnico has made its own program called FI2EPI. This program is not as in
depth as Aspen Energy Analyzer but provides the perfect platform to start learning about heat integration
and applying pinch analysis.

1.2 Pinch analysis [2]

Pinch analysis is the most famous and widely used technique to optimize the heat integration of a
process. In pinch analysis there is a pinch point. This pinch point is the point where the temperature
difference between the hot composite curve and cold composite curve is the lowest. These are curves
that set out the temperature against the enthalpy for a collection of streams. The area covered on the
x-axis by the hot composite curve but not by the cold composite curve shows the minimum energy
requirement for the cold utility while the area covered on the x-axis by the cold composite curve but not
by the hot composite curve shows the minimum energy requirement for the hot utility.

A specific minimum temperature difference is often needed for either practical, economical or safety
reasons. The minimum energy requirements for cold and hot utilities to get that minimum temperature
difference can be found by shifting the composite curves till the pinch point lands on the desired
minimum temperature difference.

Figure 1: two graphs showing the differences when working with a different minimum temperature difference.

2
In order to attain a design that has the minimum energy requirement several rules need to be followed:

1. The design gets divided by working in two parts: under and above the pinch
2. When starting with heat integration always start at the pinch and work your way from there to
the ends
3. No heat transfer across the pinch
4. No use of a cold utility above the pinch and no use of a hot utility below the pinch
5. Above the pinch the MCp of the hot stream must be lower than that of the cold stream that it is
connected with. This is the other way around for under the pinch.
6. Above the pinch the amount of hot streams can’t exceed the amount of cold streams. This is
the other way around for under the pinch.

When facing problems with the amount of streams, or MCp value, a stream can be split. This will
increase the number of streams and will divide the original MCp value over the two streams giving each
stream a lower MCp value than the original one.

Multiple pinches are also possible. While it’s possible that this is because there are, for that specific
minimum temperature difference, two or more points that where equally close to each other it’s often
more the result of using more than one hot or cold utility. These utility pinches are there to make sure
that for instance the hot utility with the lowest temperature is used as much as possible because that
utility cost less to have and this will then often lead to a cheaper costing design. The same is true for
cold utilities but here it’s then the utility that is the least cold that will be cheaper. When there are multiple
pinches the design must be done following the same pinch rules but in this case it’s important with
neighboring pinches to start from the pinches and work towards each other. The only extra rule that is
added is that there can’t be any crossing of utilities with the pinch points.

2 FI2EPI [3]
FI2EPI is a program thoroughly described in Pereira et al. [3], so for this thesis it will only the surface
matter will be described just to show the difference with AEA.

2.1 The input data

FI2EPI needs information about all the process streams that undergo heat exchange along with the
utilities that are used. This information consists of the begin and end temperatures, MCp value, heat
transfer coefficient and, in case of the utility, the cost. The existing network can also be filled in if there
is one. . The existing network is designed by typing in the temperature intervals of both streams of each
heat exchanger. The economic data must also be filled in. This includes the constants necessary to
calculate the cost of a heat exchanger, which can be done separately for the process streams and the
hot and cold utilities.

𝑐𝑜𝑠𝑡 = 𝑎 + 𝑏 ∗ 𝑎𝑟𝑒𝑎 𝑐

3
The other data is given in the economic data is to calculate the annual cost of the heat exchanger and
heat exchanger network.

2.2 Features
FI2EPI can, after filling all the data in, show a lot of information such as the composite curve, the heat
cascade along with the global composite curve and the costs for the existing HEN (=heat exchanger
network) and the HEN without integration. In diagrams it’s possible also to make up some different
scenario’s when there are multiple hot or cold utilities present. These scenario’s can be compared and,
with the information given about them, help the user decide how to construct the MER (=Minimum
Energy Requirement) Network. It’s also possible to find the optimal minimum temperature difference
for the process in terms of cost.

Figure 2: The necessary input for a stream (left), example of two scenario's getting compared (right).

2.3 Constructing the MER Network

The MER Network can be easily constructed by the use of a design matrix. A heat exchanger is created
by checking the box for the stream that lies in the same row and the stream that lies in the same column
of that box. There are two MER Networks each divided in a part under the pinch and a part above the
pinch. Both of them will also notify the user if the pinch rules are satisfied and won’t allow the user to
start working on it till all the problems are fixed. The first MER Network is a more automatic design that
recommends which streams to connect by marking the box in yellow and will automatically fill in the
value by itself when checking that box. It’s also quite restrictive. It won’t allow the user to break any of
the pinch rules and the only heat loads that can be decided are that of the non-recommended boxes.
Also, it’s not possible to use multiple utilities here. The second MER Network does not have these
restrictions and is useful for people who already have a bit more knowledge on heat integration.

Further changes to the MER Network can be made in the MER Evolution Network. Here the loops of
the heat exchanger network can be removed. Loops are circuits that can be formed in a heat exchanger
network. Figure 3 (right-down corner) shows an example of a loop in the MER Network. A loop exists if
it possible to start from one heat exchanger and follow a closed path through other heat exchangers to
then end at the same heat exchanger. The existence of a loop means that there are more heat
exchangers present than there technically need to be.

4
FI2EPI allows the user to remove the loop by automatically deleting the heat exchanger with the lowest
heat load and dividing that load over the other heat exchangers. This can cause (minimum) temperature
difference problems which FI2EPI will notify the user off and solve on its own. There are a lot of options
and it’s possible to go through each possible scenario, so that all these possible scenarios can then be
compared along with the MER, existing and network without integration in the table of scenarios, where
it can be determined then which scenario is the best.

Figure 3 MER Network (left) MER Network Evolution (right).

3 ENERGY ANALYSIS IN ASPEN PLUS [4]

Aspen Energy Analyzer works together with Aspen Plus. Any process designed in Aspen Plus can be
carried over easily to AEA where it can be heat integration network can be redesigned. In order to do
this one has to first use the Activated Energy Analysis tab. After this the program will give information
about the existing heat integration network and possible improvements it has found. It’s possible to make
some alterations in Aspen plus itself in the energy analysis environment but these options can also be
found in AEA as the retrofit options which will be later discussed.

Figure 4: The Activated Energy Analysis tab found above the workflow.

By going to the energy analysis environment and clicking on details, such as in Figure 5, a file is created
for AEA which contains all the necessary information.

Figure 5: The way to get Aspen Plus to export the design to AEA.

5
4 ASPEN ENERGY ANALYSIS [5]

4.1 How to design in AEA

By using Aspen Plus the necessary data has already been filled in. If for whatever reason some of the
data needs to be changed this can be easily done by going to the scenario and from there to data where
it’s possible to change the process and utility streams and the economics of the heat exchangers.

All the work is done on the grid diagram. Unlike FI2EPI everything is done manually. This diagram
shows all the streams on it. when a stream is a striped line or partially a striped line it means that the

heat integration for that stream still isn’t finished. The pinch points are shown by clicking on .

Heat exchangers connecting streams are added by left clicking and holding over one of the streams,
then releasing it and by right clicking then on the node that has now appeared on the stream and holding
it till it’s on the other stream.

Figure 6: Example of the diagram with an unfinished design.

Information about the heat exchanger is filled in by double clicking on either one of the nodes. This
opens a window where the data can be filled in. The boxes on the inside are the temperatures for this
heat exchanger while the boxes that lie on the outside show the temperatures for neighboring heat
exchangers or, in case there aren’t any, the end or begin temperature of the stream. It’s possible to tie
the temperatures. This will connect the temperature from the heat exchanger to that of the neighboring
heat exchanger or begin or end temperature of the stream.

There need to be three temperatures filled in, in order to fully specify the heat exchanger. By filling in
the duty or area it’s not necessary to fill in all three temperatures just one for each stream. The blue
colored text shows that these temperatures have been specified by the user.

6
Figure 7: Example of a heat exchanger window before and after filling in the data.

Splitting is done by clicking with the right mouse button on and holding the button till the mouse is
above the right stream. This will cause a node to appear onto the stream. By double clicking on the node
the split is formed. If there are already multiple heat exchangers on the streams it’s possible to just get
the part of the heat exchangers that need to be split on it. This can be done by clicking with the left
mouse button on the node, holding it, dragging it across the part that needs to be split and then releasing
it. The split editor is opened by double clicking on either the place where the streams are split or
converge back. Here the flow ratios of the streams and temperatures of the split can be filled in. The
temperatures are only able to be defined when there are also heat exchangers present outside of the
split. Extra branches for the split can also be added in the split editor. It’s also possible to split onto an

already split stream by using again. to delete a branch right click on it and select delete branch. To
delete the entire split right click either the place where the streams where split or where they converge
back and select “delete split”.

By right clicking on the design there is the option to let aspen show the loops and paths that exist in the
design. Unlike FI2EPI the loops have to be removed manually. This is done by removing the heat
exchanger with the smallest load and transferring the load of that heat exchanger onto the other heat
exchanger in the loop. In case of a breach of the minimum temperature the paths can be used to see
onto which heat exchanger, that is connected to a utility, the duty can be transferred to.

4.2 How to improve in AEA

There are several options in AEA that can help with creating a good design. One of them is
recommended designs where an entirely new design is created by AEA. To do this go to a case, then
right click on it and choose recommended designs. This will open a window where it’s possible to
overlook the data (streams, utilities, economic data) and specify if there are any streams that can’t be
matched together, the maximum amount of branches each stream can have

and how many different designs can be made. The given designs are often not as good as a self-made
design but can offer ideas on what might be changed in the existing design.

Retrofit is a good option for improving an existing design. Retrofit allows AEA to propose changes to a
design where the user can choose what type of change is made. Each type op change will add a letter
to the name so it can be easily seen what type of changes have been made to the design.

7
Table 1: the different possible steps of retrofit and their respective letters that are added to the design name.

Step Name
Modify utility heat -#U
Move one end of a heat exchanger -#S
Move both ends of a heat exchanger -#P
Add a heat exchanger -#N
Move area -#A

If the design is mostly finished and the goal is just to see if the design can be improved without changing

too much optimization is a good option. This can be done by clicking on and deciding whether it
should change the design to get the lowest possible total cost or to get the lowest possible area.

5 FI2EPI VS AEA
When improving an already existing process it’s important to take into account that there are already
existing heat exchangers. It will save a lot of money if those heat exchangers can be efficiently re-used
instead of buying new heat exchangers. This requires some flexibility in the design that FI2EPI doesn’t
allow. In FI2EPI it’s quite easy to make a design based on pinch analysis and then to go from there to a
minimum heat exchanger design that will break some pinch rules but those options are pre chosen and
often there are more options needed in order to get the best design. The best design for FI2EPI was for
instance after removing the loops and increasing the utility load on both sides. Even though the increase
in utility load was a lot, the area of a heat exchanger became just small enough in order to be used by
the largest existing heat exchanger which saved a lot of money on buying a new heat exchanger. This
design was one of the possible minimum heat exchanger designs and was found that way. With AEA
however a better design was found. The better design kept the heat exchangers from the pinch analysis
but diverted just enough heat load to the utilities to get that one heat exchanger to be just small enough
so that the existing heat exchanger could be used for it. This design was found by first getting to a
minimum heat exchanger design and then using retrofit to add extra heat exchangers while making sure
that the area of that one heat exchanger didn’t increase too much. The heat exchangers that were added
by AEA where the heat exchangers from the pinch analysis, which proves the effectiveness of pinch
analysis.

Retrofit is a really good tool to use for improving an existing design because it allows the user to specify
the max new area that can be added when making a change, ensuring that an heat exchanger won’t go
above the area of an existing heat exchanger.

8
Figure 8: The pinch analysis design (above), the better design from FIEPI (left) and the even better design from
AEA (right).

6 IMPROVING A BIOMASS PROCESS [6]

In order to see the usefulness of AEA against a real life process a proposed biomass process to create
xylitol and polylactic acid from brewer’s spent grain was used. This is a fairly complex process that uses
a lot of heat exchangers and has some type of heat integration already applied to it making it perfect to
test for AEA.

6.1 Explanation of the process

The first step of the process is getting the BSG (=brewer’s spent grain) pretreated. This is done by
hydrolyzing the hemicellulose in a reactor. This results in xylose and arabinose, both sugars, that will
be in solution while the cellulose and lignin won’t, allowing for separation via filtration.

The sugars will then get to a fermenter where the xylose will be converted to xylitol. After this the stream
gets purified, removing the toxic yeast that was used. A part of the stream gets filtered and recycled
back to the fermenter to gain a higher yield of xylitol.

The stream that contains the cellulose and lignin is brought to a fermenter where L-lactic acid is formed.
After filtration the L-lactic acid gets turned to PLA (= polylactic acid). This is done via direct condensation
polymerization. This is done in turn by using a reactor, evaporator, crystallizer and flash vessel. Multiple
solvents also have to be used like diphenyl ether, dichloromethane and methanol. The two last ones are
quite expensive and are therefore recovered at the end with the use of a distillation column.

The process is first created in Aspen plus. All the information to make this can be found in work of
George at al. [6]. Some slight alterations, like the way it does recycling, were made but for the rest it’s
the exact same design.

9
Figure 9: The biomass process shown in Aspen Plus.

6.2 Designing the MER Network

An existing process will have segmented streams. This means that the MCp value changes too much
for us to be able to use one constant MCp value as an approach for the real MCp value, so multiple
constant MCp values are taken to more closely resemble the reality. The reason for the need for these
segments is because of the heat exchangers that were used in the existing process and the phase
changes that happens in the process. This means that when making a design from the ground up, like
with pinch analysis, these segments have to go away because the only reason they are there in the first
place is because of heat exchangers that aren’t there in the new design. AEA has a simple option where
it will delete the segments automatically but this isn’t accurate. This is because the begin and end
temperature are necessary along with the total heat Q of the stream to make the unsegmented stream.
AEA however uses the MCp values of the segments of the stream to get the total heat. It’s possible to
get the total heat by using the MCp of the stream but it would be wrong to use the MCp value of the
segments. This is because the segments are an approximation of the reality, so when calculating the
heat load from these MCp values there will be a difference between the calculated and actual heat.

This Process had many different hot and cold utilities causing multiple pinch points. This design was
solved almost completely with pinch analysis. The only problem is some pinch crossing with the cold
utilities because there didn’t seem to be any other way possible. The resulting design is a true MER
(=Minimum Energy Requirement) Network though which can be seen in the performance (see Figure
10 right). The target values are the values of a true MER Network where there is no crossing as well for
the utility pinches. It can be seen in the performance that the heating and cooling is 100 percent that of
the target, so it’s an MER design. If the percentage is lower than 100 percent it would mean that for one
or more of the heat exchangers the temperature difference went under the minimum temperature
difference. In the cost indexes it can be seen that the heating costs are that of the target but not the
cooling cost because some cold utilities did cross the pinch lines.

10
Figure 10: Finished MER design (left) and the performance and cost indexes of it (right).

When comparing this design with the existing design it can be seen that there are some values that go
below the 100 percent mark of the target. This is because the process was a proposed one more to
show of how a biomass process could be made that uses a cheap raw material and results in products
that can be sold for quite a high price. Stuff such as minimum temperature difference wasn’t the main
concern here. Even with that advantage the MER Network still has a lower operating cost which is the
main point when optimizing a design.

Figure 11: Existing network (left) and the performance and cost indexes of it.

This design is on its own better than the existing one but will never be able to be applied because it’s
too different from that design. This is the main problem of making an MER Network. It’s better to work
from the existing design and use retrofit options to improve it while using the area specifications to make
sure that the biggest heat exchangers don’t go bigger so new ones have to been bought. Unfortunately
this couldn’t be done for this design because retrofit couldn’t find any possible improvements for the
existing network even though there clearly are. This could be because the process is to complex.

6.3 Improving the existing design

The existing design was manually improved as well. Most of the big heat exchangers where left
unchanged so as to not create too much capital cost. The result is a design with a capital cost of
19.155*104 €, savings of 5.3394*104 €/year and a payback time of 3.5875 years. The design didn’t make
any changes to devices such as reactors so that this design could be easily implemented back in Aspen
Plus. There where however problems with the design in Aspen Plus. This is caused by the segments.
The segments need to be kept for the heat exchangers that are unchanged but these cause problems
with the new heat exchangers.

11
Ideally the segments would be only kept for the old heat exchangers but this is difficult to do and even
then problems might show up because, unlike AEA, Aspen Plus takes the components and thus also
their phase shifts into account. These phase shifts result as well in segments.

Figure 12: Existing network with only streams (left) and improved network with only streams (right).

7 CONCLUSSION
AEA is definitely a better program, for optimizing a process, than FI2EPI. FI2EPI is a good starting point
for someone who knows nothing about heat integration with it’s simple design, intuitive controls and
helpful hints. FI2EPI is also a good program to use for simple small processes. It starts to lose its
usefulness quickly however when the amount of streams increases or when there are multiple utilities.

AEA requires more knowledge on heat integration but delivers also more information allowing for a
better control of the results of the design changes. AEA is also handy because it allows the user to apply
constraints, such as for instance which streams can’t have heat exchange. In a large process this is
definitely handy because often it would be impossible to have certain streams performed heat integration
if they are located far away from each other. This is also very useful for safety. Because the process
that was utilised for this thesis wasn’t one that was used in practice it wasn’t possible to make certain
practical constraints like that in the design. In real life this would however be the case.

From the optimization features retrofit is definitely the handiest, allowing to make several constraints
and specifications and being able to choose the type of optimization that is applied. This feature is
recommended to use when improving a complex design. Unfortunately this feature seems to have its
own problems. It’s not entirely sure if this is due to the complexity of the design but if this is the case it
would significantly decrease the practical use of this feature since most processes are this complex.
Having said this AEA is still a very handy program allowing the user to easily implement a process from
Aspen Plus.

The MER Network created for the biomass process was an improvement over the existing process but
is not the best design to replace the existing design with because it’s so different and would cause a lot
of capital cost.

The implementation of an existing network back in Aspen Plus is problematic because of the problems
with the segments. Even when starting from scratch like with the pinch analysis there may be some
problems because of phase shifts causing changes in the MCp value.

12
8 REFERENCES
1. Morar, M. and P.S. Agachi, Review: Important contributions in development and improvement
of the heat integration techniques. Computers & Chemical Engineering, 2010. 34(8): p. 1171-
1179.
2. March, L., Introduction to Pinch Technology. 1998.
3. Pereira, P.M., et al., FI2EPI: A heat management tool for process integration. Applied Thermal
Engineering, 2017. 114: p. 523-536.
4. Aspen Plus. 2017, Aspentech.
5. Aspen Energy Analysis. 2017, Aspentech.
6. Alexander George, K.S., Anthony Carradorini, Nabila Faour, Brewer
's Spent Grain to Xylitol & Polylactic Acid. 2017.

9 KEYWORDS
Aspen Energy Analyzer, Heat Exchange, pinch analysis, FI2EPI, Xylitol Proces

13