I

Pinch Technology/
Process Optimization

Volume 8: Case Study - United Refining Company

CEC Report CR-105239

Prepared by
TENSA Services, Inc.
6200 Savoy Drive, Suite 540
Houston, TX 77036-3315
May 1995

CPe 0521

REPORT SUMMARY
PINCH TECHNOLOGY/PROCESS OPTIMIZATION
Volume 8: Case Study - United Refining Company
The case study of United R e f i g Company's fuel oil refinery at
Warren, PA demonstrates how process integration or pinch
technology can identify practical and cost-effective ways to
substantially reduce energy costs. Suggested cost-saving measures
include steam and power system improvements, optimum heat
exchanger network design, furnace system improvements and use of
a heat pump. The mechanical vapor recompression heat pump
would save 11,000 pounds per hour of steam. Replacement of
steam drives with electric motors would add about 2 MW in
electricity usage.

BACKGROUND Improved industrial process efficiency is of

INTEREST CATEGORIES

Industrial
Demand-side Planning

KEYWORDS

Pinch technology
Heat recovery
Energy efficiency
Heat pumps
End use
Industry

great importance to electric utilities. It enhances customer

competitiveness and profitability, thereby fostering load
retention and strategic load growth. By understanding the
energy use pattems and options at an industrial site, the utility can
work together with its customer to define mutually beneficial
investment and operating options. The technique of
choice is pinch analysis, an innovative and effective
method for analyzing industrial sites. Since 1988,
EPRI and member utilities have cosponsored over twenty
such studies around the country in various industries, with a
high degree of success.
I
_

OBJECTIVES To identify opportunities for energy savings using

pinch technology; to develop technically and economically viable
projects to achieve these targets.

---APPROACH Project team was formed consisting of consultants,

plant and electric utility representatives. The team visited the plant
to define and collect process, utility and economic data. The
consultants developed appropriate material, heat and steam
balances, and using pinch technology, characterized each refinery's

heating and cooling needs. After quantifying the scope of potential

improvements, the site was screened for specific projects based on
processing changes, heat recovery and heat pump applications.

RESULTS The study indicates that substantial operating cost

savings could be achieved using conventional technologies and
investment payback criteria(typical1y two years or less). Projects
were identified to reduce energy cost, switch from steam to electric
drives, use heat pumps and reduce NOx.

EPRI PERSPECTIVE The study shows that process integration

or pinch technology is an effective tool to improve industrial energy
efficiency. Utilities can use pinch methodology to promote load
stability in their service territories. Information on energy use
options and interactions and their sensitivity to economic factors
can also be used to foster successful demand-side management
programs.
EPRI has published additional case studies of pinch technology on
a variety of industries. These are documented in reports TR101147, Volumes 1 through 5. Other related work is documented
in EPRI reports EM-6057, CU-6334, CU-6775, CR105237, and
CR105238. An EPRT brochure on pinch technology is numbered
BR-102466.

PROJECTS
RP3879-01
EPRI Project Manager: K. R. Amamath
Customer Systems Division
Contractor: Tensa Services, Inc.
For further information on EPRI research programs, call EPRI
Technical Information Specialists (415) 855-241 1.

PINCH TECHNOLOGY/PROCESS OPTIMIZATION

Volume 8: Case Study

--

United Refining Company

Final Report, March 1995

Prepared by
TENSA Services, Inc.
6200 Savoy Drive, Suite 540
Houston, Texas 77036-3315

Prepared for
Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304

For technical information, contact:

EPRl Chemicals & Petroleum Office
1775 St. James Place, Suite 105
Houston, Texas 77056
(713) 963-9307
Fax: (713) 963-8341

For ordering information, contact:

EPRIAMP Customer Assistance Center (ECAC)
1-800-4320-AMP

Disclaimer of Warranties and Limitation of Liabilities

NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, NOR ANY PERSON OR ORGANIZATION
ACTING ON BEHALF OF ANY OF THEM:
(A)MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED,
(I)WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR
ITEM DISCLOSED IN THIS REPORT, INCLUDING MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE, OR
(1I)THAT SUCH USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY OWNED RIGHTS,
INCLUDING ANY PARTY'S INTELLECTUAL PROPERTY, OR
(1II)THAT THIS REPORT IS SUITABLE TO ANY PARTICULAR USER'S CIRCUMSTANCE; OR
(B)ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER (INCLUDING ANY
CONSEQUENTIAL DAMAGES, EVEN IF EPRl OR ANY EPRl REPRESENTATIVE HAS BEEN ADVISED OF THE
POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR SELECTION OR USE OF THIS REPORT OR
ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS REPORT.

ACKNOWLEDGMENTS

TENSA Services, Inc. appreciates the interest and support of Pennsylvania Electric
Company (Penelec), the Electric Power Research Institute (EPRI) and the United
Refining Company (URC), throughout the course of this study. Special thanks to Mr.
Geoffrey S. Soares, Robert Ennis, David Dorn and their colleagues at URC, for their
excellent cooperation during the data collection and validation phases of this study.
We are also grateful to Gary Wareham and David A. Pascale of Penelec, who provided
the valuable electric distribution rate data.

CONTENTS

Section

Page

Introduction ..............................................................................................................

1
.

Furnaces ...................................................................................................................
Crude and Vacuum Furnaces Operation......................................................
Furnaces Excess Air Reduction ...................................................................
Recommendationsfor Flue Gas Analyzers ..................................................

2-2
2-1
2-5
2-7

Steam System .........................................................................................................

Introduction .................................................................................................
Recommendations Toward Steam System Efficiency ..................................
Summary ......................................................................................................
Description of Modeling of URC Steam System ...........................................

DIB Column Heat Pump.......................................................................................... 4 1

Introduction ..................................................................................................
4-1
Base Case ....................................................................................................
4-1
Heat Pump and Pressure Reduction Operation ........................................... 4-4
New Column Configuration ..........................................................................
4-7
New Configuration Pressure Reduction and Heat Pump Evaluation............4-7
Summary ...................................................................................................... 4-9

FCC Main Fractionation Unit.................................................................................

Process Description .....................................................................................
Analysis of the Current Base Case Operation ...........................................
Optimum Energy Recovery ..........................................................................
Discussion..................................................................................................

3-1
-3-1
3-1
3-5
3-8

5-1
5-4
5-4

5-4
5-12

Appendix A ...................................................................................................... A-1

Introduction to Pinch Technology ................................................................ A-1

EXECUTIVE SUMMARY

A detailed process efficiency optimization study using Pinch Technology of the United
Refining Companys facility at Warren, Pennsylvania was performed. The study was
co-sponsored by the Electric Power Research Institute (EPRI) and the Pennsylvania
Electric Company (Penelec).
The Scope of Work contains refinery wide steam and condensate system modeling and
optimization, heat integration of FCC unit and DIB column, operation and efficiency
improvement of crude and vacuum furnaces, and efficiency improvement of other
furnaces in the refinery. In this study, TENSA Services, Inc. applied pinch technology
along with other energy conservation strategies to identify significant savings for United
Refining Company (URC). The summary of these findings are listed below:
Heat Pump on the DIB column can save over 11,000 Ib/hr of steam usage with the
additional benefit of increasing the column capacity by 25%, and also reducing the
flare system load.
Replacing steam turbines with electric motors can save URC 68,000 Ib/hr of steam.
Modification on the reformer deaerator including piping change and flash drum
installment, can save 3,000 Ib/hr of steam and also unload the deaerator.
Modification on wash water piping can save 5,000 lblhr of steam usage.
A heat integration option in the FCC unit can save 1.6 MMBtu/hr of fuel and reduce
the wet gas compressor power consumption by 224 KW.
Excess air reduction in the furnaces can save 28 MMBtu/hr of fuel usage.
With the above projects implemented, URC can shut down some of the existing low
efficiency boilers.
The identified projects are also summarized in Table ES-1. After implementing some of
these recommendations, the refinery could reach excess-fuel gas situation. Even if the

ES-1

credit for saving only 50,000 Ib/hr of steam is considered, it will still result in a yearly
savings of over one million dollars per year for the refinery.
Total savings in steam ($/yr) = 2,150,000
Additional cost of electricity ($/yr) = 726,200
Net annual savings: $1.4 millions/yr.

ES-2

Table ES-1
Calculated Value of URCs Savings
Item

Savings in
Steam
(1000 Ib/hr)

Savings in
Fuel
(MMBtu/hr)

Additional
Electricity
Consumption
(KW)

Estimated
Installation
cost

Payback
Period

(YW

Replacing steam turbines

68

-0-

1,945

Heat pump on DIB

11

-0-

320

603,000

1.6

Reformer Deaerator

-0-

70

140,000

1.3

Water wash repiping

-0-

-0-

100,000

0.5

180,160

1.2

FCC Heat integration

1.6

-224

Flue gas analyzer

28

-0-

Total

87

29.6

*Estimated 2 years or less in payback

ES-3

2,111

2*

2*

INTRODUCTION

The objective of the study is to apply Pinch Technology to identify ways to reduce the
energy consumption in the plant and simultaneously debottleneck its capacity. Pinch
Technology is a comprehensive analytical tool for analyzing processes to calculate the
optimum amount of energy required to operate a process. These calculations are
based on basic thermodynamic concepts. Practical constraints related to the process
may be incorporated in the analysis to predict the minimum energy required to run the
process and meet these constraints at the same time. Application of Pinch Technology
also enables calculation of the capital investment required to realize these savings and
an accurate estimation of the payback period for suggested retrofit schemes. An
introduction to Pinch Technology is provided in Appendix A.
Though originally developed for optimizing Heat Exchanger Networks (HEN), pinch
technology is now being applied for the integration of the various unit operations such
as distillation, evaporation, drying, etc. Pinch technology has gained recognition as a
valuable process design tool and has pointed the way to many energy-saving ideas in
the process industries.
The Scope of Work for this study is comprised of Deisobutanizer (DIB) column in the
alkylation unit, Fluid Catalytic Cracking (FCC) main fractionation unit, steam and
condensate system, furnaces & boiler efficiencies.
Section 1 of this report describes the background information of this study. Section 2
deals with findings in the area of furnaces efficiencies. Section 3 focuses on the steam
and condensate system. DIB heat pump evaluation is presented in Section 4. Finally,
the FCC heat integration study is discussed in Section 5. The economic factors in
evaluating various project feasibility are summarized in Table 1.1. These values are
used throughout this study:
Table 1.1
Utility
Steam cost:
Cooling water cost:
Condensate cost:
Annual average electricity cost:
Annual operating hours: 8600 hr/yr

cost
$91000 Ib.
$0.3/MMBtu
$0.3/1000 I b.
$O.O4/KWH

1-1

2
FURNACES

Introduction

URC has many fired heaters including crude & vacuum heaters, Fluid Catalytic
Cracking (FCC), reformers, Naphtha HydroTreater (NHT), and Deisobutanizer (DB)
reboiler. This section discusses the operation of these heaters for efficiency
improvement through excess air reduction. More effort has been put on the crude and
vacuum furnaces, since they are the biggest fired heaters in the refinery.
Crude and Vacuum Furnaces Operation

The existing crude and vacuum furnaces configuration is shown in Figure 2-1. The
vacuum flue gas is ducked to the crude convection section. The upper 5 rows of the
process coil in the crude convection section are used to preheat the crude column
bottoms, before it gets heated further in the radiant section of the vacuum furnace.
Before the scheduled turnaround, the performance of the existing configuration was
evaluated to see the effect of furnace decoking and possible separation of the upper 5
rows from crude bottoms preheat as shown in Figure 2-2.
Table 2-1 summarizes the results of the evaluation for three different conditions: the
existing fouled (before decoking), the existing clean (after decoking), and the clean and
separate.
The evaluation was based on rigorous calculation on the radiant and convection
sections at reasonable bridge wall temperature and design heat transfer area. The
total duty provided to the crude feed was determined to be about 150 MMBtu/hr.
through feed simulation. The vacuum furnace duty was calculated to be 20 MMBtu/hr.
From Table 2-1, it is apparent that decoking does not improve the furnace efficiency
significantly in this case. However, the major benefit of the decoking procedure is to
lower the tube metal temperature. Coking rate is highly exponential and if not decoked
at the appropriate time, it could plug the tubes solid. Even before it reaches that stage,
the tube metal temperature may reach a point where it may reduce the coils life.

2- 1

0
0

0
0
0
0

0
0
0
0

Y)

0
IZ

0
0
0
0
0

0
0
0
0
0000000000000000
0
0

LL

C-I
Q
v)

0
-I

0
E
LL

- a i

iI

00

oe-

00

om-

d
4
a
W
m

Fumaces

Table 2-1
Summary Crude / Vacuum / Furnaces / Analysis

Existing Fouled

Existing Clean

Crude
Furnace
(MMBtu/hr)

Fuel Fired = 236.2

Rad. eff. = 0.494
Q crude conv = 33
Q vac mnv = 12
Rad. Flux = 12,000*

Fuel Fired = 236.6

Rad. Eff. = 0.5029
Q crude conv = 31
Q vac mnv = 12.4
Rad. Flux = 12,234*

Fuel Fired = 210.8

Rad. Eff. = 0.5171
Q crude conv = 41
Rad. Flux = 11,206"

Vacuum
Furnace
(MMBtulhr)

Fuel Fired = 11.96

Rad. Eff. = 0.6685
Rad. Flux = 2383*

Fuel Fired = 11.I

3
Rad. Eff. = 0.6829
Rad. Flux = 2264*

Fuel Fired = 32.4

Rad. Eff. = 0.6171
Q vac conv = 0
Rad. Flux = 5958*

Steam
Generation
(MMBtu/hr)

Clean Separate

Steam Generated = 27 Steam Generated = 28.3 Steam Generated = 26.5

Flue Gas
Temp.

755F

750F

735F

197

198.3

196.5

Total Heat
Released
(MMBtu/hr)

248.16

247.73

243.2

Overall
Efficiency

0.7938

0.8005

0.808

$6.454 million

$6.392 million

$6.323 million

$62,00O/yr

$131,00O/yr

Total Heat
Absorbed
(MMBtu/hr)
(CRU + VAC + STM)

Annual Fuel
Cost (8,600
hrs)
Savings
(* Unit in Btu/Hr - Sq. ft.)

2-4

, Fumaces

It is also obvious from the summary table that each service should be performed in its
own furnaces. There is a net savings of $131,00O/year, if they operate individually.
The separate configuration provided flexibility to increase capacity in the crude and
vacuum furnaces. The flux density will drop to below the designed value of 12,000 and
7,000 for crude and vacuum furnaces respectively.
From the analysis it appears that the existing crude furnace may have problems
meeting the overall efficiency shown on the data sheet.
The burners on both furnaces were also examined. There are 28 burners in the crude
furnace. Twelve of those are from John Zink with 15 MMBtu/hr. maximum heat release.
The remaining sixteen (16) are made by NAO, with 8.98 MMBtu/hr of maximum heat
release. The furnace should have only one size and type of burner to get uniform
flame pattern at the same fuel pressure and also to avoid mix up in burner tips. URC
should replace the existing sixteen (16) NAO burner with John Zink burner. This will
give a total maximum firing of 420 MMBtuIhr. Since the designed operation will be only
50% of the burner capacity, these burners will not be fired at full capacity. To achieve
the desired turndown, URC will have to shut some of the burners.
The vacuum furnace has nine (9) John Zink burners each with maximum heat release
of 5.39 MMBtu/hr. The total maximum heat release is 48.5 MMBtu/hr. By examining
the drawing of the vacuum furnace, it is noticed that the top tubes are only 17-0 above
the burners. It is recommended that only gas firing be used in the vacuum furnace to
minimize the flame length, and flame impingement on the top tubes and tube supports.
URC should explore the possibility of changing the arrangements of four (4) top tubes
and their support system, to avoid direct flame impingement on the tube. If new
burners of higher capacity are considered, they should preferably be all gas fired along
with the above changes.
Furnaces Excess Air Reduction
URC has a number of fired heaters for its process heating needs. Optimizing
combustion efficiency and minimizing exhaust emissions are vitally important for proper
operation of these furnaces.

As most combustion equipment operator knows, it is extremely undesirable to operate a

burner with less than stoichiometric combustion air. Not only is this likely to result in a
smoking stack, but it will significantly reduce the energy released by the fuel. In actual
application, it is impossible to achieve stoichiometric combustion, because burners can
not mix fuel and air perfectly. To ensure that all of the fuel is burned and that little or
no combustibles appear in the flue gas, it is common practice to supply some amount of
excess air. But too much excess air is also inefficient because it enters the burner at
ambient temperature and leaves the stack hot, thus stealing useful heat from the
process.

2-5

Furnaces

Table 2-2 illustrates the savings by reduction in excess air in the crude furnace. The
Table compares various percentage of excess air to 20% excess air. It shows that fuel
consumption can be reduced by 20 MMBtuIhr, simply by reducing excess air from 80%
to 20%, as an example.
But how is the correct amount of excess air determined? By far the most widely used
method is flue gas analysis.
Table 2-2
Effects of Excess Air On Fuel Savings For The Crude Furnace

755
755
Flue Gas Temp (OF)
% Excess Air
20
30
Efficiency (%)
81.71
80.57
Efficiency change (%)
0
1.I4
Add. heat available
0
2.83
(MMBtu/hr)
Annual Savings ($/yr)
0 85,182
Based on 180 MMBtu/hr, furnace duty.

755
40
79.14
2.57
6.38

755
50
78
3.71
9.21

755
60
76.51
5.2
12.9

755
80
73.54
8.17
20.27

192,036

277,218

388,286

610,121

By measuring the actual concentrations of several flue gas constituents, as well as the
flue gas temperature, both unburned fuel loss and flue gas heat loss can be
determined. It is then possible to control the supply of excess air so that combustion
efficiency will be maximized.
There are basically two types of flue gas analyzers available Le., extractive and in-situ.
The extractive method takes a sample from the stack, conditions it, and then analyzes
it. The in-situ method is where the sensor is mounted in the stack and typically,
because of high temperature, a Zirconium Oxide sensor is used to measure oxygen.
The in-situ instruments are normally cheaper than extractive instruments. But there are
many advantages for extractive analysis:

1. With the in-situ method, the following components may cause incorrect reading
and/or damage to the Zirconium Oxide sensor, such as: moisture, SOx, dust, freon
and fluorine gas, combustible materials, silicon compounds, HCI and Cla. With the
extractive method, harmful substances are removed by sample conditioning prior to
analysis.
2. Since the in-situ sensor is directly in the stack, process must be shut down for
sensor repair. The extractive method is remote from stack and repair will not
hinder stack operation.

2-6

,Fumaces

3. As the in-situ sensor must be calibrated remotely, this is typically done with
microprocessor-based electronics. These electronics are typically beyond the
scope of field repair and must be sent to the factory for repair, resulting in lengthy
instrument downtime.
Recommendationsfor Flue Gas Analyzers

A) An extractive flue gas analyzer system is recommended for the crude furnace,
vacuum furnace, and the debutanizer (DB) fired heater. These three furnaces are
in the same area. A single extractive analyzer with up to twelve (12) sampling point
digital sequencing, can be used to analyze the flue gas from each probe point
periodically, according to selected time intervals. This system is suitable for Class I,
Division 1 area classification. Figure 2-3 illustrates the position on the furnace
where the sample probes are to be inserted. The sample points in the radiant
section can be used to measure the excess air. The sample points in the stack are
used to evaluate the amount of air leakage to the stack.
8) Other Furnaces and Boilers. For other fired heaters and boilers in URC, a portable
analyzer is recommended for quick spot checking.
It is important to remember that just to have these analyzers installed is not sufficient to
guarantee adequate excess air percentage and improved combustion efficiency. The
operators have to be trained to pay attention to these measurement and to take proper
action once an abnormal condition occurred.
Based on a plant wide average of 50% excess air, which if reduced to 20% results in a
fuel savings of 28 MMBtu/hr. The estimated payback for the installation of flue gas
analyzer is less than 2 years.

2-7

$
m

Q)

E
a

LL

E
a
a

Q1

2
vr

E
a

I
LL

Q)
n

3
STEAM SYSTEM

Introduction

URC consumes a large amount of steam in various plant operations. The steam
system is modeled on the EPRl software APLUS (Analysis of Plant Utility System),
written by TENSA Services, to obtain an overall picture of the system. The APLUS
program solves the heat and mass balance of the entire steam system. It calculates
the BFW make-up water requirement based on the prescribed boiler steam generation,
blowdown percentage, and condensate return condition. The steam and water flow
rates are in Ib/hr in Figure 3-1. A detailed description of the modeling is attached at the
end of this section. From Figure 3-1, it is obvious that the steam system is not
operating efficiently in the following areas:
1. The percentage of condensate recovery is too low. Only 25% of the condensate
returns to the boilers. The low return causes more make-up water and steam
requirements in the deaerator, which eventually reduces the overall efficiency of the
steam system. Condensate recovery is a major area of improvement that need to
be focused on.

2. There is 49,765 Ib/hr of steam lost by the back pressure turbine exhausting to the
atmosphere. This loss of steam to the ambient also contributes to the overall low
condensate recovery of the steam system. Replacing these turbines with electrical
motors will result in big savings for the plant.
3. The estimated 69,116 Ib/hr of tracing steam consumption is large for this size of
refinery. Steam must be leaking through steam traps on the tracing.
Recommendations Toward Steam System Efficiency

Several areas in the steam system can be improved to increase its efficiency. Due to
these large losses, URC is considering installing an additional boiler to meet the
demand. The following strategies are geared toward more efficient usage of steam:

3-1

,115,800

560

20326
6712

b7500

+
All Steam Users
Condensates
Are Dumped

Reformer
Dea.

h057S.

691 16

Tracing

953

condensates
Goes to
Sour Water

43088

995

9000
Condensate Loss

4
52396

7I

373766

MainDea.
d116C

BFW Make-up (110 F)

Figure 3-1
Steam System Flow Chart

38000
1

100 GPM

Poly. Wash Alky. wash,

Gasoline Wash,OVHD. Exe.

Steam System

1. Replacing steam turbines with motor drive

a. Some turbines which are running 100% of the time, are exhausting to the
atmosphere. The following list summarizes these candidates:

I b/hr

Service
LVGO
VAC BOT
Sour Water
MCWCR
Crude Top Reflux
#8 Well
Alky Comp. Hot Well
#4 Boiler Fan
#4 Boiler Steam Drum
Total

HP
9
15
2
20
12
95
18
109
22
302

485
790
100
1,050
675
5,250
1,000
6,000
1,200
16,550

These replacements are estimated to payback in less than two years.

b. Some turbines which are used as back up and are used only 50% of the time
are listed below:
Ib/hr
3,750
525
1.965
6,240

Crude Pump
Prefrac. Reflux
610 Tank
Total

HP
68
~~

10

36
114

This replacement is estimated to payback in less than 2 years.

c. Turbines which are running only 33% of the time are listed below:

I b/hr
FCC Cooling Tower
MID Cooling Tower
Total

14975
12000
26975

HP
272
218
490

This replacement is estimated to payback in less than 2 years.

d. Air Blower Steam Turbine Replacement.
The condensing turbine to drive the air blower has an estimated actual steam
rate of 1I.02 Ib/HP-hr. This is equivalent to generating electricity at $0.074/Wh.
This is higher than the current rate offered by local utility. This turbine can be

3-3

Steam System

replaced by motor drive. This replacement is estimated to payback in less than

2 years.
There is also potential for converting some of the steam turbines to 15 psig back
pressure turbines to provide low pressure steam to the deaerator. URC should
make judgment on this option after evaluation of the proposed replacement.

In summary, the plant has a potential of saving 67,765 Ib/hr of steam

consumption, by replacing these steam turbines with electrical motors. This will
increase the electricity consumption by 1,945 KVV.
2. Condensate Recovery

From Figure 3-1 , there are several areas of condensate loss which can be
recovered to enhance the condensate return percentage and improve overall steam
system efficiency.
a) Reformer Deaerafor Area.
There is 9,000 Ib/hr of steam condensate being dumped to the river and glade in
the reformer area as shown in Figure 3-2. Condensate is being dumped
because several steam traps and the condensate return system are undersized.
URC is correcting these problems.

Currently, the reformer deaerator works as a flash drum. According to the

APLUS simulation, there is 6,712 Ib/hr of flash steam being exhausted to the
atmosphere. URC is installing a new flash drum to flash off 50 psig steam,
which will be used in the nearby sour water stripper (SWS). This will save an
estimated 3,000 Ib/hr of steam.
The reformer deaerator also needs to be modified to work properly. The
modification includes a new motor with 120 HP to replace the old motor with only
75 HP. No change is required except a large impeller.
All these changes in the reformer deaerator area will cost $140,000. Annual
savings are estimated to be $152,000.
b) Tanks 237,238 area.
There is 24,000 Ib/hr of steam usage for tank heating in Tanks 237 and 238.
The condensate is not recovered. URC is considering installing a new hot oil
heating system to replace the steam coil heating.

c) Steam Turbine Exhaust.

Many of the steam turbines are exhausting to the atmosphere, resulting in
significant loss of condensate returning to the boiler. The replacement of steam

' Steam System

turbines with electric motor will cut down the exhaust steam loss, and increase
the condensate return percentage.

3. Wash Waterusage
Figure 3-1 shows that 100 GPM of water from deaerator at 225F is used at poly
wash, overhead exchanger, alky wash, and gasoline wash. However, only at poly
wash is the hot treated water required. The other usage can be replaced with
ambient treated water. By modifying the wash water piping and adding two water
softeners for water treatment, URC can save 80 GPM of hot water from deaerator.
Annual savings are over $215,000. This change will also ease the over-loading
situation at the main deaerator.

Summary
A new APLUS simulation is summarized in Figure 3-3, assuming all the aforementioned
efficiency measures are carried out. This will leave URC with plenty of steam
generation capacity for expansion and addition of new units. The condensate return
percentage has increased to 41 % from the present 25%. Deaerators steam
requirement dropped from 57,221 Ib/Hr to 25,473 Ib/Hr. These changes also eliminate
the need for a new boiler purchase. URC is already considering recovering heat from
the reformulated gasoline to preheat BFW to 154F. This change is not included in the
simulation.

3-5

Reboilers

Various
Steam
Traps

BFW
Make Up From
Boiler House
Deaerator

N.C.

9,000 lbhr
Condensate being
dumped to river,
glade, grade, etc.,
because of condensate
return line limitations.

Reformer Deaerator
at 30 psig

N.0.

A Boiler Feed Water

7
h

Figure 3-2
Reformer Deaerator

to No. 5 Boilers
Crude
and Reformer
SteamDrums

Steam System

Description of Modeling of URC Steam System

Figure 3-3 illustrates the steam system in URC. There are nine (9) boilers shown on
the diagram to signify all the steam generation equipment in the plant. The detailed
boiler information and steam generation are listed below. This information was
extracted from an Audit Report performed by Betz Industrial Company.
BL3-BL4-BL5-RFB-SUB-CRU-DEL-FU2SSB--

Boiler 1,2,3, 225 # Pressure, I17,780 Ib/hr, 7.69% blowdown.

Boiler 4, 225 # Pressure, 115,800 Ib/hr, 7.69% blowdown.
No. 5 boiler, 300 # Pressure, 34,530 Ib/hr, 7.69% blowdown.
Reformer Boiler, 225 # Pressure, 36,620 Ib/hr, 5% blowdown.
Sulfer boilers, 225 # Pressure, 4,400 Ib/hr, 7.69% blowdown.
Steam generation in the crude unit. Including FLUX SG,HGO SG, and
crude furnace, 225 # Pressure, 40,960 Ib/hr, 7.69% blowdown.
DELTAC boiler, 225 # Pressure, 40,940 Ib/hr, 7.69% blowdown.
#2 Furnace boiler, 225 # Pressure, 6,920 Ib/hr, 7.69% blowdown.
Slurry steam generators, 225 # Pressure, 33,250 Ib/hr, 12.5% blowdown.

The steam users are grouped to simplify the diagram. For example, all steam users
whose condensate are dumped is grouped into FU2. All steam turbines used only 33%
of the time is grouped into ST6. The following list summarizes these steam users in the
plant
ST1--

Condensing steam turbine for the FCC main air blower, 18000 Ib/hr steam
flow rate. Exhausted to graham surface condenser CNI to produce 90" F
condensate.
ET1-- 2 stages steam turbine. 25,000 Iblhr steam flow rate. 16,000 Ib/hr exhaust at
40 # for process usage at PU3. 9,000 Ib/hr exhausts from the 2nd stage and
is then condensed in the condenser CN2 to generate 90" F condensate.
ST3-- ,Include two steam turbines at boiler house (8,250 Ib/hr), and FCC deaerator
(2,750 Ib/hr) to drive BFW pumps.
Steam turbines running 100% of the time and exhaust to atmosphere.
ST4Steam turbines running 50% of the time and exhaust to atmosphere.
ST5Steam turbines running 33% of the time and exhaust to atmosphere.
ST6PUI-- Represents all the stripping steam usage. Condensate goes to sour water.
40,575 Ib/hr.
PU4-- Represents the process users of lsom Regen. Vap.(560 Ib/hr), Vac. Vent
Reboiler (550 Ib/hr), and SRUI Rx. Reheat (450 Ib/hr). Condensate is
recovered and returned to the main deaerator (DAI) through Boiler feed
water mixer (BMI).
RB1-- 58,800 Ib/hr steam flow rate. For all the reboiler usage in the reformer.
Condensate are collected in a splitter SP3, from where about 9000 Ib/hr were
dumped through S21 due to limitation on the size of condensate return line.

3-8

Steam Sysfem

37,500 Ib/hr steam usage. Includes steam users where condensate are
dumped: 237TK coil (12,000 Ib/hr), 238 TK coil (12,000 Ib/hr), pretreater LSR
reboiler (12,500 Ib/hr).
PU3-- 15,350 Ib/hr steam usage. Includes Propane dryer (1,250 Ib/hr), Alky Dgp
Reb (7700 Iblhr), DIB upper reb (5,900 Ib/hr), and olefin feed heater (500
Ib/hr). These condensate are collected and flashed in FL3. Liquid from FL3
is then send to mixer (MX5) along with 90"F condensate from the graham
condenser. The vapor from FL3 is vented to the atmosphere through SH7.
FCC-- FCC deaerator. Provide 38,000 Ib/hr of BFW to SSG. It also takes I O #
steam from SH5.
DAl-- Main boiler house deaerator. Provide about 420,000 Ib/hr of BFW for boilers
and wash water.
PU2--

There are three (3) deaerators in the steam system: the main boiler deaerator,
reformer deaerator, and FCC deaerator. The APLUS simulation uses a flash drum to
represent the reformer deaerator as it is actually functioned as a drum. The boiler feed
water make-up enters the steam system at 110F after picking-up heat from the cooling
water system and boiler blowdown heat exchanger. A detailed stream information
showing the APLUS results is attached at the end of this section.

3-9

4
DIB COLUMN HEAT PUMP

Introduction
TENSAs experience in process optimization has shown that Deisobutanizer (DIB) is a
good candidate for Industrial Heat Pump (IHP) application. In this study, the DIB
column of the alkylation unit in URC is to be examined and evaluated for its IHP
potential.
Base Case
To establish a basis for comparison, the plant operating data of the DIB column is first
simulated. The attached Figure 4-1 is a simplified process flow diagram of the DIB
column. The alkylation reactor effluent, which contains alkylate and unreacted IC4and
NC4,is fed to the DIB column to obtain 70-75% purity of IC4,as overhead product of the
column to be recycled back to the reactor. There are two other feed streams to the
column, one from the alky splitter bottom and the other from IC4 storage. The NC4
component is drawn as a side vapor product from the column. Alkylate is drawn as the
bottom product.
There are two reboilers for this column. The upper reboiler provides 14.3 MMBtu/hr of
heating duty. The rest of 4.6 MMBtu/hr is provided by the lower reboiler. Table 4-1
summarizes the comparison between plant data and simulation results. The simulation
results have been consistent with the plant measured data on all crucial parameters.
The only discrepancy is in the columns bottom temperature, which is believed to be
caused by the difference in the heavy component composition of the column feed
stream. The simulation also shows that a reflux ratio of 0.215 is required compared to
the plant data of 0.17.

4-1

D1B Column Heat Pump

Table 4-1
Base Simulation of DIB Column

Reflux ratio
Upper reb. duty (MMBtu/hr)
Lower reb. duty
AP (PSI)
Top tray temp. ( O F )
NC4draw (OF)
Umer reb. return (OF)
Bottom ( O F )
Total Overhead(Ib/hr)

Plant Test Data

0.17
14.3
4.6
8
153
176
225
317
77946

Simulation
0.215
14.3
4.6
8
152.4
180
228
347
77562

2.6
0.1
70.8
24.7
1.9
0
0
6738

2.4
0.1
70.8
25.4
1.3
0
0
6757

0
0
6.8
89.9
3.2

0
0
4.5
91.2
4.0

0
34102

0.2
34159

0.1

0.1

8.3
0.3
83.I

8.8
0.3
84.2

Vol.%
CQHs
C-HF:
IC4
NC4

IC5
NC5
c6+

NCASide Draw(lb/hr)
VOl. %
C3H8
C3H6
IC1
NCA
ICs
c6+

Bottom (I b/hr)
VOl. %
C3H8
CQHF:

IC5
NC5
c6+

4-2

SGRU
DebutanherOVHD

--

Aiky Deprop. OVHD

or
Poly Deprop. OVHD

E
GI

-3IC4 Storage

From
eIC4
Storage

I
NC4
Alkylate

Figure 4-1
Existing Configuration of IC4 Separation

Dl6 Column Heat Pump

Heat Pump and Pressure Reduction Operation

The DIB column heat pump schematic is shown in Figure 4-2. The heat pump takes
the overhead vapor of the DIB column, compresses it to a higher pressure to raise its
condensing temperature. The pressure is raised high enough so that when the
overhead vapor is condensed, it will provide heat to the columns upper reboiler.
One advantage of the heat pump is that it allows the column to operate at a lower
pressure than its normal operating pressure. The two factors that may set any column
pressure are: a) the pressure required for the vapor (in case partial condenser), and b)
the temperature required for condensing the overhead vapor with cooling water or air.
The heat pump allows the column to be operated at lower pressure because the
overhead condensing is no longer restricted by the temperature differential between
the overhead vapor and cooling medium. Instead, the majority of the overhead
condensing heat will be transferred to provide reboiler duty. Due to improved
separation at lower pressure, the energy is saved by operating the column at a lower
pressure. Another advantage of column pressure reduction is that it can reduce the
column flooding by about 20%. This is much needed increase in capacity which URC
will like to have since the column is nearly flooded at present. The energy saved at low
pressure operation is obtained by comparing the difference in reboiler duty between the
base simulation and the simulation at reduced top column pressure of 60 psia. The
product specification has been kept the same in both cases. Table 4-2 summarizes the
energy savings due to pressure reduction operation along with the heat pump
placement. The heat pump is sized to take 60% of the overhead vapor through the
compressor. The remaining 40% of the overhead vapor still goes through the
condenser to be condensed. Energy savings and compression power requirement are
also summarized in Table 4-2. Heat pump operation not only saves hot utility but also
cold utility at the top condenser. A typical cost of $0.3/MMBtu has been used to
evaluate the cold utility savings. Table 4-2 summarizes the heat pump economics.
Without giving credit to the capacity increase, the heat pump system can payback in
less than 2 years.

4-4

-1

Compressor

I
I

DIB
Upper Reb.

Lower Reb.

Figure 4-2
DIB Heat Pump Schematics

DlB Column Heat Pump

Table 4-2
Heat Pump Economics For The Current Configuration
Cost of Steam ($/IO00 Ib)

Pressure Reduction Savings (MMBtuIhr)

3.7

Heat Pump Hot Utility Savings (MMBtuIhr)

7.27

Total Hot Utility Savings (MMBtuIhr)

10.97

Total Cold Utility Savings (MMBtu/hr)

7.42

Compressor Work (Hp)

420
471,710.00

Total Hot Utility Savings ($/yr)

19,143.60

Total Cold Utility Savings ($/yr)

Compression Cost ($/yr)

107,637.60

Net Savings ($/yr)

383,216.00

Investment ($)

Installed Compressor Cost

(including 50% of the equipment cost for installation)
Installed Motor Cost
(including 50% of the equipment cost for installation))
Installed Reboiler Cost
(including 50% of the equipment cost for installation))
Piping & Instrumentation
Engineering

395,100.00
54,000.00
54,000.00
40,000.00
60,000.00

Total

603,100.00
-

Payback (yrs)

1.6

4-6

 DIB Column Heat Pump

New Column Configuration

The current DIB column produces about 70% purity of IC4 at the top. There is a large
amount of NC4being carried at the top (about 25%) and is feeding back to the reactor.
The presence of NC4 in the reactor not only reduces the effective reactor volume but
also inhibits the rate.
In order to reduce the amount of NC4 being circulated in the reactor, a new way of
removing NC4 from DIB column is considered as shown in Figure 4-3. The new
configuration differs the old configurations (Fig. 4-1 ), by taking the Depropanizer
bottom to the DIB column and by taking the reactor effluent to an idle existing column
(30 trays, 42 diameter). The new configuration allows the DIB column to obtain much
purer IC4from the top, and most NC4comes out at the bottom. The effect of this better
separation of IC4and NC4are many folds:
1. It reduces the amount of NC4 being fed to the reactor (25.6 Ib molelhr, compared to
340 Ib mole/hr).
2. It unloads the DIB column for more thruput (88% flooding compared to existing

100%).
3. It increases the alky reactor capacity and reaction rate.
4. The DIB column will have a smaller temperature difference between top and bottom.
Therefore, it will be even more attractive for heat pump placement.
Table 4-3 summarizes the simulation results of the DIB and the new columns. Also
shown are the corresponding performance in the current DIB configuration.
New Configuration Pressure Reduction and Heat Pump Evaluation

The principle of heat pump and the benefits of pressure reduction in the new
configuration, are the same as described in Section 4.3. Table 4-4 summarizes the
economics of heat pump on the new DIB column. The heat pump takes 70% of the
overhead vapor to provide enough heat for the reboiler duty.

4-7

Alky Deprop. O M I D
or
Poly Deprop. OVKD

-+

SGRU
Debutanizer

OVHD

I
I
I
I
I
I

c3

I
I

IC4 Storage

I
I

I
I

I
I

---

Washwfl3
~lkywater

-a.

NC4

r-

Alkylate

Figure 4-3
New Configuration of IC4 Separation

' DIB Column Heat Pump

Table 4-3 New Configuration Simulation

Top Temp. (F")

Bottom Temp. (F")
Flooding (%)
Reflux Ratio
Reboiler Duty (MMBtu/hr)
Top Product (Iblhr)
IC4
NC4
Bottom Product (Ib/hr)

IC4
NC4

New DIB
147
170
88
2.56
12.0

Current DIB

22982
1491

54386
19732

212
6095

303*
6163*

9.8

--

100
18.9

(*Side draw product)

New Column
Reboiler Duty (MMBtu/hr)
Top Product (Ib/hr)
IC4
NC4
Bottom Product (Ib/hr)
ICr
NC4
AIkylate

23267
1041
63
31
30198

21
2266
31872

As expected, the compression power reduces to 360 HP because of smaller

temperature lift between column reboiler and condenser. At $5/MMBtu of steam cost,
the heat pump system can payback in less than 3 years. This savings estimate is
based on the new configuration already in operation. No cost for erecting the new
column and other necessary accessories are included. The comparison is solely from
the energy savings point of view. No credit has been given to the increased alky
reactor capacity and reaction rate. A more detailed study on the proposed
configuration should lead to a better assessment of these effects on the overall benefit
to the plant.
Summary

A preliminary evaluation of industrial heat pump operation on the DIB column has
indicated a very good economic payback, in both the current DIB , and a new DIB
configuration. All the comparisons are based on rigorous column simulation. URC is

4-9

DIB Column Heat Pump

currently limited by DIB capacity along with other restrictions, but increasing the DIB
capacity is certainly the direction of the immediate future.
The current configuration with heat pump installation can increase the capacity by 20%
without adding a new column and its accessories. With feed tray optimization and tray
hydraulic review, it is possible to increase the column capacity by 25%. This increase
will give URC a corresponding FCC rate of 25,000 BPD. This has been the goal of
many related upgrading projects.
Other benefits of a heat pump system including NOx,SOx,and COz reduction in flue
gas emission, which is directly proportional to the amount of fuel fired in the boiler.
Heat pump system also reduces the load on the flare system by automatically shutting
down the heat supplied to the reboiler in case of an emergency. This could be a
tremendous savings in upgrading the flare system. All these benefits have not been
credited in the economic analysis. In view of all these benefits, URC should seriously
consider implementing the heat pump system. A detailed engineering design package
should be developed to firm up all the details of the heat pump system.
Table 4-4
Heat Pump Economics For The New Configuration

New DIB column reboiler duty (MMBtuIhr)

New column reboiler duty (MMBtu/hr)
Current DIB column reboiler duty (MMBtulhr)
Heat pump savings on new DIB column (MMBtu/hr)
Net hot utility savings on new configuration
heat pump system (MMBtu/hr) 18.9 - (12.0 + 9.8 - 9.0)
Total cold utility savings (MMBtu/hr)
Compressor work (Hp)
Total hot utility savings ($/yr)
Total cold utility savings ($/yr)
Compression cost ($/yr)
Net savings ($/yr)
Investment ($)
Compressor (including 50% of the equipment cost for installation)
Motor (including 50% of the equipment cost for installation)
Reboiler (including 50% of the equipment cost for installation)
Piping & Instrumentation
Enaineerina
Total
Payback (yrs)

4-10

12.0
9.8
18.9
9.0
6.1
5.6
360
262,300.00
14,448.00
92,260.80
184,487.20
330,000.00
45,000.00
54,000.00
40,000.00
70.000.00
539,000.00
2.9

5
FCC Main Fractionation Unit

Process Description
Figure 5-1 illustrates the process flow diagram of the FCC main fractionation unit. The
fresh gas oil feed from a storage tank is heated to 355OF, by successive heat exchange
with overhead vapor, naphtha circulation, LCO product and slurry product in the
fractionation section. The feed is then further heated to 514F by a fired heater (350H102),before flowing to the reactor.
The feed to the main fractionator is the FCC reactor products which enters at the
bottom of the column. A light cycle oil (LCO) pumparound stream is taken as a side
stream to provide heat to the poly depropanizer and debutanizer reboilers. Another
heavy cycle oil (HCO) pumparound stream provides heat to the poly charge stream,
gas concentration debutanizer and gas concentration stripper reboilers. LCO product
is taken from a side stripper (TI 05) and used to preheat the feed before being sent to
storage. A naphtha pumparound stream is used for feed preheating. The bottom slurry
is used to generate steam in the steam generators (E-I02 A-C). A split stream of the
slurry is used to further preheat the column feed to 355F.
Figure 5-2 shows the existing heat exchanger network of the FCC main fractionator unit
on a grid diagram. Each stream with arrows pointing to the right is a hot stream which
needs to be cooled. Streams with arrows pointing to the left are cold streams which
needs to be heated to the desired temperature. The dumbbells represent heat
exchangers between hot and cold streams.
The overhead stream is simulated to obtain the condensing duty of heat exchangers
E117, E107, and E l l l .
The measurements on each heat exchanger inlet and outlet temperature are used to
obtain the log-mean temperature of each exchanger. The overall heat transfer
coefficient for each exchanger was calculated based on the known surface area, and
log-mean temperature difference. This information is also used in the targeting phase
of pinch analysis.

5-1

0
Y

.r(

E117

-n+

Frac. OVHD

LY1

Slurry

712

E107

Ell1

)E@z@!+
12.7

19.8

13.3

476

I
103

490

LCO Reflux

284
3703107

353

T
.$'..$
.>
.x.:.>:.,
$.?x+
....
.:.:.:.:.:2.::...:.:.
.....

Naph. Cir.
HCO Reflux

24

El05

353

Poly DeC4
Poly Charge
Gas DeC4
Gas Stripper

Figure 5-2
Existing Heat Exchanger Network

I*'

I
I

E103A

@-@+
E112

;5

........
,
:
.A
F
:
.:
..:.:.:<.:*.

PolyDeC3

13B
LCO

I
252
3513104

6.43
358

149

FCC Main Fractionation Unit

Analysis of the Current Base Case Operation

An important result of the pinch technology principles is the ability to set performance
targets prior to the actual design. It is possible from the process data alone, to
confidently predict the minimum thermal energy and surface area requirement of any
process.
Starting from the individual streams, it is possible to construct one composite curve of
all the hot streams in the process, and another for all the cold streams by simple
addition of the heat contents over the temperature intervals in the problem. Figure 5-3
shows these curves for the FCC base case. The overlap between the two composite
curves represents the maximum amount of heat recovery possible within the process.
The over-shoot of the cold composite represents the minimum amount of external
heating (hot utility) requirement. The over-shoot of the hot composite represents the
minimum amount of external cooling (cold utility) requirement. The closest point
between the two composite curves is known as the pinch point. The temperature
approach between the composite curves at the pinch is defined as DTmin)). As DTmi
increases, utility requirements increases, however, the heat exchange surface needed
to recover the energy decreases as a result of the larger approach temperature. For
any DTmi, the composite curves define the minimum utility requirements for the process
and thus establish the energy targets for the process.
The Grand Composite Curve (GCC) is a plot of the interval temperature versus heat
flows through the intervals. The GCC is derived from the composite curves. The net
heat flow is zero at the pinch. From the GCC, the levels of hot and cold utilities
required for the process can be identified (for details refer to Appendix A).
The GCC for the base case of FCC Main Fractionation is shown in Figure 5-4.This is
a typical threshold problem, no hot utility is required from outside the process. The
process itself produces enough heat to supply its own needs. From practical
application, this means that the FCC fired heater could be eliminated. However, to
meet the steam demand in the plant, the heat source available from the slurry stream
is used to generate steam. If the modification recommended in the steam system is
followed through, the steam demand in URC will be reduced drastically. This will allow
URC to use the slurry heat to preheat the reactor feed and eliminate the fired heater
totally. Even with current mode of operation, where only part of the slurry heat is used
to preheat the reactor feed, there is some energy savings potential. This can be
achieved by increasing heat exchanger surface area. The following sub-section
describes in detail the maximum savings potential.
Optimum Energy Recovery

To evaluate the optimum energy recovery potential, a pinch analysis is again

performed on those streams with potential for increased heat integration.

5-4

t
(v

a
tn

tu

u
a

u
u
Er
k

0
y.l

tn

!k
a

4J

Ot9

OOT-

FCC Main Fracfionation Unit

The streams that have been used for heat recovery like LCO reflux and HCO reflux are
removed from the analysis. Also removed are those heat sinks for LCO and HCO
reflux like Poly DeC4and Gas DeC4,etc. The heat exchanger information is
summarized in Table 5-2. Figures 5-5and 5-6show the composites curve and GCC
respectively. As the DTmi, varies, the amount of direct heat integration and the
corresponding exchanger surface areas also changes. Figure 5-7 shows the hot utility
vs. surface area plot.
Table 5-2
Heat Exchanger Information Summary

U
(MMBtu/hr) (Btu/hr- ff-OF)
12.2
12.7

Heat Exchanger

ATin
(OF)

E-I 17

57.9

Area
(ft*)
18,020

E-1 16

20.9

8,828

4.3

23.3

E-103B

134.1

1,206

1.8

11.1

E-104

261

4,692

9.2

7.51

E-107

NA

E-I05 + E103A

NA

NA

7.3

E-1 12

NA

NA

6.43

19.8

Point A on Figure 5-7 represents current level of integration. From Figure 5-7 it is
evident that the optimum heat recovery is at Point B. Beyond Point B, further
increase in the surface area will not be economical. Point B corresponds to a DTmin of
20F, with pinch interval temperature of 281OF. The pinch design rules are then
followed to obtain the optimum design change. Figure 5-8 shows the revised network
design. An increase in surface area on E-117 will allow E-1 17 to transfer more heat to
the gas oil feed. The net effect is to save 1.6 MMBtu/hr of hot utility at the FCC fired
heater and the same amount of cold utility from the cooling water. However, the big
advantage in increasing the surface area through one additional shell on E-117 is to
reduce the pressure drop of this exchanger. A 4500 HP wet gas compressor is
currently used down stream of overhead gas stream. Increasing the suction pressure
by about 2 psig can reduce the compression power by about 300 HP. The estimated
savings for utility and compressor power reduction is $14I1212/yr. An investment of
$1803 60 for installing the additional 4504 ft2 shell on E-117 is required. This system
will have a payback in 1.3 years.

5-6

>

I-

cn
a.
0

0
0

a
a

CT

El

oz9

00;-

COMPOSITE CURVES

u.
Q
W

cl
W

E8
3 N

tU

U
W

a
x
W

I-

Figure 5-5
Hot and Cold Composite Curves for FCC Modified Case

'

UJ

a
a
I-

z
H
>
>

m
a,

I-

-I
H

t-

a,

I-

m
c
m

4X

Frac. OVHD

Slurry

29 1

712

1
478

/a7

E103B

LCO

439

r\
J

353
Naph. Cir.

E105

291

@-@+
E112

E116

E103A

291

Gas Oil F e e d 4

21.7

h4

Figure 5-8 Revised Heat Exc ianger Network

149

FCC Main Fractionation Unit

Discussion
1. The calculated overall heat transfer coefficients in Table 5-2 are low compared to
most design values. For instance, the overall heat transfer coefficient calculated
between slurry and feed is 7.51 , compared to a similar design value of 49. The U
values for E-l03B and E-I 16 are expected to be about 30, but were found to be
much less. Apparently, these heat exchangers are badly fouled. Proper cleaning
and maintenance schedules should be followed to ensure satisfactory performance
on these heat exchangers.

2. There is still plenty of medium level heat available at E-107 and E-105, which are
dissipated through cooling water (E-I 07) and air (E-I 05). No low temperature
process sink can be used within FCC to recover this heat. However, the BFW
make-up can be used to pick up some heat from the overhead gas in E-107. Other
low temperature sinks such as the sat gas feed stream in the alky unit and olefin
feed to the alky depropanizer can also be used to pick up more heat from these two
heat sources.
3. The FCC fired heater can be totally eliminated if the recommendations in the steam
system are carried out and the steam demand reduced. In that case, the slurry heat
will not be used to generate steam, instead it can be entirely used to preheat the
feed to the FCC reactor.

5-12

A
INTRODUCTION TO PINCH TECHNOLOGY

This Appendix provides an introduction to the basic concepts and terminology

associated with "pinch technology". It also demonstrates the usefulness of pinchbased methods for industrial heat pump and heat engine placement. This is intended
for readers unfamiliar with these technologies, to provide the necessary background for
a general understanding of the main sections of this report.
A bibliography is attached to this appendix. It highlights recent articles that present a
more detailed discussion of pinch technology and its application to heat pump
placement and related subjects, such as:
e

Overall Energy Efficiency (I),

The Design of Heat Exchanger Networks (HEN's) (2),

Integration of Heat and Power Systems with Chemical Processes (3),

Heat integration of Distillation Systems (4), and

"Appropriate Placement'' of Heat Pumps in Chemical Processes (5,6,7,8).

Process Heating and Cooling

Within most processes in the chemical and allied industries, there are streams that
require heating and streams that requires cooling. Any stream that requires heating is
conventionally said to be "cold" (irrespective to its temperature level), and streams that
require cooling are said to be "hot".
The heating and cooling duties within a process can be provided by "utilities" such as
steam (for heating) and cooling water (for cooling) that are available on the site.
However, the load on these external utilities can often be reduced by "heat integration"
of the process - that is, by transferring heat from hot process streams to cold process
streams by means of heat exchanger networks (HEN's). This is illustrated in Figure A1. As the only heating and cooling duties that incur direct operating costs are those
associated with utilities, heat integration generally leads to a reduction in process
operating costs.

A- 1

QH

Process

b-1

Streams 1111111)

In

(from hot utility)

Process

Streams

out

QC (to cold utility)

Figure A-I : Heat

Transfer

Using A

Between

Heat

A-2

Process

Streams

Exchange Network

The Heat Transfer Pinch

In most processes, no matter how thoroughly they are heat integrated, there will always
be a residual heating duty (QH) and a residual cooling duty (Qc) that have to be met
by utility heating and cooling. The size of these residuals can be reduced by
increasing the heat transfer area within the process heat exchangers. This essentially
allows smaller temperature differences between the matched hot and cold streams. In
general, however, the residual utility loads would be finite even if the heat transfer area
were infinite.
If both utility heating and cooling are required, the process may be considered to be
made up of two parts:
e
e

A higher temperature part which, after complete heat integration, acts as a net
heat "sink" or acceptor.
A lower temperature region which, after complete heat integration, has surplus
heat to be rejected. It is thus a net heat "source".

The temperature that separates the source and sink sections of the process is called
the heat transfer "pinch" (Figure A-2). In a properly integrated process, there is no heat
transfer from above the pinch to below the pinch. Also, the "temperature driving force"
(Le., the difference in temperature between the hot and cold streams) reaches its
minimum value, designated DTmin, in the region of the pinch.
When a pinch design is implemented, the hot and cold utility requirements reach their
minimum values, QHmin, QCmin, appropriate to the selected value of Dtmin. A few
processes do not have heat transfer pinches. These require either only heating or only
cooling from external utilities, but not both. Such processes are said to exhibit
"threshold" characteristics.
Heat Pumps

Heat pumps provide a means of upgrading heat (Le. raising its temperature) by the
input of work. They may therefore be regarded as heat engines running in reverse.
The principle behind the heat pump is illustrated in Figure A-3, where an ideal heat
pump extracts an amount of heat Q from a temperature TI and elevates it to the
temperature T2 by the input of reversible work WREV. The best known real heat
pumps are reverse Rankine cycles. Low pressure vapor generated at some source
temperature, Tsl is compressed to a higher pressure at which it condenses, releasing
its heat at a higher target temperature Tt.
The work input for such a system is generally provided by mechanical compression
with the system operating in a closed cycle (Le. the working fluid repeatedly passes
through evaporation, compression and condensation stages). This is depicted in
Figure A-4.

A-3

QH

Heat
Sink
2

Pinch
Temperature

Zero Heat
Flow

Heat
Source

QC
Figure A-2:

The Heat Transfer Pinch

TQ
Figure A-3:

Basic Principles of the Heat Pump

A-4

Cold Process
Fluid

Condenser

Flash

Compressor

Evaporator
Hot Process
Fluid

Figure A 4 :

Closed Cycle Heat Pump

A- 5

Sometimes, it is possible to use a process vapor stream as the working fluid in heat
pumps. These heat pumps are called semi-open cycle heat pumps. The most common
is called the mechanical vapor recompression (MVR) heat
semi-open cycle (type I),
pump. The hot process vapors are compressed in a compressor and then condensed
in the heat pump condenser to satisfy a process heating requirement at an elevated
temperature (see Figure A-5). A less common type of semi-open cycle heat pump
(Type 2) has the opposite configuration, i.e., an evaporator instead of a condenser. A
liquid stream is vaporized in the evaporator and then compressed to a higher
temperature in a compressor (see Figure A-6). This type of heat pump cycle is
recommended when a low temperature heat source is available to evaporate a liquid
process stream which is required in the vapor phase at a higher temperature. It is
important to note that semi-open cycles are only feasible when the process fluid
undergoes a phase change; condensation for Type 1 systems, and evaporation for
Type 2 systems. These are the main types of heat pumps considered in this report. In
addition to these, there are a number of other types of heat pumps either commercially
available or under development. These include chemical heat pumps (which use
exothermic and endothermic reactions as a means of upgrading heat), absorption heat
pumps (which use low grade heat to drive an evaporationkondensation cycle to
elevate the available "waste heat" to a useful level) and electromagnetic heat pumps.
Current state-of-the-art heat pumps tend to be limited in the operating temperature
range for available working fluids. Moreover, economic considerations generally limit
the practical temperature lift in heat pumps to around 60F.

Appropriate Placement of Heat Pumps

The pinch concept leads to useful insights into the appropriate use of heat pumps in
industrial processes. Because the below pinch region is a net heat source any heat
pump must accept heat in this region if it is to reduce the external cooling requirements
of the process. By a similar argument the heat pump must reject its heat to the net heat
sink above the pinch to reduce the demands on external utility heating. A heat pump
which satisfies these criteria is said to be "appropriately placed" (Figure A-7).
If a heat pump acts wholly above the pinch, it will reduce the hot utility requirement QH
by an amount equal to the work input W of the heat pump (see Figure A-8). However,
as the unit cost of providing work is normally greater than the unit cost of heating, such
an arrangement is generally uneconomical. A heat pump acting entirely below the
pinch (Figure A-9) has the net effect of degrading the work input W into waste heat that
has to be rejected to the cold utility Le. the net heat rejected rises from QC to QC + W,
which is clearly undesirable.
Both Figures A-8 and A-9 represent "inappropriate placement" options for industrial
heat pumps.

A-6

QD
Heat
Sink

W
A
4

Tpinch

zero

Heat
Source

QA

Figure A-7:

Appropriate Heat Pump Integration

A-8

QD
W

Tpinch

v
Source

Figure A-8:

Inappropriate Heat Pump Integration-above pinch

I
QD

Q*T
I!

Sink
Heat

I
Tpinch

Heat
Source
W

Figure A-9:

Inappropriate Heat Pump Integration-below pinch

A-9

The Grand Composite Curve (GCC)

As already noted, most industrial processes can be divided at a "pinch temperature"

into net heat source and net heat sink regions with no heat flow at the pinch itself.
However, it is possible to represent the net heat flow at every temperature level within
the process by means of a "Grand Composite Curve" (GCC) or temperature enthalpy
plot. An example of such a plot is given in Figure A-10.
The ordinate of the GCC is the so-called "interval temperature." This is a convention to
put the hot and cold streams on a common temperature basis, after allowing for the
necessary minimum temperature driving force (DTmin) between them. Consider the
simplest case, where the heat transfer resistance associated with all the hot streams is
equal to that associated with all the cold streams. The interval temperature of a hot
stream at an actual temperature of TH is defined to be TH - (DTmin/2); and that of a
cold stream at an actual temperature of TC is TC + (DTminR). Where the heat transfer
resistance of the streams are different, it is necessary to ascribe an appropriate "DTmin
contribution, "DTcont, between 0 and DTmin, to each stream. Heat transfer between
such streams is permitted only if (TH - Tc) > DTmin Le. if the interval temperature of
the hot stream is greater than or equal to that of the cold stream. The abscissa on
Figure A-I 0 represents the net heat flow through the process after allowing for all
permitted heat integration of process streams. This takes the value of zero at the
pinch, as described in the earlier discussion, and has the values of QH (Le. net hot
utility requirement) at the highest interval temperature in the process and QC (Le. net
cold utility requirement) at the lowest interval temperature.
The GCC is important in evaluating heat pumping opportunities because it allows a
rapid assessment of the temperature levels available, and the amount of heat that can
be heat pumped in a process. Thus, in Figure A-I 1 an amount of heat QA can be
accepted by a heat pump from the process at an interval temperature TA below the
pinch. An amount of heat QD = QA + W can then be delivered to the process above
the pinch at interval temperature TD.
Hot and Cold Composite Curves and Area Targeting

The effect of heat integration on the process utility consumption and temperature
driving forces is shown in the form of hot and cold composite curves in Figure A-12.
The "hot composite curve" represents the summation of the heat loads associated with
all streams that have to be cooled in the process ("hot" streams) and similarly, the "cold
composite curve" represents the summation of all heating loads (Le. "cold" streams).
DTmin, the minimum temperature difference between the hot and cold composite
curves, appears as the vertical distance between the hot and cold composite curves at
their point of closest approach. DTmin is a measure of the level of heat integration.

A-10

Hot Utility Load

Cold Utility Load

Ti

pinch

Figure A-10:

The Grand Composite Curve

A-1 1

Interval Temperature
Enthalpy

QD

Figure A-11:

The Grand Composite Curve and Heat Pump Integration

A-1 2

Heat Duty

Heat Duty

Increasing
Heat
Integration

Heat Duty

Figure A-12:

Effect of Heat Integration on Utility Targets and

Temperature Driving Forces

A-1 3

Varying the extent of heat integration is represented by moving the hot and cold
composite curves horizontally relative to one another. Doing so will change the vertical
distance between them at their point of closest approach (Le. the pinch), and thus
corresponds to changing values of DTmin. The horizontal displacements between the
composite curves at their high and low temperature ends are the corresponding values
of QH and QC (the external heating and cooling requirements, respectively) for the
process. A decrease in DTmin implies an increase in the level of heat integration. In
general, as DTmin decreases the minimum hot and cold utility requirements also
decrease. Progressively smaller values of DTmin are represented in Figure A-1 2(a)
through A-I2(c). However, a decrease in DTmin also implies a decrease in the overall
driving force for heat transfer and a resulting increase in heat transfer surface area
requirements.
Figure A-I3 shows a plot of plant heat transfer area and the corresponding minimum
hot utility consumption. Curves of this type can be generated for any given process
using area targeting algorithms based on pinch technology principles (see below).
The curve shown in Figure A-I 3 separates the thermodynamically feasible region from
the infeasible region. The region above the curve and to the right represents a process
in which the available heat transfer area is greater than or equal to the minimum
needed to achieve a specified hot utility usage level. To the left and below the curve
the implied heat transfer area is less than the minimum requirement, implying that no
practical process can correspond to any point in the feasible region, e.g. point A. The
hot utility consumption and the heat transfer area requirements are directly related to
the plant operating costs and capital costs, respectively. Therefore, the inverse of the
slope of the straight line joining two points on the curve is a measure of the payback
period for going from one level of heat integration to another. This is also illustrated on
Figure A-1 3.
The subject of heat exchanger network (HEN) area requires further elaboration.
Townsend and Linnhoff (9)provides a useful algorithm for estimating required HEN
areas without having to design the HEN in detail. The principles behind the algorithm
are illustrated in Figure A-I4 in terms of heat transfer from the hot composite curve to
the cold composite curve and between the process and the hot and cold utilities. For a
given value of DTmin, there is a certain extent of horizontal "overlap" of the two
composite curves. This represents the amount of heat that can be transferred from hot
streams to cold streams within the process. Outside of the overlap region, utility
heating or cooling is required. All heat transfer is represented as vertical lines on
Figure A-14. This is an idealized representation implying that all matched hot and cold
temperatures within the process HEN must be the same as the matched temperatures
on Figure A-14. However, with this simplifying assumption, it is possible, using
appropriate stream heat transfer film coefficients, to predict the minimum area for a
HEN with surprising accuracy. For further details, reference (2)should be consulted.

A-1 4

Existing

Plant

Feasible

Q~~

Region

Decreasing AT

min

\ c

QHC

sB

sC

Heat Transfer Surface Area

(scIncremental Payback Period

cA

= Cost per Unit Area

CH

= Cost per Unit of Hot Utility

= Hours of Operation per Year

Figure A-1 3:

SB) c A

[years] =
(QHC-

Hot Utility Consumption Versus

Heat Exchanger Area

A-1 5

QHB) c H

250

200

150

II

!
I

I
Hot

Composite

Curve

!
I

50

C. Utility

Cooling Q~

1-

100

1
0

20

40

Cumulative Heat

Figure A-14:

60

80

Load (MMBtu/hr)

Area Targets for the Example Problem

A-1 6

100

Appropriate/lnappropriate Integration of Heat Engines

Consider the hypothetical process shown to the right of Figure A-I 5. The process
pinch and minimum utility requirements (QHmin, QCmin) are shown on the figure. To
the left of Figure A-I5 is a representation of a Carnot engine, for which both the heat
acceptance and rejection temperatures are hotter than the process pinch temperature.
This Carnot engine is assumed to have an efficiency (ratio of power produced to heat
absorbed) of 33.3%. This means that for each 3W units of heat absorbed, 2W units of
heat are rejected at a lower temperature and W units are converted into work. The
heat and work flows associated with the Carnot engine are shown in Figure A-I 5.
The total hot utility requirement for the Carnot engine and process is (QHmin + 3W).
This requirement can be reduced by integrating the heat engine with the process above
the pinch, see Figure A-16. In the "above pinch" integrated arrangement, the engine
exhaust heat displaces the process hot utility usage and the total utility requirement
falls to (QHmin + W), a saving of 2W units of heat. Work W has effectively been
produced at a marginal efficiency of 100% (neglecting mechanical and electrical
losses) since the waste heat from the machine is usefully used to displace process
requirements.
Figure A-I 7a illustrates the same process together with a Carnot engine for which both
the heat acceptance and rejection temperatures are colder than the process pinch
temperature. As in Figure A-I 5, a machine efficiency of 33.3% is assumed. The total
hot utility requirement for the Carnot engine and process shown in Figure A-17a is
(QHmin + 3W). This requirement reduces to QHmin when the heat engine is integrated
with the process below the pinch, see Figure A-I 7b, since the engine heat
requirements are supplied by waste process heat.

In Figure A-18, a Carnot engine is shown integrated with the process such that heat is
absorbed from above the pinch and rejected below the pinch. In this heat integrated
arrangement, the minimum utility requirement is not reduced from the total nonintegrated requirements. In other words the heat engine violates the process pinch by
transferring heat across it. As a result, the total hot utility requirement of the integrated
system (QHmin + 3W) is the sum of the two separate system requirements.
Figure A-I 6 and Figure A-I 7 illustrate the concept of "appropriate" integration of heat
engines with a process. Appropriate integration involves operating a heat engine such
that the engine heat acceptance and heat rejection are entirely above or entirely below
the process pinch but not across the pinch. As the figures illustrate, this appropriate
integration leads to substantial reductions in utility requirements over the
inappropriately integrated case, Figure A-I 8. Inappropriate integration means the heat
engine transfers heat across the process pinch.

A-I 7

Q H = 3W

Q H min
Heat
Engine
eff = 0.33

F W

2w

Pinch
Temperature

QC min

Figure A-15:

Stand Alone Heat Engine Operation

and Process Demand

A-1 8

Heat
Engine
eff = 0.33

b W

2w

Pinch
Temperature

--

Q = 0.0

Heat
Source

1
QC min

Figure A - I 6:

Appropriate Heat Engine Integration

A-1 9

Q~

min

Q H min

Sink

Sink

Pinch
Temperature

3w

I
,
Source

Engine

eff = 0.33

L
Source

Q C min

a)

2w

[ Q c min -

Before Integration

Figure A-1 7:

b)

3w

eff = 0.33

2w

After Integration

Appropriate Heat Engine Integration Below the Pinch

A-20

[ Q H min

3w

+ 3wl

Heat
Sink

I
0

Heat
Engine
eff = 0.33

= 0.0

2w

Figure A-1 8:

Inappropriate Heat Engine Integration

A-21

Placement of Distillation Columns and Evaporators

The rules for placing distillation columns and evaporators are essentially the same as
those for heat engine placement. The entire distillation or evaporation system should
be either above the pinch, or below the pinch to ensure the maximum scope for
beneficial heat integration of the condenser and reboiler heat loads. Placing the
system such that the reboiler is above the pinch and the condenser is below the pinch
(see Figure A-I 9) is "inappropriate" as it degrades heat across the pinch. The
procedure for correcting inappropriate "cross pinch" placement of distillation columns
(4) is based on the concept of "pressure shifting". Reducing the pressure of the column
lowers the evaporation and condensation temperatures, and so may allow the
temperature of an "above pinch" reboiler to be reduced to below the pinch.
Conversely, raising the pressure may allow a "below-pinch" condenser to be raised to a
temperature above the pinch. In either case, the result is that the "cross pinch"
placement is eliminated and both the reboiler and condenser are restored to the same
side of the pinch.
Correction of inappropriate evaporator placements uses precisely the same procedure.
Summary

This appendix gave a quick review of pinch technology and its relevance to heat pump
and heat engine placement in industrial processes. Methods to determine the
appropriate positions for heat pumps and heat engines in any given process have been
detailed.

A-22

Heat
Sink

1
Distillation
Column
or
Evaporator

Figure A-1 9:

Inappropriately Placed Distillation

Column or Evaporator

A-23

References for Appendix A

Boland and E. Hindmarsh. "Beyond HENS: A Total Thermodynamic Approach to
High Energy Efficiencies." Chem. Eng. Prog., July 1984.
Linnhoff and E. Hindmarsh. "The Pinch Design Method for Heat Exchanger
Networks." Chem. Eng. Sci., 1983, 38, 745-763.
Hindmarsh, D. Boland and D.W. Townsend. "Maximizing Energy Savings for Heat
Engines in Process Plant." Chem. Engineering, February 4,1985, p. 38.
Hindmarsh and D.W. Townsend. "Heat Integration of Distillation Systems into Total
Flowsheets - A Complete Approach." AlChE National Meeting, November 1984.
Townsend, J.W. Hill and D. Boland. "The Future of Heat Pumps in the Process
Industries." I. Chem E. (NW Branch) Symposium Number 3 on Heat Pumps, 1981.
Ranade, E. Hindmarsh and D. Boland. "Industrial Heat Pumps: Appropriate
Placement and Sizing Using the Grand Composite." 8th Industrial Energy
Technology Conference, Houston, Texas, June 1986.
Ranade, A. Nihalani, E. Hindmarsh and D. Boland. "Industrial Heat Pumps: A Novel
Approach to Their Placement, Sizing and Selection." Presented at the 21st
lntersociety Energy Conversion Engineering Conference, San Diego, Calif. , August
1986.
Chappell and S. J. Priebe. "Process Integration of Industrial Heat Pumps."
Presented at the 8th Annual Industrial Energy Technology Conference, Houston,
Texas, June 1986.
Townsend and B. Linnhoff. "Surface Area Targets for Heat Exchanger Networks."
11th Annual Research Meeting, The Institution of Chemical Engineers, April 1984.

A-24