Dec-2015

Heat exchanger network retrofit for energy savings

Identifying cost effective energy savings by optimising plant heat integration
in the crude preheat section

AKRAM KAMEL and MAHMOUD BAHY NOURELDIN
Saudi Aramco

Article Summary

Industrial energy use accounts for approximately one-third of the world’s energy demand. In particular, the 1970s oil crises saw how the efficient use of energy becomes a priority for policy makers in many industrialised countries. Rising concerns about climate change have sharpened the importance of energy efficiency. Energy-related emissions accounted for 9.9 gigatonnes of carbon dioxide in 2004, which represented an increase of about 65% from 1971 levels.¹ With the current best available technologies and given the huge amount of wasted heat, energy efficiency is practically regarded as the most cost effective tool to reduce CO₂ emissions and climate change.² At the company level, energy efficiency also reduces the key performance indicators of energy intensity and increases companies’ competitiveness.¹

According to the US Energy Information Administration, the industrial sector consumed 31.4 quads of energy in 2008, which was approximately one-third of all the energy used in the US that year.3 Further, more states are deploying regulations that will limit the emissions of greenhouse gases. Combined, these circumstances present a convincing opportunity for industrial manufacturers to implement energy saving projects and processes to lower their energy consumption and reduce their carbon footprint.

Crude distillation units (CDUs) are major consumers of energy in oil refineries because of the high energy consumed in the crude furnace.⁴ Retrofit of the crude preheat train was applied to find the optimal and most profitable heat exchanger network design that yields the highest net present value (NPV).

Energy conservation has become more important as public awareness and concerns regarding global warming and energy shortage continue to grow. Naturally, energy management programmes have been applied to refineries as they are the most energy intensive operations in the manufacturing industry. There are many ways to increase energy efficiency, and design and process heat integration are widely used methods. Heat transfer from hot products and pump around streams to the crude feed by the application of heat exchanger networks reduces the energy demands of both coolers and furnaces. This reduction in energy demand diminishes operating costs while increasing the capital cost of the exchanger area installation; therefore, a retrofit design is preferable to a grassroots design for oil refineries. In a real situation, the uncertain quality of crude oil in the market and changes in the quality of crude from traditional sources motivate the heat exchanger network retrofit to be operated in multiple periods for greater flexibility. The application of energy cost and capital cost trade-off from the retrofit technique of the pinch design method proposed by Tjoe and Linnhoff5, integrated with thermodynamic properties, provides the energy saving plots and optimum target.

In the present scenario, the retrofit of the heat exchanger network is an important way of improving energy efficiency in process industries. An industrial plant of 40 years plus lifetime may need to be retrofitted several times to improve energy efficiency and/or to meet increased production. There are several approaches to achieve energy savings in a retrofit study: for example, reducing utility use, modifying the network topology, upgrading heat transfer units, installing additional heat transfer area, repiping streams, and reassigning heat recovery matches. The retrofit’s objective is to identify a cost effective heat exchanger network, subject to design and operating constraints, that will not hinder any future retrofits. Implementing such retrofit strategies in practice may be difficult due to constraints related to topology, safety and maintenance, which often exist in a complex network.⁶

As well, the capital cost is usually high because of the considerable piping and civil works required for the retrofit and potential production losses during process modification. Since the late 1970s, continuous escalation in energy prices has outpaced plant equipment costs. This warrants continual modification of the facility’s heat exchanger network to enhance energy efficiency during its lifetime. In such cases, the retrofit’s objective/ task is to produce a practically implementable cost effective design modification that satisfies new process objectives and new operating constraints. There are many possible modifications to an existing network to retrofit the original design to the new objective. It can include a combination of all possible process operating and design condition modifications, existing topological/structural modifications, and existing unit design modifications and parametric modifications (such as heat transfer enhancement to improve the overall heat transfer coefficient) as well.

This article illustrates through a case study the ‘plant lifetime retrofitability’ concept in a typical oil refinery ADU/ VDU application.

Retrofitability and pinch technology

In crude distillation units, crude oil is preheated in two stages before entering the distillation column. The first stage is a heat exchanger network, where the oil is heated to an intermediate temperature by exchanging heat with hot process streams that require cooling and recovery of this heat from condensers. Subsequently, the crude oil enters a fired heater to reach its target processing temperature. The more fuel consumed in a furnace, the larger the operating cost. Any heat recovered from the hot distillates streams reduces the fuel consumed in the furnace. The energy efficiency of crude distillation units can be improved by optimising the heat exchanger network to maximise heat recovery and minimise fuel consumption in the crude furnace.

Retrofit targets are preferably achieved by reusing existing equipment more efficiently, repiping process streams, reassigning heat recovery matches, and installing additional heat transfer areas if required. While fulfilling these retrofit objectives, existing equipment constraints, such as hydraulic capacity, available heat transfer area, and heat exchanger ΔT must be checked.

Pinch technology/analysis was developed in the late 1970s as a technique for optimisation of thermal heat recovery, and rapidly gained wide acceptance as a theoretically elegant, yet practical, approach to the design of heat exchanger networks. Since then, it has evolved into a general methodology for optimisation, based on the principles of process integration. It has been applied successfully not only to energy systems (heat recovery, pressure drop recovery and power generation), but also to freshwater conservation, wastewater minimisation, production capacity debottlenecking, and management of chemical species in complex processes.

By applying pinch analysis to heat exchanger network synthesis and retrofit, engineers can calculate the energy requirement for any process, and produce thermally efficient and practical designs. Energy savings are typically 20% or more compared to previous best designs.

Pinch analysis also applies to optimisation of the supply-side, consisting of on-site utilities such as boilers, furnaces, steam and gas turbines, cogeneration, heat pumps, and refrigeration systems. Pinch technology can play an important role in analysis of the efficiency of an oil refinery. The technology has been developed over the last decade by Linnhoff and others^7-13, and has been applied to the analysis of heat recovery and process-utility interaction in a number of industries. The oil refining industry is one of the major users of energy and is also highly heat integrated. It is therefore a good candidate for the application of pinch technology.

The first step is to separate the streams into two groups: hot streams that need to be cooled; and cold streams that require heating. In the case of small temperature intervals in each group, all required enthalpy changes are added to produce a composite temperature- enthalpy curve. In this way, we can model a heat recovery problem in terms of a single composite hot stream and a single composite cold stream (see Figure 1). As we are only interested in enthalpy changes, rather than absolute enthalpies, we can move the two composite curves horizontally with respect to each other. The (vertical) temperature difference between the curves represents the ideal temperature driving force for heat exchange, and must be greater than a certain minimum value. The two composite streams are moved horizontally toward each other until this minimum temperature difference is reached at one point. This point is known as ‘the pinch’.

As Figure 1 shows, composite temperature enthalpy graphs can be used to set targets for process-to-process heat recovery, furnace duty, and cooling loads. Process-to-process heat recovery is possible wherever the hot composite stream is vertically above the cold composite stream. As this procedure involves only simple summations over the streams, it can be applied to a single process unit, a group of process units, or even an entire chemical plant or refinery. A number of commercial computer programs^{14, 15} are available to make this task relatively easy to perform. Although it is not possible to heat integrate some units because of operational, safety or piping cost constraints, there are usually enough opportunities for interunit integration to make the overall target calculations valid. A major advantage of using pinch technology for target setting is that a structured approach can be used to survey a large chemical plant or refinery. The approach ensures that attractive possibilities for inter-unit integration are identified and evaluated first. Once this is accomplished, attractive intra-unit heat recovery improvements are identified and evaluated further. The structured approach not only ensures that no attractive improvements are missed, but also avoids unnecessary effort in screening intra-unit heat integration possibilities when inter-unit integration is either a better alternative or an additional part of a revamp effort. These features are major advantages over previous approaches based on multiple case studies. Pinch analysis is used to identify energy cost and heat exchanger network capital cost targets for a process, and for recognising the pinch point. The procedure first predicts, ahead of design, the minimum requirements of external energy, network area, and the number of units for a given process at the pinch point. Next a heat exchanger network design that satisfies these targets is synthesised. Finally the network is optimised by comparing energy costs and the capital cost of the network so that the total annual cost is minimised. Thus, the prime objective of pinch analysis is to achieve financial savings by better process heat integration (maximising process-to-process heat recovery and reducing the external utility loads).

Case study

The primary focus of this study was to identify the most cost effective potential energy saving and enhancement initiatives through optimising refinery ADU/VDU plant heat integration, waste heat recovery and utilities consumption in the crude preheat section at minimum capital/operating cost requirements. This would enable a considerable reduction in the required furnace duties and corresponding reduction in fuel gas consumption. This study considered modifications to both unit design and operation to achieve the ultimate goal of improved energy efficiency for the ADU/VDU and an overall improvement of the refinery’s energy intensity index.

The main driver for this study has been the expected rise in the cost of energy over time and new initiatives aimed at improving the refinery margins and energy intensity index performance. At the time this study was conducted, the corporate price for fuel gas was $5.90/MMBtu. This price is projected to rise significantly within the coming years according to corporate planning quarterly reporting. In fact, this study considers a starting price of $7.40/MMBtu, which is the projected price by the beginning of year 2017 as the established price utilised for all new projects starting up that year. This is the earliest expected time frame in which potential design modifications would be commissioned in the unit.

According to company project cost forecasting practices, the net present value (NPV) will be based on a minimum 20-year operational period. In addition, an assumed discount rate of 6.5% is used for economic analysis as an estimate for a large national corporation oil refinery.

The study is a complete and definitive energy assessment along with a preliminary design of the new required heat exchangers and crude preheat configuration.

ADU/VDU overview

A plant’s crude distillation unit (atmospheric and vacuum distillation) currently processes 325 000 b/d of crude oil feed consisting of 65 LV% (211 000 b/d) Arabian Light Crude (AL) and 35 LV%(114 000 b/d) Arabian Heavy Crude (AH). The plant also has asphalt producing facilities. The sustainable asphalt production (paving and cutback) with 100% AL charged to the plant is 18 000 b/d, and 20 000 b/d when a mixed crude of AL and 100 000 b/d of AM or AL and 30 000 b/d of AH crude is processed. The products from the plant are: LPG, stabilised whole naphtha, kerosene, diesel (DGO), heavy diesel, and vacuum residuum. The atmospheric column receives the crude charge and separates it into overhead product, kerosene, DGO, and reduced crude. The naphtha stabiliser receives the atmospheric overhead stream and separates it into LPG and stabilised naphtha. The reduced crude is charged to the vacuum tower where it is further separated into heavy diesel, vacuum gas oils, and vacuum residuum. Two vacuum gas oil streams feed the hydrocracker plant. As much as 20 000 b/d of the vacuum residuum can be charged to the asphalt section with the balance of the residuum going to a visbreaker and later to fuel oil blending. A plant overview block diagram is shown in Figure 2. The plant overview process flow diagram is shown in Figure 3.

ADU/VDU hot and cold stream data

The cold streams to be heated and hot streams to be cooled for the ADU/VDU plant used in this study are shown in Table 1.

Study approach

ADU/VDU plant process variables (data collection)

The plant information system was used to obtain relevant plant historical process data such as flow rates, temperatures, and pressures over different time intervals, and yearly average figures were used in building the primary design simulation model. This step also involved comparing and consolidating actual plant data to the simulated data, until an appropriate level of confidence was obtained in the base case simulation of the plant. The model output material and heat balance were compared to real plant data to ensure the percentage error is accepted within engineering limits.

Lab data was used to assure the simulation model output product specification matches the real product specifications.

Simulation model

Once the base design model case was established and the output figures matched real data, there commenced conversion of the model to a more comprehensive rating model by using the heat exchanger actual rating data. This step involved extensive troubleshooting in order to ensure non-modified parameters at the same values as the base case to maintain the benchmark for accurate comparisons.

At this stage the model was ready for engineers to propose improvements and alterations to the plant, thereby developing the modification case.

Equipment sizing

Once the potential value and savings of the proposed improvements were quantified and validated, a preliminary equipment sizing exercise estimated required capital expenditure and operating expenditure as inputs to the economic analysis of the proposed improvements. As an outcome of the economic analysis, the NPV and internal rate of return were calculated to validate the proposed project.

Technical assessment

Base case description and model

The current configuration of the secondary crude preheat train can be described as follows (see Figure 4). The vacuum residue stream currently leaves the bottom of the C-200 vacuum distillation column at approximately 688°F (364°C) and exchanges heat with flashed crude stream in exchangers E-291 A/B/C & E-292 A-F, thereby reaching a temperature of approximately 424°F (218°C). Following that, the stream is split into three. Stream 1 returns back as vacuum bottoms quench to the bottom of the column. Stream 2 is routed to the asphalt production section. Stream 3 is cooled to 352°F using 60# steam generator E-293 A/B/C/D (producing LPS) before being cooled by a tempered water system in E-294 A-F and routed to the visbreaker plant as feedstock.

The side-cut 8 heavy vacuum gas oil (HVGO) stream currently leaves C-200 vacuum distillation column at approximately 641°F (338°C) and is split into two streams. Stream 1 returns to the column as wash oil. Stream 2 exchanges heat with the flashed crude stream in E-281 A/B/C and E-282 A/B/C, reaching a temperature of approximately 419°F (215°C). Following that, Stream 2 is split into two more streams. Stream 2A returns to the column as vacuum bottom circulating reflux (VBCR). Stream 2B is the net S/C 8 HVGO product. This is cooled by further preheating the flashed crude stream in E-283 A/B, followed by cooling with tempered water in E-284 A/B/C/D and routed to the hydrocracker plant as feedstock.

The atmospheric bottom pumparound (ABPA) stream currently leaves C-100 crude distillation column at approximately 574°F (301°C) and exchanges heat with the flashed crude stream in E-145 A/B/C, reaching a temperature of approximately 488°F (253°C). Following that, the stream exchanges heat with unstabilised naphtha (C300 BTM) in E-305 (stabiliser reboiler), reaching a temperature of approximately 455°F (235°C), then is cooled to 417°F (214°C) using 150# steam generator E-146 (producing medium pressure steam) before being routed back to C-100.

The material balance showed only 0.03% losses, attributed primarily to hydrocarbon losses in the aqueous phase (desalter and overhead systems of the atmospheric and vacuum columns). These losses were found to be well within acceptable engineering limits (1-2% maximum) for crude distillation units.

A heat balance performed on the ADU/VDU plant showed that the crude requires a heating duty of approximately 999 MMBtu/h from the battery limit conditions to the atmospheric distillation column (C-100), of which the crude preheat (primary and secondary heat exchanger network) contributes approximately 622 MMBtu/hr (62.1%) and the furnaces contribute the remaining 378 MMBtu/hr (37.9%). The primary aim of this study was to modify the existing heat exchanger network integration to be capable of increasing the crude preheating operation. In turn, this will reduce the required furnace duties during normal operations and will also reduce fuel gas consumption.

Crude preheat: heat exchanger network modifications

A total of approximately 56.8 MMBtu/hr can be saved in crude furnace F-100 A/B by raising the temperature of the mixed crude feed going to F-100 A/B by 20.1°F; by modifying the circuits of the vacuum bottom/residue product, side cut 8 (SC-8) heavy vacuum gas oil (HVGO) and atmospheric bottom pumparound (ABCR) streams, with the introduction of three new heat exchangers to the secondary crude preheat train.

Modifications to vacuum bottom/residue product stream

Putting a new heat exchanger on the heat-load path (E-293 A/B/C/D, E-292 A-F, E-291 A/B/C, NHX1, F-100 A/B) allows some load transfer from E-293 A/B/C/D to NHX1 and so reduces F-100 heat load. To implement this initiative, another new heat exchanger (NHX2) must be installed in parallel with E-292 A-F to maintain the vacuum bottoms quench temperature returning to C-200. Vacuum residue is cooled to 395°F (202°C) using 150# steam generator E-146 (producing medium pressure steam) before being cooled to 345°F (174°C) using 60# steam generator E-293 A/B/C/D (producing low pressure steam) before finally being cooled by tempered water system in E-294 A-F and routed to the visbreaker plant as feedstock. This modification will result in a change in the vacuum bottom stream temperature profiles into E-291 and E-292, which proved to be acceptable within the current exchanger’s area of heat transfer. It is worth noting that this case also takes into account the modification following completion of the Clean Fuels & Aromatics Project, in particular regarding the 5 500 b/d vacuum overflash stream (675°F) being mixed with the vacuum residue (C-200 bottoms product) stream upstream of NHX-1.

Two new heat exchangers are required – NHX1 and NHX2. Modifications to atmospheric bottom pumparound (ABPA) The ABPA will be rerouted to exchange with crude in E-282 downstream E-145, then to E-305 (unstabilised naphtha reboiler) to exchange heat with unstabilised naphtha from C-300 bottom to reach 417°F, then back to the atmospheric column. This requires no new heat exchangers.

Modification to E-146 (not in use) will be required. Modifications to side cut 8 HVGO product streams

A total load of E-282 will be transferred to a new heat exchanger (NHX3) between the SC8 and furnace feed. E-282 will be used to exchange heat between the bottom ABCR and crude feed as described above. One new heat exchanger is required – NHX3.

It should be noted that the proposed modifications maintain all column returned temperatures and so the column temperature profile, as well as the product battery limit temperatures, will have no impact on the current process operation. It is worth mentioning that a hydraulic study was performed and validated the capability of the entire hydraulic system, especially the operation of some product pumps, due to the increased pressure drops resulting from the additional exchangers on the product and crude circuits. The modifications are summarised in Figure 5.

Crude preheat modification: potential savings

The potential fuel savings are considered to be about 56.8 MMBtu/h of fuel gas consumption in the crude furnaces, equivalent to approximately 1.8 MMSCF/day. A net present value of $21 832 900.00 and a rate of return of 24.01% is expected. The potential fuel gas savings summary for F-100A/B is shown in Table 2.

Crude preheat modification: steam requirement

In the base case, E-146 was used to generate about 24 700 lb/hr of medium pressure steam by exchanging heat with the ABPA stream with a total duty of 21 MMBtu/h, while E-293 was used to generate about 42 700 lb/hr of low pressure steam by exchanging heat with the vacuum residue product stream with a total duty of 40 MMBtu/h.

The modification proposed changing the steam generation configuration to the following:

• E-146 to be used to generate 19 100 lb/hr of medium pressure steam by exchanging heat with the vacuum residue product upstream E-293 with a total duty of 16.6 MMBtu/hr
• E-293 to be used to generate 33 500 lb/hr of low pressure steam by exchanging heat with the vacuum residue product downstream E-146 with a total duty of 32.1 MMBtu/hr.

A steam requirements summary and modified temperatures are shown in Tables 3 and 4, respectively.

New heat exchanger equipment sizing

The study provides detailed equipment sizing for the proposed modifications to the heat exchanger network. The proposed new heat exchangers were designed using Aspen Exchanger Design and Rating (EDR) software with stream process data obtained from the Aspen Hysys simulation for ADU/VDU plants. Designed exchangers were then linked with the Aspen Hysys rating model of ADU/ VDU plants to simulate the effects of these new exchangers on the crude preheat train.

A total of 10 additional shells have an effective surface area of 73 850.90 ft2. A summary of the design data and cost of these exchangers is shown in Table 5.

Another method was used to estimate the new heat exchangers’ effective area and capital cost using Aspen Hysys calculated duties, Ft correction factor, LMTD, and estimated overall heat transfer coefficient (U). A summary of the design data and cost of these exchangers is shown in Table 6.

Economic study

Table 7 summarises the values used for the project’s economic evaluation.

Sensitivity analysis for net present value

A sensitivity analysis was performed on the project’s NPV for fuel gas savings (see Figure 6). The base case data is shown in blue, with a fuel gas saving of 56.8 MMBtu/h and a discount rate of 6.5%. The break-even case is shown in red for the sensitivity test performed. The case showed excellent project economics, which strengthened justification of the investment.

Impact of fuel gas saving

In the second sensitivity case, NPV was tested against varying fuel gas savings, which were the primary source of revenue to the project. As can be seen from the data in Figure 5, the project enjoys a very good position in terms of fuel gas savings, since the project’s NPV would be positive for fuel gas savings values higher than the break-even value of 22.28 MMBtu/h, which is approximately 60.78% below the base value of 56.8 MMBtu/hr. As a result, the estimated savings would need to drop in value by approximately 60.78% before causing the NPV to turn negative and rendering the project uneconomical. This situation is of course highly unlikely due to the conservative estimates developed for project capital inflated by 20% to cater for expected extra piping and fittings costs.

Results and conclusion

The primary focus of this study was to identify potential energy savings and enhancement of a ADU/VDU plant initiative by optimising plant heat integration.

The main recommendation for energy optimisation from this study is to enhance heat recovery (by retrofit) within the crude secondary preheat trains by re-routing three streams: atmospheric bottom pump around, SC-8 heavy vacuum gas oil, and vacuum residue The modifications save 56.8 MMBtu/h of fuel gas consumption in the atmospheric furnaces, equivalent to approximately 1.8 MMSCFD. The modifications were simulated in Aspen Hysys and the new heat exchangers were designed utilising Aspen EDR software. Following that, cost estimation of capex/opex was utilised for the economic analysis, which clearly points to the economic viability and feasibility of this proposed project with a NPV of $21.8 million and an internal rate of return of 24.01% based on the future projected rise in fuel gas prices. A sensitivity analysis was performed on the project’s NPV for both the discount rate and fuel gas saving, showing excellent project economics and so strengthening justification for the investment.

The benefit of fuel gas savings is two-fold: a reduction in the refinery’s environmental impact and carbon emissions to the atmosphere; and an economic value saving, in terms of the ability to reduce fuel gas imports from the national fuel gas grid, and/or passing this allocation to another user for power or industrial use. With current production shortages in fuel gas, the latter point becomes of significant importance.

The investment required for the project amounts to approximately $10.6 million for 10 new heat exchangers in the ADU/ VDU plant. The estimates developed for this study were quite conservative, and it is expected that the total investment value would be less than originally planned.

References

1. Worrell et al, 2009.
2. Energetics Study, PNNL Study, 2004.
3. Energy Information Administration Table S1 Energy Consumption Estimates by Source and End-Use Sector, 2008.
4. Chem. Eng. Transactions, Vol. 32, 2013.
5. Tjoe T N, Linnhoff B, Using pinch technology for process retrofit, Chemical Engineering, 93(8), 1986, 47-60.
6. Noureldin M B, Methods for Heat Exchanger Network Energy Efficiency Assessment and Lifetime Retrofit, USP 8,150,560, 2012.
7. Hohmann E L, Optimum Networks for Heat Exchange, Ph.D. Thesis Univ. S. Calif., 1971.
8. Linnhoff B, Flower J R, Synthesis of heat exchanger networks, AIChE Journal 24, 1978, 633.
9. Linnhoff B, Kotjabasakis E, Downstream paths for operable process design, CEP, May 1986, 23.
10. Kotjabasakis E, Linnhoff B, Sensitivity Tables for the Design of Flexible Processes (1) - How much contingency in heat exchanger networks is cost effective?, Chem. Eng. Res. Des. 64, 1986, 197.
11.  Umeda T, ltoh J, Shiroko K, Heat exchanger systems synthesis, D.E.P., July 1978, 70.
12. Sargent R W H, Grossmann I E, Optimum design of heat exchanger networks, Comp. Chem. Eng. 2, 1978, 1.
13. Linnhoff B, Turner J A, Heat recovery networks: new insights yield big savings, Chern. Eng. 2, 1981, 56
14. HEXTRAN Manual, Simulation Sciences, Fullerton, CA.
15. SUPERTARGET, available from Linnhoff-March, Manchester, UK.