Sunteți pe pagina 1din 7

_______________________________________________________________

_______________________________________________________________

Report Information from ProQuest


September 11 2012 06:48 _______________________________________________________________

Table of Contents
1. Selecting Streams In Pinch Systems............................................................................................................ 1 Bibliography...................................................................................................................................................... 5

11 September 2012

ii

ProQuest

Document 1 of 1

Selecting Streams In Pinch Systems


Author: Gill, B S. Publication info: Chemical Engineering 112. 5 (May 2005): 55-57. ProQuest document link Abstract: When pinch technology is used, design of the heat-transfer network must be approached differently from how it is with systems that use standalone hot and cold sources. Perhaps the most noticeable difference with pinch systems is the need to select matches between hot and cold streams while developing a train of exchangerse. Up to now, most engineers have solved each task on a case by case basis, sometimes going to lengths that verge upon reinventing the wheel. The better strategy, however, is to use a principles-based method that can be followed consistently with ease, a method termed the maximum flux method. When multiple choices exist for possible streams, the maximum flux method allows the engineer to determine the stream with the maximum Q/A, provided that the corresponding heat exchanger being is not unrealistically small. Full Text: Headnote This method simplifies the process and converts it into an easy-to-follow procedure (ProQuest Information and Learning: ... denotes formulae omitted.) When pinch technology is used, design of the heat-transfer network must be approached differently from how it is with systems that use standalone hot and cold sources [1]. Perhaps the most noticeable difference with pinch systems is the need to select matches between hot and cold streams while developing a train of exchangers - in a petroleum refinery preheat train, for instance. Up to now, most engineers have solved each task on a case by case basis, sometimes going to lengths that verge upon reinventing the wheel. The better strategy, however, is to use a principles-based method that can be followed consistently with ease, a method termed the maximum flux method. When multiple choices exist for possible streams, the maximum flux method allows the engineer to determine the stream with the maximum Q/A, provided that the corresponding heat exchanger being is not unrealistically small. Method of calculation Heat exchangers are designed using the equation Q = U.A.F^sub T^.T^sub LM^ (1) Imagine a situation wherein a hot stream available at a temperature T^sub 1^ is to be cooled by a cold stream available at temperature t^sub 1^. As the heat transfer proceeds, the hot stream gets cooler whereas the cold stream heats up (Figure 1). To what extent the hot stream will be cooled (or the cold stream will be heated) will be decided by the amount of area provided for heat transfer. Higher area will result in cooler T^sub 2^ (or hotter t^sub 2^). The log mean temperature difference is ... (2) If the heat flow capacities (product of heat flowrate and mean specific heat) of the hot and cold streams are H and C, respectively, we can write the following relationship: ... (3) Equation (2) can then be re-written: ... (4) where R = C/H Since T^sub 1^ and t^sub 1^ are fixed, T^sub LM^ is a function of t^sub 2^ only. (This method does not apply when R=1.) The overall heat transfer coefficient U is a function of physical properties of the fluids and flow velocities and is usually not significantly dependent on temperature changes. Equation (1) can now be rewritten as follows: ... (5) Obviously a heat exchanger with larger value of Q/A would be considered more efficient and economical since it would transfer higher amount of heat per unit heat-transfer area A; or, in other words, it would need smaller heat transfer area for transferring a given amount of heat. As heat transfer area increases, t^sub 2^ increases. Figure 2 shows a typical dependence of Q/A on t^sub 2^. The dependence of Q/A on A will follow a similar trend with Q/A remaining finite, as A approaches zero (Figure 3). It can be easily shown that ... (6) Example 1 Consider a stage in the development of a heat exchanger train where a cold stream is selected that may be matched with any one of four hot streams. This is typical of network development below pinch point. (Above the pinch point, the problem would be matching a selected hot stream with one of several possible cold streams.) Stream data for these cases are given in Table 1. For illustration purposes the outlet temperatures of the hot streams have been adjusted so that each hot stream completely meets the heat duty of the cold stream on its own. The goal is to 11 September 2012 Page 1 of 5 ProQuest

select the most 'economical' hot stream match for the selected cold stream. The most economical match is defined here as one that requires smallest heat-transfer area. The heat duty required by the cold stream (the product of its temperature difference and its heat capacity) is 14.7 Gcal/h. For now, we assume F^sub T^ = 1.0 for all possible matches. (After selection of the stream, the heat transfer area can be adjusted for the true value of F^sub T^). T^sub LM^ for all possible matches are as follows: T^sub LM^ (C1-H1) = 66.4C T^sub LM^ (C1-H2) = 17.3C T^sub LM^ (C1-H3) = 67.9C T^sub LM^ (C1-H4) = 50.1C Overall heat transfer coefficients for all matches are as follows: U ^sub (C1-H1)^ = 87.4 kcaym^sup 2^.h.C U ^sub (C1-H2)^ = 105.8 kcal/m^sup 2^.h.C U ^sub (C1-H3)^ = 89.7 kcal/m^sup 2^.h.C U ^sub (C1-H4)^ = 91.9 kcal/m^sup 2^.h.C Thus, the product of T^sub LM^ and U is equivalent to Q/A. Q/A ^sub (C1-H1)^ = 5,803.4 kcal/m^sup 2^.h Q/A ^sub (C1-H2)^ = 1,830.3 kcal/m^sup 2^.h Q/A ^sub (C1-H3)^ = 6,090.6 kcal/m^sup 2^.h Q/A ^sub (C1-H1)^ = 4,604.2 kcal/m^sup 2^.h This shows that the C1-H3 match would be most economical (since its Q/A value is highest) as it will result in the smallest exchanger size. By comparison, the match C1-H2 would be the worst. Conveniently, this procedure bypasses the actual calculation of required heat transfer area on the front end. So, the heat exchanger area is now calculated as Q/(U.T^sub LM^) ... (This match may need multiple shells as it may be difficult to accommodate so much area in one shell). Just for interest, other match areas are as follows: A ^sub (C1-H1)^ = 2,533 m^sup 2^ A ^sub (C1-H2)^ = 8,031 m^sup 2^ A ^sub (C1-H4)^ = 3,193 m^sup 2^ A computer program that automates the above procedure can easily take the effect of fouling and wall resistance in the calculation of U. Also standard equations are available in literature for the calculation of F^sub T^ factors based on the knowledge of the stream temperatures and the ratio of their heat-capacity rates [2]. Example 2 Now imagine that in Example 1, the stream H2 needs to be cooled to only 150C (Let the new stream be named as H2', and note that the C1-H2' match no longer satisfies the original specifications of stream C1. The C1 stream outlet temperature in this case would only be 128.6C.) ... Thus ... Thus the match C1-H2' becomes most desirable since its Q/A value is now highest. Heat exchanger area for such a match would be 753 m^sup 2^. One can provide this match and reanalyze the balance problem and locate the next best match and so on. Since the heat exchanger network is developed by selecting one match at a time, the problem remaining to be solved after placement of a selected match is referred to as the balance problem. In an alternative scenario, if the C1 stream outlet temperature needs to be kept at its original value of 200 C, its inlet temperature would have to be 131.4 C. Then ... This shows that the C1-H2' match is the least preferred since its Q/A value is now lowest. A further improvement via the suggested method is especially applicable where several shells are required for a match. In Example 1, heat exchanger area required for the selected match is 2,413 m^sup 2^ which may require about six shells of industrial size exchangers of area 400 m^sup 2^ each. If we presume F^sub T^ = 1 for the time being, Equation (5) reduces to ... (5a) Minimizing the number of shells. Let us try to determine the value of Q/A when a hypothetical "zero-area exchanger" is provided. In this case no heat is transferred, and t^sub 1^=t^sub 2^. This makes the right-hand side of Equation (5a) indeterminate with respect to the numerator. To find its value, we differentiate the numerator and the denominator. There are two possibilities: when the cold stream inlet temperature is fixed (differentiating with respect to t^sub 2^) and when the cold stream outlet temperature is fixed (differentiating with respect to t^sub 1^). In the former case ... (6a) In the latter case ... (6b) [Below the pinch point, relationship (6b) would be relevant]. For the data in Example 1, values of Q/A at A=0 are listed in Table 2. This data suggests that at this hypothetical condition, a match between C1 and H1 should be made (since its Q/A value is highest). If the required heat-transfer area is more than what can be accommodated by a single-shell exchanger, limit the area to a single-shell unit. Reanalyze the balance problem for the next match, again limiting the area to a single-shell area. Proceed for further matches on similar lines. As the selected match becomes bigger (heat exchanger area is increased), the value of Q/A keeps going down. At the same time, the potential of other possible matches increases, eventually overtaking the Q/A value of the selected match. Figure 4 represents this phenomenon for the data in Example 1 when the C1-H1 match is selected in the first place. A study of such a figure will guide as to how much area the chosen 11 September 2012 Page 2 of 5 ProQuest

heat exchange match needs to have. If such logic is followed over the total problem, the resultant train of exchangers will be an ideal one from the point of view of highest heat recovery at lowest capital cost. As can be seen, in the proposed procedure, physical properties of streams (which go into calculation of stream heat transfer coefficients) play a significant role in the development of heat exchanger networks; this is not the case with the existing methods. As an illustration, if the hot stream H1 were a heavy viscous stream with a heat transfer coefficient of 50 instead of 210, Q/A at A = 0 for the match C1-H1 would be only 8,114 kcal/m^sup 2^.h (t^sub 1^ fixed case) and would be the least desired match since its Q/A value would then be least. Edited by Rebekkah Marshall Sidebar NOMENCLATURE Q Heat transferred from the hot stream to the cold stream U Overall heat transfer coefficient A Effective heat transfer area of the heat exchanger F^sub T^ Log mean temperature difference correction factor T^sub LM^ Log mean temperature difference between the heat exchanging streams References References 1. Linhoff, B. and Flower, J.R., Synthesis of heat exchanger networks, Part 1: systematic generation of energy optimal networks, AIChEJ, Vol 24, No. 4, July, 1978. 2. Kern, D.Q. "Process Heat Transfer", McGraw Hill, New York, 1990. AuthorAffiliation B.S.Gill Engineers India Ltd. AuthorAffiliation Author Bhupinder Gill is currently general manager of the environmental engineering division of Engineers India Ltd. (H-45, Sarita Vihar, New Delhi - 110066, India; Phone: 91-11-610-7001; Fax:91-11-6186734; Email: bsgill39@hotmail.com). Previously, he was head of the research and development division of EIL. His areas of specialization include fired heaters, process integration, heat transfer and fluid flow. Dr. Gill has also worked at the Central Mechanical Engineering Research Institute in Durgapur, India and the Research and development division of Bharat Heavy Electricals Ltd., in Hyderabad, India. He holds a Ph.D in mechanical engineering. Subject: Methods; Chemical engineering; Heat transfer; Mathematics Classification: 8640: Chemical industry Publication title: Chemical Engineering Volume: 112 Issue: 5 Pages: 55-57 Number of pages: 3 Publication year: 2005 Publication date: May 2005 Year: 2005 Section: Engineering Practice Publisher: Access Intelligence LLC Place of publication: New York Country of publication: United States Journal subject: Chemistry, Engineering--Chemical Engineering ISSN: 00092460 Source type: Scholarly Journals Language of publication: English Document type: General Information

11 September 2012

Page 3 of 5

ProQuest

ProQuest document ID: 194460029 Document URL: http://search.proquest.com/docview/194460029?accountid=27542 Copyright: Copyright Chemical Week Associates May 2005 Last updated: 2010-08-21 Database: ABI/INFORM Complete

11 September 2012

Page 4 of 5

ProQuest

Bibliography
Citation style: APA 6th - American Psychological Association, 6th Edition B, S. G. (2005). Selecting streams in pinch systems. Chemical Engineering, 112(5), 55-57. Retrieved from http://search.proquest.com/docview/194460029?accountid=27542

_______________________________________________________________
Contact ProQuest

Copyright 2012 ProQuest LLC. All rights reserved. - Terms and Conditions

11 September 2012

Page 5 of 5

ProQuest

S-ar putea să vă placă și