1、2010 ASHRAE 525ABSTRACTThis paper describes the results of an analysis of theopportunity for industrial waste heat to power in the UnitedStates using the organic Rankine cycle. The EPA NationalEmissions Inventory databases are used to quantify the avail-able heat content and temperature of the sourc
2、es. By frequency,the majority of waste heat sources are at temperatures below450F (232C) however, more than half of the total opportunityfor waste heat to power comes from sources with exhaust gastemperature between 500F and 1000F (260 and 538C).While these temperatures are not high enough to make s
3、teambased generation attractive they are high enough that workingfluid decomposition must be considered in the opportunityanalysis. For sources under 1000F (538C) including thelimitations of working fluid decomposition brings the techni-cally recoverable power from 44 to 32 GW. Total opportunity,inc
4、luding all sources over 300F (149C) is estimated to be 51GW. In addition to opportunity analysis the kinetics of workingfluid decomposition are discussed and calculated for severalwidely used fluids as a function of temperature.INTRODUCTIONLawrence Berkeley National Laboratory (LBNL) pre-dicts there
5、 is 100 GW of waste heat to electric power potentialin the United States (Baily and Worrell 2005). This figure rep-resents over 10% of the current installed capacity of electricgenerators in the country. This potential presents an opportu-nity to produce low cost, virtually zero emissions, local gen
6、-eration to assist in meeting power quality needs and pendingclean energy regulations such as Renewable Portfolio Stan-dards.In many cases industrial process heat is discharged to theatmosphere still containing 60+% of the heat from thecombustion process. Unless there is an opportunity for use ofthi
7、s heat at the industrial site, it is almost certainly wasted.However, conversion of the waste heat to electricity providesenergy that can be used either at the site, or it can be econom-ically transported over long distances to another customer. Aslong ago as 1979, studies investigated technology ca
8、pable oftransforming waste heat into electricity (General Electric1979). With increases in fuel and electricity prices, theeconomic case for harvesting this wasted energy has becomeeven more compelling.Industrial heat sources vary widely with respect to size,exhaust temperature, primary fuel source,
9、 duty cycle, and con-taminant content. Depending on the relative influence of thesefactors, different technologies may be more appropriate thanothers for converting the heat to power. Rankine cycle powergeneration is a well known technology and, with steam as aworking fluid, is the basis for the vas
10、t majority of power gen-eration worldwide. It remains the premier choice for powergeneration for waste heat to electric power conversion in manyhigh temperature applications. Because of the thermal stabil-ity of steam it can be used in cases where the source temper-atures are very high without fear
11、of thermal decomposition.However, because of turbine size, vacuum conditions in thecondenser, and the need to avoid condensation in the turbine,steam is most appropriate for the largest sources of high tem-perature waste heat. Also, because of design considerationswith small molecular weight working
12、 fluids, steam turbinesgenerally have lower efficiency than organic working fluidturbines for sizes below several megawatts (Table 1 afterAbbin and Leuenberger 1974). For waste heat sources belowOrganic Rankine Cycle Working Fluid Considerations for Waste Heat to Power ApplicationsDavid J. Schroeder
13、, PhD Neil Leslie, PEMember ASHRAEDavid Schroeder is assistant professor in the Department of Engineering Technology, Northern Illinois University, Dekalb, IL. Neil Leslie isR Paul and Marek1934; Morgan et al. 1935; Frey and Hepp 1933; Pease 1928).More recent measurements were taken in the temperatu
14、rerange of interest but were not performed over a range oftemperatures (Andersen and Bruno 2005). The results of thesestudies have been re-plotted in Figure 3. Reaction rateconstants as a function of temperature were determined usingthe Arrhenius equation:k = AeEa/RT(1)wherek = reaction rate constan
15、t in s-1A = pre-exponential constantEa= activation energyR = gas constantT = temperature in absolute unitsThe rate of decomposition may then be determined by:Figure 1 T-s diagram for several working fluids.Figure 2 T-s behavior of the linear alkanes and toluene.Note that critical point increases wit
16、h chain length. 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions 2010, Vol. 116, Part 1. For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permit
17、ted without ASHRAEs prior written permission. 528 ASHRAE Transactions(2)wherer = decomposition rateC = concentration of the working fluidt = timeThe amount of working fluid remaining is then given by:C = C0ekt(3)where C0is the initial concentration of the working fluid.The early temperature dependen
18、t studies cited here haveconcluded that the decomposition mechanism is a simplefirst-order decomposition reaction for all molecules studied.While there may be other, faster, processes that dominate de-composition at lower temperatures first order decompositionshould provide a lower bound of the deco
19、mposition rate. Ex-trapolating the results into a lower temperature regime allowsan estimate of decomposition rate at the temperatures of in-terest for ORCs.There are several trends of interest from Figure 3. First,smaller molecules decompose more slowly than larger ones.This is expected assuming en
20、tropy is the driving force for firstorder thermal decomposition. Branched molecules decom-pose more rapidly than straight chains, and ring structures,particularly benzene, are more stable than linear chains. Thisis also expected because of the added stability introduced bycarbons resonant structures
21、. Finally, when similar moleculeswere tested at lower temperatures, their stability was signifi-cantly worse than predicted by high temperature data extrap-olation. This may indicate that some mechanism other thanthermally activated first order decomposition becomes domi-nant at lower temperatures.
22、One possibility is that the surfaceof the reaction vessel catalyzes the decomposition at lowtemperatures.In Figure 3, k is the reaction rate constant, the dashedvertical line near 1000F is the low temperature bound of theearly temperature dependent studies, and the horizontaldashed lines represent t
23、he threshold for the rate constant where1% of the fluid decays in the time period indicated. These canbe used as a guide to estimate the maximum allowable work-ing fluid temperature for a given fluid decomposition toler-ance. However, the lack of data in the temperature range ofinterest for technolo
24、gically important working fluids is aconcern that needs to be addressed by the research community.Figure 3 Decomposition rate constant for various hydrocarbon fluids. Lines represent calculated values based on earlierwork done over 1000F individual data points represent more recent lower temperature
25、 work.rdC()dt- kC()= 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions 2010, Vol. 116, Part 1. For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not p
26、ermitted without ASHRAEs prior written permission. ASHRAE Transactions 529An ORC working fluid is circulated at temperaturesbetween ambient and the maximum temperature of thesystem. So if, for example, one percent of the fluid decompos-ing is tolerable and the maximum temperature of the system issuc
27、h that one percent is expected to decompose per year at thattemperature, satisfactory operation for ten or more years maybe possible because the fluid is not exposed to the maximumtemperature for a large fraction of the cycle. To determinemore accurate measures, these temperature dependent rateconst
28、ants could be combined with modeling that takes intoaccount the temperature profile of the fluid with time. Thisalso highlights the importance of avoiding hot spots in a heatexchanger that is near the maximum allowable temperature.In addition to the data presented in Figure 3, there is somevariation
29、 in reports regarding acceptable working temperaturefor some fluids. For example, Andersen and Bruno (2005)concluded that toluene has an unacceptable decompositionrate at 600F, but Marciniak et al. (1981) reported that it canbe used to 750F (399C), though they do so without referenceor supporting da
30、ta. Cole found toluene to be a stable workingfluid to 677F (358C) and was expected to be stable at leastto 750F (399C) provided that oxygen was excluded from thesystem (Cole et al. 1987). Researchers in that study suggestedthat years of operation should be possible between fluidchanges. Baton (2000)
31、 reported that in one facility operatingwith toluene, working fluid decomposition products of tolu-ene were found after several thousand hours of operation at700F (371C) hot side temperature. But another facility oper-ating at 750F (399C) hot side temperature had not shown anysigns of decomposition.
32、 Differences between laboratorymeasurements of decomposition and field observations mightbe explained by differences in measurement methodology.Specifically, the field workers are looking for visibly obvioussigns of decomposition such as black chunks or residue, whilelaboratory workers are using ins
33、truments to measure concen-trations of decomposition products quantitatively. If this is thecase, it may indicate that some of the decomposition productsare largely benign.Figure 4 highlights the conflicting needs of high criticalpoint and high resistance to thermal decomposition for the n-alkanes.
34、For example, if 1%/year degradation is the maximumallowable for a heat source at 500F (260C) a fluid wouldhave to be chosen which is working above its critical point.The data displayed for decomposition is meant to illustrate thetradeoff that exists between stability and critical point, the dataused
35、 is based on the early high temperature work which maysignificantly overestimate stability. For example, Figure 3shows more recent low temperature data for n-pentane whichshows decomposition of pentane faster than 1%/30 days at600F (316C) while extrapolating higher temperature datainto this range su
36、ggests it would be closer to 1%/year.Fluorinated refrigerants such as R-245fa (1,1,1,3,3-pentafluoropropane) are also actively used as ORC fluids.Angelino and Invernizzi (2003) reported that this compound isstable for at least 50 hours at 572F (300C) but at 626F(330C) decomposition is rapid (Angelin
37、o and Invernizzi2003). Their results are in contrast with a representative of themanufacturer of R-245fa, who suggests that working fluidtemperatures much above 300F (149C) should be avoideddue to observations of fluorine formation attributed to decom-position (Zyhowski 2008). Further, manufacturers
38、 of ORCequipment using R-245fa generally limit the maximum work-ing fluid temperature to 300F. While the C-F bond strength ishigh, the contributions of entropy driving the decomposition isalso larger for the refrigerants than it is for alkanes of the samechain length due to the larger number of atom
39、 types. This iswhat drives the auto ignition temperature of the HCFCs to belower than alkanes of the same chain length (The EngineeringToolbox 2008; BOC Gases 2008). Therefore, a lower decom-position temperature should be expected for these compoundsthan their non-halogenated relatives. An explanati
40、on for thedifference in opinion regarding operating temperature maysimply be the length of time over which the experiments wereconducted and the sensitivity limits available to measuredecomposition. Because the decomposition process is a ther-mally activated one, moving from 572 to 626F (300 to 330C
41、)could increase the decomposition rate from 1%/day to 1%/hour (which was the stated maximum sensitivity for Angelinoand Invernizzi 2003). A change this dramatic is consistentwith the activation energies for alkane thermal decomposition.The greater bond strength of the C-F bond compared to the C-H or
42、 C-C bond would likely make the activation energy largerand the change in decomposition rate even more abrupt.Figure 4 Calculated temperature where decomposition rateis 1% per year for n-alkanes and critical point vs.number of carbons in the chain. Note that thesecalculated values are based on first
43、 order decom-position only and yield significantly higher de-composition temperatures compared to lowertemperature experimental data. 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions 2010, Vol. 116, Part 1. For p
44、ersonal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted without ASHRAEs prior written permission. 530 ASHRAE TransactionsThermal decomposition of ORC working fluids is nor-mally avoided by using an intermediate heat transfer fluid tos
45、eparate the working fluid from the high temperature exhauststreams. This has the disadvantage of reducing the attainableconversion efficiency by reducing the maximum temperatureof the cycle. It also increases capital cost by requiring an addi-tional heat exchanger. Even when the maximum temperaturei
46、s limited by use of a secondary heat exchanger and thermaloil, decomposition of pentane has been noted in some facilitieswhere low winter operating temperatures cause vacuum con-ditions in part of the system (Sweetser and Leslie 2007).Vacuum conditions can result in air infiltration into the system,
47、particularly if it was designed for positive pressure operation.Entrained air will significantly lower decomposition temper-ature.In addition to the concern of loss of the working fluiditself to decomposition, there is the concern of potential safetyhazards and equipment damage depending on the natu
48、re ofdecomposition products. In the case of alkanes as workingfluids, products of decomposition will likely be smaller alkanechains, hydrogen gas, and eventually carbon (Marek andMcCluer 1931; Paul and Marek 1934; Morgan et al. 1935;Frey and Hepp 1933; Pease 1928). These products are notparticularly
49、 concerning unless the working fluid is diluted tothe point that system performance is affected or carbon isdeposited as a solid. Solid carbon can reduce the efficiency ofheat exchangers as well as eventually clogging them. Also, theturbine could be damaged by carbon particulates. Fluorinatedhydrocarbons, on the other hand, may form HF or F2duringdecomposition in addition to possibly forming carbon andhydrogen. There is a significantly greater safety concern inthis case. This is an interesting point considering that safety,by virtue o
