Transcript
Vacuum reliability analysis of parabolic trough receiver Jian Li a , Zhifeng Wang a, n , Jianbin Li b , Dongqiang Lei a a Key Laboratory of Solar Thermal Energy and Photovoltaic System, Institute of Electrical Engineering, Chinese Academy of Sciences, No. 6 Beiertiao, Zhongguancun, Beijing 100190, China b Himin Solar Co., Ltd., Dezhou, Shandong 253090, PR China a r t i c l e i n f o Article history: Received 13 February 2012Received in revised form7 June 2012Accepted 19 June 2012Available online 11 July 2012 Keywords: Vacuum reliabilityParabolic trough receiverHydrogen permeationOutgassingLeakGetter a b s t r a c t Experience has shown that the vacuum reliability and lifetime of parabolic trough receiver are one of the most significant issues for existing and future parabolic trough plants, affecting the plant efficiencyand revenue. The gas sources within parabolic trough receiver and absorption characteristics of thegetter were evaluated to analyze the factors affecting vacuum reliability. The outgassing rate and thegas species within parabolic trough receiver were tested to show that besides hydrogen permeationcaused by thermal degradation of the organic heat transfer fluid, outgassing may also lead to hydrogenaccumulation. The analysis further shows that both high and low temperatures affect the getter abilityto absorb H 2 . O 2 , CO, N 2 and CO 2 , introduced into the receiver vacuum annulus by outgassing orleakage, significantly reduce the getter surface activity, which reduces the H 2 absorption capacity. & 2012 Elsevier B.V. All rights reserved. 1. Introduction The parabolic trough receiver is the key component of solarparabolic trough systems for electric power production, which isthe most mature and prominent concentrating solar powertechnology. There are over 1.3 GW of parabolic trough powerplants in operation and 1.7 GW more under construction [1]. Thereceiver is placed in the focal line of the parabolic trough collectorto absorb the solar radiation reflected by the parabolic trough. Aheat transfer fluid circulates through the receiver tube to trans-port the thermal energy to a heat exchanger system to generatehigh-pressure superheated steam which is used to drive a steamturbine to produce electricity.The parabolic trough receiver is a 4.06 m long, 70 mm outsidediameter (O.D.) stainless steel tube sputtered with a solar-selective coating, surrounded by a 115/120 mm O.D. glass tube.The tube has incorporated conventional glass-to-metal seals andmetal bellows to achieve the necessary vacuum-tight enclosureand to accommodate the thermal expansion difference betweenthe steel tube and the glass envelope. The vacuum enclosureserves primarily to significantly reduce heat losses at highoperating temperatures and to protect the solar-selective absorbersurface from oxidation. The initial vacuum is typically about10 2 Pa. The parabolic trough receiver has a graded multilayer‘‘Cermet’’ (ceramic-metal) coating that is sputtered, using vacuumdeposition techniques, onto the steel tube. The sputtered Cermetcoating provides good selective optical–thermal properties withhigh solar absorptance of direct-beam solar radiation and a lowthermal emissivity at the operating temperature to reduce thethermal radiation [2]. Fig. 1 shows a schematic of the parabolic trough receiver.Price et al. [3] found that the thermal losses from a parabolictrough receiver after most of the vacuum has been lost are muchhigher than with a good vacuum, especially for receivers filledwith hydrogen whose thermal losses can become as high as4 times those of a good vacuum. The annual plant revenue canthen be reduced by as much as 20% by receivers infiltrated withhydrogen as shown by parabolic trough plant simulations withthe Solar Advisor Model (SAM) [4]. Furthermore, receivers arewelded together to form solar collector loops in the parabolicpower plant, so replacing failed receivers is a very complicated,expensive process. Experience has shown that the vacuum relia-bility and its effect on the lifetime of parabolic trough receivers isone of the most significant issues for existing and future parabolictrough plants [2].However, there are few studies of vacuum reliability and itseffect on the lifetime of parabolic trough receivers, with mostworks focusing on the thermal performance of the tubes [3–8] and their solar-selective coating [9–14]. This study analyzes the sources leading to pressure increases in parabolic trough receivers.Then, based on the outgassing experiments and operation Contents lists available at SciVerse ScienceDirectjournal homepage: www.elsevier.com/locate/solmat Solar Energy Materials & Solar Cells 0927-0248/$-see front matter & 2012 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.solmat.2012.06.034 n Corresponding author. Tel.: þ 86 10 62520684; fax: þ 86 10 62587946. E-mail addresses:
[email protected] (J. Li),
[email protected] (Z. Wang).Solar Energy Materials & Solar Cells 105 (2012) 302–308 conditions in parabolic trough plant, the factors affecting vacuumreliability of parabolic trough receiver are analyzed, including theeffects of getters. 2. Sources of gas in parabolic trough receiver The effect of gases and their importance in evacuated devicesdepend on the purpose of the vacuum. Some devices need avacuum to avoid interception of an electron beam and reactionswith electron emitters, such as CRTs and FEDs. Some need avacuum to introduce a specific functional gas and maintain itspurity to avoid changes in certain parameters, such as lamps andplasma displays. Still others need high thermal insulation, such asthermos bottles and parabolic trough receivers. For the thermalinsulation, the thermal conductivity of the residual gas and thetotal residual pressure in the parabolic trough receiver are veryimportant. To maintain good thermal performance, the gasspecies and the total pressure in the parabolic trough receivermust be controlled.The vacuum degradation of a vacuum device is generally dueto the following:(1) gas resulting from evaporation of materials in the system.(2) outgassing of the materials in the system.(3) gas permeation through walls or windows in the vacuumsystem.(4) gas penetrating into the vacuum system as a result of leaks.The main materials used in parabolic trough receivers arestainless steel and borosilicate glass, which are both good vacuummaterials often used in high vacuum systems. The solar selective-coating sputtered on the absorber tube surface is composed of various metals, such as Al, Ti, W or Mo [10,13–15]. The maximum operating temperature of the parabolic trough receiver is about400 1 C and the vapor pressure of these materials at this tempera-ture is so low that this vapor source can be neglected.The other sources are discussed here in detail. 2.1. Outgassing Any material surface exposed to normal air at atmosphericpressure for some time will be in equilibrium with the air andcovered by an adsorbed phase. The adsorbed gas will then desorbunder low pressure conditions. The material also contains gasestrapped inside the material during manufacture. In a vacuum,these trapped gases diffuse slowly to the surface and desorb intothe low pressure region. The combination of adsorption, diffusionand desorption during the outgassing process is quite complex,with no studies of outgassing of parabolic trough receivers.Experiments are given to measure the outgassing rates and gasspecies in parabolic trough receivers. 2.1.1. Outgassing rate of a parabolic trough receiver 2.1.1.1. Sample preparation. 0.7 m long parabolic trough receiversamples were made by Himin Solar Corporation, the only receiversupplier for the first Chinese 50 MW commercial parabolic troughpower plant. One side of the samples had a evacuation nozzleconnected to vacuum pump system while the other side had asmall tube connected to a vacuum gage as shown in Fig. 1. 2.1.1.2. Apparatus and procedure. The throughput method [16]was used to measure the outgassing rate of the parabolic troughreceiver sample. a schematic diagram of the system for themeasurements is shown in Fig. 2. An oven with a maximumtemperature of 600 1 C was used for the heat treatment. Theparabolic trough receiver was connected to a vacuum chamberby a glass evacuation tube. The pressures in the parabolic troughreceiver sample and the vacuum chamber were measured by hotcathode gages.All the outgassing rate measurements were carried out withthe following procedures. Before the tests, the dimensions of the glass evacuation tube were measured to calculate the gasconductance C . Then, the parabolic trough receiver sample was Fig. 1. Schematic of a parabolic trough receiver. Fig. 2. Schematic diagram for outgassing rate tests. J. Li et al. / Solar Energy Materials & Solar Cells 105 (2012) 302–308 303 evacuated at room temperature. When the pressure in the tubewas lower than 1 10 3 Pa, the oven was heated according to thesettings listed in Table 1 while the pressures were measured inthe vacuum tube and the chamber. 2.1.1.3. Results and discussions. The outgassing rate was calculatedbased on the measured pressure difference across the evacuationglass tube [16]: q ¼ C ð P tube P cham Þ L ð 1 Þ where q is the outgassing rate of the parabolic trough receiver(Pa m 3 s 1 m 1 ), C is the conductance of the glass evacuationtube (m 3 s 1 ), P tube is the pressure in the parabolic troughreceiver (Pa), P cham is the pressure in the vacuum chamber (Pa),and L is the length of the parabolic trough receiver sample (m).The outgassing rates of three parabolic trough receiver sampleswere measured, and the test results are shown in Figs. 3–5. Theresults indicate that when the heating temperature was raised, theoutgassing rates of the parabolic trough receiver rose rapidly to apeak, with the outgassing rate then falling exponentially with timeafter long pumping times. However, even with a very low out-gassing rate at low temperatures, the outgassing rate still roserapidly with increasing temperature because the gas desorptionand diffusion are both thermally activated processes whose ratesincrease dramatically with temperature. The desorption is givenby: dndt ¼ Kf 0 ð y Þ exp E des RT ð 2 Þ where dn/dt is the molecules desorption rate, f 0 ( y ) is a function of the surface coverage, K is a constant, E des is the desorptionactivation energy and R is the universal gas constant. Gas diffusionin a solid is proportional to the concentration gradient of the gas inthe solid by Fick’s law [17]: q dif ¼ Ddc dx D ¼ D 0 exp E dif RT ð 3 Þ where q dif is the gas diffusion rate in the x direction, dc/dx is theconcentrationgradient, D isthediffusioncoefficient attemperature T , D 0 is the diffusion constant and E dif is the activation energy fordiffusion.Another reason for the rapid desorption increase with tem-perature is that gases in the vacuum materials exist in multiplebulk states [18]. During outgassing, only the occupied diffusionstate with the lowest energy will be significantly involved at lowtemperatures. At high temperatures, strongly bound gas mole-cules partially evolve while the rest are redistributed to differentbinding states in the material bulk. The gases in the materialswith high bound energies cannot be degassed at low tempera-tures, so the degassing baking temperature of parabolic troughreceivers cannot be lower than the operating temperature;otherwise, the outgassing rate at the operating temperature willnot be reduced.The experimental results for the outgassing rate at onetemperature can be represented by the empirical equation [19] q h ¼ q 1 t h g ð 4 Þ where q h and q 1 are the outgassing rates at h hours and at 1 hafter starting (Pa m 3 s 1 m 1 ) and t h is the time in hours (h). Thetotal outgassing amount for a parabolic trough receiver during itslifetime with isothermal degassing conditions would then be Q o ¼ Z t l t o q h L dt ð 5 Þ where Q o is the outgassing amount of the parabolic troughreceiver during its lifetime (Pa.m 3 ), t o is the degassing baking Table 1 Heat settings for outgassing rate test.Temperature ( 1 C) Time20–200 20 min200–200 5 h200–400 20 min400–400 5 h Fig. 3. Outgassing rate of parabolic trough receiver sample #1 during heating. Fig. 4. Outgassing rate of parabolic trough receiver sample #2 during heating. Fig. 5. Outgassing rate of parabolic trough receiver sample #3 during heating. J. Li et al. / Solar Energy Materials & Solar Cells 105 (2012) 302–308 304 time at the selected temperature (s), and t l is total operating timeof the parabolic trough receiver (s). Thus, the total outgassing isdetermined not only by the outgassing rate, but also by thebaking time during the degassing process.The lifetime of a typical parabolic trough receiver is expectedto be 25 years operating 8 h per day, with a total operating time of about 73,000 h. Assuming the degassing baking settings for thereceiver are the same as in Table 1, the outgassing rates of thereceiver at 400 1 C during operation can be calculated by Eq. (4)using average value of parameters determined by fitting theoutgassing test results at 400 1 C shown in Figs. 6–8, and thefitting results of three samples are listed in Table 2. Then theoutgassing rates of the receiver at 400 1 C during operation are q h ¼ 8 : 04 10 5 t 0 : 933 ð 6 Þ The temperature of the parabolic trough receiver at night isvery low, so the outgassing rate is much lower than at theoperating temperature and can be neglected. Thus, the overalloutgassing amount for the parabolic trough receiver at 400 1 Cduring its operating lifetime will be Q o ¼ Z 73 , 000 36005 3600 8 : 04 10 5 t 3600 0 : 933 4 : 06 dt ¼ 17 : 6 Pam 3 ð 7 Þ Generally, the outside diameters of the steel absorber tube andthe glass envelope are 0.07 m and 0.12 m respectively, so thevolume of the vacuum annulus in the receiver is about 0.03 m 3 .Thus, the overall outgassing amount can result in the pressure inthe receiver increasing to 586 Pa, which is large enough tosignificantly degrade the performance of the parabolic troughreceiver. This analysis indicates that the parabolic trough receivermust be very well degassed during the manufacture process toreduce the outgassing during operation. However, the receiveroutgassing rates decrease exponentially with time after longpumping times, so after a few hours the outgassing rate will bevery slow, and the degassing efficiency will be decreased. Thus, acompromise is needed between the cost and the baking time formass production. 2.1.2. Outgassing species Different gases have different thermal conductivities whichaffect the thermal performance of the parabolic trough receiver.More importantly, the getter absorption capabilities for differentgases also differ. The outgassing species of glass are mainly H 2 O,CO and CO 2 , with most H 2 O removed during the heat treatmentprocess [20]. The outgassing species from the stainless steel withthe solar-selective coating have not been tested, so tests wereconducted using the gas accumulation method [16].Stainless steel samples were prepared with and without solar-selective coatings to compare the effect of the coatings on theoutgassing species. All the samples were cut into small flat pieces20 mm 20 mm 3 mm. The samples were then polished andcleaned by acetone and ethanol in an ultrasonic bath.The experiments were performed in a UHV apparatus providedwith a Quadruple Mass Spectrometer (QMS). The temperaturewas monitored using a thermocouple. At first, the vacuumchamber with the samples was evacuated to 10 4 Pa and then Fig. 6. Fitting curve of sample #1 test results at 400 1 C. Fig. 7. Fitting curve of sample #2 test results at 400 1 C. Fig. 8. Fitting curve of sample #3 test results at 400 1 C. Table 2 Fitting results of three samples. q 1 g Sample #1 6.95 10 5 0.95Sample #2 1.15 10 4 0.94Sample #3 5.66 10 5 0.91Average 8.04 10 - 5 0.933 J. Li et al. / Solar Energy Materials & Solar Cells 105 (2012) 302–308 305 isolated from the pumping system as the samples started to beheated. The released gases during the thermal treatment accu-mulated in the vacuum chamber and at specified times a smallvolume of the gases was collected in a sampling volume andanalyzed on the QMS system.Table 3 lists the outgassing species when the samples wereheated at 400 1 C for 4 h. The gas species desorbed by the sampleswithout the coating were hydrogen, hydrocarbons, carbon oxideand water. The coated samples also desorbed nitrogen and argon,which were probably trapped during the coating process anddesorbed during the thermal treatment. For all the samples,hydrogen is the primary species released, accounting for over90% of the total, so the outgassing can be assumed to be onesource for hydrogen built up in the receiver.Since Ar is not absorbed by the getters and the N 2 absorptioncapacity of the getters is very low, the parabolic trough receiverdegassing process should be emphasized to reduce the outgassingrates of Ar and N 2 to as low as possible. 2.2. Hydrogen permeation Moens and Blake [21] reviewed the most relevant chemicaland patent literature relating to the thermal decomposition of biphenyl and diphenyl oxides to support the notion that hydrogenformation occurs at temperatures around 400 1 C as a result of thermal degradation of the organic heat transfer fluid in parabolictrough plants. The degradation process proceeds at a very lowrate at 400 1 C, but accelerates considerably between 400 1 C and425 1 C. The process involves a radical reaction wherein hydrogengas is formed as a by-product, as well as aromatic oligomers withhigher molecular weights (high boilers). Organic impurities intro-duced during the production process can catalyze the thermalbreakdown of the heat transfer fluid through the thermal forma-tion of highly reactive hydrogen atoms. The inherent thermalinstability of the chemical bonds in the DPO/biphenyl fluidprecludes any chemical strategies to mitigate the formation of hydrogen gas. This hydrogen gas can then permeate through thesteel absorber tube into the vacuum annulus of the parabolictrough receiver, affecting the vacuum reliability and lifetime.The hydrogen permeation parameters of solar-selective coat-ings were measured using a gas phase permeation technique todetermine the hydrogen permeability of the receiver tube anddevelop a hydrogen permeation model of the parabolic troughreceiver, which facilitates measures to reduce hydrogen permea-tion into the receiver annulus [22].The test specimen was a thin 304 austenitic stainless steel(304SS) layer sputtered with solar-selective coating, with totalthickness 0.24 mm. The test results gave the correlation betweenthe hydrogen permeability of the specimen and temperature: F c ¼ 5 : 9 10 6 exp 57 , 500 RT ð 8 Þ where F c is the hydrogen permeability of the specimen(mol m 1 s 1 MPa 0.5 ), R is the universal gas constant, and T isthe temperature (K). The solar absorber tube can be considered tobe composed of a pure stainless steel layer and a stainlesssteel layer sputtered with the solar-selective coating like thespecimen in the test, and the hydrogen permeability of pureSS304 steel [23]is F s ¼ 2 : 85 10 4 exp 62 , 430 RT ð 9 Þ where F s is the hydrogen permeability of the specimen(mol m 1 s 1 MPa 0.5 ). Thus, the hydrogen permeability of the2.5 mm thick SS304 steel absorber tube sputtered with selectivecoatings can be calculated based on Eqs. (8) and (9) [24] F ab ¼ l ab F c F s F c l s þ F s l c ð 10 Þ l ab ¼ l s þ l c ð 11 Þ where F ab is the hydrogen permeability of the absorbertube (mol m 1 s - 1 MPa 0.5 ), l ab is absorber tube thickness (m), l s is thickness of the pure 304SS layer (m), and l c is thicknessof the SS304 layer sputtered with the solar-selective coating (m).From Eqs (8)–(11), we can know that the hydrogen permeabilityof 304SS absorber tube at 400 1 C is about 1.44 10 9 mol m 1 s 1 MPa 0.5 .According to the hydrogen permeation model, the effects of the hydrogen generation rate, the hydrogen pressure in thereceiver pipes and hydrogen barrier coating on the hydrogenpermeation process were analyzed. The model shows that thehydrogen generation rate significantly affects the hydrogen per-meation process, with the hydrogen pressure, permeability andadsorbed area all influencing how the hydrogen permeation rateapproaches the hydrogen generation rate. The hydrogen permea-tion amount into the parabolic trough receiver at low tempera-tures is much less than at high temperatures as shown in Fig. 9.The amount of hydrogen permeation into annulus over 25 years isabout 2 10 4 Pa m 3 at 390 1 C based on the ideal gas equation. Thehydrogen permeation amount is then much greater than theoutgassing, so permeation is the major source of hydrogenaccumulation in the annulus of parabolic trough receivers. 2.3. Leakage Vacuum systems cannot possibly have a zero leakage. Air willenter the vacuum system as a result of the pressure difference Table 3 Outgassing species fractions (%).H 2 CO N 2 CH 4 H 2 O C 2 H 6 C 3 H 8 Ar CO 2 Sample without coating 92.48 1.5 0 3.92 0.15 0.18 0.74 0 1.03Sample with coating 95.55 0.42 1.34 1.36 0.05 0.25 0.19 0.05 0.8 Fig. 9. Amount of hydrogen permeation amount into annulus over 25 years. J. Li et al. / Solar Energy Materials & Solar Cells 105 (2012) 302–308 306
