Investigation on Performance Prediction and Operation Characteristics of a Parabolic Trough Solar Thermal Collecting System

Transcription

1 rd International Conference on Engineering Technology and Application (ICETA 2016) ISBN: Investigation on Performance Prediction and Operation Characteristics of a Parabolic Trough Solar Thermal Collecting System Jingxiao Han SunCan Co., Ltd., Beijing, China Xiangyan Wang & Lingzhi Zhu China Electric Power Research Institute, Nanjing, Jiangsu, China Yanan Zhang*, Feibiao Wang & Zhihao Yao SunCan Co., Ltd., Beijing, China ABSTRACT: Based on the 1MWt parabolic trough concentrating solar power (CSP) pilot plant, the operation characteristics are investigated in three typical weather conditions around the vernal equinox. A simplified performance prediction model is proposed, and the simulation results are compared with the measured data. The results show that the change of the heat transfer oil temperature at the loop inlet will be transferred to the loop outlet after a while. The fluctuation of direct normal irradiance (DNI) has little effect on the outlet temperature. The model is able to accurately predict the performance of the parabolic trough systems at stable flow rate and stable inlet temperature. Installation accuracy of parabolic trough collectors as well as the cleanliness of mirrors has a great influence on the collectors solar-to-thermal efficiency, therefore, high standard of installation and maintenance is necessary. Keywords: concentrating solar power; parabolic trough; performance prediction; operation characteristics 1 PREFACE China s solar energy resources are widely distributed in the northwest region, where the solar radiation intensity is high and suitable for the development of concentrating solar power (CSP). Recently our country is going to substantially reduce the power consumption, and obtain the full implementation of coal-fired power plant ultra-low emission and energy saving, thus CSP as a kind of environmentally friendly and clean power generation technology is drawing extensive attention. As the most mature CSP technology, parabolic trough technology [1-4] has been in successfully commercial operation for many years, moreover the California SEGS project [5] has been persistently running for nearly 30 years, and the electricity sale price is only about 6 cents/kwh. Compared *Corresponding author: zhangyanan@suncancn.com with thermal power plant, parabolic trough power plant (PTPP) has a good application prospect. PTPP consists of a solar field filled with many parabolic trough solar collector assemblies (SCAs) including parabolic reflector mirrors, receiver tubes, metal support structure and tracking systems. Cold heat transfer fluid (HTF, usually heat transfer oil) comes in the receiver tubes which locate in the focal line of the trough, and picks up the heat. Then HTF exits from the solar field at a high temperature, which carries energy to the water in a boiler heat exchanger, and produces high temperature and pressure steam to drive turbine. [6]. It s important to improve the solar-to-thermal efficiency that contributes to reducing of the operation cost and commercialization of PTPP. For a good solar-to-thermal efficiency, the accuracy of installation and control of the tracking system should be enhanced, and a performance prediction model should be built. 1116

2 Regarding tracking control system, Gang Pei [10] investigated the heat exchange efficiency of different tracking modes and found that the single axis tracking collector with axis oriented north-south following sun yields higher annual thermal efficiency. Wang Jin [4] developed a solar tracking controller, and carried out experiments in four typical whether conditions. It is pointed out that tracking speed of different seasons is different, where the speed is the highest on the winter solstice and the lowest on the summer solstice. Regarding operation performance, Xiong Yaxuan, et al. [11] investigated parabolic trough solar thermal performance with Direct Current Heating (DCH) method, and derived the heat loss curve of the collector. Xu Li [13] from the Chinese Academy of Sciences proposed a dynamic testing method for a parabolic trough loop, so that the temperature of HTF can be predicted with the DNI fluctuation. Under the following conditions, HTF temperature rises more than 100 C and the flow rate is stable, an ideal prediction result is achieved using a complex mathematical process. The outlet HTF temperature of solar field is affected by DNI. In order to improve the solar-to-thermal efficiency and to establish a scientific and reasonable control strategy, a fast prediction model is set up. Based on the data measured from SunCan Tianjin 1MWt parabolic trough pilot plant, this paper studies the plant operating characteristics and the outlet HTF temperature predicted. 2 PROJECT INTRODUCTION SunCan Tianjin 1MWt parabolic trough pilot plant is shown in Figure 1, including solar field, stream generator, power block and HTF antifreeze system, etc. As shown in Figure 2, in the solar field there are 4 rows SCAs, namely 1#, 2#, 3# and 4#, respectively. The four SCAs are laid out on linear axis from north to south, with tracking controllers and hydraulic driving systems [14, 15]. The length of each SCA is 100m, and the width of SCA is 5.75m. The inlet and outlet of the loop are installed with instruments to measure the HTF temperature, pressure and flow rate. (a) SunCan 1MW t pilot plant (b) Steam generation Figure 1. SunCan 1MW t parabolic trough plant in Tianjin. Figure 2. Schematic of the parabolic trough pilot plant. The plant installation and commissioning were completed at the end of 2013, realizing automatic electricity generation in May of The plant has been in experimental operation for more than one year, including function verification of automatic operation and continuous power generation. There is a weather station in the plant which is used to collect the relevant meteorological data, such as environmental temperature, relative humidity, wind speed and direction, and sun radiation including global horizontal irradiance, DNI and diffuse horizontal irradiance [16]. If the technology was proved feasible and mature, large scale parabolic trough plants will be constructed in the northwest of China that is rich solar irradiance [17]. 3 PREDICTION MODEL In parabolic trough power plant, it is expected that the HTF temperature at the outlet of loop is steady and equals the designed temperature. The outlet temperature is not only related to the solar radiation intensity, the HTF temperature at the inlet and HTF flow rate, but also the environmental temperature and wind speed and wind direction, etc. The European Union released the dynamic solar collector performance test standard in 2006, which has looser test conditions for evaluating the performance of collectors, compared with the performance test standard in steady state [18]. The Equation (1) is a 1117

3 mathematical equation of dynamic test based on the steady-state test equation: ' mc f ( Tout Tin ) F ( ) enk b( ) GeAa 2 dtm c1( Tm Ta) c2( Tm Ta) c3 dt where m is the mass flow rate of heat transfer fluid (kg/s); cf is the specific heat capacity of heat transfer fluid (J/(kg C)); Tout is the loop outlet temperature ( C); Tin is the loop inlet temperature ( C); Aa is the solar collector area of the loop (m 2 ); F (τα)en is the optical efficiency of collector, which includes specular reflectance of the collector mirror, cleanliness, transmittance of collector tube, absorptivity, surface cleanness and other parameters for parabolic trough power plant; Kθb(θ) is a function of the angle of incidence of DNI; Ge is the solar radiation used in the model which also is DNI, since parabolic trough collectors can only take advantage of DNI instead of the global horizontal irradiance and diffuse irradiance, the formula does not include reflect the global and diffuse irradiance; c1, c2 represent the coefficient of heat dissipation of the collector tubes and connecting pipes; c3 is the effective thermal capacitance. The collector model is also written as: dt mc m f( Tout Tin ) K b( ) GeA a Qloss Aa Qloss,pipeAa Cloop dt (2) (1) In the actual operation, besides the normal operation mode, there are other modes such as start, cloud cover, night standby, emergency stop, and multi-model switching. The performance prediction model will be used to calculate the outlet temperature under different operation modes. In the paper, the testing and calculation will be developed under three typical weather conditions. 4 RESULTS AND DISCUSSION Figure 3 shows the DNI with time under three different durations of time. Figure 3(a) shows the DNI with time in the morning on April 3 rd and the value is steady but sometimes jumping. Figure 3(b) shows the decrease curve of DNI with jumping in the late afternoon. Before midday on March 22 nd, the DNI increases with time. Under the stable HTF flow, the first 15 min experiment data are used to estimate the optical efficiency η with minimal variance, and then the dynamic performance prediction model can be used to predict the subsequent outlet HTF temperature of the loop. where, following the experience equations from Sandia Lab. the heat dissipation of the collector tubes Qloss is written as: Q T (1.296E 9) T (3) 4 loss m m (a) April 3 rd The heat dissipation of the connecting pipes Qloss,pipe is written as: Q loss,pipe 2 ( Tm Ta) 1 Do 2 ln (4) D D i o where Do is the outside diameter of the pipe wrapped insulation material; Di is the inside diameter of the pipe; λ is the conductivity of the insulation material; α is the heat transfer coefficient of pipe surface. During the commissioning period of the loop, the optical efficiency η in the Equation (2) was unknown, but the value was considered as a constant that can be calculated using measured testing data during a continuous period of time. Then the calculated value was used in the Equation (5) to predict the HTF temperature Tout,cal in the subsequent period of time: (b) April 3 rd K b( ) GeAa Q Q loss loss,pipe C loop dt T m out,cal Ti n mc f mc f mc f mc f dt (5) (c) March 22 nd Figure 3. DNI curve during experiment. 1118

4 Figure 4(a) shows the measured temperatures of the inlet and outlet of the loop in the testing from 9:30 to 11:00 in the morning on April 3 rd and the calculated outlet temperature. Based on the first 15 minutes experimental data, the optical efficiency of the loop was calculated, that is Substituted the temperature at the inlet and DNI with time and the optical efficiency into equation (5), the outlet temperatures with time were calculated, lagging behind the time when HTF flow into the solar field. The predicted HTF outlet temperature was generally consistent with the outlet temperature measured. The maximum deviation occurs at 10:30, reaching 4.86 C. After about 3 min, the deviation gradually reduced, and the average deviation was 1.07 C during the experiment period. The installation precision of SCAs may result in the deviation in short time (a) April 3 rd (b) April 3 rd (c) March 22 nd Figure 4. Comparison of the predicted value and the measured value of outlet temperature. In April 3 rd afternoon, due to the sun is sinking in the west, the DNI decreased gradually in Figure 4(b). The optical efficiency of the loop was 0.277, which increased slightly relative to the value in the morning. At 15:34, the HTF inlet temperature dropped obviously because of feeding water to the steam generator which was connected with the inlet of the loop to cool the HTF. During operation, the maximum deviation between the measured value and predicted value of outlet temperature of the loop was 2.86 C occurred at 16:14. The average deviation was 1.09 C during the experiment period. In March 22 nd morning, only 1# and 2# SCA were in operation, while 3# and 4# SCA were in defocusing. At 11:06, the steam generator was fed with water, resulting in the decrease of HTF inlet temperature, shown in the figure 4(c). At 11:12, the maximum deviation between the predicted temperature and the measured temperature was 9.07 C. At 11:19, the loop reached thermal equilibrium state, and the predicted temperature and the measured temperature difference was 2.16 C. The time at the beginning of the HTF outlet temperature reducing is later than the time calculated. The predicted temperature is lower than the measured temperature in most time, due to the thermal inertia of the loop. Since on March 22 nd, only the first two arrays (1#, 2# SCA) the HTF are being heated, but the four SCAs were in a series loop and the HTF also flows through 3# and 4# SCA, while, the thermal inertia of the whole loop has to be taken into consideration in the calculation. The optical efficiency of the three cases are 0.32 (on March 22 nd ), 0.24 (in the morning on April 3 rd ) and 0.26 (in the afternoon on April 3 rd ), respectively. It is inferred that the installation precision and the meteorological factors lead to the difference of the efficiency. The solar collector vibrates intensively with strong wind, affecting the tracking accuracy. On March 22 nd, the wind in the daytime was less than grade 3. On April 1 st, it was rainy. The dirty mirrors decrease the reflectivity of mirrors. On April 3 rd, the wind in the morning was grade 4-5 and decreased to grade 3-4 in the afternoon, so that the DNI fluctuated obviously in the Figure 3. To obtain high solar-to-thermal efficiency, the structure of SCAs should be optimum designed for decreasing the wind load, and the cleanliness of the mirrors should be maintained for a high reflectivity and absorptivity. 5 CONCLUSIONS A performance prediction model is adopted to calculate the optical efficiency and predict the HTF outlet temperature with time. Compared the HTF measured temperature at the outlet of the loop with the predicted value, the following conclusions are drawn: 1) The fluctuation of the DNI effect weakly on the HTF temperature of the loop. 1119

5 2) The performance prediction model can be used to predict the HTF outlet temperature, and the average deviation rates are about 2% when the HTF flow rate and inlet temperature are stable. 3) The wind-resistant performance and the cleanliness of mirrors have a great influence on the solar-to-thermal efficiency. ACKNOWLEDGEMENT This work is supported by the Science and Technology Project of State Grid Corporation of China (NY ). REFERENCES [1] Giostri A., Binotti M., Astolfi M., et al Comparison of different solar plants based on parabolic trough technology. Solar Energy, 86(5): [2] Kearney D., Kelly B., Herrmann U., et al Engineering aspects of a molten salt heat transfer fluid in a trough solar field. Energy, 29(5-6): [3] Yang Y., Han J., Li P., et al Thermal Energy Storage Characteristics of Synthetic Oil and Sand Mixture for Thermocline Single Tank. Proceedings of the CSEE, 35(3): [4] Wang J., Wang J., Feng W., et al Development and application of sun-tracking control system for parabolic trough solar collector. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 31(2): [5] Yan Y., Gao.Z, Xu D., et al Performance investigation of integrated MS-DSG Parabolic solar electric generation system. Journal of Engineering Thermopysics, 34(6): [6] Qu H., Zhao J., Yu X Simulation of parabolic trough solar power generating system for typical Chinese sites. Proceedings of the CSEE, 28(11): [7] Tang Q., Cao W., Zhang.L, et al Analysis on the thermal economy of solar aided coal-fired electricity generation system. Acta Energiae Solaris Sinica, 36 (3): [8] Wang Y., Liu Q., Sui J., et al Research on the solar collector field layout of a solar thermal aided coal-fired power generation system. Journal of Engineering Thermopysics, 35(1): 1-5. [9] Zhou L., Li Y., Hu E., et al Comparison in net solar efficiency between the use of concentrating and non-concentrating solar collectors in solar aided power generation systems. Applied Thermal Engineering, 75: [10] Pei G., Fu H., Ji J., et al Performance comparison of a trough solar concentration system in different tracking modes. Acta Energiae Solaris Sinica, 31(10): [11] Xiong Y., Wu Y., Cui W., et al Performance study of parabolic trough receivers with a novel method. Journal of Engineering Thermopysics, 36(3): [12] Chen H., Wang H., Yu Z., et al Experimental investigation of a natural circulation parabolic trough collector system for medium-high temperature steam generation. Journal of Zhejiang University (Engineering Science), 46(9): [13] Xu L., Wang Z., Li X., et al Dynamic test model for the transient thermal performance of parabolic trough solar collectors. Solar Energy, 95(5): [14] Xu S., Li B., Ma S Research on tracking and control for solar parabolic trough thermal power system based on microcomputer. Computer Measurement & Control. 16(11): [15] Shu J., Rao Z., Zhang X., et al Tracing control for parabolic trough thermal power generation. J. Huazhong Univ. of Sci. & Tech. (Natural Science Edition), 39(2): [16] Perez-Astudillo D., Bachour D DNI, GHI and DHI ground measurements in Doha, Qatar. Energy Procedia, 49: [17] Chen S., Zhang K., Xing X., et al Climatic change of sunshine duration in northwest China during the last 47 years. Journal of Natural Resources, 25(7): [18] Patnode A.M Simulation and performance evaluation of parabolic trough solar power plants, University of Wisconsin-Madison, USA. [19] Burkholder F., Kutscher C Heat Loss Testing of Schott s 2008 PTR70 Parabolic Trough Receiver. Golden, Colorado, USA: National Renewable Energy Laboratories. [20] Wagner M.J., Gilman P Technical Manual for the SAM Physical Trough Model. Golden, Colorado, USA: National Renewable Energy Laboratories. 1120