A dynamic prediction method for thermal performance of parabolic trough solar collector and its testing device

Abstract

The present invention provides a method for dynamically predicting the thermal performance of a parabolic trough solar collector and a testing device thereof. The testing device comprises a mirror reflectance measuring instrument, an inlet temperature sensor, an outlet temperature sensor, a flow meter, an automatic tracking direct heat radiation meter, an ambient temperature sensor and an anemometer. The dynamic prediction method comprises the following steps: firstly, measuring the structure and performance parameters of the collector itself, then performing a dynamic test of the thermal performance of the collector, obtaining the inlet temperature, outlet temperature, volume flow data of the heat transfer medium in the collector and ambient meteorological data, and calculating the solar altitude angle and azimuth angle data, then generating a training sample data set from the obtained data, establishing a neural network model on a computer, and performing training and optimization using the established training sample data set, and finally predicting the outlet temperature of the collector under the specific working conditions to be analyzed, thereby obtaining the thermal performance of the collector.

Description

Technical Field

[0001] The present invention belongs to the technical field of solar thermal utilization, and in particular relates to a method for dynamically predicting the thermal performance of a parabolic trough solar collector and a testing device thereof. Background Art

[0002] Parabolic trough solar collectors are important components for converting light into heat in trough solar thermal power generation systems and low- and medium-temperature solar thermal utilization systems. Their thermal performance affects the heat gain and economic efficiency of the entire system. The thermal performance of trough solar collectors is not only a core indicator for the acceptance of power plant systems and low- and medium-temperature thermal utilization systems, but also an important parameter for the expected benefits of the system. However, currently, no relevant domestic or international standards provide methods for dynamically predicting and evaluating the thermal performance of trough solar collectors in the system. Existing methods are generally based on mathematical and physical prediction models, which predict the thermal performance of the collector in a steady-state or quasi-steady-state manner. However, since mathematical and physical models require many parameters and all relevant parameters cannot be calculated, the constructed equations are relatively complex. In particular, the actual trough solar collector system is difficult to test due to the instability of meteorological parameters such as solar irradiation, solar position, ambient air temperature, wind speed, and wind direction, as well as the non-steady-state effects of the heat transfer medium flow rate and inlet temperature. In particular, the requirement for direct irradiance in the normal direction of the sun to be no less than 700W / m 2 , the heat transfer medium inlet temperature rise rate during the test should not exceed 2.5°C / min, and the heat transfer medium flow rate should not vary by more than ±2.0% of the measured value. Actual field-based trough solar collector systems struggle to meet these technical requirements, resulting in excessively long test wait times or the inability to conduct real-time tests. Therefore, it is highly desirable to develop a method for dynamically predicting the thermal performance of trough solar collectors under actual field-based, non-steady-state conditions. Summary of the Invention

[0003] In response to the problem that parabolic trough solar collectors are difficult to test for thermal performance under actual field conditions in accordance with the technical requirements of existing test standards and test methods, the present invention proposes a method for dynamically predicting the thermal performance of parabolic trough solar collectors and a test device thereof. The method is a method for dynamically predicting the thermal performance of parabolic trough solar collectors and a test device thereof based on a neural network. The present invention only needs to provide several days of actual field operation test data with the required characteristics of the sample data set to predict the outlet temperature of the trough solar collector and its thermal performance under different working conditions, which significantly reduces the technical requirements for thermal performance testing of the collector field and has wider adaptability.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is:

[0005] A method for dynamically predicting the thermal performance of a parabolic trough solar collector comprises the following steps:

[0006] Step 1: measure the structure and performance parameters of the parabolic trough solar thermal collector itself, and obtain at least measurement data including the length, opening width and reflectance ratio of the parabolic trough solar thermal collector.

[0007] Step 2: Perform a dynamic thermal performance test of the parabolic trough solar collector, and obtain the inlet temperature, outlet temperature, volume flow data of the heat transfer medium in the parabolic trough solar collector and the ambient meteorological data during the dynamic thermal performance test; the ambient meteorological data includes at least the direct solar normal irradiance, ambient temperature and wind speed of the working environment.

[0008] Furthermore, the conditions that need to be met for the dynamic test of thermal performance of the parabolic trough solar collector in step 2 include: during the dynamic test of thermal performance, the average wind speed does not exceed 8m / s, the inlet temperature of the heat transfer medium has at least different temperature conditions or a temperature dynamic change process, and the outlet temperature of the heat transfer medium should meet the maximum operating temperature requirement of the heat transfer medium; the volume flow rate of the heat transfer medium is greater than the minimum safe operating volume flow setting value of the parabolic trough solar collector; the total time during the dynamic test of thermal performance should be no less than 3 hours, and the time interval for collecting each data tested should not exceed 1 minute.

[0009] Step 3: Calculate and obtain real-time solar altitude angle and solar azimuth angle data during the thermal performance dynamic test according to the longitude and latitude of the parabolic trough solar collector and the thermal performance dynamic test time.

[0010] Step 4: perform preprocessing and data correlation analysis on the data obtained in steps 2 and 3 to determine the parameter types of the input prediction model and generate a training sample data set. The preprocessing operation includes at least outlier removal and normalization operations.

[0011] Furthermore, the training sample data set described in step 4 should include data from at least two dynamic thermal performance tests, and the data in the training sample set should meet the following conditions: the sample set should at least include the solar normal direct irradiance, ambient temperature, wind speed, solar altitude angle, solar azimuth angle, heat transfer fluid inlet temperature and volume flow parameter types of the parabolic trough solar collector to be evaluated; the variation range of the ambient meteorological data, heat transfer fluid inlet temperature and volume flow data should at least include the ambient meteorological data, inlet temperature value and volume flow value required for evaluation of the parabolic trough solar collector.

[0012] Step 5: Establish a neural network prediction model on a computer, and use the generated training sample data set to train and optimize the neural network model.

[0013] Furthermore, the neural network model described in step 5 is a hybrid neural network model based on a long short-term memory neural network and an artificial neural network model. The neural network model first uses the long short-term memory network model to extract data time features and avoid the data gradient explosion problem, and then uses the artificial neural network model to speed up the calculation;

[0014] Furthermore, in step 5, the neural network model is trained based on an activation function, the activation function is a Tanh function, the loss function is a mean absolute error, the weights are updated using a back propagation algorithm, and the learning rate is optimized using an Adam optimization algorithm.

[0015] Furthermore, when the relative error between the predicted outlet temperature parameter obtained by the neural network model training in step 5 and the outlet temperature parameter of the actual thermal performance dynamic test is ≤5%, the neural network model training is completed.

[0016] Step 6: Based on the trained neural network model obtained in step 5, the outlet temperature of the parabolic trough solar collector is predicted under the specific environmental meteorological data, location, heat transfer medium inlet temperature and volume flow conditions to be evaluated, and the thermal performance of the parabolic trough solar collector is analyzed based on the predicted outlet temperature.

[0017] Furthermore, the thermal performance of the parabolic trough solar collector described in step 6 is based on the inlet temperature T in , volume flow rate V, density ρ, specific heat capacity C p , predicted outlet temperature T out , the solar normal direct irradiance G and the length L and opening width W of the parabolic trough solar collector are calculated according to the following formula:

[0018] (1)Q=V*ρ*C p *(T out -T in )

[0019] (2)

[0020] Wherein, Q is the heat gain of the parabolic trough solar collector, and η is the thermal efficiency of the parabolic trough solar collector.

[0021] A parabolic trough solar collector thermal performance dynamic test device adapted for the parabolic trough solar collector thermal performance dynamic prediction method of the present invention comprises a mirror reflectance meter, an inlet temperature sensor, an outlet temperature sensor, a flow meter, an automatic tracking direct pyrheliometer, an ambient temperature sensor, and an anemometer. The mirror reflectance meter is placed on the reflector of the parabolic trough solar collector during testing; the inlet temperature sensor is installed on a pipe within 2 meters of the parabolic trough solar collector inlet, the outlet temperature sensor is installed on a pipe within 2 meters of the parabolic trough solar collector outlet, the flow meter is installed on the parabolic trough solar collector inlet pipe, and the automatic tracking direct pyrheliometer, ambient temperature sensor, and anemometer are all installed near the parabolic trough solar collector.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] (1) The prediction based on the neural network model does not need to be tested according to the technical requirements of the existing test standards or test methods. It only needs to provide at least 2 days of actual operation data and environmental meteorological data to make a prediction, which solves the problem of the direct irradiance of the sun being lower than 700W / m during the test. 2 , and the problem that the test requirements cannot be met when the inlet temperature and flow rate of the heat transfer fluid change. It significantly expands the available days for testing, reduces the requirements for test technical conditions, and has wide adaptability.

[0024] (2) Compared with the traditional mathematical physics-based prediction model method, it has the advantages of self-learning, nonlinear mapping, fast dynamic response, and fewer required parameters. It has low test data requirements, fast and more accurate prediction speed, and can be adapted to predictions under various working conditions. It has strong adaptability and provides a powerful tool for the prediction of the dynamic thermal performance of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a flow chart of a method for dynamically predicting thermal performance of a parabolic trough solar collector according to the present invention;

[0026] Figure 2 It is a schematic diagram of a parabolic trough solar collector thermal performance testing device of the present invention.

[0027] In the figure, 1 is a mirror reflectance measuring instrument, 2 is an inlet temperature sensor, 3 is an outlet temperature sensor, 4 is a flow meter, 5 is an automatic tracking direct heat meter, 6 is an ambient temperature sensor, 7 is an anemometer, and 8 is a parabolic trough solar collector. DETAILED DESCRIPTION

[0028] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. However, this embodiment does not limit the present invention, and any method changes made by a person skilled in the art based on this embodiment are all within the scope of protection of the present invention.

[0029] like Figure 1 and Figure 2 As shown, a method for dynamically predicting thermal performance of a parabolic trough solar collector according to the present invention comprises the following steps:

[0030] Step 1: Measure the length and opening width of the parabolic trough solar collector 8, and measure the reflectance on the reflector using a mirror reflectance measuring instrument 1, measuring at intervals of at least 12 meters, and taking the average value as the reflectance data of the parabolic trough solar collector 8;

[0031] Step 2, perform a dynamic test of the thermal performance of the parabolic trough solar collector 8. The dynamic test device for thermal performance includes a mirror reflectance meter 1, an inlet temperature sensor 2, an outlet temperature sensor 3, a flow meter 4, an automatic tracking direct heat radiation meter 5, an ambient temperature sensor 6 and an anemometer 7. The inlet temperature sensor 2 is installed on a pipeline within 2m close to the inlet of the parabolic trough solar collector 8, the outlet temperature sensor 3 is installed on a pipeline within 2m close to the outlet of the parabolic trough solar collector 8, the flow meter 4 is installed on the inlet pipeline of the parabolic trough solar collector 8, and the automatic tracking direct heat radiation meter 5, the ambient temperature sensor 6 and the anemometer 7 are all installed near the parabolic trough solar collector 8. During the dynamic thermal performance test, the inlet temperature of the heat transfer medium is required to have different temperature conditions or a dynamic process of temperature increasing and decreasing. The volume flow rate of the heat transfer medium meets the minimum safe operating volume flow rate setting value of the parabolic trough solar collector 8, and the outlet temperature of the heat transfer medium is controlled not to exceed its maximum operating temperature requirement. The average wind speed does not exceed 8m / s. The total time during the dynamic thermal performance test is from 8:00 am to 5:00 pm. The time interval for each data collection is set to 5s. Finally, the inlet temperature, outlet temperature, volume flow rate, solar normal direct irradiance, ambient temperature and wind speed data of the heat transfer medium in the parabolic trough solar collector 8 are obtained.

[0032] Step 3, calculating and obtaining real-time solar altitude angle and solar azimuth angle data during the dynamic thermal performance test according to the longitude and latitude of the location of the parabolic trough solar collector 8 and the time of the dynamic thermal performance test;

[0033] Step 4: Conduct the dynamic thermal performance test of step 2 for 2 days, remove outliers from the data obtained in steps 2 and 3, and then perform normalization preprocessing on the data to unify the dimensions of the input parameters to provide a high-quality data set for training. Perform correlation analysis on the data based on the Pearson correlation coefficient to determine the types of parameters input into the prediction model and generate a training sample data set. The generated training sample data set should include the solar normal direct irradiance, ambient temperature, wind speed, solar altitude angle, solar azimuth angle, heat transfer medium inlet temperature, and volume flow rate parameters of the parabolic trough solar collector 8 to be evaluated.

[0034] At the same time, analyze the data of the training sample dataset to see whether the following conditions are met:

[0035] The range of variation of the ambient meteorological data, heat transfer medium inlet temperature, and volume flow rate data should at least include the ambient meteorological data, inlet temperature values, and volume flow rate values required for evaluation of the parabolic trough solar collector 8. If the training sample dataset meets the above requirements, the training sample dataset step is completed. If not, the number of dynamic thermal performance test days needs to be increased until the above requirements are met.

[0036] Step 5: Establish a neural network prediction model on a computer and train and optimize the neural network model using the generated training sample data set. The neural network model is a hybrid neural network model based on a long short-term memory neural network and an artificial neural network model. The neural network model first uses the long short-term memory network model to extract data time features and avoid the data gradient explosion problem, and then uses the artificial neural network model to accelerate the calculation speed. The neural network includes an input layer, a hidden layer, and an output layer. The input layer extracts the implicit features of the training sample set data and serves as the input of the hidden layer. The hidden layer first uses the long short-term memory network structure to extract the data time features, and then uses the artificial neural network structure to accelerate the calculation speed. The hidden layer analyzes the input data to obtain a nonlinear relationship between the output parameter and the input parameter. Finally, the output layer outputs the predicted value of the parabolic trough solar collector outlet temperature. The neural network model is trained based on an activation function, which is the Tanh function. The loss function is the mean absolute error. The backpropagation algorithm is used to update the weights, and the Adam optimization algorithm is used to optimize the learning rate. The neural network model training is completed when the predicted outlet temperature parameter obtained by the neural network model training meets the outlet temperature parameter of the actual thermal performance dynamic test to meet the relative error of ≤5%.

[0037] Step 6: Based on the trained neural network model obtained in step 5, the outlet temperature of the parabolic trough solar collector 8 is predicted under the specific environmental meteorological data, location, heat transfer medium inlet temperature and volume flow rate conditions to be evaluated, and the inlet temperature T of the heat transfer medium to be evaluated is used as the output temperature. in , volume flow rate V, density ρ, specific heat capacity C p , predicted outlet temperature T out , the solar normal direct irradiance G and the length L and opening width W of the parabolic trough solar collector are calculated according to the following formula:

[0038] (1)Q=V*ρ*C p *(T out -T in )

[0039] (2)

[0040] Wherein, Q is the heat gain of the parabolic trough solar collector, and η is the thermal efficiency of the parabolic trough solar collector.

[0041] It will be easily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for dynamically predicting the thermal performance of a parabolic trough solar collector, characterized in that: By simply providing several days of actual field test data with the required characteristics of the sample data set, the outlet temperature and thermal performance of the trough solar collector under different working conditions can be predicted, including the following steps: Step 1: measuring the structure and performance parameters of the parabolic trough solar collector, obtaining at least measurement data including the length, opening width and reflectance of the parabolic trough solar collector; Step 2: Perform a dynamic thermal performance test of the parabolic trough solar collector, and obtain inlet temperature, outlet temperature, volume flow rate data of the heat transfer medium in the parabolic trough solar collector and ambient meteorological data during the dynamic thermal performance test; the ambient meteorological data at least includes the normal solar irradiance, ambient temperature, and wind speed of the working environment; Step 3: Calculate and obtain real-time solar altitude angle and solar azimuth angle data during the thermal performance dynamic test according to the longitude and latitude of the parabolic trough solar collector and the thermal performance dynamic test time; Step 4: preprocessing the data obtained in steps 2 and 3 and performing data correlation analysis to determine the parameter types for input into the prediction model and generate a training sample data set. The preprocessing operation includes at least outlier removal and normalization. The training sample data set includes data from at least two thermal performance dynamic tests, and the data in the training sample set meets the following conditions: (1) The training sample set at least includes the solar normal direct irradiance, ambient temperature, wind speed, solar altitude angle, solar azimuth angle, heat transfer medium inlet temperature and volume flow rate parameters of the parabolic trough solar collector to be evaluated; (2) The range of variation of the environmental meteorological data, heat transfer medium inlet temperature and volume flow rate data at least includes the environmental meteorological data, inlet temperature value and volume flow rate value required for evaluation of the parabolic trough solar collector; Step 5: Establishing a neural network prediction model on a computer. The neural network model is a hybrid neural network model based on a long short-term memory neural network and an artificial neural network model. The neural network model first uses the long short-term memory neural network model to extract data time features and avoid the data gradient explosion problem, and then uses the artificial neural network model to speed up the calculation speed. The generated training sample data set is used to train and optimize the neural network model. Step 6: Based on the trained neural network model obtained in step 5, the outlet temperature of the parabolic trough solar collector is predicted under the specific environmental meteorological data, location, heat transfer medium inlet temperature and volume flow conditions to be evaluated, and the thermal performance of the parabolic trough solar collector is analyzed based on the predicted outlet temperature.

2. A method for dynamic prediction of thermal performance of a parabolic trough solar collector according to claim 1, characterized in that: In step 2, the conditions satisfied by the dynamic test of thermal performance of the parabolic trough solar collector include: (1) During the dynamic thermal performance test, the average wind speed does not exceed 8m / s; (2) During the dynamic test of thermal performance, the inlet temperature of the heat transfer medium has at least different temperature conditions or a temperature dynamic change process, and the outlet temperature of the heat transfer medium should meet the maximum operating temperature requirement of the heat transfer medium; the volume flow rate of the heat transfer medium is greater than the minimum safe operating volume flow rate setting value of the parabolic trough solar collector; (3) The total duration of the thermal performance dynamic test is not less than 3 hours, and the time interval for collecting the various data tested is not more than 1 minute.

3. The method for dynamic prediction of thermal performance of a parabolic trough solar collector according to claim 1, characterized in that: In step 5, the neural network model is trained based on an activation function, wherein the activation function is a Tanh function, the loss function is a mean absolute error, and the learning rate is optimized using the Adam optimization algorithm.

4. The method for dynamic prediction of thermal performance of a parabolic trough solar collector according to claim 1, characterized in that: In step 5, when the maximum relative error between the predicted outlet temperature parameter obtained by the neural network model training and the outlet temperature parameter of the actual thermal performance dynamic test is ≤±5%, the neural network model training is completed.

5. The method for dynamic prediction of thermal performance of a parabolic trough solar collector according to claim 1, characterized in that: In step 6, the thermal performance of the parabolic trough solar collector is determined based on the inlet temperature T in , volume flow rate V, density ρ, specific heat capacity C p , predicted outlet temperature T out , the solar normal direct irradiance G and the length L and opening width W of the parabolic trough solar collector are calculated according to the following formula: ； ； in, is the heat gain of the parabolic trough solar collector, is the thermal efficiency of the parabolic trough solar collector.

6. A testing device for implementing the method for dynamically predicting thermal performance of a parabolic trough solar collector according to any one of claims 1 to 5, characterized in that: The test device comprises a mirror reflectance meter (1), an inlet temperature sensor (2), an outlet temperature sensor (3), a flow meter (4), an automatic tracking direct solar radiation meter (5), an ambient temperature sensor (6) and an anemometer (7); the mirror reflectance meter (1) performs a reflectance test on a reflector of a parabolic trough solar collector (8) before a dynamic thermal performance test; the inlet temperature sensor (2) is installed on a pipeline within 2 m of an inlet of the parabolic trough solar collector (8); the outlet temperature sensor (3) is installed on a pipeline within 2 m of an outlet of the parabolic trough solar collector (8); the flow meter (4) is installed on the inlet pipeline of the parabolic trough solar collector (8); the automatic tracking direct solar radiation meter (5), the ambient temperature sensor (6) and the anemometer (7) are all installed near the parabolic trough solar collector (8).

Images