Intelligent regulation and control method for PVT system with reflector

The LSTM-Attention algorithm and WOA whale optimization algorithm optimize the inclination angle and reflector angle of the PVT system, which solves the problem of low thermal efficiency of the PVT system under outdoor conditions, and achieves more efficient thermoelectric efficiency and economic benefits.

CN119065403APending Publication Date: 2024-12-03CHONGQING UNIV +2
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411213587.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-30
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The existing PVT systems have low thermal efficiency under outdoor conditions and cannot effectively respond to radiation changes, resulting in low economic returns. The control methods are mostly pure PV systems, which are not suitable for the characteristics of PVT systems.

Method used

The LSTM-Attention algorithm is used to predict future radiation data, and combined with the WOA whale optimization algorithm and the tank circulation system model, the inclination angle and reflector angle of the PVT system are optimized to improve the thermoelectric efficiency.

Benefits of technology

By dynamically adjusting the inclination angle and reflector angle of the PVT system, the thermoelectric efficiency is improved, the operating cost is reduced, and it is suitable for the utilization of clean energy in rural households.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119065403A_ABST
    Figure CN119065403A_ABST
Patent Text Reader

Abstract

The invention discloses an intelligent regulation and control method for a PVT system with a reflector, and the method is characterized in that the method comprises the steps: obtaining the historical horizontal plane radiation data of an area where the PVT system is located, and predicting the hourly total radiation intensity data of a horizontal plane in a few days in the future through an LSTM-Attention algorithm; constructing a multiple regression equation of thermoelectric efficiency based on each control factor of system operation and the electrical efficiency and thermal efficiency of the system; according to the method, the cumulative thermoelectric energy benefit in several days in the future is taken as an optimization target, a whale algorithm and a water tank circulation system model are utilized, an optimal adjustable parameter result is obtained through calculation according to predicted radiation data and a daily first water inlet temperature value, and the PVT system is adjusted and controlled according to the obtained parameters. According to the invention, the operation of the PVT equipment can be better controlled and adjusted according to the current environmental climate change, and the thermoelectric efficiency of the PVT equipment is improved, so that the PVT equipment is particularly suitable for optimized use of a rural household residence PV / T system.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of comprehensive utilization of solar photovoltaic and thermal energy, and in particular to an intelligent control method for a PVT system with a reflector. Background Art

[0002] Renewable energy is an important component of clean and low-carbon building energy systems. Taking advantage of the endless and pollution-free characteristics of solar energy resources, photovoltaic power generation systems are the most widely used and most mature energy technology in China. However, among the radiation from the sun, only light waves in a certain bandwidth can stimulate the photoelectric effect. According to professional estimates, for silicon-based cells, about 54% of the incident radiation is effective, and the corresponding spectral band is about 500-1200 nm, concentrated in the ultraviolet and visible light regions. For more advanced organic cells, the effective spectrum is 400-1800 nm, accounting for only about 77% of the incident radiation. At the same time, unusable radiation will eventually accumulate in the form of heat, causing the battery temperature to rise and the output efficiency to decrease.

[0003] The PVT system is a system that can use solar energy to generate electricity and hot water at the same time. The system generally uses a heat-absorbing plate installed on the back of the solar panel to achieve a cooling effect on the battery to improve the power generation efficiency. The inner cavity of the heat-absorbing plate is connected to a water pipe to achieve heat exchange. However, for many current studies, the average thermal efficiency of different PVT systems under outdoor conditions rarely exceeds 50%, while the thermal efficiency under the same radiation under indoor conditions exceeds 80%. Dynamically changing environmental factors have a great impact on energy output. Therefore, the energy gain of the PVT system under outdoor conditions depends not only on advanced equipment but also has a lot to do with operation regulation.

[0004] Under weaker radiation (about 400 W / m2), the convection and radiation heat transfer between the PVT system and the environment causes the battery temperature to drop significantly. However, the water system with good thermal insulation performance and water with good thermal inertia causes the temperature of the fluid medium to decay slower than that of the battery, so that heat is transferred from the fluid medium to the battery, causing heat loss and a decrease in electrical energy efficiency. Increasing the incident radiation to maintain a higher battery surface temperature is the most convenient and simple method, but excessive radiation will cause a sharp drop in electrical efficiency. However, for bare photovoltaic modules under experimental conditions, about 400-900 W / m 2In the range of , the gradient of electrical efficiency attenuation is very small. For example, when the radiation varies in the range of 500 - 850 W / m2, 400 - 800 W / m2, and 500 - 850 W / m2, the electrical efficiency in different studies varies in the range of 5.9%-5.8%[1], 11.5%-10.9%[2], and 19.5%-18.5%[3]. Therefore, there is a PVT system with a reflector, whose structure includes a PVT plate, which is rotatably mounted on a fixed bracket and can control its rotation angle by a first rotating motor, a water interlayer is arranged on the rear side of the PVT plate, a water inlet end of the water interlayer is connected to a water inlet pipe with a water inlet control valve, and a water outlet end is connected to a water outlet pipe, a reflector is rotatably mounted on the lower end of the PVT plate and can control the rotation angle by a second rotating motor, and a controller is also included, which is respectively connected to the water inlet control valve, the first rotating motor, and the second rotating motor.

[0005] This PVT system is equipped with a reflector with a concentration ratio of less than 2, which increases the incident radiation under weak light conditions while preventing the electrical efficiency from rapidly decreasing under high radiation conditions. It is particularly suitable for rural household applications. However, the high cost of construction and equipment maintenance costs of this existing PVT system means that most systems usually adopt a fixed installation method, which cannot effectively respond to changes in radiation, resulting in low economic benefits of the system. At the same time, for systems equipped with tracking devices, frequent adjustments reduce the service life of the control equipment, resulting in an ineffective performance improvement effect. On the other hand, most current control methods are only applicable to pure PV systems, and there are few patents that design corresponding control optimization methods based on the characteristics of PVT, resulting in an effective performance improvement effect.

[0006] In summary, the applicant considers that if the dual tilt control can be optimized and adjusted better to improve its power exchange and heat exchange efficiency, it can better utilize clean energy, expand the system's thermal power output, and better save costs and improve living standards for rural families. Therefore, how to design an intelligent control method that can be adjusted according to environmental changes, better optimize the system's thermal power efficiency from the operation stage, make it low-cost, widely applicable, and high in energy utilization efficiency, has become a problem that technicians in this field need to consider and solve. Summary of the invention

[0007] In view of the above-mentioned deficiencies of the prior art, the technical problem to be solved by the present invention is: how to provide an intelligent control method for a PVT system with a reflector that can better control and adjust the operation of the equipment according to the current environmental climate changes and improve its thermoelectric efficiency.

[0008] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0009] A method for intelligent control of a PVT system with a reflector is characterized by obtaining historical horizontal plane radiation data of the area where the PVT system is located, using the LSTM-Attention algorithm (Long Short-Term Memory Network Algorithm-Attention Mechanism) to predict the hourly total radiation intensity data of the horizontal plane in the next few days; constructing a multivariate regression equation of PVT thermoelectric efficiency based on various control factors of system operation and system electrical efficiency and thermal efficiency; taking the accumulated thermoelectric energy income in the next few days as the optimization goal, using the WOA whale optimization algorithm and the thermal model of the water tank circulation system, according to the predicted radiation data and the first water inlet temperature value of the day, calculating to obtain the best adjustable parameter results, and adjusting and controlling the PVT system according to the obtained parameters.

[0010] Furthermore, the PVT system with a reflector includes a PVT board, which is rotatably mounted on a fixed bracket and can control its rotation angle by a first rotating motor, a water layer is arranged on the rear side of the PVT board, a water inlet end of the water layer is connected to a water inlet pipe with a water inlet control valve, and a water outlet end is connected to a water outlet pipe, a reflector is rotatably mounted on the lower end of the PVT board and can control its rotation angle by a second rotating motor, and also includes a controller, which is respectively connected to the water inlet control valve, the first rotating motor and the second rotating motor.

[0011] In this way, the rotation angle of the PVT panel can be controlled by the first rotating motor so that it can better receive sunlight. At the same time, the angle of the reflector can be adjusted by the second rotating motor so that the sunlight it emits can better irradiate the PVT panel, maximizing the reception of sunlight and improving the efficiency of light energy absorption and conversion. The scheme of the present invention is based on the above-mentioned PVT system, which can predict the radiation data of the next few days based on historical radiation data and obtain the optimal adjustment parameters and control them according to the WOA whale optimization algorithm, so as to realize the intelligent regulation of the PVT system and better improve its thermoelectric efficiency.

[0012] Furthermore, the method comprises the following steps:

[0013] a. Obtain the historical horizontal surface radiation data of the area where the PVT system is located, take the total horizontal surface radiation data of any hour in the area where the PVT system is located as the output factor, take the hourly total horizontal surface radiation data of x hours before the hour as the corresponding input factor, and build an LSTM-Attention model for the hourly total horizontal surface radiation data of the area where the PVT system is located;

[0014] b. In the constructed LSTM-Attention model, several groups of output factors and input factors formed by data at least 21 days before the current time y days constitute the training set, and several groups of output factors and input factors formed by data within the current time y days constitute the test set. The training and testing of the LSTM-Attention model are completed. After multiple trainings, the model with the best performance in the test set is selected as the LSTM-Attention model used;

[0015] c. Use the obtained LSTM-Attention model to complete the prediction of the total horizontal radiation data of any hour in the area where the PVT system is located in the next z days one by one;

[0016] d. Based on the empirical equation of electrical efficiency of the PVT system, the water tank circulation system model and the empirical equation of thermal efficiency, an adaptive multiple regression model of electrical efficiency, thermal efficiency, dry bulb temperature, water inlet temperature and total radiation intensity of the PVT panel under different water flow rates was constructed;

[0017] e. Obtain temperature data for the next z days according to the forecast information of the local meteorological bureau, and obtain the water inlet temperature data at the first moment of each day in the next z days according to the operation settings of the PVT system, so as to obtain the maximum power generation and heat collection in the next z days as the operation control target of the system. Based on the daily maximum water temperature of the PVT system water tank as the limiting condition, the total horizontal radiation data of any hour in the area where the PVT system is located in the next z days predicted in step c, as well as the conversion relationship between the total horizontal radiation intensity, PVT inclination angle and reflector inclination angle and the total radiation intensity of the PVT board, are brought into the multivariate regression model in step d; the WOA whale optimization algorithm is used to calculate and obtain the PVT inclination angle, reflector inclination angle, circulating water flow rate and daily water pump stop time in the next z days;

[0018] f. Control the operation of the PVT system within the next z days based on the calculation results in step e.

[0019] Furthermore, in step a, the historical horizontal radiation data of the area where the PVT system is located is obtained by field measurement in advance. This can better ensure the authenticity of the data and improve the accuracy of the prediction. During implementation, a solar radiation sensor is used to collect data. The data collection time interval is 10 minutes. The straight-line distance between the solar radiation sensor and the system should be less than 30 meters to ensure the authenticity of the data.

[0020] Furthermore, in step a, x is 10-15 hours, preferably 12 hours.

[0021] This is because if the time is longer than this, the correlation will decrease, and if the time is shorter than this, the predictive power will be weak and a stable model cannot be trained.

[0022] Furthermore, in step b, y is 3 days.

[0023] This is because when the time is greater than 3 days, the accuracy of radiation prediction decreases rapidly and it is difficult for the WOA algorithm to converge when executing the optimization procedure. When it is less than 3 days, it is easy to cause large deviations due to the non-radiation condition, and the fault tolerance and robustness of the model are insufficient.

[0024] Furthermore, z is set to 3 days in the method, which is consistent with the length of the test set. If it is too long, the prediction accuracy will decrease.

[0025] During implementation, the LSTM-Attention model (Long Short-Term Memory Network Algorithm-Attention Mechanism Model) is an existing algorithm model for realizing data time series prediction. The model can solidify the internal connection between historical time series data and corresponding target values ​​in the model by training, so that it can predict future trends. During implementation, in the best performing LSTM-Attention model selected in step b, the Attention (attention mechanism) mechanism is used before the LSTM (Long Short-Term Memory Neural Network Model). The Q, K, and V (query, key, value) matrices of the Attention mechanism are the same, the number of neurons of the LSTM is 96, the optimizer is Adam (adaptive moment estimation optimization), the dropout (random inactivation coefficient) is 0.2, the output layer activation function is 'Tanh' (hyperbolic tangent function), and the maximum number of iterations is 1000.

[0026] Furthermore, the adaptive multivariate regression model constructed in step d includes the empirical equations of electrical efficiency and thermal efficiency under different water flow rates and the water tank water temperature calculation model, and the expression is as follows:

[0027]

[0028] in, is thermal efficiency; is the thermal efficiency; the subscript is the current water flow rate, in units of ; is the inlet water temperature; is the total radiation intensity of the PVT panel; is the dry bulb temperature; are fitting parameters; is the inlet temperature at the current moment; is the inlet temperature at the next moment; is the pipeline heat loss; is the specific heat capacity of the fluid material used; is the density of the fluid material used; is the volume of the water tank;

[0029] In the above expressions, the empirical equations of electrical efficiency and thermal efficiency under different water flow rates are derived based on the PVT one-dimensional heat transfer model; the expression is: thermal efficiency 0.1L / s =b1+ b2* {(water inlet temperature-dry bulb temperature) / total radiation intensity of PVT board}, electrical efficiency 0.1L / s = b3+ b4* total radiation intensity of PVT board + b5* inlet water temperature + b6* dry bulb temperature, b1~b6 are fitting coefficients. Among them, the water tank water temperature calculation model is a one-dimensional node model. The temperature difference is calculated according to the heat collection at the current moment to obtain the next incoming water temperature. The water tank circulation system model itself is an existing technology and can be used to calculate the continuity of the inlet water temperature. The derivation process is not introduced in detail here. The heat loss of the pipeline in the water tank circulation system model should be set according to the specific pipeline material. When there is no reference, it can be estimated to be 15%-30% for a well-insulated system.

[0030] Furthermore, the inlet water temperature data at the first moment of each day in the future z days in step e may be consistent with the inlet water temperature at the first moment of each day during the field measurement of the previous z days, because the tap water temperature fluctuates less within a period of time.

[0031] Among them, the maximum power generation and heat collection in step e are the products of electrical efficiency, thermal efficiency and total radiation intensity of the PVT panel.

[0032] Furthermore, in step e, the daily maximum water temperature of the PVT system water tank is set to 50°C because the hot water produced by this system is used for domestic hot water.

[0033] Furthermore, the conversion relationship between the total radiation intensity of the horizontal plane, the PVT inclination angle and the reflector inclination angle and the total radiation intensity of the PVT panel in step e is derived based on the HDKR model using a trigonometric function model. The specific derivation process is prior art and will not be described in detail here.

[0034] Preferably, in the WOA whale optimization algorithm adopted in step e, the dimension value m of the input variable is 4, which represents the PVT inclination angle, the reflector inclination angle, the circulating water flow rate and the daily pump stop time respectively.

[0035] The WOA whale optimization algorithm itself is a mature existing algorithm. When applied to this method, the corresponding parameters are determined according to the above requirements, and the dimension of the input variable is m; the initial whale population size is 2m, which is set to twice the dimension of the input variable to achieve the highest search efficiency; the maximum number of iterations is 500 to avoid failing to obtain the best search results due to too low an iteration number. In this way, the use of the WOA whale optimization algorithm can adapt to large-scale optimization problems, improve the optimization efficiency through parallel computing and distributed computing, and overcome the problems of large number of calculations and low efficiency when using the traditional grid method for optimization.

[0036] Furthermore, the inclination angle variation range is [0,90] and is an integer multiple of 5, so that the step-by-step regulation is used to reduce the cost of hardware equipment; the water flow rate is the mass flow rate (kg / s), and the water pump stop time is the hour.

[0037] What this means is that the inclination angle will only be 25, 30, 35, not 31, 32, because the higher the accuracy, the more expensive it is, but for our adjustment, the difference between 30 and 31 is very small. Similarly, the water pump will stop at 17:00, not 17:02. We collect data for 10 minutes, and need to control the stop at a larger time span.

[0038] Therefore, the present invention can adjust and control the PVT panels and reflector inclination angles of the PVT system, so that it can be dynamically adjusted according to environmental changes. It not only solves the problem of rapid attenuation of the electrical efficiency of photovoltaic cells under high radiation, but also avoids the reverse heat exchange phenomenon under low radiation, and can expand the thermal power benefits of the system and create better economic benefits.

[0039] Therefore, in summary, the present invention can better realize the control and adjustment of the operation of the PVT equipment according to the current environmental climate changes, improve its thermal power efficiency, and make it particularly suitable for the optimized use of PV / T systems in rural household residential buildings. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a schematic diagram of the method flow of the present invention.

[0041] Figure 2 It is a schematic structural diagram of a PVT system with a reflector targeted by the present invention. DETAILED DESCRIPTION

[0042] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0043] Embodiment: A method for intelligent control of a PVT system with a reflector, characterized in that historical horizontal plane radiation data of the area where the PVT system is located is obtained, and the LSTM-Attention algorithm (Long Short-Term Memory Network Algorithm-Attention Mechanism) is used to predict the hourly total radiation intensity data of the horizontal plane in the next few days; a multivariate regression equation of PVT thermoelectric efficiency is constructed based on various control factors of system operation and the system electrical efficiency and thermal efficiency; with the accumulated thermoelectric energy income in the next few days as the optimization goal, the WOA whale optimization algorithm and the thermal model of the water tank circulation system are used to calculate the best adjustable parameter results according to the predicted radiation data and the first water inlet temperature value of the day, and the PVT system is adjusted and controlled according to the obtained parameters.

[0044] See implementation Figure 2 The PVT system with reflector includes a PVT board 1, which is rotatably mounted on a fixed bracket 2 and can control its rotation angle by a first rotating motor 3. A water interlayer is arranged on the rear side of the PVT board 1, and a water inlet pipe 7 with a water inlet control valve is connected to the water inlet end of the water interlayer, and a water outlet pipe 8 is connected to the water outlet end. A reflector 4 is rotatably mounted on the lower end of the PVT board 1 and can control its rotation angle by a second rotating motor 5. The system also includes a controller 6, which is respectively connected to the water inlet control valve, the first rotating motor and the second rotating motor.

[0045] In this way, the rotation angle of the PVT panel can be controlled by the first rotating motor so that it can better receive sunlight. At the same time, the angle of the reflector can be adjusted by the second rotating motor so that the sunlight it emits can better irradiate the PVT panel, maximizing the reception of sunlight and improving the efficiency of light energy absorption and conversion. The scheme of the present invention is based on the above-mentioned PVT system, which can predict the radiation data of the next few days based on historical radiation data and obtain the optimal adjustment parameters and control them according to the WOA whale optimization algorithm, so as to realize the intelligent regulation of the PVT system and better improve its thermoelectric efficiency.

[0046] Specifically, see Figure 1 , the method comprises the following steps:

[0047] a. Obtain the historical horizontal surface radiation data of the area where the PVT system is located, take the total horizontal surface radiation data of any hour in the area where the PVT system is located as the output factor, take the hourly total horizontal surface radiation data of x hours before the hour as the corresponding input factor, and build an LSTM-Attention model for the hourly total horizontal surface radiation data of the area where the PVT system is located;

[0048] b. In the constructed LSTM-Attention model, several groups of output factors and input factors formed by data at least 21 days before the current time y days constitute the training set, and several groups of output factors and input factors formed by data within the current time y days constitute the test set. The training and testing of the LSTM-Attention model are completed. After multiple trainings, the model with the best performance in the test set is selected as the LSTM-Attention model used;

[0049] c. Use the obtained LSTM-Attention model to complete the prediction of the total horizontal radiation data of any hour in the area where the PVT system is located in the next z days one by one;

[0050] d. Based on the empirical equation of electrical efficiency of the PVT system, the water tank circulation system model and the empirical equation of thermal efficiency, an adaptive multiple regression model of electrical efficiency, thermal efficiency, dry bulb temperature, water inlet temperature and total radiation intensity of the PVT panel under different water flow rates was constructed;

[0051] e. Obtain temperature data for the next z days according to the forecast information of the local meteorological bureau, and obtain the water inlet temperature data at the first moment of each day in the next z days according to the operation settings of the PVT system, so as to obtain the maximum power generation and heat collection in the next z days as the operation control target of the system. Based on the daily maximum water temperature of the PVT system water tank as the limiting condition, the total horizontal radiation data of any hour in the area where the PVT system is located in the next z days predicted in step c, as well as the conversion relationship between the total horizontal radiation intensity, PVT inclination angle and reflector inclination angle and the total radiation intensity of the PVT board, are brought into the multivariate regression model in step d; the WOA whale optimization algorithm is used to calculate and obtain the PVT inclination angle, reflector inclination angle, circulating water flow rate and daily water pump stop time in the next z days;

[0052] f. Control the operation of the PVT system within the next z days based on the calculation results in step e.

[0053] During implementation, historical horizontal radiation data of the area where the PVT system is located is obtained in step a by field measurement in advance. This can better ensure the authenticity of the data and improve the accuracy of the prediction. During implementation, solar radiation sensors are used to collect data. The data collection interval is 10 minutes. The straight-line distance between the solar radiation sensor and the system should be less than 30 meters to ensure the authenticity of the data.

[0054] During implementation, in step a, x is 10-15 hours, preferably 12 hours.

[0055] This is because if the time is longer than this, the correlation will decrease, and if the time is shorter than this, the predictive power will be weak and a stable model cannot be trained.

[0056] During implementation, y in step b is 3 days.

[0057] This is because when the time is greater than 3 days, the accuracy of radiation prediction decreases rapidly and it is difficult for the WOA algorithm to converge when executing the optimization procedure. When it is less than 3 days, it is easy to cause large deviations due to the non-radiation condition, and the fault tolerance and robustness of the model are insufficient.

[0058] When implemented, z in the method is 3 days. This is consistent with the length of the test set. If it is too long, the prediction accuracy will decrease.

[0059] During implementation, the LSTM-Attention model (Long Short-Term Memory Network Algorithm-Attention Mechanism Model) is an existing algorithm model for realizing data time series prediction. The model can solidify the internal connection between historical time series data and corresponding target values ​​in the model by training, so that it can predict future trends. During implementation, in the best performing LSTM-Attention model selected in step b, the Attention (attention mechanism) mechanism is used before the LSTM (Long Short-Term Memory Neural Network Model). The Q, K, and V (query, key, value) matrices of the Attention mechanism are the same, the number of neurons of the LSTM is 96, the optimizer is Adam (adaptive moment estimation optimization), the dropout (random inactivation coefficient) is 0.2, the output layer activation function is 'Tanh' (hyperbolic tangent function), and the maximum number of iterations is 1000.

[0060] During implementation, the adaptive multivariate regression model constructed in step d includes the empirical equations of electrical efficiency and thermal efficiency under different water flow rates and the water tank water temperature calculation model, and the expression is as follows:

[0061]

[0062] in, is the thermal efficiency; is the thermal efficiency; the subscript is the current water flow rate, in units of ; is the inlet water temperature; is the total radiation intensity of the PVT panel; are fitting parameters; is the inlet temperature at the current moment; is the inlet temperature at the next moment; is the pipeline heat loss; is the specific heat capacity of the fluid material used; is the density of the fluid material used; is the volume of the water tank.

[0063] In the above expressions, the empirical equations of electrical efficiency and thermal efficiency under different water flow rates are derived based on the PVT one-dimensional heat transfer model; the expression is: thermal efficiency 0.1L / s =b1+ b2* {(water inlet temperature-dry bulb temperature) / total radiation intensity of PVT board}, electrical efficiency 0.1L / s= b3+ b4* total radiation intensity of PVT board + b5* inlet water temperature + b6* dry bulb temperature, b1~b6 are fitting coefficients. Among them, the water tank water temperature calculation model is a one-dimensional node model. The temperature difference is calculated according to the heat collection at the current moment to obtain the next incoming water temperature. The water tank circulation system model itself is an existing technology and can be used to calculate the continuity of the inlet water temperature. The derivation process is not introduced in detail here. The heat loss of the pipeline in the water tank circulation system model should be set according to the specific pipeline material. When there is no reference, it can be estimated to be 15%-30% for a well-insulated system.

[0064] During implementation, the inlet water temperature data at the first moment of each day for the next z days in step e may be consistent with the inlet water temperature at the first moment of each day during field measurement for the previous z days, because the tap water temperature fluctuates less over a period of time.

[0065] Among them, the maximum power generation and heat collection in step e are the products of electrical efficiency, thermal efficiency and total radiation intensity of the PVT panel.

[0066] During implementation, the maximum daily water temperature of the PVT system water tank in step e is set to 50°C because the hot water produced by this system is used for domestic hot water.

[0067] During implementation, the conversion relationship between the total radiation intensity of the horizontal plane, the PVT inclination angle and the reflector inclination angle in step e and the total radiation intensity of the PVT panel is derived based on the HDKR model using a trigonometric function model. The specific derivation process is prior art and will not be described in detail here.

[0068] During implementation, the HDKR model derived in step e is as follows:.

[0069]

[0070] During implementation, the WOA whale optimization algorithm used in step e has a dimension m of 4, which represents the PVT inclination angle, the reflector inclination angle, the circulating water flow rate and the daily pump stop time.

[0071] The WOA whale optimization algorithm itself is a mature existing algorithm. When applied to this method, the corresponding parameters are determined according to the above requirements, and the dimension of the input variable is m; the initial whale population size is 2m, which is set to twice the dimension of the input variable to achieve the highest search efficiency; the maximum number of iterations is 500 to avoid failing to obtain the best search results due to too low an iteration number. In this way, the use of the WOA whale optimization algorithm can adapt to large-scale optimization problems, improve the optimization efficiency through parallel computing and distributed computing, and overcome the problems of large number of calculations and low efficiency when using the traditional grid method for optimization.

[0072] During implementation, the inclination angle variation range is [0,90] and is an integer multiple of 5, so that the step-by-step regulation is used to reduce the cost of hardware equipment; the water flow rate is the mass flow rate (kg / s), and the water pump stop time is the hour.

[0073] What this means is that the inclination angle will only be 25, 30, 35, not 31, 32, because the higher the accuracy, the more expensive it is, but for our adjustment, the difference between 30 and 31 is very small. Similarly, the water pump will stop at 17:00, not 17:02. We collect data for 10 minutes, and need to control the stop at a larger time span.

[0074] Therefore, the present invention can adjust and control the PVT panels and reflector inclination angles of the PVT system, so that it can be dynamically adjusted according to environmental changes. It not only solves the problem of rapid attenuation of the electrical efficiency of photovoltaic cells under high radiation, but also avoids the reverse heat exchange phenomenon under low radiation, and can expand the thermal power benefits of the system and create better economic benefits.

Claims

1. A smart control method for a PVT system with a reflector. It is characterized in that Obtain the historical horizontal surface radiation data of the area where the PVT system is located, and use the LSTM-Attention algorithm to predict the hourly total radiation intensity data of the horizontal surface in the next few days; The multivariate regression equation of PVT thermal power efficiency is constructed based on the control factors of system operation and the system electrical efficiency and thermal efficiency; Taking the accumulated thermal power energy gain in the next few days as the optimization goal, the WOA whale optimization algorithm and the thermal model of the water tank circulation system are used to calculate the best adjustable parameter results according to the predicted radiation data and the first water inlet temperature value of the day, and the PVT system is adjusted and controlled according to the obtained parameters.

2. The intelligent control method of a PVT system with a reflector as claimed in claim 1, It is characterized in that The PVT system with a reflector includes a PVT board, which is rotatably mounted on a fixed bracket and can control its rotation angle by a first rotating motor. A water interlayer is arranged on the rear side of the PVT board, a water inlet end of the water interlayer is connected to a water inlet pipe with a water inlet control valve, and a water outlet end is connected to a water outlet pipe. A reflector is rotatably mounted on the lower end of the PVT board and can control its rotation angle by a second rotating motor. The system also includes a controller, which is respectively connected to the water inlet control valve, the first rotating motor and the second rotating motor.

3. The intelligent control method of a PVT system with a reflector as claimed in claim 1, It is characterized in that The method comprises the following steps: a. Obtain the historical horizontal surface radiation data of the area where the PVT system is located, take the total horizontal surface radiation data of any hour in the area where the PVT system is located as the output factor, take the hourly total horizontal surface radiation data of x hours before the hour as the corresponding input factor, and build an LSTM-Attention model for the hourly total horizontal surface radiation data of the area where the PVT system is located; b. In the constructed LSTM-Attention model, several groups of output factors and input factors formed by data at least 21 days before the current time y days constitute the training set, and several groups of output factors and input factors formed by data within the current time y days constitute the test set. The training and testing of the LSTM-Attention model are completed. After multiple trainings, the model with the best performance in the test set is selected as the LSTM-Attention model used; c. Use the obtained LSTM-Attention model to complete the prediction of the total horizontal radiation data of any hour in the area where the PVT system is located in the next z days one by one; d. Based on the empirical equation of electrical efficiency of the PVT system, the water tank circulation system model and the empirical equation of thermal efficiency, an adaptive multiple regression model of electrical efficiency, thermal efficiency, dry bulb temperature, water inlet temperature and total radiation intensity of the PVT panel under different water flow rates was constructed; e. Obtain temperature data for the next z days according to the forecast information of the local meteorological bureau, and obtain the water inlet temperature data at the first moment of each day in the next z days according to the operation settings of the PVT system, so as to obtain the maximum power generation and heat collection in the next z days as the operation control target of the system. Based on the daily maximum water temperature of the PVT system water tank as the limiting condition, the total horizontal radiation data of any hour in the area where the PVT system is located in the next z days predicted in step c, as well as the conversion relationship between the total horizontal radiation intensity, PVT inclination angle and reflector inclination angle and the total radiation intensity of the PVT board, are brought into the multivariate regression model in step d; the WOA whale optimization algorithm is used to calculate and obtain the PVT inclination angle, reflector inclination angle, circulating water flow rate and daily water pump stop time in the next z days; f. Control the operation of the PVT system within the next z days based on the calculation results in step e.

4. The intelligent control method of a PVT system with a reflector as claimed in claim 3, It is characterized in that In step a, the historical horizontal radiation data of the area where the PVT system is located is obtained by field measurement in advance.

5. The intelligent control method of a PVT system with a reflector as claimed in claim 3, It is characterized in that In step a, x is 10-15 hours.

6. The intelligent control method of a PVT system with a reflector as claimed in claim 3, It is characterized in that In step b, y is 3 days.

7. The intelligent control method of a PVT system with a reflector as claimed in claim 3, It is characterized in that In the method, z is 3 days.

8. The intelligent control method of a PVT system with a reflector as claimed in claim 3, It is characterized in that The adaptive multivariate regression model constructed in step d includes the empirical equations of electrical efficiency and thermal efficiency under different water flow rates and the water tank water temperature calculation model, and the expression is as follows: Among them, TE is thermal efficiency; EE is thermal efficiency; the subscript is the current water flow rate, in L / s; T in is the inlet water temperature; T is the total radiation intensity of the PVT panel; T a is the dry bulb temperature; b1-b6 are the fitting parameters; T in,i is the inlet temperature at the current moment; T in,i+1 is the inlet temperature at the next moment; η L is the heat loss of the pipeline; C is the specific heat capacity of the fluid material used; ρ is the density of the fluid material used; V is the volume of the water tank.

9. The intelligent control method of a PVT system with a reflector as claimed in claim 3, It is characterized in that The inlet water temperature data at the first moment of each day in the next z days in step e is consistent with the inlet water temperature at the first moment of each day during the field measurement in the previous z days.

10. The intelligent control method of a PVT system with a reflector as claimed in claim 3, It is characterized in that The maximum power generation and heat collection in step e are the product of the electrical efficiency, thermal efficiency and the total radiation intensity of the PVT panel; In step e, the maximum daily water temperature of the PVT system water tank is set to 50°C; In the WOA whale optimization algorithm used in step e, the dimension value m of the input variable is 4, which represents the PVT inclination angle, the reflector inclination angle, the circulating water flow rate and the daily pump stop time in sequence; The inclination angle variation range is [0,90] and is an integer multiple of 5; the water flow rate is the mass flow rate, and the water pump stop time is the hour.

Citation Information

Cited By

  • Building micro-grid multi-element collaborative optimization method and system, electronic equipment and medium

    CN120806299A