Photovoltaic photo-thermal device based on LSTM prediction and intelligent decision

Abstract

The invention relates to a photovoltaic photo-thermal device based on LSTM prediction and intelligent decision, and belongs to the technical field of photovoltaic power generation, and the device comprises a photovoltaic photo-thermal module which is used for achieving the conversion from light energy to electric energy, and collecting heat at the same time; the circulating storage module is used for realizing cooling water circulation, hot water storage and photovoltaic power generation panel cooling; the data acquisition and control module is used for monitoring various state parameters in real time and sending the state parameters to the AI main control module after processing; the control module is used for receiving a control instruction of the AI main control module and controlling the circulating storage module; and the AI main control module is used for receiving the parameters acquired by the data acquisition and control module, performing state prediction by utilizing a long-short-term memory recurrent neural network, making a decision by utilizing a global grid search algorithm, obtaining the variable frequency water pump flow rate with the maximum value, and sending a control instruction to the data acquisition and control module.

Description

Technical Field

[0001] This invention belongs to the field of photovoltaic power generation technology and relates to a photovoltaic thermal device based on LSTM prediction and intelligent decision-making. Background Technology

[0002] With the development of the times, the global demand for renewable energy is increasing, and photovoltaic (PV) power generation is one of the important sources of renewable energy. PV power generation technology refers to the use of photovoltaic panels to convert light energy into electrical energy using the photovoltaic effect. However, PV power generation technology has an inherent physical bottleneck: PV panels absorb a large amount of solar heat during operation, causing the panel temperature to rise. The higher the temperature, the lower the power generation efficiency of the PV panel. Therefore, heat dissipation devices for PV panels have emerged.

[0003] Existing photovoltaic panels have three main problems: First, they have low heat dissipation efficiency. The common method is to attach copper pipes to the back of the photovoltaic panel, which only contact one side of the panel, resulting in a small contact area and low heat dissipation efficiency. Second, they waste water resources. The common method is to spray cooling water onto the back panel, but the cooling water cannot be recycled, leading to water waste. Third, they lack intelligence. Fixed heat dissipation control cannot adapt to dynamic changes in the environment, resulting in a lag in system response and an inability to always operate at the optimal efficiency point. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a photovoltaic thermal device based on LSTM prediction and intelligent decision-making.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A photovoltaic thermal device based on LSTM prediction and intelligent decision-making includes: Photovoltaic thermal modules are used to convert light energy into electrical energy while simultaneously collecting heat. The circulating storage module is used to realize the circulation of cooling water, the storage of hot water, and the cooling of photovoltaic panels; The data acquisition and control module is used to monitor various status parameters in real time, process them, and send them to the AI ​​main control module; it also receives control commands from the AI ​​main control module to control the circular storage module. The AI ​​main control module receives parameters collected by the data acquisition and control module, uses a long short-term memory recurrent neural network for state prediction, uses a global grid search algorithm for decision-making, obtains the most valuable variable frequency pump flow rate, and sends control commands to the data acquisition and control module.

[0006] Furthermore, the photovoltaic thermal module consists of a photovoltaic panel and an aluminum heat dissipation backplate; the photovoltaic panel is used to convert light energy into electrical energy; the aluminum heat dissipation backplate is used to collect the heat absorbed by the photovoltaic panel and cool it down; the aluminum heat dissipation backplate is connected to the bottom of the photovoltaic panel by thermally conductive adhesive; the aluminum heat dissipation backplate includes multiple heat exchange channels.

[0007] Furthermore, the circulating storage module consists of a cold water storage tank, a hot water storage tank, a solenoid valve, a variable frequency water pump, and pipes, including an inlet pipe, an outlet pipe, and pipes for the cold and hot water tanks. One end of the water inlet pipe is connected to a cold water storage tank, and the other end is connected to a heat exchange channel, which is used to transport cold water from the cold water storage tank to the channel. One end of the outlet pipe is connected to the heat exchange channel, and the other end is connected to the hot water storage tank, which is used to transport the water that has absorbed heat in the heat exchange channel to the hot water storage tank. One end of the hot and cold water tank pipe is connected to the cold water storage tank, and the other end is connected to the hot water storage tank; The solenoid valve is installed at the inlet of the cold water storage tank to control the entry of external cold water into the cold water storage tank. The variable frequency water pump is installed on the inlet pipe and is used to pump cold water from the cold water storage tank into the heat exchange channel and push hot water out of the heat exchange channel to the hot water storage tank.

[0008] Furthermore, the data acquisition and control module consists of an I / O controller and multiple sensors; The I / O controller is used to receive data from various sensors and send the processed data to the AI ​​main control module, and receive control commands from the AI ​​main control module to control the variable frequency water pump and solenoid valve; The sensor includes: Outlet water temperature sensor: installed on the outlet water pipe to monitor the hot water temperature in the outlet water pipe in real time; Electronic flow meter for water outlet: installed on the water outlet pipe to monitor the water flow rate in the pipe in real time; Inlet water temperature sensor: installed on the inlet water pipe to monitor the temperature of the cold water in the inlet water pipe in real time; Cold water tank temperature sensor: installed inside the cold water storage tank to monitor the temperature of the cold water in the tank in real time; Hot water tank temperature sensor: installed inside the hot water storage tank to monitor the temperature of the hot water in the tank in real time; Water level gauge: Installed inside the hot water storage tank to monitor the water level in the tank in real time; Electronic water flow meter: installed at the outlet of the hot water storage tank to monitor the flow rate of hot water flowing out of the tank in real time; Anemometer: Used for real-time monitoring of ambient wind speed around the device; Ambient temperature sensor: used to monitor the temperature around the device in real time; Light intensity sensor: used to monitor the solar irradiance received by the photovoltaic thermal module in real time; Back panel temperature sensor: installed between the photovoltaic panel and the aluminum heat dissipation back panel to monitor the temperature of the back of the photovoltaic panel in real time; Furthermore, the AI ​​main control module includes an edge computing device, a human-computer interaction module, and a power supply module. The edge computing device is connected to the I / O controller, receives sensor data transmitted by the I / O controller, and constructs an AI prediction system based on a long short-term memory recurrent neural network to predict the future trend of backplate temperature and outlet water temperature. It also constructs an AI decision-making system based on a global grid search algorithm to calculate the optimal flow rate of the variable frequency water pump and send control commands to the variable frequency water pump and solenoid valve to the I / O controller. The human-computer interaction module is used for status information display and command input. The power supply module is used to power the edge computing device.

[0009] Furthermore, the method of predicting the future trends of backplate temperature and outlet water temperature based on a long short-term memory recurrent neural network specifically includes the following steps: A0: Model training and deployment phase: Train the long short-term memory recurrent neural network model on a high-performance computing workstation and deploy it in an edge computing device; A1: Input Phase: Continuously receives status data sent by the I / O control system, including the current backplane temperature. Current inlet water temperature Current water temperature Current water flow speed Current light intensity Current ambient temperature Current ambient wind speed And synthesize the data into the input feature vector at the current time. This is combined with historical data to form a sequence window that meets the model input requirements; A2: Prediction model solution stage: Load the long short-term memory recurrent neural network model, take the sequence window in A1 as input, and solve the prediction result through forward propagation; A3: Output stage: The predicted results are sent to the AI ​​decision-making system and the human-computer interaction module.

[0010] Furthermore, the calculation of the optimal flow rate of the variable frequency water pump based on the global grid search algorithm specifically includes the following steps: B1: Input Phase: Continuously receives mode information sent by the human-machine interface system and status data sent by the I / O controller, and reads the photovoltaic panel efficiency parameters. Specific heat capacity of water Water pump power consumption curve ; B2: Decision Model Solving Stage: Using a global grid search algorithm, all feasible flow velocities are traversed, the prediction results obtained from the AI ​​prediction system are called, the value function is calculated based on the input information, and the optimal decision is obtained through comparison. B3: Decision Output and Execution Stage: The AI ​​decision system sends the optimal decision to the I / O controller and the human-computer interaction system.

[0011] Furthermore, in step B2, the global grid search algorithm includes the following sub-steps: B21: Decision Space Definition: Define the complete legal operating range of the variable frequency water pump, forming a discretized set of candidate decisions. :

[0012] in, This is the minimum flow rate for the water pump. This is the maximum flow rate of the water pump. This is the search step size; B22: Traversal Prediction Phase: For Each of them This is used as the current set water flow velocity in the AI ​​prediction model. Call the AI ​​prediction model to calculate the current state. Predicted backplate temperature and outlet water temperature at flow rates ; B23: Valuation Phase The AI ​​decision-making system calculates each based on the prediction results. The value, the value function is defined as:

[0013] in, For net profit from power generation, For heating profits; The formula for net profit from power generation is:

[0014] in, Average power generation:

[0015]

[0016] in, Power consumption of the water pump:

[0017] in, As the weight of power generation; The formula for heating profit is:

[0018] in, Average heat production power:

[0019]

[0020] in, The heat generation weight is determined by the mode selected by the user. B24: Determining the Optimal Decision: AI Decision-Making Systems from All The result is the one that maximizes the total value. The corresponding flow rate :

[0021] This flow rate is the optimal flow rate determined by the decision.

[0022] Furthermore, based on the device status, the AI ​​main control module controls the opening and closing of the solenoid valve to switch the cooling water circulation between mode one and mode two, as detailed below: Mode 1: When the hot water storage tank is not full, the solenoid valve of the cold water storage tank inlet pipe opens, and cooling water enters the cold water storage tank from the outside. It is then drawn into the heat dissipation channel by the variable frequency water pump, absorbs heat, becomes hot water, and enters the hot water storage tank. In this mode, the solenoid valve between the cold and hot water storage tanks is closed, and there is no water flow between the cold and hot water storage tanks. When the hot water storage tank is full, the water level gauge is triggered, and then the mode switches to Mode 2. Mode 2: When the hot water storage tank is full, the solenoid valve on the inlet pipe of the cold water storage tank is closed, and the solenoid valve between the hot and cold water storage tanks is opened. Hot water flows out of the hot water storage tank, flows into the cold water storage tank through the hot and cold water pipes, dissipates heat during this process, and becomes cold water again. The cold water is drawn into the heat dissipation channel by the variable frequency water pump, absorbs heat, becomes hot water again, and re-enters the hot water storage tank. In this mode, cooling water circulates throughout the device, ensuring that the hot water storage tank is always full. When the user uses hot water, the hot water storage tank is no longer full, and the system returns to Mode 1.

[0023] Furthermore, the photovoltaic thermal module has a cascading function. One end of the water inlet pipe and the water outlet pipe is equipped with a pipe cap. By removing the pipe cap and connecting it to the water inlet pipe and the water outlet pipe of another photovoltaic power generation module, a set of AI main control module, loop storage module, and data acquisition and control module can control multiple photovoltaic power generation modules.

[0024] The beneficial effects of this invention are that this device can both efficiently dissipate heat to improve power generation efficiency and recover and store heat energy in hot water for utilization, thus having high practicality and promotional value.

[0025] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0026] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein: Figure 1 A block diagram of a photovoltaic thermal device based on LSTM prediction and intelligent decision-making; Figure 2 This is a structural diagram of a photovoltaic thermal device based on LSTM prediction and intelligent decision-making. Figure 3 A top view of a photovoltaic thermal module; Figure 4 This is a software architecture diagram.

[0027] Reference numerals: 1-Photovoltaic power generation panel, 2-Aluminum heat dissipation backplate, 3-Outlet pipe, 4-Inlet pipe, 5-Bracket, 6-Outlet water temperature sensor, 7-Outlet electronic flow meter, 8-Inlet water temperature sensor, 9-Variable frequency water pump, 10-Cold water storage tank, 11-Cold water tank temperature sensor, 12-Solenoid valve, 13-Cold water tank inlet pipe, 14-Hot water storage tank, 15-Hot water tank temperature sensor, 16-Water level gauge, 17-Water electronic flow meter, 18-Hot water tank outlet pipe, 19-Cold and hot water tank pipes, 20-Solenoid valve, 21-Edge computing device, 22-Anemometer, 23-Ambient temperature sensor, 24-Light intensity sensor, 25-Ground, 26-Thermal conductive adhesive, 27-Flow channel, 28-Backplate temperature sensor, 29-I / O controller, 30-Pipe cap. Detailed Implementation

[0028] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0029] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0030] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0031] Example 1: like Figure 1 As shown, the present invention provides a photovoltaic thermal device based on LSTM prediction and intelligent decision-making, which consists of four core modules: an AI main control module, a photovoltaic thermal module, a data acquisition and control module, and a circulating storage module.

[0032] The photovoltaic (PV) thermal module consists of a photovoltaic panel 1 and an aluminum heat dissipation backplate 2. Its function is to convert solar energy into electrical energy while simultaneously collecting heat. The photovoltaic panel 1 is used to convert solar energy into electrical energy. It is composed of tempered glass and photovoltaic cells. Under sunlight, the "photovoltaic effect" excites the directional movement of electrons to generate current, which is ultimately converted into electrical energy and transmitted to the power module and the power grid. The aluminum heat dissipation backplate 2 is used to cool the photovoltaic panel 1 and collect heat. The aluminum heat dissipation backplate 2 is made of highly thermally conductive aluminum alloy and is formed into multiple trapezoidal structures through a profile extrusion process, with built-in heat exchange channels 27. The aluminum heat dissipation backplate is tightly bonded to the photovoltaic panel 1 by thermally conductive adhesive 26. The lower end of the heat exchange channel 27 is connected to the water inlet pipe 4, and the upper end is connected to the water outlet pipe 3. The aluminum heat dissipation backplate 2 is a structural support component of the photovoltaic module. The aluminum heat dissipation backplate 2 efficiently absorbs and dissipates the heat energy absorbed by the photovoltaic panel 1, which is then absorbed by the cooling water in the heat exchange channel 27. The trapezoidal structure of the aluminum heat dissipation backplate 2 reduces the amount of aluminum used while ensuring extremely high structural strength, thereby reducing the overall weight of the device. The trapezoidal structure of the aluminum heat dissipation backplate 2 significantly increases the contact area between the backplate and the outside air, facilitating air circulation and dissipating heat that is not completely absorbed by the cooling water. Specifically, the thermally conductive adhesive 26 is used to bond the photovoltaic panel and the aluminum heat dissipation backplate, conduct heat, and provide electrical insulation. The thermally conductive adhesive is a modified EVA film, applied between the photovoltaic panel 1 and the aluminum heat sink backplate 2. The adhesive has extremely low thermal resistance, conducting the heat absorbed by the photovoltaic panel 1 to the aluminum heat sink backplate 2. It is also a high-strength dielectric, electrically isolating the high-voltage circuit of the photovoltaic panel 1 from the aluminum heat sink backplate 2. The photovoltaic thermal module 1 is fixed and supported by a bracket 5, which is made of high-strength metal profiles and has its base fixed to the ground at a 25° angle. The bracket fixes the photovoltaic thermal module at the optimal tilt angle facing the sun.

[0033] The circulating storage module consists of a cold water storage tank 10, a hot water storage tank 14, a solenoid valve 12, a variable frequency water pump 9, and pipelines. It is responsible for the cooling water circulation, hot water storage, and cooling of the photovoltaic panels for the entire device. The cold water storage tank 10 stores the cooling water that has not absorbed heat; the hot water storage tank 14 stores the hot water that has absorbed heat; the solenoid valve 12 controls the flow and closure of its corresponding pipeline; the variable frequency water pump 9 controls the flow and closure of its corresponding pipeline and the flow rate of the cooling water; and the pipelines are responsible for the flow of water.

[0034] Specifically, the cold water storage tank 10 is used to store unheated cold water. The cold water storage tank is an insulated container installed on the ground 25. Its inlet is connected to the cold water tank inlet pipe 13, and its outlet is connected to the inlet pipe 4. It is also connected to the hot water storage tank 14 via a hot and cold water tank pipe 19. The hot water storage tank 14 is used to store heated hot water. The hot water storage tank is also an insulated container installed on the ground 25. Its inlet is connected to the outlet pipe 3, and its outlet is connected to the hot water tank outlet pipe 18. It is also connected to the cold water storage tank 10 via the hot and cold water tank pipe 19.

[0035] The outlet pipe 3 is used to transport the water that has absorbed heat in the flow channel to the hot water storage tank. The outlet pipe is made of high-temperature resistant and corrosion-resistant pipe material and is wrapped with an external insulation layer. The outlet pipe is connected to the upper end of the flow channel 27, one end is connected to the hot water storage tank 14, and the other end is equipped with a pipe cap 30. The outlet pipe is equipped with an outlet water temperature sensor 6 and an outlet water electronic flow meter 7. The outlet pipe transports the heat-absorbing hot water to the hot water storage tank 14 for storage and keeps it warm during the transportation process. After removing the pipe cap 30, the outlet pipe can be cascaded with the outlet pipes of other photovoltaic thermal modules.

[0036] The inlet pipe 4 is used to transport cold water from the cold water storage tank 10 to the flow channel 27. The inlet pipe 4 is made of corrosion-resistant pipe material; the inlet pipe 4 is connected to the lower end of the flow channel 27, one end is connected to the cold water storage tank 10, and the other end is equipped with a pipe cap 30; a variable frequency water pump 9 is installed on the inlet pipe 4; the inlet pipe transports the cold water from the cold water storage tank 10 to the flow channel 27 for heat dissipation; after removing the pipe cap 30, the inlet pipe can be cascaded with the inlet pipes of other photovoltaic thermal modules.

[0037] The cold water tank inlet pipe 13 is used to transport external cold water to the cold water storage tank. The cold water tank inlet pipe is made of corrosion-resistant pipe material; one end of the cold water tank inlet pipe is connected to the external water line, and the other end is connected to the cold water storage tank 10; a solenoid valve 12 is installed on the cold water tank inlet pipe.

[0038] The hot water tank outlet pipe 18 is used to transport hot water from the hot water storage tank to the user end. The hot water tank outlet pipe is made of high temperature resistant and corrosion resistant pipe material; one end of the hot water tank outlet pipe is connected to the hot water storage tank 14, and the other end is connected to the user end water line; an electronic flow meter 17 is installed on the hot water tank outlet pipe.

[0039] The hot and cold water tank pipe 19 is used to transport hot water from the hot water storage tank to the cold water storage tank, and dissipates heat in the process. The hot and cold water tank pipe is made of high-temperature resistant and corrosion-resistant tubing, and multiple pipes are arranged in parallel. One end of the hot and cold water tank pipe is connected to the hot water storage tank 14, and the other end is connected to the cold water storage tank 10. A solenoid valve 20 is installed on the hot and cold water tank pipe. The hot and cold water tank pipe is used to transport hot water from the hot water storage tank 14 to the cold water storage tank, and utilizes the parallel structure of the multiple pipes to dissipate heat, causing the water entering the cold water storage tank 10 to become cold water again.

[0040] The variable frequency water pump 9 is used to circulate the cooling water. The variable frequency water pump 9 is installed on the water inlet pipe 4; the variable frequency water pump 9 draws cold water from the cold water storage tank 10 and delivers it to the aluminum heat dissipation back plate 2, and pushes hot water out of the aluminum heat dissipation back plate 2 and delivers it to the hot water storage tank 14; the variable frequency water pump is controlled by the I / O controller 29, which controls the on / off state and flow rate of the water pump by changing its frequency.

[0041] Solenoid valve 12 is used to control the entry of external cold water into the cold water storage tank. The solenoid valve is installed on the cold water tank inlet pipe 13; the solenoid valve is controlled by I / O controller 29, which controls the entry of external cold water into the cold water storage tank 10 by controlling the opening and closing of the solenoid valve.

[0042] Solenoid valve 20 is used to control the flow of hot water from hot water storage tank 14 into cold water storage tank 10. The solenoid valve is installed on the hot and cold water tank pipe 19; the solenoid valve is controlled by I / O controller 29, which controls the flow of hot water from hot water storage tank 14 into cold water storage tank 10 by controlling the opening and closing of the solenoid valve.

[0043] The data acquisition and control module consists of an I / O controller and all sensors: a temperature sensor, an electronic flow meter, an anemometer, an electronic water level gauge, and a light intensity sensor. Its function is to monitor all status information in real time, process the data, and send it to the AI ​​main control module. It also receives decisions from the AI ​​main control module to control the variable frequency water pump and solenoid valves. The I / O controller is responsible for receiving and processing data, sending it to the AI ​​main control module, and receiving control commands from the AI ​​main control module to control the variable frequency water pump and solenoid valves. The temperature sensor monitors temperature data; the electronic flow meter monitors flow rate data; the anemometer monitors wind speed data; the electronic water level gauge monitors water level data; and the light intensity sensor monitors light intensity data. Specifically, it includes: The I / O controller 29 is used to receive sensor signals and control the variable frequency water pump and solenoid valve. The I / O controller is an industrial-grade MCU, which is connected to the edge computing device 21 via a high-speed communication bus, and is connected to all sensors, solenoid valves, and variable frequency water pumps. The I / O controller receives all sensor data in real time and sends the processed data to the edge computing device 21. The I / O controller also receives control commands from the edge computing device in real time and controls the behavior of the solenoid valves and variable frequency water pumps according to the commands.

[0044] The outlet water temperature sensor 6 is used to monitor the temperature of the hot water leaving the heat sink backplate in real time. The outlet water temperature sensor is a temperature probe installed in the outlet water pipe 3; the outlet water temperature sensor monitors the temperature of the hot water in the outlet water pipe 3 in real time and sends the data to the I / O controller 29.

[0045] The electronic flow meter 7 is used to monitor the real-time flow rate of cooling water flowing through the pipeline. The electronic flow meter is a turbine flow meter and is installed on the outlet pipeline 3; the electronic flow meter monitors the real-time flow rate of the cooling water circulation and sends the data to the I / O controller 29.

[0046] The inlet water temperature sensor 8 is used to monitor the temperature of the cold water entering the heat sink backplate in real time. The inlet water temperature sensor is a temperature probe installed in the inlet water pipe 4; the inlet water temperature sensor monitors the temperature of the cold water in the inlet water pipe 4 in real time and sends the data to the I / O controller 29.

[0047] The cold water tank temperature sensor 11 is used to monitor the cold water temperature in the cold water storage tank in real time. The cold water tank temperature sensor is a temperature probe installed at the bottom of the cold water storage tank 10; the cold water tank temperature sensor monitors the cold water temperature in the cold water storage tank 10 in real time and sends the data to the I / O controller 29.

[0048] The hot water tank temperature sensor 15 is used to monitor the hot water temperature in the hot water storage tank in real time. The hot water tank temperature sensor is a temperature probe installed at the bottom of the hot water storage tank 14; the hot water tank temperature sensor monitors the hot water temperature in the hot water storage tank 14 in real time and sends the data to the I / O controller 29.

[0049] The water level gauge 16 is used to monitor the water level in the hot water storage tank in real time. The water level gauge is a capacitive water level gauge and is installed on the top of the hot water storage tank 14; the water level gauge monitors the hot water level in the hot water storage tank 14 in real time and sends the data to the I / O controller 29.

[0050] The water electronic flow meter 17 is used to monitor the real-time flow rate of hot water flowing through the hot water tank outlet pipe. The water electronic flow meter is a turbine flow meter, installed on the hot water tank outlet pipe 18; the outlet electronic flow meter monitors the real-time flow rate of hot water in the hot water tank outlet pipe, thereby monitoring the hot water usage on the user side, and sends the data to the I / O controller 29.

[0051] An anemometer 22 is used to monitor the ambient wind speed around the device in real time. The anemometer is an ultrasonic anemometer, installed on the ground 25 or above the edge computing device 21 chassis; the anemometer monitors the ambient wind speed around the device in real time and sends the data to the I / O controller 29.

[0052] An ambient temperature sensor 23 is used to monitor the temperature around the device in real time. The ambient temperature sensor is a digital temperature sensor, installed on the ground 25 or above the chassis of the edge computing device 21; the ambient temperature sensor monitors the ambient temperature around the device in real time and sends the data to the I / O controller 29.

[0053] The light intensity sensor 24 is used to monitor the solar irradiance received by the photovoltaic thermal module in real time. The light intensity sensor is a radiometer, which is installed on the ground 25 or above the chassis of the edge computing device 21, and has the same tilt angle and orientation as the photovoltaic panel 1; the light intensity sensor monitors the total solar irradiance in real time and sends the data to the I / O controller 29.

[0054] The AI ​​main control module consists of an edge computing device, a screen, and a power module, and its function is to control the entire device. The edge computing device is responsible for receiving data, running AI prediction and decision-making model algorithms, and outputting control commands to control the entire device; the power module is responsible for supplying power to the entire device; and the screen is responsible for displaying the device status and facilitating human-computer interaction.

[0055] In this embodiment, the edge computing device is used to control the entire device. The edge computing device is an NVIDIA Jetson Nano, connected to a power module and a touchscreen display, and connected to the I / O controller 29 via a high-speed communication bus. The edge computing device receives data from the I / O controller 29 and sends control commands to it. The edge computing device connects to the Internet via Wi-Fi. The edge computing device runs AI algorithms. The power module supplies power to the edge computing device. The touchscreen display is used for status information display and command input.

[0056] Edge computing devices are equipped with AI prediction and decision-making models. The AI ​​prediction system utilizes a Long Short-Term Memory (LSTM) recurrent neural network algorithm to predict the trends in backplate temperature and outlet water temperature over the next 15 minutes based on the current state, including the following steps: A0: Model Training and Deployment Phase: This phase takes place on a high-performance computing workstation, where the long short-term memory recurrent neural network model is trained and deployed on edge computing devices.

[0057] A1: Input Phase: The AI ​​prediction system continuously receives status data from the I / O control system: current backplane temperature Current inlet water temperature Current water temperature Current water flow speed Current light intensity Current ambient temperature Current ambient wind speed And synthesize the data into the input feature vector at the current time. This is combined with historical data to form a sequence window that meets the model input requirements.

[0058] A2: Prediction Model Solving Stage: The AI ​​prediction system loads the Long Short-Term Memory Recurrent Neural Network model, takes the sequence window in A1 as input, and solves the prediction result through forward propagation.

[0059] A3: Output stage: The AI ​​prediction system sends the predicted results to the AI ​​decision-making system and the human-computer interaction system.

[0060] Step A0 only needs to be executed once beforehand, while steps A1, A2, and A3 are executed sequentially each time the AI ​​prediction system is invoked.

[0061] Step A0 includes the following sub-steps: A01: Data Acquisition and Preparation: Collect a massive historical database from the long-term operation of this device.

[0062] A02: Model Building and Training: Construct a Multiple-Input Multiple-Output (MIMO) Long Short-Term Memory Recurrent Neural Network model. This model is configured to receive the multi-dimensional input sequence from A1 and output two prediction vectors from A3. Supervised learning training is performed using the historical database from A01. The training objective is to minimize a root mean square error loss function, whose formula is:

[0063] in, The total number of samples, For sample index.

[0064] A03: The trained model is optimized using NVIDIA TensorRT tools to make it smaller and faster, making it suitable for edge computing environments. The model file is then deployed to the AI ​​prediction system on edge computing devices.

[0065] In step A2, the core of the Long Short-Term Memory Recurrent Neural Network model is the "Cell State". The forward propagation solution of the structure consists of three gates and includes the following sub-steps: A21: Forget Gate This decision requires consideration of the "long-term memory" (cellular state) from the previous moment. The formula for how much information is forgotten in a cell is:

[0066] in, This is the short-term output from the previous moment. Input the feature vector at the current time. It is the Sigmoid activation function. and The weights and biases of the forget gate.

[0067] A22: Input Gate This gate determines the new information at the current moment. The formula for determining how much of it is important and needs to be added to "long-term memory" is as follows:

[0068]

[0069] in, This is a candidate state (candidate memory). These are the weights and biases of the input gate and the candidate state, respectively. It is the hyperbolic tangent activation function.

[0070] A23: Cell State Update: This step combines the decisions of the "forget gate" and the "input gate" to generate the "long-term memory" (cell state) at the current moment. Its formula is:

[0071] in, Represents the Hadamard product.

[0072] A24: Output Gate This gate decides to start from the updated "long-term memory" (cellular state): In the given information, which information should be extracted as the short-term output for the current moment? The formula is:

[0073]

[0074] Among them, and The weights and biases of the output gate.

[0075] A25: Output Layer Decoupling: Hidden State ( The sequence is fed into two independent fully connected output layers, decoupling two different prediction targets, as shown in the formula:

[0076]

[0077] in, Predict the backplate temperature and outlet water temperature for each time point within the next 15 minutes.

[0078] The AI ​​decision-making system uses a global grid search algorithm to calculate the optimal flow rate of the variable frequency water pump, including the following steps: B1: Input Phase: The AI ​​decision-making system continuously receives pattern information from the human-computer interaction system and status data from the I / O control system, and reads physical constants from the system configuration: photovoltaic panel efficiency parameters. Specific heat capacity of water Water pump power consumption curve .

[0079] B2: Decision Model Solving Stage: The AI ​​decision system uses a global grid search algorithm to traverse all feasible flow rates, calls the AI ​​prediction system to obtain prediction results, calculates the value function based on the input information, and compares them to obtain the optimal decision.

[0080] B3: Decision Output and Execution Stage: The AI ​​decision-making system sends the optimal decision to the I / O control system and the human-computer interaction system.

[0081] B1, B2, and B3 are executed cyclically in the edge computing device, with an execution cycle of 1 minute.

[0082] In step B2, the global grid search algorithm includes the following sub-steps: B21: Decision Space Definition: The AI ​​decision-making system defines the complete legal operating range of the variable frequency water pump, forming a discretized "candidate decision set". :

[0083] in, This is the minimum flow rate for the water pump. This is the maximum flow rate of the water pump. This is the search step size.

[0084] B22: Traversal Prediction Phase: For Each of them This is used as the current set water flow velocity in the AI ​​prediction model. Call the AI ​​prediction model to calculate the current state. Predicted backplate temperature and outlet water temperature at flow rates .

[0085] B23: Valuation Phase The AI ​​decision-making system calculates each based on the prediction results. The value, the value function is defined as:

[0086] in, For net profit from power generation, For heating profit.

[0087] The formula for net profit from power generation is:

[0088] in, Average power generation:

[0089] in, Power consumption of the water pump:

[0090] in, This is the weighting for power generation.

[0091] The formula for heating profit is:

[0092] in, Average heat production power:

[0093]

[0094] in, The heat generation weight is determined by the mode selected by the user.

[0095] B24: Determining the Optimal Decision: AI Decision-Making Systems from All The result is the one that maximizes the total value. The corresponding flow rate :

[0096] The flow rate is the optimal flow rate determined by the decision.

[0097] The photovoltaic modules of this device abandon the traditional frame or plastic backplate and adopt an aluminum heat dissipation backplate. The backplate is tightly attached to the photovoltaic panel with thermally conductive adhesive, which not only provides structural support, but also efficiently conducts and dissipates heat. The backplate is integrally formed by the "profile extrusion" process and has built-in heat exchange channels, which has higher heat conduction efficiency than traditional adhesive pipes. The trapezoidal structure of the backplate reduces the weight of the device while ensuring strength, increases the air contact area, and facilitates air circulation.

[0098] This device uses two water storage tanks, one for cold water and one for hot water, with parallel pipes installed between the tanks. It has two cooling water circulation modes, which can ensure that the hot water tank is quickly replenished with hot water and that cold water is continuously supplied for heat dissipation.

[0099] The photovoltaic and solar thermal modules of this device have a cascading function, which can enable one AI main control module, a loop storage module, and a data acquisition and control module to control multiple photovoltaic power generation modules, thereby improving power generation efficiency and heating efficiency.

[0100] Example 2: This embodiment provides the working principle of the photovoltaic thermal device based on LSTM prediction and intelligent decision-making described in Embodiment 1. The photovoltaic panel generates electricity using light energy and absorbs heat during power generation. This heat is efficiently conducted to the internal heat dissipation channel by the aluminum heat dissipation backplate and absorbed by the cooling water. The edge computing device receives data sent by the I / O controller in real time, uses the Long Short-Term Memory Recurrent Neural Network (LSTM) algorithm for state prediction, and uses a global grid search algorithm for decision-making to obtain the most valuable variable frequency water pump flow rate. The hot water generated by heat dissipation is stored in a hot water storage tank for user use. The cooling water circulation switches between mode one (external water supply) and mode two (internal circulation) according to the device status. The photovoltaic thermal modules of the device can be cascaded. Various status data of the device are displayed in real time through a touch screen and uploaded to the cloud via WIFI connection to the Internet for users to view in real time on their mobile devices.

[0101] The working process of this device includes the following steps: When the power module in the AI ​​main control module is turned on, the entire device starts up.

[0102] In edge computing devices, the main control system is initialized, including the initialization of the human-computer interaction system, the initialization of the AI ​​prediction system (loaded with the long short-term memory recurrent neural network model trained and deployed in the A0 step beforehand), and the initialization of the AI ​​decision-making system; in the I / O controller, the I / O control system is initialized, including the initialization of the device control system and the initialization of the sensing data system.

[0103] When sunlight shines on a photovoltaic panel, the panel begins to generate electricity. Part of the electrical energy is stored in the power module of the device for the device's use, while the other part is stored in the power grid. During the power generation process, the photovoltaic panel absorbs solar heat, and its temperature gradually rises, reducing its power generation efficiency and requiring cooling.

[0104] Photovoltaic panels abandon traditional aluminum frames or plastic back panels, instead using aluminum heat dissipation back panels, which provide both support and heat dissipation. This back panel can efficiently conduct heat from the photovoltaic panel. The aluminum heat dissipation back panel has heat dissipation channels, where cooling water enters from the bottom and flows out from the top. During the flow, it absorbs the heat conducted by the back panel, raising its own temperature while lowering the temperature of the photovoltaic panel, thus improving power generation efficiency. The cooling water is heated back into hot water and stored in a hot water storage tank.

[0105] There are two modes of cooling water circulation: Mode 1: When the hot water storage tank is not full: The solenoid valve of the inlet pipe of the cold water storage tank opens, and cooling water enters the cold water storage tank from the outside. It is then sucked into the heat dissipation channel by the variable frequency water pump, absorbs heat and becomes hot water, which then enters the hot water storage tank. In this mode, the solenoid valve between the cold and hot water storage tanks is closed, and there is no water flow between the cold and hot water storage tanks. When the hot water storage tank is full, the water level gauge is triggered, and then the mode switches to mode 2.

[0106] Mode 2: When the hot water storage tank is full: The solenoid valve on the inlet pipe of the cold water storage tank is closed, while the solenoid valve between the hot and cold water storage tanks is open. Hot water flows out of the hot water storage tank and into the cold water storage tank through the parallel-arranged pipes. During this process, it dissipates heat and becomes cold water again. The cold water is drawn into the heat dissipation channel by the variable frequency water pump, absorbs heat, becomes hot water again, and re-enters the hot water storage tank. In this mode, cooling water circulates throughout the entire device, ensuring that the hot water storage tank is always full. When the user uses hot water, the hot water storage tank is no longer full, and the system returns to Mode 1.

[0107] During the cooling cycle, the switching between the two modes, i.e. the opening and closing of the solenoid valve, is controlled by the main control system.

[0108] The AI ​​algorithm controls the flow rate of the variable frequency water pump, thereby controlling the cooling water circulation speed, which in turn changes the power generation efficiency and hot water temperature to maximize the weighted benefit of the entire device. The AI ​​decision-making system operates in 1-minute cycles. At the beginning of each cycle, the AI ​​decision-making system makes a decision and calls the AI ​​prediction system during the decision-making process to find the optimal flow rate of the variable frequency water pump. The variable frequency water pump runs at this flow rate for 1 minute. At the beginning of the next cycle, the AI ​​decision-making system and the AI ​​prediction system discard all previous predictions and calculations and start prediction and decision-making again. The weighted benefit is set according to user needs. Under normal circumstances, the device prioritizes power generation, and when users need water, the device slightly increases its heat generation capacity.

[0109] The touch screen provides human-machine interaction for the device; the device's hot water storage tank temperature, cold water storage tank temperature, hot water storage tank level, user water consumption, photovoltaic panel back panel temperature, environmental data, and variable frequency water pump flow rate are all displayed on the screen; users can set their needs through the touch screen, allowing the device to decide whether to generate electricity at full capacity or slightly increase its heat production capacity.

[0110] The photovoltaic power generation module has a cascading function; the inlet and outlet pipe caps of the photovoltaic power generation module can be removed and connected to the inlet and outlet pipes of another photovoltaic power generation module. By connecting them in sequence, one set of AI main control module, loop storage module, and data acquisition and control module can control multiple sets of photovoltaic power generation modules.

[0111] The device can connect to the Internet via WIFI and upload information such as the temperature of the hot water storage tank, the temperature of the cold water storage tank, the water level of the hot water storage tank, the user's water consumption, the temperature of the photovoltaic power generation panel back panel, environmental data, and the flow rate of the variable frequency water pump to the cloud. Users can view the device status in real time on their mobile devices.

[0112] This invention has the following characteristics: (1) Structural aspects: The photovoltaic modules of this device abandon the traditional frame or plastic backplate and adopt an aluminum heat dissipation backplate. The backplate is tightly attached to the photovoltaic panel with thermally conductive adhesive, which not only provides structural support, but also efficiently conducts and dissipates heat. The backplate is integrally formed by the "profile extrusion" process and has built-in heat exchange channels, which has higher heat conduction efficiency than traditional adhesive pipes. The trapezoidal structure of the backplate reduces the weight of the device while ensuring strength, increases the air contact area, and facilitates air circulation.

[0113] This device uses two water storage tanks, one for cold water and one for hot water, with parallel pipes installed between the tanks. It has two cooling water circulation modes, which can ensure that the hot water tank is quickly replenished with hot water and that cold water is continuously supplied for heat dissipation.

[0114] The photovoltaic and solar thermal modules of this device have a cascading function, which can enable one AI main control module, a loop storage module, and a data acquisition and control module to control multiple photovoltaic power generation modules, thereby improving power generation efficiency and heating efficiency.

[0115] (2) Algorithm aspect: The AI ​​prediction system of this device uses the Long Short-Term Memory (LSTM) network algorithm to predict the backplate temperature and outlet water temperature in the next 15 minutes. This not only solves the problem of the physical model being unable to be accurately constructed due to too many variables, but also improves the accuracy of the prediction.

[0116] The AI ​​decision-making system of this device adopts a global grid search algorithm to traverse all feasible variable frequency pump flow rates, find the optimal variable frequency pump flow rate, and ensure that the global optimal solution is found.

[0117] (3) Reliability: This device can select different operating modes. Users can choose between two modes: full power generation or slightly increased heat generation capacity. The AI ​​decision-making system will adjust decision preferences by changing the weight of the value function, and intelligently adapt to different business needs.

[0118] This device is equipped with a touch screen and WIFI functionality, displaying all device statuses on the screen and uploading them to the cloud internet, allowing users to view the device's working status in real time.

[0119] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A photovoltaic thermal device based on LSTM prediction and intelligent decision-making, characterized in that: include: Photovoltaic thermal modules are used to convert light energy into electrical energy while simultaneously collecting heat. The circulating storage module is used to realize the circulation of cooling water, the storage of hot water, and the cooling of photovoltaic panels; The data acquisition and control module is used to monitor various status parameters in real time, process them, and send them to the AI ​​main control module; it also receives control commands from the AI ​​main control module to control the circular storage module. The AI ​​main control module receives parameters collected by the data acquisition and control module, uses a long short-term memory recurrent neural network for state prediction, uses a global grid search algorithm for decision-making, obtains the most valuable variable frequency pump flow rate, and sends control commands to the data acquisition and control module.

2. The photovoltaic thermal device based on LSTM prediction and intelligent decision-making according to claim 1, characterized in that: The photovoltaic thermal module consists of a photovoltaic panel and an aluminum heat dissipation backplate. The photovoltaic panel is used to convert light energy into electrical energy. The aluminum heat dissipation backplate is used to collect the heat absorbed by the photovoltaic panel and cool it down. The aluminum heat dissipation backplate is connected to the bottom of the photovoltaic panel by thermally conductive adhesive. The aluminum heat dissipation backplate includes multiple heat exchange channels.

3. The photovoltaic thermal device based on LSTM prediction and intelligent decision-making according to claim 1, characterized in that: The circulating storage module consists of a cold water storage tank, a hot water storage tank, a solenoid valve, a variable frequency water pump, and pipes. The pipes include an inlet pipe, an outlet pipe, and pipes for the cold and hot water tanks. One end of the water inlet pipe is connected to a cold water storage tank, and the other end is connected to a heat exchange channel, which is used to transport cold water from the cold water storage tank to the channel. One end of the outlet pipe is connected to the heat exchange channel, and the other end is connected to the hot water storage tank, which is used to transport the water that has absorbed heat in the heat exchange channel to the hot water storage tank. One end of the hot and cold water tank pipe is connected to the cold water storage tank, and the other end is connected to the hot water storage tank; The solenoid valve is installed at the inlet of the cold water storage tank to control the entry of external cold water into the cold water storage tank. The variable frequency water pump is installed on the inlet pipe and is used to pump cold water from the cold water storage tank into the heat exchange channel and push hot water out of the heat exchange channel to the hot water storage tank.

4. The photovoltaic thermal device based on LSTM prediction and intelligent decision-making according to claim 1, characterized in that: The data acquisition and control module consists of an I / O controller and multiple sensors; The I / O controller is used to receive data from various sensors and send the processed data to the AI ​​main control module, and receive control commands from the AI ​​main control module to control the variable frequency water pump and solenoid valve; The sensor includes: Outlet water temperature sensor: installed on the outlet water pipe to monitor the hot water temperature in the outlet water pipe in real time; Electronic flow meter for water outlet: installed on the water outlet pipe to monitor the water flow rate in the pipe in real time; Inlet water temperature sensor: installed on the inlet water pipe to monitor the temperature of the cold water in the inlet water pipe in real time; Cold water tank temperature sensor: installed inside the cold water storage tank to monitor the temperature of the cold water in the tank in real time; Hot water tank temperature sensor: installed inside the hot water storage tank to monitor the temperature of the hot water in the tank in real time; Water level gauge: Installed inside the hot water storage tank to monitor the water level in the tank in real time; Electronic water flow meter: installed at the outlet of the hot water storage tank to monitor the flow rate of hot water flowing out of the tank in real time; Anemometer: Used for real-time monitoring of ambient wind speed around the device; Ambient temperature sensor: used to monitor the temperature around the device in real time; Light intensity sensor: used to monitor the solar irradiance received by the photovoltaic thermal module in real time; Back panel temperature sensor: installed between the photovoltaic panel and the aluminum heat dissipation back panel to monitor the temperature of the back of the photovoltaic panel in real time.

5. The photovoltaic thermal device based on LSTM prediction and intelligent decision-making according to claim 1, characterized in that: The AI ​​main control module includes an edge computing device, a human-computer interaction module, and a power supply module. The edge computing device is connected to the I / O controller, receives sensor data transmitted by the I / O controller, and constructs an AI prediction system based on a long short-term memory recurrent neural network to predict the future trend of backplate temperature and outlet water temperature. It also constructs an AI decision-making system based on a global grid search algorithm to calculate the optimal flow rate of the variable frequency water pump and send control commands to the variable frequency water pump and solenoid valve to the I / O controller. The human-computer interaction module is used for status information display and command input. The power supply module is used to power the edge computing device.

6. The photovoltaic thermal device based on LSTM prediction and intelligent decision-making according to claim 5, characterized in that: The method of predicting the future trends of backplate temperature and outlet water temperature based on a long short-term memory recurrent neural network specifically includes the following steps: A0: Model training and deployment phase: Train the long short-term memory recurrent neural network model on a high-performance computing workstation and deploy it in an edge computing device; A1: Input Phase: Continuously receives status data sent by the I / O control system, including the current backplane temperature. Current inlet water temperature Current water temperature Current water flow speed Current light intensity Current ambient temperature Current ambient wind speed And synthesize the data into the input feature vector at the current time. This is combined with historical data to form a sequence window that meets the model input requirements; A2: Prediction model solution stage: Load the long short-term memory recurrent neural network model, take the sequence window in A1 as input, and solve the prediction result through forward propagation; A3: Output stage: The predicted results are sent to the AI ​​decision-making system and the human-computer interaction module.

7. The photovoltaic thermal device based on LSTM prediction and intelligent decision-making according to claim 6, characterized in that: The calculation of the optimal flow rate of the variable frequency water pump based on the global grid search algorithm specifically includes the following steps: B1: Input Phase: Continuously receives mode information sent by the human-machine interface system and status data sent by the I / O controller, and reads the photovoltaic panel efficiency parameters. Specific heat capacity of water Water pump power consumption curve ; B2: Decision Model Solving Stage: Using a global grid search algorithm, all feasible flow velocities are traversed, the prediction results obtained from the AI ​​prediction system are called, the value function is calculated based on the input information, and the optimal decision is obtained through comparison. B3: Decision Output and Execution Stage: The AI ​​decision system sends the optimal decision to the I / O controller and the human-computer interaction system.

8. The photovoltaic thermal device based on LSTM prediction and intelligent decision-making according to claim 7, characterized in that: In step B2, the global grid search algorithm includes the following sub-steps: B21: Decision Space Definition: Define the complete legal operating range of the variable frequency water pump, forming a discretized set of candidate decisions. : in, This is the minimum flow rate for the water pump. This is the maximum flow rate of the water pump. This is the search step size; B22: Traversal Prediction Phase: For Each of them This is used as the current set water flow velocity in the AI ​​prediction model. Call the AI ​​prediction model to calculate the current state. Predicted backplate temperature and outlet water temperature at flow rates ; B23: Valuation Phase The AI ​​decision-making system calculates each based on the prediction results. The value, the value function is defined as: in, For net profit from power generation, For heating profits; The formula for net profit from power generation is: in, Average power generation: in, Power consumption of the water pump: in, As the weight of power generation; The formula for heating profit is: in, Average heat production power: in, The heat generation weight is determined by the mode selected by the user. B24: Determining the Optimal Decision: AI Decision-Making Systems from All The result is the one that maximizes the total value. The corresponding flow rate : This flow rate is the optimal flow rate determined by the decision.

9. The photovoltaic thermal device based on LSTM prediction and intelligent decision-making according to claim 1, characterized in that: The AI ​​main control module controls the opening and closing of the solenoid valves according to the device status, switching the cooling water circulation between mode one and mode two, as detailed below: Mode 1: When the hot water storage tank is not full, the solenoid valve of the cold water storage tank inlet pipe opens, and cooling water enters the cold water storage tank from the outside. It is then drawn into the heat dissipation channel by the variable frequency water pump, absorbs heat, becomes hot water, and enters the hot water storage tank. In this mode, the solenoid valve between the cold and hot water storage tanks is closed, and there is no water flow between the cold and hot water storage tanks. When the hot water storage tank is full, the water level gauge is triggered, and then the mode switches to Mode 2. Mode 2: When the hot water storage tank is full, the solenoid valve on the inlet pipe of the cold water storage tank is closed, and the solenoid valve between the hot and cold water storage tanks is opened. Hot water flows out of the hot water storage tank, flows into the cold water storage tank through the hot and cold water pipes, dissipates heat during this process, and becomes cold water again. The cold water is drawn into the heat dissipation channel by the variable frequency water pump, absorbs heat, becomes hot water again, and re-enters the hot water storage tank. In this mode, cooling water circulates throughout the device, ensuring that the hot water storage tank is always full. When the user uses hot water, the hot water storage tank is no longer full, and the system returns to Mode 1.

10. The photovoltaic thermal device based on LSTM prediction and intelligent decision-making according to claim 3, characterized in that: The photovoltaic thermal module has a cascading function. One end of the water inlet pipe and the water outlet pipe is equipped with a pipe cap. By removing the pipe cap and connecting it to the water inlet pipe and the water outlet pipe of another photovoltaic power generation module, a set of AI main control module, loop storage module, and data acquisition and control module can control multiple photovoltaic power generation modules.

Images