Coordinated adaptive variable flow photovoltaic module liquid cooling system and control method

Abstract

The application discloses a kind of self-adapting variable flow photovoltaic module liquid cooling systems and control methods of synergic regulation, with bionic tree-shaped V-shaped cold plate, contain V-shaped fin and multistage tree-shaped fin, when temperature rises, tree-shaped fin self-adapting expansion, increase the flow of coolant in heat concentration area, realize passive distribution and component surface temperature homogenization of flow.The control method, through long short-term memory neural network predicts photovoltaic module heat load and temperature distribution in future time;Based on the prediction result, a multi-objective particle swarm optimization model is constructed, and a feedforward control instruction is obtained by solving;At the same time, the current temperature deviation is input into the PID feedback controller to calculate the feedback correction amount;Synergistically regulate mechanical pump frequency and heater power.The application realizes passive self-adapting flow regulation through bionic flow channel structure, combined with active synergistic control strategy, accurately stabilizes photovoltaic module operating temperature in optimal interval, improves photovoltaic module power generation efficiency, temperature uniformity and system energy comprehensive utilization rate.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic module thermal management technology, and in particular to a collaboratively regulated adaptive variable flow photovoltaic module liquid cooling system and control method. Background Technology

[0002] As a crucial component of renewable energy, photovoltaic (PV) power generation's conversion efficiency and lifespan are directly affected by operating temperature. During the process of converting solar energy into electricity, only about 20% of the incident energy is converted into electrical energy; the majority is converted into heat, leading to a significant increase in module temperature. Prolonged operation at high temperatures accelerates the aging of encapsulation materials, increases the risk of hot spot effects, and in severe cases, may cause permanent damage to the modules. Therefore, efficient and reliable PV module cooling technology is of great significance for improving power generation and ensuring stable system operation.

[0003] Currently, the cooling methods for photovoltaic modules are mainly divided into two categories: passive cooling and active cooling. Passive cooling includes natural air cooling, heat dissipation fins, phase change materials, etc., which have simple structures but limited heat dissipation capacity and are difficult to meet the heat dissipation requirements under high irradiance or high temperature environments. Active cooling includes forced air cooling and liquid cooling. Forced air cooling increases airflow by using a fan, but the fan consumes a lot of power and has a relatively low heat transfer coefficient. Liquid cooling systems use a circulating cooling medium to remove heat, which has the advantages of high heat transfer coefficient and good temperature control. However, existing photovoltaic liquid cooling systems still have the following shortcomings: (1) The cooling channel structure design is simple. Traditional liquid cooling systems mostly use straight pipes or serpentine pipe channels, and the coolant is unevenly distributed in the channel, resulting in local high temperature areas on the surface of the photovoltaic module. The uneven temperature distribution not only reduces the overall power generation efficiency, but also accelerates local aging and forms hot spots; (2) The thermal management strategy is lagging. Existing control methods are mostly switch control or PID feedback control based on the current temperature. When the heat load changes suddenly, the response is lagging and it is impossible to accurately control the module temperature in the optimal operating range. Meanwhile, the lack of predictive ability for future heat load of photovoltaic modules makes it difficult to achieve forward-looking regulation. (3) High system energy consumption and low energy utilization. The cooling system is usually powered by an external power source, which increases additional energy consumption. At the same time, the heat absorbed in the cooling loop is often directly dissipated into the environment without being recovered and reused, resulting in energy waste. (4) Single control dimension. Most systems only change the cooling intensity by adjusting the flow rate of the circulating pump, and fail to coordinate the adjustment of other key parameters such as the working fluid inlet temperature, thus limiting the temperature control accuracy and response speed.

[0004] Therefore, how to design an adaptive variable flow photovoltaic module liquid cooling system that can achieve uniform cooling of photovoltaic modules, has the ability to predict heat load, and can coordinate the flow rate and inlet temperature has become an urgent technical problem to be solved. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by providing a liquid cooling system and its control method that can achieve uniform cooling of photovoltaic modules and adaptive variable flow rate coordinated regulation according to changes in heat load, so as to solve the problems of uneven temperature distribution and control lag in existing liquid cooling systems.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In the first aspect, an adaptive variable flow photovoltaic module liquid cooling system with coordinated regulation is provided, including a photovoltaic module, a cooling circulation loop, a data acquisition module, and a coordinated control module; The photovoltaic module includes stacked photovoltaic cells, a heat-absorbing layer, and a biomimetic tree-shaped V-shaped cold plate. The biomimetic tree-shaped V-shaped cold plate is provided with a number of spaced tree-shaped ribs and V-shaped ribs. The heat generated by the photovoltaic cells is transferred to the biomimetic tree-shaped V-shaped cold plate through the heat-absorbing layer and exchanges heat through the cooling medium flowing between the tree-shaped ribs and V-shaped ribs. The cooling circulation loop includes a variable frequency circulation pump, a heat exchanger, and a liquid storage tank. The variable frequency circulation pump, the bionic tree-shaped V-shaped cold plate, the heat exchanger, and the liquid storage tank are connected in series and a cooling medium is passed through them. A heater is installed in the liquid storage tank. The variable frequency circulation pump drives the cooling medium to flow through the bionic tree-shaped V-shaped cold plate, the heat exchanger, and the liquid storage tank in sequence and then return to the variable frequency circulation pump to form a closed loop. The data acquisition module includes an ambient light sensor, an ambient temperature sensor, a photovoltaic module temperature sensor, a cold plate inlet temperature sensor, a cold plate outlet temperature sensor, and a flow meter. The photovoltaic module temperature sensor is installed on the photovoltaic cells. The cold plate inlet temperature sensor and the cold plate outlet temperature sensor are respectively installed at the inlet and outlet ends of the biomimetic tree-shaped V-shaped cold plate. The flow meter is installed in the cooling circulation loop. The collaborative control module is electrically connected to the data acquisition module, the variable frequency circulating pump, and the heater. The collaborative control module incorporates a collaborative control strategy that combines long short-term memory neural network prediction, multi-objective particle swarm optimization, and PID feedback controller. These strategies are superimposed to form a composite control signal, which is used to collaboratively adjust the frequency of the variable frequency circulating pump to change the circulation flow rate of the cooling medium, and to adjust the power of the heater in the storage tank to change the inlet temperature of the cooling medium.

[0007] Furthermore, the tree-like ribs include a first tree-like rib, a second tree-like rib, and a third tree-like rib that are symmetrically distributed along the central axis. The V-shaped ribs include a first V-shaped rib distributed on the outside of the tree-like ribs and a second V-shaped rib distributed in the middle of the tree-like ribs. The second and third tree-like ribs are made of shape memory alloy. When the temperature rises, the inlet ends of the second and third tree-like ribs gradually expand to both sides to increase the cooling medium flow rate in the middle of the biomimetic tree-like V-shaped cold plate.

[0008] Furthermore, the length ratio of the first, second, and third tree-shaped ribs is 2:3:5, and the angles with the length direction are 75°, 35°, and 25°, respectively.

[0009] Furthermore, the first V-shaped rib includes a straight edge section, a middle section, and a beveled edge section from the outside to the inside, with a length ratio of 2:3:1. The straight edge section, the middle section, and the beveled edge section are respectively provided with a first circular hole, a second circular hole, and a third circular hole. The number ratio of the first circular hole, the second circular hole, and the third circular hole is 2:3:1, and the hole diameter ratio is 4:3:4.

[0010] Furthermore, a fourth circular hole is provided on the second V-shaped rib.

[0011] Furthermore, the long short-term memory neural network prediction includes constructing a feature matrix from normalized time-series data according to a fixed time window, and then inputting the feature matrix into a pre-trained long short-term memory neural network to predict the heat load and temperature distribution of the photovoltaic module at future times. ,in X t In order to be in t The feature matrix constructed at each time step, T t-k:t For the temperature of each zone of the photovoltaic module at tk arrive t Time series data for this time window I rad,t-k:t This is time-series data of ambient light intensity. P PV,t-k: Time series data of photovoltaic module output power. T amb,t-k:t For time series data of ambient temperature, T in,t-k:t This is time-series data of the inlet temperature of the cooling medium. T out,t-k:t This is time-series data of the outlet temperature of the cooling medium.

[0012] Furthermore, the multi-objective particle swarm optimization, based on the prediction results of the long short-term memory neural network, constructs a multi-objective particle swarm optimization model with the dual objectives of minimizing temperature deviation and minimizing system energy consumption. The feedforward control command is obtained by solving this model. The objective function of the multi-objective particle swarm optimization model is: ,in , Decision variables It includes frequency regulation of variable frequency circulating pump and power regulation of heater. The Pareto front solution set is obtained through particle swarm iteration, and the optimal compromise solution is selected as the feedforward control command using fuzzy set theory.

[0013] Furthermore, the PID feedback controller receives the current photovoltaic module temperature and a set value as input. deviation ,according to Calculate the feedback correction amount, where, T pv (t) represents the measured temperature of the photovoltaic module at the current moment. T set Set the temperature value. U fb (t) represents the PID feedback correction value. K p This is the proportionality coefficient. K i The integral coefficient is... K d This represents the differential coefficient.

[0014] Furthermore, a heat storage condenser is provided in the cooling circulation loop. The interior of the heat storage condenser is divided into several flow channels by porous foam metal. A third V-shaped fin is provided in the flow channel. The inlet of the heat storage condenser is connected to the outlet of the biomimetic tree-shaped V-shaped cold plate, and the outlet of the heat storage condenser is connected to the inlet of the heat exchanger.

[0015] Secondly, a collaboratively regulated adaptive variable flow photovoltaic module liquid cooling control method is provided, based on the system described in the first aspect, including the following steps: Data Acquisition and Preprocessing: Real-time acquisition of temperature in each zone of the photovoltaic module. Photovoltaic module output power Total flow rate of cooling medium Cooling medium inlet temperature Cooling medium outlet temperature Ambient light intensity and ambient temperature ; The multi-source data collected above is preprocessed, including using moving average filtering to remove outliers and normalizing the data. Heat load prediction: Construct a feature matrix from normalized time series data according to a fixed time window. , among which, among which X t In order to be in t The feature matrix constructed at each time step, T t-k:t For the temperature of each zone of the photovoltaic module at tk arrive t Time series data for this time window I rad,t-k:t This is time-series data of ambient light intensity. P PV,t-k:Time series data of photovoltaic module output power. T amb,t-k:t For time series data of ambient temperature, T in,t-k:t This is time-series data of the inlet temperature of the cooling medium. T out,t-k:t The time series data of the cooling medium outlet temperature is used; the feature matrix is ​​input into a pre-trained long short-term memory neural network to predict the heat load and temperature distribution at future times. Feedforward control optimization: A multi-objective particle swarm optimization model is established with the dual objectives of minimizing temperature deviation and system energy consumption. The objective function is... ,in Indicates temperature deviation; Represents system energy consumption; decision variables Includes the frequency of the variable frequency circulating pump and the power of the heater. T pv,i Indicates the first Temperature of photovoltaic modules in each zone T opt This indicates the optimal operating temperature for photovoltaic modules. n Indicates the total number of photovoltaic module zones. Q Δ represents the volumetric flow rate of the cooling medium. p This represents the total voltage drop in the loop. η pump Indicates the efficiency of the variable frequency circulating pump. f pump Indicates the operating frequency of the variable frequency circulating pump. α The power regulation coefficient of the heater is represented by the Pareto front solution set obtained through particle swarm optimization, and the feedforward control command is obtained by solving for the Pareto front solution set. ; Feedback correction: Input the deviation between the current component temperature and the set value into the PID feedback controller, and adjust according to the formula. Calculate the feedback correction amount, where, T pv (t) represents the measured temperature of the photovoltaic module at the current moment. T set Indicates the temperature setpoint. U fb (t) represents the PID feedback correction amount. K p Represents the proportionality coefficient. K i Represents the integral coefficient. K d Represents the differential coefficient; The feedforward control command and the feedback control quantity are superimposed to form a composite control signal. After analysis, the frequency of the variable frequency circulating pump and the power of the heater are adjusted to achieve variable flow adaptive cooling.

[0016] Compared with the prior art, the beneficial effects of the present invention are: (1) High heat exchange efficiency and good temperature uniformity: Through the biomimetic tree-shaped V-shaped cold plate structure, the adaptive characteristics of shape memory alloy and the turbulence effect of V-shaped fins are utilized to realize the on-demand distribution of cooling flow, effectively solving the problem of uneven temperature distribution of photovoltaic modules and significantly improving the overall heat exchange effect.

[0017] (2) High control accuracy, fast response, and low energy consumption: A collaborative control strategy combining long short-term memory neural network prediction, multi-objective particle swarm optimization, and PID feedback is adopted, combining the predictability of feedforward with the accuracy of feedback to form a collaborative control of prediction, optimization, and feedback. The long short-term memory neural network enables accurate prediction of future heat load, giving the system predictability; the multi-objective particle swarm optimization dynamically balances the two objectives of minimizing temperature deviation and minimizing system energy consumption to obtain the optimal feedforward control command, ensuring that the system energy consumption is minimized while accurately controlling the temperature, overcoming the shortcomings of traditional control methods such as lag and high energy consumption; finally, PID feedback correction is added to form a composite control combining feedforward and feedback. This collaborative control strategy not only has a faster response speed and higher temperature control accuracy, but also significantly reduces the energy consumption of the cooling system itself by taking energy consumption as one of the optimization objectives and finding the most energy-efficient operating mode while meeting the temperature control requirements by simultaneously adjusting the pump frequency and the inlet temperature of the cooling medium.

[0018] (3) High energy utilization rate: The system uses the photovoltaic modules to drive the cooling system, achieving energy self-sufficiency. At the same time, the heat storage condenser can effectively store and buffer heat, avoiding energy waste, and can recover and utilize waste heat through the heat exchanger, playing a role in peak shaving and valley filling, and smoothing the impact of heat load fluctuations on the system.

[0019] (4) The system operates stably and reliably: By coordinating the control of flow rate and inlet temperature, the operating temperature of photovoltaic modules is precisely stabilized within the optimal range, which not only improves power generation efficiency but also extends the service life of the modules. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the overall structure of Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the photovoltaic module structure according to Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the biomimetic tree-shaped V-shaped cold plate structure of Embodiment 1 of the present invention; Figure 4 This is a front view of the biomimetic tree-shaped V-shaped cold plate according to Embodiment 1 of the present invention; Figure 5 (a) is a side view of the first V-shaped rib in Embodiment 1 of the present invention; Figure 5 (b) is a side view of the second V-shaped rib in Embodiment 1 of the present invention; Figure 6 This is a schematic diagram of the heat storage condenser structure according to Embodiment 1 of the present invention; Figure 7 This is a front view of the heat storage condenser structure according to Embodiment 1 of the present invention; Figure 8 This is a flowchart of the adaptive variable flow photovoltaic module liquid cooling control method of coordinated regulation according to Embodiment 2 of the present invention; In the diagram: 1-Cooperative control module, 2-Photovoltaic module, 3-Variable frequency circulating pump, 4-Heat exchanger, 5-Storage tank, 6-Heater, 7-Ambient light sensor, 8-Ambient temperature sensor, 9-Photovoltaic module temperature sensor, 10-Cold plate inlet temperature sensor, 11-Cold plate outlet temperature sensor, 12-Flow meter, 13-First dendritic fin, 14-Second dendritic fin, 15-Third dendritic fin, 16-First V-shaped fin, 17-Second V-shaped fin, 18-Heat storage condenser, 19-Solar controller, 20-Inverter, 21-Photovoltaic cell, 22-Heat absorption layer, 23-Bionic dendritic V-shaped cold plate, 24- EVA film, 161-straight edge section, 162-middle section, 163-slanted edge section, 164-first round hole, 165-second round hole, 166-third round hole, 171-fourth round hole, 181-porous foam metal, 182-flow channel, 183-third V-shaped fin, 184-heat storage condenser inlet, 185-heat storage condenser outlet. Detailed Implementation

[0021] To enhance understanding of the present invention, we will now describe it in further detail with reference to the accompanying drawings. These embodiments are for illustrative purposes only and do not constitute a limitation on the scope of protection of the present invention.

[0022] Example 1 like Figure 1 As shown, an adaptive variable flow photovoltaic module liquid cooling system with coordinated control includes a photovoltaic module 2, a cooling circulation loop, a data acquisition module, and a coordinated control module 1.

[0023] Among them, such as Figure 2 As shown, the photovoltaic module 2 includes stacked photovoltaic cells 21, EVA film 24, heat-absorbing layer 22, and biomimetic tree-shaped V-shaped cold plate 23. The EVA film 24 is located below the photovoltaic cells 21 and is mainly used to bond the photovoltaic cells 21, glass, and backsheet, protecting the cells and extending the module's lifespan. The heat-absorbing layer 22 is located below the EVA film 24, and heat is transferred to the biomimetic tree-shaped V-shaped cold plate 23 through the heat-absorbing layer 22.

[0024] like Figure 3 , 4 As shown, the biomimetic tree-shaped V-shaped cold plate 23 is provided with several spaced-apart tree-shaped ribs and V-shaped ribs. The heat generated by the photovoltaic cells 21 is transferred to the biomimetic tree-shaped V-shaped cold plate 23 through the heat absorption layer 22, and heat is exchanged through the cooling medium flowing between the tree-shaped ribs and V-shaped ribs. The flow channel layout on the biomimetic tree-shaped V-shaped cold plate 23 imitates the branching shape of trees in nature through the tree-shaped ribs, and combines it with the turbulence of the V-shaped ribs to achieve efficient distribution of the cooling medium and enhanced heat exchange.

[0025] The tree-like ribs include a first tree-like rib 13, a second tree-like rib 14, and a third tree-like rib 15 symmetrically distributed along the central axis. The V-shaped ribs include a first V-shaped rib 16 distributed on the outer side of the tree-like ribs and a second V-shaped rib 17 distributed in the middle of the tree-like ribs. The second tree-like ribs 14 and the third tree-like ribs 15 are made of shape memory alloy. They can undergo plastic deformation in the low-temperature phase (martensitic phase) and automatically recover to their original shape (austenitic phase) when the temperature rises above the phase transformation temperature. Alternatively, bimetallic sheets or materials with different coefficients of thermal expansion can be used instead of shape memory alloys to achieve the same effect of changing the flow cross-sectional area when the temperature rises.

[0026] As the temperature of the photovoltaic module 2 rises, the inlet ends of the second tree-shaped fin 14 and the third tree-shaped fin 15 gradually expand to both sides, increasing the flow cross-sectional area of ​​the heat concentration area in the middle of the bionic tree-shaped V-shaped cold plate 23. This increases the flow rate of the cooling medium in the middle of the bionic tree-shaped V-shaped cold plate 23, improves the heat exchange capacity of the heat concentration area in the middle, realizes adaptive variable flow distribution, strengthens local heat exchange, and makes the temperature distribution of the photovoltaic module 2 more uniform.

[0027] After being diverted by the first tree-shaped fin 13, the cooling medium enters the edge region of the cold plate, where the heat exchange between the fluid and the hot wall surface is enhanced by the first V-shaped fin 16. The cooling medium is then diverted by the second tree-shaped fin 14 and the third tree-shaped fin 15 into the middle region of the cold plate, where the turbulence is enhanced by the second V-shaped fin 17, disrupting the fluid boundary layer and further improving the heat exchange efficiency.

[0028] Preferably, the length ratio of the first tree-shaped rib 13, the second tree-shaped rib 14, and the third tree-shaped rib 15 is 2:3:5, and the angles between the main body and the length direction are 75°, 35°, and 25°, respectively.

[0029] like Figure 5As shown in (a), the first V-shaped rib 16 includes a straight edge section 161, a middle section 162, and a beveled edge section 163 from the outside to the inside, with a length ratio of 2:3:1. A first circular hole 164, a second circular hole 165, and a third circular hole 166 are respectively provided on the straight edge section 161, the middle section 162, and the beveled edge section 163. The ratio of the number of the first circular holes 164, the second circular holes 165, and the third circular holes 166 is 2:3:1, and the diameter ratio is 4:3:4. Figure 5 As shown in (b), a fourth circular hole 171 is provided on the second V-shaped rib 17. The large-diameter circular hole helps to turbulent the cooling medium at the edge of the cold plate, thereby enhancing edge heat transfer.

[0030] like Figure 1 As shown, the cooling circulation loop includes a variable frequency circulating pump 3, a heat exchanger 4, a liquid storage tank 5, and a heat storage condenser 18. The variable frequency circulating pump 3, the biomimetic tree-shaped V-shaped cold plate 23, the heat storage condenser 18, the heat exchanger 4, and the liquid storage tank 5 are connected in series and circulated with a cooling medium. A heater 6 is installed in the liquid storage tank 5.

[0031] The cooling medium in the storage tank 5 is driven by the variable frequency circulating pump 3 to flow through the biomimetic tree-shaped V-shaped cold plate 23. The low-temperature fluid absorbs the heat released by the photovoltaic cells 21, causing the temperature of the cooling medium to rise. It then enters the thermal storage condenser 18 through the inlet 184, exchanges heat, and then passes through the outlet 185 of the thermal storage condenser to exchange heat with the cooling water via the heat exchanger 4. Finally, it returns to the storage tank 5 and then to the variable frequency circulating pump 3, forming a closed loop. The photovoltaic cells 21 output direct current through photoelectric conversion. The direct current is converted into alternating current by the solar controller 19 and the inverter 20. The alternating current is used to operate the variable frequency circulating pump 3 and the heater 6.

[0032] like Figure 6 , 7 As shown, the interior of the thermal storage condenser 18 is divided into three flow channels 182 by porous foam metal 181. Each flow channel 182 is provided with a third V-shaped fin 183. The inlet 184 of the thermal storage condenser is connected to the outlet of the biomimetic tree-shaped V-shaped cold plate 23, and the outlet 185 of the thermal storage condenser is connected to the inlet of the heat exchanger 4. Porous foam metal 181, as a high-porosity metal material with a three-dimensional network pore structure, is commonly made of copper, aluminum, or their alloys. It has a large specific surface area and excellent thermal conductivity, and can be used as a carrier for phase change materials, significantly improving the heat transfer rate of the thermal storage unit.

[0033] The high-temperature cooling medium flowing out of the biomimetic tree-shaped V-shaped cold plate 23 flows into three flow channels 182 through the inlet 184 of the thermal storage condenser. A third V-shaped fin 183 is arranged in the flow channels 182 to enhance turbulence, thereby disrupting the fluid boundary layer and strengthening heat exchange between the cooling medium and the side phase change material. The phase change material is located in the porous foam metal 181. The solid phase change material absorbs heat from the high-temperature cooling medium and undergoes a melting phase change. The cooling medium in the flow channels 182 experiences a temperature reduction through heat exchange and flows out through the outlet 185 of the thermal storage condenser. This composite structure utilizes the high latent heat of the phase change material to absorb and release heat, acting as a thermal buffer, and enhances the heat exchange efficiency of the energy storage and release processes through the high thermal conductivity of the porous foam metal 181.

[0034] like Figure 1 As shown, the data acquisition module includes an ambient light sensor 7, an ambient temperature sensor 8, a photovoltaic module temperature sensor 9, a cold plate inlet temperature sensor 10, a cold plate outlet temperature sensor 11, and a flow meter 12. The ambient light sensor 7 and ambient temperature sensor 8 are used to collect ambient light intensity and ambient temperature, respectively. The photovoltaic module temperature sensor 9 is installed on the photovoltaic cell 21 to collect the temperature distribution on the photovoltaic cell 21. The cold plate inlet temperature sensor 10 and cold plate outlet temperature sensor 11 are installed at the inlet and outlet ends of the biomimetic tree-shaped V-shaped cold plate 23, respectively, to collect the inlet and outlet temperatures of the cooling medium. The flow meter 12 is installed in the cooling circulation loop to collect the circulating flow rate of the cooling medium.

[0035] The collaborative control module 1 is electrically connected to the data acquisition module, the variable frequency circulating pump 3, and the heater 6. The collaborative control module 1 incorporates a collaborative control strategy that combines long short-term memory neural network prediction, multi-objective particle swarm optimization, and PID feedback controller: with the dual objectives of minimizing temperature deviation and minimizing system energy consumption, a multi-objective particle swarm optimization model is established to obtain the feedforward control command. The deviation between the current component temperature and the set value is input into the PID feedback controller to calculate the feedback correction amount. Superimposed to form a composite control signal After analysis, it is used to coordinate the adjustment of the frequency of the variable frequency circulating pump 3 to change the circulation flow rate of the cooling medium, and to adjust the power of the heater 6 in the liquid storage tank 5 to change the inlet temperature of the cooling medium.

[0036] The Long Short-Term Memory (LSTM) neural network prediction process involves constructing a feature matrix from normalized time-series data within a fixed time window, then inputting this feature matrix into a pre-trained LSM neural network to predict the heat load and temperature distribution of the photovoltaic module at future moments. ,in X t In order to be in t The feature matrix constructed at each time step, Tt-k:t For the temperature of each zone of the photovoltaic module at tk arrive t Time series data for this time window I rad,t-k:t This is time-series data of ambient light intensity. P PV,t-k: Time series data of photovoltaic module output power. T amb,t-k:t For time series data of ambient temperature, T in,t-k:t This is time-series data of the inlet temperature of the cooling medium. T out,t-k:t This is time-series data of the outlet temperature of the cooling medium.

[0037] Multi-objective particle swarm optimization (MPS) is based on predictions from a long short-term memory neural network. A multi-objective PMS model is constructed with the dual objectives of minimizing temperature deviation and minimizing system energy consumption. The feedforward control command is obtained by solving this model. The objective function of the multi-objective PMS model is: ,in , Decision variables It includes frequency regulation of variable frequency circulating pump 3 and power regulation of heater 6. The Pareto front solution set is obtained through particle swarm optimization, and the optimal compromise solution is selected as the feedforward control command using fuzzy set theory. .

[0038] The PID feedback controller uses the current photovoltaic module temperature and the set value as input. deviation ,according to Calculate the feedback correction amount, where, T pv (t) represents the measured temperature of the photovoltaic module at the current moment. T set Set the temperature value. U fb (t) represents the PID feedback correction value. K p This is the proportionality coefficient. K i The integral coefficient is... K d This represents the differential coefficient.

[0039] Example 2 like Figure 8 As shown, an adaptive variable flow photovoltaic module liquid cooling control method with coordinated regulation is provided. This control method is based on the system described in Embodiment 1 and includes the following steps: (1) Data acquisition and preprocessing: The temperature of each zone of the photovoltaic module is collected in real time through temperature sensors, flow meters and other devices in the data acquisition module. Photovoltaic module output power Total flow rate of cooling medium Cooling medium inlet temperature Cooling medium outlet temperature Ambient light intensity and ambient temperature Data from multiple sources.

[0040] The collected raw data is filtered using a moving average to remove outliers, and then processed according to the formula... Normalization is performed to eliminate the influence of dimensions. Among these, x This refers to a specific value in the collected raw data, including measurable physical quantities such as temperature, flow rate, light intensity, and power. x min This represents the minimum value of the physical quantity within a certain data collection period. x max This indicates the maximum value of the physical quantity within a certain data collection period. x norm This represents the dimensionless value obtained after normalization, thus limiting the range of this value to the interval [0, 1].

[0041] (2) Heat load prediction: Construct a feature matrix from the normalized time series data according to a fixed time window. This enables the fusion of heterogeneous data from multiple sources. X t In order to be in t The feature matrix constructed at each time step, T t-k:t For the temperature of each zone of the photovoltaic module at tk arrive t Time series data for this time window I rad,t-k:t This is time-series data of ambient light intensity. P PV,t-k: Time series data of photovoltaic module output power. T amb,t-k:t For time series data of ambient temperature, T in,t-k:t This is time-series data of the inlet temperature of the cooling medium. T out,t-k:t This is time-series data of the outlet temperature of the cooling medium.

[0042] The feature matrix is ​​input into a pre-trained weighted long short-term memory neural network to predict the heat load at future time points. and temperature distribution ,in, Indicates the predicted t + Δt The heat load at any given time. Indicates the predicted t + Δt The temperature of the photovoltaic module at any given time. Long Short-Term Memory (LSTM) neural networks, as a special type of recurrent neural network, excel at processing time-series data. They control long-term dependencies of information through a "gate" structure (input gate, forget gate, output gate), enabling them to learn dynamic patterns from historical data.

[0043] (3) Feedforward control optimization: Based on the prediction results, a multi-objective particle swarm optimization model with the dual objectives of minimizing temperature deviation and minimizing system energy consumption is constructed. As a heuristic optimization algorithm, the multi-objective particle swarm optimization model simulates the foraging behavior of bird flocks and optimizes two or more conflicting objective functions simultaneously through the cooperative search of multiple particles in the solution space.

[0044] The objective function is ,in The first sub-objective function represents the temperature deviation, which is the sum of the absolute values ​​of the differences between the temperatures of each partition and the optimal temperature. The second sub-objective function represents the system energy consumption, specifically the power consumption of the variable frequency circulating pump; decision variables. Includes a variable frequency circulating pump (3 frequencies) and a heater (6 power). T pv,i Indicates the first Temperature of photovoltaic modules in each zone T opt This indicates the optimal operating temperature for photovoltaic modules. n Indicates the total number of photovoltaic module zones. Q Δ represents the volumetric flow rate of the cooling medium. p This represents the total voltage drop in the loop. η pump This indicates the efficiency of the variable frequency circulating pump 3. f pump This indicates the operating frequency of the variable frequency circulating pump 3. α This represents the power regulation coefficient of heater 6. The Pareto front solution set is obtained through particle swarm optimization, representing the optimal trade-off between different objectives. Fuzzy set theory is then used to select the optimal compromise solution as the feedforward control command. .

[0045] (4) Feedback correction: Compare the current photovoltaic module temperature with the set value deviation Input to the PID feedback controller, according to the formula Calculate the feedback correction amount, where,T pv (t) represents the measured temperature of the photovoltaic module at the current moment. T set Indicates the temperature setpoint. U fb (t) represents the PID feedback correction amount. K p Represents the proportionality coefficient. K i Represents the integral coefficient. K d This represents the differential coefficient.

[0046] (5) Composite control: feedforward control command With feedback correction amount U fb Superimposed to form a composite control signal .in, U ff (t) represents the feedforward control command. U (t) represents the composite control signal. Composite control is a control strategy that combines predictive control and feedback control. Feedforward control provides the control quantity in advance based on the predictive model to deal with known or predictable disturbances, while feedback control corrects the output based on the deviation from the setpoint. In this method, the two are superimposed, balancing response speed and control accuracy.

[0047] After being analyzed, the signal generates dual execution instructions: one is to control the frequency of the variable frequency circulating pump 3. To regulate the total flow rate; on the other hand, to adjust the power of the heater 6 inside the storage tank 5, thereby changing the inlet temperature of the cooling medium. Within the set range. After the system parameters are adjusted, return to step (1), and the data is re-acquired and fed back through the data acquisition module to form a closed-loop control.

[0048] The above specific embodiments are only for illustrating the technical concept and structural features of the present invention, and are intended to enable those skilled in the art to implement them. However, the above content does not limit the scope of protection of the present invention. Any equivalent changes or modifications made in accordance with the spirit and essence of the present invention should fall within the scope of protection of the present invention.

Claims

1. A synergistically controlled adaptive variable flow photovoltaic module liquid cooling system, characterized in that: Includes photovoltaic modules (2), cooling circulation loop, data acquisition module, and collaborative control module (1); The photovoltaic module (2) includes stacked photovoltaic cells (21), a heat absorption layer (22), and a biomimetic tree-shaped V-shaped cold plate (23). The biomimetic tree-shaped V-shaped cold plate (23) is provided with a number of tree-shaped ribs and V-shaped ribs arranged at intervals. The heat generated by the photovoltaic cells (21) is transferred to the biomimetic tree-shaped V-shaped cold plate (23) through the heat absorption layer (22), and the heat is exchanged through the cooling medium flowing between the tree-shaped ribs and V-shaped ribs. The cooling circulation loop includes a variable frequency circulation pump (3), a heat exchanger (4), and a liquid storage tank (5). The variable frequency circulation pump (3), the bionic tree-shaped V-shaped cold plate (23), the heat exchanger (4), and the liquid storage tank (5) are connected in series and a cooling medium is passed through them. A heater (6) is installed in the liquid storage tank (5). The variable frequency circulation pump (3) drives the cooling medium to flow through the bionic tree-shaped V-shaped cold plate (23), the heat exchanger (4), and the liquid storage tank (5) in sequence and then returns to the variable frequency circulation pump (3) to form a closed loop. The data acquisition module includes an ambient light sensor (7), an ambient temperature sensor (8), a photovoltaic module temperature sensor (9), a cold plate inlet temperature sensor (10), a cold plate outlet temperature sensor (11), and a flow meter (12). The photovoltaic module temperature sensor (9) is installed on the photovoltaic cell (21). The cold plate inlet temperature sensor (10) and the cold plate outlet temperature sensor (11) are respectively installed at the inlet and outlet ends of the biomimetic tree-shaped V-shaped cold plate (23). The flow meter (12) is installed on the cooling circulation loop. The collaborative control module (1) is electrically connected to the data acquisition module, the variable frequency circulating pump (3), and the heater (6). The collaborative control module (1) incorporates a collaborative control strategy of long short-term memory neural network prediction, multi-objective particle swarm optimization, and PID feedback controller, which are superimposed to form a composite control signal. This signal is used to collaboratively adjust the frequency of the variable frequency circulating pump (3) to change the circulation flow rate of the cooling medium, and to adjust the power of the heater (6) in the liquid storage tank (5) to change the inlet temperature of the cooling medium.

2. The adaptive variable flow photovoltaic module liquid cooling system with coordinated regulation according to claim 1, characterized in that: The tree-like ribs include a first tree-like rib (13), a second tree-like rib (14), and a third tree-like rib (15) symmetrically distributed along the central axis. The V-shaped ribs include a first V-shaped rib (16) distributed on the outside of the tree-like ribs and a second V-shaped rib (17) distributed in the middle of the tree-like ribs. The second tree-like rib (14) and the third tree-like rib (15) are made of shape memory alloy. When the temperature rises, the inlet ends of the second tree-like rib (14) and the third tree-like rib (15) gradually expand to both sides to increase the cooling medium flow rate in the middle of the biomimetic tree-like V-shaped cold plate (23).

3. The adaptive variable flow photovoltaic module liquid cooling system with coordinated regulation according to claim 2, characterized in that: The length ratio of the first tree-shaped rib (13), the second tree-shaped rib (14), and the third tree-shaped rib (15) is 2:3:5, and the angles with the length direction are 75°, 35°, and 25°, respectively.

4. The adaptive variable flow photovoltaic module liquid cooling system with coordinated regulation according to claim 2, characterized in that: The first V-shaped rib (16) includes a straight edge section (161), a middle section (162) and a beveled edge section (163) from the outside to the inside, with a length ratio of 2:3:

1. The straight edge section (161), the middle section (162) and the beveled edge section (163) are respectively provided with a first round hole (164), a second round hole (165) and a third round hole (166). The number ratio of the first round hole (164), the second round hole (165) and the third round hole (166) is 2:3:1 and the diameter ratio is 4:3:

4.

5. The adaptive variable flow photovoltaic module liquid cooling system with coordinated regulation according to claim 2, characterized in that: The second V-shaped rib (17) has a fourth round hole (171).

6. The adaptive variable flow photovoltaic module liquid cooling system with coordinated regulation according to claim 1, characterized in that: The Long Short-Term Memory (LSTM) neural network prediction involves constructing a feature matrix from normalized time-series data within a fixed time window, and then inputting the feature matrix into a pre-trained LSM neural network to predict the heat load and temperature distribution of the photovoltaic module at future times. ,in X t In order to be in t The feature matrix constructed at each time step, T t-k:t For the temperature of each zone of the photovoltaic module at tk arrive t Time series data for this time window I rad,t-k:t This is time-series data of ambient light intensity. P PV,t-k: Time series data of photovoltaic module output power. T amb,t-k:t For time series data of ambient temperature, T in,t-k:t This is time-series data of the inlet temperature of the cooling medium. T out,t-k:t This is time-series data of the outlet temperature of the cooling medium.

7. The adaptive variable flow photovoltaic module liquid cooling system with coordinated regulation according to claim 1, characterized in that: The multi-objective particle swarm optimization is based on the prediction results of the long short-term memory neural network. A multi-objective particle swarm optimization model is constructed with the dual objectives of minimizing temperature deviation and minimizing system energy consumption. The feedforward control command is obtained by solving the model. The objective function of the multi-objective particle swarm optimization model is: ,in , Decision variables It includes frequency regulation of variable frequency circulating pump (3) and power regulation of heater (6). The Pareto front solution set is obtained through particle swarm iteration, and the optimal compromise solution is selected as the feedforward control command using fuzzy set theory.

8. The adaptive variable flow photovoltaic module liquid cooling system with coordinated regulation according to claim 1, characterized in that: The PID feedback controller is activated by inputting the current photovoltaic module temperature and a set value. deviation ,according to Calculate the feedback correction amount, where, T pv (t) represents the measured temperature of the photovoltaic module at the current moment. T set Set the temperature value. U fb (t) represents the PID feedback correction value. K p This is the proportionality coefficient. K i The integral coefficient is... K d This represents the differential coefficient.

9. The adaptive variable flow photovoltaic module liquid cooling system with coordinated regulation according to claim 1, characterized in that: The cooling circulation loop is provided with a heat storage condenser (18). The interior of the heat storage condenser (18) is divided into several flow channels (182) by porous foam metal (181). A third V-shaped fin (183) is provided in the flow channel (182). The inlet (184) of the heat storage condenser is connected to the outlet of the biomimetic tree-shaped V-shaped cold plate (23), and the outlet (185) of the heat storage condenser is connected to the inlet of the heat exchanger (4).

10. A method for adaptive variable flow photovoltaic module liquid cooling control with coordinated regulation, characterized in that, The system based on any one of claims 1-9 includes the following steps: Data Acquisition and Preprocessing: Real-time acquisition of temperature in each zone of the photovoltaic module. Photovoltaic module output power Total flow rate of cooling medium Cooling medium inlet temperature Cooling medium outlet temperature Ambient light intensity and ambient temperature ; The collected multi-source data were preprocessed, including using moving average filtering to remove outliers and normalizing the data. Heat load prediction: Construct a feature matrix from normalized time series data according to a fixed time window. ,in, X t In order to be in t The feature matrix constructed at each time step, T t-k:t For the temperature of each zone of the photovoltaic module at tk arrive t Time series data for this time window I rad,t-k:t This is time-series data of ambient light intensity. P PV,t-k: Time series data of photovoltaic module output power. T amb,t-k:t For time series data of ambient temperature, T in,t-k:t This is time-series data of the inlet temperature of the cooling medium. T out,t-k:t The time series data of the cooling medium outlet temperature is used; the feature matrix is ​​input into a pre-trained long short-term memory neural network to predict the heat load and temperature distribution at future times. Feedforward control optimization: A multi-objective particle swarm optimization model is established with the dual objectives of minimizing temperature deviation and system energy consumption. The objective function is... ,in Indicates temperature deviation; Represents system energy consumption; decision variables Includes the frequency of the variable frequency circulating pump (3) and the power of the heater (6), T pv,i Indicates the first Temperature of photovoltaic modules in each zone T opt This indicates the optimal operating temperature for photovoltaic modules. n This indicates the total number of photovoltaic module zones. Q Δ represents the volumetric flow rate of the cooling medium. p This represents the total voltage drop in the circulating circuit. η pump Indicates the efficiency of the variable frequency circulating pump (3). f pump This indicates the operating frequency of the variable frequency circulating pump (3). α The power regulation coefficient of heater (6) is represented by the Pareto front solution set obtained through particle swarm optimization, and the feedforward control command is obtained by solving the Pareto front solution set. ; Feedback correction: Input the deviation between the current component temperature and the set value into the PID feedback controller, and adjust according to the formula. Calculate the feedback correction amount, where, T pv (t) represents the measured temperature of the photovoltaic module at the current moment. T set Indicates the temperature setpoint. U fb (t) represents the PID feedback correction amount. K p Represents the proportionality coefficient. K i Represents the integral coefficient. K d Represents the differential coefficient; The feedforward control command and the feedback control quantity are superimposed to form a composite control signal. After analysis, the frequency of the variable frequency circulating pump (3) and the power of the heater (6) are adjusted to achieve variable flow adaptive cooling.

Images