Temperature equalization control method and device for liquid cooling over-charging system

Abstract

The invention relates to the technical field of temperature equalization control, and discloses a temperature equalization control method and device for a liquid cooling over-charging system. The method comprises the following steps: carrying out grid division on the liquid cooling overcharge system, inputting thermophysical parameters, and obtaining a temperature field calculation matrix through finite difference time domain analysis; establishing a two-layer distributed control network of a station-level control layer and a pile-level control layer based on the temperature field calculation matrix; acquiring real-time temperature data of the two layers of distributed control networks, and performing temperature field mapping to obtain temperature field distribution data; performing linear prediction and nonlinear prediction calculation on the temperature field distribution data to obtain a temperature change trend and a temperature extreme point; the target temperature is calculated according to the temperature change trend and the temperature extreme point, the cascade PID controller is adopted to adjust the rotating speed of the cooling liquid pump, and the temperature balance control parameters are obtained, station-level and pile-level cooperative control is achieved, cooling resources among the multiple charging piles are reasonably distributed, and the overall cooling efficiency of the system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of temperature balance control, and in particular to a temperature balance control method and device for a liquid-cooled supercharging system. Background Art

[0002] With the rapid development of the new energy vehicle industry, supercharging technology has become the key to improving charging efficiency. At present, the supercharging power has reached more than 350kW, which poses a huge challenge to the heat dissipation of the charging system. Traditional air-cooling heat dissipation methods can no longer meet the needs of high-power charging, and liquid-cooled supercharging systems have gradually become the mainstream solution.

[0003] However, the existing liquid-cooled supercharging system has problems such as low temperature control accuracy, slow system response speed, and difficulty in coordinated control of multiple charging piles. Especially during high-power charging, the temperature distribution of key components such as charging guns, charging cables, and interfaces is uneven, which can easily cause local overheating and affect charging safety and efficiency. At the same time, due to the lack of accurate temperature field prediction and real-time monitoring methods, it is difficult to achieve active temperature control of the charging system. Summary of the invention

[0004] The present invention provides a temperature balancing control method and device for a liquid-cooled supercharging system. The present invention realizes coordinated control at the station level and the pile level, so that cooling resources among multiple charging piles are reasonably allocated, thereby improving the overall cooling efficiency of the system.

[0005] In a first aspect, the present invention provides a temperature balancing control method for a liquid-cooled supercharging system, the temperature balancing control method for a liquid-cooled supercharging system comprising: The liquid-cooled supercharging system is meshed and thermophysical parameters are input, and the temperature field calculation matrix is ​​obtained through finite difference time domain analysis; Based on the temperature field calculation matrix, a two-layer distributed control network of a station-level control layer and a pile-level control layer is established; Acquire the real-time temperature data of the two-layer distributed control network, and perform temperature field mapping through heat conduction equation and convection heat transfer calculation to obtain temperature field distribution data; Performing linear prediction and nonlinear prediction calculations on the temperature field distribution data to obtain temperature change trends and temperature extreme points; The target temperature is calculated according to the temperature variation trend and the temperature extreme point, and the rotation speed of the coolant pump is adjusted by using a cascade PID controller to obtain a temperature balance control parameter.

[0006] In a second aspect, the present invention provides a temperature balancing control device for a liquid-cooled supercharging system, the temperature balancing control device for a liquid-cooled supercharging system comprising: The analysis module is used to mesh the liquid-cooled supercharging system and input the thermophysical parameters, and obtain the temperature field calculation matrix through finite difference time domain analysis; An establishment module is used to establish a two-layer distributed control network of a station-level control layer and a pile-level control layer based on the temperature field calculation matrix; A mapping module is used to obtain the real-time temperature data of the two-layer distributed control network, and perform temperature field mapping through heat conduction equation and convection heat transfer calculation to obtain temperature field distribution data; A calculation module, used for performing linear prediction and nonlinear prediction calculation on the temperature field distribution data to obtain temperature change trends and temperature extreme points; The regulating module is used to calculate the target temperature according to the temperature change trend and the temperature extreme point, and use a cascade PID controller to adjust the speed of the coolant pump to obtain a temperature balance control parameter.

[0007] In the technical solution provided by the present invention, by constructing a digital twin model and a finite difference time domain method, high-precision mapping of the temperature field of the charging system is achieved, providing an accurate data basis for the system temperature control; a two-layer distributed control network architecture is adopted to realize station-level and pile-level coordinated control, so that cooling resources among multiple charging piles are reasonably allocated, thereby improving the overall cooling efficiency of the system; a temperature prediction method based on a two-stage prediction model, combined with the advantages of a linear predictor and a nonlinear predictor, improves the accuracy of temperature prediction, and enables the system to respond to temperature changes in advance; a control strategy combining a cascade PID controller with a model predictive control is used to improve the dynamic response characteristics of the system and reduce temperature fluctuations; a two-layer deep reinforcement learning evaluation model is introduced to realize intelligent evaluation and optimization of system performance, and the control strategy is continuously optimized through adaptive learning under dynamic charging conditions; FPGA is used to realize parallel computing, which reduces the computing load, ensures rapid data acquisition and processing, and meets real-time control requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0009] Figure 1 A schematic diagram of a flow chart of a temperature balancing control method for a liquid-cooled supercharging system provided in an embodiment of the present application; Figure 2 A schematic block diagram of the structure of a temperature balancing control device for a liquid-cooled supercharging system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0010] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0011] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change based on actual conditions.

[0012] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0013] It should be further understood that the term “and / or” used in the specification and appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0014] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0015] See also Figure 1 , Figure 1 A flow chart of a temperature balancing control method for a liquid-cooled supercharging system provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, the temperature balancing control method of the liquid-cooled supercharging system provided in the embodiment of the present application includes steps S100 to S600.

[0016] Step S100, meshing the liquid-cooled supercharging system and inputting thermophysical parameters, and obtaining a temperature field calculation matrix through finite difference time domain analysis; It is understandable that the execution subject of the present invention may be a temperature balancing control device for a liquid-cooled supercharging system, or may be a terminal or a server, which is not specifically limited here. The embodiment of the present invention is described by taking a server as the execution subject as an example.

[0017] Specifically, the liquid-cooled supercharging system is meshed, and its key components, including charging guns, charging cables, and charging interfaces, are structurally discretized to form the grid units required for calculation. In this process, combined with the actual physical structure characteristics, ensure that the grid units completely cover the heat conduction path of the entire system, so that subsequent numerical calculations can accurately describe the distribution and changes of heat inside the system. The grid units are arranged with nodes, that is, N, M, and L temperature acquisition nodes are set in the X-axis, Y-axis, and Z-axis directions of the three-dimensional space, respectively, to construct a temperature acquisition node array. The design of the node array needs to comprehensively consider the geometric characteristics, thermophysical properties, and temperature gradient distribution of the liquid-cooled supercharging system to ensure the accuracy of temperature acquisition. The temperature acquisition node array is constructed into a group of differential equations according to the finite difference time domain method, and a three-dimensional temperature field discrete model is established in combination with the thermophysical parameters of the liquid-cooled supercharging system, such as the thermal conductivity, specific heat capacity, density of the material, and the convective heat transfer coefficient of the fluid. The finite difference time domain method converts the continuous heat conduction equation into a discrete numerical calculation form through discretization, so that the dynamic evolution of the temperature field is approximated by iterative calculation. When constructing a three-dimensional temperature field discrete model, ensure the uniformity of the grid division to reduce numerical errors, and pay attention to the accuracy of the thermophysical parameters to ensure that the model truly reflects the heat transfer characteristics of the liquid-cooled supercharging system. The temperature gradient of the three-dimensional temperature field discrete model is calculated, and the grid density is adjusted according to the calculation results. In areas where the temperature changes drastically, the grid is appropriately encrypted to improve the calculation accuracy, while in areas where the temperature changes are relatively stable, the grid density is appropriately reduced to reduce the amount of calculation and improve the overall calculation efficiency. Based on the optimized calculation grid, a finite difference time domain iterative equation is constructed, and the calculation area is divided into multiple parallel calculation units to form a parallel calculation matrix. Since the temperature field calculation of the liquid-cooled supercharging system involves a large number of numerical calculations, the traditional serial calculation method is difficult to meet the needs of efficient real-time calculations. The parallel calculation method is used to effectively improve the calculation speed, so that the system can obtain accurate temperature field distribution in a shorter time. The parallel calculation matrix is ​​mapped to FPGA, and a large number of finite difference iterative operations are allocated to hardware resources for execution through the parallel computing capability of FPGA to achieve efficient temperature field calculation. The temperature field calculation matrix is ​​obtained, which contains the temperature distribution information of the entire liquid-cooled supercharging system and reflects the temperature change trend over time in real time.

[0018] Step S200, establishing a two-layer distributed control network of a station-level control layer and a pile-level control layer based on the temperature field calculation matrix; Specifically, the temperature field calculation matrix is ​​input into the industrial PLC of the dual-core processor to build a station-level control processing unit. The industrial PLC has high reliability and real-time control capabilities. Its dual-core architecture enables the system to simultaneously handle temperature calculation and control instruction distribution tasks, ensuring that temperature balance control can operate efficiently. Inside the PLC, the temperature field calculation matrix is ​​parsed and converted into the data format required for control, so as to facilitate subsequent temperature analysis and adjustment calculations. At the same time, the PLC and external devices establish a communication connection through the RS485 bus to ensure that data can be transmitted stably and with low latency between different control units. In order to improve the system response speed, a data acquisition interval is set at the station-level control layer so that the system obtains temperature data according to a predetermined time period to form a station-level data acquisition system. The station-level data acquisition system is loaded with a distributed temperature balance control algorithm to optimize the temperature and flow of the charging pile coolant. A calculation model for the temperature and flow of the charging pile coolant is established to ensure that the control system makes the best adjustment based on the current temperature state. The establishment process of this calculation model includes multiple aspects. Combined with the data of the temperature field calculation matrix, the temperature distribution of different charging piles is determined, and the heat exchange model of the coolant is constructed based on the theory of heat conduction and fluid dynamics to calculate the temperature change trend of each part of the charging pile. The flow regulation mechanism is introduced to achieve effective intervention in the temperature distribution through precise control of the coolant flow, so that the system can dynamically adjust the flow rate and flow distribution of the coolant according to the temperature change to form a station-level temperature control scheme. The station-level temperature control scheme is imported into the ARM architecture processor to achieve pile-level temperature control. Under the control of the ARM architecture processor, temperature sensors, flow sensors and actuators are configured so that the pile-level control unit can accurately sense and adjust the temperature and flow of the coolant. The temperature sensor collects the temperature data inside and around the charging pile in real time and transmits it to the ARM processor for analysis, while the flow sensor is used to monitor the real-time flow of the coolant to ensure that the fluid distribution meets the set target of the station-level temperature control scheme. The actuator includes a coolant pump and a valve regulation system, whose function is to adjust the flow state of the coolant according to the control calculation results so as to achieve temperature balance in the shortest time. Inside the ARM processor, the data is preprocessed to reduce noise interference and improve the accuracy of temperature regulation. The PWM modulation control program is set for the pile-level control unit to control the speed of the coolant pump within the target range. PWM (pulse width modulation) control achieves stepless control of the motor speed by adjusting the duty cycle, so that the coolant pump automatically adjusts the speed according to the temperature control requirements, thereby accurately controlling the flow rate of the coolant. In this process, the optimal parameters of the PWM modulation signal are calculated based on the flow set value provided by the station-level temperature control solution, so that the coolant pump can provide the best heat dissipation effect while ensuring energy efficiency.A speed feedback mechanism is set up, that is, a speed sensor is added to the control system of the coolant pump so that the ARM processor can monitor the actual operating status of the pump in real time and adjust the PWM parameters according to the feedback data to ensure that the coolant flow rate is always in the ideal range. The pile-level temperature control execution unit and the station-level temperature control solution are distributed and coordinated to form a two-layer distributed control network. The station-level control layer performs macro optimization on the temperature distribution of the entire charging station, while the pile-level control layer accurately adjusts the temperature status of a single charging pile. The two exchange data through the RS485 bus and work together through distributed computing algorithms.

[0019] Step S300, obtaining real-time temperature data of the two-layer distributed control network, and performing temperature field mapping through heat conduction equation and convection heat transfer calculation to obtain temperature field distribution data; Specifically, the temperature sensors are used to collect the temperature of each key part of the charging pile and station-level cooling system. These temperature sensors map the real-time collected temperature data to the spatial position of the three-dimensional grid unit of the liquid-cooled supercharging system to avoid calculation deviations caused by data dislocation or interpolation errors, and obtain preliminary temperature field data. The initial temperature field data is input into the piecewise linear fitting thermal conductivity calculation model for processing. In the liquid-cooled supercharging system, due to the large differences in the thermal conductivity characteristics of different material regions, the traditional uniform thermal conductivity cannot accurately describe the actual thermal conduction behavior. The piecewise linear fitting method is used to subdivide and fit the thermal conductivity according to the material properties, so that the thermal conduction characteristics of each region are more accurately expressed. Based on the thermal conductivity and other physical properties, the calculation model simulates the heat transfer between grid units to obtain the heat conduction distribution data. Based on the heat conduction distribution data, the convective heat transfer coefficient generated by the coolant flow is calculated by the Nusselt criterion. The Nusselt criterion describes the relationship between liquid flow and heat exchange, and determines the convective heat transfer coefficient according to the flow state, flow rate and temperature difference. By calculating the convective heat transfer coefficients of different flow areas, the heat exchange capacity of the coolant at various positions is evaluated. The charging current and power parameters are combined to perform Joule heat calculation, accurately estimate the heat generated during the charging process, obtain the heat source item data, and incorporate it into the system's temperature calculation as a heat source item. The heat source item data is stored in the FPGA onboard memory. FPGA has parallel computing capabilities and can process large amounts of data at the same time. After the data is stored, the data is transmitted to the control system through the DMA (direct memory access) channel to obtain the comprehensive temperature field data. By processing the comprehensive temperature field data, the temperature abnormality area is identified. When the temperature in a certain area exceeds the safety threshold, an early warning signal is automatically triggered so that necessary remedial measures can be taken in time.

[0020] Step S400, performing linear prediction and nonlinear prediction calculations on the temperature field distribution data to obtain the temperature change trend and temperature extreme points; Specifically, a two-stage prediction model is constructed, which includes a linear predictor and a nonlinear predictor. The two predictors each undertake different computing tasks and use different algorithm advantages to accurately predict temperature changes. The linear predictor adopts an autoregressive moving average model structure, in which the autoregressive module and the moving average module work together to model and predict time series. The autoregressive module predicts the temperature value at the current moment based on the autocorrelation of historical temperature data, assuming that the current temperature change is directly related to the temperature change at certain moments in the past. The moving average module smoothes the noise of temperature changes, thereby eliminating short-term fluctuations in the data and making the prediction results more stable. The nonlinear predictor adopts a long short-term memory (LSTM) network structure, which is suitable for processing time series data with long-term dependencies. In the liquid-cooled supercharging system, the temperature change has nonlinear characteristics, such as changes in coolant flow rate, charging power fluctuations, etc., which cause the system temperature to show complex nonlinear changes. The LSTM network captures the long-term dependencies in temperature changes through its special memory units and models these complex nonlinear patterns. The LSTM structure has three layers of networks, each of which adjusts the weight according to the state and input data of the previous moment, thereby continuously optimizing the prediction accuracy of the model. In this way, the nonlinear predictor effectively handles the nonlinear component of temperature change. The prediction results of the linear predictor and the nonlinear predictor are weighted and fused. According to the prediction accuracy, stability and adaptability of each model to the current data set, the appropriate weight coefficient is set, and the linear and nonlinear prediction results are combined to obtain a comprehensive temperature prediction sequence. The temperature change rate and extreme point extraction are performed on the comprehensive prediction sequence. The prediction sequence is differentially operated, that is, the difference between the temperatures at adjacent moments is calculated to evaluate the temperature change rate. The temperature change rate reflects the trend of temperature change and helps to determine whether the temperature control system is in a stable state and whether it needs to be adjusted. The temperature extreme points are extracted, which represent the high and low points in the temperature fluctuation. When the temperature reaches the extreme value, it means that the system is close to the critical value of the operating temperature, triggering the early warning mechanism and taking corresponding adjustment measures.

[0021] Step S500, calculating the target temperature according to the temperature change trend and the temperature extreme point, and using a cascade PID controller to adjust the speed of the coolant pump to obtain a temperature balance control parameter.

[0022] Specifically, a model predictive controller is constructed at the station-level control layer. The controller optimizes the target temperature model by establishing a cost function including a temperature tracking error term, a control quantity change rate constraint term, and a terminal constraint term. The temperature tracking error term is used to measure the deviation between the current temperature and the target temperature. The control quantity change rate constraint term is used to limit the change speed of control quantities such as the coolant pump speed, thereby avoiding excessive control and causing system instability. The terminal constraint term ensures that the system can reach the desired target temperature within the specified time. The introduction of these constraints effectively balances the temperature tracking accuracy and the stability of the system during system operation, and obtains the target temperature optimization model. The temperature change trend and the temperature extreme point are input into the target temperature optimization model, and the optimal control sequence is calculated by the quadratic programming solver. The target temperature control instruction is calculated by predicting the temperature change trend and identifying the temperature extreme point. A cascade PID controller is set at the pile-level control layer to achieve more accurate temperature regulation. The inner and outer loop design of the cascade PID controller enables the system to regulate both flow and temperature at the same time, with the inner loop controlling the flow and the outer loop controlling the temperature. The inner loop of flow control sets the first proportional coefficient, the first integral time and the first differential time, while the outer loop of temperature control sets the second proportional coefficient, the second integral time and the second differential time. The settings of the inner and outer loop PID parameters are precisely adjusted according to the actual working conditions to ensure that the system responds quickly and accurately controls the flow and temperature of the coolant under various operating conditions. In order to improve the stability and accuracy of the system, the parameters of the PID controller are dynamically adjusted. A self-tuning rule based on temperature deviation is established so that the PID controller automatically adjusts the control parameters according to the current temperature error. The adaptive PID controller adjusts its control parameters according to the real-time temperature deviation, thereby ensuring that the system always maintains the best control performance under different working environments and temperature change conditions. The adaptive adjustment rules dynamically optimize the PID parameters according to the temperature fluctuation of the system, avoiding the problems of response lag or over-adjustment caused by the traditional fixed parameter control method. After the target temperature control command is input into the adaptive PID controller, the speed of the coolant pump is precisely controlled by the variable frequency speed regulation technology. The variable frequency speed regulation adjusts the speed of the pump according to the real-time temperature data and control instructions to control the flow rate and temperature of the coolant within the target range. The speed control signal is modulated and output via PWM (Pulse Width Modulation) to accurately adjust the speed of the coolant pump, ensuring that it always maintains the optimal operating point during dynamic adjustment, thereby achieving balanced temperature control.

[0023] A two-layer deep reinforcement learning evaluation model is constructed. The evaluation model consists of two levels: the station-level evaluation layer and the pile-level evaluation layer. The station-level evaluation layer uses a deep deterministic policy gradient network, while the pile-level evaluation layer uses a dual Q-learning network. The design of the two-layer architecture aims to divide the temperature balance control task into multiple decision problems at different levels, so as to optimize them at the station level and the pile level respectively, and improve the control performance of the entire system. Deep deterministic policy gradient is a reinforcement learning algorithm for continuous action space, which can effectively handle continuous variables involved in temperature control, such as the flow rate or speed of the coolant pump. By using a deep deterministic policy gradient network, the station-level controller takes appropriate control actions under different charging conditions according to temperature changes and other related factors to maximize the effect of the temperature control strategy. In the pile-level evaluation layer, a dual Q-learning network is used to solve the over-estimation problem in Q learning and improve the stability of control decisions by reducing the deviation of the value function. In the face of complex charging environments and different charging pile states, the dual Q-learning network more accurately evaluates the value of each action, thereby helping the pile-level control layer to adjust the temperature in a local range. In the application process of temperature balance control, the first temperature response data under static charging conditions is input into the state space of the two-layer deep reinforcement learning evaluation model. The score of the temperature control strategy is calculated through the action value function to generate the evaluation data under static conditions. Under static conditions, the temperature change of the system is relatively stable, and the evaluation of the control strategy mainly depends on the accurate prediction of temperature change and the evaluation of control effect. Therefore, the temperature data in the state space is used to estimate the impact of different control strategies on temperature and help the model select the most suitable control action. When entering the dynamic charging condition, the temperature change of the system is more complicated, and the charging process will be disturbed by factors such as instantaneous load fluctuations and coolant flow rate adjustment. In this case, the temperature response data is passed as input to the reward function of the two-layer deep reinforcement learning evaluation model. Through iterative optimization of the policy network in the model, the system continuously adjusts its temperature control strategy to minimize temperature fluctuations and optimize the cooling effect. The reward function is designed to encourage the system to select control actions that can effectively stabilize the temperature, while suppressing actions that cause excessive temperature fluctuations or overheating. Through this dynamic adjustment, the system maintains good temperature control performance under changing charging loads and environmental conditions. Based on the evaluation data of static and dynamic conditions, the timing features related to temperature rise time, overshoot and steady-state error are extracted. These timing features reflect the response characteristics of the system under different control conditions, helping to evaluate whether the system can effectively reach the target temperature within the specified time, whether there are excessive temperature fluctuations, and the stability of the temperature when reaching steady state. The temperature rise time indicates the time required from the start of charging to reaching the target temperature, the overshoot indicates the extent to which the temperature exceeds the target value, and the steady-state error reflects the difference between the system and the target temperature under long-term operation.The performance evaluation results of the cooling system power consumption, charging efficiency, controller load and communication delay are calculated based on the timing characteristics. In the calculation of the cooling system power consumption, the power consumption of the coolant pump and the change in the coolant flow rate are considered to ensure that the system minimizes energy consumption while maintaining temperature balance. The charging efficiency focuses on the energy transfer efficiency during the charging process to ensure that the charging process is completed within a reasonable time with minimal energy loss. The controller load reflects whether the control system's computing and decision-making capabilities are overloaded during the temperature control process, thereby affecting the control's response speed and accuracy. Communication delay refers to the information transmission time between the various parts of the system. Too long a communication delay causes a system response lag, which in turn affects the temperature control effect.

[0024] In the embodiment of the present invention, by constructing a digital twin model and a finite difference time domain method, high-precision mapping of the temperature field of the charging system is achieved, providing an accurate data basis for system temperature control; a two-layer distributed control network architecture is adopted to achieve coordinated control at the station level and the pile level, so that the cooling resources among multiple charging piles are reasonably allocated, thereby improving the overall cooling efficiency of the system; the temperature prediction method based on the two-stage prediction model combines the advantages of linear predictors and nonlinear predictors to improve the accuracy of temperature prediction, so that the system can respond to temperature changes in advance; through the control strategy combining the cascade PID controller with the model predictive control, the dynamic response characteristics of the system are improved and the temperature fluctuations are reduced; the introduction of a two-layer deep reinforcement learning evaluation model realizes the intelligent evaluation and optimization of the system performance, and the control strategy is continuously optimized through adaptive learning under dynamic charging conditions; the use of FPGA to implement parallel computing reduces the computing load, ensures the rapid acquisition and processing of data, and meets the real-time control requirements.

[0025] In a specific embodiment, the process of executing step S100 may specifically include the following steps: The charging gun, charging cable and charging interface of the liquid-cooled supercharging system are structurally discretized to obtain grid units. Arrange nodes on the grid units, set N, M, and L temperature acquisition nodes in the X-axis, Y-axis, and Z-axis directions respectively, and obtain a temperature acquisition node array; The temperature acquisition node array is used to construct a differential equation group according to the finite difference time domain method, and the thermal physical parameters of the liquid-cooled supercharging system are input to obtain a three-dimensional temperature field discrete model; The temperature gradient is calculated and the mesh density is adjusted for the three-dimensional temperature field discrete model to obtain the optimized computational mesh; The finite difference time-domain iterative equation is constructed based on the optimized computing grid, and the computing area is divided into multiple parallel computing units to obtain a parallel computing matrix. The parallel computing matrix is ​​then mapped using FPGA to obtain a temperature field computing matrix.

[0026] Specifically, when discretizing the structure, the key components of the liquid-cooled supercharging system are divided into multiple small grid units, so that the heat transfer in different areas can be accurately described in subsequent calculations. Assuming that the geometric shapes of the charging gun, charging cable, and charging interface are relatively complex, a non-uniform grid division method is used, so that the grid division is denser in areas with large temperature gradients, and the grid division is sparser in areas where the temperature changes relatively slowly. The size of the grid unit is set to ,in , and Represents the grid spacing in the X-axis, Y-axis and Z-axis directions respectively, discretizing the entire physical space of the liquid cooling supercharging system into a series of grid units. Arrange temperature collection nodes in each grid unit, and set temperature acquisition nodes, build a temperature acquisition node array, record temperature data, and use it to solve the heat transfer equation in the subsequent calculation process. Assume that the temperature distribution of the temperature field is ,in Indicates temperature, is the spatial coordinate, For time, at each acquisition node, the temperature value needs to be continuously updated to reflect the heat transfer process. In order to calculate the evolution of the temperature field, a differential equation system is constructed based on the finite difference time domain method, and the thermophysical parameters of the liquid cooling supercharging system are input. The expression of the heat conduction equation in three-dimensional space is: ; in, Indicates temperature; For time; is the thermal conductivity of the material; is the density of the material; is the specific heat capacity. The equation is discretized using the finite difference method, the time derivative is discretized using the forward difference method, and the space derivative is discretized using the central difference method, and we get: ; in, Represents a grid cell At time step Temperature at Represents a grid cell At time step Temperature at is the thermal diffusivity. After establishing the three-dimensional temperature field discrete model, the temperature gradient is calculated and the grid density is optimized to improve the calculation accuracy. The temperature gradient is calculated by: ; in Represents the temperature gradient. If the temperature gradient in a certain area is large, it means that the temperature in the area changes dramatically, and the grid density needs to be increased to improve the accuracy of the calculation. In areas where the temperature changes more slowly, the grid density is reduced to reduce the amount of calculation and improve the calculation efficiency. The optimized calculation grid is used to construct the finite difference time domain iterative equation, and the calculation area is divided into multiple parallel calculation units to form a parallel calculation matrix. The entire calculation area is divided into multiple independent sub-areas so that the calculation is executed in parallel to improve the calculation efficiency. Each parallel calculation unit processes the data of a part of the grid, and multiple calculation units communicate through boundary conditions to ensure the consistency of the calculation. The parallel calculation matrix is ​​mapped to the FPGA (field programmable gate array) to achieve efficient hardware accelerated calculation and obtain the temperature field calculation matrix. On the FPGA, the core computing unit of the temperature field calculation is executed in parallel through a pipeline structure, so that the finite difference iterative calculation is completed at an extremely fast speed. The DMA (direct memory access) channel of the FPGA is used to transmit data at high speed, so that the calculation results are updated in real time and provided to the temperature balance control system, thereby achieving precise temperature regulation.

[0027] In a specific embodiment, the process of executing step S200 may specifically include the following steps: The temperature field calculation matrix is ​​input into the industrial PLC of the dual-core processor to obtain the station-level control processing unit, and the RS485 bus communication connection is established for the station-level control processing unit, and the data collection interval is set to obtain the station-level data collection system; The distributed temperature balance control algorithm is installed on the station-level data acquisition system, and a calculation model for the temperature and flow of the charging pile coolant is established to obtain the station-level temperature control solution; The station-level temperature control solution is imported into the ARM architecture processor, and the temperature sensor, flow sensor and actuator are configured to obtain the pile-level control unit; A PWM modulation control program is set for the pile-level control unit to control the speed of the coolant pump within a target range, thereby obtaining a pile-level temperature control execution unit; The pile-level temperature control execution unit and the station-level temperature control scheme are distributed and coordinated to obtain a two-layer distributed control network.

[0028] Specifically, the temperature field calculation matrix is ​​input into the industrial-grade PLC (programmable logic controller) of the dual-core processor to form a station-level control processing unit. The temperature field calculation matrix provides the three-dimensional temperature distribution of the entire liquid cooling system, reflecting the temperature gradient and heat transfer of different parts. The PLC adopts a dual-core architecture, one core is used for real-time temperature data acquisition and logic control, and the other core is used to perform calculation tasks to ensure the response speed and stability of the system. An RS485 bus communication connection is established inside the station-level control processing unit, so that the various parts of the entire liquid-cooled supercharging system can exchange data through a stable communication protocol. RS485 is a serial communication protocol used in the field of industrial automation, which supports long-distance data transmission and multi-node communication, and can ensure a stable connection between the station-level control layer and multiple pile-level control units. After the communication connection is completed, the data collection interval is set to ensure that the system collects information such as temperature, flow and charging status at an appropriate frequency to form a station-level data acquisition system. The station-level data acquisition system is loaded with a distributed temperature balancing control algorithm, and a calculation model of the charging pile coolant temperature and flow is constructed based on the temperature field data to obtain a station-level temperature control solution. The distributed temperature balance control algorithm adjusts the coolant flow rate through real-time temperature data to ensure that the temperature of each charging pile is evenly distributed. The calculation model of coolant temperature and flow is composed of the following two equations: The flow heat transfer equation of the coolant is: in, Represents the heat removed by the coolant. represents the mass flow rate of the coolant, is the specific heat capacity of the coolant, and They represent the inlet temperature and outlet temperature of the coolant respectively. The temperature change of the charging pile is described by the heat transfer equation: in, Represents the temperature of the charging pile, is the Joule heat generated during the charging process. The heat removed by the cooling system, is the mass of the charging pile. Through these two equations, the temperature change trend under different coolant flow rates and environmental conditions is calculated, so as to optimize the coolant flow rate control strategy. The station-level temperature control solution is imported into the ARM architecture processor to achieve more refined pile-level temperature control. The ARM architecture processor has efficient computing power and low power consumption characteristics, and is suitable for real-time control systems. Under the control of the ARM architecture, multiple temperature sensors, flow sensors and actuators are configured to ensure that the system monitors and adjusts the flow state of the coolant in real time. The temperature sensor is used to measure the temperature of the charging pile and transmit the data to the processor for calculation, while the flow sensor is used to detect the flow rate of the coolant so as to adjust the parameters of the cooling system in real time. The actuator includes a coolant pump and a valve control device, whose function is to adjust the flow of the coolant according to the calculated target flow rate to maintain the optimal heat dissipation effect. In the pile-level control unit, a PWM (pulse width modulation) modulation control program is set to accurately adjust the speed of the coolant pump to ensure that it works within the target range. PWM modulation is a method of controlling the speed of the motor by adjusting the duty cycle, and its control signal is expressed as: in, represents the average supply voltage of the coolant pump, Represents the maximum voltage of the power supply, Represents the duty cycle of the PWM signal. By adjusting the duty cycle of the PWM signal, the pump speed can be adjusted steplessly so that the flow of the coolant accurately matches the temperature control requirements, thereby achieving the optimal temperature balance control effect. The pile-level temperature control execution unit and the station-level temperature control scheme are distributed and coordinated to form a two-layer distributed control network. At the station-level control layer, the control system analyzes the temperature data of all charging piles and calculates the optimal coolant flow distribution plan. At the pile-level control layer, the execution unit adjusts the speed of the coolant pump in real time according to the station-level instructions to ensure that the temperature of each charging pile is evenly distributed.

[0029] In a specific embodiment, the process of executing step S300 may specifically include the following steps: The temperature sensor data collected in the two-layer distributed control network is spatially mapped to obtain real-time temperature data, and the real-time temperature data is matched with the position information of the grid unit to obtain the initial temperature field data; The initial temperature field data is input into the thermal conductivity calculation model of piecewise linear fitting, and the heat transfer between grid units is calculated through the three-dimensional heat conduction equation to obtain the heat conduction distribution data; Based on the heat conduction distribution data, the convective heat transfer coefficient generated by the coolant flow is calculated by the Nusselt criterion to obtain the convective heat transfer data, and the convective heat transfer data is subjected to Joule heat calculation according to the charging current and power parameters to obtain the heat source term data; The heat source item data is stored in the FPGA onboard memory and transmitted to the control system through the DMA channel to obtain the comprehensive temperature field data. The temperature abnormal area identification and early warning signal triggering judgment are performed on the comprehensive temperature field data to obtain the temperature field distribution data.

[0030] Specifically, high-precision temperature sensors are placed in key parts of the system, such as charging guns, charging cables, and charging interfaces, and grid division rules are followed to ensure that the collected data points can fully cover the entire three-dimensional space. After data collection, set the spatial coordinates The position of each temperature measurement point is calibrated, and the data of the uncovered grid cells are filled with interpolation algorithms to construct the initial temperature field data. The initial temperature field data is input into the thermal conductivity calculation model of piecewise linear fitting to calculate the heat transfer between different grid cells. Since the liquid-cooled supercharging system involves a variety of materials, such as metal conductors, insulating materials, and coolants, the thermal conductivity of each area is There are large differences. The thermal conductivity of different materials is modeled using the piecewise linear fitting method. The thermal conductivity coefficient in a certain area is set to , then the piecewise linear function is expressed as: ; in, Represent the benchmark thermal conductivity in different temperature ranges, is the fitting coefficient, is the temperature, is the temperature dividing point. Through this model, the thermal conductivity is dynamically adjusted in different temperature ranges to improve the accuracy of heat transfer calculation. After obtaining the thermal conductivity, the three-dimensional heat conduction equation is used to calculate the heat transfer between grid units. The basic form of the heat conduction equation is: ; in, is the temperature, For time, is the thermal conductivity, is the material density, is the specific heat capacity. The finite difference method is used to perform numerical solution, and the heat conduction distribution data is calculated to optimize the temperature field calculation. Based on the heat conduction distribution data, the convective heat transfer coefficient generated by the coolant flow is calculated by the Nusselt criterion to analyze the heat dissipation capacity of the cooling system. Characterizes the convective heat transfer intensity, and its calculation formula is: in, is an empirical constant, is the Reynolds number, is the Prandtl number. Convective heat transfer coefficient Calculated by the following formula: in, is the convective heat transfer coefficient, is the thermal conductivity of the coolant, is the characteristic length. After calculating the convective heat transfer data, the Joule heat calculation is performed in combination with the charging current and power parameters to obtain the heat source data of the system. The calculation formula for Joule heat is: in, is Joule heat, is the charging current, is the resistance, for time. The calculation of Joule heat helps determine the main source of temperature rise, and combines the convective heat transfer data to optimize the control strategy of the cooling system. After the heat source term calculation is completed, the data is stored in the FPGA (field programmable gate array) on-board memory to use hardware acceleration for real-time calculation, and the data is transmitted to the control system through the DMA (direct memory access) channel. FPGA quickly processes temperature data through parallel calculations and realizes efficient data transmission to ensure the real-time and stability of the system during the charging process. After obtaining the comprehensive temperature field data, the abnormal temperature area is identified and the early warning signal is triggered to ensure the safety of the charging system. The identification of abnormal areas is based on temperature gradient and historical data analysis. For example, if the temperature change rate in a certain area is too fast or the temperature exceeds the set threshold, a warning signal is triggered, and an adjustment instruction is sent to the control system to increase the coolant flow or reduce the charging power to ensure the stable operation of the system.

[0031] In a specific embodiment, the process of executing step S400 may specifically include the following steps: Construct a two-stage prediction model, which includes a linear predictor and a nonlinear predictor. The linear predictor adopts an autoregressive moving average structure, and the nonlinear predictor adopts a long short-term memory network structure. The linear predictor includes an autoregressive module and a moving average module, and the nonlinear predictor uses a three-layer long short-term memory network structure; The temperature field distribution data is input into the two-stage prediction model, and the prediction results are outputted by the linear predictor and the nonlinear predictor respectively, and the comprehensive prediction sequence is obtained by weighted fusion; The temperature change rate is calculated and the extreme points are extracted for the comprehensive prediction sequence to obtain the temperature change trend and temperature extreme points.

[0032] Specifically, a two-stage prediction model is constructed, which includes a linear predictor and a nonlinear predictor to exploit the characteristics of time series data and provide more accurate temperature predictions. The linear predictor adopts an autoregressive moving average structure, which includes an autoregressive module and a moving average module, while the nonlinear predictor adopts a three-layer long short-term memory (LSTM) network structure to capture long-term dependencies in temperature changes. By combining these two methods, fast and stable prediction results are obtained in a short time scale, while complex temperature change patterns are accurately captured in a long time scale. In the construction of the linear predictor, the autoregressive module uses past temperature data to predict the current temperature, and its mathematical expression is: in, Indicates the current temperature. is the autoregressive order, is the autoregressive coefficient, Indicates the past Temperature data for each time step, is a white noise term. This formula is based on the past The temperature value at each moment, combined with the weight coefficient , perform linear regression calculation to obtain the estimated value of the current temperature. The moving average module smoothes the prediction error to reduce short-term fluctuations in the data. Its mathematical expression is: in, is the moving average order, is the moving average coefficient, Representing the past The prediction error of each time step is calculated. In this way, the moving average module adjusts the prediction results to make them more robust to short-term disturbances. The long short-term memory (LSTM) network is introduced to enhance the system's ability to model long-term time dependencies. The core of the LSTM network lies in its gating mechanism, including the input gate, forget gate, and output gate, which control the updating, forgetting, and output of information respectively, thereby optimizing temperature prediction over a longer time scale. The core calculation process of the LSTM network includes: Forget gate calculation: in, is the output of the forget gate, and are the weight matrix and bias term respectively, and Respectively represent the temperature data and hidden state of the previous moment, Represents the Sigmoid activation function. Input gate calculation: in, The output of the input gate determines whether the current information needs to be stored in the LSTM cell. Cell state update: in, Represents the current cell state, is the cell state at the previous moment. Output gate calculation: Final output: in, is the output of the LSTM network, i.e. the predicted temperature value. The output results of the linear predictor and the nonlinear predictor are weighted and fused to obtain the temperature prediction sequence. The mathematical expression of weighted fusion is as follows: in, is the final predicted temperature value, is the prediction result of the linear predictor, is the prediction result of the LSTM network, and is the weighting coefficient, satisfying The weighting coefficient is determined by historical data training or optimization algorithm to minimize the error of the final prediction result. The temperature change rate is calculated for the comprehensive prediction sequence to analyze the change trend of the system temperature. The temperature change rate is expressed as: in, represents the rate of temperature change, represents the time step. If If it is too large, it means that the system temperature is changing rapidly and the coolant flow rate or charging power needs to be adjusted. At the same time, the extreme point is extracted. By calculating the second-order derivative, the inflection point of the temperature change is determined: If the value of the second-order derivative changes from positive to negative at a certain moment, it means that the temperature has reached the local maximum value. Conversely, if it changes from negative to positive, it means that the temperature has reached the local minimum value. Through calculation, the temperature extreme points are extracted and adjusted in combination with the system control strategy to optimize the temperature balance control of the liquid-cooled supercharging system.

[0033] In a specific embodiment, the process of executing step S500 may specifically include the following steps: A model predictive controller is constructed at the station-level control layer, and a cost function including temperature tracking error term, control quantity change rate constraint term and terminal constraint term is established to obtain the target temperature optimization model; The temperature change trend and temperature extreme point are input into the target temperature optimization model, and the optimal control sequence is calculated by the quadratic programming solver to obtain the target temperature control instruction; A cascade PID controller is set at the pile-level control layer, and the first proportional coefficient, the first integral time and the first differential time of the flow control inner loop, and the second proportional coefficient, the second integral time and the second differential time of the temperature control outer loop are set to obtain PID control parameters; Dynamically tune the PID control parameters, establish the self-tuning rules based on temperature deviation, and obtain the adaptive PID controller; The target temperature control command is input into the adaptive PID controller, and the speed of the coolant pump is controlled to change within the target range through variable frequency speed regulation to obtain a speed control signal, and the speed control signal is modulated and output by PWM to control the dynamic adjustment of the coolant pump speed and obtain the temperature balance control parameter.

[0034] Specifically, a target temperature optimization model is established at the station-level control layer. The model optimizes the temperature regulation strategy of the system by establishing a cost function that includes a temperature tracking error term, a control quantity change rate constraint term, and a terminal constraint term. The temperature tracking error term is used to measure the deviation between the current temperature and the target temperature, the control quantity change rate constraint term is used to limit the rate of change of the coolant pump speed or flow rate to prevent sudden adjustments from causing system instability, and the terminal constraint term ensures that the system can reach the desired target temperature within the specified time range. The cost function is expressed as: ; in, is the objective function, i.e., the temperature optimization target, is the number of time steps in the prediction domain, For the The actual temperature of the time step, For the The target temperature for each time step is is the rate of change of the controlled variable, i.e. the rate of change of the coolant pump speed or flow rate, and are weight parameters, which are used to balance the stability of the control variable change and the accuracy of the final temperature. The temperature change trend and the temperature extreme point are input into the target temperature optimization model, and the optimal control sequence is calculated by the quadratic programming solver to obtain the target temperature control instruction. The quadratic programming solver is used to optimize the control variables while satisfying the constraints. , to minimize the objective function The optimization process is expressed as: ; ; in, Indicates that while optimizing the objective function J, the variables must be satisfied The value range is and Between, and the rate of change Must be in and The constraints between Represents the speed or flow rate of the coolant pump. and are the minimum and maximum allowable values ​​of pump speed or flow, respectively, and are the minimum and maximum change rates allowed, respectively. By solving the above optimization problem, a series of optimal control instructions are obtained, so that the system can achieve optimal temperature regulation in a dynamic environment. A cascade PID controller is set at the pile level control layer to accurately control the flow and temperature of the coolant. The cascade PID controller includes two closed-loop control systems, in which the inner loop controls the flow of the coolant and the outer loop controls the temperature to ensure the stability and rapid response of the system. In the inner loop of the flow control, the first proportional coefficient is set , First integration time and the first differential time , which is used to adjust the response characteristics of the coolant flow rate. In the temperature control outer loop, set the second proportional coefficient , Second integration time and the second differential time , which is used to optimize the dynamic response of temperature. The PID control law is expressed as: in, To control the output, i.e. the coolant pump speed or flow rate regulation signal, is the temperature error, that is, the difference between the target temperature and the current temperature, is the proportional gain, is the integral gain, is the differential gain. The control signal of the inner loop is It is calculated from the flow error and used as the execution variable of the outer loop control. The control signal of the outer loop The temperature error is calculated and used to adjust the flow setting value so that the temperature can reach the target value faster. In order to improve the adaptability of the system, the PID control parameters are dynamically tuned to establish a self-tuning rule based on temperature deviation to obtain an adaptive PID controller. The adaptive PID dynamically adjusts the current temperature error and error change rate. , and To improve the responsiveness of the system. For example, when the temperature error is large, increase the proportional coefficient To improve the response speed, and when the error is small, appropriately reduce the integral coefficient To avoid the integral saturation problem. The adaptive PID parameter adjustment strategy is expressed as: in, is the initial parameter, is the adjustment coefficient. In this way, the PID controller is ensured to maintain good control performance under different temperature change conditions. The target temperature control command is input into the adaptive PID controller, and the speed of the coolant pump is controlled by variable frequency speed regulation technology to change it within the target range to ensure that the coolant flow rate matches the temperature control requirements in real time. Variable frequency speed regulation control is to accurately adjust the speed of the coolant pump by changing the motor input voltage and frequency. Its mathematical expression is: in, is the motor speed, is the frequency of the input voltage, is the input voltage, is the characteristic constant of the motor. The speed control signal is output through PWM modulation to ensure that the coolant pump can operate stably. The duty cycle calculation formula of the PWM signal is: in, represents the duty cycle of the PWM signal, is the output voltage, is the maximum power supply voltage. By adjusting the PWM duty cycle, the speed of the coolant pump is controlled to achieve balanced temperature regulation.

[0035] In a specific embodiment, executing the temperature balancing control method of the liquid-cooled supercharging system further includes the following steps: Construct a two-layer deep reinforcement learning evaluation model. The station-level evaluation layer adopts a deep deterministic policy gradient network, and the pile-level evaluation layer adopts a dual Q-learning network. Applying the temperature balance control parameter, the first temperature response data under the static charging condition is input into the state space of the double-layer deep reinforcement learning evaluation model, and the temperature control strategy score is calculated through the action value function to obtain the static condition evaluation data; Apply the temperature balance control parameters, input the second temperature response data under the dynamic charging condition into the reward function of the double-layer deep reinforcement learning evaluation model, and iteratively optimize the temperature control strategy through the policy network to obtain the dynamic condition evaluation data; The timing characteristics of temperature rise time, overshoot and steady-state error are extracted based on the static operating condition evaluation data and the dynamic operating condition evaluation data, and the performance evaluation results of the cooling system power consumption, charging efficiency, controller load and communication delay are calculated according to the timing characteristics.

[0036] Specifically, a two-layer deep reinforcement learning evaluation model is constructed, so that the system can adaptively adjust the temperature control parameters under complex working conditions to improve the overall cooling efficiency and stability. The model includes a station-level evaluation layer and a pile-level evaluation layer. The station-level evaluation layer adopts a deep deterministic policy gradient network to optimize the global temperature control strategy of the entire charging station, while the pile-level evaluation layer adopts a dual Q-learning network to optimize the local temperature adjustment strategy of a single charging pile. Through a two-layer architecture, the temperature control strategy is dynamically adjusted at both the global and local levels to ensure that the system can achieve optimal temperature balance control under different charging conditions. At the station-level evaluation layer, the deep deterministic policy gradient network uses a policy gradient-based method to optimize the control problem of the continuous action space. The core of this method is to learn a policy function To maximize the cumulative reward. The goal of the deep deterministic policy gradient network is to find an optimal policy so that in any state Next, the actions taken Ability to maximize long-term rewards . Its policy update is obtained by the following gradient calculation: in, is the policy optimization objective function, is the action-value function, are policy network parameters, is the Q network parameter. The calculation of the policy gradient shows that each time the policy is updated, the policy is adjusted according to the gradient of the Q value function so that the actions taken in the future can obtain higher cumulative rewards. At the stake-level evaluation layer, in order to enhance the stability of Q learning, a dual Q learning network is used, which uses two independent Q networks to reduce the over-estimation problem of Q value. The way to update the Q value is as follows: in, Indicates the current status Take action The Q value, is the current reward, is the discount factor Through iterative optimization of reinforcement learning strategies, the system automatically adjusts cooling parameters during dynamic charging to maintain temperature balance. Based on static and dynamic operating condition evaluation data, key timing features are extracted to calculate the overall performance indicators of the system. The temperature rise time is expressed as: in, The system temperature changes from the initial temperature Rise to target temperature The overshoot is a measure of how much the temperature exceeds the target value and is calculated as: in, is the maximum overshoot, Indicates the highest point of the temperature curve. Steady-state error is used to measure the deviation between the final temperature of the system and the target temperature. The calculation formula is: After extracting these timing features, the overall performance indicators of the system are calculated, including cooling system power consumption, charging efficiency, controller load, and communication delay. The cooling system power consumption is expressed as: in, is the cooling system power consumption, is the coolant flow rate, is the change in coolant temperature, is the specific heat capacity of the coolant. The charging efficiency is calculated as follows: in, For charging efficiency, is the effective power received by the battery, is the input power. The controller load and communication delay are measured by the data packet processing time and the instruction transmission time to ensure that the system response speed meets the real-time requirements.

[0037] See also Figure 2 , Figure 2 A schematic block diagram of the structure of the temperature balancing control device 200 for the liquid-cooled supercharging system provided in an embodiment of the present application is shown in FIG. Figure 2 As shown, the temperature balancing control device 200 of the liquid-cooled supercharging system includes: The analysis module 210 is used to mesh the liquid-cooled supercharging system and input thermophysical parameters, and obtain a temperature field calculation matrix through finite difference time domain analysis; Establishing module 220, for establishing a two-layer distributed control network of a station-level control layer and a pile-level control layer based on the temperature field calculation matrix; A mapping module 230 is used to obtain real-time temperature data of the two-layer distributed control network, and perform temperature field mapping through heat conduction equations and convection heat transfer calculations to obtain temperature field distribution data; The calculation module 240 is used to perform linear prediction and nonlinear prediction calculation on the temperature field distribution data to obtain the temperature change trend and the temperature extreme point; The regulating module 250 is used to calculate the target temperature according to the temperature change trend and the temperature extreme point, and use the cascade PID controller to adjust the speed of the coolant pump to obtain the temperature balance control parameter.

[0038] Through the collaborative cooperation of the above-mentioned components, by building a digital twin model and a finite difference time domain method, high-precision mapping of the temperature field of the charging system is achieved, providing an accurate data basis for system temperature control; a two-layer distributed control network architecture is adopted to realize station-level and pile-level collaborative control, so that cooling resources among multiple charging piles are reasonably allocated, and the overall cooling efficiency of the system is improved; the temperature prediction method based on the two-stage prediction model combines the advantages of linear predictors and nonlinear predictors to improve the accuracy of temperature prediction, so that the system can respond to temperature changes in advance; the control strategy that combines the cascade PID controller with model predictive control improves the dynamic response characteristics of the system and reduces temperature fluctuations; the introduction of a two-layer deep reinforcement learning evaluation model realizes intelligent evaluation and optimization of system performance, and continuously optimizes the control strategy through adaptive learning under dynamic charging conditions; the use of FPGA to implement parallel computing reduces the computing load, ensures rapid data acquisition and processing, and meets real-time control requirements.

[0039] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0040] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0041] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A temperature balance control method for a liquid-cooled supercharging system, characterized in that: include: The liquid-cooled supercharging system is meshed and thermophysical parameters are input, and the temperature field calculation matrix is ​​obtained through finite difference time domain analysis; Based on the temperature field calculation matrix, a two-layer distributed control network of a station-level control layer and a pile-level control layer is established; Acquire the real-time temperature data of the two-layer distributed control network, and perform temperature field mapping through heat conduction equation and convection heat transfer calculation to obtain temperature field distribution data; Performing linear prediction and nonlinear prediction calculations on the temperature field distribution data to obtain temperature change trends and temperature extreme points; The target temperature is calculated according to the temperature variation trend and the temperature extreme point, and the rotation speed of the coolant pump is adjusted by using a cascade PID controller to obtain a temperature balance control parameter.

2. The temperature balancing control method of the liquid-cooled supercharging system according to claim 1, characterized in that: The liquid-cooled supercharging system is meshed and thermophysical parameters are input, and a temperature field calculation matrix is ​​obtained by finite difference time domain analysis, including: Performing structural discretization processing on the charging gun, charging cable, and charging interface of the liquid-cooled supercharging system to obtain grid units; Arrange nodes on the grid unit, set N, M, and L temperature acquisition nodes in the X-axis, Y-axis, and Z-axis directions respectively, and obtain a temperature acquisition node array; The temperature acquisition node array is used to construct a differential equation group according to a finite difference time domain method, and the thermophysical parameters of the liquid-cooled supercharging system are input to obtain a three-dimensional temperature field discrete model; Performing temperature gradient calculation and mesh density adjustment on the three-dimensional temperature field discrete model to obtain an optimized calculation mesh; A finite difference time-domain iterative equation is constructed based on the optimized computing grid, the computing area is divided into a plurality of parallel computing units, a parallel computing matrix is ​​obtained, and FPGA mapping processing is performed on the parallel computing matrix to obtain a temperature field computing matrix.

3. The temperature balancing control method of the liquid-cooled supercharging system according to claim 1, characterized in that: The two-layer distributed control network of the station-level control layer and the pile-level control layer is established based on the temperature field calculation matrix, including: Input the temperature field calculation matrix into an industrial-grade PLC of a dual-core processor to obtain a station-level control processing unit, establish an RS485 bus communication connection for the station-level control processing unit, set a data collection interval, and obtain a station-level data collection system; The station-level data acquisition system is loaded with a distributed temperature balancing control algorithm, a calculation model for the temperature and flow of the charging pile coolant is established, and a station-level temperature control solution is obtained; The station-level temperature control solution is introduced into the ARM architecture processor, and a temperature sensor, a flow sensor and an actuator are configured to obtain a pile-level control unit; A PWM modulation control program is set for the pile-level control unit to control the rotation speed of the coolant pump within a target range, thereby obtaining a pile-level temperature control execution unit; The pile-level temperature control execution unit and the station-level temperature control solution are subjected to distributed collaborative computing to obtain a two-layer distributed control network.

4. The temperature balancing control method of the liquid-cooled supercharging system according to claim 1, characterized in that: The real-time temperature data of the two-layer distributed control network is obtained, and the temperature field is mapped by heat conduction equation and convection heat transfer calculation to obtain temperature field distribution data, including: Performing spatial position mapping on the temperature sensor data collected in the two-layer distributed control network to obtain real-time temperature data, and matching the real-time temperature data with the position information of the grid unit to obtain initial temperature field data; The initial temperature field data is input into a piecewise linear fitting thermal conductivity calculation model, and the heat transfer between grid units is calculated by a three-dimensional heat conduction equation to obtain heat conduction distribution data; Based on the heat conduction distribution data, a convective heat transfer coefficient generated by the coolant flow is calculated by the Nusselt criterion to obtain convective heat transfer data, and Joule heat calculation is performed on the convective heat transfer data according to charging current and power parameters to obtain heat source term data; The heat source item data is stored in the FPGA onboard memory and transmitted to the control system through the DMA channel to obtain comprehensive temperature field data, and the temperature abnormal area identification and early warning signal triggering judgment are performed on the comprehensive temperature field data to obtain temperature field distribution data.

5. The temperature balancing control method of the liquid-cooled supercharging system according to claim 1, characterized in that: The linear prediction and nonlinear prediction calculation of the temperature field distribution data are performed to obtain the temperature change trend and the temperature extreme point, including: Constructing a two-stage prediction model, the two-stage prediction model includes a linear predictor and a nonlinear predictor, the linear predictor adopts an autoregressive moving average structure, and the nonlinear predictor adopts a long short-term memory network structure; The linear predictor includes an autoregressive module and a moving average module, and the nonlinear predictor adopts a three-layer long short-term memory network structure; Inputting the temperature field distribution data into the two-stage prediction model, outputting prediction results through the linear predictor and the nonlinear predictor respectively, and obtaining a comprehensive prediction sequence through weighted fusion; The temperature change rate is calculated and the extreme point is extracted for the comprehensive prediction sequence to obtain the temperature change trend and the temperature extreme point.

6. The temperature balancing control method of the liquid-cooled supercharging system according to claim 1, characterized in that: The target temperature is calculated according to the temperature change trend and the temperature extreme point, and the coolant pump speed is adjusted by a cascade PID controller to obtain the temperature balance control parameter, including: A model predictive controller is constructed at the station-level control layer, and a cost function including temperature tracking error term, control quantity change rate constraint term and terminal constraint term is established to obtain the target temperature optimization model; Inputting the temperature variation trend and the temperature extreme point into the target temperature optimization model, calculating the optimal control sequence through a quadratic programming solver, and obtaining a target temperature control instruction; A cascade PID controller is set at the pile-level control layer, and the first proportional coefficient, the first integral time and the first differential time of the flow control inner loop, and the second proportional coefficient, the second integral time and the second differential time of the temperature control outer loop are set to obtain PID control parameters; Dynamically tuning the PID control parameters, establishing a self-tuning rule based on temperature deviation, and obtaining an adaptive PID controller; The target temperature control instruction is input into the adaptive PID controller, and the speed of the coolant pump is controlled to change within the target range through variable frequency speed regulation to obtain a speed control signal, and the speed control signal is modulated and output by PWM to control the dynamic adjustment of the coolant pump speed to obtain the temperature balance control parameter.

7. The temperature balance control method of the liquid-cooled supercharging system according to claim 1, characterized in that: The liquid-cooled supercharging system temperature balancing control method further includes: Constructing a two-layer deep reinforcement learning evaluation model, wherein the station-level evaluation layer adopts a deep deterministic policy gradient network, and the pile-level evaluation layer adopts a dual Q-learning network; Applying the temperature balance control parameter, inputting the first temperature response data under the static charging condition into the state space of the double-layer deep reinforcement learning evaluation model, calculating the temperature control strategy score through the action value function, and obtaining the static condition evaluation data; Applying the temperature balance control parameter, inputting the second temperature response data under the dynamic charging condition into the reward function of the double-layer deep reinforcement learning evaluation model, iteratively optimizing the temperature control strategy through the policy network, and obtaining dynamic condition evaluation data; The timing characteristics of the temperature rise time, overshoot and steady-state error are extracted based on the static operating condition evaluation data and the dynamic operating condition evaluation data, and the performance evaluation results of the cooling system power consumption, charging efficiency, controller load and communication delay are calculated according to the timing characteristics.

8. A temperature balance control device for a liquid-cooled supercharging system, characterized in that: A method for controlling a temperature balance of a liquid-cooled supercharging system according to any one of claims 1 to 7, comprising: The analysis module is used to mesh the liquid-cooled supercharging system and input the thermophysical parameters, and obtain the temperature field calculation matrix through finite difference time domain analysis; An establishment module is used to establish a two-layer distributed control network of a station-level control layer and a pile-level control layer based on the temperature field calculation matrix; A mapping module is used to obtain the real-time temperature data of the two-layer distributed control network, and perform temperature field mapping through heat conduction equation and convection heat transfer calculation to obtain temperature field distribution data; A calculation module, used for performing linear prediction and nonlinear prediction calculation on the temperature field distribution data to obtain the temperature change trend and temperature extreme point; The regulating module is used to calculate the target temperature according to the temperature change trend and the temperature extreme point, and use a cascade PID controller to adjust the speed of the coolant pump to obtain a temperature balance control parameter.

Images