Hybrid power station heat dissipation management system and method

Abstract

The invention relates to the technical field of energy storage power stations, in particular to a hybrid power station heat dissipation management system and method.The hybrid power station heat dissipation management system comprises a power generation and energy storage system and a heat dissipation system, and the power generation and energy storage system comprises a plurality of photovoltaic inverters, a gas turbine and a battery pack; the heat dissipation system comprises a plurality of air cooling motors, liquid cooling heat dissipation assemblies and radiant panels, the photovoltaic inverter, the gas turbine and the battery pack are electrically connected with the control system, the air cooling motors, the liquid cooling heat dissipation assemblies and the radiant panels are electrically connected with the control system, and the control system adjusts and optimizes data of the power generation and energy storage system and then transmits the data to the heat dissipation system. According to the hybrid power station heat dissipation management system and method, a thermodynamic partial differential equation is used as a hard constraint to be embedded into a space-time diagram convolutional network, discrete-continuous hybrid action space design and a dynamic multi-target award function, and air cooling / liquid cooling / radiation collaborative optimization is completed; the energy consumption is reduced by 23.8%, and the temperature difference fluctuation is reduced by 18.5%.

Description

Technical Field

[0001] The present invention relates to the technical field of energy storage power stations, and in particular to a hybrid power station heat dissipation management system and method. Background Art

[0002] A hybrid power plant is a power plant that combines multiple different types of power generation resources or combines power generation and energy storage. The problems of hybrid power plants at this stage are as follows:

[0003] 1. Dynamic fluctuation of heat load: Hybrid power stations (such as solar thermal-gas complementation, battery-diesel hybrid) often face intermittent heat load fluctuations, which makes it difficult for traditional cooling systems to match in real time and reduces efficiency.

[0004] 2. Multiple heat source coupling interference: Different heat sources (such as photovoltaic inverters, gas turbines, and battery packs) have very different thermal characteristics and conflicting heat dissipation requirements, which can easily cause local overheating or energy waste.

[0005] 3. Insufficient environmental adaptability: The heat dissipation performance is attenuated under extreme climates (high temperature, dust), and the traditional air cooling / liquid cooling system has high energy consumption and relies on external resources (such as water resources).

[0006] Current patents such as CN202311572568.6 use phase change heat storage technology to recycle heat from the conversion link of the energy storage system, and adopt a liquid cooling solution. There is no mention of the heat dissipation method between different equipment in the hybrid power station, and the heat dissipation effect is poor; Patent CN202211399301.7 uses liquid cooling to dissipate heat from the battery pack, and does not mention the impact of other heat sources such as photovoltaic inverters and gas turbines.

[0007] To this end, we propose a hybrid power station heat dissipation management system and method. Summary of the invention

[0008] A technical problem to be solved by this application is:

[0009] 1. Dynamic fluctuation of heat load: Hybrid power stations (such as solar thermal-gas complementation, battery-diesel hybrid) often face intermittent heat load fluctuations, which makes it difficult for traditional cooling systems to match in real time and reduces efficiency.

[0010] 2. Multiple heat source coupling interference: Different heat sources (such as photovoltaic inverters, gas turbines, and battery packs) have very different thermal characteristics and conflicting heat dissipation requirements, which can easily cause local overheating or energy waste.

[0011] 3. Insufficient environmental adaptability: The heat dissipation performance is attenuated under extreme climates (high temperature, dust), and the traditional air cooling / liquid cooling system has high energy consumption and relies on external resources (such as water resources).

[0012] Current patents consider liquid cooling, but give less consideration to the impact of other heat sources such as photovoltaic inverters and gas turbines.

[0013] In order to solve the above technical problems, the embodiment of the present application provides a hybrid power station heat dissipation management system, including a power generation and energy storage system and a heat dissipation system, characterized in that: it also includes a control 2 system;

[0014] The power generation and energy storage system includes a certain number of photovoltaic inverters, gas turbines, and battery packs;

[0015] The heat dissipation system includes a number of air-cooled motors, liquid-cooled heat dissipation components and radiation panels;

[0016] The photovoltaic inverter, gas turbine and battery pack are electrically connected to the control system;

[0017] The air-cooled motor, the liquid-cooled heat dissipation component and the radiation plate are electrically connected to the control system;

[0018] The control system adjusts and optimizes the data of the power generation and energy storage system and transmits it to the heat dissipation system to dissipate the heat of the power generation and energy storage system.

[0019] In some embodiments, a PT100 sensor is arranged on the surface of the photovoltaic inverter to collect the surface temperature;

[0020] K-type thermocouples are arranged on the surface of the gas turbine to collect the exhaust gas temperature;

[0021] NTC thermistors are arranged at the battery terminals of the battery pack to collect the battery temperature;

[0022] A Hall encoder is arranged on the surface of the air-cooled motor shaft to collect the air-cooled motor speed data;

[0023] An electromagnetic flowmeter is arranged in the liquid cooling heat dissipation component to collect the liquid cooling flow of the liquid cooling heat dissipation component;

[0024] A photoelectric encoder is arranged on the surface of the radiation plate to collect the opening degree of the radiation plate.

[0025] In some embodiments, the operation process of the control system is as follows:

[0026] S101, data acquisition and synchronization: obtain the thermodynamic state, environmental parameters and control signals of the hybrid power station. The heat source parameters include: photovoltaic inverter surface temperature, battery cell temperature, gas turbine exhaust temperature. The heat dissipation system parameters include: air-cooled motor speed, liquid cooling flow, radiation panel opening; environmental parameters: ambient temperature and humidity, wind speed, dust concentration;

[0027] S201, data preprocessing and feature engineering: clean abnormal data, align spatiotemporal scales, construct high-dimensional physical features, use Z-score filtering to process outliers, perform cross-sensor verification, upsample environmental parameters from .Hz to Hz through cubic spline interpolation, divide the power station into m×m D grids, generate virtual temperature nodes through inverse distance weighted IDW interpolation, extract physical features of the data, such as heat flux density and heat dissipation efficiency, and extract statistical features of the data, such as temperature variance and maximum temperature difference;

[0028] S301, spatiotemporal thermal field prediction: predict the thermal field distribution in the next minute, embed thermodynamic physical constraints, build graph structure, build graph structure of internal equipment, between equipment, and environment interaction, and aggregate hierarchical features of internal equipment, between equipment, and environment interaction, and force the prediction result to satisfy the residual of heat conduction equation;

[0029] S401, multi-objective strategy optimization: generate air cooling / liquid cooling / radiation coordinated control strategies, balance energy consumption, temperature difference, equipment life, design hybrid action space, build dynamic reward function, and consider factors such as equipment energy consumption, equipment temperature, and equipment stress;

[0030] S501, online adaptive compensation: quickly adapt to environmental mutations such as sandstorms and extreme high temperatures, detect environmental mutations, and perform adaptive adjustment and compensation if the ambient temperature change rate is greater than ℃ / min or the dust concentration is greater than μg / m, meta-learning fast fine-tuning, building fine-tuning data sets, calculating meta-gradients, and single-step parameter updates;

[0031] S601, control execution and feedback: execute control instructions and collect feedback data to form a closed-loop optimization. The air cooling speed, liquid cooling pump frequency, and radiation panel angle are adjusted through the PLC controller. When the maximum temperature of the equipment exceeds the critical value, the maximum cooling power is forced, feedback data is collected, and actual temperature changes and energy consumption are recorded.

[0032] In some embodiments, ambient temperature and humidity data are collected by an SHT sensor, wind speed data are collected by an ultrasonic anemometer, and dust concentration is collected by a laser scattering particle sensor;

[0033] All data is synchronized to edge computing nodes via industrial Ethernet, with timestamps aligned to milliseconds and stored as a time series database;

[0034] The collected data set becomes the original data set D raw ={T i (t),ω(t),q(t),θ(t),E env (t)}, where T i is the temperature of the ith sensor, E env is the environmental parameter.

[0035] In some embodiments, Z-score filtering is used to process data outliers, and the mean and standard deviation of the temperature data sliding window are calculated, and outliers with |z| greater than three are removed, using the following formula:

[0036] where u window (t) is the mean value in the window, σ window (t) is the standard deviation within the window;

[0037] The standard for cross-sensor verification is that if the temperature difference between adjacent sensors continues to be >15°C for more than 10 seconds, it is marked as abnormal;

[0038] The formula for generating virtual temperature nodes by inverse distance weighted IDW interpolation is: where d i is the Euclidean distance from the grid point to the i-th sensor, k is the number of nearest neighbor sensors, and the default value is k = 5;

[0039] The heat dissipation density calculation formula is: where k is the thermal conductivity, is the temperature gradient;

[0040] The heat dissipation efficiency calculation formula is: where q cool is the effective heat dissipation, P fan is the air cooling power consumption, P pump is the power consumption of liquid cooling, P rad is the radiation power consumption;

[0041] After data processing is completed, the cleaned and aligned spatiotemporal feature tensor is obtained Where N grid The number of grids, T = 300 means the time step is within 5 minutes, C = 10 means including temperature, heat flow, wind speed, etc.

[0042] In some embodiments, the electrical graph structure construction includes the microscopic scale, i.e., the inside of the device, and the edge weight calculation formula is: where x i 、x j are the coordinates of spatial grid points i and j, ||x i -x j || is the Euclidean distance between two nodes;

[0043] The mesoscale is between devices, and the edge weight calculation formula is: Where Q i , Q j is the heat source power density at grid points i and j;

[0044] The macroscopic scale is the environmental interaction, and the edge weight calculation formula is: Where T i ,T jis the temperature time series of grid points i and j;

[0045] The formula for hierarchical aggregation is Where D s is the degree matrix D s,ii =∑ j A s,ij Used to normalize the adjacency matrix, is the trainable weight matrix corresponding to the lth layer, A s is the adjacency matrix corresponding to the micro / meso / macro scale, H (l) is the node feature matrix of the lth layer, σ is the activation function such as ReLU, which introduces nonlinearity;

[0046] Heat conduction equation residual calculation formula Where λ is the physical loss weight, ρ is the material density, c p is the specific heat capacity, k is the thermal conductivity, q cool,i Effective heat dissipation;

[0047] The predicted thermal field is obtained by spatiotemporal thermal field prediction The format is a grid point temperature matrix.

[0048] In some embodiments, the input data of the multi-objective strategy optimization is the predicted thermal field and device status t Such as battery SOC and gas turbine load.

[0049] The mixed action space includes air cooling speed ω, liquid cooling flow rate q, and radiation opening θ;

[0050] The air cooling speed ω is discrete data, and the air cooling speed ω range is {0,800,1200,1600}RPM. The physical constraint condition of the air cooling speed ω is when T max When the temperature is ≥70℃, it is forbidden to select ω=0;

[0051] The liquid cooling flow rate q is continuous data, and the value range of the liquid cooling flow rate q is [0.2, 2.5] m / s. The physical constraint condition of the liquid cooling flow rate q is that the flow rate change rate is limited to

[0052] The radiation opening θ is discrete data, and the value range of the radiation opening θ is [0, 30%, 60%, 100%]. The physical constraint condition of the radiation opening θ is that when the dust concentration is greater than 50 μg / m, θ ≤ 60%;

[0053] The dynamic reward function formula is R(t) = α(t)R enegery +β(t)R temp +γ(t)R stress, where α(t), β(t), and γ(t) are dynamically adjusted weights, and the adjustment rules are as follows:

[0054] Where SOC(t) is the current state of charge of the battery;

[0055] in To predict the maximum temperature, T safe is the safe temperature, T crit is the critical temperature of the equipment;

[0056] Among them, stress cum The accumulated stress of the equipment, such as the number of motor starts and stops, max is the maximum stress of the equipment;

[0057] Where R enegery is the energy consumption item, and the calculation formula is Where P fan is the real-time power consumption of the air-cooled motor, is the maximum allowable power consumption of the air-cooled motor, P pump The real-time power consumption of the liquid cooling components. is the maximum allowable power consumption of the liquid cooling component, P rad is the real-time power consumption of the radiation panel, is the maximum permissible power consumption of the radiator;

[0058] Where R temp is the temperature difference term, and the calculation formula is In the formula To predict the maximum temperature in the thermal field, T opt is the optimal operating temperature of the device, the denominator 10 is the scaling factor of the temperature difference, which controls the sensitivity of the exponential function;

[0059] Where R stress is the life term, and the calculation formula is The current control action air cooling speed, liquid cooling flow, and radiation opening are the allowable range of the control action, such as air cooling speed 0-1600RPM, and the coefficient 1 / 10 is the normalization factor to balance the impact of action changes of different dimensions on life.

[0060] The multi-objective strategy update mechanism includes physical rule guidance, that is, it is forbidden to reduce the liquid cooling flow in the high-temperature area, the adjacent air-cooling nodes need to adjust the speed synchronously, and the improved proximal strategy optimization algorithm. The calculation formula is as follows:

[0061]

[0062] Where θ is the policy network parameter, η is the learning rate, etc., which controls the parameter update step size. is the probability ratio of the new and old strategies, limiting strategy mutation, A b is the generalized advantage estimate GAE, which measures the advantage of an action relative to the average, λ = 0.95 controls the temporal difference weight, ε is the clipping threshold, and 0.2 is used to enforce r b (θ) prevents excessive updates between [.,.];

[0063] After multi-objective strategy optimization, the control action a can be obtained t =[ω,q,θ], where ω is the air cooling speed, q is the liquid cooling flow rate, and θ is the radiation opening.

[0064] In some embodiments, the online adaptive compensation input real-time data window {X t-w ,...,X t}, where the sliding window length is w=1000;

[0065] Construct a fine-tuning dataset D adapt ={X t-w ,X t};

[0066] The formula for calculating the meta-gradient is: Where w is the sliding window length time step, w =, To predict the temperature, T i real is the actual temperature, To predict the gradient of temperature on the policy network parameters;

[0067] The calculation formula for single-step parameter update is where η meta is the meta-learning rate, which controls the model fine-tuning step size and is set to 0.001;

[0068] Online adaptive compensation in the input real-time data window {X t-w ,...,X t}Get the updated strategy parameters θ'.

[0069] In some embodiments, the control and execution module inputs a control action a t =[ω,q,θ];

[0070] The maximum critical value of temperature is T crit -5℃;

[0071] The maximum cooling power is a emergency =(ω max ,q max ,θ max ),ω max ,q max ,θ max is the maximum fan speed, liquid cooling flow, and radiation opening;

[0072] Control and execution module, input control action a t =[ω,q,θ], and get the updated data set D new , return to S201 and start the next round of optimization cycle.

[0073] In some embodiments, a heat dissipation method for a hybrid power plant is characterized in that: applied to any hybrid power plant heat dissipation management system in claim 1 to 2, the ventilation and heat dissipation method comprises: collecting the thermodynamic state, environmental parameters and control signals of the hybrid plant and integrating them into an original data set D raw ={T i (t),ω(t),q(t),θ(t),E env (t)};

[0074] The spatiotemporal feature tensor is obtained by preprocessing the original data set

[0075] Through the spatiotemporal feature tensor To predict the thermal field distribution in the next minute, embed thermodynamic physical constraints to obtain the predicted thermal field

[0076] By inputting the predicted thermal field and equipment status such as battery SOC and gas turbine load, multi-objective strategy optimization is performed to obtain the control action a t =[ω,q,θ], at this time, the air-cooled motor, liquid-cooled heat dissipation components and radiation plates are adjusted by the control system to dissipate heat and cool down;

[0077] Through the real-time data window {X t-w ,...,X t} Perform online adaptive compensation to obtain the updated strategy parameter θ';

[0078] Controlling actions through input a t =[ω,q,θ], and get the updated data set D new , and perform data preprocessing for the next round of optimization cycle.

[0079] The present invention has at least the following beneficial effects:

[0080] 1. By embedding the thermodynamic partial differential equations as hard constraints into the space-time graph convolutional network, the high-precision prediction of the thermal field of the hybrid power plant is integrated with the physical laws, and a reliable prediction model is obtained in small sample scenarios.

[0081] 2. Through the discrete-continuous hybrid action space design and dynamic multi-objective reward function, the coordinated optimization of air cooling / liquid cooling / radiation was completed, and the comprehensive performance improvement was achieved with energy consumption reduced by 23.8% and temperature difference fluctuation reduced by 18.5%.

[0082] 3. Through sliding window meta-learning and edge-cloud collaborative architecture, rapid adaptation to sudden environmental changes is achieved, and real-time response capabilities for model fine-tuning within 5 minutes are obtained, breaking through the migration bottleneck of traditional reinforcement learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] Figure 1 This is a flow chart of the operation of the control system of the present invention;

[0084] Figure 2 It is a schematic diagram of the overall structural framework of the present invention.

[0085] In the figure: 1-power generation and energy storage system; 2-heat dissipation system; 3-control system; 11-photovoltaic inverter; 12-gas turbine; 13-battery pack; 21-air-cooled motor; 22-liquid-cooled heat dissipation component; 23-radiation plate. DETAILED DESCRIPTION

[0086] 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 only 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.

[0087] Example 1: Please refer to Figure 1-2 ,The present invention provides a technical solution: a hybrid power station heat dissipation management system and method, including a power generation and energy storage system 1, a heat dissipation system 2, characterized in that: it also includes a control system 3;

[0088] The power generation and energy storage system 1 includes a number of photovoltaic inverters 11, a gas turbine 12 and a battery pack 13;

[0089] The heat dissipation system 2 includes a number of air-cooled motors 21, liquid-cooled heat dissipation components 22 and radiation plates 23;

[0090] The photovoltaic inverter 11, the gas turbine 12 and the battery pack 13 are electrically connected to the control system 3;

[0091] The air-cooled motor 21, the liquid-cooled heat dissipation assembly 22 and the radiation plate 23 are electrically connected to the control system 3;

[0092] The control system 3 adjusts and optimizes the data of the power generation and energy storage system 1 and transmits the data to the heat dissipation system 2 to perform heat dissipation processing on the power generation and energy storage system 1 .

[0093] Furthermore, a PT100 sensor is arranged on the surface of the photovoltaic inverter 11 to collect the surface temperature;

[0094] A K-type thermocouple is arranged on the surface of the gas turbine 12 to collect the exhaust gas temperature;

[0095] An NTC thermistor is arranged at the battery terminal of the battery pack 13 to collect the battery temperature;

[0096] A Hall encoder is arranged on the surface of the rotating shaft of the air-cooled motor 21 to collect the speed data of the air-cooled motor 21;

[0097] An electromagnetic flowmeter is arranged inside the liquid cooling heat dissipation component 22 to collect the liquid cooling flow of the liquid cooling heat dissipation component 22;

[0098] A photoelectric encoder is arranged on the surface of the radiation plate 23 to collect the opening degree of the radiation plate 23 .

[0099] Furthermore, the operation process of the control system 3 is as follows:

[0100] S101, data acquisition and synchronization: obtain the thermodynamic state, environmental parameters and control signals of the hybrid power station. The heat source parameters include: the surface temperature of the photovoltaic inverter 11, the temperature of the battery pack 13, the exhaust temperature of the gas turbine 12, and the heat dissipation system 2 parameters include: the speed of the air-cooled motor 21, the liquid cooling flow rate, and the opening of the radiation panel 23; environmental parameters: ambient temperature and humidity, wind speed, and dust concentration;

[0101] S201, Data preprocessing and feature engineering: Clean abnormal data, align spatiotemporal scales, construct high-dimensional physical features, use Z-score filtering to process outliers, perform cross-sensor verification, upsample environmental parameters from 0.1Hz to 1Hz through cubic spline interpolation, divide the power station into 1m×1m 3D grids, generate virtual temperature nodes through inverse distance weighted IDW interpolation, extract physical features of data, such as heat flux density and heat dissipation efficiency, and extract statistical features of data, such as temperature variance and maximum temperature difference;

[0102] S301, spatiotemporal thermal field prediction: predict the thermal field distribution in the next 5 minutes, embed thermodynamic physical constraints, build graph structures, build graph structures for internal equipment, between equipment, and the environment, and aggregate hierarchical features of internal equipment, between equipment, and the environment, forcing the prediction results to satisfy the residual of the heat conduction equation;

[0103] S401, multi-objective strategy optimization: generate air cooling / liquid cooling / radiation coordinated control strategies, balance energy consumption, temperature difference, equipment life, design hybrid action space, build dynamic reward function, and consider factors such as equipment energy consumption, equipment temperature, and equipment stress;

[0104] S501: Online adaptive compensation: Rapidly adapt to sudden environmental changes such as sandstorms and extreme high temperatures, detect sudden environmental changes, and perform adaptive adjustment and compensation if the ambient temperature change rate is greater than 2°C / min or the dust concentration is greater than 200μg / m3, perform meta-learning fast fine-tuning, build fine-tuning data sets, calculate meta-gradients, and update parameters in a single step;

[0105] S601, control execution and feedback: execute control instructions and collect feedback data to form a closed-loop optimization. The air cooling speed, liquid cooling pump frequency, and radiation plate 23 angle are adjusted through the PLC controller. When the maximum temperature of the equipment exceeds the critical value, the maximum cooling power is forced, feedback data is collected, and the actual temperature changes and energy consumption are recorded.

[0106] Furthermore, the ambient temperature and humidity data were collected by SHT35 sensors, the wind speed data were collected by ultrasonic anemometers, and the dust concentration was collected by laser scattering particle sensors;

[0107] All data is synchronized to edge computing nodes via industrial Ethernet, with timestamps aligned to milliseconds and stored as a time series database;

[0108] The collected data set becomes the original data set D raw ={T i (t),ω(t),q(t),θ(t),E env (t)}, where T i is the temperature of the ith sensor, E env is the environmental parameter.

[0109] Furthermore, Z-score filtering is used to process data outliers. The mean and standard deviation of the temperature data sliding window of 60 seconds are calculated, and outliers with |z| greater than three are removed. The formula used is as follows:

[0110] where u window (t) is the mean value in the window, σ window (t) is the standard deviation within the window;

[0111] The standard for cross-sensor verification is that if the temperature difference between adjacent sensors continues to be >15°C for more than 10 seconds, it is marked as abnormal;

[0112] The formula for generating virtual temperature nodes by inverse distance weighted IDW interpolation is: where d i is the Euclidean distance from the grid point to the i-th sensor, k is the number of nearest neighbor sensors, and the default value is k = 5;

[0113] The heat dissipation density calculation formula is: where k is the thermal conductivity, is the temperature gradient;

[0114] The heat dissipation efficiency calculation formula is: where q cool is the effective heat dissipation, P fan is the air cooling power consumption, P pump is the power consumption of liquid cooling, P rad is the radiation power consumption;

[0115] After data processing is completed, the cleaned and aligned spatiotemporal feature tensor is obtained Where N grid The number of grids, T = 300 means the time step is within 5 minutes, C = 10 means including temperature, heat flow, wind speed, etc.

[0116] Furthermore, the graph structure construction includes the microscopic scale, i.e., the inside of the device, and the edge weight calculation formula is where x i 、x j are the coordinates of spatial grid points i and j, ||x i -x j || is the Euclidean distance between two nodes;

[0117] The mesoscale is between devices, and the edge weight calculation formula is: Where Q i , Q j is the heat source power density at grid points i and j;

[0118] The macroscopic scale is the environmental interaction, and the edge weight calculation formula is: Where T i ,T j is the temperature time series of grid points i and j;

[0119] The formula for hierarchical aggregation is Where D s is the degree matrix D s,ii =Σ j A s,ij Used to normalize the adjacency matrix, is the trainable weight matrix corresponding to the lth layer, A s is the adjacency matrix corresponding to the micro / meso / macro scale, H (l) is the node feature matrix of the lth layer, σ is the activation function such as ReLU, which introduces nonlinearity;

[0120] Heat conduction equation residual calculation formula Where λ is the physical loss weight and is set to 0.3, ρ is the material density, and c p is the specific heat capacity, k is the thermal conductivity, q cool,i Effective heat dissipation;

[0121] The predicted thermal field is obtained by spatiotemporal thermal field prediction The format is a grid point temperature matrix.

[0122] Furthermore, the input data of the multi-objective strategy optimization is the predicted thermal field and device status t Such as battery SOC and gas turbine 12 load.

[0123] The mixed action space includes air cooling speed ω, liquid cooling flow rate q, and radiation opening θ;

[0124] The air cooling speed ω is discrete data, and the air cooling speed ω range is {0,800,1200,1600}RPM. The physical constraint condition of the air cooling speed ω is when T max When the temperature is ≥70℃, it is forbidden to select ω=0;

[0125] Liquid cooling flow rate q is continuous data, and the value range of liquid cooling flow rate q is [0.2, 2.5] m 3 / s, the physical constraint of the liquid cooling flow rate q is that the flow rate change rate is limited to

[0126] The radiation opening θ is discrete data, and the value range of the radiation opening θ is [0, 30%, 60%, 100%]. The physical constraint condition of the radiation opening θ is that when the dust concentration is greater than 50μg / m3, θ≤60%;

[0127] The dynamic reward function formula is R(t) = α(t)R enegery +β(t)R temp +γ(t)R stress , where α(t), β(t), and γ(t) are dynamically adjusted weights, and the adjustment rules are as follows:

[0128] Where SOC(t) is the current state of charge of the battery;

[0129] in To predict the maximum temperature, T safe is the safe temperature, T crit is the critical temperature of the equipment;

[0130] Among them, stress cum The accumulated stress of the equipment, such as the number of motor starts and stops, max is the maximum stress of the equipment;

[0131] Where R enegery is the energy consumption item, and the calculation formula is Where P fan is the real-time power consumption of the air-cooled motor 21, is the maximum allowable power consumption of the air-cooled motor 21, P pump is the real-time power consumption of the liquid cooling component 22, is the maximum allowable power consumption of the liquid cooling component 22, P rad is the real-time power consumption of the radiation panel 23, is the maximum allowable power consumption of the radiation panel 23;

[0132] Where R temp is the temperature difference term, and the calculation formula is In the formula To predict the maximum temperature in the thermal field, T opt is the optimal operating temperature of the device, the denominator 10 is the scaling factor of the temperature difference, which controls the sensitivity of the exponential function;

[0133] Where R stress is the life term, and the calculation formula is The current control action air cooling speed, liquid cooling flow, and radiation opening are the allowable range of the control action, such as air cooling speed 0-1600RPM, and the coefficient 1 / 10 is the normalization factor to balance the impact of action changes of different dimensions on life.

[0134] The multi-objective strategy update mechanism includes physical rule guidance, that is, it is forbidden to reduce the liquid cooling flow in the high-temperature area, the adjacent air-cooling nodes need to adjust the speed synchronously, and the improved proximal strategy optimization algorithm. The calculation formula is as follows:

[0135]

[0136] Where θ is the policy network parameter, η is the learning rate such as 0.001 to control the parameter update step size, is the probability ratio of the new and old strategies, limiting strategy mutation, A b is the generalized advantage estimate GAE, which measures the advantage of an action relative to the average, λ = 0.95 controls the temporal difference weight, ε is the clipping threshold, and 0.2 is used to enforce r b (θ) is between [0.8, 1.2] to prevent excessive updates;

[0137] After multi-objective strategy optimization, the control action a can be obtained t =[ω,q,θ], where ω is the air cooling speed, q is the liquid cooling flow rate, and θ is the radiation opening.

[0138] Furthermore, the online adaptive compensation input real-time data window {X t-w ,...,X t}, where the sliding window length is w=1000;

[0139] Construct a fine-tuning dataset D adapt ={X t-w ,X t};

[0140] The formula for calculating the meta-gradient is: Where w is the sliding window length time step, w = 1000, To predict the temperature, T i real is the actual temperature, To predict the gradient of temperature on the policy network parameters;

[0141] The calculation formula for single-step parameter update is where η meta is the meta-learning rate, which controls the model fine-tuning step size and is set to 0.001;

[0142] Online adaptive compensation in the input real-time data window {X t-w ,...,X t}Get the updated strategy parameters θ'.

[0143] Furthermore, the control and execution module inputs a control action a t =[ω,q,θ];

[0144] The maximum critical value of temperature is T crit -5℃;

[0145] The maximum cooling power is a emergency =(ω max ,q max ,θ max ),ω max ,q max ,θ max is the maximum fan speed, liquid cooling flow, and radiation opening;

[0146] Control and execution module, input control action a t =[ω,q,θ], and get the updated data set D new , return to S201 and start the next round of optimization cycle.

[0147] Furthermore, a heat dissipation method of a hybrid power plant is provided, and a heat dissipation management system of a hybrid power plant is applied, wherein the ventilation and heat dissipation method comprises: collecting the thermodynamic state, environmental parameters and control signals of the hybrid plant and integrating them into an original data set

[0148] D raw ={T i (t),ω(t),q(t),θ(t),E env (t)};

[0149] The spatiotemporal feature tensor is obtained by preprocessing the original data set

[0150] Through the spatiotemporal feature tensor To predict the thermal field distribution in the next 5 minutes, embed thermodynamic physical constraints to obtain the predicted thermal field

[0151] By inputting the predicted thermal field and equipment status such as battery SOC and gas turbine 12 load, multi-objective strategy optimization is performed to obtain the control action a t =[ω,q,θ], at this time, the air-cooled motor 21, the liquid-cooled heat dissipation component 22 and the radiation plate 23 are adjusted by the control system 3 to dissipate heat and cool down;

[0152] Through the real-time data window {X t-w ,...,X t} Perform online adaptive compensation to obtain the updated strategy parameter θ';

[0153] Controlling actions through input a t =[ω,q,θ], and get the updated data set D new , and perform data preprocessing for the next round of optimization cycle.

[0154] Combine the following Figure 1-Figure 2 Introducing the hybrid power station heat dissipation management system and method:

[0155] First, a PT100 sensor is arranged on the surface of the photovoltaic inverter 11 to collect the surface temperature, a K-type thermocouple is arranged on the surface of the gas turbine 12 to collect the exhaust temperature, an NTC thermistor is arranged at the battery ear of the battery pack 13 to collect the battery temperature, a Hall encoder is arranged on the shaft surface of the air-cooled motor 21 to collect the speed data of the air-cooled motor 21, an electromagnetic flowmeter is arranged in the liquid-cooled heat dissipation component 22 to collect the liquid cooling flow of the liquid-cooled heat dissipation component 22, a photoelectric encoder is arranged on the surface of the radiation plate 23 to collect the opening of the radiation plate 23, the ambient temperature and humidity data are collected by the SHT35 sensor, the wind speed data are collected by the ultrasonic anemometer, and the dust concentration is collected by the laser scattering particle sensor. These raw data are transmitted to the control system 3 through the sensor, and the control system 3 processes and integrates the received data into D raw ={T i (t),ω(t),q(t),θ(t),E env (t)}, and the spatiotemporal feature tensor is obtained by preprocessing the original data set Through the spatiotemporal feature tensor To predict the thermal field distribution in the next 5 minutes, embed thermodynamic physical constraints to obtain the predicted thermal field By inputting the predicted thermal field and equipment status such as battery SOC and gas turbine 12 load, multi-objective strategy optimization is performed to obtain the control action a t =[ω,q,θ], at this time, the air-cooled motor 21, the liquid-cooled heat dissipation component 22 and the radiation plate 23 are adjusted by the control system 3 to dissipate heat and cool down, and the real-time data window {X t-w,...,X t} Perform online adaptive compensation to obtain the updated strategy parameter θ', and input the control action a t =[ω,q,θ], and get the updated data set D new , and perform data preprocessing for the next round of optimization cycle to obtain a new control action a t =[ω,q,θ], the air-cooled motor 21, the liquid-cooled heat dissipation component 22 and the radiation plate 23 are adjusted by the control system 3 to achieve the effect of heat dissipation and temperature reduction.

[0156] The effect of this method compared with other methods is shown in the following table:

[0157] index Traditional PID control This method Improvement effect Average daily energy consumption 1450kWh 1102kWh 23.8% reduction Maximum temperature difference 17.3℃ 14.1℃ 18.5% reduction Equipment stress fluctuations 0.78 0.52 33.3% reduction

[0158] It can be seen from the above table that compared with the traditional method, this method has smaller daily average energy consumption, maximum temperature difference, equipment stress fluctuation, and better heat dissipation effect.

[0159] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0160] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. A hybrid power station heat dissipation management system, comprising a power generation and energy storage system (1) and a heat dissipation system (2), characterized in that: Also includes a control system (3); The power generation and energy storage system (1) comprises a plurality of photovoltaic inverters (11), a gas turbine (12) and a battery pack (13); The heat dissipation system (2) comprises a number of air-cooled motors (21), liquid-cooled heat dissipation components (22) and radiation plates (23); The photovoltaic inverter (11), the gas turbine (12) and the battery pack (13) are electrically connected to the control system (3); The air-cooled motor (21), the liquid-cooled heat dissipation component (22) and the radiation plate (23) are electrically connected to the control system (3); The control system (3) adjusts and optimizes the data of the power generation and energy storage system (1) and transmits the data to the heat dissipation system (2) to perform heat dissipation processing on the power generation and energy storage system (1).

2. The hybrid power station heat dissipation management system according to claim 1, characterized in that: A PT100 sensor is arranged on the surface of the photovoltaic inverter (11) to collect the surface temperature; A K-type thermocouple is arranged on the surface of the gas turbine (12) to collect the exhaust gas temperature; An NTC thermistor is arranged at the battery tab of the battery pack (13) to collect the battery temperature; A Hall encoder is arranged on the surface of the rotating shaft of the air-cooled motor (21) to collect the rotating speed data of the air-cooled motor (21); An electromagnetic flowmeter is arranged in the liquid-cooling heat dissipation component (22) to collect the liquid cooling flow of the liquid-cooling heat dissipation component (22); A photoelectric encoder is arranged on the surface of the radiation plate (23) to collect the opening degree of the radiation plate (23).

3. The hybrid power station heat dissipation management system according to claim 1, characterized in that: The operation process of the control system (3) is as follows: S101, data acquisition and synchronization: obtaining the thermodynamic state, environmental parameters and control signals of the hybrid power station, the heat source parameters include: the surface temperature of the photovoltaic inverter (11), the cell temperature of the battery pack (13), the exhaust temperature of the gas turbine (12), the heat dissipation system (2) parameters include: the speed of the air-cooled motor (21), the liquid cooling flow rate, the opening of the radiation panel (23); environmental parameters include: ambient temperature and humidity, wind speed, and dust concentration; S201, Data preprocessing and feature engineering: Clean abnormal data, align spatiotemporal scales, construct high-dimensional physical features, use Z-score filtering to process outliers, perform cross-sensor verification, upsample environmental parameters (0.1Hz) to 1Hz through cubic spline interpolation, divide the power station into 1m×1m 3D grids, generate virtual temperature nodes through inverse distance weighted (IDW) interpolation, extract physical features of the data, such as heat flux density and heat dissipation efficiency, and extract statistical features of the data, such as temperature variance and maximum temperature difference; S301, spatiotemporal thermal field prediction: predict the thermal field distribution in the next 5 minutes, embed thermodynamic physical constraints, build graph structures, build graph structures for internal equipment, between equipment, and the environment, and aggregate hierarchical features of internal equipment, between equipment, and the environment, forcing the prediction results to satisfy the residual of the heat conduction equation; S401, multi-objective strategy optimization: generate air cooling / liquid cooling / radiation coordinated control strategies, balance energy consumption, temperature difference, equipment life, design hybrid action space, build dynamic reward function, and consider factors such as equipment energy consumption, equipment temperature, and equipment stress; S501: Online adaptive compensation: Rapidly adapt to sudden environmental changes (such as sandstorms, extreme high temperatures), detect sudden environmental changes, and perform adaptive adjustment and compensation if the ambient temperature change rate is greater than 2°C / min or the dust concentration is greater than 200μg / m3, perform meta-learning fast fine-tuning, build fine-tuning data sets, calculate meta-gradients, and update parameters in a single step; S601, control execution and feedback: execute control instructions and collect feedback data to form a closed-loop optimization. The air cooling speed, liquid cooling pump frequency, and radiation plate (23) angle are adjusted through the PLC controller. When the maximum temperature of the equipment exceeds the critical value, the maximum cooling power is forced. Feedback data is collected to record the actual temperature change and energy consumption.

4. The hybrid power station heat dissipation management system according to claim 3, characterized in that: The environmental temperature and humidity data are collected by an SHT35 sensor, the wind speed data are collected by an ultrasonic anemometer, and the dust concentration is collected by a laser scattering particle sensor; All data is synchronized to edge computing nodes via industrial Ethernet, with timestamps aligned to milliseconds and stored as a time series database; The collected data set becomes the original data set D raw ={T i (t),ω(t),q(t),θ(t),E env (t)}, where T i is the temperature of the ith sensor, E env is the environmental parameter.

5. The hybrid power station heat dissipation management system according to claim 3, characterized in that: Use Z-score filtering to process data outliers, calculate the mean and standard deviation of the temperature data sliding window (60 seconds), and remove outliers with |z| greater than three. The formula used is as follows: where u window (t) is the mean value in the window, σ window (t) is the standard deviation within the window; The standard for cross-sensor verification is that if the temperature difference between adjacent sensors continues to be >15°C for more than 10 seconds, it is marked as abnormal; The formula for generating virtual temperature nodes by inverse distance weighted (IDW) interpolation is: where d i is the Euclidean distance from the grid point to the i-th sensor, k is the number of nearest neighbor sensors (default k = 5); The heat dissipation density calculation formula is: where k is the thermal conductivity, is the temperature gradient; The heat dissipation efficiency calculation formula is: where q cool is the effective heat dissipation, P fan is the air cooling power consumption, P pump is the power consumption of liquid cooling, P rad is the radiation power consumption; After data processing is completed, the cleaned and aligned spatiotemporal feature tensor is obtained Where N grid The number of grids, T = 300 means the time step is within 5 minutes, C = 10 means including temperature, heat flow, wind speed, etc.

6. The hybrid power station heat dissipation management system and method according to claim 3, characterized in that: The graph structure construction includes the microscopic scale, i.e., the inside of the device, and the edge weight calculation formula is: where x i 、x j are the coordinates of spatial grid points i and j, ||x i -x j || is the Euclidean distance between two nodes; The mesoscale is between devices, and the edge weight calculation formula is: Where Q i , Q j is the heat source power density at grid points i and j; The macroscopic scale is the environmental interaction, and the edge weight calculation formula is: Where T i ,T j is the temperature time series of grid points i and j; The formula for the hierarchical aggregation is Where D s is the degree matrix D s,ii =∑ j A s,ij Used to normalize the adjacency matrix, is the trainable weight matrix corresponding to the lth layer, A s is the adjacency matrix of the corresponding scale (micro / meso / macro), H (l) is the node feature matrix of the lth layer, σ is the activation function (such as ReLU), which introduces nonlinearity; The heat conduction equation residual calculation formula Where λ is the physical loss weight, which is 0.3, ρ is the material density, c p is the specific heat capacity, k is the thermal conductivity, q cool,i Effective heat dissipation; The spatiotemporal thermal field prediction obtains a predicted thermal field The format is a grid point temperature matrix.

7. A hybrid power station heat dissipation management system and method according to claim 3, characterized in that: The multi-objective strategy optimizes the input data to predict the thermal field and device status t Such as battery SOC, the gas turbine (12) load. The hybrid action space includes air cooling speed ω, liquid cooling flow rate q, and radiation opening θ; The air cooling speed ω is discrete data, the air cooling speed ω range is {0, 800, 1200, 1600} RPM, and the physical constraint condition of the air cooling speed ω is when T max When the temperature is ≥70℃, it is forbidden to select ω=0; The liquid cooling flow rate q is continuous data, and the value range of the liquid cooling flow rate q is [0.2, 2.5] m 3 / s, the physical constraint condition of the liquid cooling flow rate q is that the flow rate change rate is limited to The radiation opening θ is discrete data, and the value range of the radiation opening θ is [0, 30%, 60%, 100%]. The physical constraint condition of the radiation opening θ is that when the dust concentration is greater than 50 μg / m3, θ≤60%; The dynamic reward function formula is R(t)=α(t)R enegery +β(t)R temp +γ(t)R stress , where α(t), β(t), and γ(t) are dynamically adjusted weights, and the adjustment rules are as follows: Where SOC(t) is the current state of charge of the battery; in To predict the maximum temperature, T safe is the safe temperature, T crit is the critical temperature of the equipment; Among them, stress cum is the cumulative stress of the equipment (such as the number of motor starts and stops), stress max is the maximum stress of the equipment; Where R enegery is the energy consumption item, and the calculation formula is Where P fan is the real-time power consumption of the air-cooled motor (21), is the maximum allowable power consumption of the air-cooled motor (21), P pump is the real-time power consumption of the liquid cooling component (22), is the maximum allowable power consumption of the liquid cooling heat dissipation component (22), P rad is the real-time power consumption of the radiation panel (23), is the maximum allowable power consumption of the radiation panel (23); Where R temp is the temperature difference term, and the calculation formula is In the formula To predict the maximum temperature in the thermal field, T opt is the optimal operating temperature of the device, the denominator 10 is the scaling factor of the temperature difference, which controls the sensitivity of the exponential function; Where R stress is the life term, and the calculation formula is The current control action (air cooling speed, liquid cooling flow, radiation opening) is the allowable range of the control action (such as air cooling speed 0-1600RPM), and the coefficient 1 / 10 is the normalization factor to balance the impact of action changes of different dimensions on life. The multi-objective strategy update mechanism includes physical rule guidance, that is, it is forbidden to reduce the liquid cooling flow in the high-temperature area, the adjacent air-cooling nodes need to adjust the speed synchronously, and the improved proximal strategy optimization algorithm. The calculation formula is as follows: Where θ is the policy network parameter, η is the learning rate (such as 0.001) to control the parameter update step size, is the probability ratio of the new and old strategies, limiting strategy mutation, A b is the generalized advantage estimate (GAE), which measures the advantage of an action relative to the average, λ=0.95 controls the temporal difference weight, ε is the clipping threshold, and 0.2 is used to enforce r b (θ) is between [0.8, 1.2] to prevent excessive updates; After the multi-objective strategy is optimized, the control action a can be obtained. t =[ω,q,θ], where ω is the air cooling speed, q is the liquid cooling flow rate, and θ is the radiation opening.

8. A hybrid power station heat dissipation management system and method according to claim 5, characterized in that: The online adaptive compensation input real-time data window {X t-w ,...,X t }, where the sliding window length is w=1000; The fine-tuning dataset is constructed as D adapt ={X t-w ,X t }; The meta-gradient calculation formula is: Where w is the sliding window length (time steps), w = 1000, To predict the temperature, T i real is the actual temperature, To predict the gradient of temperature on the policy network parameters; The calculation formula for the single-step parameter update is: where η meta is the meta-learning rate, which controls the model fine-tuning step size and is set to 0.001; The online adaptive compensation is performed in the input real-time data window {X t-w ,...,X t }Get the updated policy parameters θ'.

9. A hybrid power station heat dissipation management system and method according to claim 3, characterized in that: The control and execution module inputs a control action t =[ω,q,θ]; The maximum critical value of the temperature is T crit -5℃; The maximum cooling power is a emergency =(ω max ,q max ,θ max ),ω max ,q max ,θ max is the maximum fan speed, liquid cooling flow, and radiation opening; The control and execution module inputs a control action t =[ω,q,θ], and get the updated data set D new , return to S201 and start the next round of optimization cycle.

10. A heat dissipation method for a hybrid power station, characterized in that: The heat dissipation management system for a hybrid power plant according to any one of claims 1 to 9, wherein the ventilation and heat dissipation method comprises: collecting the thermodynamic state, environmental parameters and control signals of the hybrid plant and integrating them into an original data set D raw ={T i (t),ω(t),q(t),θ(t),E env (t)}; The spatiotemporal feature tensor is obtained by preprocessing the original data set Through the spatiotemporal feature tensor To predict the thermal field distribution in the next 5 minutes, embed thermodynamic physical constraints to obtain the predicted thermal field By inputting the predicted thermal field, and the equipment status such as the battery SOC and the load of the gas turbine (12), a multi-objective strategy optimization is performed to obtain the control action a t =[ω, q, θ], at which time the air-cooled motor (21), the liquid-cooled heat dissipation component (22) and the radiation plate (23) are adjusted by the control system (3) to dissipate heat and cool down; Through the real-time data window {X t-w ,...,X t } Perform online adaptive compensation to obtain the updated strategy parameter θ'; Controlling actions through input a t =[ω,q,θ], and get the updated data set D new , and perform data preprocessing for the next round of optimization cycle.

Images