Thermal management system and method for wind driven generator

Abstract

The invention relates to the technical field of wind driven generator control, in particular to a thermal management system and method for a wind driven generator, and the system comprises a data collection layer, a prediction module, a digital twin model, an intelligent control and diagnosis engine, and a heat dissipation execution mechanism. In the prior art, an open-loop or simple feedback control scheme based on a real-time temperature threshold value is mainly adopted, the problem of response lag exists, and all heat dissipation parts always work independently and lack cooperation, so that the energy efficiency is low; according to the invention, a closed-loop intelligent thermal management scheme based on a digital twin model is adopted, and real-time simulation and future state prediction are carried out by constructing virtual mapping of a generator thermal system, so that a control system can actively and cooperatively manage all heat dissipation resources according to a simulation result; the method has the core advantage of converting passive into active, and can intervene in advance before excessive accumulation of heat, so that collaborative optimal control of all parts of the system is realized, and the accuracy, timeliness and overall energy efficiency of heat management are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of wind turbine control technology, and in particular to a wind turbine thermal management system and method. Background Technology

[0002] The thermal management system for wind turbine generators is a key technology system for ensuring the safe and stable operation of wind turbine generator sets. Its core task is to effectively manage the heat generated by high-heat components such as generators, gearboxes, and converters during operation, preventing equipment performance degradation or damage due to overheating. With the continuous increase in the single-unit capacity of wind turbine generator sets, the heat generation has also increased significantly, placing higher demands on the heat dissipation efficiency, reliability, and energy efficiency of the thermal management system.

[0003] Existing thermal management solutions, such as single air-cooling systems, independent and decentralized water-cooling systems, or simple combined heat dissipation, generally have significant limitations. These technologies are mostly passive response controls, relying on judgment of the current temperature threshold, making it difficult to anticipate future changes in operating conditions, resulting in response lag. Each heat dissipation component often operates independently, lacking coordination, leading to low overall system energy efficiency. Furthermore, traditional solutions are insufficiently adaptable to complex operating conditions such as extreme weather and sudden load changes, and generally lack effective fault diagnosis and fault-tolerant operation mechanisms, resulting in low energy utilization and failure to effectively recover and utilize waste heat. Control strategies are also relatively rigid, unable to precisely adjust according to the dynamic changes in the three-dimensional temperature field inside the equipment.

[0004] This invention proposes a wind turbine thermal management system and method to systematically solve the aforementioned problems. The invention introduces digital twin technology to create a high-fidelity virtual mapping of the physical heat generation and dissipation system. By integrating future weather forecasts and power load prediction data, the system can proactively simulate the processes of heat generation, transfer, and dissipation, and dynamically generate the optimal heat dissipation strategy based on a multi-objective optimization algorithm, achieving a fundamental shift from "passive response" to "active prediction and optimization." Simultaneously, by fusing measured data from temperature sensors with simulated data from the digital twin model, and combining this with acoustic signature data for correlation analysis, the system achieves early and accurate diagnosis of potential faults and model self-calibration, thereby triggering fault-tolerant control strategies and significantly improving the system's reliability and intelligence level. Summary of the Invention

[0005] To overcome the problems mentioned in the background art, the present invention proposes a thermal management system and method for wind turbine generators.

[0006] The technical solution of this invention is: a wind turbine thermal management system, comprising: The data acquisition layer includes a temperature sensor array deployed on the heat-generating components and heat dissipation circuit of the wind turbine generator, and an acoustic fingerprint sensor for collecting generator operating noise. The forecasting module is used to obtain future weather forecast data for the location of the wind turbine and future power load forecast data for the turbine. The digital twin model is a virtual mapping of the heat generation and heat dissipation system of a wind turbine. It has a built-in thermodynamic simulation engine to simulate and predict the heat generation, transfer, accumulation and heat dissipation process inside the generator based on the generator's real-time operating status data, the status of the heat dissipation system actuators and the prediction data from the prediction module. The intelligent control and diagnostic engine is used to optimize the control of heat dissipation schemes based on simulation results of digital twin models, and to perform fault diagnosis and model calibration based on multi-source data fusion. The heat dissipation actuator receives and executes control commands issued by the heat dissipation optimization control unit to adjust the operating status of the heat dissipation system.

[0007] As a preferred option, the intelligent control and diagnostic engine specifically includes: A11: Heat dissipation optimization control unit, used to dynamically generate the optimal heat dissipation control strategy based on the simulation prediction results of the digital twin model, with the optimization goal of balancing heat distribution and controlling the temperature of key points to not exceed the limit; A12: Multi-source data fusion fault diagnosis unit, used to compare the measured temperature data collected by the temperature sensor group with the simulated temperature data corresponding to the digital twin model in real time, and to analyze the voiceprint data collected by the voiceprint sensor in order to realize fault detection and model calibration.

[0008] Preferably, the prediction module specifically includes: A21: Meteorological data acquisition unit, used to access forecast data streams from external meteorological services and local meteorological stations to acquire future weather forecast data for the location of the wind turbine. A22: Power load forecasting unit, used to generate future power load forecasting data for wind turbines based on historical operating data and grid dispatching plans; A23: Predictive data fusion and processing unit, used to perform time alignment, format standardization and reliability verification on the acquired future weather forecast data and future power load forecast data, and output a structured future operating condition prediction sequence.

[0009] As a preferred option, digital twin models specifically include: A31: Three-dimensional geometry and mesh model, which is the three-dimensional geometric model and computational mesh generation for the physical entities in a wind turbine that require thermal management; A32: Material and physical property parameter library, which stores and defines the physical property parameters of the materials used in the various components that constitute a physical entity, including density, specific heat capacity and thermal conductivity; A33: A thermodynamics simulation engine that loads and runs 3D geometric and mesh models, and calls upon material and physical property parameter libraries to simulate the heat transfer process inside physical entities by solving energy conservation and heat transfer control equations.

[0010] Preferably, the heat dissipation optimization control unit specifically includes: A111: Optimization target setting module, used to set a comprehensive optimization objective function that includes three sub-objectives; A112: Control strategy generation module, used to generate a set of candidate heat dissipation control strategies covering different heat dissipation intensities based on the current state and prediction information; A113: Forward simulation and evaluation module, used to call digital twin model to perform forward simulation prediction of candidate heat dissipation control strategies, and evaluate and score the simulation results based on the optimization objective function; A114: Optimal Decision and Output Module, used to select the candidate strategy with the highest evaluation score as the optimal control strategy and convert the strategy into a specific control command output.

[0011] Preferably, the heat dissipation optimization control unit is also used for forward-looking control using predictive data, the specific process of which is as follows: S11: Predictive state assessment, receiving a predicted data sequence for the future control cycle from the prediction module, including the predicted ambient temperature sequence and the predicted heat source power sequence; S12: Forward control mode decision, based on the predicted data sequence, judges the future operating condition change trend and determines the forward control mode according to predefined rules; S13: Generate a forward-looking control strategy. In the control strategy generation module of the heat dissipation optimization control unit, the predicted data sequence is used as the boundary condition input, and the generation rules of the candidate control strategy and the weight coefficients in the optimization objective function are set or adjusted according to the determined forward-looking control mode. S14: Forward Simulation and Decision Making. The forward simulation and evaluation module evaluates the comprehensive performance of candidate strategies over the entire prediction time domain based on a complete simulation containing future prediction data. The optimal decision making and output module selects and outputs the optimal forward control strategy.

[0012] As a preferred option, the rules for forward-looking control mode decision-making specifically include: If the forecast data indicates that the ambient temperature will continue to rise, the wind speed will continue to decrease, and the heat source power will continue to increase during the future control period, then it will be determined as a heat-prone operating condition, and the preventive enhanced heat dissipation mode will be activated. If the forecast data indicates that the ambient temperature will continue to decrease and the heat source power will continue to decrease during the future control period, it will be determined as a cooling condition, and the economical heat dissipation mode will be activated. If the forecast data does not show a clear trend of continued heating or cooling, then the regular optimization mode will be activated.

[0013] As a preferred embodiment, the multi-source data fusion fault diagnosis unit specifically includes: A41: Real-time data synchronization and comparison module, used to receive the measured temperature data stream of the temperature sensor group after time synchronization and the simulated temperature data stream output by the digital twin model, and calculate the real-time deviation of key measuring points. A42: Anomaly detection and diagnosis trigger module, used to monitor real-time deviations and activate the fault diagnosis process when preset anomaly trigger conditions are met; A43: Multi-source feature extraction and fusion analysis module, used to synchronously retrieve the voiceprint sensor data within the corresponding time period after the diagnostic process is activated, and extract temperature abnormality features and voiceprint abnormality features for correlation fusion analysis; A44: Fault Reasoning and Output Module, with a built-in fault feature knowledge base, is used to match fault modes based on the fusion analysis results and output diagnostic conclusions and confidence levels.

[0014] Preferably, the diagnostic conclusions of the fault reasoning and output module will be input as feedback signals to the heat dissipation optimization control unit. This is used to adjust the control strategy for fault-tolerant operation after a specific fault is confirmed. The specific process is as follows: S21: Diagnostic conclusion reception and parsing. The heat dissipation optimization control unit receives diagnostic conclusions from the fault reasoning and output module in real time, including fault type identification, fault component location and fault severity level. S22: Fault-control mode mapping, which maps the specific fault type and severity received to a specific fault-tolerant operation control mode according to predefined mapping rules; S23: The control strategy is adaptively adjusted. Based on the fault-tolerant operation control mode obtained by mapping, the objective function, constraints of the target setting module, and candidate strategy generation rules of the control strategy generation module are dynamically adjusted and optimized. S24: Fault-tolerant strategy execution and monitoring. Based on the adjusted objectives and constraints, regenerate and select the optimal control strategy, issue it to the heat dissipation actuator for execution, and strengthen the monitoring of fault-related parameters.

[0015] A method for thermal management of a wind turbine generator includes the following steps: S31: Real-time temperature data and operating acoustic data of the wind turbine's heating components and heat dissipation circuit are acquired through the data acquisition layer, and future weather forecast data and future power load forecast data of the generator are acquired through the prediction module. S32: Input real-time data and predicted data into the digital twin model. The digital twin model uses its thermodynamic simulation engine to simulate and predict the heat generation, transfer, accumulation and heat dissipation process inside the generator under the future operating conditions corresponding to the predicted data, and generate simulation results including the future temperature field and heat flow distribution. S33: The intelligent control and diagnostic engine executes the following parallel sub-steps: S331: Heat dissipation optimization control aims to balance heat distribution, control the temperature of key points to keep them within limits, and optimize energy consumption. Based on the simulation results of the digital twin model, it dynamically generates the optimal heat dissipation control strategy and sends it to the heat dissipation actuator. S332: Multi-source data fusion fault diagnosis, real-time comparison of measured data from temperature sensors with simulated data from digital twin models, calculation of deviation, and when the deviation is continuously abnormal, feature fusion analysis is performed in conjunction with synchronized acoustic data to diagnose potential equipment faults and trigger parameter calibration of the digital twin model. S34: The optimal heat dissipation control strategy is executed through the heat dissipation actuator to adjust the operating state of the heat dissipation system. At the same time, the actual state of the system after execution is fed back to the data acquisition layer and the digital twin model to form a closed-loop control.

[0016] The beneficial effects of this invention are: 1. Compared to existing technologies that mainly employ open-loop or simple feedback control schemes based on real-time temperature thresholds, the disadvantages of these schemes are that the control behavior is entirely dependent on the current or past state, the system can only passively respond to temperature changes, resulting in response lag, and the heat dissipation components often work independently without coordination, leading to low energy efficiency. This invention adopts a closed-loop intelligent thermal management scheme based on a digital twin model. By constructing a virtual mapping of the generator thermal system for real-time simulation and future state prediction, the control system can proactively and collaboratively manage all heat dissipation resources based on the simulation results. This scheme has the core advantage of turning passive into active, as it can intervene in advance before excessive heat accumulation, achieving optimal coordinated control of all parts of the system, thereby significantly improving the accuracy, timeliness, and overall energy efficiency of thermal management. 2. Compared to existing technologies that generally rely on real-time sensor data for control, the disadvantage of such solutions is that the control system is like the blind men and the elephant, unable to foresee critical operating conditions such as sudden changes in ambient temperature or large fluctuations in generator load in the short term. This results in a slow response to sudden heat loads, requiring only drastic post-event compensation, which is detrimental to equipment lifespan and operational economy. This invention adopts a forward-looking control scheme that deeply integrates meteorological and power load forecast data across multiple time scales. The system can obtain the trend of operating condition changes several hours to several days in advance and intelligently switch between different control modes, such as preventative enhanced heat dissipation or economical maintenance. This scheme has the strategic advantage of being prepared for unforeseen circumstances. It enables the thermal management system to anticipate the future, smoothly transition to thermal shocks caused by changes in operating conditions, and maximize energy-saving potential while ensuring safety, achieving forward-looking energy consumption optimization. 3. Compared to existing thermal management systems that typically lack in-depth fault diagnosis capabilities or employ limited diagnostic methods, which are disadvantageous in that they are insensitive to potential faults, struggle to distinguish between model errors and actual physical faults, and cannot adaptively adjust when sensor or heat dissipation component performance deteriorates, potentially leading to false alarms or missed alarms, resulting in untimely or excessive maintenance, this invention employs a multi-source data fusion-based integrated fault diagnosis and fault-tolerant control scheme. By comparing the predicted temperature from the digital twin model with the measured temperature in real time and combining this with acoustic signature features for correlation analysis, it can accurately locate early faults. This scheme has the outstanding advantages of intelligent self-diagnosis and self-adaptation. The diagnostic conclusions can be directly fed back to the control unit, triggering specific fault-tolerant operation strategies, enabling the system to maintain safe operation even when the performance of some components deteriorates or experiences minor faults, greatly improving the system's reliability and availability. 4. Compared to existing technologies that typically focus solely on preventing temperature exceedances, which neglects the impact of temperature field uniformity on equipment thermal stress and fails to incorporate the energy consumption of the cooling system itself into the optimization objective, potentially leading to localized overheating or overcooling, resulting in energy waste and shortened equipment lifespan, this invention employs a multi-objective dynamic collaborative optimization scheme. This scheme simultaneously incorporates the highest temperature at key points, the uniformity of the entire temperature field, and the total power consumption of the cooling system into the optimization framework, dynamically adjusting the weights of each objective based on the operating mode. This scheme offers the advantage of globally optimal comprehensive decision-making; it goes beyond simply meeting the safety baseline of not exceeding limits, striving to find the optimal balance between safety, reliability, and energy efficiency. This ensures long-term equipment health while significantly reducing the operation and maintenance costs of wind farms. Attached Figure Description

[0017] Figure 1 The diagram shown is a structural schematic of the wind turbine thermal management system of the present invention. Figure 2 The diagram shown is a flowchart of the wind turbine thermal management method of the present invention. Detailed Implementation

[0018] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0019] Please see Figure 1-2 The present invention provides an embodiment: a thermal management system for a wind turbine generator, comprising: I. Data Acquisition Layer The data acquisition layer includes a temperature sensor array deployed on the heat-generating components and heat dissipation circuit of the wind turbine, and an acoustic fingerprint sensor for collecting the operating noise of the generator. II. Prediction Module The forecasting module is used to obtain future weather forecast data for the location of the wind turbine and future power load forecast data for the turbine, specifically including: A21: Meteorological data acquisition unit, used to access forecast data streams from external meteorological services and local meteorological stations to acquire future weather forecast data for the location of the wind turbine. A22: Power load forecasting unit, used to generate future power load forecasting data for wind turbines based on historical operating data and grid dispatching plans; A23: Predictive data fusion and processing unit, used to perform time alignment, format standardization and reliability verification on the acquired future weather forecast data and future power load forecast data, and output a structured future operating condition prediction sequence.

[0020] The meteorological data acquisition unit acquires future weather forecast data, including forecast data for the next 6 to 72 hours, and data elements include ambient temperature, wind speed, wind direction, air pressure, humidity, and solar radiation intensity.

[0021] The process by which the power load forecasting unit generates future power load forecasting data includes: The system collects input data, which includes historical power curves, short-term grid dispatch instructions, wind turbine health status data, and future wind speed prediction data from the prediction module. Based on physical formulas, the input data is processed into a sequence of expected generator output power for the corresponding future time period; Based on the generator's efficiency characteristic curve, the expected output power sequence is converted into a predicted power sequence for the heat source of the heat-generating components.

[0022] The process of processing input data into a sequence of expected generator output power for a corresponding future time period based on physical formulas leverages the fundamental aerodynamics and mechanical transmission laws of wind energy conversion. Its core lies in using the standard power curve of a wind turbine, which characterizes the functional relationship between wind speed and theoretical output power under rated operating conditions. In the processing, future wind speed prediction data is first used as the main input, and an initial theoretical power generation sequence is calculated using the standard power curve. Subsequently, grid dispatch instructions are introduced as a hard constraint on power output, limiting the theoretical power. Simultaneously, health status data reflecting wind turbine operating efficiency is incorporated to perform efficiency corrections on the theoretical power or the limited power, ultimately generating a sequence of expected output power that conforms to physical laws, reflects the true capacity of the equipment, and meets grid requirements.

[0023] The workflow of the predictive data fusion and processing unit is as follows: Receive raw weather forecast data from the meteorological data acquisition unit and raw load forecast data from the power load forecasting unit; The two sets of data are synchronized in time, unified onto the same future time coordinate axis, and interpolated to generate data points with the same time resolution. Perform quality checks on the data, remove outliers that are clearly outside the reasonable physical range, and fill in the missing time periods by interpolating adjacent data. The output contains a structured data sequence including timestamps, predicted environmental parameters, and predicted heat source power, which serves as the input for the predicted boundary conditions of the digital twin model.

[0024] III. Digital Twin Model The digital twin model is a virtual mapping of the heat generation and heat dissipation system of a wind turbine. It incorporates a built-in thermodynamic simulation engine to simulate and predict the heat generation, transfer, accumulation, and dissipation processes within the generator based on real-time operating data, the status of the heat dissipation system actuators, and prediction data from the prediction module. Specifically, this includes: A31: Three-dimensional geometry and mesh model, which is the three-dimensional geometric model and computational mesh generation for the physical entities in a wind turbine that require thermal management; A32: Material and physical property parameter library, which stores and defines the physical property parameters of the materials used in the various components that constitute a physical entity, including density, specific heat capacity and thermal conductivity; A33: A thermodynamics simulation engine that loads and runs 3D geometric and mesh models and calls up material and physical property parameter libraries to simulate the heat transfer process inside physical entities by solving energy conservation and heat transfer control equations. A34: Real-time data interface, used to receive sensor data from the data acquisition layer, current status data of the heat dissipation actuator, and prediction data from the prediction module in real time; A35: Heat source power mapping module, used to dynamically map the real-time operating load and predicted load of the generator to the volumetric heat generation rate and surface heat flux density of the corresponding heat source components in the three-dimensional geometry and mesh model, based on the component efficiency loss characteristics.

[0025] The governing equations solved by the thermodynamic simulation engine include the heat conduction equation in the solid domain and the flow-heat transfer coupling equation in the fluid domain, in order to simulate heat conduction, convection heat transfer and fluid flow processes.

[0026] Among them, the three-dimensional geometry and mesh model performs local mesh refinement processing on key components with concentrated heat and temperature-sensitive points.

[0027] The predicted data from the prediction module serves as the input to the time-varying boundary conditions of the thermodynamic simulation engine, including ambient temperature, wind speed, and solar radiation intensity for future periods.

[0028] The predictive operation process of the digital twin model includes: Initialize the model based on the current measured data; The forward time-step solution is obtained by combining future prediction data with a thermodynamic simulation engine. Output the temperature field, heat flux density distribution, and temperature change curves at key points across the entire system over a future period.

[0029] IV. Intelligent Control and Diagnostic Engine The intelligent control and diagnostic engine is used to optimize and control the heat dissipation scheme based on simulation results from the digital twin model, and to perform fault diagnosis and model calibration based on multi-source data fusion. Specifically, it includes: A11: Heat dissipation optimization control unit, used to dynamically generate the optimal heat dissipation control strategy based on the simulation prediction results of the digital twin model, with the optimization goal of balancing heat distribution and controlling the temperature of key points to not exceed the limit; A12: Multi-source data fusion fault diagnosis unit, used to compare the measured temperature data collected by the temperature sensor group with the simulated temperature data corresponding to the digital twin model in real time, and to analyze the voiceprint data collected by the voiceprint sensor in order to realize fault detection and model calibration.

[0030] The heat dissipation optimization control unit specifically includes: A111: Optimization target setting module, used to set a comprehensive optimization objective function that includes three sub-objectives; A112: Control strategy generation module, used to generate a set of candidate heat dissipation control strategies covering different heat dissipation intensities based on the current state and prediction information; A113: Forward simulation and evaluation module, used to call digital twin model to perform forward simulation prediction of candidate heat dissipation control strategies, and evaluate and score the simulation results based on the optimization objective function; A114: Optimal Decision and Output Module, used to select the candidate strategy with the highest evaluation score as the optimal control strategy and convert the strategy into a specific control command output.

[0031] The comprehensive optimization objective function set in the optimization objective setting module adopts a weighted summation form, and its expression is: ; in, Let be the total cost function. A penalty for the equipment's predicted maximum temperature. Predict the highest temperature for the equipment. This is a penalty term related to the temperature field uniformity index. As an index of temperature field uniformity, This is a cost item in the total power consumption of the cooling system. This represents the total power consumption of the cooling system. , and The weighting coefficients for each sub-objective are dynamically adjusted according to the operating mode.

[0032] The method by which the control strategy generation module generates candidate heat dissipation control strategies is as follows: Based on the current state of the heat dissipation actuators, within their respective safe adjustment ranges, multiple discrete or continuous combinations of control parameters are generated according to preset step sizes. Each combination represents a specific heat dissipation control strategy, and each heat dissipation control strategy includes a control sequence for each heat dissipation actuator in a future control cycle.

[0033] The workflow of the forward simulation and evaluation module is as follows: Obtain a candidate heat dissipation control strategy from the control strategy generation module; The candidate strategy is used as the control input, combined with future operating condition data from the prediction module, to drive the digital twin model to perform forward simulation, with the simulation time window covering at least one control cycle. Extract the predicted maximum temperature of the device, temperature field uniformity index, and cumulative power consumption of the heat dissipation system within the entire time window from the simulation results; The extracted predicted maximum temperature of the device, temperature field uniformity index, and cumulative power consumption of the heat dissipation system are substituted into the optimization objective function to calculate the evaluation score of the candidate strategy. Repeat the above process for all candidate strategies to obtain the evaluation scores for all strategies.

[0034] The optimal decision and output module also includes a feedback adjustment mechanism, specifically: After executing the optimal control strategy for one control cycle, the predicted temperature of the digital twin model during that cycle is compared with the actual measured temperature of the data acquisition layer. If the deviation between the two exceeds the preset tolerance range, a strategy re-evaluation is triggered, and the weight coefficients in the objective function are adaptively adjusted to adapt to dynamic changes in model error or unmodeled data.

[0035] When generating candidate strategies, the control strategy generation module's search space is limited by operational constraints, which include: the action rate limit of the heat dissipation actuator, the single adjustment range limit, and the equipment's minimum temperature protection constraint.

[0036] The heat dissipation optimization control unit is also used for forward-looking control using predictive data, and the specific process is as follows: S11: Predictive state assessment, receiving a predicted data sequence for the future control cycle from the prediction module, including the predicted ambient temperature sequence and the predicted heat source power sequence; S12: Forward control mode decision, based on the predicted data sequence, judges the future operating condition change trend and determines the forward control mode according to predefined rules; S13: Generate a forward-looking control strategy. In the control strategy generation module of the heat dissipation optimization control unit, the predicted data sequence is used as the boundary condition input, and the generation rules of the candidate control strategy and the weight coefficients in the optimization objective function are set or adjusted according to the determined forward-looking control mode. S14: Forward Simulation and Decision Making. The forward simulation and evaluation module evaluates the comprehensive performance of candidate strategies over the entire prediction time domain based on a complete simulation containing future prediction data. The optimal decision making and output module selects and outputs the optimal forward control strategy.

[0037] The specific rules for forward-looking control mode decision-making include: If the forecast data indicates that the ambient temperature will continue to rise, the wind speed will continue to decrease, and the heat source power will continue to increase during the future control period, then it will be determined as a heat-prone operating condition, and the preventive enhanced heat dissipation mode will be activated. If the forecast data indicates that the ambient temperature will continue to decrease and the heat source power will continue to decrease during the future control period, it will be determined as a cooling condition, and the economical heat dissipation mode will be activated. If the forecast data does not show a clear trend of continued heating or cooling, then the regular optimization mode will be activated.

[0038] When the preventative enhanced heat dissipation mode is activated, the specific operation of step S13 is as follows: Adjust and optimize the weighting coefficients in the objective function setting module, increase the penalty weight for the device's predicted maximum temperature approaching the safety threshold, and reduce the cost weight for the total power consumption of the cooling system. In the control strategy generation module, when generating candidate strategies, the search range for increasing the power and flow of the active cooling actuator is expanded to ensure that the cooling capacity is improved in advance before heat accumulation occurs.

[0039] When the economical heat dissipation mode is activated, the specific operation of step S13 is as follows: Adjust and optimize the weight coefficients in the module setting of the objective function, introduce a penalty term for the device temperature falling below the lower limit of the base temperature, and increase the cost weight of the total power consumption of the heat dissipation system. In the control strategy generation module, when generating candidate strategies, the power lower limit of the active heat dissipation actuator is limited, and the consideration of temperature stability is added to the optimization objective, so as to minimize heat dissipation energy consumption and delay heat dissipation action while ensuring that the temperature is above the lower limit of the temperature threshold.

[0040] The forward simulation and decision-making process is as follows: The forward simulation and evaluation module calls the digital twin model, taking the current state as the starting point and the predicted data sequence as the time-varying boundary condition, to perform forward dynamic simulation of each candidate control strategy across multiple future control cycles; During the evaluation, not only the instantaneous state at the end of the control cycle is considered, but also the comprehensive cost integral of temperature safety, uniformity and energy consumption is calculated over the entire simulation time domain. The optimal decision and output module selects the candidate strategy with the minimum integrated cost over the entire prediction time domain as the optimal strategy output.

[0041] The heat dissipation optimization control unit, when using predictive data for forward-looking control, also includes a prediction uncertainty compensation step: In step S11, uncertainty or confidence information of the prediction data is obtained simultaneously; When generating candidate strategies in step S13 and evaluating strategies in step S14, a conservatism offset is introduced into the prediction data. The offset is dynamically adjusted according to the uncertainty: the higher the uncertainty, the more conservative cooling is adopted in the preventive enhanced heat dissipation mode, and the more conservative maintenance strategy is adopted in the economical maintenance heat dissipation mode.

[0042] The multi-source data fusion fault diagnosis unit specifically includes: A41: Real-time data synchronization and comparison module, used to receive the measured temperature data stream of the temperature sensor group after time synchronization and the simulated temperature data stream output by the digital twin model, and calculate the real-time deviation of key measuring points. A42: Anomaly detection and diagnosis trigger module, used to monitor real-time deviations and activate the fault diagnosis process when preset anomaly trigger conditions are met; A43: Multi-source feature extraction and fusion analysis module, used to synchronously retrieve the voiceprint sensor data within the corresponding time period after the diagnostic process is activated, and extract temperature abnormality features and voiceprint abnormality features for correlation fusion analysis; A44: Fault Reasoning and Output Module, with a built-in fault feature knowledge base, is used to match fault modes based on the fusion analysis results and output diagnostic conclusions and confidence levels; A45: Model calibration trigger module, used to trigger the parameter self-calibration process of the digital twin model when the fault reasoning and output module determines that the system is fault-free, but the real-time data synchronization and comparison module detects a systematic, global long-term and fixed real-time deviation.

[0043] The workflow of the real-time data synchronization and comparison module includes: Obtain real-time measured temperature values ​​from specific physical measurement points from the data acquisition layer; Obtain simulated temperature values ​​for the same physical location from a digital twin model; Timestamp alignment is performed between real-time measured temperature values ​​and simulated temperature values; Calculate the real-time deviation between the measured temperature value and the simulated temperature value, and perform short-time filtering.

[0044] The preset exception triggering conditions in the exception detection and diagnosis triggering module are as follows: The mean absolute value of the same measurement point exceeds the static threshold within N consecutive sampling periods; Within M consecutive sampling periods, the rate of change of the real-time deviation between the measured real-time temperature value and the simulated temperature value exceeds the dynamic threshold. The spectral characteristics of the real-time deviation between the measured and simulated temperature values ​​show abnormal energy peaks within a specific frequency band.

[0045] The anomaly detection and diagnosis triggering module also includes a preliminary model mismatch exclusion function, specifically: When an abnormal deviation between the real-time measured temperature value and the simulated temperature value is detected, first check whether the input boundary conditions of the digital twin model are consistent with the actual working conditions. If the inputs are consistent, it is judged as a suspected physical fault and the subsequent diagnostic process is triggered; if there is a significant deviation in the inputs, it is judged as a model input error first, and an abnormal operating condition alarm is generated instead of a fault alarm.

[0046] The workflow of the multi-source feature extraction and fusion analysis module includes: Extract temperature anomaly features, including the spatial distribution pattern of anomaly measurement points, the time-domain variation curve of real-time deviation, and the spectral characteristics of real-time deviation anomalies. Acoustic features are extracted from synchronized voiceprint data, including energy variations in specific frequency bands, voiceprint envelope features, and pulse features. Temperature anomaly features and acoustic features are correlated, aligned, and fused in the time and frequency domains to form a multimodal anomaly feature vector.

[0047] The fault reasoning and output module has a built-in fault feature knowledge base that stores the temperature and voiceprint feature mapping relationships of various typical faults. Its fault reasoning process includes: The multimodal anomaly feature vectors generated by the multi-source feature extraction and fusion analysis module are compared with the fault feature templates in the knowledge base for similarity matching calculation. Based on the matching results, multiple candidate fault types and their confidence levels are output. Fault types include: generator bearing wear or damage, gearbox bearing or gear failure, coolant line blockage or leakage, cooling fan failure, converter power module overheating, and poor electrical connection point contact. Output a diagnostic report, which includes the fault type, suspected faulty component, confidence level, and recommended maintenance measures.

[0048] The diagnostic conclusions of the fault reasoning and output module will be input as feedback signals to the heat dissipation optimization control unit. This is used to adjust the control strategy for fault-tolerant operation after a specific fault is confirmed. The specific process is as follows: S21: Diagnostic conclusion reception and parsing. The heat dissipation optimization control unit receives diagnostic conclusions from the fault reasoning and output module in real time, including fault type identification, fault component location and fault severity level. S22: Fault-control mode mapping, which maps the specific fault type and severity received to a specific fault-tolerant operation control mode according to predefined mapping rules; S23: The control strategy is adaptively adjusted. Based on the fault-tolerant operation control mode obtained by mapping, the objective function, constraints of the target setting module, and candidate strategy generation rules of the control strategy generation module are dynamically adjusted and optimized. S24: Fault-tolerant strategy execution and monitoring. Based on the adjusted objectives and constraints, regenerate and select the optimal control strategy, issue it to the heat dissipation actuator for execution, and strengthen the monitoring of fault-related parameters.

[0049] The predefined rules for fault-control mode mapping include: If the diagnosis is insufficient coolant flow and partial blockage of the heat dissipation pipes, it will be mapped to enhanced auxiliary cooling and load reduction mode. If the diagnosis is that the temperature sensor is drifting or malfunctioning, it is mapped to a model-based fault-tolerant control mode. If the diagnosis is a risk of bearing overheating, it is mapped to a preventative load reduction and focused heat dissipation mode.

[0050] When mapped to enhanced auxiliary heat dissipation and load reduction mode, the control strategy adaptively adjusts as follows: In the optimization target setting module, the penalty weight for the temperature of the faulty component and its associated area is significantly increased, and the safety threshold for the temperature of the faulty component and its associated area is set to a value lower than the global safety threshold. In the generation rules of the control strategy generation module, the enhanced utilization of non-faulty heat dissipation paths is added, and candidate strategy combinations that reduce generator output power are generated to actively reduce thermal load.

[0051] When mapped to a model-based fault-tolerant control mode, the adaptive adjustment of the control strategy includes: In the optimization target setting module, the dependence weight on the measured temperature data from the faulty sensor is reduced, and the trust weight on the simulated temperature output by the digital twin model is increased accordingly, and it is used as the main control feedback. In the control strategy generation module, a robustness constraint based on the model prediction error range is introduced to ensure that the control strategy remains robust in the event of sensor failure.

[0052] V. Heat dissipation mechanism The heat dissipation actuator is used to receive and execute control commands issued by the heat dissipation optimization control unit to adjust the operating status of the heat dissipation system, including the coolant circulation pump, proportional control valve, cooling fan and its drive controller.

[0053] Among them, the control commands generated by the intelligent control and diagnostic engine are adjustment commands for coolant flow rate, flow speed, fan start / stop and speed.

[0054] like Figure 2 As shown, this embodiment also provides a thermal management method for wind turbine generators, including the following steps: S31: Real-time temperature data and operating acoustic data of the wind turbine's heating components and heat dissipation circuit are acquired through the data acquisition layer, and future weather forecast data and future power load forecast data of the generator are acquired through the prediction module. S32: Input real-time data and predicted data into the digital twin model. The digital twin model uses its thermodynamic simulation engine to simulate and predict the heat generation, transfer, accumulation and heat dissipation process inside the generator under the future operating conditions corresponding to the predicted data, and generate simulation results including the future temperature field and heat flow distribution. S33: The intelligent control and diagnostic engine executes the following parallel sub-steps: S331: Heat dissipation optimization control aims to balance heat distribution, control the temperature of key points to keep them within limits, and optimize energy consumption. Based on the simulation results of the digital twin model, it dynamically generates the optimal heat dissipation control strategy and sends it to the heat dissipation actuator. S332: Multi-source data fusion fault diagnosis, real-time comparison of measured data from temperature sensors with simulated data from digital twin models, calculation of deviation, and when the deviation is continuously abnormal, feature fusion analysis is performed in conjunction with synchronized acoustic data to diagnose potential equipment faults and trigger parameter calibration of the digital twin model. S34: The optimal heat dissipation control strategy is executed through the heat dissipation actuator to adjust the operating state of the heat dissipation system. At the same time, the actual state of the system after execution is fed back to the data acquisition layer and the digital twin model to form a closed-loop control.

[0055] Example 1: Intelligent Thermal Management of a 3MW Onshore Wind Turbine 1. System Configuration and Data Acquisition This embodiment focuses on a 3MW onshore wind turbine installed in a wind farm in northern China. Its thermal management system is configured as follows: Data acquisition layer: PT100 platinum resistance temperature sensors were deployed in 12 key heat-generating components, including the generator stator winding, gearbox bearings, and converter power modules; a total of 4 acoustic fingerprint sensors were installed on the gearbox and the non-drive ends of the generator, with a sampling frequency of 16kHz.

[0056] Forecasting module: Accesses local weather forecast data for wind farms from the China Meteorological Administration to obtain forecasts of ambient temperature, wind speed, air pressure, and humidity for the next 72 hours with a time resolution of 15 minutes. Simultaneously, based on the power grid dispatch plan and wind turbine operation model, it generates a future power load forecast sequence.

[0057] Digital twin model: A three-dimensional geometric model including the generator, gearbox, converter, and cooling circuit was established, and the mesh of key components was locally refined, with a total of approximately 2.5 million meshes. Material parameter library defines the thermal properties of copper windings, silicon steel sheets, bearing steel, coolant, etc.

[0058] Cooling mechanism: The system includes a coolant circulation pump with a maximum flow rate of 50L / min, three proportional valves with adjustable opening, and four cooling fans.

[0059] 2. Proactive control process At 14:00 on a certain summer day, the system initiated an optimization decision-making process for a control cycle (set to 15 minutes).

[0060] Predictive Status Assessment: The prediction module outputs data for the next hour, showing that the ambient temperature will continue to rise from the current 32℃ to 35℃, and the wind speed will decrease from 8m / s to 7m / s. At the same time, the generator load will increase by 10% due to grid demand. The predictive data fusion unit processes this information into a time-aligned structured sequence.

[0061] Forward-looking control mode decision: Based on the rules of continuously rising ambient temperature, decreasing wind speed, and increasing heat source power, the intelligent control engine determines that the future operating conditions are trending towards heat and automatically activates the preventive enhanced heat dissipation mode.

[0062] Strategy Generation and Evaluation: Based on the current state, the control strategy generation module generates five candidate control strategies (such as combining the adjustment of the circulating pump speed and the proportional valve opening) within the safe range of the heat dissipation actuator. The forward simulation and evaluation module calls the digital twin model and uses the predicted environmental data as boundary conditions to perform dynamic simulations of these five strategies over the next 45 minutes (three control cycles). In the optimization objective function, the penalty weight for the highest equipment temperature is automatically increased from 0.6 to 0.8, while the weight for power consumption is decreased from 0.3 to 0.1.

[0063] Optimal Decision and Execution: Simulation results show that the strategy of increasing the main pump speed to 85% of the rated speed in advance received the highest score. This strategy was output to the actuator. Although the current critical point temperature was only 72℃ (below the alarm value of 80℃), the system had already begun to actively enhance cooling. Actual operating data after 15 minutes showed that the critical point temperature stabilized at 76℃, successfully avoiding the potential risk of overheating. Compared with traditional hysteresis control, energy savings of approximately 5% are expected.

[0064] 3. Interactive Fault Diagnosis The acoustic signature sensor detected an abnormal energy peak in the 850Hz frequency band near the gearbox. Simultaneously, the digital twin model's predicted temperature for this area showed a sustained deviation of approximately 5°C from the actual measured value. The multi-source data fusion fault diagnosis unit was triggered, and by comparing with the fault feature knowledge base, it diagnosed early wear of the gearbox bearing with 85% confidence. This diagnosis was fed back to the control engine, which then switched to a fault-tolerant operation mode focused on preventative load reduction and prioritized heat dissipation. The weight of the bearing area's temperature was significantly increased in the optimization objectives, and a candidate strategy to slightly reduce the generator output power (3%) was generated to ensure equipment safety.

[0065] Example 2: Intelligent Thermal Management of a 5MW Offshore Wind Turbine 1. System Configuration and Data Acquisition This embodiment focuses on a 5MW offshore direct-drive wind turbine installed along the southeastern coast. The system is specially designed for high humidity and high salt spray environments.

[0066] Data acquisition layer: In addition to basic temperature sensors, humidity sensors have been added to areas prone to condensation, such as the converter cooling plate and generator air gap. The acoustic fingerprint sensor has enhanced corrosion resistance.

[0067] Forecasting module: In addition to regular meteorological data, sea surface salt spray concentration forecasts from local marine meteorological buoys have been specially introduced to predict the impact on radiator efficiency.

[0068] Digital twin model: The model specifically considers the airtightness of the cabin and the internal air convection effect. The boundary conditions of the thermodynamic simulation engine have been updated to include a reduction factor for the predicted salt spray concentration on the condenser heat transfer efficiency.

[0069] Heat dissipation mechanism: The system adopts a pump-driven two-phase flow circulation, which utilizes the latent heat of phase change of the working fluid to efficiently transfer heat, and uses seawater as the final cold source.

[0070] 2. Economical operation and fault diagnosis At 20:00 on a certain winter day, the system ran as follows: Predicted conditions assessment: Forecast data shows that the ambient temperature will drop from 8°C to 2°C in the next 6 hours, the wind speed will be stable, and the generator load will remain high due to the favorable wind speed forecast.

[0071] Forward-looking control mode decision: Based on the rule that the ambient temperature continues to decrease and the heat source power is stable, the system determines that it is in a cooling condition and activates the economical heat dissipation maintenance mode.

[0072] Strategy Generation and Evaluation: When generating candidate strategies, the control strategy generation module limits the lower power limit of the two-phase circulating pump. In the optimization objective function, the weight of the heat dissipation system power consumption is increased to 0.7, and a penalty term is introduced for equipment temperatures falling below the lower limit of the base operating temperature to prevent overcooling. Forward simulation evaluates the cumulative energy consumption during the entire nighttime low-load period.

[0073] Optimal Decision and Execution: The system ultimately adopts a strategy of maintaining the main circulation pump speed at 40% of its rated speed and utilizing low-temperature seawater to meet the heat dissipation requirements. Under this strategy, the digital twin model predicts that the temperature of key components will ideally fluctuate within the range of 68℃±2℃ over the next 8 hours, and the total power consumption of the cooling system is reduced by 18% compared to the conventional mode.

[0074] Fault-tolerant control under sensor failure: A generator inlet temperature sensor experienced abnormal reading drift (displaying -10℃) due to condensation. The fault diagnosis unit quickly identified the sensor as faulty by comparing readings from other adjacent sensors with the simulated temperature (15℃) output by the digital twin model. The diagnosis triggered a model-based fault-tolerant control mode. The control engine then reduced its reliance on the faulty sensor data and primarily relied on the simulated temperature output by the digital twin model for control feedback, ensuring continuous and stable system operation even under sensor malfunctions.

[0075] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

Claims

1. A thermal management system for a wind power generator, characterized in that: The method comprises the following steps: A data acquisition layer comprising a temperature sensor group deployed on heat generating components and heat dissipation circuits of the wind turbine and an acoustic fingerprint sensor for collecting operating noise of the wind turbine; A prediction module for obtaining future weather forecast data of the location of the wind turbine and future power load prediction data of the wind turbine; A digital twin model for virtual mapping of the heat generation and dissipation system of the wind turbine, which comprises a thermal dynamics simulation engine for simulating and predicting the heat generation, transfer, accumulation and dissipation process in the wind turbine according to real-time operating state data of the wind turbine, state of the heat dissipation system execution mechanism and prediction data from the prediction module; An intelligent control and diagnosis engine for optimizing control of the heat dissipation scheme based on the simulation results of the digital twin model and for fault diagnosis and model calibration based on multi-source data fusion; A heat dissipation execution mechanism for receiving and executing control instructions issued by the heat dissipation optimization control unit to adjust the operating state of the heat dissipation system.

2. A thermal management system for a wind power generator according to claim 1, characterized in that: The intelligent control and diagnosis engine specifically comprises: A11: a heat dissipation optimization control unit for dynamically generating an optimal heat dissipation control strategy based on the simulation prediction results of the digital twin model, with the optimization goal of balancing heat distribution and controlling key point temperature not to exceed the limit; A12: a multi-source data fusion fault diagnosis unit for comparing the measured temperature data collected by the temperature sensor group with the simulated temperature data corresponding to the digital twin model in real time, and analyzing the acoustic fingerprint data collected by the acoustic fingerprint sensor to realize fault detection and model calibration.

3. A thermal management system for a wind power generator according to claim 2, characterized in that: The prediction module specifically comprises: A21: a weather data acquisition unit for accessing the prediction data stream of external weather services and local weather stations to obtain future weather forecast data of the location of the wind turbine; A22: a power load prediction unit for generating future power load prediction data of the wind turbine based on the operating history data of the wind turbine and the power grid dispatching plan; A23: a prediction data fusion and processing unit for time alignment, format standardization and credibility verification of the obtained future weather forecast data and future power load prediction data, and outputting a structured future working condition prediction sequence.

4. A thermal management system for a wind generator as claimed in claim 3, wherein: The digital twin model specifically comprises: A31: a three-dimensional geometry and grid model, which is a three-dimensional geometric model and a computational grid division established for physical entities in the wind turbine that need to be heat managed; A32: a material and physical property parameter library, which stores and defines the physical property parameters of the materials used by each component of the physical entity, including density, specific heat capacity and thermal conductivity; A33: a thermal dynamics simulation engine, which loads and runs the three-dimensional geometry and grid model, and calls the material and physical property parameter library to simulate the heat transfer process in the physical entity by solving the energy conservation and heat transfer control equations.

5. A thermal management system for a wind power generator according to claim 4, characterized in that: The heat dissipation optimization control unit specifically comprises: A111: an optimization target setting module for setting a comprehensive optimization objective function comprising three sub-targets; A112: a control strategy generation module for generating a set of candidate heat dissipation control strategies covering different heat dissipation intensities according to the current state and prediction information; A113: a forward simulation and evaluation module, configured to call the digital twin model to perform forward simulation prediction on the candidate heat dissipation control strategy, and evaluate and score the simulation result based on an optimization objective function; A114: an optimal decision and output module, configured to select the candidate strategy with the highest evaluation score as the optimal control strategy, and convert the optimal control strategy into specific control instructions for output.

6. A thermal management system for a wind power generator according to claim 5, characterized in that: The heat dissipation optimization control unit is also configured to perform prospective control by using the prediction data, and the specific process is as follows: S11: prediction state evaluation, receiving a prediction data sequence in a future control period from the prediction module, including a predicted ambient temperature sequence and a predicted heat source power sequence; S12: prospective control mode decision, determining the future working condition change trend based on the prediction data sequence, and deciding the prospective control mode according to a predefined rule; S13: generating a prospective control strategy, inputting the prediction data sequence as a boundary condition into the control strategy generation module of the heat dissipation optimization control unit, and setting or adjusting the generation rule of the candidate control strategy and the weight coefficient in the optimization objective function according to the determined prospective control mode; S14: prospective simulation and decision, evaluating the comprehensive performance of the candidate strategy in the entire prediction time domain based on the complete simulation containing the future prediction data by the forward simulation and evaluation module, and selecting and outputting the optimal prospective control strategy by the optimal decision and output module.

7. A thermal management system for a wind power generator according to claim 6, characterized in that: The rules of the prospective control mode decision specifically include: If the prediction data indicates that the ambient temperature continues to rise, the wind speed continues to decrease, and the heat source power continues to increase in the future control period, it is determined that the working condition tends to be hot, and the preventive strengthening heat dissipation mode is enabled; If the prediction data indicates that the ambient temperature continues to decrease and the heat source power continues to decrease in the future control period, it is determined that the working condition tends to be cold, and the economic maintenance heat dissipation mode is enabled; If the prediction data does not show a clear continuous trend of heating and cooling, the regular optimization mode is enabled.

8. A thermal management system for a wind power generator according to claim 7, characterized in that: The multi-source data fusion fault diagnosis unit specifically includes: A41: real-time data synchronization and comparison module, configured to receive the measured temperature data stream of the temperature sensor group after time synchronization and the simulated temperature data stream output by the digital twin model, and calculate the real-time deviation of the key measurement points; A42: abnormality detection and diagnosis triggering module, configured to monitor the real-time deviation, and activate the fault diagnosis process when the preset abnormal triggering condition is met; A43: multi-source feature extraction and fusion analysis module, configured to synchronously call the voiceprint sensor data in the corresponding period after the diagnosis process is activated, extract temperature abnormal features and voiceprint abnormal features, and perform correlation fusion analysis; A44: fault reasoning and output module, with a built-in fault feature knowledge base, configured to match the fault mode according to the fusion analysis result, and output the diagnosis conclusion and confidence.

9. A thermal management system for a wind power generator according to claim 8, characterized in that: The diagnosis conclusion of the fault reasoning and output module will be input as a feedback signal to the heat dissipation optimization control unit for fault-tolerant operation after confirming a specific fault, and the specific process is as follows: S21: diagnosis conclusion receiving and analysis, the heat dissipation optimization control unit receives the diagnosis conclusion output by the fault reasoning and output module in real time; S22: Fault-control mode mapping, according to the pre-defined mapping rule, the received specific fault type and severity are mapped to a specific fault-tolerant operation control mode; S23: Control strategy adaptive adjustment, according to the fault-tolerant operation control mode obtained by mapping, the objective function of the optimization target setting module and the constraint condition are dynamically adjusted, and the candidate strategy generation rule of the control strategy generation module is dynamically adjusted; S24: Fault-tolerant strategy execution and monitoring, based on the adjusted target and constraint, the optimal control strategy is regenerated and selected, and is issued to the heat dissipation execution mechanism for execution, and the monitoring of the fault-related parameters is strengthened.

10. A method of thermal management of a wind power generator, characterized in that: Comprising the following steps: S31: Real-time acquisition of temperature data and operation voiceprint data of heat generating components and heat dissipation circuits of wind driven generators through the data acquisition layer, acquisition of future weather forecast data and generator future power load prediction data through the prediction module; S32: Input the real-time data and prediction data into the digital twin model, and the digital twin model simulates and predicts the heat generation, transmission, accumulation and dissipation process in the generator under the future working condition corresponding to the prediction data through its thermal dynamics simulation engine, and generates simulation results including future temperature field and heat flow distribution; S33: The intelligent control and diagnosis engine performs the following parallel sub-steps: S331: Heat dissipation optimization control, taking balancing heat distribution, controlling key point temperature not exceeding limit and optimizing energy consumption as the comprehensive goal, based on the simulation results of the digital twin model, dynamically generating the optimal heat dissipation control strategy and issuing it to the heat dissipation execution mechanism; S332: Multi-source data fusion fault diagnosis, real-time comparison of temperature sensor measured data and simulation data of the digital twin model, calculation of deviation, when the deviation is continuously abnormal, combined with the synchronous voiceprint data for feature fusion analysis, diagnosis of potential equipment faults and triggering of parameter calibration of the digital twin model; S34: Execute the optimal heat dissipation control strategy through the heat dissipation execution mechanism to adjust the operation state of the heat dissipation system, and at the same time, feedback the actual state of the system after execution to the data acquisition layer and the digital twin model, forming a closed loop control.

Images