Energy efficiency optimization method of new energy motor under dynamic regulation

Abstract

The present invention discloses a method for optimizing the energy efficiency of new energy motors under dynamic regulation, which relates to the field of electronic control technology. The method comprises: when an electric vehicle is driving, monitoring the vehicle load, navigation information, outside temperature, ambient humidity, and current motor temperature distribution within a predetermined future period, and predicting the driving speed change based on the navigation information to obtain a speed prediction sequence; then combining the data to predict the temperature distribution of the permanent magnet synchronous motor and generate a motor temperature prediction distribution field; calculating the motor temperature difference distribution, using this as a heat dissipation demand to analyze the heat dissipation plan, generating a global liquid cooling heat dissipation control plan containing multiple local control parameters, and performing heat dissipation control on the permanent magnet synchronous motor within a predetermined future period. The method solves the technical problem in the prior art that the liquid cooling system does not match the motor heat dissipation demand, resulting in low heat dissipation efficiency and affecting the performance and energy efficiency of the electric vehicle motor, and achieves the technical effect of improving heat dissipation efficiency, optimizing energy utilization efficiency, and extending the service life of the motor.

Description

Technical Field

[0001] The present application relates to the field of electronic control technology, and in particular to a method for optimizing energy efficiency of new energy motors under dynamic regulation. Background Art

[0002] With growing global environmental awareness and the deepening adoption of sustainable development concepts, the market for electric vehicles, as a clean and efficient means of transportation, is experiencing rapid growth. Among the key technologies in electric vehicles, the performance of the motor system plays a decisive role in the vehicle's power, range, and reliability. Permanent magnet synchronous motors (PMSMs) are widely used in electric vehicles due to their high power density, high efficiency, and excellent speed regulation. However, during operation, motors generate significant heat due to factors such as winding resistance, iron losses, and mechanical friction. Excessive motor temperature not only affects the magnetic properties of the permanent magnets, reducing the motor's output power and efficiency, but also accelerates the aging of the motor's internal insulation materials, shortening the motor's service life. Furthermore, uneven temperature distribution in the motor can cause internal thermal stresses, impacting the motor's structural stability. Currently, liquid cooling systems are a common method for dissipating heat from electric vehicle motors. However, traditional liquid cooling systems are not well-suited to the motor's cooling requirements. They employ a fixed cooling strategy that cannot be flexibly adjusted based on the motor's real-time operating status and environmental conditions. This often results in insufficient or excessive heat dissipation in practical applications. Furthermore, the continuous advancement of sensor, communication, and data processing technologies, along with the increasing adoption of intelligent control technology in the automotive sector, has provided new possibilities and directions for solving the heat dissipation problem of electric vehicle motors. In this context, optimizing the heat dissipation system for electric vehicle motors has become a critical issue that the industry urgently needs to address.

[0003] At present, relevant technologies have technical problems such as the mismatch between the liquid cooling system and the motor's heat dissipation requirements, resulting in low heat dissipation efficiency and affecting the performance and energy efficiency of electric vehicle motors. Summary of the Invention

[0004] This application solves the technical problem in the prior art that the liquid cooling system does not match the motor heat dissipation requirements, resulting in low heat dissipation efficiency and affecting the performance and energy efficiency of electric vehicle motors, by providing a new energy motor energy efficiency optimization method under dynamic regulation.

[0005] This application provides a method for optimizing the energy efficiency of new energy motors under dynamic regulation, including:

[0006] While an electric vehicle is traveling, the vehicle load, navigation information, outside temperature, ambient humidity, and current motor temperature distribution within a predetermined future time period are monitored and obtained, and a driving speed change is predicted based on the navigation information to obtain a speed prediction sequence; the motor temperature distribution of the permanent magnet synchronous motor is predicted based on the vehicle load, outside temperature, ambient humidity, current motor temperature distribution, and speed prediction sequence to generate a motor temperature prediction distribution field; the motor temperature difference distribution is calculated based on the standard motor temperature distribution and the motor temperature prediction distribution field, and a motor heat dissipation scheme is analyzed based on the motor temperature difference distribution as the heat dissipation requirement to generate a global liquid cooling heat dissipation control scheme, and heat dissipation control is performed on the permanent magnet synchronous motor within the predetermined future time period, wherein the global liquid cooling heat dissipation control scheme includes multiple local liquid cooling heat dissipation control parameters.

[0007] The method for optimizing the energy efficiency of new energy motors under dynamic regulation proposed in this application is to first monitor the vehicle load, navigation information, outside temperature, ambient humidity and current motor temperature distribution in a predetermined future period when the electric vehicle is driving, and predict the driving speed change based on the navigation information to obtain a speed prediction sequence; then combine these data to predict the temperature distribution of the permanent magnet synchronous motor and generate a motor temperature prediction distribution field; calculate the motor temperature difference distribution based on the standard and predicted temperature distribution, and use this as the heat dissipation demand to analyze the heat dissipation plan, generate a global liquid cooling heat dissipation control plan containing multiple local control parameters, and perform heat dissipation control on the permanent magnet synchronous motor in a predetermined future period. By accurately matching the liquid cooling system with the motor heat dissipation demand, the technical effect of improving heat dissipation efficiency, optimizing energy utilization efficiency and extending the service life of the motor is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] To more clearly illustrate the technical solutions of the embodiments of the present disclosure, the accompanying drawings of the embodiments of the present disclosure are briefly introduced below. Flowcharts are used in this application to illustrate the operations performed by the systems according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in precise order. Instead, various steps may be processed in reverse order or simultaneously as needed. Furthermore, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0009] Figure 1 A flow chart of a method for optimizing energy efficiency of a new energy motor under dynamic regulation provided in an embodiment of the present application;

[0010] Figure 2 Schematic diagram of the speed prediction sequence output process of the new energy motor energy efficiency optimization method under dynamic regulation provided in an embodiment of the present application. DETAILED DESCRIPTION

[0011] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0012] In order to make the purpose, technical solutions and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limiting this application. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0013] In the following description, reference is made to “some embodiments”, which describes a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict, and the terms “first\second” involved are merely used to distinguish similar objects and do not represent a specific ordering of the objects. The terms “including” and “having” and any variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or that are inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application belongs. The terms used herein are for the purpose of describing the embodiments of this application only.

[0014] The embodiment of the present application provides a method for optimizing the energy efficiency of a new energy motor under dynamic regulation, such as Figure 1 As shown, the method includes:

[0015] Step S100 monitors and obtains vehicle load, navigation information, exterior temperature, ambient humidity, and current motor temperature distribution for a predetermined future time period while the electric vehicle is in motion. Based on the navigation information, the system predicts driving speed changes and obtains a speed prediction sequence. Specifically, during the electric vehicle's driving process, the system first uses pressure sensors, load cells, and other devices installed in key locations such as the chassis and suspension to accurately sense vehicle load changes in real time. Because vehicle load directly correlates to motor workload and heat generation, accurate data is essential for subsequent analysis. Simultaneously, the system utilizes the vehicle's navigation system, connected to a satellite positioning system, to obtain navigation information for the predetermined future time period, including rich features such as road type, slope, curve curvature, and traffic conditions. Furthermore, temperature and humidity sensors are installed at appropriate locations on the vehicle's exterior to monitor exterior temperature and ambient humidity in real time. These sensors, which affect the motor's heat dissipation efficiency and internal insulation performance, respectively, provide a reference for subsequent analysis. Furthermore, multiple temperature sensors are deployed on key motor components, such as the stator winding, rotor, and housing. Combined with the layout of the electric vehicle's motor components, these sensors monitor the temperature of different motor parts in real time, providing a direct understanding of the motor's current heating status. After completing data monitoring, the driving speed change prediction phase begins. First, the acquired navigation information is preprocessed, the road type is classified and coded, the road slope and curve curvature are converted into numerical form, and the traffic condition information is quantified to make it structured data suitable for model processing. Then, a large amount of sample data is collected based on the historical driving records of electric vehicles. The data contains different navigation information features and the corresponding actual driving speeds. Then, the speed change prediction model is trained using machine learning algorithms (such as random forest algorithms) using the sample data. The model learns the complex relationship between navigation information features and driving speed during training and adjusts parameters to improve prediction accuracy. Finally, the preprocessed navigation information is input into the trained speed change prediction model. After calculation and analysis, the model outputs a speed prediction sequence for a predetermined future time period. This sequence can reflect the changing trend of the vehicle's future driving speed and provide important input parameters for subsequent motor temperature distribution prediction.

[0016] In one possible implementation, Figure 2During the driving of the electric vehicle, the vehicle load, navigation information, exterior temperature, ambient humidity, and current motor temperature distribution within a predetermined future time period are monitored and obtained. Based on the navigation information, driving speed changes are predicted to obtain a speed prediction sequence. Step S100 further includes step S110, where road features are extracted from the navigation information according to a predetermined distance and predetermined characteristic indicators to obtain a road feature sequence. Specifically, the road feature extraction process systematically processes the navigation information to obtain key road features for subsequent analysis and prediction. First, the predetermined distance and predetermined characteristic indicators must be determined. Based on actual needs and application scenarios, an appropriate fixed road segment distance is selected as the interval scale for feature extraction, such as every 200 meters. The distance selection requires comprehensive consideration to ensure that the extracted features are consistent and representative while not resulting in excessive data volume and computational burden. Key features to be extracted from the navigation information are also determined, including road type (highway, corridor, etc.), slope, and traffic lights. Road type affects vehicle speed and driving rules, slope is related to vehicle power demand and energy consumption, and traffic lights cause frequent vehicle starts and stops, significantly affecting driving speed and rhythm. Next, navigation information is parsed. Using the onboard navigation system or related positioning service interfaces, navigation information is obtained in real time during vehicle travel. This information is presented in the form of electronic map data, positioning data, and other forms. Professional data parsing algorithms and tools are then used to conduct in-depth analysis and processing, classifying and extracting road information according to predetermined rules. This converts complex navigation data into a structured, easily processable format, preparing for subsequent feature extraction. Finally, road features are extracted based on predetermined distances and indicators. Starting from the vehicle's current position, feature extraction points are set at predetermined distance intervals along the route determined by the navigation information. At each extraction point, road information is extracted and recorded based on predetermined characteristic indicators. The road type is determined and recorded. Slope information is obtained and recorded using terrain data or a slope sensor. The number and location of traffic lights within a certain range are counted and recorded. As the vehicle drives, feature extraction points are continuously advanced, and each feature is extracted and combined into a sequence in order, ultimately resulting in a road feature sequence.

[0017] Step S120, based on the historical driving records of the electric vehicle, collect a sample road feature sequence set, and mark the vehicle driving speed under different sample road feature sequences to obtain a sample speed sequence set. Specifically, valuable sample data is mined from the historical driving records of the electric vehicle. First, the data source of the historical driving records must be determined, mainly from the vehicle's on-board data recorder, the vehicle computer system, and the relevant cloud data storage platform. The on-board data recorder can record key parameters such as time, location, and speed in real time. The vehicle computer system can provide vehicle operating status data. The cloud data storage platform gathers a large amount of historical data from different time periods and different scenarios. Then, from the large amount of historical driving records obtained, data samples that meet the requirements are screened out, and abnormal data caused by sensor failure, emergency situations, etc. are removed. The screened data are sorted in chronological order, and data from different sources are integrated to form a complete and organized historical driving record data set. Then, based on the previously determined predetermined distance and predetermined characteristic indicators, road features are extracted from the navigation information in the sorted historical driving records. Each driving route is analyzed in turn according to a fixed road section distance (such as every 200 meters), and key feature information such as road type, slope, and traffic lights are extracted and combined into a road feature sequence. Road features are extracted for each historical driving record in this way to obtain a series of road feature sequences, which together constitute a sample road feature sequence set. The collected sample road feature sequence set is then properly stored using an appropriate data structure and database management system. A unique identifier is added to each sequence, and relevant information such as driving time and driving route is recorded for easy traceability and management. Finally, while collecting sample road feature sequences, the vehicle speed corresponding to each sample road feature sequence is annotated. Based on the timestamp information in the historical driving records, the vehicle speed corresponding to each feature extraction point is accurately matched. An appropriate method is used to determine the representative speed of each feature extraction point, such as taking the average speed over a period of time before and after the feature extraction point. The annotated speed is associated with the corresponding sample road feature sequence to ensure data accuracy and consistency. All sample road feature sequences with annotated speeds are then sorted and arranged in a certain order to form a sample speed sequence set. This set contains a large amount of actual vehicle speed data under different road characteristics, reflecting the speed change pattern of vehicles under various driving conditions.

[0018] In step S130, a random forest algorithm is trained using the sample road feature sequence set and the sample speed sequence set to obtain a converged speed change prediction model. The road feature sequence is then input into the speed change prediction model to output the speed prediction sequence. Specifically, the speed change prediction model is trained and applied using the random forest algorithm to accurately predict vehicle speeds. Data preprocessing begins by carefully cleaning the collected sample road feature sequence set and sample speed sequence set, identifying and addressing missing values ​​and outliers, and selecting appropriate imputation or correction methods based on data characteristics. Data accuracy is also verified to ensure that it truly reflects actual driving conditions. Because different features have varying dimensions and value ranges, to prevent certain features from unbalancedly affecting model training, the data is normalized. All feature values ​​are mapped to a uniform interval, ensuring more stable model training and improving convergence speed and prediction accuracy. Next, the random forest model is trained. An appropriate random forest algorithm library is selected. Key hyperparameters are set during model initialization, such as the number of decision trees, maximum depth, and the number of features considered for splitting nodes. The optimal combination of hyperparameters is determined through experimentation and tuning. The preprocessed sample road feature sequence set is then used as input features, and the sample speed sequence set is used as the target variable for model training. During training, the random forest model, based on the principles of decision tree construction, randomly selects samples and features from the sample data to construct multiple decision trees. During the growth process, each decision tree splits samples according to node features to maximize the distinction between samples of different speeds. Through ensemble learning of multiple decision trees, the model continuously adjusts parameters to improve its fit to the sample data. Model performance indicators are continuously monitored during training. When the performance indicators no longer decrease significantly or decrease by less than a set threshold over multiple training iterations, the model is considered to have reached convergence, indicating that the model has fully grasped the patterns of the sample data and can effectively predict vehicle speeds. Next, the speed prediction sequence is generated. The predicted road feature sequence is preprocessed using the same operations as the training data, including cleaning and normalization, to ensure that the input data format and range are consistent with the training data. The processed road feature sequence is then input into the trained and converged model. Based on the learned relationships, the model calculates and outputs speed predictions, which are arranged into a speed prediction sequence to reflect the possible future changes in vehicle speed. Finally, the model is evaluated and optimized using an independent test dataset. The road feature sequence from the test data is fed into the model to generate a predicted speed sequence. This is then compared with the actual speed sequence, and the prediction error is measured by calculating evaluation metrics. If the metrics are unsatisfactory, the model is either overfitting or underfitting, requiring further optimization. In the case of overfitting, the number of decision trees can be reduced, the maximum depth can be lowered, and the model simplified to improve generalization. In the case of underfitting, the number of decision trees can be increased, and the amount of training data can be expanded to enhance fitting capabilities. After multiple rounds of evaluation and optimization, the model achieves optimal prediction performance.

[0019] Step S200 predicts the motor temperature distribution of the permanent magnet synchronous motor based on the vehicle load, outside temperature, ambient humidity, current motor temperature distribution and speed prediction sequence, and generates a motor temperature prediction distribution field. Specifically, the accurate prediction of the temperature distribution of the permanent magnet synchronous motor is carried out to provide a strong basis for the formulation of the subsequent heat dissipation plan and ensure the stable operation and performance of the motor. The first step is data collection and integration. Various sensors installed on the electric vehicle are used to collect data such as vehicle load, outside temperature, ambient humidity and current motor temperature distribution in real time and accurately. The vehicle load sensor is installed on the chassis and other key parts, and the outside temperature and ambient humidity sensors are installed in a well-ventilated place outside the vehicle. The current motor temperature distribution is collected by multiple temperature sensors arranged on the key components of the motor. At the same time, the speed prediction sequence obtained by the speed change prediction model is combined to integrate the data according to the time series and establish a unified data structure for storage. During this period, quality inspection is carried out to handle missing or abnormal data. Next comes model selection and construction. Various existing motor temperature prediction models, such as heat conduction models based on physical principles, empirical formula models, and neural network models based on machine learning, are investigated and evaluated. The most suitable model is selected, taking into account factors such as accuracy, computational efficiency, and data requirements. Then, a motor temperature prediction model is constructed based on the selected model type. If a neural network model is selected, the network structure and training parameters are determined. If a heat conduction model based on physical principles is selected, various motor physical parameters are obtained and a reasonable heat conduction equation is established, fully considering the actual operating conditions of the electric vehicle. Model training and optimization are then carried out. The integrated data is divided into training and test sets according to a certain ratio. The training set is then input into the constructed model to begin training. During training, the model adjusts parameters based on the input data to reduce the error between the predicted and actual values. The training results are regularly monitored and the model is optimized based on the evaluation results. If overfitting occurs, regularization techniques are used to reduce model complexity and improve generalization. If underfitting occurs, methods such as increasing model complexity are tried to improve fitting ability. After multiple training and optimization cycles, the model achieves optimal prediction performance. Finally, the motor temperature prediction distribution field is generated. The integrated and preprocessed data are input into the trained and optimized model to ensure that the input data format and range are consistent with those during training. The model predicts the temperature distribution of the permanent magnet synchronous motor in the future based on the input data, and outputs a matrix or array containing the temperature prediction values ​​of various parts of the motor. Combined with the layout of motor components, the temperature prediction values ​​are mapped to the actual three-dimensional spatial structure of the motor. Visualization technology is used to generate an intuitive motor temperature prediction distribution field, providing a clear reference for the subsequent formulation of heat dissipation solutions.

[0020] In one possible implementation, a permanent magnet synchronous motor's motor temperature distribution is predicted based on the vehicle load, exterior temperature, ambient humidity, current motor temperature distribution, and speed prediction sequence to generate a predicted motor temperature distribution field. Step S200 further includes step S210 of pre-training a temperature distribution prediction plug-in, wherein the plug-in includes Q temperature distribution prediction branches, where Q is an integer greater than 10. Specifically, pre-training the plug-in and embedding multiple temperature distribution prediction branches is a key step in improving the accuracy and comprehensiveness of motor temperature distribution prediction. In plug-in design and construction, the overall plug-in architecture must be clearly defined, and a framework that can accommodate multiple branches must be established to ensure good scalability and compatibility. Furthermore, a data exchange mechanism and coordinated workflow between branches must be designed. Based on different prediction principles and methods, various temperature distribution prediction branches based on physical principles, machine learning, empirical formulas, and other methods can be selected and combined to leverage the strengths of different methods. The number of branches, Q, is set as an integer greater than 10 as required. The specific number should be determined based on factors such as prediction accuracy requirements, computing resource constraints, and data size. During pre-training data preparation, extensive data collection is required, including vehicle load, exterior temperature, ambient humidity, and current motor temperature distribution, recorded by onboard sensors. This data, along with route and speed information provided by the navigation system and various operating condition data from historical driving records, should cover diverse driving scenarios, weather conditions, and load conditions. After collection, the data is cleaned to remove outliers and noise. Normalization and standardization are then performed to ensure consistent scale and distribution. The data is also labeled to clearly identify the actual motor temperature distribution for each sample. Finally, the pre-processed data is divided into training, validation, and test sets in appropriate proportions. Entering the pre-training process, for each temperature distribution prediction branch, the corresponding model parameters are initialized according to its model type and structure, and then the appropriate training algorithm and optimizer are selected according to the model type of different branches. Then, the training set data is used to pre-train each branch. During training, the loss function between the prediction result and the true value is continuously calculated, and the model parameters are updated according to the training algorithm and optimizer. During this period, the validation set is used to evaluate the model performance and adjust the hyperparameters regularly. When the performance of the model on the validation set no longer improves or reaches the preset target, the training is stopped. After the pre-training is completed, the test set is used to evaluate the performance of each branch. If the performance of a branch does not meet the standard, further adjustment and optimization are required until all branches reach the expected performance indicators.

[0021] Step S220, perform speed mean calculation and speed fluctuation analysis based on the speed prediction sequence, determine the speed mean and speed variance, and configure the number of branch calls of the temperature distribution prediction branch based on the speed mean and speed variance. Specifically, by deeply analyzing the speed prediction sequence, the number of calls of the temperature distribution prediction branch is scientifically and reasonably configured, thereby improving the accuracy and efficiency of the motor temperature distribution prediction. First, the speed mean is calculated. Before the calculation, it is necessary to ensure that the acquired speed prediction sequence is complete and accurate. The sequence is based on the previous prediction model or algorithm and can reflect the trend of changes in the vehicle's driving speed over a period of time in the future. Before the calculation, a preliminary inspection is required to exclude abnormal data points caused by sensor failure or data transmission errors. If abnormal data is found, it must be corrected or eliminated according to the actual situation to ensure the reliability of the calculation results. Then the common arithmetic mean calculation method is used to determine the speed mean. All speed values ​​in the speed prediction sequence are added together and then divided by the total number of speed values. The mathematical formula is: , the speed mean can reflect the average speed level of the vehicle in the future, providing an important reference for subsequent analysis and decision-making. Then, speed fluctuation analysis is carried out. The speed variance is used to measure the degree of dispersion of each speed value in the speed prediction sequence relative to the speed mean, reflecting the speed fluctuation. When calculating, the difference between each speed value and the speed mean is first calculated, then the difference is squared, and finally the average of all the squared differences is calculated. The mathematical formula is: The larger the speed variance, the more drastic the speed fluctuation. Conversely, the smaller the speed fluctuation, the more stable the vehicle's speed. Following this formula, the speed variance is calculated by sequentially calculating the difference, squaring it, summing it, and dividing it by the total. This calculation can be aided by computer software or programming language functions to improve efficiency and accuracy. Finally, the number of calls to the temperature distribution prediction branch is configured. Based on actual needs and experience, association rules are established between the speed mean, speed variance, and the number of temperature distribution prediction branch calls. Generally, when the speed mean and speed variance are high, the vehicle's driving state is complex and changeable, and the motor's operating conditions are diverse, requiring more branch calls to improve prediction accuracy. When both the speed mean and variance are low, the vehicle's driving state is stable, and the motor's operating conditions are simple. The number of branch calls can be appropriately reduced to reduce computational cost and improve efficiency. The specific number of calls can be determined by setting a series of thresholds. The calculated speed mean and speed variance are then substituted into the rules to determine the final number of calls, ensuring that the number is within a reasonable range. The association rules can also be adjusted and optimized based on the characteristics and requirements of the actual application scenario, laying a solid foundation for more accurate and efficient motor temperature distribution prediction.

[0022] Step S230 randomly selects the temperature distribution prediction branches with the number of branch calls from the Q temperature distribution prediction branches, performs motor temperature distribution prediction based on the vehicle load, outside temperature, ambient humidity, current motor temperature distribution and speed prediction sequence, and outputs the motor temperature prediction distribution field after averaging the output multiple prediction results. Specifically, the accuracy and reliability of the motor temperature distribution prediction are improved by fusion of multi-branch prediction and results. First, review the number of branch calls determined based on the speed mean and speed variance to ensure that it is within a reasonable range, neither exceeding the total number of branches Q nor being too small to affect the accuracy and diversity of the prediction. Then, a random number generation algorithm is used to randomly select from the Q temperature distribution prediction branches. The random number function library of common programming languages ​​can assist in setting the value range and generating non-repeating random numbers that are the same as the number of branch calls. Each random number corresponds to a branch index, thereby determining the branches involved in the prediction, avoiding the deviation caused by fixed selection branches, and increasing the reliability and diversity of the prediction results. Each branch then independently predicts the motor temperature distribution. Data such as vehicle load, exterior temperature, ambient humidity, current motor temperature distribution, and speed prediction sequence are first organized into the required input format for each branch. Different branch types have different data format and preprocessing requirements, ensuring that the input data is accurate and complete. The prepared data is then fed into randomly selected branches. The physics-based branch calculates the temperature distribution based on physical laws such as heat conduction, combined with motor parameters and operating conditions. The machine learning-based branch extracts features through neural network learning for prediction. The empirical-based branch estimates the temperature distribution using empirical formulas. Each branch outputs an independent prediction result. The multiple prediction results are then averaged. The prediction results from each branch are first collected and organized into a unified format. Next, for each motor component, the predicted temperature values ​​from all branches are summed and divided by the number of branch calls to obtain the average prediction value. This combines the prediction strengths of multiple branches, reducing the error and uncertainty of individual branch predictions and improving the accuracy and stability of the final prediction result. Finally, before outputting the predicted motor temperature distribution, the calculated mean value is verified and optimized, compared with historical data or actual measurement data. If significant deviations are found, the cause is analyzed and adjustments are made, such as evaluating the prediction model for abnormal branches or adjusting parameter settings. After verification and optimization, the predicted motor temperature distribution is output as the final result, which can be presented in the form of a 2D or 3D temperature distribution graph, data table, or other formats.

[0023] In one possible implementation, a pre-trained temperature distribution prediction plug-in is embedded with Q temperature distribution prediction branches, where Q is an integer greater than 10. Step S210 further includes step S211, where a sample dataset is collected based on the electric vehicle's historical driving history, constrained by a predetermined future time period. The sample data includes sample input data and sample supervision data. The sample input data includes sample load, sample temperature, sample humidity, sample initial motor temperature distribution, and sample speed sequence, and the sample supervision data includes sample motor temperature distribution. Specifically, a sample dataset that meets the predetermined future time period is accurately and comprehensively collected from the electric vehicle's historical driving history, laying a solid data foundation for subsequent model training. First, the application requirements must be clearly defined, and the overall project scenario and objectives must be thoroughly analyzed. The predetermined future time period is then determined based on the timescale requirements for electric vehicle motor temperature prediction. If the goal is to adjust the electric vehicle's cooling system in real time to ensure short-term stable motor operation, the time period can be set to 5-15 minutes. If the goal is to evaluate and optimize the electric vehicle's motor performance over longer trips, the time period can be extended to 30 minutes or even several hours. At the same time, comprehensive consideration must be given to the driving characteristics of electric vehicles. The motor's operating state and temperature changes vary significantly under different driving conditions. The motor's inherent thermal characteristics, such as its heat capacity and heat dissipation efficiency, determine the rate of temperature change. Furthermore, the timeliness and availability of data must be considered. Excessively long time periods reduce data timeliness, while too short a time period prevents sufficient and effective information from being obtained. After weighing these factors, an appropriate future time period must be determined. Next, the sources of historical electric vehicle driving records must be identified. These sources include onboard data recording systems, which can record key driving parameters in real time; cloud-based data storage platforms, which ensure complete and reliable data; and third-party data collection devices or platforms, ensuring data accuracy and legality. A collection strategy must then be developed based on the data sources. For onboard data recording systems, understand their data storage formats and readout methods, and develop reading tools or interfaces. For cloud-based data storage platforms, negotiate with the provider to obtain access rights and download data according to the specified interfaces and specifications. When collecting data, set a reasonable timeframe, collect sufficient records to ensure data diversity and representativeness, and perform preliminary screening and filtering to remove erroneous or anomalous data. Then we start extracting sample data. For sample input data, we obtain sample load from historical driving records through on-board weighing sensors or estimates to ensure data accuracy and consistency. We use an external temperature sensor to extract sample temperature, paying attention to sensor accuracy and calibration, and recording the acquisition time. We use a humidity sensor to obtain sample humidity, evaluate and calibrate sensor performance, and record the acquisition time. We use temperature sensors at key motor locations to obtain sample initial motor temperature distribution to ensure normal sensor operation and data accuracy. We extract sample speed sequences from vehicle speed sensors or navigation systems, calibrate and verify the speed sensors, and organize data in chronological order.In terms of sample supervision data, at the end of the predetermined future time period, the motor temperature distribution data is obtained through the motor temperature sensor to ensure that the sensor measurement range and accuracy meet the requirements, and the acquisition time corresponds to the end time of the time period. Finally, the extracted sample input data and sample supervision data are formatted uniformly, and data in different formats and units are converted to a unified format and unit, such as unifying temperature to degrees Celsius and speed to kilometers per hour. The sample input data and sample supervision data are aligned and associated based on the timestamp information to ensure that the input and supervision data of each sample correspond to the same time period and driving status, forming a complete sample data pair, checking data integrity and consistency, and handling missing or erroneous data. The sorted sample data is organized according to a certain structure and format, and a sample data set is constructed using common data storage formats such as CSV files and database tables. A unique identifier is added to each sample to facilitate management and query, and the data set is backed up to prevent data loss or damage.

[0024] Step S212, the sample data set is divided into Q equal parts, and the BP neural network is supervised and trained respectively until convergence, and Q temperature distribution prediction branches are obtained, which are combined to generate the temperature distribution prediction plug-in. Specifically, the sample data set is used to supervise the training of multiple BP neural networks to generate multiple temperature distribution prediction branches, which are then combined to form a temperature distribution prediction plug-in, providing a powerful tool for motor temperature distribution prediction. First, the sample data set is divided into equal parts. Before the operation, the total amount of the sample data set is determined. Assuming that there are N samples in total, Q equal parts are required, and the number of samples in each part is n=N / Q. When calculating, it is necessary to consider whether the number of samples can be divided evenly. If not, the excess samples can be randomly distributed to each part to ensure that the number of samples in each data set is as balanced as possible. Then, the data processing tools of the programming language are used to distribute the samples to Q different subsets in a sequential or random sampling manner. When distributing, it is ensured that the data distribution of each subset is similar to avoid special data or deviation from the overall distribution, so as to provide representative data for subsequent model training. Then the BP neural network is initialized. A BP neural network is initialized for each equally divided sample data set. The network structure is determined during initialization. The number of neurons in the input layer is consistent with the dimension of the sample input data. The number of layers and neurons in the hidden layer are determined based on experience or experiments. Adding hidden layers can improve the fitting ability, but it will also increase the training time and the risk of overfitting. The number of neurons in the output layer is consistent with the dimension of the sample supervision data. At the same time, the network parameters are set, such as the learning rate determines the parameter update step size, the activation function introduces nonlinear factors, and the number of iterations determines the number of learning rounds. Then the supervised training process begins. The sample input data in each sample data set is input into the corresponding BP neural network. The neural network performs forward propagation. The data goes from the input layer to the hidden layer and then to the output layer. The data in each layer is multiplied by the weight matrix, and the bias term is added and then the output result is obtained by nonlinear transformation of the activation function, that is, the predicted value of the current sample. The predicted value is then compared with the sample supervision data to calculate the error. The error function such as the mean square error is commonly used. Back propagation is performed based on the error. The error signal is transmitted from the output layer in the reverse direction layer by layer. The gradient of the neuron error to the weight and bias is calculated. The weight and bias are updated by the gradient descent algorithm to reduce the error. During training, it is constantly judged whether the convergence conditions are met. The error threshold or the maximum number of iterations can be set. After each iteration, it is checked whether it is met. If not, the next round of training is continued.Finally, the temperature distribution prediction branches are combined and plug-ins are generated. When all BP neural network training converges, these Q trained neural networks are collected as temperature distribution prediction branches. Each branch has the ability to predict the motor temperature distribution and may have different focuses. Then the plug-in architecture is designed, including an input interface that receives external input data and passes it to each branch, an output interface that collects the prediction results of each branch and outputs the motor temperature prediction distribution field after fusion according to rules, and internal call logic to coordinate the work of each branch. Finally, according to the designed architecture, the Q branches are integrated to generate a temperature distribution prediction plug-in. After generation, it is tested and optimized with sample data that did not participate in the training or new actual data. If the prediction result has a large error or is unstable, further adjust the fusion rules and retrain some neural networks.

[0025] In one possible implementation, speed mean calculation and speed fluctuation analysis are performed based on the speed prediction sequence to determine the speed mean and speed variance, and the number of branch calls of the temperature distribution prediction branch is configured based on the speed mean and speed variance. Step S220 further includes step S221, respectively calculating the ratio of the speed mean to the maximum speed limit of the vehicle, and the speed variance to the historical maximum speed variance. The speed demand analysis scale is obtained after the mean calculation. Specifically, by deeply analyzing the speed-related parameters and comprehensively considering the relationship between the speed mean and the maximum speed limit of the vehicle, and the speed variance and the historical maximum speed variance, the speed demand analysis scale with key guiding significance for subsequent decision-making is obtained. Before calculating the speed demand analysis scale, the method of obtaining the speed mean is first clarified. It is based on the speed prediction sequence and uses the formula , sum all speed values ​​in the sequence and divide by the total. During calculations, ensure the sequence is accurate and complete to avoid biased results due to data anomalies or missing data. The mean speed value is then properly recorded and stored. Next, determine the vehicle's maximum speed limit. Since speed limits vary by road type, this can be obtained by consulting traffic regulations, consulting traffic management departments, or using professional traffic information query platforms. Combined with real-time location information obtained from the vehicle's navigation system and onboard positioning devices, the vehicle's road type can be determined and the maximum speed limit for the current road can be determined. At the same time, be aware of temporary speed limit adjustments due to special circumstances such as road construction and inclement weather. After determining the mean speed value and the vehicle's maximum speed limit, divide the mean speed value by the maximum speed limit to obtain a ratio that intuitively reflects the relative relationship between the two. A smaller ratio indicates that the vehicle's average speed is further from the maximum speed limit, the more stable the driving state, and the potentially lower prediction accuracy requirements. This ratio is then recorded. To determine the historical maximum speed variance, a large amount of historical speed data covering different driving scenarios, time periods, and weather conditions should be collected from vehicle data recorders and driving data management platforms. The data should be preliminarily sorted and screened to remove abnormal and erroneous data, and then the formula should be used. Calculate the speed variance for each data set, find the maximum value, and record it as the historical maximum speed variance. Then, divide the current speed variance by the historical maximum speed variance to obtain a ratio that reflects the relative relationship between the current speed fluctuation and the historical maximum fluctuation. The smaller the ratio, the smaller the current speed fluctuation, the more stable the driving state, and the lower the prediction accuracy requirement. This ratio is recorded. Finally, average the ratio of the mean speed to the vehicle's maximum speed limit and the ratio of the speed variance to the historical maximum speed variance to determine the speed demand analysis scale. This scale comprehensively reflects the average level and fluctuation of vehicle speeds. Based on its size, appropriate decisions can be made. A small scale indicates stable driving conditions and low prediction accuracy requirements, which can reduce computing resources and improve efficiency. A large scale indicates complex driving conditions and high prediction accuracy requirements, requiring more computing resources to ensure accuracy.

[0026] In step S222, the speed demand analysis scale is multiplied by Q and rounded to the nearest integer to obtain the number of branch calls. Specifically, the number of branch calls of the temperature distribution prediction plug-in is determined based on the speed demand analysis scale, which can reasonably allocate computing resources and improve prediction efficiency while ensuring prediction accuracy. First, the basic data must be clarified. First, the accuracy and validity of the speed demand analysis scale data must be ensured. This data is obtained by calculating the ratio of the comprehensive speed mean to the maximum speed limit of the vehicle, and the ratio of the speed variance to the historical maximum speed variance, and averaging them. It reflects the degree of demand for prediction accuracy of the current vehicle speed. It needs to be checked again before use to check and correct errors or abnormal data. Second, the total number of temperature distribution prediction branches Q in the temperature distribution prediction plug-in must be clearly known. Q is an integer greater than 10, which has been determined during the plug-in design and construction phase, and limits the upper limit of the number of branches that can be used for motor temperature distribution prediction. Then perform a multiplication operation to multiply the confirmed speed demand analysis scale by Q to dynamically adjust the number of branch calls based on the speed demand analysis scale. Since the multiplication result may be a decimal, and the actual number of branch calls must be an integer, a rounding operation is required. You can choose a suitable rounding method such as rounding down, rounding up, or rounding up according to the actual situation and project requirements. For example, if the multiplication result is 8.3, the number of branch calls after rounding up is 8; if the multiplication result is 8.6, the number of branch calls after rounding up is 9. The rounded result is the final number of branch calls. Finally, verify and adjust the results. After obtaining the number of branch calls, verify its rationality and check whether it is within a reasonable range greater than 0 and less than or equal to Q. If it exceeds the range, recheck the calculation steps and data. At the same time, combine the actual prediction needs and computing power resources to determine whether the number of branch calls can meet the prediction accuracy requirements. If the prediction accuracy is too low or the computing power resources are seriously wasted, adjust the rounding method or the calculation method of the speed demand analysis scale. For example, if the number of branch calls is too small and the prediction accuracy does not meet the standard, consider rounding up or increasing the calculation weight of the speed demand analysis scale; if the number of branch calls is too large and the computing power resources are wasted, try rounding down or reducing the calculation weight of the speed demand analysis scale. Through continuous verification and adjustment, ensure that the determined number of branch calls can not only meet the prediction accuracy requirements, but also reasonably utilize computing power resources to improve prediction efficiency.

[0027] Step S300 calculates the motor temperature differential distribution based on the standard motor temperature distribution and the predicted motor temperature distribution field, and uses the motor temperature differential distribution as the heat dissipation requirement to analyze the motor heat dissipation solution, generate a global liquid cooling heat dissipation control solution, and perform heat dissipation control on the permanent magnet synchronous motor during the predetermined future period, wherein the global liquid cooling heat dissipation control solution includes multiple local liquid cooling heat dissipation control parameters. Specifically, the motor temperature differential distribution is calculated by analyzing the difference between the standard motor temperature distribution and the predicted motor temperature distribution field. This is used as the basis for the heat dissipation requirement to develop a global liquid cooling heat dissipation control solution that locally dissipates heat at specific heat-generating locations in the permanent magnet synchronous motor, thereby achieving efficient heat dissipation control during the predetermined future period. First, calculate the motor temperature difference distribution. Obtain the standard temperature ranges for key motor components during normal operation from the motor manufacturer's technical documentation. Verify and supplement the standards with data from past experimental tests and actual operating experience to ensure the data is reliable and comprehensive. Refer to the previously generated predicted motor temperature distribution field to check its completeness and accuracy. Compare the standard motor temperature distribution with the predicted distribution field point by point. For key components such as the stator winding, rotor, and motor controller, calculate the temperature difference by subtracting the predicted temperature value from the standard temperature value. This determines the degree to which each component's temperature deviates from the standard and the areas where heat dissipation is most urgent. Next, conduct a heat dissipation demand analysis. Based on the motor's temperature difference distribution, identify high-heating areas, focusing on areas where electromagnetic conversion heats up rapidly due to high-speed rotation or high-load operation. Combined with the motor's operating characteristics, operating conditions, and the temperature sensitivity of each component, assess the degree of heat dissipation demand and rationally allocate heat dissipation resources. Then, a global liquid cooling control scheme is designed, and the cooling source positioning is planned according to the high-heating area to ensure that the cooling source effectively covers and does not interfere with the normal operation of the motor; an independent cooling circuit is designed for each high-heating area, and the cooling circuit parameters are customized. Each circuit is equipped with an independent control device; local liquid cooling control parameters are determined for each independent cooling module, and temperature changes are monitored in real time and adjusted dynamically. Parameters are optimized according to different operating conditions and environmental conditions to ensure efficient operation of the system. Finally, heat dissipation control is implemented, and liquid cooling system components are installed according to the plan to ensure that the installation position is accurate, the pipe connections are firm and well sealed, the pipes are reasonably arranged to reduce resistance, and sensors are installed to monitor the temperature of various parts of the motor and the operating parameters of the cooling system; after starting the system, data is monitored in real time, and the control system automatically adjusts the control parameters according to the preset strategy to ensure that the high-heating area is within the appropriate temperature range; during the heat dissipation control process, the motor operation data and the operation effect of the heat dissipation system are continuously collected and analyzed, and the scheme is optimized according to the actual situation to reduce energy consumption while ensuring the heat dissipation effect.

[0028] In one possible implementation, a motor temperature differential distribution is calculated based on a standard motor temperature distribution and the predicted motor temperature distribution field. Using the motor temperature differential distribution as the heat dissipation requirement, a motor heat dissipation solution is analyzed to generate a global liquid cooling control solution. Heat dissipation control is then performed on the permanent magnet synchronous motor during the predetermined future period. The global liquid cooling control solution includes multiple local liquid cooling control parameters. Step S300 further includes step S310 of configuring a predetermined future period and obtaining periodic navigation information for the predetermined future period, wherein the interval of the predetermined future period is greater than the predetermined future period. Specifically, configuring the predetermined future period and obtaining navigation information for the predetermined future period are key foundations for subsequent analysis and decision-making. Reasonable period setting and comprehensive information acquisition provide strong data support for accurate prediction and effective control. First, the configuration basis must be clearly defined, and actual electric vehicle usage scenarios must be thoroughly studied. For urban commuting, where roads are complex and vehicles frequently start and stop, the predetermined future period can be set to 15-30 minutes to promptly respond to emergencies and adjust motor heat dissipation and performance in real time. For long-distance transportation, where the driving environment is stable, the period can be extended to 1-2 hours. Consider the motor's thermal characteristics and thermal stability. Motors with small heat capacity, weak heat dissipation capabilities, and rapid temperature rise require more frequent monitoring and adjustment of cooling measures, and the scheduled future cycle should be shorter. Motors with good thermal stability and slow temperature changes can have longer cycles. The response time and efficiency of the cooling system should also be considered. Systems with fast response times can have shorter scheduled future cycles, while systems with long response times require longer cycles. Next, determine the interval for the scheduled future cycles, taking into account the length of the scheduled future period and the aforementioned factors. Ensure that the interval is greater than the scheduled future period. For example, if the scheduled future period is 10 minutes, considering urban commuting, average motor thermal stability, and the short response time of the cooling system, a 30-minute interval can be used. Once determined, verify the appropriateness of this interval to ensure it meets actual needs and effectively monitors and controls the motor's operating status. If the interval is unreasonable, adjust it accordingly. Next, choose a method for obtaining periodic navigation information. You can use the vehicle's navigation system to obtain route planning, estimated travel time, real-time traffic conditions, and other information. When interacting with the vehicle's navigation system, ensure its accuracy and stability, and promptly update maps and traffic data. Alternatively, you can use third-party map data services, choosing a reliable provider and obtaining information according to interface specifications, while also paying attention to data security and privacy. Finally, collect and organize periodic navigation information to obtain information such as the route, estimated travel time, road type, and real-time traffic conditions for the planned future period. The route affects vehicle speed and time, as well as the motor's operating status and cooling requirements. Estimated travel time helps optimize cooling system operations. Road type and real-time traffic conditions affect the motor's workload and cooling requirements. Obtaining this information in a timely manner allows for proactive adjustments to cooling strategies to ensure proper motor operation.

[0029] Step S320 : using the speed change prediction model to predict and acquire a periodic speed prediction sequence according to the periodic navigation information, and setting a heat dissipation time constraint according to the periodic speed prediction sequence.Specifically, before using the speed change prediction model for prediction, the model must be fully checked and confirmed: if the model is trained based on a machine learning algorithm, the generalization ability must be verified to prevent overfitting or underfitting; if it is established based on a physical model and empirical formula, it is necessary to check whether the model assumptions are in line with reality and whether the parameters are accurately calibrated to ensure that the model is complete and accurate, the parameters are reasonable, and the training data is sufficient and representative; at the same time, the acquired periodic navigation information is sorted and preprocessed. The periodic navigation information includes the driving route, estimated driving time, road type, real-time traffic conditions, etc. The format conversion and data cleaning are performed according to the model input requirements to remove noise and outliers, correct and supplement incomplete data, and encode road type information to ensure that it can be accurately input. Enter the model; then make a prediction to obtain the cycle speed prediction sequence, input the sorted cycle navigation information into the model, and combine or input them separately according to the model type and requirements. For example, the machine learning model needs to extract features and combine them into vector inputs, and the physical model needs to input relevant information into the corresponding formulas separately. Ensure that the information order and format are correct when inputting, and then start the model calculation and prediction. The model analyzes the impact of various factors on vehicle speed based on the internal algorithm logic and outputs a sequence. Pay close attention to the operating status during the prediction, and handle errors and exceptions in time. If problems such as long running time or memory overflow occur, optimize the model or adjust the input data scale; after obtaining the prediction sequence, verify and evaluate the results, and compare the prediction results with actual historical data or Reliable prediction data is compared, and the prediction error is quantified using indicators such as mean square error (MSE) and mean absolute error (MAE). If the error is large, the cause is analyzed and corresponding measures are taken to improve it, such as adjusting model parameters, optimizing input data, or improving model structure; then the periodic speed prediction sequence is analyzed, statistical analysis is performed, and the mean of the predicted driving speed is calculated, that is, all speed values ​​are added and divided by the total number to obtain the average driving speed to reflect the approximate driving speed level, and then the predicted speed fluctuation information is calculated, and the speed fluctuation degree is measured by variance or standard deviation. The larger the variance or standard deviation, the more severe the speed fluctuation and the more unstable the driving. At the same time, the trend of sequence changes is observed, and a curve of speed change over time is drawn to judge the vehicle driving status. The driving state will affect the motor workload and heat dissipation requirements; finally, the heat dissipation time constraint is set. When setting, the motor thermal characteristics and heat dissipation system performance must be fully considered. Different motors have different thermal characteristic parameters, and the heat dissipation system performance indicators are also different. These all affect the setting of the heat dissipation time constraint. According to the predicted average driving speed, predicted speed fluctuation information and speed change trend, a reasonable heat dissipation time constraint is set. When the average speed is high and the fluctuation is large, the motor workload is large and the heat generation is high. The heat dissipation time interval needs to be shortened and the working time of the cooling system needs to be increased. Conversely, the heat dissipation time interval can be extended and the working time can be reduced. For unstable speed change trends, dynamic adjustments are made according to specific changes to ensure that the motor remains within the appropriate temperature range.

[0030] Step S330: Based on the heat dissipation time constraint and the motor temperature difference distribution as the heat dissipation requirement, the motor heat dissipation solution is analyzed, and the global liquid cooling heat dissipation control solution is output. Specifically, the heat dissipation requirements are sorted out, the motor temperature difference distribution is sorted out, the temperature difference size and distribution of each part are clarified, the temperature difference data of each part is recorded and analyzed, and the parts with high temperatures and large heat dissipation requirements are determined. For key parts such as the stator winding and the rotor, large temperature differences indicate high heat generation and require special attention, providing a basis for subsequent heat dissipation solutions. At the same time, the heat dissipation time constraint is studied to understand the working requirements of the heat dissipation system at different time periods. Based on factors such as the cycle speed prediction sequence, the working intensity and duration of the heat dissipation system at different times are clarified. For example, when the vehicle is traveling at a high speed and the motor load is high, a large amount of heat dissipation is required in a short time. When the vehicle is traveling stably and the load is light, the working intensity and duration can be appropriately reduced. Next, a cooling strategy is planned. Different cooling time periods are divided according to cooling time constraints. The cooling requirements for each time period are kept relatively consistent. This can be done by time intervals or driving state changes. Then, combined with the motor's temperature differential distribution, a cooling strategy is developed for each time period. When cooling demand is high, the cooling system's workload is increased, the coolant flow rate and velocity are increased, and the temperature is lowered for rapid heat dissipation. When demand is low, the workload is reduced, the flow rate and velocity are reduced, and the temperature is increased to conserve energy. Different measures should also be taken based on the differences in cooling requirements for different parts. Next, local liquid cooling control parameters are determined. Based on the cooling strategy for each time period, appropriate control parameters are set for each key part of the motor. For example, when the stator winding has high cooling demand, high flow and low temperature are set. When the rotor has low cooling demand, appropriately lower flow and pressure and increase temperature. Sensors are used to monitor the motor temperature and coolant flow in real time, and parameters are dynamically adjusted based on the monitoring data to meet cooling requirements. Finally, a global liquid cooling control plan is integrated to summarize the control parameters of each time period and location to ensure accuracy and consistency. The plan is recorded and organized in detail for implementation adjustments. The implementation process and control logic are formulated, and the operation steps and parameter adjustment methods of starting, stopping, and adjusting the cooling system are clarified. After the plan is formed, its cooling effect and energy consumption are evaluated through simulation, experimental testing, etc. If there are problems, the causes are analyzed in a timely manner, the control parameter setting values ​​are optimized, the structural layout of the cooling system is improved, and the system performance and efficiency are improved.

[0031] In one possible implementation, the speed change prediction model is used to predict a cyclical speed sequence based on the cyclical navigation information, and a heat dissipation time constraint is set based on the cyclical speed prediction sequence. Step S320 further includes step S321, calculating the speed mean and speed variance of the cyclical speed prediction sequence to obtain the cyclical speed mean and cyclical speed variance. Specifically, before calculating the cyclical speed mean and cyclical speed variance, the data in the cyclical speed prediction sequence is first inspected to identify missing values ​​and outliers, and the location and number of missing values ​​are determined. Outliers are often caused by data acquisition equipment failures, signal interference, and other factors. For example, a speed value of 200 kilometers per hour may appear when the normal speed range is 30-60 kilometers per hour. If there are few missing values, the average value of adjacent data can be used to fill the gap. If there are many missing values ​​and they are regular, more complex methods can be used for estimation. If outliers are due to acquisition errors, correction is required. If the data is extreme and has a significant impact, statistical methods can be used to adjust the data. After data inspection and processing, the cyclical speed mean is calculated. The sum of the speed values ​​in the sequence is then divided by the total number of speed values ​​to obtain the mean, which can reflect the approximate vehicle speed. After calculating the mean, calculate the variance. First, find the difference between each speed value and the mean and square it. Then, add the squared differences and divide by the total number of speed values ​​to obtain the variance. The variance reflects the fluctuation of vehicle speed. After calculating the mean and variance, record them in detail, noting the calculation time, data source, and calculation method. You can use a table or document to record the data. After analyzing the results, the mean reflects the overall speed, while the variance reflects the stability of driving speed.

[0032] Step S322, input the periodic speed mean and periodic speed variance into the speed-heat dissipation time mapping library, and match and obtain the heat dissipation time constraint, wherein the heat dissipation time constraint is negatively correlated with the periodic speed mean and periodic speed variance. Specifically, before inputting data into the speed-heat dissipation time mapping library, it is necessary to carefully check the periodic speed mean and periodic speed variance, and check the calculation process again to ensure that the calculation method is correct and the data processing is rigorous, and to ensure that the periodic speed prediction sequence used for the calculation is complete and accurate, without data omissions or errors. For example, when calculating the mean, it is necessary to ensure that all speed values ​​are included in the summation, and when calculating the variance, it is necessary to ensure that the difference, square, and summation operations of each speed value and the mean are accurate. At the same time, check whether the formats of the two data meet the requirements of the mapping library. Different mapping libraries may have specific regulations on the number of decimal places, data types, and other formats of values. It is also necessary to ensure that the data unit is consistent with the mapping library preset. If there is any inconsistency, the unit conversion is performed to ensure that the input data format and unit are compatible with the mapping library. Next, it is necessary to have a deep understanding of the speed-heat dissipation time mapping library. The structure and characteristics of the heat dissipation time mapping library. The mapping library may be stored in the form of tables, databases, hash tables, etc. If it is stored in a database, its table structure and query method must be mastered, and the characteristics of the mapping relationship must be analyzed. Since the heat dissipation time constraint is negatively correlated with the cycle speed mean and cycle speed variance, that is, the larger the cycle speed mean and variance, the smaller the heat dissipation time constraint, and vice versa, when using the mapping library, it is necessary to understand the data distribution law based on this characteristic; then, the checked cycle speed mean and cycle speed variance must be accurately input according to the requirements of the mapping library. If implemented through programming, it is necessary to ensure that the code logic is correct and the data can be accurately passed to the query interface. If operated manually, the data input box must be carefully filled in to avoid errors. After entering the data, the matching query function of the mapping library is started. According to the storage structure and query algorithm, the best matching record is found in the data. If there is a completely matching record, the corresponding heat dissipation time constraint value is directly obtained. If not, then according to the mapping library design, use interpolation method, nearest neighbor algorithm, etc. to estimate the appropriate heat dissipation time constraint value. For example, the interpolation method can be calculated based on adjacent known data points; after obtaining the heat dissipation time constraint, check its rationality and judge whether the obtained value meets expectations based on the negative correlation. For example, if the input cycle speed mean and variance are large, and the heat dissipation time constraint value is also large, it does not meet the characteristics and there may be problems. At the same time, combined with the actual motor heat dissipation requirements and system operation conditions, judge whether the constraint can meet the heat dissipation requirements at different driving speeds. If it is found to be unreasonable, it needs to be adjusted. The reasons may be inaccurate mapping library data, defects in the matching algorithm, or errors in the input data. If it is a mapping library data problem, the mapping library can be updated and improved. If it is an algorithm problem, it may be necessary to optimize or select a more appropriate algorithm. If it is an input data error, the calculation process and data of the corrected cycle speed mean and cycle speed variance must be rechecked.

[0033] In one possible implementation, based on the heat dissipation time constraint and taking the motor temperature difference distribution as the heat dissipation requirement, a motor heat dissipation solution is analyzed, and the global liquid cooling control solution is output. Step S330 further includes step S331, obtaining a global liquid cooling structure for the permanent magnet synchronous motor, wherein the global liquid cooling structure includes multiple local liquid cooling components. Specifically, first, technical data and documents are consulted. Technical data and documents are collected from the manufacturer, including key content such as the overall motor design, detailed description of the heat dissipation system, and specifications of each component. The technical manual can provide a comprehensive and professional introduction to the global liquid cooling structure, clarifying the functions, locations, and connection relationships of each local liquid cooling component. Product manuals and installation guides can also be consulted to provide additional information from different perspectives. At the same time, relevant industry standards and specifications are referenced. These standards clearly stipulate the design and performance requirements of the permanent magnet synchronous motor heat dissipation system, covering all aspects of the liquid cooling structure. By consulting these standards, general industry design principles, component selection requirements, and performance testing methods can be understood, ensuring that the information obtained is accurate and complies with industry standards. Next, if conditions permit, physically inspect the motor and disassemble the cooling system on-site. This is an effective way to obtain intuitive and accurate information. Disassembly should be performed in a sequential and methodical manner to avoid damaging the motor and cooling components. Each component should be carefully observed for its appearance, structural features, and connections to other components. Each local liquid cooling component should be labeled and documented in detail, using numbers and labels to record the component name, model, specifications, installation location, and connection method. Photos or sketches can also be taken to assist with documentation. This information will provide a basis for subsequent analysis. Next, specialized testing equipment should be used to obtain more in-depth information. Ultrasonic flaw detectors should be used to inspect the cooling ducts for defects such as cracks and holes to ensure their tightness and reliability. Infrared thermal imaging cameras should be used to monitor the temperature distribution during motor operation to understand the heat generation in various components and assess the cooling effectiveness. Non-destructive testing equipment captures information that is invisible to the naked eye. This data is then analyzed to assess the performance and operating status of the overall liquid cooling structure, allowing for targeted optimization and improvement of the cooling structure. Finally, if you encounter problems in the process of obtaining information, you can consult motor design experts. With their professional knowledge and practical experience, they can analyze and interpret the global liquid cooling structure from a professional perspective, answer questions and provide improvement suggestions; you can also communicate with motor maintenance technicians. They are familiar with the motor cooling system in their daily work and can share problems encountered in maintenance and solutions, and provide actual component use and maintenance experience, so as to obtain practical information and further deepen your understanding of the global liquid cooling structure.

[0034] In step S332, based on the heat dissipation time constraint and the motor temperature differential distribution as the heat dissipation requirement, and with the goal of minimizing heat dissipation energy consumption, the heat dissipation control parameters of the multiple local liquid-cooled heat dissipation components are optimized and analyzed, outputting multiple local liquid-cooled heat dissipation control parameters. Specifically, basic data is first compiled and analyzed to clarify the heat dissipation time constraint. Based on the motor load under different driving conditions, the operating requirements and time constraints of the heat dissipation system are understood, providing a time dimension constraint for subsequent optimization. The motor temperature differential distribution is then thoroughly analyzed to determine the heat dissipation priority of key components such as the stator winding and rotor, with a particular focus on areas with large temperature differentials. Information such as the model, specifications, performance parameters, and location of the local liquid-cooled heat dissipation components is summarized to lay the foundation for parameter optimization. Next, a heat dissipation energy consumption model is established. With the goal of minimizing heat dissipation energy consumption, factors influencing heat dissipation energy consumption, such as cooling water pump power, radiator heat dissipation efficiency, coolant flow rate, and temperature, are analyzed to identify their intrinsic relationships with heat dissipation energy consumption. A specific heat dissipation energy consumption calculation model is constructed to quantify the relationship between each factor and heat dissipation energy consumption, providing a quantitative basis for parameter optimization. Then, the initial control parameter range is set. Based on component performance and operating characteristics, and in conjunction with the actual motor operation, the initial value ranges for each component's control parameters are determined to avoid poor cooling or excessive energy consumption due to unreasonable parameter selection. Then, using an appropriate optimization algorithm, such as a genetic algorithm or particle swarm optimization algorithm, the cooling time constraint, the motor temperature differential distribution, the cooling energy consumption model, and the initial control parameter ranges are used as input. During operation, the control parameters of each component are adjusted, the cooling effect and energy consumption under different parameter combinations are calculated, the optimal parameter combination is selected, and the algorithm's performance is monitored in real time. The optimization results are then evaluated and verified to check whether the parameters meet the motor cooling requirements. The reasonableness of the motor temperature distribution is verified through simulation or testing. The optimized energy consumption is calculated based on the energy consumption model, and the energy consumption before and after optimization is compared to assess whether the minimization goal is achieved. If the results are not satisfactory, the parameters are readjusted. The stability and reliability of the cooling system under the optimized parameters are also considered to ensure the normal operation of each component. Finally, the finalized local liquid cooling control parameters are recorded, detailing the control parameter values, applicable conditions, and time ranges. The report is then output as a report that includes parameter descriptions, optimization results, cooling effect, energy consumption evaluation, and system stability analysis.

[0035] The embodiment of the present application adopts the method of monitoring the vehicle load, navigation information, outside temperature, ambient humidity and current motor temperature distribution in a predetermined future time period when the electric vehicle is traveling, and predicting the driving speed change based on the navigation information to obtain a speed prediction sequence; then combining these data to predict the temperature distribution of the permanent magnet synchronous motor, and generating a motor temperature prediction distribution field; calculating the motor temperature difference distribution based on the standard and the predicted temperature distribution, and analyzing the heat dissipation plan based on this as the heat dissipation demand, generating a global liquid cooling heat dissipation control plan containing multiple local control parameters, and controlling the heat dissipation of the permanent magnet synchronous motor in a predetermined future time period, thereby achieving the technical effect of improving heat dissipation efficiency, optimizing energy utilization efficiency, and extending the service life of the motor by accurately matching the liquid cooling system with the heat dissipation demand of the motor.

[0036] The above specific embodiments do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application shall be included within the scope of protection of this application.

Claims

1. A method for optimizing energy efficiency of new energy motors under dynamic regulation, characterized in that: Methods include: While the electric vehicle is traveling, monitoring and obtaining vehicle load, navigation information, outside temperature, ambient humidity, and current motor temperature distribution within a predetermined future period, predicting driving speed changes based on the navigation information, and obtaining a speed prediction sequence; Predicting the motor temperature distribution of the permanent magnet synchronous motor according to the vehicle load, the outside temperature, the ambient humidity, the current motor temperature distribution and the speed prediction sequence to generate a motor temperature prediction distribution field; Calculating a motor temperature difference distribution based on a standard motor temperature distribution and the predicted motor temperature distribution field, performing a motor heat dissipation scheme analysis based on the motor temperature difference distribution as a heat dissipation requirement, generating a global liquid cooling heat dissipation control scheme, and performing heat dissipation control on the permanent magnet synchronous motor within the predetermined future period, wherein the global liquid cooling heat dissipation control scheme includes a plurality of local liquid cooling heat dissipation control parameters; Predicting a driving speed change based on the navigation information to obtain a speed prediction sequence includes: Extracting road features from the navigation information according to a predetermined distance and a predetermined feature index to obtain a road feature sequence; Based on the historical driving records of electric vehicles, a sample road feature sequence set is collected, and the vehicle driving speeds under different sample road feature sequences are marked to obtain a sample speed sequence set; Using the sample road feature sequence set and the sample speed sequence set to train a random forest to obtain a converged speed change prediction model, and inputting the road feature sequence into the speed change prediction model to output the speed prediction sequence; The motor temperature distribution of the permanent magnet synchronous motor is predicted according to the vehicle load, the outside temperature, the ambient humidity, the current motor temperature distribution, and the speed prediction sequence, and a motor temperature prediction distribution field is generated, including: A pre-trained temperature distribution prediction plug-in, wherein the temperature distribution prediction plug-in has Q temperature distribution prediction branches embedded therein, where Q is an integer greater than 10; Performing speed mean calculation and speed fluctuation analysis according to the speed prediction sequence to determine the speed mean and speed variance, and configuring the number of branch calls of the temperature distribution prediction branch according to the speed mean and speed variance; The temperature distribution prediction branches of the number of branch calls are randomly selected from the Q temperature distribution prediction branches, the motor temperature distribution is predicted according to the vehicle load, the outside temperature, the ambient humidity, the current motor temperature distribution and the speed prediction sequence, and the motor temperature prediction distribution field is output after the mean of the output multiple prediction results is calculated.

2. The method for optimizing energy efficiency of new energy motors under dynamic regulation according to claim 1, characterized in that: Pre-trained temperature distribution prediction plug-in, including: Collecting a sample data set based on the historical driving record of the electric vehicle and taking the predetermined future time period as a constraint, wherein the sample data includes sample input data and sample supervision data, the sample input data includes sample load, sample temperature, sample humidity, sample initial motor temperature distribution and sample speed sequence, and the sample supervision data includes sample motor temperature distribution; The sample data set is equally divided into Q parts, and the BP neural network is supervised and trained respectively until convergence, to obtain Q temperature distribution prediction branches, which are combined to generate the temperature distribution prediction plug-in.

3. The method for optimizing energy efficiency of new energy motors under dynamic regulation according to claim 1, characterized in that: Configuring a temperature distribution according to the speed mean and the speed variance to predict the number of branch calls of a branch includes: Calculating the ratio of the speed mean to the maximum speed limit of the vehicle, and the speed variance to the historical maximum speed variance respectively, and obtaining the speed demand analysis scale after the mean calculation; The speed requirement analysis scale is multiplied by Q and rounded to an integer to obtain the number of branch calls.

4. The method for optimizing energy efficiency of new energy motors under dynamic regulation according to claim 1, characterized in that: Based on the motor temperature difference distribution as the heat dissipation requirement, the motor heat dissipation solution is analyzed to generate a global liquid cooling control solution, including: Configuring a predetermined future period and acquiring periodic navigation information of the predetermined future period, wherein the time interval of the predetermined future period is greater than the predetermined future time period; Using the speed change prediction model, a periodic speed prediction sequence is obtained according to the periodic navigation information prediction, and a heat dissipation time constraint is set according to the periodic speed prediction sequence; Based on the heat dissipation time constraint and taking the motor temperature difference distribution as the heat dissipation requirement, a motor heat dissipation solution analysis is performed to output the global liquid cooling control solution.

5. The method for optimizing energy efficiency of new energy motors under dynamic regulation according to claim 4 is characterized in that: Setting a heat dissipation time constraint according to the periodic speed prediction sequence includes: Calculating the speed mean and speed variance of the periodic speed prediction sequence to obtain the periodic speed mean and periodic speed variance; The periodic speed mean and periodic speed variance are input into a speed-heat dissipation time mapping library, and the heat dissipation time constraint is obtained by matching. The heat dissipation time constraint is negatively correlated with the periodic speed mean and periodic speed variance.

6. The method for optimizing energy efficiency of new energy motors under dynamic regulation according to claim 4, characterized in that: Based on the heat dissipation time constraint and taking the motor temperature difference distribution as the heat dissipation requirement, a motor heat dissipation solution analysis is performed, including: Acquire a global liquid cooling and heat dissipation structure of the permanent magnet synchronous motor, wherein the global liquid cooling and heat dissipation structure includes a plurality of local liquid cooling and heat dissipation components; Based on the heat dissipation time constraint, the motor temperature difference distribution as the heat dissipation requirement, and minimizing heat dissipation energy consumption as the goal, the heat dissipation control parameter optimization analysis is performed on the multiple local liquid cooling heat dissipation components, and multiple local liquid cooling heat dissipation control parameters are output.

Images