Motor temperature prediction method and system, storage medium and vehicle
By combining the LSTM neural network model with the motor operating status parameters, the problem of low accuracy in existing motor temperature prediction has been solved, achieving efficient and accurate motor temperature prediction, reducing costs and improving motor utilization efficiency.
Patent Information
- Application Number
- CN202410947378.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-15
- Publication Date
- 2024-11-12
AI Technical Summary
Existing motor temperature prediction algorithms fail to effectively utilize historical data, resulting in low prediction accuracy. Furthermore, their reliance on temperature sensors increases costs, impacting motor efficiency and user experience.
An LSTM neural network model is used, which combines various operating parameters such as motor rotation frequency, copper loss, iron loss, and coolant temperature. The model is trained using historical data to predict motor temperature, reducing reliance on temperature sensors.
It improves the accuracy of motor temperature prediction, reduces reliance on temperature sensors, lowers overall vehicle application costs, prevents motor overheating and burnout, and improves motor efficiency.
Smart Images

Figure CN118937993A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of DC motor temperature prediction in a vehicle, and in particular to a motor temperature prediction method, system, storage medium and vehicle. Background Art
[0002] Electric vehicles are developing rapidly. For electric vehicles, it is very important to improve the maximum efficiency of motors. Currently, most of the motor temperature acquisition methods still rely on temperature sensors. There are hundreds of motors in a car. If the temperature is collected entirely by temperature sensors, the cost of using the car will be greatly increased, which is not easy to control.
[0003] Therefore, some motor temperature prediction algorithms have also appeared in the prior art. However, most of the current motor prediction algorithms only consider the impact of the current time slice on the motor temperature, and do not effectively utilize a large amount of historical data, resulting in data waste. At the same time, the prediction accuracy of the motor temperature is low;
[0004] In addition, in some existing technologies, the motor thermal protection is directly performed according to the rotation frequency of the motor, that is, when the motor rotates a preset fixed number of times, the motor stops working; in other cases, the temperature prediction is performed when the motor temperature is normal, which may result in the motor temperature being within the normal temperature range when the motor stops, resulting in a waste of computing resources. Therefore, these existing methods will cause the motor to be used less efficiently and affect the user experience. Summary of the invention
[0005] The technical problem to be solved by the present invention is to provide a motor temperature prediction method, system, storage medium and vehicle, which can quickly and accurately predict the motor temperature based on historical data, improve the prediction accuracy, reduce the dependence on temperature sensors, reduce the application cost of the whole vehicle, and improve the use efficiency of the motor.
[0006] As one aspect of the present invention, a method for predicting motor temperature is provided, which comprises the following steps:
[0007] Acquire the rotation frequency information of the motor, and compare the rotation frequency information with a predetermined frequency threshold to determine whether it is necessary to trigger the motor temperature prediction process;
[0008] When it is determined that the motor temperature prediction process needs to be triggered, the time parameters of the current cycle and a predetermined number of previous cycles of the motor under various operating conditions, the initial temperature of the motor to be tested, and the operating time of the motor to be tested are obtained, and the time parameters at least include: copper loss and iron loss information of the corresponding cycle;
[0009] The acquired information is input into a trained neural network prediction model to perform motor temperature prediction, obtain the motor temperature prediction result and output it.
[0010] The time parameters of the current cycle and a predetermined number of previous cycles of the motor under various operating conditions are obtained, including:
[0011] Obtain the time parameters of the motor at the current time t, the previous time parameters at the time t-1, and the previous time parameters at the time t-2 under various operating conditions, where:
[0012] The time parameters at the current time t include: current ambient temperature T1, current copper loss P cu 、Current iron loss P fe , current coolant temperature T2;
[0013] The time parameters at time t-1 include: ambient temperature T1 at time t-1 t-1 , copper loss P at time t-1 cu t-1 、Iron loss P at time t-1 fe t-1 , Motor coolant temperature T2 at time t-1 t-1 ;
[0014] The time parameters at time t-2 include: ambient temperature T1 at time t-2 t-2 , copper loss P at time t-2 cu t-2 , Iron loss P at time t-2 fe t-2 , Motor coolant temperature T2 at time t-2 t-2 ;
[0015] Among them, the copper loss information at each moment is calculated using the following formula:
[0016] P cu =r*i 2
[0017] Where P cu is the copper loss of the motor to be measured, i is the stator current of the motor, and r is the winding resistance of the motor;
[0018] The iron loss information at each moment is calculated using the following formula:
[0019] P fe =K h *B max 2 *f+K e *B max 2 *f
[0020] Where Pfe is the iron loss of the motor to be tested, K h is the motor hysteresis loss coefficient, K e is the motor eddy current loss coefficient, B max is the maximum magnetic flux density of the motor, and f is the motor speed.
[0021] The method further includes the steps of establishing a neural network prediction model and training the model, including:
[0022] Build a neural network prediction model using LSTM;
[0023] Collect the time parameters of multiple cycles of the motor under various operating conditions, the initial temperature of the motor to be tested and the running time of the motor to be tested corresponding to each cycle to form a training set and a prediction set, wherein the time parameters at least include: copper loss and iron loss information of the corresponding cycle;
[0024] The current time parameter, t-1 time parameter, t-2 time parameter, the initial temperature of the motor to be tested, and the running time of the motor to be tested in the training set are used as input training sample data, and the corresponding motor end temperature in the training set is used as output training sample data to train the neural network prediction model;
[0025] After the training stop condition is reached, the trained neural network prediction model is obtained.
[0026] Wherein, obtaining and outputting the motor temperature prediction result further includes:
[0027] Output the predicted temperature obtained by the neural network prediction model to the CAN bus;
[0028] The predicted temperature is compared with a preset maximum allowable temperature. When the predicted temperature exceeds the maximum allowable temperature, an alarm operation is triggered, and a shutdown instruction is sent to the motor controller to stop the motor from working.
[0029] Accordingly, as another aspect of the present invention, a motor temperature prediction system is also provided, comprising:
[0030] A prediction trigger judgment module is used to obtain the rotation frequency information of the motor and compare the rotation frequency information with a predetermined frequency threshold to determine whether the motor temperature prediction process needs to be triggered;
[0031] The motor related information acquisition module is used to obtain the time parameters of the current cycle and a predetermined number of previous cycles of the motor under various operating states, the initial temperature of the motor to be tested, and the running time of the motor to be tested when the prediction trigger judgment module determines that the motor temperature prediction needs to be triggered. The time parameters at least include: copper loss and iron loss information of the corresponding cycle;
[0032] The temperature prediction processing module is used to input the acquired information into a trained neural network prediction model to perform motor temperature prediction, obtain the motor temperature prediction result and output it.
[0033] Wherein, the motor related information acquisition module includes:
[0034] The copper loss information calculation module is used to calculate the copper loss information at each moment using the following formula:
[0035] P cu =r*i 2
[0036] Where P cu is the copper loss of the motor to be measured, i is the stator current of the motor, and r is the winding resistance of the motor;
[0037] The iron loss information calculation module is used to calculate the iron loss information at each moment using the following formula:
[0038] P fe =K h *B max 2 *f+K e *B max 2 *f
[0039] Where P fe is the iron loss of the motor to be tested, K h is the motor hysteresis loss coefficient, K e is the motor eddy current loss coefficient, B max is the maximum magnetic flux density of the motor, and f is the motor speed;
[0040] The time parameter acquisition module is used to obtain the time parameters of the motor at the current time t, the time parameters at the previous time t-1, and the time parameters at the previous time t-2 under various operating conditions, wherein:
[0041] The time parameters at the current time t include: current ambient temperature T1, current copper loss P cu 、Current iron loss P fe , current coolant temperature T2;
[0042] The time parameters at time t-1 include: ambient temperature T1 at time t-1 t-1 , copper loss P at time t-1 cu t-1 、Iron loss P at time t-1 fe t-1 , Motor coolant temperature T2 at time t-1 t-1 ;
[0043] The time parameters at time t-2 include: ambient temperature T1 at time t-2 t-2 , copper loss P at time t-2 cu t-2 , Iron loss P at time t-2 fe t-2 , Motor coolant temperature T2 at time t-2 t-2 ;
[0044] The other information acquisition module is used to obtain the initial temperature of the motor to be tested and the running time of the motor to be tested.
[0045] Among them, further include:
[0046] The model training processing module is used to establish a neural network prediction model and perform training, which includes:
[0047] Model building module, used to build a neural network prediction model using LSTM;
[0048] The training set acquisition module is used to collect the time parameters of multiple cycles of the motor under various operating conditions, the initial temperature of the motor to be tested, and the running time of the motor to be tested corresponding to each cycle to form a training set and a prediction set, wherein the time parameters at least include: copper loss and iron loss information of the corresponding cycle;
[0049] A training processing module, used to use the current time parameter, t-1 time parameter, t-2 time parameter, the initial temperature of the motor to be tested, and the running time of the motor to be tested in the training set as input training sample data, and the corresponding motor end temperature in the training set as output training sample data, train the neural network prediction model, and use the prediction set for prediction processing;
[0050] The training end module is used to obtain the trained neural network prediction model after the training stop condition is reached.
[0051] Wherein, the temperature prediction processing module further includes a result output processing module, and the result output processing module further includes:
[0052] An output module, used for outputting the predicted temperature obtained by the neural network prediction model to the CAN bus;
[0053] The alarm processing module is used to compare the predicted temperature with a preset maximum allowable temperature. When the predicted temperature exceeds the maximum allowable temperature, an alarm operation is triggered and a shutdown command is sent to the motor controller to stop the motor from working.
[0054] Accordingly, as another aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the motor temperature prediction method as described above are implemented.
[0055] Accordingly, as another aspect of the present invention, a vehicle is provided, in which the motor temperature prediction system as described above is deployed.
[0056] Implementing the embodiments of the present invention has the following beneficial effects:
[0057] The present invention provides a motor temperature prediction method, system, storage medium and vehicle. When predicting the motor temperature, the rotation frequency information of the motor is first obtained, and the rotation frequency information is compared with a predetermined frequency threshold to determine whether the motor temperature prediction process needs to be triggered, which can save computing resources, avoid some invalid calculations, and play a role in giving full play to the working capacity of the motor components;
[0058] At the same time, in the scheme adopted by the present invention, the influence of factors such as time accumulation effect, ambient temperature, initial motor temperature, motor coolant temperature, motor copper loss, motor iron loss, motor running time, etc. on the temperature at the end of the motor is considered at the same time, and the above are input into the prediction model as parameters to improve the prediction accuracy; the present invention also effectively utilizes historical data (vehicle parameters of multiple cycles), uses the LSTM algorithm, and fully considers the influence of historical data on the motor temperature, which can improve the prediction accuracy, reduce the dependence on temperature sensors, and reduce the application cost of the whole vehicle; thereby preventing the motor from overheating and burning, and improving the use efficiency of the motor. BRIEF DESCRIPTION OF THE DRAWINGS
[0059] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For a person skilled in the art, other drawings obtained based on these drawings still belong to the scope of the present invention without creative labor.
[0060] Figure 1 A schematic diagram of the main process of an embodiment of a motor temperature prediction method provided by the present invention;
[0061] Figure 2 A schematic diagram of the principles involved in the method provided by the present invention;
[0062] Figure 3 A schematic diagram of the structure of an embodiment of a motor temperature prediction system provided by the present invention;
[0063] Figure 4 for Figure 3 A schematic diagram of the structure of the module for obtaining the motor related information;
[0064] Figure 5 for Figure 3 The structural diagram of the model training processing module in the figure;
[0065] Figure 6 for Figure 3 The structural diagram of the temperature prediction processing module;
[0066] Figure 7 for Figure 6 Schematic diagram of the structure of the result output processing module in . DETAILED DESCRIPTION
[0067] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings.
[0068] like Figure 1 FIG. 1 is a schematic diagram showing a main flow chart of an embodiment of a motor temperature prediction method provided by the present invention. Figure 2 As shown, in this embodiment, the motor temperature prediction method includes the following steps:
[0069] Step S10, obtain the rotation frequency information of the motor, and compare the rotation frequency information with the predetermined frequency threshold to determine whether it is necessary to trigger the motor temperature prediction process; it is understandable that since motor overheating usually occurs after the motor has been working continuously at an excessively high frequency for a certain period of time, and in order to save computing resources, avoid some invalid calculations, and give full play to the working capacity of components, in the present invention, it is necessary to judge the triggering conditions of the motor temperature prediction process. When the motor rotation frequency n is lower than the frequency threshold N, the motor temperature prediction process is not triggered; if the motor rotation frequency n is greater than or equal to the frequency threshold N, the motor temperature prediction process is triggered to start;
[0070] Step S11, when it is determined that the motor temperature prediction process needs to be triggered, obtaining time parameters of the current cycle and a predetermined number of previous cycles of the motor under various operating conditions, the initial temperature of the motor to be tested, and the operating time of the motor to be tested, wherein the time parameters at least include: copper loss and iron loss information of the corresponding cycle;
[0071] In a specific example, in step S11, obtaining time parameters of a current cycle and a predetermined number of previous cycles of the motor in various operating states specifically includes:
[0072] Obtain the time parameters of the motor at the current time t, the previous time parameters at the time t-1, and the previous time parameters at the time t-2 under various operating conditions, where:
[0073] The time parameters at the current time t include: current ambient temperature T1, current copper loss P cu 、Current iron loss P fe , current coolant temperature T2;
[0074] The time parameters at time t-1 include: ambient temperature T1 at time t-1 t-1 , copper loss P at time t-1 cu t-1 、Iron loss P at time t-1 fe t-1 , Motor coolant temperature T2 at time t-1 t-1 ;
[0075] The time parameters at time t-2 include: ambient temperature T1 at time t-2 t-2 , copper loss P at time t-2 cu t-2 , Iron loss P at time t-2 fe t-2 , Motor coolant temperature T2 at time t-2 t-2 ;
[0076] Among them, the copper loss information at each moment is calculated using the following formula:
[0077] P cu =r*i 2
[0078] Where P cu is the copper loss of the motor to be measured, i is the stator current of the motor, and r is the winding resistance of the motor;
[0079] The iron loss information at each moment is calculated using the following formula:
[0080] P fe =K h *B max 2 *f+K e *B max 2 *f
[0081] Where P fe is the iron loss of the motor to be tested, K h is the motor hysteresis loss coefficient, K e is the motor eddy current loss coefficient, B max is the maximum magnetic flux density of the motor, and f is the motor speed.
[0082] It is understandable that copper loss is affected by motor current, and iron loss is affected by motor speed. If the current and speed are constantly changing in the last few seconds, the copper loss and iron loss data will also be constantly changing. If the sudden change in current in the last few seconds leads to a sudden change in copper loss, and the sudden change in speed leads to a sudden change in iron loss. If the historical copper loss and iron loss data are not considered, the accuracy of the predicted data will be reduced; and the historical temperature of the motor coolant has a greater impact on the motor temperature.
[0083] Step S12, inputting the acquired information into a trained neural network prediction model, performing motor temperature prediction, obtaining the motor temperature prediction result and outputting it.
[0084] It is understandable that before step S12, the method further includes: step S00, the step of establishing a neural network prediction model and training the model, including:
[0085] Step S001, establishing a neural network prediction model using LSTM;
[0086] Step S002, collecting time parameters of multiple cycles of the motor under various operating conditions, the initial temperature of the motor to be tested, and the running time of the motor to be tested corresponding to each cycle to form a training set and a prediction set, wherein the time parameters at least include: copper loss and iron loss information of the corresponding cycle;
[0087] Step S003, using the current time parameter, t-1 time parameter, t-2 time parameter, the initial temperature of the motor to be tested, and the running time of the motor to be tested in the training set as input training sample data, and using the corresponding motor end temperature in the training set as output training sample data to train the neural network prediction model;
[0088] Step S004, after the training stop condition is reached, obtain the trained neural network prediction model.
[0089] It can be understood that in the present invention, the prediction algorithm adopted by the neural network prediction model is the LSTM algorithm. LSTM is a neural network with the ability to memorize long-term and short-term information. It has the ability to use hidden nodes with time variables such as t-1 and t-2 as the input of the current time slice t, that is, it can use the information of the previous time slice to calculate the content of the current time slice, which improves data utilization and prediction accuracy, and solves the problem that the hidden node input of the traditional motor prediction model only depends on the input characteristics of the current time slice.
[0090] Therefore, in step S13, according to the trained neural network prediction model, the current time parameter, t-1 time parameter, t-2 time parameter, the initial temperature of the motor to be tested, and the running time of the motor to be tested in the prediction set are input into the neural network model, and the temperature prediction result predicted by the model (i.e., the predicted motor end temperature) can be obtained; specifically, the input and output details of the neural network prediction model can refer to Figure 2 shown.
[0091] More specifically, in a specific example, in step S13, obtaining and outputting the motor temperature prediction result further includes:
[0092] Output the predicted temperature obtained by the neural network prediction model to the CAN bus;
[0093] The predicted temperature is compared with a preset maximum allowable temperature. When the predicted temperature exceeds the maximum allowable temperature, an alarm operation is triggered, and a shutdown instruction is sent to the motor controller to stop the motor from working.
[0094] It can be understood that in the motor temperature prediction method provided by the present invention, the time accumulation effects of copper loss, iron loss and coolant temperature are taken into account, so that the prediction accuracy is higher; at the same time, the motor frequency is considered and the LSTM prediction algorithm is adopted, which can reduce the waste of computing resources and obtain more accurate prediction results.
[0095] like Figure 3 FIG. 1 is a schematic diagram showing a structure of an embodiment of a motor temperature prediction system provided by the present invention. Figures 4 to 7 As shown, in this embodiment, the motor temperature prediction system 1 at least includes:
[0096] The prediction trigger judgment module 10 is used to obtain the rotation frequency information of the motor and compare the rotation frequency information with a predetermined frequency threshold to determine whether the motor temperature prediction process needs to be triggered;
[0097] The motor related information acquisition module 11 is used to acquire the time parameters of the current cycle and a predetermined number of previous cycles of the motor under various operating states, the initial temperature of the motor to be tested, and the running time of the motor to be tested when the prediction trigger judgment module determines that the motor temperature prediction needs to be triggered, wherein the time parameters at least include: copper loss and iron loss information of the corresponding cycle;
[0098] The temperature prediction processing module 12 is used to input the acquired information into a trained neural network prediction model to perform motor temperature prediction, obtain the motor temperature prediction result and output it;
[0099] The model training processing module 13 is used to establish a neural network prediction model and perform training to obtain a trained neural network prediction model.
[0100] Further, such as Figure 4 As shown, in a specific example, the motor related information acquisition module 11 includes:
[0101] The copper loss information calculation module 110 is used to calculate the copper loss information at each moment using the following formula:
[0102] P cu =r*i 2
[0103] Where P cu is the copper loss of the motor to be measured, i is the stator current of the motor, and r is the winding resistance of the motor;
[0104] The iron loss information calculation module 111 is used to calculate the iron loss information at each moment using the following formula:
[0105] P fe =K h *B max 2 *f+K e *B max 2 *f
[0106] Where P fe is the iron loss of the motor to be tested, K h is the motor hysteresis loss coefficient, K e is the motor eddy current loss coefficient, B max is the maximum magnetic flux density of the motor, and f is the motor speed;
[0107] The time parameter acquisition module 112 is used to obtain the time parameters of the motor at the current time t, the time parameters at the previous time t-1, and the time parameters at the previous time t-2 under various operating conditions, wherein:
[0108] The time parameters at the current time t include: current ambient temperature T1, current copper loss P cu 、Current iron loss P fe , current coolant temperature T2;
[0109] The time parameters at time t-1 include: ambient temperature T1 at time t-1 t-1 , copper loss P at time t-1 cu t-1 、Iron loss P at time t-1 fe t-1 , Motor coolant temperature T2 at time t-1 t-1 ;
[0110] The time parameters at time t-2 include: ambient temperature T1 at time t-2 t-2 , copper loss P at time t-2 cu t-2 , Iron loss P at time t-2 fe t-2 , Motor coolant temperature T2 at time t-2 t-2 ;
[0111] The other information obtaining module 113 is used to obtain the initial temperature of the motor to be tested and the running time of the motor to be tested.
[0112] like Figure 5 As shown, in a specific example, the model training processing module 13 further includes:
[0113] A model building module 130, used to build a neural network prediction model using LSTM;
[0114] The training set acquisition module 131 is used to collect time parameters of multiple cycles of the motor under various operating conditions, the initial temperature of the motor to be tested, and the running time of the motor to be tested corresponding to each cycle to form a training set and a prediction set, wherein the time parameters at least include: copper loss and iron loss information of the corresponding cycle;
[0115] The training processing module 132 is used to use the current time parameter, the t-1 time parameter, the t-2 time parameter, the initial temperature of the motor to be tested, and the running time of the motor to be tested in the training set as input training sample data, and use the corresponding motor end temperature in the training set as output training sample data to train the neural network prediction model, and use the prediction set for prediction processing;
[0116] The training end module 133 is used to obtain the trained neural network prediction model after the training stop condition is met.
[0117] like Figure 6 As shown, in a specific example, the temperature prediction processing module 12 further includes: a prediction algorithm task management module 124, and a data acquisition module 120, a data cache module 121, a data calculation module 122, an algorithm prediction module 125 and a result output processing module 123 respectively connected thereto.
[0118] The above motor related parameters are acquired through the data acquisition module 120 and transmitted to the data cache module 121 in real time by category.
[0119] The data cache module 121 caches relevant data according to parameter categories. The cache categories can be divided into ambient temperature T1, motor coolant temperature T2, current i, winding resistance r, motor speed f, motor hysteresis loss coefficient K h , Motor eddy current loss coefficient Ke 、Maximum magnetic flux density of the motor B max , the motor running time t, and each cached data has a timestamp, which makes it easy to confirm the data at time t, time t-1, and time t-2, and provides an interface for the data calculation module 122 to obtain data.
[0120] The prediction algorithm task management module 124 is responsible for controlling the startup of the data calculation module 122 and the algorithm prediction module 125, receiving the motor frequency n from the data acquisition module 120, and when it is determined whether the current motor movement frequency n is greater than or equal to the preset frequency threshold N, controlling the startup of the data calculation module 122 and the algorithm prediction module 125 to improve the use efficiency of the motor.
[0121] The data calculation module 122 is also controlled by the prediction algorithm task management module. When the start condition is met, the data calculation module 122 is started to calculate the cache data to obtain the current ambient temperature T1, the current copper loss P cu 、Current iron loss P fe , current coolant temperature T2, ambient temperature T1 at time t-1 t-1 , copper loss P at time t-1 cu t-1 、Iron loss P at time t-1 fe t-1 , Motor coolant temperature T2 at time t-1 t-1 , Ambient temperature T1 at time t-2 t-2 , copper loss P at time t-2 cu t-2 , Iron loss P at time t-2 fe t-2 , Motor coolant temperature T2 at time t-2 t-2 , motor initial temperature, current motor running time and other information, and outputs them to the algorithm prediction module 125.
[0122] The algorithm prediction module 125 receives the output data of the data calculation module 122 as the prediction input parameter of the LSTM neural network, then performs LSTM neural network operation and outputs the temperature prediction result (i.e. the current motor predicted temperature); when the predicted temperature is higher than the preset maximum acceptable temperature, an alarm instruction is triggered.
[0123] The result output processing module 123 further includes:
[0124] The output module is used to output the predicted temperature obtained by the neural network prediction model to the CAN bus for other systems or modules to obtain;
[0125] The alarm processing module is used to compare the predicted temperature with a preset maximum allowable temperature. When the predicted temperature exceeds the maximum allowable temperature, an alarm operation is triggered, and a stop command is sent to the motor controller in a service manner to stop the motor from working, so as to effectively protect the motor.
[0126] For more details, please refer to and combine the above Figure 1 to Figure 2 The description is not repeated here.
[0127] Accordingly, as another aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned Figure 1 and Figure 2 For more details, please refer to and combine with the above Figure 1 to Figure 2 The description is not repeated here.
[0128] Accordingly, as another aspect of the present invention, a vehicle is also provided, in which the above-mentioned Figures 3 to 7 For more details, please refer to and combine the above Figures 3 to 7 The description is not repeated here.
[0129] Implementing the embodiments of the present invention has the following beneficial effects:
[0130] The present invention provides a motor temperature prediction method, system, storage medium and vehicle. When predicting the motor temperature, the rotation frequency information of the motor is first obtained, and the rotation frequency information is compared with a predetermined frequency threshold to determine whether the motor temperature prediction process needs to be triggered, which can save computing resources, avoid some invalid calculations, and play a role in giving full play to the working capacity of the motor components;
[0131] At the same time, in the scheme adopted by the present invention, the influence of factors such as time accumulation effect, ambient temperature, initial motor temperature, motor coolant temperature, motor copper loss, motor iron loss, motor running time, etc. on the temperature at the end of the motor is considered at the same time, and the above are input into the prediction model as parameters to improve the prediction accuracy; the present invention also effectively utilizes historical data (vehicle parameters of multiple cycles), uses the LSTM algorithm, and fully considers the influence of historical data on the motor temperature, which can improve the prediction accuracy, reduce the dependence on temperature sensors, and reduce the application cost of the whole vehicle; thereby preventing the motor from overheating and burning, and improving the use efficiency of the motor.
[0132] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0133] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0134] The above disclosure is only a preferred embodiment of the present invention, which certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.
Claims
1. A method for predicting motor temperature, characterized in that: The steps include: Acquire the rotation frequency information of the motor, and compare the rotation frequency information with a predetermined frequency threshold to determine whether it is necessary to trigger the motor temperature prediction process; When it is determined that the motor temperature prediction process needs to be triggered, the time parameters of the current cycle and a predetermined number of previous cycles of the motor under various operating conditions, the initial temperature of the motor to be tested, and the operating time of the motor to be tested are obtained, and the time parameters at least include: copper loss and iron loss information of the corresponding cycle; The acquired information is input into a trained neural network prediction model to perform motor temperature prediction, obtain the motor temperature prediction result and output it.
2. The method according to claim 1, characterized in that Further comprising: establishing a neural network prediction model and performing training steps, including: Build a neural network prediction model using LSTM; Collecting time parameters of multiple cycles of the motor under various operating conditions, the initial temperature of the motor to be tested, and the running time of the motor to be tested corresponding to each cycle to form a training set, wherein the time parameters at least include: copper loss and iron loss information of the corresponding cycle; The current time parameter, the t-1 time parameter, the t-2 time parameter, the initial temperature of the motor to be tested, and the running time of the motor to be tested in the training set are used as input training sample data, and the corresponding motor end temperature in the training set is used as output training sample data, the neural network prediction model is trained, and the prediction set is used for prediction processing; After the training stop condition is reached, the trained neural network prediction model is obtained.
3. The method according to claim 1 or 2, characterized in that Get the time parameters of the current cycle and a predetermined number of previous cycles of the motor in various operating states, including: Obtain the time parameters of the motor at the current time t, the previous time parameters at the time t-1, and the previous time parameters at the time t-2 under various operating conditions, where: The time parameters at the current time t include: current ambient temperature T1, current copper loss P cu 、Current iron loss P fe , current coolant temperature T2; The time parameters at time t-1 include: ambient temperature T1 at time t-1 t-1 , copper loss P at time t-1 cut-1 、Iron loss P at time t-1 fet-1 , Motor coolant temperature T2 at time t-1 t-1 ; The time parameters at time t-2 include: ambient temperature T1 at time t-2 t-2 , copper loss P at time t-2 cut-2 , Iron loss P at time t-2 fet-2 , Motor coolant temperature T2 at time t-2 t-2 ; Among them, the copper loss information at each moment is calculated using the following formula: P cu =r*i 2 Where P cu is the copper loss of the motor to be measured, i is the stator current of the motor, and r is the winding resistance of the motor; The iron loss information at each moment is calculated using the following formula: P fe =K h *B max 2 *f+K e *B max 2 *f Where P fe is the iron loss of the motor to be tested, K h is the motor hysteresis loss coefficient, K e is the motor eddy current loss coefficient, B max is the maximum magnetic flux density of the motor, and f is the motor speed.
4. The method according to claim 3, characterized in that The step of obtaining and outputting the motor temperature prediction result further includes: Output the predicted temperature obtained by the neural network prediction model to the CAN bus; The predicted temperature is compared with a preset maximum allowable temperature. When the predicted temperature exceeds the maximum allowable temperature, an alarm operation is triggered, and a shutdown instruction is sent to the motor controller to stop the motor from working.
5. A motor temperature prediction system, characterized in that: include: A prediction trigger judgment module is used to obtain the rotation frequency information of the motor and compare the rotation frequency information with a predetermined frequency threshold to determine whether the motor temperature prediction process needs to be triggered; A motor related information acquisition module is used to acquire the time parameters of the current cycle and a predetermined number of previous cycles of the motor under various operating states, the initial temperature of the motor to be tested, and the operating time of the motor to be tested when the prediction trigger judgment module determines that the motor temperature prediction process needs to be triggered. The time parameters at least include: copper loss and iron loss information of the corresponding cycle; The temperature prediction processing module is used to input the acquired information into a trained neural network prediction model to perform motor temperature prediction, obtain the motor temperature prediction result and output it.
6. The system according to claim 5, characterized in that Further including: The model training processing module is used to establish a neural network prediction model and perform training, which includes: Model building module, used to build a neural network prediction model using LSTM; The training set acquisition module is used to collect the time parameters of multiple cycles of the motor under various operating conditions, the initial temperature of the motor to be tested, and the running time of the motor to be tested corresponding to each cycle to form a training set and a prediction set, wherein the time parameters at least include: copper loss and iron loss information of the corresponding cycle; A training processing module, used to use the current time parameter, t-1 time parameter, t-2 time parameter, the initial temperature of the motor to be tested, and the running time of the motor to be tested in the training set as input training sample data, and the corresponding motor end temperature in the training set as output training sample data, train the neural network prediction model, and use the prediction set for prediction processing; The training end module is used to obtain the trained neural network prediction model after the training stop condition is reached.
7. The system according to claim 5 or 6, characterized in that The motor related information acquisition module includes: The copper loss information calculation module is used to calculate the copper loss information at each moment using the following formula: P cu =r*i 2 Where P cu is the copper loss of the motor to be measured, i is the stator current of the motor, and r is the winding resistance of the motor; The iron loss information calculation module is used to calculate the iron loss information at each moment using the following formula: P fe =K h *B max 2 *f+K e *B max 2 *f Where P fe is the iron loss of the motor to be tested, K h is the motor hysteresis loss coefficient, K e is the motor eddy current loss coefficient, B max is the maximum magnetic flux density of the motor, and f is the motor speed; The time parameter acquisition module is used to obtain the time parameters of the motor at the current time t, the time parameters at the previous time t-1, and the time parameters at the previous time t-2 under various operating conditions, wherein: The time parameters at the current time t include: current ambient temperature T1, current copper loss P cu 、Current iron loss P fe , current coolant temperature T2; The time parameters at time t-1 include: ambient temperature T1 at time t-1 t-1 , copper loss P at time t-1 cut-1 、Iron loss P at time t-1 fet-1 , Motor coolant temperature T2 at time t-1 t-1 ; The time parameters at time t-2 include: ambient temperature T1 at time t-2 t-2 , copper loss P at time t-2 cut-2 , Iron loss P at time t-2 fet-2 , Motor coolant temperature T2 at time t-2 t-2 ; The other information acquisition module is used to obtain the initial temperature of the motor to be tested and the running time of the motor to be tested.
8. The system according to claim 7, characterized in that The temperature prediction processing module further includes a result output processing module, and the result output processing module further includes: An output module, used for outputting the predicted temperature obtained by the neural network prediction model to the CAN bus; The alarm processing module is used to compare the predicted temperature with a preset maximum allowable temperature. When the predicted temperature exceeds the maximum allowable temperature, an alarm operation is triggered and a shutdown command is sent to the motor controller to stop the motor from working.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the motor temperature prediction method according to any one of claims 1 to 4 are implemented.
10. A vehicle, characterized in that: The motor temperature prediction system according to any one of claims 5 to 8 is deployed in the vehicle.
