Vehicle motor temperature control method and device and vehicle motor temperature prediction method

Through graph convolution feature extraction and iterative optimization, the operation and cold liquid flow data of vehicle motors are solved, and the problem of low temperature control efficiency of vehicle motors in the prior art is achieved, and more efficient temperature control and motor life extension are achieved.

CN118971730AActive Publication Date: 2024-11-15CHONGQING JINKANG POWER NEW ENERGY CO LTD

Patent Information

Application Number
CN202410935994.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-12
Publication Date
2024-11-15
Estimated Expiration
2044-07-12

AI Technical Summary

Technical Problem

The existing vehicle motor temperature control methods are relatively low in efficiency, which makes it difficult to effectively control the motor temperature, affecting the working efficiency and service life of the motor.

Method used

By obtaining the target operation timing data of the vehicle motor and the cooling liquid flow timing data, iterative optimization is performed using graph convolution feature extraction processing and temperature prediction model, and the running subsequence and cooling liquid flow quantum sequence are divided in the time window to optimize the cooling liquid flow rate. To control the motor temperature.

Benefits of technology

It improves the efficiency of vehicle motor temperature control, reduces the calculation amount, improves control speed and accuracy, and maintains the motor temperature more stably at a lower level, extending the service life of the motor.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118971730A_ABST
    Figure CN118971730A_ABST
Patent Text Reader

Abstract

The invention relates to a vehicle motor temperature control method and device and a vehicle motor temperature prediction method. The method comprises the steps of performing iterative optimization processing by minimizing output data of a motor temperature prediction model based on target operation time sequence data corresponding to a vehicle motor at a target moment and multiple pieces of cold liquid flow acquisition data, and obtaining target flow control data for controlling the cold liquid flow for cooling the vehicle motor; the motor temperature prediction model is used for dividing input operation time sequence data and cold liquid flow time sequence data into a plurality of operation subsequences and a plurality of cold liquid flow subsequences, and part of data in each subsequence is used for graph convolution feature extraction processing and temperature prediction processing; cold liquid flow data included in a part of cold liquid flow subsequences in cold liquid flow time sequence data input by the motor temperature prediction model in each iterative optimization process is cold liquid flow candidate data in the current iterative optimization process. By adopting the method, the vehicle motor temperature control efficiency can be improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of vehicle motor control, and in particular to a vehicle motor temperature control method, device and vehicle motor temperature prediction method. Background Art

[0002] Permanent magnet synchronous motors have gradually become the mainstream choice for new energy vehicle drive motors due to their own advantages such as high power density, high efficiency, fast response speed, and wide speed regulation range. However, there will inevitably be energy loss in the process of converting electrical energy into mechanical energy. For motor systems with rated power ranging from tens of kilowatts to hundreds of kilowatts, a small energy loss (for example, less than 5%) is enough to cause a high temperature rise in the motor. Among them, thermal energy will cause permanent damage such as aging to the motor insulation system, and will also cause demagnetization of the permanent magnets in the permanent magnet synchronous motor, affecting the control accuracy of the motor controller and the transmission efficiency of the motor.

[0003] The vehicle motor temperature control method in the related technology constructs a digital twin model of the motor temperature field, and then predicts the temperature data of the motor at future times through the digital twin model of the motor temperature field and the real-time operating data collected under the current working conditions, and solves the optimal cooling liquid flow under the current working conditions, so that the maximum temperature of the motor in the future period of time is maintained at a low level.

[0004] However, the above vehicle motor temperature control method has low efficiency. Summary of the invention

[0005] Based on this, it is necessary to provide a vehicle motor temperature control method, device and vehicle motor temperature prediction method that can improve efficiency in response to the above technical problems.

[0006] In a first aspect, the present application provides a vehicle motor temperature control method, comprising:

[0007] Obtaining target operation timing data and a plurality of cooling liquid flow collection data corresponding to the vehicle motor at a target time, wherein the target operation timing data includes a plurality of operation collection data;

[0008] Based on the target operating timing data and multiple cold liquid flow collection data, iterative optimization processing is performed by minimizing the output data of the motor temperature prediction model to obtain the target flow control data. The motor temperature prediction model is used to divide the input operating timing data and cold liquid flow timing data into multiple operating subsequences and multiple cold liquid flow subsequences according to a preset time window, and use part of the data in each operating subsequence and each cold liquid flow subsequence to perform graph convolution feature extraction processing and temperature prediction processing, and output temperature prediction data corresponding to the input operating timing data and cold liquid flow timing data, wherein, in each iterative optimization process, the cold liquid flow data included in the part of the cold liquid flow subsequence in the cold liquid flow timing data input by the motor temperature prediction model is the cold liquid flow candidate data corresponding to the current iterative optimization process;

[0009] A flow rate of coolant for cooling a motor of a vehicle is controlled based on the target flow rate control data.

[0010] In a second aspect, the present application also provides a vehicle motor temperature prediction method, comprising:

[0011] Obtaining the operation timing data and coolant flow timing data corresponding to the vehicle motor at the target time;

[0012] Dividing the operation time series data and the cooling liquid flow time series data into a plurality of operation subsequences and a plurality of cooling liquid flow subsequences according to a preset time window;

[0013] Perform graph convolution feature extraction processing on part of the data in each operation subsequence and each cold liquid flow subsequence to obtain the target operation convolution feature and target flow convolution feature corresponding to the target time;

[0014] Prediction processing is performed based on the target operation convolution characteristics and the target flow convolution characteristics to obtain the target predicted temperature data corresponding to the vehicle motor at the target time.

[0015] In a third aspect, the present application also provides a vehicle motor temperature control device, the device comprising:

[0016] A data acquisition module, used to acquire target operation timing data and a plurality of cooling liquid flow acquisition data corresponding to the vehicle motor at a target time, wherein the target operation timing data includes a plurality of operation acquisition data;

[0017] A flow optimization module is used to obtain target flow control data by iteratively optimizing the target operating timing data and multiple cold liquid flow acquisition data by minimizing the output data of the motor temperature prediction model, wherein the motor temperature prediction model is used to divide the input operating timing data and cold liquid flow timing data into multiple operating subsequences and multiple cold liquid flow subsequences according to a preset time window, and use part of the data in each operating subsequence and each cold liquid flow subsequence to perform graph convolution feature extraction and temperature prediction processing, and output temperature prediction data corresponding to the input operating timing data and cold liquid flow timing data; wherein, in each iterative optimization process, the cold liquid flow data included in the part of the cold liquid flow subsequence in the cold liquid flow timing data input by the motor temperature prediction model is the cold liquid flow candidate data corresponding to the current iterative optimization process;

[0018] A flow control module is used to control a flow of a coolant for cooling a motor of a vehicle based on target flow control data.

[0019] The above-mentioned vehicle motor temperature control method, device and vehicle motor temperature prediction method obtain the target operating timing data and multiple cooling liquid flow collection data corresponding to the vehicle motor at the target time, wherein the target operating timing data includes multiple operating collection data, and based on the target operating timing data and multiple cooling liquid flow collection data, iterative optimization processing is performed by minimizing the output data of the motor temperature prediction model to obtain the target flow control data; based on the target flow control data, the cooling liquid flow used to cool the vehicle motor is controlled; wherein the motor temperature prediction model is used to divide the input operating timing data and cooling liquid flow timing data into multiple operating subsequences and multiple cooling liquid flow subsequences according to a preset time window, and use part of the data in each operating subsequence and each cooling liquid flow subsequence to perform graph convolution feature extraction processing and temperature prediction processing, and output temperature prediction data corresponding to the input operating timing data and cooling liquid flow timing data, wherein, in each iterative optimization process, part of the cooling liquid flow subsequences in the cooling liquid flow timing data input by the motor temperature prediction model include The cooling liquid flow data is the cooling liquid flow candidate data corresponding to the current iterative optimization process; in the process of determining the target flow control data, multiple iterations are required, and each iterative process needs to use the motor temperature prediction model to obtain the temperature prediction data corresponding to the cooling liquid flow candidate data of the current iterative process, that is, in the process of determining the target flow control data, it is necessary to call the motor temperature prediction model multiple times for calculation and processing. The motor temperature prediction model provided in the above embodiment divides the input time series data into multiple subsequences according to a preset time window, and uses part of the data in each subsequence for graph convolution feature extraction and temperature prediction processing, which can avoid the need for the motor temperature field digital twin model in the related technology to perform feature extraction processing on each data in the input time series data, which has a large amount of calculation and leads to low efficiency of the control process. The motor temperature prediction model provided by the above embodiment can reduce the amount of calculation corresponding to each iterative process, thereby reducing the overall amount of calculation for determining the target flow control data, and improving the speed of obtaining the target flow control data, so that the control efficiency is higher. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 A diagram showing an application environment of a vehicle motor temperature control method in one embodiment;

[0022] Figure 2A schematic flow chart of a vehicle motor temperature control method according to an embodiment;

[0023] Figure 3 A schematic diagram of a flow chart of steps for obtaining target flow control data in one embodiment;

[0024] Figure 4 A schematic diagram of a flow chart of a step of obtaining candidate temperature prediction data in one embodiment;

[0025] Figure 5 A schematic flow chart of a vehicle motor temperature control method in another embodiment;

[0026] Figure 6 A schematic diagram of a process of obtaining a speed diagram structure, a torque diagram structure and a flow diagram structure in one embodiment;

[0027] Figure 7 It is a schematic diagram of a process of obtaining an intermediate speed convolution feature, an intermediate torque convolution feature and an intermediate flow convolution feature in one embodiment;

[0028] Figure 8 It is a schematic diagram of a process of obtaining a target speed convolution feature, a target torque convolution feature and a candidate flow convolution feature in one embodiment;

[0029] Fig. 9 A schematic diagram of a process flow of a vehicle motor temperature prediction method in one embodiment;

[0030] Fig.10 is a structural block diagram of a vehicle motor temperature control device provided in one embodiment;

[0031] Fig.11 A structural block diagram of a vehicle motor temperature prediction device provided in one embodiment;

[0032] Fig.12 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0034] The vehicle motor temperature control method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown. The on-board controller 102 is connected to the motor 104, and the on-board controller 102 is used to obtain the real-time operation data corresponding to the motor 104 and control the cooling liquid flow corresponding to the motor 104; the on-board controller 102 can be, but is not limited to, an electronic control unit (Electronic Control Unit, referred to as ECU) or a microcontroller unit (Microcontroller Unit, referred to as MCU); in some embodiments, the motor can be a permanent magnet synchronous motor, and in other embodiments, the motor 104 can also be other types of motors, such as an AC asynchronous motor.

[0035] Among them, the on-board controller 102 obtains the target operating timing data and multiple cold liquid flow collection data corresponding to the vehicle motor at the target time, wherein the target operating timing data includes multiple operating collection data; based on the target operating timing data and multiple cold liquid flow collection data, iterative optimization processing is performed by minimizing the output data of the motor temperature prediction model to obtain target flow control data, and the motor temperature prediction model is used to divide the input operating timing data and cold liquid flow timing data into multiple operating subsequences and multiple cold liquid flow subsequences according to a preset time window, and use part of the data in each operating subsequence and each cold liquid flow subsequence to perform graph convolution feature extraction processing and temperature prediction processing, and output temperature prediction data corresponding to the input operating timing data and cold liquid flow timing data, wherein, in each iterative optimization process, the cold liquid flow data included in part of the cold liquid flow subsequences in the cold liquid flow timing data input by the motor temperature prediction model is the cold liquid flow candidate data corresponding to the current iterative optimization process; based on the target flow control data, the cold liquid flow used to cool the vehicle motor is controlled.

[0036] In an exemplary embodiment, Figure 2 As shown, a vehicle motor temperature control method is provided, and the method is applied to Figure 1 The vehicle-mounted controller 102 in FIG. 1 is taken as an example to illustrate, and the following steps 202 to 206 are included. Among them:

[0037] Step 202 , obtaining target operating timing data corresponding to the vehicle motor at the target time and a plurality of cooling liquid flow rate collection data.

[0038] The target time refers to the time when the cooling liquid flow needs to be determined, which may be a periodic time; illustratively, the vehicle-mounted controller updates the cooling liquid flow size every 10 seconds.

[0039] The target operation timing data includes a plurality of operation collection data. The vehicle controller 102 continuously collects the operation status data of the vehicle motor according to a certain sampling frequency; the target operation timing data refers to a plurality of operation collection data within a preset time period before the target time period. Exemplarily, the target operation timing data includes a first number of operation collection data; Exemplarily, the first number is determined by the collection frequency and the preset time period.

[0040] In a possible implementation, the operation acquisition data includes speed acquisition data and torque acquisition data, and correspondingly, the target operation timing data includes target speed timing data and target torque timing data. The speed refers to the motor speed, and the speed acquisition data can be obtained through a rotary encoder; the torque refers to the motor output torque, and the torque speed acquisition data can be obtained through a torque sensor. The motor speed and output torque can reflect the load and working conditions of the motor.

[0041] In another possible implementation, the operation acquisition data may include speed acquisition data, torque acquisition data, motor bus voltage data, and motor bus current data.

[0042] In another possible implementation, the operation collection data includes rotation speed collection data and motor output power.

[0043] Among them, the coolant flow rate refers to the amount of coolant flowing through the coolant pipe between units; in the embodiment of the present application, two types of coolant flow data are included, one is the coolant flow collection data; illustratively, the coolant flow collection data can be obtained by a coolant flow sensor; the other is the coolant flow control data, which is calculated by the motor temperature control method provided in this embodiment and is used to adjust the coolant flow. Exemplarily, multiple cold night flow collection data can refer to the cold night flow collection data obtained by the vehicle controller 102 within a preset time period before the target time and the target time.

[0044] Step 204 , based on the target operating timing data and the plurality of cooling liquid flow rate collection data, iterative optimization processing is performed by minimizing the output data of the motor temperature prediction model to obtain the target flow rate control data.

[0045] Among them, the motor temperature prediction model is used to divide the input operation timing data and cooling liquid flow timing data into multiple operation subsequences and multiple cooling liquid flow subsequences according to a preset time window, and use part of the data in each operation subsequence and each cooling liquid flow subsequence to perform graph convolution feature extraction and temperature prediction processing, and output temperature prediction data corresponding to the input operation timing data and cooling liquid flow timing data.

[0046] Among them, the output data of the motor temperature prediction model includes the temperature corresponding to at least one position in the vehicle motor at at least one prediction moment. In one possible implementation, the output data of the motor temperature prediction model includes the surface temperature value of the motor at a prediction moment corresponding to the target moment, and exemplarily, the prediction moment is a moment after the target moment. In one possible implementation, the output data of the motor temperature prediction model includes the temperature values ​​corresponding to multiple positions on the surface of the motor at the prediction moment corresponding to the target moment. In one possible implementation, the output data of the motor temperature prediction model includes the surface temperature values ​​of the motor at multiple prediction moments corresponding to the target moment, wherein the multiple prediction moments are multiple moments after the target moment, and exemplarily, the multiple prediction moments are 10 prediction moments, and the output data is in the form of [(prediction moment 1, temperature 1), (prediction moment 2, temperature 2),…, (prediction moment 10, temperature 10)]. In one possible implementation, the output data of the motor temperature prediction model includes the temperature values ​​corresponding to multiple positions on the surface of the motor at multiple prediction moments corresponding to the target moment; exemplarily, the output data of the motor temperature prediction model includes the temperature values ​​corresponding to 5 positions on the surface of the motor at 10 prediction moments corresponding to the target moment.

[0047] In a possible implementation manner, when the output data of the motor temperature prediction model includes multiple temperature values, minimization in the iterative process may refer to minimizing the maximum value of the multiple temperature prediction values ​​output by the motor temperature prediction model.

[0048] In this embodiment, the operation time series data and the cold liquid flow time series data are first divided into multiple operation subsequences and multiple cold liquid flow subsequences according to a preset time window; illustratively, the number of data in each operation subsequence is the same, and the number of data in each cold liquid flow subsequence is the same; the amount of data in the operation subsequence and the number of data in the cold liquid subsequence can be the same or different, and can be specifically determined according to the collection frequency of the operation collection data and the collection frequency of the cold liquid flow collection data.

[0049] In this embodiment, considering that the motor operating state and the cooling liquid flow rate do not change much in a very short period of time, for each time window, part of the data is selected from the operating subsequence and the cooling liquid flow subsequence as representative data to participate in the subsequent graph convolution feature extraction process and temperature prediction process, so as to save the computational complexity of the feature extraction process and the temperature prediction process and improve the efficiency of temperature prediction.

[0050] In a possible implementation, the time window is less than 3 seconds. Optionally, the time window is 2 seconds.

[0051] In each iterative optimization process, the cooling liquid flow data included in a part of the cooling liquid flow subsequences in the cooling liquid flow time series data input by the motor temperature prediction model is the cooling liquid flow candidate data corresponding to the current iterative optimization process.

[0052] Among them, the output data of the motor temperature prediction model that is minimized is used as the objective function of iterative optimization. The cold liquid flow candidate data in the iterative process needs to be used as a partial input of the motor temperature prediction model to determine the motor temperature prediction data corresponding to the cold liquid flow candidate data as the fitness value of the cold liquid flow candidate data, and to judge whether the cold liquid flow candidate data is the optimal solution of the current iterative process.

[0053] In each iterative optimization process, the running time data input by the motor temperature prediction model is the target time series running data, and the input cold liquid flow time series data is determined based on multiple cold liquid flow collection data and the cold liquid flow candidate data of the current iterative process. In one possible implementation, the cold liquid flow candidate data of the current iterative process is used as the cold liquid flow data in a cold liquid flow subsequence that is closest to the target moment in time in the cold liquid flow time series data, and the cold liquid flow data in the remaining cold liquid subsequences are cold liquid flow collection data, so as to ensure that the cold liquid flow candidate data participates in the graph convolution feature extraction processing process and the temperature prediction processing process. In another possible implementation, the cold liquid flow candidate data corresponding to the current iterative process is used as the cold liquid flow data in two or more cold liquid flow subsequences that are closest to the target moment in time in the cold liquid quantity time series data, and the cold liquid flow data in the remaining cold liquid subsequences are cold liquid flow collection data, so as to ensure that the cold liquid flow candidate data participates in the graph convolution feature extraction processing process and the temperature prediction processing process.

[0054] For ease of description, the number of data in each cold liquid subsequence is m, there are n cold liquid subsequences in total, and the number of data in the entire cold liquid flow time series data is n×m. In a possible implementation, the data close to the target moment in the cold liquid flow time series data (n×m in total) input based on the motor temperature prediction model is filled with cold liquid flow candidate data. Exemplarily, the cold liquid flow data in a cold liquid flow subsequence closest to the target moment is the cold liquid flow candidate data, so that the multiple cold liquid flow acquisition data obtained in step 204 can be (n-1)×m data that are far away from the target moment in time series within a preset time length before the target moment.

[0055] In another possible implementation, the number of cooling liquid flow acquisition data acquired in step 204 may be n×m. When constructing the input data for the motor temperature prediction model, the cooling liquid flow candidate data is used to replace the cooling liquid flow data in the corresponding cooling liquid subsequence.

[0056] Exemplarily, the iterative optimization process can be described as follows:

[0057] minimize max(f(x 1 ,x 2 )), the constraints are: x 2L <x 2 <x 2H ;

[0058] Among them, x 1 Refers to the runtime data, x 2 Refers to the cooling liquid timing data, x 2L Refers to the minimum value that the coolant flow can be set to, x 2H Refers to the maximum value that the cooling liquid flow can be set to, f(x 1 ,x 2 ) represents the output data of the motor temperature prediction model.

[0059] Among them, the optimization algorithm can be used to find the optimal cooling liquid flow control data as the target flow control data corresponding to the target time based on the target operation data corresponding to the target time and multiple cooling liquid flow collection data. Based on the target flow control data as a part of the cooling liquid flow timing data, the temperature value output by the motor temperature prediction model is minimized.

[0060] In a possible implementation, the whale optimization algorithm is used for iterative optimization to obtain the target flow control data, wherein the whale optimization algorithm includes three actions: surrounding the prey, spirally attacking the prey, and randomly searching for the prey.

[0061] Among them, encircling the prey means taking the current optimal candidate solution as the target prey (i.e., the optimal solution); after knowing the position of the prey, the whale group starts to update the position according to the current relationship between itself and the prey; the action of encircling the prey can be adopted by formula 1 and formula

[0062] Formula 2 indicates:

[0063]

[0064] Among them, t represents the current iteration number, and represents the position of the best candidate solution before t generations, and represent the solutions of generation t and the next generation respectively. and is the coefficient vector and needs to be calculated separately. and The calculation methods are shown in Formula 3 and Formula 4 respectively:

[0065]

[0066] in, and Represents two random numbers between 0 and 1. is a number that decreases linearly from 2 to 0 with each iteration, The calculation formula is shown in Formula 5:

[0067]

[0068] Where G represents the total number of iterations.

[0069] As a system vector, it represents a control factor, also known as the amplification factor (Amplitude). The amplification factor is an important parameter used to adjust the movement step size of whale group members. During the update process of the whale algorithm, each whale updates its position based on the relationship between its current position and the position of the target prey (optimal solution). Amplification factor In this process, it plays the role of scaling step. By adjusting the size of the magnification factor, the range and speed of the whale's movement can be controlled.

[0070] The spiral attack refers to the simulation of the spiral process of the humpback whale, that is, the foaming net attack process; by establishing a spiral equation between the whale and the prey, the spiral movement of the humpback whale is simulated; the total work of the spiral can be expressed by formula 6 and formula 7:

[0071]

[0072] Where b is a constant, which is usually set to 1. l is a random number between [-1,1], which can be determined according to Formula 8:

[0073] l=(a 2 -1)·r 3 +1, Formula 8, where where r 3 A random number between 0 and 1.

[0074] The choice of these two behaviors, encircling the prey and spiral attack, is randomly selected. When the probability p<0.5, encircling the prey is selected, otherwise spiral attack is selected. It can be expressed by formula 9:

[0075]

[0076] Among them, the random search action is used in the whale algorithm for global search and a way to jump out of the local optimum; the random search action can be expressed by formula 10 and formula 11:

[0077]

[0078] in, is a randomly selected whale (solution). It can be seen that the formula for random search is very similar to the formula for encircling prey. The only difference is that the direction is adjusted from the current optimal candidate solution to a random whale. Whether encircling prey and random search behavior occur depends on the value of A. When Use siege behavior; when Adopt random search behavior; that is, in formula 9, when p < 0.5, if Then the random search action expressed by formula 10 and formula 11 is adopted. If Then the action of surrounding the prey expressed by Formula 1 and Formula 2 is adopted.

[0079] Among them, when When The value of is close to 0 and between -1 and 1. At this time, choosing to surround the prey behavior means adopting a strategy to approach, surround or concentrate the target. When When the value of exceeds 1 and is greater than any value close to 0, choosing random search behavior means adopting an exploratory strategy to find new possibilities or solutions.

[0080] In a possible implementation, a bat optimization algorithm or a fruit fly optimization algorithm or other swarm intelligence optimization algorithms are used to perform iterative optimization to obtain target flow control data.

[0081] Step 206 : Control the flow rate of the coolant for cooling the motor of the vehicle based on the target flow rate control data.

[0082] Among them, the target flow control data is the cooling liquid flow control value that can minimize the output data of the motor temperature prediction model at the target time; controlling the cooling liquid flow for cooling the vehicle motor based on the target flow control data can keep the temperature value of the vehicle motor at the prediction time at as low a level as possible, thereby improving the working efficiency and service life of the motor.

[0083] The vehicle motor temperature control method provided in the above embodiment includes obtaining target operating timing data and multiple cooling liquid flow collection data corresponding to the vehicle motor at the target time, wherein the target operating timing data includes multiple operating collection data, and based on the target operating timing data and multiple cooling liquid flow collection data, iterative optimization processing is performed by minimizing the output data of the motor temperature prediction model to obtain target flow control data; based on the target flow control data, the cooling liquid flow used to cool the vehicle motor is controlled; wherein the motor temperature prediction model is used to divide the input operating timing data and cooling liquid flow timing data into multiple operating subsequences and multiple cooling liquid flow subsequences according to a preset time window, and use part of the data in each operating subsequence and each cooling liquid flow subsequence to perform graph convolution feature extraction processing and temperature prediction processing, and output temperature prediction data corresponding to the input operating timing data and cooling liquid flow timing data, wherein in each iterative optimization process, part of the cooling liquid flow subsequences in the cooling liquid flow timing data input by the motor temperature prediction model include cooling liquid flow The data is the candidate data of the cooling liquid flow corresponding to the current iterative optimization process; in the process of determining the target flow control data, multiple iterations are required, and each iterative process needs to use the motor temperature prediction model to obtain the temperature prediction data corresponding to the candidate data of the cooling liquid flow of the current iterative process, that is, in the process of determining the target flow control data, it is necessary to call the motor temperature prediction model multiple times for calculation and processing. The motor temperature prediction model provided in the above embodiment divides the input time series data into multiple subsequences according to a preset time window, and uses part of the data in each subsequence for graph convolution feature extraction and temperature prediction. This can avoid the problem that the motor temperature field digital twin model in the related technology needs to perform feature extraction on each data in the input time series data, which has a large amount of calculation and leads to low efficiency of the control process; the motor temperature prediction model provided by the above embodiment can reduce the amount of calculation corresponding to each iterative process, thereby reducing the overall amount of calculation for determining the target flow control data, and improving the speed of obtaining the target flow control data, so that the control efficiency is higher.

[0084] In an exemplary embodiment, based on Figure 2 The embodiment shown in the figure is related to a process of obtaining target flow control data by iteratively optimizing the output data of the motor temperature prediction model based on the target operation timing data and multiple cooling liquid flow collection data. Please refer to Figure 3 , the process includes steps 302 to 304, wherein:

[0085] Step 302, based on multiple cold liquid flow collection data and cold liquid flow candidate data in the current iteration process, obtain candidate flow time series data, and the flow data corresponding to at least one time window close to the target time in the candidate flow time series data are all cold liquid flow candidate data.

[0086] In a possible implementation, the flow data corresponding to the first time window close to the target time in time series among the candidate flow time series data are all cold liquid flow candidate data.

[0087] In a possible implementation, the flow data corresponding to the first time window and the second time window in the candidate flow time series data that are close to the target time in time series are both cold liquid flow data.

[0088] Step 304: input the target operation timing data and the candidate flow timing data into the motor temperature prediction model to obtain the candidate temperature prediction data corresponding to the coolant flow candidate data.

[0089] In a possible implementation, when inputting the target operation timing data and the candidate flow timing data into the motor temperature prediction model, the target operation timing data and the candidate flow timing data are first normalized, and the normalized target operation timing data and the normalized candidate flow timing data are obtained and input into the motor temperature prediction model. By normalizing the target operation timing data and the candidate flow timing data, the dimensioned data is converted into dimensionless data, which facilitates the unification of the data scale in the subsequent feature extraction stage.

[0090] In a possible implementation manner, the motor temperature prediction model includes a normalization layer for normalizing input data.

[0091] Step 306: Obtain target flow control data based on the cold liquid flow candidate data and the candidate temperature prediction data corresponding to the cold liquid flow candidate data.

[0092] In a possible implementation, if the current iteration process satisfies the iteration end condition, and the candidate temperature prediction data is better than the optimal candidate solution before the current iteration process, the cold liquid flow candidate data is the target flow control data; if the current iteration process satisfies the iteration end condition, and the prediction data is better than the optimal candidate solution before the current iteration process, the target flow control data is determined based on the optimal candidate solution. Exemplarily, the iteration end condition includes reaching the maximum number of iterations or other preset end conditions.

[0093] In one possible implementation, if the current iteration process does not meet the iteration end condition, the candidate temperature prediction data is compared with the optimal candidate solution before the current iteration process, and the subsequent iteration process is continued, that is, step 302 to step 304 are executed for the cold liquid flow candidate data corresponding to the next iteration process until the iteration end condition is met to obtain the target flow control data.

[0094] In this embodiment, by making the flow data corresponding to at least one time window close to the target moment in time series all cold liquid flow candidate data, the motor temperature prediction model divides the input candidate flow time series data according to the preset time window to obtain multiple candidate flow subsequences in which at least one candidate flow subsequence is all cold liquid candidate flow data. In this way, when using part of the data in the candidate flow subsequence for graph convolution feature extraction processing, it is guaranteed that the cold liquid candidate flow data participates in the graph convolution feature extraction processing.

[0095] In an exemplary embodiment, based on Figure 3 In the embodiment shown, the motor temperature prediction model in this embodiment includes an input layer, a graph convolution layer and a prediction layer. Figure 4 This embodiment involves a process of inputting target operation timing data and candidate flow timing data into a motor temperature prediction model to obtain candidate temperature prediction data corresponding to the cold liquid flow candidate data. Figure 4 As shown, the process includes:

[0096] Step 402 : dividing the target operation time series data and the candidate flow time series data according to time windows through the input layer of the motor temperature prediction model to obtain a plurality of target operation subsequences and a plurality of candidate flow subsequences.

[0097] In one possible implementation, the input layer includes a normalization sublayer and a partitioning sublayer; the corresponding step 402 further includes: normalizing the target operation time series data and the candidate flow time series data through the normalization sublayer to obtain normalized target operation time series data and normalized candidate flow time series data; partitioning the normalized target operation time series data and the normalized candidate flow time series data according to time windows through the partitioning sublayer to obtain multiple target operation subsequences and multiple candidate flow subsequences.

[0098] In a possible implementation, the target runtime time series data and the candidate traffic time series data input to the input layer are normalized data.

[0099] Step 404, performing graph convolution feature extraction processing on part of the data in each target operation subsequence and each candidate flow subsequence through the graph convolution layer of the motor temperature prediction model to obtain the target operation convolution features corresponding to the target operation time series data and the candidate flow convolution features corresponding to the candidate flow time series data.

[0100] In a possible implementation, the graph convolution layer includes a node sublayer, a first graph convolution sublayer, and a second graph convolution sublayer, and step 402 further includes:

[0101] Step A1, through the node sublayer in the graph convolution layer, determine at least one operation collection data from each target operation subsequence as each target operation representative node, and determine at least one flow collection data or flow candidate data from each candidate flow subsequence as each candidate flow representative node.

[0102] Exemplarily, taking a target operation subsequence as an example, one operation acquisition data is randomly determined from the target operation subsequence as the target operation representative node of the target operation subsequence. Another exemplary example, taking a target operation subsequence as an example, two operation acquisition data are randomly determined from the target operation subsequence as the target operation representative nodes of the target operation subsequence.

[0103] Step A2, performing graph convolution feature extraction processing on each target operation representative node and each candidate flow representative node through the first graph convolution sublayer in the graph convolution layer, to obtain each intermediate operation convolution feature corresponding to each target operation representative node and each intermediate flow convolution feature corresponding to each candidate flow representative node.

[0104] In a possible implementation, the process of performing graph convolution feature extraction processing on each target operation representative node through the first graph convolution sublayer to obtain each intermediate operation convolution feature corresponding to each target operation representative node includes:

[0105] Step A21, based on the distance priority principle, determine a first number of cooling liquid flow data from the candidate flow subsequence belonging to the same time window as the target operation representative node, wherein the first number of cooling liquid flow data includes cooling liquid flow collection data or cooling liquid flow candidate data.

[0106] Exemplarily, the Euclidean distance between each cooling liquid flow data in the candidate flow subsequence and the target operation representative node is calculated, and a first number of cooling liquid flow data is obtained by a TOP-K nearest neighbor method, where K represents the first number.

[0107] In another exemplary embodiment, the cosine similarity data between each cooling liquid flow data in the candidate flow subsequence and the target operation representative node is calculated, and a first number of cooling liquid flow data are obtained by a TOP-K nearest neighbor method.

[0108] Step A22: obtaining a first adjacency matrix corresponding to the target operation representative node based on the distance data between the target operation representative node and the first number of cold liquid flow data.

[0109] Among them, the value of each element in the first adjacency matrix is ​​the distance data between the target operation representative node and the corresponding cold liquid flow data. Compared with the case where the value of each element in the adjacency matrix in the related technology is 0 or 1, the first adjacency matrix in this embodiment is used for subsequent convolution aggregation processing, which can better represent the spatial relationship between (operation data and cold liquid flow data) under different perspectives, thereby improving the accuracy of subsequent temperature prediction.

[0110] Exemplarily, each operation representative node is an operation collection data, and five cold liquid flow data closest to the operation collection data are determined through step A21. Correspondingly, the first adjacency matrix is ​​a matrix with a dimension of 11×11.

[0111] Step A23, performing convolution aggregation processing based on the first adjacency matrix corresponding to the target operation representative node, the target operation representative node and the first number of traffic data to obtain intermediate operation convolution features.

[0112] Exemplarily, the process of convolution aggregation processing can be expressed by Formula 12:

[0113]

[0114] Among them, H 运行i represents the intermediate running convolutional feature corresponding to the i-th target running representative node, A 运行i is the adjacency matrix corresponding to the i-th target running representative node, X 运行i It is a node feature corresponding to the i-th target operation representative node, and the node feature includes a specific value corresponding to the i-th target operation representative node and a specific value corresponding to the first number of cold liquid flow data. Represents the symmetric normalization operation on the adjacency matrix. 运行i represents the weight matrix corresponding to the i-th running representative node, and σ is the Relu activation function.

[0115] Correspondingly, the same method is adopted, that is, a method similar to steps A21 to A23, to obtain the convolution features of each intermediate flow.

[0116] Step A3, performing graph convolution feature extraction processing on each intermediate running convolution feature and each intermediate flow convolution feature through the second graph convolution sublayer in the graph convolution layer to obtain the target running convolution feature and the candidate flow convolution feature.

[0117] In a possible implementation, a process of performing graph convolution feature extraction processing on each intermediate running convolution feature and each intermediate flow convolution feature through a second graph convolution sublayer in a graph convolution layer to obtain a target running convolution feature and a candidate flow convolution feature includes:

[0118] Step A31, concatenate each intermediate operation convolution feature and each intermediate flow convolution feature according to the time window to obtain each concatenated feature corresponding to each time window.

[0119] Among them, for each time window, the intermediate running convolution feature and the intermediate flow convolution feature corresponding to the time window are spliced ​​to obtain the spliced ​​feature corresponding to the time window. In the second graph convolution sublayer, each time window is regarded as a node, and the spliced ​​feature corresponding to the time window is used as the node feature of the node.

[0120] Step A32, performing similarity calculation processing on each splicing feature to obtain a second adjacency matrix.

[0121] The value of each element in the second adjacency matrix is ​​the similarity data between the corresponding two splicing features. Exemplarily, there are n time windows in total, and the second adjacency matrix is ​​a matrix with a dimension of n×n.

[0122] Exemplarily, similarity data between each splicing feature is calculated by cosine similarity as the value of each element in the second adjacency matrix. Also exemplary, distance data between each splicing feature is calculated by cosine similarity as the value of each element in the second critical matrix.

[0123] Step A33, performing convolution aggregation processing based on each intermediate running convolution feature, each intermediate flow convolution feature and the second adjacency matrix to obtain the target running convolution feature and the candidate flow convolution feature.

[0124] Among them, convolution aggregation processing is performed based on the second adjacency matrix and each intermediate running convolution feature to obtain the target running convolution feature, and convolution aggregation processing is performed based on the second adjacency matrix and each intermediate flow convolution feature to obtain the candidate flow convolution feature.

[0125] In this implementation, the intermediate running convolution features and the intermediate flow convolution features in each time window are spliced ​​according to the time window as a node, and then the second adjacency matrix between the spliced ​​features is calculated. In this way, the second adjacency matrix can be used as a high-order matrix to better capture the connection between different time windows and the commonalities between multiple features. Then, by performing convolution aggregation processing on the second adjacency matrix and each intermediate running convolution feature and each intermediate flow convolution feature, the obtained target running convolution features and candidate flow convolution features can well integrate the information of different time windows and capture the commonalities and connections between features from different perspectives, providing a more accurate and comprehensive basis for subsequent temperature prediction processing.

[0126] In another possible implementation, the graph convolution layer includes a node sublayer and a first graph convolution layer; compared with the previous implementation, the intermediate running convolution features in the previous implementation are concatenated to obtain the target running convolution features, and the intermediate flow convolution features are concatenated to obtain the candidate flow convolution features.

[0127] In another possible implementation, the graph convolution layer includes a node sublayer, a first graph convolution layer, a second graph convolution layer, and a third graph convolution layer; in this implementation, the output data of the second graph convolution layer in step A33 is input into the third graph convolution layer for convolution aggregation processing again to obtain target operation convolution features and candidate flow convolution features.

[0128] Step 406, predicting the target operation convolution feature and the candidate flow convolution feature through the prediction layer of the motor temperature prediction model to obtain candidate temperature prediction data.

[0129] In one possible implementation, the process of performing prediction processing on the target operation convolution feature and the candidate flow convolution feature through the prediction layer of the motor temperature prediction model to obtain the candidate temperature prediction data includes:

[0130] Step B1, the target running convolution feature and the candidate flow convolution feature are concatenated through the concatenation sublayer in the prediction layer to obtain the concatenated candidate feature.

[0131] Among them, the target operation convolutional features and the candidate flow convolutional features are spliced ​​to obtain the spliced ​​candidate features as a comprehensive feature representation and input into the prediction sublayer.

[0132] Step B2, predicting the spliced ​​candidate features through the prediction sublayer in the prediction layer to obtain candidate temperature prediction data, wherein the candidate temperature prediction data includes the temperature corresponding to at least one position in the vehicle motor at at least one prediction time.

[0133] Exemplarily, the prediction sublayer is a fully connected layer.

[0134] In an exemplary embodiment, the vehicle motor temperature control method provided further includes a training process for the motor temperature prediction model, including steps C1 to C3, wherein:

[0135] Step C1, collecting the timing data of the vehicle motor under different working conditions to obtain a training data set and a test data set, wherein the timing data includes operation collection timing data, coolant flow timing data and motor temperature timing data.

[0136] Step C2, using the training data set to iteratively learn the initial prediction model until the iterative learning end condition is reached to obtain the target prediction model.

[0137] Among them, the loss function of iterative learning is the root mean square error, which is used to measure the deviation between the predicted value and the true value; the calculation formula of the loss function of iterative learning is: Among them, y i represents the predicted value; represents the true value; N represents the number of predictions, an exemplary number of predictions is the temperature values ​​of 10 positions on the surface of the motor, and the corresponding N is 10.

[0138] Step C3, input the test data set into the target prediction model, obtain the output data of the target prediction model, and judge whether the output data of the target prediction model reaches the preset evaluation index. If it does not reach the preset evaluation index, the target prediction model is used as the initial prediction model to continue iterative learning until the preset evaluation index is reached, and the target prediction model is used as the motor temperature prediction model.

[0139] In an exemplary embodiment, Figure 5 As shown, a vehicle motor temperature control method is provided, and the method is applied to Figure 1 The vehicle controller 102 in the example is used for explanation, including the following steps 502 to 518. The motor temperature prediction model involved in the motor temperature control method includes an input layer, a graph convolution layer and a prediction layer, wherein the graph convolution layer includes a node sublayer, a first graph convolution layer and a second graph convolution layer, and the prediction layer includes a splicing sublayer and a prediction sublayer, wherein:

[0140] Step 502, obtaining target speed timing data, target torque timing data and multiple cooling fluid flow acquisition data corresponding to the vehicle motor at the target time, wherein the target speed timing data includes multiple speed acquisition data, and the target torque timing data includes multiple torque acquisition data.

[0141] Step 504, obtaining candidate flow time series data based on multiple cold liquid flow collection data and cold liquid flow candidate data in the current iteration process, wherein the flow data corresponding to at least one time window in the candidate flow time series data that is close to the target moment in time series are all cold liquid flow candidate data.

[0142] Step 506, dividing the target speed time series data, the target torque time series data and the candidate flow time series data according to a preset time window through the input layer of the motor temperature prediction model to obtain multiple target speed subsequences, multiple target torque subsequences and multiple candidate flow subsequences.

[0143] Step 508, through the node sublayer in the graph convolution layer of the motor temperature prediction model, at least one speed acquisition data is determined from each target speed subsequence as each target speed representative node, at least one torque acquisition data is determined from each target torque subsequence as each target torque representative node, and at least one flow acquisition data or flow candidate data is determined from each candidate flow subsequence as each candidate flow representative node.

[0144] Step 510, performing graph convolution feature extraction processing on each target speed representative node, each target torque representative node and each candidate flow representative node through the first graph convolution sublayer in the graph convolution layer, to obtain each intermediate speed convolution feature corresponding to each target speed representative node, each intermediate torque convolution feature corresponding to each target torque representative node and each intermediate flow convolution feature corresponding to each candidate flow representative node.

[0145] Optionally, step 510 includes: determining a first number of cold liquid flow data from a candidate flow subsequence belonging to the same time window as the target operation representative node based on a distance priority principle, wherein the first number of cold liquid flow data includes cold liquid flow collection data or cold liquid flow candidate data, obtaining a first adjacency matrix corresponding to the target operation representative node based on the distance data between the target operation representative node and the first number of cold liquid flow data, and performing convolution aggregation processing based on the first adjacency matrix corresponding to the target operation representative node, the target operation representative node and the first number of cold liquid flow data to obtain intermediate operation convolution features.

[0146] Please combine Figure 6 and Figure 7 , is a schematic diagram of the implementation process of step 506 to step 510 in an optional implementation manner. Figure 6 As shown, taking the target speed subsequence corresponding to the i-th time window as an example, a speed acquisition data is determined from the target speed subsequence as the target speed representative node, and then the first quantity ( Figure 6 The first number is 5 as an example) data is connected to the target speed representative node to obtain the speed graph structure G in the i-th time window 转i =(A 转i ,X 转i ), where i = 1, 2, ..., n, n represents the number of time windows, A 转i represents the target speed corresponding to the i-th time window, represents the first adjacency matrix corresponding to the node, X 转i Represents the node characteristics corresponding to all nodes in the speed graph. Please refer to Figure 6, using the same method as the target speed sub-series, the torque graph structure G corresponding to the i-th time window is obtained 扭i =(A 扭i ,X 扭i ) and the flow graph structure G corresponding to the i-th time window 冷i =(A 冷i ,X 冷i ).

[0147] like Figure 7 As shown, the three graph structures of each time window are subjected to graph convolution feature fusion extraction, and the spatial features guided by different graph structures in each time window are extracted through the graph convolution layer, so as to obtain the relationship information between the three time series data under different perspectives, and make full use of various graph structure information to achieve feature fusion and extraction. Among them, the graph convolution feature fusion extraction process in the implementation process of step 510 can be expressed by the following formula:

[0148]

[0149] Among them, H 转i , H 扭i and H 冷i They respectively represent the intermediate speed convolution feature, intermediate torque convolution feature and intermediate flow convolution feature corresponding to the i-th time window.

[0150] Step 512, performing graph convolution feature extraction processing on each intermediate speed convolution feature, each intermediate torque convolution feature, and each intermediate flow convolution feature through the second graph convolution sublayer in the graph convolution layer to obtain a target speed convolution feature, a target torque convolution feature, and a candidate flow convolution feature.

[0151] Optionally, step 512 further includes: splicing each intermediate speed convolution feature, each intermediate torque convolution feature and each intermediate flow convolution feature according to the time window to obtain each spliced ​​feature corresponding to each time window; performing similarity calculation processing on each spliced ​​feature to obtain a second adjacency matrix; performing convolution aggregation processing based on each intermediate speed convolution feature, each intermediate torque convolution feature, each intermediate flow convolution feature and the second adjacency matrix to obtain a target speed convolution feature, a target torque convolution feature and a candidate flow convolution feature.

[0152] Please refer to Figure 8 , is a schematic diagram of the implementation process of step 512 in an optional implementation manner. Figure 8 As shown in the figure, n time windows are regarded as n nodes, and the intermediate speed convolution features, intermediate torque convolution features and intermediate flow convolution features corresponding to each time window are concatenated as the node features of each time window. Then, the similarity between each node is calculated by cosine similarity to obtain an n×n similarity matrix as the second adjacency matrix A.H ; In this way, the second adjacency matrix obtained by calculating the similarity of the three features can better capture the connection between different time periods and the commonality between the three features; then the intermediate speed convolution features corresponding to the n time windows and the second adjacency matrix form a high-order speed graph structure G 转 =(A H , H 转i ), the intermediate torque convolution features corresponding to the n time windows are combined with the second adjacency matrix to form a high-order torque graph structure G 扭 =(A H , H 扭i ), the intermediate flow convolution features corresponding to the n time windows are combined with the second adjacency matrix to form a high-order flow graph structure G 冷 =(A H , H 冷i ); then, the graph convolution is used to aggregate and obtain three aggregated high-order features, namely, the target speed convolution feature, the target torque convolution feature and the candidate flow convolution feature; wherein, the process of obtaining the three aggregated high-order features can be expressed by the following formula:

[0153]

[0154] Among them, H 1 represents the target speed convolution feature, H 2 represents the target torque convolution feature, H 3 W represents the candidate flow convolution feature. 高1 , W 高2 and W 高3 They are the weight matrices corresponding to the high-order speed graph structure, the high-order torque graph structure and the high-order flow graph structure respectively.

[0155] Step 514, the target operation convolution feature and the candidate flow convolution feature are concatenated by the concatenation sublayer in the prediction layer to obtain a concatenated candidate feature.

[0156] Step 516, predicting the spliced ​​candidate features through the prediction sublayer in the prediction layer to obtain candidate temperature prediction data, wherein the candidate temperature prediction data includes the temperature corresponding to at least one position in the vehicle motor at at least one prediction time.

[0157] Step 518: based on the coolant flow candidate data and the candidate temperature prediction data corresponding to the coolant flow candidate data, the target flow control data is obtained, wherein the target flow control data is the coolant flow candidate data that minimizes the candidate temperature prediction data output by the motor temperature prediction model in each iteration process.

[0158] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0159] It is understood that the term "based on" used in this application is used to describe one or more factors that affect the determination, and does not exclude other factors that may affect the determination. For example, the phrase "determine A based on B" means that the determination of A can be based entirely or at least partially on factor B, that is, B is a factor that affects the determination of A, but does not exclude that the determination of A is also based on C.

[0160] In an exemplary embodiment, please refer to Fig. 9 , a vehicle motor temperature prediction method is provided. This method is applied to Figure 1 The vehicle-mounted controller 102 in FIG. 1 is taken as an example to illustrate, and the steps include the following steps 902 to 906. Among them:

[0161] Step 902, obtaining the operation timing data and coolant flow timing data corresponding to the vehicle motor at the target time.

[0162] Step 906: Divide the operation time series data and the cooling liquid flow time series data into a plurality of operation subsequences and a plurality of cooling liquid flow subsequences according to a preset time window.

[0163] Step 908: Perform graph convolution feature extraction processing on part of the data in each operation subsequence and each cold liquid flow subsequence to obtain the target operation convolution feature and the target flow convolution feature corresponding to the target time.

[0164] Step 910, performing prediction processing based on the target operation convolution feature and the target flow convolution feature to obtain target predicted temperature data corresponding to the vehicle motor at the target time.

[0165] The vehicle motor temperature prediction method provided in the above embodiment uses a data-driven approach, by dividing the input time series data into multiple subsequences according to a preset time window, and using part of the data in each subsequence for graph convolution feature extraction and temperature prediction. This can avoid the need to perform feature extraction on each data in the input time series data when the motor temperature field digital twin model in the related art predicts the motor temperature, which results in large amount of calculation and low prediction efficiency. The vehicle motor temperature prediction method provided in the above embodiment can improve the efficiency of temperature prediction.

[0166] In an exemplary embodiment, based on Fig. 9 In the embodiment shown, this embodiment involves a process of performing graph convolution feature extraction processing on part of the data in each operation subsequence and each cold liquid flow subsequence to obtain the target operation convolution feature and the target flow convolution feature corresponding to the target time. The process includes:

[0167] Step D1, determining at least one operation collection data from each operation subsequence as each operation representative node, and determining at least one flow collection data from each flow subsequence as each flow node.

[0168] Step D2, performing graph convolution feature extraction processing on each operation representative node and each flow representative node, to obtain each intermediate operation convolution feature corresponding to each operation representative node and each intermediate flow convolution feature corresponding to each flow representative node.

[0169] Step D3, performing graph convolution feature extraction processing on each intermediate operation convolution feature and each intermediate flow convolution feature to obtain a target operation convolution feature and a target flow convolution feature.

[0170] In an exemplary embodiment, the process of performing graph convolution feature extraction processing on each running representative node in step D2 to obtain each intermediate running convolution feature corresponding to each running representative node includes:

[0171] Step D21 , determining a first number of flow collection data from a flow subsequence belonging to the same time window as the running representative node based on a distance priority principle.

[0172] Step D22: obtaining a first adjacency matrix corresponding to the running representative node based on the distance data between the running representative node and the first number of traffic collection data.

[0173] Step D23, performing convolution aggregation processing based on the first adjacency matrix corresponding to the running representative node, the running representative node and the first number of traffic collection data to obtain intermediate running convolution features.

[0174] In an exemplary embodiment, the process of performing graph convolution feature extraction processing on each intermediate running convolution feature and each intermediate flow convolution feature in step D3 to obtain a target running convolution feature and a target flow convolution feature includes:

[0175] Step D31, concatenates each intermediate operation convolution feature and each intermediate flow convolution feature according to the time window to obtain each concatenated feature corresponding to each time window.

[0176] Step D32, performing similarity calculation processing on each splicing feature to obtain a second adjacency matrix.

[0177] Step D33, performing convolution aggregation processing based on each intermediate running convolution feature, each intermediate flow convolution feature and the second adjacency matrix to obtain a target running convolution feature and a target flow convolution feature.

[0178] In an exemplary embodiment, based on Fig. 9 The embodiment shown in the present embodiment relates to a process of performing prediction processing according to the target operation convolution feature and the target flow convolution feature in step 908 to obtain the target predicted temperature data corresponding to the vehicle motor at the target time; the process includes:

[0179] Step E1, concatenate the target running convolution feature and the target flow convolution feature to obtain a concatenated candidate feature.

[0180] Step E2, performing prediction processing based on the spliced ​​candidate features to obtain target predicted temperature data, wherein the target predicted temperature data includes the temperature corresponding to at least one position in the vehicle motor at at least one prediction time.

[0181] In an exemplary embodiment, based on Fig. 9 In an embodiment of the vehicle motor temperature prediction method provided, the operating time series data includes speed time series data and torque time series data.

[0182] Based on the same inventive concept, the embodiment of the present application also provides a vehicle motor temperature control device for implementing the vehicle motor temperature control method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more vehicle motor temperature control device embodiments provided below can refer to the limitations of the vehicle motor temperature control method above, and will not be repeated here.

[0183] In an exemplary embodiment, Fig.10 As shown, a vehicle motor temperature control device is provided, including: a data acquisition module 1002, a flow optimization module 1004 and a flow control module 1006, wherein:

[0184] The data acquisition module 1002 is used to acquire target operation timing data and a plurality of cooling liquid flow acquisition data corresponding to the vehicle motor at a target time, wherein the target operation timing data includes a plurality of operation acquisition data;

[0185] The flow optimization module 1004 is used to obtain target flow control data by iteratively optimizing the output data of the motor temperature prediction model based on the target operating timing data and multiple cold liquid flow acquisition data, wherein the motor temperature prediction model is used to divide the input operating timing data and cold liquid flow timing data into multiple operating subsequences and multiple cold liquid flow subsequences according to a preset time window, and use part of the data in each operating subsequence and each cold liquid flow subsequence to perform graph convolution feature extraction and temperature prediction processing, and output temperature prediction data corresponding to the input operating timing data and cold liquid flow timing data; wherein, in each iterative optimization process, the cold liquid flow data included in the part of the cold liquid flow subsequence in the cold liquid flow timing data input by the motor temperature prediction model is the cold liquid flow candidate data corresponding to the current iterative optimization process.

[0186] A flow control module 1006, for controlling a flow of a cooling liquid for cooling a motor of a vehicle based on target flow control data;

[0187] In an exemplary embodiment, the flow optimization module 1004 is used to obtain candidate flow timing data based on multiple cold liquid flow collection data and cold liquid flow candidate data in the current iteration process, and the flow data corresponding to at least one time window in the candidate flow timing data that is close to the target time in timing are all cold liquid flow candidate data; the target operation timing data and the candidate flow timing data are input into the motor temperature prediction model to obtain candidate temperature prediction data corresponding to the cold liquid flow candidate data; based on the cold liquid flow candidate data and the candidate temperature prediction data corresponding to the cold liquid flow candidate data, the target flow control data is obtained.

[0188] In an exemplary embodiment, the traffic optimization module 1004 is used to divide the target operation timing data and the candidate traffic timing data according to time windows through the input layer of the motor temperature prediction model to obtain multiple target operation subsequences and multiple candidate traffic subsequences; perform graph convolution feature extraction processing on part of the data in each target operation subsequence and each candidate traffic subsequence through the graph convolution layer of the motor temperature prediction model to obtain the target operation convolution features corresponding to the target operation timing data and the candidate traffic convolution features corresponding to the candidate traffic timing data; perform prediction processing on the target operation convolution features and the candidate traffic convolution features through the prediction layer of the motor temperature prediction model to obtain candidate temperature prediction data.

[0189] In an exemplary embodiment, the traffic optimization module 1004 is used to determine at least one operation collection data from each target operation subsequence as each target operation representative node, and to determine at least one flow collection data or flow candidate data from each candidate flow subsequence as each candidate flow representative node through the node sublayer in the graph convolution layer; perform graph convolution feature extraction processing on each target operation representative node and each candidate flow representative node through the first graph convolution sublayer in the graph convolution layer to obtain each intermediate operation convolution feature corresponding to each target operation representative node and each intermediate flow convolution feature corresponding to each candidate flow representative node; perform graph convolution feature extraction processing on each intermediate operation convolution feature and each intermediate flow convolution feature through the second graph convolution sublayer in the graph convolution layer to obtain target operation convolution features and candidate flow convolution features.

[0190] In an exemplary embodiment, the flow optimization module 1004 is used to determine a first number of cold liquid flow data from a candidate flow subsequence belonging to the same time window as the target operation representative node based on a distance priority principle, wherein the first number of cold liquid flow data includes cold liquid flow collection data or cold liquid flow candidate data; based on the distance data between the target operation representative node and the first number of cold liquid flow data, obtain a first adjacency matrix corresponding to the target operation representative node; based on the first adjacency matrix corresponding to the target operation representative node, the target operation representative node and the first number of cold liquid flow data, perform convolution aggregation processing to obtain intermediate operation convolution features.

[0191] In an exemplary embodiment, the traffic optimization module 1004 is used to perform splicing processing on each intermediate operation convolution feature and each intermediate flow convolution feature according to a time window to obtain each splicing feature corresponding to each time window; perform similarity calculation processing on each splicing feature to obtain a second adjacency matrix; perform convolution aggregation processing based on each intermediate operation convolution feature, each intermediate flow convolution feature and the second adjacency matrix to obtain a target operation convolution feature and a candidate flow convolution feature.

[0192] In an exemplary embodiment, the traffic optimization module 1004 is used to perform splicing processing on the target operation convolution feature and the candidate traffic convolution feature through the splicing sublayer in the prediction layer to obtain the spliced ​​candidate feature; perform prediction processing on the spliced ​​candidate feature through the prediction sublayer in the prediction layer to obtain candidate temperature prediction data, wherein the candidate temperature prediction data includes the temperature corresponding to at least one position in the vehicle motor at at least one prediction time.

[0193] In an exemplary embodiment, the operation acquisition data includes rotation speed acquisition data and torque acquisition data, and the target operation timing data includes target rotation speed timing data and target torque timing data.

[0194] Based on the same inventive concept, the embodiment of the present application also provides a vehicle motor temperature prediction device for implementing the vehicle motor temperature prediction method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more vehicle motor temperature prediction device embodiments provided below can refer to the limitations of the vehicle motor temperature prediction method above, and will not be repeated here.

[0195] In an exemplary embodiment, Fig.11 As shown, a vehicle motor temperature prediction device is provided, comprising: a data input module 1102, a data division module 1104, a feature extraction module 1106 and a temperature prediction module 1108, wherein:

[0196] The data input module 1102 is used to obtain the operation timing data and the cooling liquid flow timing data corresponding to the vehicle motor at the target time;

[0197] A data partitioning module 1104 is used to partition the operation time series data and the cooling liquid flow time series data into a plurality of operation subsequences and a plurality of cooling liquid flow subsequences according to a preset time window;

[0198] The feature extraction module 1106 is used to perform graph convolution feature extraction processing on part of the data in each operation subsequence and each cold liquid flow subsequence to obtain the target operation convolution feature and the target flow convolution feature corresponding to the target time;

[0199] The temperature prediction module 1108 is used to perform prediction processing based on the target operation convolution feature and the target flow convolution feature to obtain the target predicted temperature data corresponding to the vehicle motor at the target time.

[0200] Each module in the above-mentioned vehicle motor temperature control device and vehicle motor temperature prediction device can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0201] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal vehicle-mounted controller, and its internal structure diagram may be as shown in FIG. Fig.12As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface and an input device. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a vehicle motor temperature control method or a vehicle motor temperature prediction method is implemented. The input device of the computer device can be a key, a trackball or a touchpad set on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0202] Those skilled in the art will understand that Fig.12 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0203] In one embodiment, a vehicle-mounted controller is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0204] In one embodiment, 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 in the above-mentioned method embodiments are implemented.

[0205] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0206] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0207] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0208] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0209] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A vehicle motor temperature control method, characterized in that: The method comprises: Acquire target operation timing data and a plurality of cooling liquid flow collection data corresponding to the vehicle motor at a target time, wherein the target operation timing data includes a plurality of operation collection data; Based on the target operating timing data and the multiple cold liquid flow acquisition data, iterative optimization processing is performed by minimizing the output data of the motor temperature prediction model to obtain the target flow control data, the motor temperature prediction model is used to divide the input operating timing data and cold liquid flow timing data into multiple operating subsequences and multiple cold liquid flow subsequences according to a preset time window, and use part of the data in each of the operating subsequences and each of the cold liquid flow subsequences to perform graph convolution feature extraction processing and temperature prediction processing, and output temperature prediction data corresponding to the input operating timing data and cold liquid flow timing data, wherein, in each iterative optimization process, the cold liquid flow data included in the part of the cold liquid flow subsequences in the cold liquid flow timing data input by the motor temperature prediction model is the cold liquid flow candidate data corresponding to the current iterative optimization process; A flow rate of coolant for cooling the vehicle motor is controlled based on the target flow rate control data.

2. The method according to claim 1, characterized in that The target flow control data is obtained by performing iterative optimization processing based on the target operation timing data and the plurality of cooling liquid flow acquisition data by minimizing the output data of the motor temperature prediction model, including: According to the plurality of cold liquid flow acquisition data and the cold liquid flow candidate data in the current iteration process, candidate flow time series data are obtained, wherein the flow data corresponding to at least one of the time windows close to the target time in time series among the candidate flow time series data are all the cold liquid flow candidate data; Inputting the target operation timing data and the candidate flow timing data into the motor temperature prediction model to obtain candidate temperature prediction data corresponding to the cold liquid flow candidate data; The target flow control data is obtained based on the cold liquid flow candidate data and the candidate temperature prediction data corresponding to the cold liquid flow candidate data.

3. The method according to claim 2, characterized in that The step of inputting the target operation timing data and the candidate flow timing data into the motor temperature prediction model to obtain the candidate temperature prediction data corresponding to the cold liquid flow candidate data includes: The target operation time series data and the candidate flow time series data are divided according to the time window through the input layer of the motor temperature prediction model to obtain multiple target operation subsequences and multiple candidate flow subsequences; Performing graph convolution feature extraction processing on part of the data in each of the target operation subsequences and each of the candidate flow subsequences through the graph convolution layer of the motor temperature prediction model to obtain target operation convolution features corresponding to the target operation time series data and candidate flow convolution features corresponding to the candidate flow time series data; The target operation convolution feature and the candidate flow convolution feature are predicted and processed by the prediction layer of the motor temperature prediction model to obtain the candidate temperature prediction data.

4. The method according to claim 3, characterized in that The graph convolution layer of the motor temperature prediction model performs graph convolution feature extraction processing on part of the data in each of the target operation subsequences and each of the candidate flow subsequences to obtain the target operation convolution features corresponding to the target operation time series data and the candidate flow convolution features corresponding to the candidate flow time series data, including: Determine, through the node sublayer in the graph convolution layer, at least one operation collection data from each of the target operation subsequences as each target operation representative node, and determine, from each of the candidate flow subsequences, at least one flow collection data or flow candidate data as each candidate flow representative node; Performing graph convolution feature extraction processing on each of the target operation representative nodes and each of the candidate flow representative nodes through the first graph convolution sublayer in the graph convolution layer, to obtain each intermediate operation convolution feature corresponding to each of the target operation representative nodes and each intermediate flow convolution feature corresponding to each of the candidate flow representative nodes; The second graph convolution sublayer in the graph convolution layer performs graph convolution feature extraction processing on each of the intermediate running convolution features and each of the intermediate flow convolution features to obtain the target running convolution features and the candidate flow convolution features.

5. The method according to claim 4, characterized in that The process of performing graph convolution feature extraction processing on each target operation representative node through the first graph convolution sublayer in the graph convolution layer to obtain each intermediate operation convolution feature corresponding to each target operation representative node includes: Determine a first number of cold liquid flow data from the candidate flow subsequence belonging to the same time window as the target operation representative node based on a distance priority principle, wherein the first number of cold liquid flow data includes the cold liquid flow collection data or the cold liquid flow candidate data; Based on the distance data between the target operation representative node and the first number of cold liquid flow data, obtaining a first adjacency matrix corresponding to the target operation representative node; Convolution aggregation processing is performed based on a first adjacency matrix corresponding to the target operation representative node, the target operation representative node, and the first number of cold liquid flow data to obtain the intermediate operation convolution feature.

6. The method according to claim 4, characterized in that The step of performing graph convolution feature extraction processing on each of the intermediate running convolution features and each of the intermediate flow convolution features through the second graph convolution sublayer in the graph convolution layer to obtain the target running convolution feature and the candidate flow convolution feature includes: Performing splicing processing on each of the intermediate operation convolution features and each of the intermediate flow convolution features according to the time window to obtain each splicing feature corresponding to each of the time windows; Performing similarity calculation processing on each of the splicing features to obtain a second adjacency matrix; Convolution aggregation processing is performed based on each of the intermediate running convolution features, each of the intermediate flow convolution features and the second adjacency matrix to obtain the target running convolution feature and the candidate flow convolution feature.

7. The method according to claim 3, characterized in that The step of performing prediction processing on the target operation convolution feature and the candidate flow convolution feature through the prediction layer of the motor temperature prediction model to obtain the candidate temperature prediction data includes: The target operation convolution feature and the candidate flow convolution feature are concatenated by a concatenation sublayer in the prediction layer to obtain a concatenated candidate feature; The candidate temperature prediction data is obtained by performing prediction processing on the splicing candidate features through the prediction sublayer in the prediction layer, wherein the candidate temperature prediction data includes the temperature corresponding to at least one position in the vehicle motor at at least one prediction time.

8. The method according to claim 1, characterized in that The operation acquisition data includes rotation speed acquisition data and torque acquisition data, and the target operation timing data includes target rotation speed timing data and target torque timing data.

9. A vehicle motor temperature prediction method, characterized in that: The method comprises: Obtaining the operation timing data and coolant flow timing data corresponding to the vehicle motor at the target time; Dividing the operation time series data and the cooling liquid flow time series data into a plurality of operation subsequences and a plurality of cooling liquid flow subsequences according to a preset time window; Performing graph convolution feature extraction processing on part of the data in each of the operation subsequences and each of the cold liquid flow subsequences to obtain target operation convolution features and target flow convolution features corresponding to the target time; Prediction processing is performed based on the target operation convolution feature and the target flow convolution feature to obtain target predicted temperature data corresponding to the vehicle motor at the target time.

10. A vehicle motor temperature control device, characterized in that: The device comprises: A data acquisition module, used to acquire target operation timing data and a plurality of cooling liquid flow acquisition data corresponding to the vehicle motor at a target time, wherein the target operation timing data includes a plurality of operation acquisition data; A flow optimization module, for obtaining target flow control data by performing iterative optimization processing based on the target operating timing data and the multiple cold liquid flow acquisition data by minimizing the output data of the motor temperature prediction model, wherein the motor temperature prediction model is used to divide the input operating timing data and cold liquid flow timing data into multiple operating subsequences and multiple cold liquid flow subsequences according to a preset time window, and use part of the data in each of the operating subsequences and each of the cold liquid flow subsequences to perform graph convolution feature extraction processing and temperature prediction processing, and output temperature prediction data corresponding to the input operating timing data and cold liquid flow timing data; wherein, in each iterative optimization process, the cold liquid flow data included in the part of the cold liquid flow subsequences in the cold liquid flow timing data input by the motor temperature prediction model is the cold liquid flow candidate data corresponding to the current iterative optimization process; A flow control module is used to control the flow of cooling liquid for cooling the vehicle motor based on the target flow control data.

Citation Information

Patent Citations

  • Mixed model multivariate time sequence anomaly detection method based on graph neural network

    CN113516226A

  • New energy automobile motor heat control method

    CN115313763A

  • Industrial product testing system applied to intelligent manufacturing

    CN115345256A

  • Cable temperature prediction method and device, computer equipment and storage medium

    CN115470976A

  • Permanent magnet synchronous motor temperature prediction method and system based on temperature time sequence input

    CN116317800A

Cited By

  • Motor temperature prediction method and device, storage medium and equipment

    CN121365226A

  • Motor cooling system real-time intelligent control method based on edge calculation

    CN121690008A

  • Motor cooling system real-time intelligent control method based on edge computing

    CN121690008B