Motor temperature estimation method and device based on GPR-RC model

Abstract

The invention provides a motor temperature estimation method and device based on a GPR-RC model, the GPR-RC model comprises two sub-models of a GPR model and an RC filter model, and the method comprises the steps: obtaining the working condition parameters of a motor at a plurality of moments in a continuous operation process; the working condition parameters at the multiple moments are input into the GPR model obtained through training in the training stage to obtain the steady-state temperature values of the motor at the multiple moments, the multiple working condition parameters are optimized based on the grey correlation degree in the training stage, and the steady-state temperature values of the motor at the multiple moments are obtained. Working condition parameter categories more related to the motor temperature are screened to serve as training data input into the GPR model, and then the GPR model is trained; selecting an adaptive RC filter model, and determining parameters of the RC filter model; and estimating the real-time temperature of the motor based on the steady-state temperature values of the motor at the plurality of moments output by the GPR model and the RC filter model.

Description

Technical Field

[0001] The present invention relates to the field of motor temperature estimation, and particularly to a method and device for motor temperature estimation based on a GPR-RC model. Background Art

[0002] Due to the advantages of energy conservation, environmental protection, green and low-carbon, the market penetration rate of new energy vehicles is continuously increasing, and the market share is also increasing. As the core component of the electric drive system of new energy vehicles, the drive motor has received more attention and research. During the operation of the motor, losses will occur, causing its temperature to rise. Excessive temperature will accelerate the aging of the stator winding and the demagnetization of the rotor permanent magnet, and in severe cases, it will lead to the scrapping of the motor. Therefore, the research on how to estimate the motor temperature is of great significance.

[0003] In the prior art, the methods for motor temperature estimation mainly include: the finite element method, the lumped parameter thermal network method, the signal injection method, and the flux observer, etc. However, due to the complex heat transfer relationship inside the automotive drive motor, it is difficult to accurately analyze its heat transfer path, and these motor temperature estimation methods face various difficulties and challenges in practical applications. At the same time, in recent years, with the extensive research and application of artificial intelligence technology in data mining, complex system modeling, parameter identification, etc., there have begun to appear solutions for using artificial intelligence technology to estimate the temperature of automotive drive motors to assist the research work on motor temperature.

[0004] At the same time, Gaussian Process Regression (GPR) is a non-parametric probability model based on the Bayesian method and is a powerful tool for simulating the nonlinear behavior of a system. However, when using the GPR model to establish a real-time motor temperature estimation model, there are problems such as many parameters and too large a model size. In addition, the RC filter is often used to model the heat transfer process inside the motor, where the heat resistance represents the heat exchange relationship between two nodes, and the heat capacity represents the thermal inertia of the node. Using the RC filter can better simulate the transient change process of the motor temperature, but this method requires accurate calculation of the motor losses.

[0005] There are many parameters that can affect the temperature during the operation of the motor, and the selection of model input parameters is crucial for modeling. However, the artificial feature selection has great blindness and often has unsatisfactory results. In addition, there are many hyperparameters to be determined in the GPR model, which directly affect the performance of the model. The parameter optimization methods of existing solutions in the prior art often fail to achieve ideal results.

[0006] In order to overcome the above-mentioned defects existing in the prior art, there is an urgent need in the art for a method and device for estimating the temperature of an electric motor based on a GPR-RC model, which establishes an electric motor thermal model based on a hybrid algorithm of Gaussian process regression and low-pass filtering. The GPR model is used to predict the steady-state temperature of the electric motor under a certain working condition, and the RC filter predicts the transient temperature change between two steady-state temperature points, so as to realize the real-time temperature estimation of the electric motor. At the same time, in the modeling process, it is possible to avoid the complex analysis of the heat transfer path of the electric motor and the identification process of the thermal network parameters. Aiming at the problem that it is difficult to determine the model input variables and hyperparameters, the PSO algorithm is used to select the model input variables and optimize the hyperparameters, and the optimized GPR-RC thermal model of the electric motor is used to realize the temperature estimation of the electric motor, which is especially suitable for occasions where the physical model is particularly complex, improves the development efficiency of the electric motor, and thus reduces the vehicle development cost. Summary of the Invention

[0007] A brief overview of one or more aspects is given below to provide a basic understanding of these aspects. This overview is not an exhaustive survey of all contemplated aspects, and is neither intended to identify key or decisive elements of any or all aspects nor to attempt to define the scope of any or all aspects. Its sole purpose is to present some concepts of one or more aspects in a simplified form as a prelude to the more detailed description that follows.

[0008] In order to overcome the above-mentioned defects existing in the prior art, the present invention provides a method for estimating the temperature of an electric motor based on a GPR-RC model. The GPR-RC model includes two sub-models, namely a GPR model and an RC filter model. The method for estimating the temperature of the electric motor includes: obtaining the working condition parameters of the electric motor at multiple moments during continuous operation; respectively inputting the working condition parameters at the multiple moments into the GPR model trained in the training stage to obtain the steady-state temperature values of the electric motor at the multiple moments. Among them, in the training stage, multiple working condition parameters are optimized based on the grey relational degree to screen out the categories of working condition parameters more relevant to the electric motor temperature as the training data input into the GPR model, and then the GPR model is trained; selecting a suitable RC filter model and determining the parameters of the RC filter model; and estimating the real-time temperature of the electric motor based on the steady-state temperature values of the electric motor at the multiple moments output by the GPR model and the RC filter model.

[0009] In one embodiment, preferably, the optimization of multiple working condition parameters based on the grey relational degree may include: calculating the grey relational degree values between multiple working condition parameters and the electric motor temperature respectively by using the following formula:

[0010]

[0011] where y(k) is the electric motor temperature, x i(k) are multiple operating condition parameters related to the motor temperature, ρ is the grey discrimination coefficient; and sorting the multiple operating condition parameters in descending order according to their corresponding grey correlation degree values to obtain the hyperparameter U to be optimized for determining the number of input variables of the GPR model:

[0012] U = [U 1 , U 2 ,..., U M

[0013] wherein, U 1 represents the operating condition parameter with the largest grey correlation degree value with the motor temperature, U 2 represents the operating condition parameters with the top two grey correlation degree values with the motor temperature, and so on, U M represents the operating condition parameters with the top M grey correlation degree values with the motor temperature.

[0014] In one embodiment, preferably, the training stage includes: collecting multiple sets of operating condition data of the motor during actual operation, dividing the multiple sets of operating condition data into multiple data sets, and performing optimized sorting based on the grey correlation degree, and the multiple data sets include a training set and a validation set; selecting an appropriate kernel function for the GPR model, inputting the operating condition data of the training set sorted by grey correlation degree into the GPR model; and determining the hyperparameters of the GPR model by using the PSO optimization algorithm, thereby completing the model training work.

[0015] In one embodiment, preferably, the multiple sets of operating condition data include the operating condition parameters of multiple operating conditions and the steady-state temperature of the motor under each operating condition, and the training stage further includes: preprocessing the multiple sets of operating condition data sorted by grey correlation degree and organizing them into a time series form of (x t , T t ), where x t is the operating condition parameter of the motor at time t, T t is the steady-state temperature of the motor under the corresponding operating condition at time t, and x t is expressed as:

[0016] x t = [t coolt , V coolt , U dc , …, I s , T q , n]

[0017] wherein, the operating condition parameters include coolant temperature t coolt , coolant flow rate V coolt , DC bus voltage U dc , line current amplitude I s , motor torque T q and motor speed n.​

[0018] In one embodiment, preferably, the steady-state temperature T of the motor under the corresponding operating conditions at time t t is expressed as:

[0019] T t = f(x t ) + ε t

[0020] where f(·) represents the GPR model, and ε t is Gaussian white noise with a mean of zero added, and ε t satisfies the following expression to prevent model overfitting:

[0021]

[0022] where σ y is the standard deviation of ε t .

[0023] In one embodiment, preferably, the GPR model also satisfies the following expression:

[0024] f = f(x t ) ~ N(f|0, K)

[0025] where K is the covariance matrix, and K ij = k(x i , x j ), and k(·,·) is the kernel function; an appropriate kernel function is selected for the GPR model, including: selecting the squared exponential function as the kernel function, and representing the kernel function with the following formula:

[0026]

[0027] where D is the data dimension, and σf, l are hyperparameters to be optimized in the kernel function.

[0028] In one embodiment, preferably, the hyperparameters of the GPR model are θ = [U, σ f , l, σ y , and the hyperparameters of the GPR model are determined using the PSO optimization algorithm, which may include: determining the dimension of each particle according to the number of variables of the hyperparameters, randomly generating an initial population of the hyperparameters of the GPR model, including the initial velocity and initial position of each particle; calculating the fitness value of each particle individual based on the fitness function, and the fitness value of each particle individual is the mean squared error f id of the motor temperature prediction; if the mean squared error f id of the motor temperature prediction corresponding to a particle is less than the minimum temperature mean squared error f(P id), then use the mean square error of the motor temperature f id to replace f(P in the previous round id ). Replace the particle in the previous round with this particle, and update the velocity and position of the particle; if the minimum mean square error of the motor temperature f(P id ) of a particle is less than the global minimum mean square error of the motor temperature f(P gd ) of all particles, then replace the original global minimum mean square error of the motor temperature with the minimum mean square error of this particle, and save the current position of this particle at the same time; and when the preset optimization end condition is met, use the current position of the particle as the optimal input and hyperparameters of the GPR model obtained by the optimization.

[0029] In one embodiment, preferably, the updating of the velocity and position of the particle may include: updating the velocity and position of the particle using the following formula:

[0030] V id (k + 1) = wV id (k) + c 1 r 1 (k)(P id (k) - X id (k)) + c 2 r 2 (k)(P gd (k) - X id (k))

[0031] X id (k + 1) = X id (k) + V id (k + 1)

[0032] Among them, V id (k) is the velocity of the i-th particle in the d-th dimension at the k-th iteration, X id (k) is the position of the i-th particle in the d-th dimension at the k-th iteration, P id is the individual historical optimal position of the i-th particle in the d-th dimension, P gd is the optimal position experienced by all particles, c 1 and c 2 are learning factors determined according to practical experience, r 1 , r 2 are random numbers uniformly distributed between [0, 1], w is the inertia weight that linearly decreases according to the number of iterations, and is expressed by the following formula:

[0033]

[0034] Among them, w max , w min are the maximum and minimum inertia weights respectively, t is the current iteration number, kmax is the maximum number of iterations.

[0035] In one embodiment, preferably, the multiple data sets further include a test set, and the motor temperature estimation method further includes: after the GPR model is trained through the training phase, the trained GPR model is verified through a test phase; if the trained GPR model meets the preset verification conditions in the test phase, the operating condition parameters and the initial temperature value are input into the GPR model, and the output value of the model is used as the temperature value of the motor at the corresponding moment of the input operating condition; if the trained GPR model does not meet the preset verification conditions in the test phase, the GPR model is retrained.

[0036] In one embodiment, preferably, the preset verification conditions include: using the following error function RMSE as an evaluation index to measure the prediction accuracy of the GPR model:

[0037]

[0038] where N represents the number of data groups in the test set, T t represents the steady-state temperature value of the motor at time t of the predicted output of the GPR model, and T t * represents the steady-state temperature value of the motor measured by the sensor at time t.

[0039] In one embodiment, preferably, inputting the operating condition parameters at the multiple moments into the GPR model trained through the training phase may include: inputting the operating condition parameter data at the multiple moments of the operating condition parameter categories optimized and screened based on the grey relational degree in the training phase into the GPR model to obtain the steady-state temperature values of the motor at the multiple moments.

[0040] In one embodiment, preferably, determining the parameters of the RC filter model includes: using the RC filter model to fit the actual temperature rise curve of the motor, thereby determining the parameters of the RC filter model.

[0041] In one embodiment, preferably, the motor temperature estimation method further includes: after the motor is powered on, performing initialization work to obtain the initial temperature value of the motor; initializing the RC filter model based on the initial temperature value to determine the parameters of the RC filter model.

[0042] In one embodiment, preferably, the initialization work includes: obtaining the vehicle power-off time, comparing the vehicle power-off time with a preset time threshold; in response to the vehicle power-off time being not less than the preset time threshold, using the current temperature of the motor cooling water inlet as the initial temperature value; and in response to the vehicle power-off time being lower than the preset time threshold, selecting a motor cooling curve based on the real-time ambient temperature, and calculating the current temperature of the motor according to the vehicle power-off time and the motor cooling curve as the initial temperature value.

[0043] Another aspect of the present invention also provides a motor temperature estimation device based on the GPR-RC model, including: a memory; and a processor coupled to the memory, the processor being configured to execute the motor temperature estimation method based on the GPR-RC model as described in any one of the above.

[0044] The present invention also provides a computer-readable medium, on which a computer program is stored, and when the computer program is executed by a processor, the motor temperature estimation method based on the GPR-RC model as described in any one of the above is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] After reading the detailed description of the embodiments of the present disclosure in conjunction with the following drawings, the above features and advantages of the present invention can be better understood. In the drawings, the components are not necessarily drawn to scale, and components having similar related characteristics or features may have the same or similar reference numerals.

[0046] Figure 1 is a schematic flowchart of a method for estimating motor temperature based on the GPR-RC model according to one aspect of the present invention;

[0047] Figure 2 is a schematic diagram of the principle of the GPR-RC model according to an embodiment of the present invention;

[0048] Figure 3A is a schematic flowchart of a method for estimating motor temperature based on the GPR-RC model in the GPR model training and testing stages according to an embodiment of the present invention;

[0049] Figure 3B is a data fitting result graph of fitting a first-order RC filter model based on the actual temperature rise curve of the motor according to an embodiment of the present invention;

[0050] Figure 4 is a schematic flowchart of the initialization work in the motor temperature estimation method according to an embodiment of the present invention;

[0051] Figure 5It is a schematic structural diagram of a motor temperature estimation device based on the GPR-RC model shown according to another aspect of the present invention; and

[0052] Figure 6 It is a schematic structural diagram of a hardware device for deploying the GPR-RC model shown according to an embodiment of the present invention.

[0053] For clarity, the following gives a brief description of the reference numerals:

[0054] 601 AMU unit

[0055] 602 DMA unit

[0056] 603 SRAM unit

[0057] 604 Non-volatile storage

[0058] 605 First path

[0059] 606 Second path Detailed implementation manners

[0060] The following specific embodiments illustrate the implementation manners of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Although the description of the present invention will be introduced in conjunction with preferred embodiments, this does not mean that the features of this invention are limited to this implementation manner. On the contrary, the purpose of introducing the invention in conjunction with the implementation manner is to cover other alternatives or modifications that may be extended based on the claims of the present invention. In order to provide a deep understanding of the present invention, many specific details will be included in the following description. The present invention can also be implemented without using these details. In addition, in order to avoid confusing or obscuring the focus of the present invention, some specific details will be omitted in the description.

[0061] In the description of the present invention, it should be noted that unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0062] In addition, the "upper", "lower", "left", "right", "top", "bottom", "horizontal", and "vertical" used in the following description should be understood as the orientations shown in this paragraph and the related drawings. This relative term is only for the convenience of description, and it does not mean that the device described needs to be manufactured or operated in a specific orientation, so it should not be understood as a limitation to the present invention.

[0063] It is understood that although terms such as "first", "second", "third", etc. may be used herein to describe various components, regions, layers, and / or parts, these components, regions, layers, and / or parts should not be limited by these terms, and these terms are only used to distinguish different components, regions, layers, and / or parts. Therefore, the first component, region, layer, and / or part discussed below may be referred to as the second component, region, layer, and / or part without departing from some embodiments of the present invention.

[0064] To overcome the above-mentioned defects existing in the prior art, the present invention provides a method and device for estimating the temperature of an electric motor based on a GPR-RC model. An electric motor thermal model is established based on a hybrid algorithm of Gaussian process regression and low-pass filtering. The GPR model is used to predict the steady-state temperature of the electric motor under a certain working condition, and the RC filter predicts the transient temperature change between two steady-state temperature points, thereby realizing real-time temperature estimation of the electric motor. At the same time, in the modeling process, complex analysis of the heat transfer path of the electric motor and the identification process of thermal network parameters can be avoided. For the problem that it is difficult to determine the model input variables and hyperparameters, the PSO algorithm is used to select the model input variables and optimize the hyperparameters, and the optimized GPR-RC thermal model of the electric motor is used to estimate the temperature of the electric motor, which is especially suitable for occasions where the physical model is particularly complex, improves the development efficiency of the electric motor, and thereby reduces the vehicle development cost.

[0065] Figure 1 It is a schematic flowchart of a method for estimating the temperature of an electric motor based on a GPR-RC model shown according to one aspect of the present invention. The GPR-RC model includes two sub-models, namely a GPR model and an RC filter model.

[0066] Please refer to Figure 1 , the method 100 for estimating the temperature of an electric motor based on a GPR-RC model provided by the present invention may include:

[0067] Step 101: Obtain the working condition parameters of the electric motor at multiple moments during continuous operation;

[0068] Step 102: Input the working condition parameters at the multiple moments into the GPR model trained in the training stage respectively to obtain the steady-state temperature values of the electric motor at the multiple moments. Among them, in the training stage, multiple working condition parameters are optimized based on the grey relational degree to screen out the category of working condition parameters more relevant to the electric motor temperature as the training data input into the GPR model, and then the GPR model is trained;

[0069] Step 103: Select a suitable RC filter model and determine the parameters of the RC filter model; and

[0070] Step 104: Estimate the real-time temperature of the motor based on the steady-state temperature values of the motor at the multiple moments output by the GPR model and the RC filter model.

[0071] Figure 2 It is a schematic diagram of the model principle of the GPR-RC model shown according to an embodiment of the present invention.

[0072] It can be combined with reference to Figure 2 , the motor temperature estimation method provided by the present invention uses an algorithm that combines Gaussian process regression and low-pass filtering to establish a motor thermal model to achieve the estimation of the real-time temperature of the motor. Gaussian Process Regression (GPR) is a non-parametric probability model based on Bayesian methods and is a powerful tool for simulating the non-linear behavior of systems. However, when using this method to establish a motor real-time temperature estimation model, there are problems such as many parameters and an overly large model size.

[0073] At the same time, RC filters are often used to model the heat transfer process inside the motor, where the heat resistance represents the heat exchange relationship between two nodes, and the heat capacity represents the thermal inertia of the node. Using the RC filter model can simulate the transient change process of the motor temperature. However, using this model requires accurate calculation of the motor losses. As mentioned above, the temperature of the automotive drive motor is affected by various factors. When using only one of the above algorithms to estimate the real-time temperature of the motor, there are certain difficulties and challenges. Therefore, the present invention proposes a motor thermal model established based on an algorithm that combines Gaussian process regression and low-pass filtering, and realizes the estimation of the real-time temperature of the motor by passing the temperature prediction value output by the GPR model through the RC filter model. Among them, the GPR model is used to predict the steady-state temperature that the motor can reach when continuously operating under a certain working condition. If the motor operating condition changes, the model prediction output changes, and the transient temperature change process is described by the output of the RC filter, effectively improving the calculation efficiency and reducing the motor development cost.

[0074] It is easy to understand that an overly high temperature of the automotive drive motor will cause changes in its parameters and affect its performance. Specific manifestations include, for example, an increase in the stator winding resistance, an increase in copper loss and an exacerbation of the temperature rise, demagnetization of the rotor permanent magnet, and a decrease in the output torque under the same current. Those skilled in the art can understand that the motor temperature change is related to the operating conditions of the motor itself, such as: speed, torque, cooling conditions, bus voltage, etc. Therefore, the motor temperature estimation method provided by the present invention uses these operating condition parameters as the input of the GPR model to estimate the real-time temperature of the motor.

[0075] More specifically, for example, in Figure 2 the shown embodiment, the input of the GPR model is the operating condition parameters at multiple moments during the continuous operation of the motor, such as the coolant temperature t coolt, coolant flow rate V coolt , DC bus voltage U dc , three-phase line current amplitude I s , torque T q , rotational speed n, etc. The GPR outputs the predicted steady-state temperature value T at the corresponding moment t-1 and T t Then, a continuous temperature curve is obtained through the RC filter model to estimate the real-time temperature of the motor at any moment.

[0076] Figure 3A is a schematic flow chart of the method in the GPR model training and testing stages in the motor temperature estimation method based on the GPR-RC model illustrated according to an embodiment of the present invention.

[0077] In a preferred embodiment, the training stage of the GPR model may first include: Step 301, collecting multiple groups of operating condition data of the motor during actual operation and preprocessing the data. The multiple groups of operating condition data may include the operating condition parameters of multiple operating conditions and the steady-state temperature of the motor under each operating condition.

[0078] More specifically, for example, an electric drive test bench can be used to conduct motor temperature measurement tests under different operating conditions, and then synchronously collect relevant electrical, mechanical, and temperature signals affecting the motor temperature, as well as temperature data of different parts of the motor during the test. The different operating conditions may be different ambient temperatures, different coolant temperatures, different coolant flow rates, different DC bus voltages, different rotational speeds, different torques, Active Short Circuit (ASC) mode, locked-rotor mode. The parts for measuring the motor temperature may be the stator three-phase windings, stator core, rotor magnet, rotor core, bearing, coolant outlet. The collected electrical, mechanical, and temperature signals may be bus voltage, bus current, power, motor d / q axis voltage, d / q axis current, three-phase line voltage, three-phase line current, modulation index, switching frequency, power factor, rotational speed, torque, coolant flow rate, coolant inlet temperature, electric drive system oil pump rotational speed, oil pump current, etc. Through the above tests, the motor temperature field data under different operating conditions are obtained, and a motor temperature field data set is established.

[0079] Considering the high cost of bench tests, in addition to using a test bench, in one embodiment, data acquisition can also be combined with a simulation model. For example, first select typical operating conditions for bench tests to obtain actual measurement data, then calibrate the simulation model using the test data to obtain a high-precision simulation model. Subsequently, different operating condition parameters can be used as inputs to the simulation model, and the temperature data of different parts of the motor under different operating conditions can be obtained through simulation. This data acquisition method can reduce the cost of bench tests, make it easier to obtain data for operating condition points that are difficult to measure in bench tests, and achieve cost reduction and efficiency improvement.

[0080] One of the above two data acquisition methods or a combination of the two can be adopted to obtain the motor temperature field data under different working conditions and establish a motor temperature field data set.

[0081] Understandably, after establishing the motor temperature field data set, it is necessary to perform data preprocessing on the data set, such as data filtering, missing value filling, etc. In particular, in the motor temperature estimation method provided by the present invention, multiple working condition parameters are optimized based on the grey relational grade during the training stage to screen out the working condition parameter categories more relevant to the motor temperature as the training data input into the GPR model, and then the GPR model is trained.

[0082] For example, in one embodiment, after collecting the motor temperature field data set, calculate the grey relational grade between variables and sort them. The output of the model is the temperature of a certain part of the motor to be predicted, but there are many working condition parameters related to the motor temperature, and selecting the optimal model input becomes a problem to be optimized. Here, we use the grey relational grade analysis method to select the model input. If there are M motor working condition parameters, first calculate the grey relational grade values between these working condition parameters and the motor temperature, as shown in the following formula:

[0083]

[0084] where y(k) is the motor temperature, x i (k) are multiple such working condition parameters related to the motor temperature, and ρ is the grey discrimination coefficient, which can take a value of 0.5 in this embodiment, for example.

[0085] After calculating the grey relational grade values of each working condition parameter with respect to the motor temperature, sort them from large to small according to the grey relational grade values to obtain [ζ 1 , ζ 2 ,..., ζ M .

[0086] Therefore, the multiple groups of working condition data mentioned in the present invention include the working condition parameters under multiple working conditions sorted by grey relational grade and the real-time temperature of the motor under each working condition. In one embodiment, the training stage may further include: preprocessing the multiple groups of working condition data sorted by grey relational grade and organizing them into a time series form of (x t , T t ), where x t is the working condition parameter of the motor at time t, T t is the steady-state temperature of the motor under the corresponding working condition at time t, and x t is expressed as:

[0087] x t = [t coolt , Vcoolt , U dc , …, I s , T q , n]

[0088] Among them, the operating condition parameters include the coolant temperature t coolt , the coolant flow rate V coolt , the DC bus voltage U dc , the line current amplitude I s , the motor torque T q and the motor speed n.

[0089] It should be noted that the selection of these operating condition parameters here is only for illustrative purposes, aiming to clearly explain the method steps in the GPR model training stage of the motor temperature estimation method provided by the present invention, rather than for limiting the protection scope of the present invention. In fact, other operating condition parameters that affect the motor temperature can also be selected according to actual needs for temperature estimation work.

[0090] Please continue to refer to Figure 3A , after data preprocessing, execute step 302: divide the multiple groups of operating condition data into multiple data sets, and the multiple data sets can include a training set and a test set. For example, the proportions of the training set and the test set can be 70% and 30% respectively. Among them, the training set is used to train the model to obtain better hyperparameters so as to obtain a better alternative model, and the test set is used to verify the performance of the model after training to determine the optimal model.

[0091] Then execute step 303: initialize the GPR-RC model and select a kernel function. It is easy to understand that since the determination of the parameters of the RC filter model in the motor temperature estimation method provided by the present invention relies on the fitting work of the actual temperature rise curve of the motor, the initialization work can obtain the initial temperature value of the motor, so as to provide a temperature rise curve to determine the RC model parameters. At the same time, a suitable kernel function also needs to be selected for the model, and then the hyperparameters of the model are solved.

[0092] In a preferred embodiment, the model input x t corresponding to the steady-state temperature T of the motor under the operating conditions at time t t can be expressed as:

[0093] T t = f(x t ) + ε t

[0094] Among them, f(·) represents the function of the GPR model output-input relationship, and ε t is the added Gaussian white noise with a mean of zero, and ε t satisfies the following expression to prevent model overfitting:

[0095]

[0096] where σ y is the standard deviation of ε t .

[0097] Furthermore, in a preferred embodiment, when using the GPR model to model the input and output data, it can be assumed that the function f(·) has a zero-mean multivariate Gaussian prior distribution, which can be expressed, for example, by the following expression:

[0098] f = f(x t ) ~ N(f|0, K)

[0099] In the formula, K is the covariance matrix, where K ij = k(x i , x j ), and k(·,·) is the kernel function, that is, the covariance function.

[0100] Next, an appropriate kernel function is selected for the GPR model. The selection of the kernel function is crucial for the performance of the model and can be flexibly selected according to the actual problem. For example, in a preferred embodiment, an appropriate kernel function is selected for the GPR model, which can include, for example: selecting the Squared Exponential Kernel (SE Kernel) as the kernel function, and the kernel function can be expressed by the following formula:

[0101]

[0102] where D is the data dimension, σ f , l are the hyperparameters to be optimized in the kernel function.

[0103] In a preferred embodiment, for the motor temperature estimation method provided by the present invention, after selecting an appropriate kernel function for the model, the working condition data of the training set sorted by the grey relational grade is input into the GPR model; and the PSO optimization algorithm is used to determine the hyperparameters of the GPR model, thereby completing the model training work.

[0104] It can be understood that in the previous steps, the motor temperature estimation method provided by the present invention has sorted multiple working condition parameters based on the grey relational grade. Subsequently, in an embodiment, preferably, according to the grey relational grade values between the sorted motor working condition parameters and the motor temperature, the parameter with the largest grey relational grade value, the parameters with the top two grey relational grade values are selected in sequence until all parameters are selected as the model input. In this way, we obtain the hyperparameter U to be optimized for determining the number of model input variables as follows:

[0105] U = [U 1 , U 2 ,..., UM

[0106] Among them, U 1 represents the operating condition parameter with the largest grey correlation degree value with the motor temperature, and U 2 represents the operating condition parameters with the grey correlation degree values with the motor temperature ranking in the top two, and so on, U M represents the operating condition parameters with the grey correlation degree values with the motor temperature ranking in the top M.

[0107] At this time, including U that determines the number of model input variables, the hyperparameters of the SE kernel function also include σ f , l, and the standard deviation σ y of the Gaussian white noise of the model. That is to say, a set of hyperparameters to be optimized in the GPR thermal model in the motor temperature estimation method provided by the present invention can be expressed as θ = [U, σ f , l, σ y .

[0108] Next, execute step 304: PSO initialization, and randomly generate an initial population of the hyperparameters of the GPR model. In one embodiment, first, the dimension of each particle can be determined according to the number of variables of the hyperparameters, and an initial population of the hyperparameters of the GPR model is randomly generated, including the initial velocity and the initial position of each particle.

[0109] For example, in one embodiment, there are 4 variables in total for the hyperparameters θ to be optimized in the model, so the dimension of each particle is 4-dimensional, and each particle vector represents a set of hyperparameters of the GPR model, and different GPR models can be determined. Set the particle swarm size to N. Optionally, for example, take N = 20, and the maximum number of iterations is K. Optionally, for example, take K = 1000. The initial position and the initial velocity of the particle can be determined by the following formulas respectively:

[0110] x i0 = rand([U, σ f , l, σ y )

[0111] V i0 = rand(V i )

[0112] In the formula, x i0 is the initial position of the particle randomly generated within the hyperparameter range, and V i0 is the initial velocity of the randomly generated particle.

[0113] Subsequently, execute step 305: Calculate the fitness value of each particle individual based on the fitness function, and the fitness value of each particle individual is the mean square error f id of the motor temperature prediction.

[0114] ​For example, the following formula can be used to calculate the mean square error of motor temperature prediction:

[0115]

[0116] In the formula, f is the mean square error of the model's prediction of the motor temperature, f(x i ) is the predicted motor temperature value of the i-th sample during model training, y i is the actual motor temperature value of the i-th sample, and M is the number of samples.

[0117] Further, step 306 is executed: Determine whether the optimization end condition is satisfied; if not, execute step 307: Update the velocity and position of the particle, and update the global optimum and local optimum.

[0118] Specifically, for example, if the mean square error f of the motor temperature prediction corresponding to a particle id is less than the minimum mean square error f(P id ) of the particle in the previous iteration history, then use this mean square error f of the motor temperature id to replace the previous round of f(P id ), use this particle to replace the previous round of particle, and update the velocity and position of the particle. In one embodiment, optionally, the updating of the velocity and position of the particle may include: updating the velocity and position of the particle using the following formula:

[0119] V id (k + 1) = wV id (k) + c 1 r 1 (k)(P id (k) - X id (k)) + c 2 r 2 (k)(P gd (k) - X id (k))

[0120] X id (k + 1) = X id (k) + V id (k + 1)

[0121] Among them, V id (k) is the velocity of the i-th particle in the d-th dimension at the k-th iteration, X id (k) is the position of the i-th particle in the d-th dimension at the k-th iteration, P id is the individual historical optimal position of the i-th particle in the d-th dimension, P gd is the optimal position experienced by all particles, c 1 and c 2 are learning factors determined according to practical experience, r1 , r 2 is a random number uniformly distributed between [0, 1], and w is an inertial weight that decreases linearly according to the number of iterations, which can be expressed by the following formula:

[0122]

[0123] where w max , w min are the maximum and minimum inertial weights respectively, t is the current iteration number, and k max is the maximum number of iterations.

[0124] If the mean square error of the minimum motor temperature f(P id ) of a particle is less than the global minimum mean square error of the motor temperature f(P gd ) of all particles, then use the mean square error of the minimum motor temperature of this particle to replace the original global minimum mean square error of the motor temperature, and at the same time save the current position of this particle:

[0125] x gd = x id = [U id , n u*id , n y*id , σ f*id , l id , σ y*id

[0126] In the formula, U is the optimal model input variable currently searched, n u and n y are the parameters of the optimal regression step length of the model output and the input step length currently searched respectively, which determine the modeling ability of the model for the dynamic change characteristics of temperature; σ f and l are the optimal hyperparameters of the model kernel function currently searched, which determine the ability of the model to explore and map the law of motor temperature change contained in the training data; σ y is the standard deviation of the Gaussian white noise of the model currently searched, which can enable the model to overcome the influence of measurement noise in the original data acquisition and improve the generalization ability of the model.

[0127] ​Then, return to step 305 to calculate the next particle, and iterate and update in turn until the optimization termination condition is met. For example, when the maximum number of iterations is reached or global convergence occurs, end the program and return the position of the particle with the minimum mean square error of the current motor temperature, that is, the optimal input and hyperparameters of the model obtained by PSO optimization, which are also the optimal model output step size, the optimal input step size, the optimal kernel function parameters, and the optimal Gaussian white noise standard deviation. If the termination condition is not met, perform a new round of iteration. At this time, update the position and velocity of the particle, generate a new particle, and return to step 305 until the maximum number of iterations or global convergence is reached. The algorithm ends, returns the optimal input variables and hyperparameters of the GPR model, and determines the optimal GPR model, which has the highest prediction accuracy for the motor temperature.

[0128] Use the PSO optimization algorithm to optimize the hyperparameters of the GPR model, reduce the difficulty of parameter optimization, and accelerate the model training speed.

[0129] So far, the model training stage is completed, and transfer to step 309: enter the test stage according to the determined GPR model.

[0130] In the test stage, the prior distribution and marginal likelihood distribution of the GPR model can be calculated first; then calculate the posterior distribution of the target value. According to the Bayesian posterior formula, the posterior distribution of the predicted output T of the GPR model for the input x * satisfies: * In the formula, μ

[0131]

[0132] is the mean of the predicted value, * and is the variance of the predicted value, which can be calculated by the following formula:

[0133]

[0134]

[0135] When taking the 95% confidence interval, the boundaries of the confidence interval can be described as:

[0136]

[0137] Next, use the test set determined in the previous steps to test the model accuracy. As shown in Figure 3, execute step 310: determine whether the accuracy of the model meets the requirements. For example, in a preferred embodiment, the preset verification condition may include: using the following error function RMSE as an evaluation index to measure the prediction accuracy of the GPR model:

[0138]

[0139] Among them, N represents the number of data groups in the test set, and T t represents the steady-state temperature value of the motor at time t of the predicted output of the GPR model, and T t * represents the steady-state temperature value of the motor measured by the sensor at time t.

[0140] It should be noted that here RMSE as an evaluation index for prediction accuracy is only for illustrative purposes, aiming to more clearly show the specific steps of the motor temperature estimation method provided by the present invention in the model testing stage, rather than being used to limit the protection scope of the present invention. In fact, other evaluation indexes can also be selected to measure the prediction accuracy of the model.

[0141] In this test stage, the motor temperature estimation method may further include: after the GPR model is trained in the training stage, then in the test stage, the trained GPR model is verified; if the trained GPR model meets the preset verification conditions in the test stage, the working condition parameters and the initial temperature value are input into the GPR model, and the output value of the model is used as the temperature value of the motor at the corresponding moment of the input working condition; if the trained GPR model does not meet the preset verification conditions in the test stage, the GPR model is retrained.

[0142] It is easy to understand that the evaluation process of the model prediction accuracy is the content of Figure 3A step 310 in Figure 3A As shown, if it is judged that the prediction accuracy cannot meet the preset requirements, then return to step 302 to retrain the model. For example, the proportion or order of the training set and the test set in the vehicle dataset can be changed and then the training work is carried out again; if after testing, the prediction accuracy of the model can meet the preset requirements, then terminate the training, output the model configuration file and save it for use in the formal motor temperature estimation work.

[0143] After the GPR model is trained, it is combined with the RC filter model to estimate the motor temperature, that is, Figure 3A step 311 in

[0144] Please return to Figure 1 , where step 103: select an appropriate RC filter model and determine the parameters of the RC filter model, which may further include: using the RC filter model to fit the actual temperature rise curve of the motor, so as to determine the parameters of the RC filter model.

[0145] For example, the RC filter model can be a first-order RC filter, and its expression can be as follows:

[0146]

[0147] Wherein, k is a proportionality coefficient, and τ is a time constant.

[0148] Figure 3B It is a data fitting result diagram of the first-order RC filter model fitting the actual temperature rise curve of the motor shown according to an embodiment of the present invention.

[0149] It can be combined with reference to Figure 3B , fitting the RC filter model based on the temperature values of the motor at the multiple moments to determine the parameters of the RC filter model, which may include: fitting the actual temperature rise curve of the motor using the above expression to obtain the parameters of the first-order RC filter. After obtaining the parameters of the RC filter through curve fitting, the transient temperature value at any point between the two steady-state temperature points output by the GPR model of the motor can be calculated according to the time from the RC curve, thereby realizing the estimation of the real-time temperature of the motor.

[0150] In a preferred embodiment, the method for estimating the motor temperature provided by the present invention may further include: after the motor is powered on, performing initialization work to obtain the initial temperature value of the motor; initializing the RC filter model based on the initial temperature value to determine the parameters of the RC filter model.

[0151] More specifically, for example, in the embodiment shown in Figure 3B , the temperature curve fitting work is carried out with the 0 moment as the initial moment. Among them, the darker scatter line represents the actual temperature rise curve of the motor, and the lighter curve represents the fitting result diagram of the parameter fitting based on the first-order RC model. The temperature value at the 0 moment is provided by the initialization work.

[0152] It should be noted that the RC filter model here is selected as a first-order RC filter, which is only for illustrative purposes and not for limiting the protection scope of the present invention. In fact, the RC filter can also be a second-order RC filter, a third-order RC filter or other suitable filters, which can all be applied to the method for estimating the motor temperature based on the GPR model provided by the present invention and should also be included in the protection scope of the present invention.

[0153] In another embodiment, the parameters in the RC filter model can also be obtained by fitting through the GPR method. The input is the working condition information, and the output is the parameters of the RC filter model. When actually used, the model of the RC filter parameters can be deployed to the CPU or can be assisted in calculation by means of a hardware acceleration unit.

[0154] Figure 4 It is a schematic flow diagram of the initialization work in the method for estimating the motor temperature shown according to an embodiment of the present invention.

[0155] AsFigure 4 As shown, in the motor temperature estimation method provided by the present invention, the initialization work may include: First, step 401: The vehicle starts and runs, and the vehicle is powered on; after the motor is powered on, step 402 is executed: Obtain the vehicle power-off time \(t_{EcuOff}\); then step 403 is executed: Compare the vehicle power-off time \(t_{EcuOff}\) with a preset time threshold \(t_{Max\_C}\); in response to the vehicle power-off time being not lower than the preset time threshold, enter step 404: Take the current temperature \(t\) of the motor coolant inlet coolt as the initial temperature value \(t\) init , that is, \(t\) init = \(t\) coolt ; at the same time, in response to the vehicle power-off time being lower than the preset time threshold, enter step 405: Select a motor cooling curve based on the real-time ambient temperature, and calculate the current temperature of the motor according to the vehicle power-off time and the motor cooling curve to be used as the initial temperature value \(t\) init ; finally, step 406: Determine the initial temperature value \(t\) of the model based on the above process init as the initial feedback input of the model.

[0156] After completing the initialization work of the GPR-RC model, in the motor temperature estimation method provided by the present invention, preferably, inputting the operating condition parameters at the multiple moments into the GPR model obtained through the training stage may further include: Inputting the operating condition parameter data at the multiple moments of the operating condition parameter categories optimized and screened based on the grey relational degree in the training stage into the GPR model to obtain the steady-state temperature values of the motor at the multiple moments. In other words, in the training stage, the operating condition parameter categories that are more relevant to the motor temperature are optimized and screened using the grey relational degree, and in the model usage stage, the parameters of these categories are also selected according to the relational degree for the real-time estimation work of the motor temperature, thereby further improving the estimation accuracy and the model operation efficiency.

[0157] It should be additionally noted that in the motor temperature estimation method based on the GPR-RC model provided by the present invention, the GPR-RC thermal model is not only applicable to the motors of new energy vehicle drive systems, but also applicable to other occasions where the motor is used as the power source, and its modeling objects may include various types such as permanent magnet synchronous motors, asynchronous motors, reluctance motors, brushless DC motors, etc.

[0158] In addition, the input of the motor GPR-RC thermal model may be the operating condition information related to the motor temperature, the temperature predicted by the model, or a combined form of these operating condition information. The predicted temperature may be any part of the motor that needs to monitor the temperature, not limited to the temperatures of the stator three-phase windings, stator core, rotor permanent magnet, rotor core, bearing, coolant outlet, etc.

[0159] The motor temperature estimation method based on the GPR-RC model provided by the present invention uses a data-driven modeling algorithm, avoiding the process of complex motor heat transfer path analysis and thermal network parameter identification, and is particularly suitable for occasions where the physical model is extremely complex, improving the development efficiency of the motor. At the same time, the above-mentioned modeling process takes into account the uncertainty between input and output, and can give the confidence interval of the output prediction result; and multi-source information can be fused during model training, improving the applicability of the model under various working conditions.

[0160] In addition, the present invention combines the GPR model and the RC filter model. A motor steady-state temperature prediction model is established through the GPR model, and the RC filter model predicts the continuous temperature change of the motor. Since the size of the motor GPR steady-state thermal model is significantly smaller than that of the motor GPR dynamic thermal model, and the number of parameters of the RC filter model is small, the overall model size of the motor temperature estimation method combining the GPR steady-state thermal model and the RC model is small, which helps to improve the calculation efficiency, simplify the operation process, and further improve the development efficiency. Moreover, once the model training is completed, based on this data-driven model, it can replace the traditional physical temperature sensor, effectively achieving cost savings in the vehicle development stage.

[0161] Furthermore, the working condition parameters related to the motor temperature are optimally selected as the model input variables based on the grey relational degree, and the PSO optimization method is used to optimize the solution of the model parameters, making the motor temperature estimation work more accurate, fast and efficient.

[0162] The motor GPR-RC thermal model on which the motor temperature estimation method provided by the present invention is based is a hybrid model. The GPR sub-model establishes a motor steady-state temperature prediction model, and the RC filter sub-model predicts the transient temperature change of the motor. The size of the motor GPR steady-state thermal model is significantly smaller than that of the motor GPR dynamic thermal model, and the number of parameters of the RC filter model is small. Therefore, the size of the motor GPR-RC thermal model based on the hybrid algorithm is small, further accelerating the operation efficiency of the temperature estimation work.

[0163] Although the above methods are illustrated and described as a series of actions to simplify the explanation, it should be understood and appreciated that these methods are not limited by the order of the actions, because according to one or more embodiments, some actions may occur in a different order and / or concurrently with other actions that are illustrated and described herein or not illustrated and described herein but are understandable to those skilled in the art.

[0164] Figure 5 It is a schematic diagram of the device structure of the motor temperature estimation device based on the GPR-RC model shown according to another aspect of the present invention.

[0165] According to another aspect of the present invention, an embodiment of a motor temperature estimation device 500 based on the GPR-RC model is also provided herein.

[0166] As Figure 5 shown, the above-mentioned motor temperature estimation device 500 based on the GPR-RC model provided in this embodiment may include a memory 501 and a processor 502 coupled to the memory 501. The processor 502 may be configured to implement any of the above-mentioned motor temperature estimation methods based on the GPR-RC model.

[0167] According to another aspect of the present invention, an embodiment of a computer storage medium is also provided herein. A computer program is stored on the computer storage medium. When the computer program is executed by a processor, the steps of any of the above-mentioned motor temperature estimation methods based on the GPR-RC model can be implemented.

[0168] Those skilled in the art will further appreciate that the various illustrative logical blocks, modules, circuits, and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or a combination of both. To clearly illustrate this interchangeability of hardware and software, the various illustrative components, blocks, modules, circuits, and steps are described above in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and the design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention. The processors described in this case may be implemented using electronic hardware, computer software, or any combination thereof. Whether such a processor is implemented as hardware or software will depend upon the particular application and the overall design constraints imposed on the system. By way of example, the processors presented in this disclosure, any part of a processor, or any combination of processors may be implemented using a microprocessor, a microcontroller, a digital signal processor (DSP), a field programmable gate array (FPGA), a programmable logic device (PLD), a state machine, gated logic, discrete hardware circuits, and other suitable processing components configured to perform the various functions described throughout this disclosure. The functionality of the processors presented in this disclosure, any part of a processor, or any combination of processors may be implemented using software executed by a microprocessor, a microcontroller, a DSP, or other suitable platform.

[0169] The steps of the methods or algorithms described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. The software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor such that the processor can read from, and write to, the storage medium. In an alternative, the storage medium may be integrated into the processor. The processor and the storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In an alternative, the processor and the storage medium may reside as discrete components in a user terminal.

[0170] In one or more exemplary embodiments, the described functionality may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software as a computer program product, the functions may be stored on or transmitted via a computer-readable medium as one or more instructions or code. The computer-readable medium includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. A storage medium may be any available medium that can be accessed by a computer. By way of example and not limitation, such computer-readable medium can comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer. Any connection is properly termed a computer-readable medium.

[0171] For example, if the software is transmitted from a web site, server, or other remote source using a coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of medium. As used herein, disk and disc include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk, and Blu-ray disc, where disk generally reproduces data magnetically, while disc uses lasers optically to reproduce data. Combinations of the above should also be included within the scope of computer-readable medium.

[0172] More specifically, for example, the model file optimized and trained by PSO in the previous step can be deployed to the micro - control unit (MCU) of the motor controller, and the motor GPR - RC thermal model can be deployed to run in the central processing unit (CPU) within the MCU.

[0173] Since the computing power required to run the motor GPR - RC model is relatively large, the model can also be accelerated by means of the hardware acceleration unit inside the MCU. The CPU calls the driver of the hardware acceleration unit to assist in the calculation of the motor GPR - RC thermal model through the hardware acceleration unit, so as to accelerate the operation speed of the thermal model.

[0174] Figure 6 It is a schematic structural diagram of a hardware device for deploying the GPR - RC model according to an embodiment of the present invention.

[0175] As Figure 6 shown, it is the structure of the common hardware acceleration device AMU unit 601 (Advanced Modelling Unit) inside the vehicle - used MCU, which may include: several AMU computing engines AMU0..n (n = 1, 2, 3...), DMA unit 602 (Direct Memory Access), and SRAM unit 603 (System RAM, volatile storage). Among them, the AMU unit 601 can directly obtain the parameters of the GPR sub - model in the motor GPR - RC thermal model from non - volatile storage 604, such as FLASH, for model operation, that is, the first path 605 shown by the arrow; it can also transfer the model parameters from the FLASH area to the SRAM unit 603 of the hardware acceleration unit through the DMA unit 602 of the AMU unit 601, and the AMU unit 601 obtains the parameters of the motor GPR part model from the SRAM unit 603 for model operation, that is, the second path 606 shown by the arrow, which can shorten the time for the AMU computing engine to obtain the model parameters and further improve the model operation efficiency.

[0176] Meanwhile, the parameters in the RC filter can also be fitted by the GPR method. The input is the working condition information, and the output is the parameters of the RC filter. The model of the RC filter parameters can be deployed to the CPU or assisted in calculation by means of the hardware acceleration unit.

[0177] In addition, the GPR-RC model can also be deployed to processing units other than the MCU within the motor controller, including: a micro-processing unit (MPU), and other controllers outside the motor controller, such as a regional controller, a domain controller, an in-vehicle computer, and a cloud server, etc. Data interaction can be carried out between them and the motor controller MCU through a communication medium, such as SPI (Serial Peripheral Interface), UART (Universal Asynchronous Receiver / Transmitter), CAN (Controller Area Network), LIN (Local Interconnect Network), ETH (Ethernet), and wireless communication, etc., so as to realize real-time temperature prediction of the motor.

[0178] In one embodiment, after the model is deployed, data can be transmitted to the cloud, and through big data analysis, it can be judged whether there are problems such as excessive scatter and aging of the motor, and the model can be corrected.

[0179] The foregoing description of the disclosure has been provided to enable any person skilled in the art to make or use the disclosure. Various modifications to the disclosure will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other variations without departing from the spirit or scope of the disclosure. Thus, the disclosure is not intended to be limited to the examples and designs described herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A motor temperature estimation method based on a GPR-RC model, wherein the GPR-RC model includes two sub-models, a GPR model and an RC filter model, and the motor temperature estimation method includes: Obtain the operating parameters of the motor at multiple moments during continuous operation; Inputting the operating parameters at the multiple moments into the GPR model trained in the training phase respectively, so as to obtain the steady-state temperature values ​​of the motor at the multiple moments, wherein in the training phase, the multiple operating parameters are optimized based on the grey correlation degree, so as to select the operating parameter categories that are more relevant to the motor temperature as the training data input to the GPR model, thereby training the GPR model; Selecting the adapted RC filter model and determining parameters of the RC filter model; and The real-time temperature of the motor is estimated based on the steady-state temperature values ​​of the motor at the multiple moments output by the GPR model and the RC filter model.

2. The motor temperature estimation method according to claim 1, characterized in that: The optimization of multiple operating parameters based on grey correlation degree includes: The following formulas are used to calculate the grey correlation values ​​of multiple operating parameters and motor temperature: Among them, y(k) is the motor temperature, x i (k) is a plurality of operating parameters related to the motor temperature, ρ is a gray resolution coefficient; and Sort multiple operating condition parameters in descending order according to their corresponding grey correlation values ​​to obtain the hyperparameter U to be optimized for determining the number of input variables of the GPR model: U=[U1,U2,…,U M ] Among them, U1 represents the operating parameter with the largest gray correlation value with the motor temperature, U2 represents the operating parameters with the top two gray correlation values ​​with the motor temperature, and so on. M Indicates the operating parameters with the top M gray correlation values ​​with the motor temperature.

3. The motor temperature estimation method according to claim 2, characterized in that: The training phase includes: Collecting multiple groups of working condition data of the motor in actual operation, dividing the multiple groups of working condition data into multiple data sets, and performing optimization sorting based on the grey correlation degree, the multiple data sets including a training set and a verification set; Selecting an adaptive kernel function for the GPR model, and inputting the operating condition data of the training set sorted by grey correlation degree into the GPR model; and The PSO optimization algorithm is used to determine the hyperparameters of the GPR model, thereby completing the model training work.

4. The motor temperature estimation method according to claim 3, characterized in that: The multiple groups of operating condition data include operating condition parameters of multiple operating conditions and the steady-state temperature of the motor under each operating condition. The training stage also includes: The multiple groups of working condition data sorted by grey correlation degree are preprocessed and sorted into (x t ,T t ) in time series form, x t is the operating parameter of the motor at time t, T t is the steady-state temperature of the motor under the corresponding working condition at time t, x t It is expressed as: x t =[t coolt ,V coolt ,U dc ,…,I s ,T q ,n] Among them, the operating parameters include the coolant temperature t coolt , Coolant flow V coolt , DC bus voltage U dc , line current amplitude I s , motor torque T q and motor speed n.

5. The motor temperature estimation method according to claim 4, characterized in that: The steady-state temperature T of the motor under the working condition corresponding to the time t t It is expressed as: T t =f(x t )+ε t Where, f(·) represents the GPR model, ε t is the added Gaussian white noise with zero mean, and ε t The following expression is satisfied to prevent the model from overfitting: where σ y is t The standard deviation of .

6. The motor temperature estimation method according to claim 5, characterized in that: The GPR model also satisfies the following expression: f=f(x t )~N(f|0,K) Among them, K is the covariance matrix, K ij = k(x i ,x j ), k(·,·) is the kernel function; The step of selecting an adapted kernel function for the GPR model comprises: The square exponential function is selected as the kernel function, and the kernel function is expressed by the following formula: Where D is the data dimension, σ f , l are the hyperparameters to be optimized in the kernel function.

7. The motor temperature estimation method according to claim 6, characterized in that: The hyperparameters of the GPR model are θ = [U, σ f ,l,σ y ], the PSO optimization algorithm is used to determine the hyper parameters of the GPR model, including: Determine the dimension of each particle according to the number of variables of the hyperparameters, and randomly generate an initial population of the hyperparameters of the GPR model, including the initial velocity and initial position of each particle; The fitness value of each individual particle is calculated based on the fitness function, and the fitness value of each individual particle is the mean square error f of the motor temperature prediction. id ; If the mean square error of the motor temperature prediction corresponding to a particle is f id Less than the minimum temperature mean square error f(P id ), then use the motor temperature mean square error f id Replace the previous round of f(P id ), use this particle to replace the particle in the previous round, and update the particle's speed and position; If the minimum motor temperature mean square error f(P id ) is less than the global minimum motor temperature mean square error f(P gd ), the original global minimum motor temperature mean square error is replaced by the minimum motor temperature mean square error of the particle, and the current position of the particle is saved; and When the preset optimization end condition is met, the current position of the particle is used as the optimal GPR model input and hyperparameters obtained by optimization.

8. The motor temperature estimation method according to claim 7, characterized in that: The updating of the speed and position of the particle includes: Use the following formula to update the particle's velocity and position: V id (k+1)=wV id (k)+c1r1(k)(P id (k)-X id (k))+c2r2(k)(P gd (k)-X id (k)) X id (k+1)=X id (k)+V id (k+1) Among them, V id (k) is the velocity of the ith particle in the dth dimension at the kth iteration, X id (k) is the position of the ith particle in the dth dimension at the kth iteration, P id is the individual historical optimal position of the ith particle in the dth dimension, P gd is the optimal position experienced by all particles, c1 and c2 are learning factors determined based on practical experience, r1 and r2 are uniformly distributed random numbers between [0,1], and w is the inertia weight that decreases linearly according to the number of iterations, expressed as follows: Among them, w max 、w min are the maximum and minimum inertia weights respectively, t is the current iteration number, k max is the maximum number of iterations.

9. The motor temperature estimation method according to claim 3, characterized in that: The plurality of data sets also include a test set, and the motor temperature estimation method further includes: After the GPR model is trained in the training phase, the trained GPR model is verified in the testing phase; If the trained GPR model meets the preset verification condition in the test phase, the operating condition parameter and the initial temperature value are input into the GPR model, and the output value of the model is used as the temperature value of the motor at the time corresponding to the input operating condition; and If the trained GPR model does not meet the preset verification condition during the testing phase, the GPR model is retrained.

10. The motor temperature estimation method according to claim 9, characterized in that: The preset verification conditions include: The following error function RMSE is used as an evaluation index to measure the prediction accuracy of the GPR model: Where N represents the number of data sets in the test set, T t represents the steady-state temperature value of the motor at time t, which is the predicted output of the GPR model. t * It represents the steady-state temperature value of the motor at time t measured by the sensor.

11. The motor temperature estimation method according to claim 1, characterized in that: The step of inputting the operating parameters at the plurality of moments into the GPR model obtained through training in the training phase comprises: The operating parameter data at multiple moments of the operating parameter category selected based on the grey correlation degree optimization in the training phase are input into the GPR model to obtain the steady-state temperature values ​​of the motor at the multiple moments.

12. The motor temperature estimation method according to claim 1, characterized in that: The determining of the parameters of the RC filter model comprises: The RC filter model is used to fit the measured temperature rise curve of the motor, so as to determine the parameters of the RC filter model.

13. The motor temperature estimation method according to claim 1, characterized in that: The motor temperature estimation method further includes: After the motor is powered on, an initialization operation is performed to obtain an initial temperature value of the motor; The RC filter model is initialized based on the initial temperature value to determine parameters of the RC filter model.

14. The motor temperature estimation method according to claim 13, characterized in that: The initialization work includes: Obtaining a vehicle power-off time, and comparing the vehicle power-off time with a preset time threshold; In response to the vehicle power-off time being not less than the preset time threshold, taking the current temperature of the motor cooling water inlet as the initial temperature value; and In response to the vehicle power-off time being lower than the preset time threshold, a motor cooling curve is selected based on the real-time ambient temperature, and the current temperature of the motor is calculated according to the vehicle power-off time and the motor cooling curve as the initial temperature value.

15. A motor temperature estimation device based on a GPR-RC model, comprising: Memory; as well as A processor coupled to the memory, wherein the processor is configured to execute the motor temperature estimation method based on the GPR-RC model according to any one of claims 1 to 14.

16. A computer readable medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the motor temperature estimation method based on the GPR-RC model according to any one of claims 1 to 14 is implemented.

Images