Cooling liquid flow estimation method, bus capacitor temperature estimation method and storage medium

Abstract

The invention relates to a cooling liquid flow estimation method for a motor controller. The cooling liquid flow estimation method comprises the following steps: acquiring at least one working parameter related to the current operation condition of a motor; and inputting the obtained at least one working parameter into a neural network model to determine the coolant flow of the motor controller, the training process of the neural network model comprising: creating a plurality of data sets to be used for training; inputting each created data set into a neural network model to obtain a coolant flow estimated value; a flow meter is used for collecting a cooling liquid flow true value of the motor controller under each operation condition; and carrying out iteration and optimization on model parameters of the neural network model by utilizing the collected cooling liquid flow true value so as to enable a prediction result of the neural network model to approach the cooling liquid flow true value. The invention further relates to a bus capacitor temperature estimation method for the motor controller and a computer storage medium.

Description

Technical Field

[0001] The present invention relates to the field of motor controllers, and more specifically, to a coolant flow estimation method for a motor controller, a busbar capacitor temperature estimation method for a motor controller, and a computer storage medium. Background Art

[0002] In electric vehicles, the vehicle control unit (VCU) is usually used to obtain the coolant flow information of the motor controller, or to calculate the coolant flow of the motor controller based on the known coolant temperature and loss. Usually, the cooling system of the whole vehicle is not designed separately for the motor controller. Therefore, for a cooling system with multiple cooling channels, the flow information provided by the vehicle VCU is not necessarily the flow information for the motor controller, so the information may be inaccurate.

[0003] In addition, the coolant flow rate can be estimated by the electronic pump of the motor cooling system. However, when the coolant of the motor controller comes from the vehicle coolant, the coolant flow rate of the motor controller cannot be accurately calculated based on the total electronic pump information, and the process also requires a lot of calibration at the vehicle level.

[0004] In addition, in current technology, estimating the temperature of the bus capacitor in the motor controller not only requires designing a complex thermal capacitance and thermal resistance model, but also requires calibrating the losses under different switching modes and correcting the estimated temperature, which consumes a lot of manpower costs and computing resources. Summary of the invention

[0005] In view of this, according to a first aspect of the present invention, a method for estimating coolant flow of a motor controller is proposed, the method comprising the following steps:

[0006] Acquire at least one operating parameter related to the current operating condition of the motor; and

[0007] inputting the obtained at least one operating parameter into a neural network model having pre-trained model parameters to determine a coolant flow rate of the motor controller,

[0008] The training process of the neural network model includes the following steps:

[0009] - creating a plurality of data sets to be used for training, each data set comprising at least one operating parameter of the motor under a predetermined operating condition;

[0010] - Input each created data set into a neural network model to obtain an estimated coolant flow rate corresponding to the motor controller under each operating condition;

[0011] -Use flow meter to collect the actual value of coolant flow of motor controller under each operating condition; and

[0012] -Use the collected real value of the coolant flow rate to iterate and optimize the model parameters of the neural network model so that the prediction result of the neural network model is close to the real value of the coolant flow rate.

[0013] According to an alternative embodiment,

[0014] The at least one operating parameter is selected from the group consisting of motor speed, DC bus voltage, motor modulation mode, power module temperature, motor D-axis current, motor Q-axis current, and a switching frequency of a motor controller and a coolant temperature.

[0015] According to an optional embodiment, the neural network model adopts a multi-layer perceptron architecture, and the multi-layer perceptron architecture includes:

[0016] An input layer, the input layer is used to receive a plurality of data sets to be used for training and perform normalization processing on the plurality of data sets;

[0017] A hidden layer, wherein the hidden layer is used to process and transform the normalized data set to extract feature information of the data set; and

[0018] An output layer is used to output the prediction result of the neural network model according to the feature information provided by the hidden layer.

[0019] According to an optional embodiment, the hidden layer includes a plurality of nodes, each node including an activation function with a predetermined weight coefficient.

[0020] According to an optional embodiment, the neural network model optimizes and iterates the weight coefficients of the activation function using a back propagation algorithm to determine model parameters of the neural network model.

[0021] According to an optional embodiment, the motor controller includes a main drive inverter for an electric vehicle.

[0022] According to a second aspect of the present invention, a bus capacitance temperature estimation method for a motor controller is also provided, the bus capacitance temperature estimation method comprising the following steps:

[0023] Acquire at least one operating parameter related to the current operating condition of the motor; and

[0024] inputting the obtained at least one operating parameter into a neural network model having pre-trained model parameters to determine a bus capacitor temperature of the motor controller,

[0025] The training process of the neural network model includes the following steps:

[0026] - creating a plurality of data sets to be used for training, each data set comprising at least one operating parameter of the motor under a specified operating condition;

[0027] - Input each created data set into a neural network model to obtain an estimated value of the bus capacitor temperature corresponding to each operating condition of the motor controller;

[0028] - Using a thermometer to collect the actual value of the busbar capacitor temperature under each operating condition; and

[0029] -Using the collected real value of the bus capacitor temperature to iterate and optimize the model parameters of the neural network model, so that the prediction result of the neural network model is close to the real value.

[0030] According to an optional embodiment, the at least one operating parameter is selected from a group including motor speed, DC bus voltage, motor modulation mode, power module temperature, motor D-axis current, motor Q-axis current, motor D-axis voltage, motor Q-axis voltage, and coolant flow, switching frequency and coolant temperature of the motor controller.

[0031] According to an optional embodiment, the at least one operating parameter includes a coolant flow rate of the motor controller, and the coolant flow rate is determined using the coolant flow rate estimation method as described above.

[0032] According to a third aspect of the present invention, there is also provided a computer-readable storage medium on which a computer program is stored. The computer program includes program instructions, and when the program instructions are executed by a processor, the coolant flow estimation method as described above or the bus capacitance temperature estimation method as described above is implemented.

[0033] The coolant flow estimation method and bus capacitance temperature estimation method for a motor controller according to the present invention can achieve at least one of the following advantages:

[0034] - No need to design complex thermal capacitance and thermal resistance algorithms, only different neural network models can be obtained through training to estimate the busbar capacitance temperature, which reduces the complexity of algorithm design and improves calculation efficiency;

[0035] - The estimation of coolant flow and bus capacitor temperature can be realized without relying on additional sensors and information outside the motor controller, saving generation cost;

[0036] - For different models of motor controllers, there is no need to redesign the algorithm logic. Only model training based on different motor parameters is required to estimate the coolant flow and bus capacitor temperature of different motor controllers, which improves production efficiency; and

[0037] -For cooling systems with multiple cooling channels, it does not rely on the flow information provided by the vehicle VCU, thus ensuring the accuracy of the estimation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] By incorporating the accompanying drawings and Figure 1 With reference to the specific embodiments used to illustrate certain principles of the present invention, other features and advantages of the method of the present invention will become clear or be described in more detail.

[0039] Figure 1 A flow chart of a method for estimating coolant flow rate (or bus capacitor temperature) of a motor controller according to an exemplary embodiment of the present invention is shown.

[0040] Figure 2 A schematic diagram showing a neural network model architecture used in the coolant flow estimation method according to the present invention.

[0041] Figure 3 A schematic diagram showing a neural network model architecture used in the busbar capacitance temperature estimation method according to the present invention is shown. DETAILED DESCRIPTION

[0042] The coolant flow estimation method and bus capacitor temperature estimation method for a motor controller according to the present invention will be described below with reference to the accompanying drawings and by way of example. In the following description, many specific details are set forth so that those skilled in the art can more fully understand the present invention. However, it is obvious to those skilled in the art that the present invention may be implemented without some of these specific details. Instead, it may be considered to implement the present invention with any combination of the following features and elements, regardless of whether they relate to different embodiments. Therefore, the following aspects, features, embodiments and advantages are for illustrative purposes only and should not be regarded as elements or limitations of the claims.

[0043] In view of the defects of the prior art methods for calculating coolant flow and / or bus capacitance, such as the complicated calibration process and model design and inaccurate calculation results, the present invention proposes a new method for estimating the coolant flow and bus capacitance temperature of a motor controller. This method does not increase the cost of the sensor, does not require a complex thermal model structure and data calibration process, and can achieve improved flow and bus capacitance temperature estimation accuracy only through a parameter model trained by the collected data set. This method is targeted at different motor controllers, does not require changing the algorithm logic, and achieves optimized training and compatibility for a variety of motor controllers, making the estimation results of the coolant flow and bus capacitance temperature more accurate and stable.

[0044] Figure 1 A flow chart of a method for estimating the coolant flow rate (or bus capacitor temperature) of a motor controller according to an exemplary embodiment of the present invention is shown. The core of the method is to estimate the coolant flow rate or bus capacitor temperature of the motor controller using a neural network model. In this article, the "motor controller" may include, in particular, a main drive inverter in an electric vehicle.

[0045] The model parameters of the neural network model need to be determined in advance during the training process. In the training process, multiple data sets to be used for training need to be created first, each data set includes at least one operating parameter of the motor under predetermined operating conditions, and the operating parameter includes but is not limited to: motor speed, DC bus voltage, motor modulation mode, power module temperature, motor D-axis current, motor Q-axis current, and the switching frequency and coolant temperature of the motor controller.

[0046] Each created data set is fed into the initial neural network model to obtain an estimate of the coolant flow rate for the motor controller under each operating condition.

[0047] Subsequently, a flow meter is used to collect the actual value of the coolant flow of the motor controller under each operating condition. This actual value can be used as the target result to be compared with the output result of the neural network model to iterate and optimize the model parameters of the neural network model.

[0048] Specifically, the optimization process of the neural network model can be implemented using a loss function. For example, a gradient descent algorithm can be used to make the loss function converge to a minimum threshold. This process ensures that the prediction result of the neural network model is infinitely close to the true value collected by the flow meter. For example, the difference between the predicted value and the true value is less than a threshold range. If this ideal state is reached, the current model parameters can be loaded into the neural network model as the final training result.

[0049] After the final model parameters and neural network model are determined, the online estimation of coolant flow or bus capacitor temperature can be performed.

[0050] In the process of online estimation of coolant flow, it is first necessary to obtain at least one working parameter related to the current operating condition of the motor. The working parameter can be a physical quantity collected in real time by a sensor, or an existing parameter read from the microcontroller of the motor controller. For example, the working parameter can include motor speed, DC bus voltage, motor modulation mode, power module temperature, motor D-axis current, motor Q-axis current, and at least one of the switching frequency and coolant temperature of the motor controller. Substituting these working parameters as input into the pre-trained neural network model, the coolant flow (or bus capacitor temperature) of the motor controller can be predicted.

[0051] Figure 2 FIG. 2 is a schematic diagram showing a neural network model architecture used in the coolant flow estimation method according to the present invention. Figure 2 As shown in , the neural network model can adopt a multilayer perceptron (MLP) architecture. The basic architecture of MLP consists of three levels: input layer, hidden layer and output layer.

[0052] Among them, the input layer is the bottom layer, which is used to receive input data. The input layer may include multiple input nodes, each input node corresponds to a feature or attribute of the input data. Therefore, the number of nodes in the input layer can be determined by the number of features of the input data.

[0053] exist Figure 2 In the architecture shown in , the input layer can receive multiple data sets to be used for training, each of which may include at least one operating parameter of the motor under a predetermined operating condition, including motor speed, DC bus voltage, motor modulation mode, power module temperature, motor D-axis current, motor Q-axis current, and the switching frequency and coolant temperature of the motor controller, etc. These data also need to be normalized in the input layer to ensure that the data between each feature is in the same range, so as to avoid the difference between different features affecting the training effect of the model.

[0054] The hidden layer is the core of the MLP architecture, which can process and transform the normalized data set to extract the characteristic information of the data set. The hidden layer usually includes a plurality of nodes, each of which includes an activation function with a predetermined weight coefficient. For example, in the flow estimation method of the present invention, the activation function can be a ReLu function. In the training process of the neural network model, the focus is on determining the weight coefficient of the activation function of each node in the MLP architecture.

[0055] The output layer is used to output the prediction result of the neural network model according to the feature information provided by the hidden layer. The output layer usually has only one node for outputting a single prediction result.

[0056] During the training process, the back propagation algorithm (BP algorithm for short) can be used to continuously optimize and iterate the weight coefficients of the activation function in the hidden layer to determine the final model parameters. The BP algorithm is a common neural network learning algorithm, which aims to optimize the weight parameters of the neural network through the training data set so that the output of the neural network is as close to the true value as possible.

[0057] Figure 3 FIG. 1 is a schematic diagram showing a neural network model architecture used in the busbar capacitance temperature estimation method according to the present invention. Figure 2 Similarly, the neural network model can also be trained using the MLP architecture, and its training process is roughly the same as the training process in the coolant flow estimation method. Figure 2 The difference between the MLP architecture in Figure 3 The number of input nodes of the MLP architecture in is larger.

[0058] like Figure 3 As shown in, the operating parameters received by the input layer may include motor speed, DC bus voltage, motor modulation mode, power module temperature, motor D-axis current, motor Q-axis current, motor D-axis voltage, motor Q-axis voltage, and the coolant flow, switching frequency and coolant temperature of the motor controller.

[0059] It is particularly noteworthy that Figure 3 The coolant flow rate in the input layer can be utilized Figure 2 By superimposing two MLP architectures, the bus capacitor temperature in the motor controller can be directly obtained based on the input parameters.

[0060] In addition, in the busbar capacitance temperature estimation method of the present invention, the activation function in the hidden layer can be selected as a tanh function.

[0061] In summary, the coolant flow estimation method and bus capacitance temperature estimation method for a motor controller according to the present invention can achieve at least one of the following advantages:

[0062] - No need to design complex thermal capacitance and thermal resistance algorithms, only different neural network models can be obtained through training to estimate the busbar capacitance temperature, which reduces the complexity of algorithm design and improves calculation efficiency;

[0063] - The estimation of coolant flow and bus capacitor temperature can be realized without relying on additional sensors and information outside the motor controller, saving generation cost;

[0064] - For different models of motor controllers, there is no need to redesign the algorithm logic. Only model training based on different motor parameters is required to estimate the coolant flow and bus capacitor temperature of different motor controllers, which improves production efficiency; and

[0065] -For cooling systems with multiple cooling channels, it does not rely on the flow information provided by the vehicle VCU, thus ensuring the accuracy of the estimation results.

[0066] In an exemplary embodiment of the present application, a computer-readable storage medium is also provided, on which a computer program is stored, the program including executable program instructions, which, when executed by, for example, a processor, can implement the various steps of the coolant flow estimation method or the bus capacitance temperature estimation method described in any embodiment of the present invention.

[0067] It can be understood by those skilled in the art that the example implementation of this document can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the implementation of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, or a network device, etc.) to execute the coolant flow estimation method or the busbar capacitance temperature estimation method according to the implementation of the present application.

[0068] Although the present invention has been disclosed as above with preferred embodiments, the present invention is not limited thereto. Any changes and modifications made by any person skilled in the art without departing from the spirit and scope of the present invention should be included in the protection scope of the present invention, and therefore the protection scope of the present invention should be subject to the scope defined by the claims.

Claims

1. A method for estimating the coolant flow rate of a motor controller, It is characterized in that The method comprises the following steps: Acquire at least one operating parameter related to the current operating condition of the motor; and inputting the obtained at least one operating parameter into a neural network model having pre-trained model parameters to determine a coolant flow rate of the motor controller, The training process of the neural network model includes the following steps: - creating a plurality of data sets to be used for training, each data set comprising at least one operating parameter of the motor under a predetermined operating condition; - Input each created data set into a neural network model to obtain an estimated coolant flow rate corresponding to the motor controller under each operating condition; -Use flow meter to collect the actual value of coolant flow of motor controller under each operating condition; and -Use the collected real value of the coolant flow rate to iterate and optimize the model parameters of the neural network model so that the prediction result of the neural network model is close to the real value of the coolant flow rate.

2. The method for estimating the coolant flow rate of a motor controller according to claim 1, It is characterized in that The at least one operating parameter is selected from the group consisting of motor speed, DC bus voltage, motor modulation mode, power module temperature, motor D-axis current, motor Q-axis current, and a switching frequency of a motor controller and a coolant temperature.

3. The method for estimating the coolant flow rate of a motor controller according to claim 1 or 2, It is characterized in that The neural network model adopts a multi-layer perceptron architecture, which includes: An input layer, the input layer is used to receive a plurality of data sets to be used for training and perform normalization processing on the plurality of data sets; A hidden layer, wherein the hidden layer is used to process and transform the normalized data set to extract feature information of the data set; and An output layer is used to output the prediction result of the neural network model according to the feature information provided by the hidden layer.

4. The method for estimating the coolant flow rate of a motor controller according to claim 3, It is characterized in that The hidden layer includes a plurality of nodes, and each node includes an activation function with a predetermined weight coefficient.

5. The method for estimating the coolant flow rate of a motor controller according to claim 4, It is characterized in that The neural network model uses a back propagation algorithm to optimize and iterate the weight coefficients of the activation function to determine the model parameters of the neural network model.

6. The method for estimating the coolant flow rate of a motor controller according to claim 1 or 2, It is characterized in that The motor controller includes a main drive inverter for an electric vehicle.

7. A busbar capacitance temperature estimation method for a motor controller, It is characterized in that The busbar capacitance temperature estimation method comprises the following steps: Acquire at least one operating parameter related to the current operating condition of the motor; and inputting the acquired at least one operating parameter into a neural network model having pre-trained model parameters to determine a bus capacitor temperature of the motor controller, The training process of the neural network model includes the following steps: - creating a plurality of data sets to be used for training, each data set comprising at least one operating parameter of the motor under a specified operating condition; - Input each created data set into a neural network model to obtain an estimated value of the bus capacitor temperature corresponding to each operating condition of the motor controller; - Using a thermometer to collect the actual value of the busbar capacitor temperature under each operating condition; and -Using the collected real value of the bus capacitor temperature to iterate and optimize the model parameters of the neural network model, so that the prediction result of the neural network model is close to the real value.

8. The bus capacitance temperature estimation method for a motor controller according to claim 7, It is characterized in that The at least one operating parameter is selected from a group including motor speed, DC bus voltage, motor modulation mode, power module temperature, motor D-axis current, motor Q-axis current, motor D-axis voltage, motor Q-axis voltage, and coolant flow, switching frequency and coolant temperature of the motor controller.

9. The bus capacitance temperature estimation method for a motor controller according to claim 7, It is characterized in that The at least one operating parameter comprises a coolant flow rate of the motor controller, and the coolant flow rate is determined using the coolant flow rate estimation method according to any one of claims 1 to 6.

10. A computer-readable storage medium having a computer program stored thereon, the computer program comprising program instructions, It is characterized in that When the program instructions are executed by a processor, the program instructions implement the coolant flow estimation method according to any one of claims 1 to 6 or the bus capacitance temperature estimation method according to any one of claims 7 to 9.

Images