Motor parameter identification method
The motor parameter identification method based on artificial intelligence and neural network models solves the problems of time-consuming, labor-intensive and inaccurate traditional methods, realizes efficient and accurate identification and optimization of motor parameters, and improves the performance and maintenance efficiency of electric motorcycles.
Patent Information
- Application Number
- CN202410351900.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-26
- Publication Date
- 2025-09-26
AI Technical Summary
Traditional motor parameter identification methods are time-consuming and labor-intensive, and it is difficult to guarantee parameter accuracy, which cannot meet the needs of optimizing electric motorcycle performance and extending its life.
A motor parameter identification method based on artificial intelligence and big data analysis is adopted. By collecting motor operation data, a neural network model is established, and data preprocessing and training are performed to monitor and optimize motor parameters in real time.
It achieves efficient, accurate and real-time identification of motor parameters, improves the performance and reliability of electric motorcycles, optimizes maintenance processes and reduces maintenance costs.
Smart Images
Figure CN120705567A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of parameter identification, and in particular relates to a method for identifying motor parameters. Background Art
[0002] With the rapid development of the electric motorcycle market, the performance requirements for motors are becoming increasingly demanding. Motor parameter identification plays a crucial role in optimizing motor performance, improving the efficiency of electric motorcycles, and extending their service life. Traditional motor parameter identification methods rely primarily on experimental testing and empirical estimation, which is time-consuming and labor-intensive, and struggles to guarantee parameter accuracy. Therefore, the development of a motor parameter identification method based on artificial intelligence and big data analysis has become an urgent need in the electric motorcycle industry.
[0003] In view of this, the present invention is proposed. Summary of the Invention
[0004] In order to solve the above technical problems, the basic concept of the technical solution adopted by the present invention is:
[0005] A motor parameter identification method includes the following steps: collecting data of an electric motorcycle motor during operation; preprocessing the data;
[0006] Step 2: Use artificial intelligence technology to establish a mapping relationship model between motor parameters and operating data;
[0007] Step 3: Train the model to accurately predict the motor parameters based on the motor’s operating data.
[0008] Step 4: During the actual operation of the electric motorcycle, the motor operating data is collected in real time and the trained model is used to predict and identify parameters.
[0009] Step 5: Optimize the motor parameters based on the identified parameters.
[0010] As a preferred embodiment of the present invention, the data collection in the first step includes but is not limited to collecting the voltage, current, speed and temperature data of the motor, and the data preprocessing includes but is not limited to data cleaning, denoising and normalization.
[0011] 9. As a preferred embodiment of the present invention, the mapping relationship model adopts a neural network model. The basic structure of the neural network model includes an input layer, a hidden layer and an output layer. The forward propagation process formula of the neural network is:
[0012] Linear transformation from input layer to hidden layer (z^{(1)}=W^{(1)}x+b^{(1)})
[0013] Where (z^{(1)}) is the input of the hidden layer (or called the activation value before), (W^{(1)}) is the weight matrix from the input layer to the hidden layer, (x) is the input data (i.e., the motor operation data), and (b^{(1)}) is the bias term of the hidden layer;
[0014] The activation function of the hidden layer is:
[0015] (a^{(1)}=f(z^{(1)}))
[0016] Where (a^{(1)}) is the output of the hidden layer (or called the activation value), (f) is the activation function;
[0017] Linear transformation from hidden layer to output layer (z^{(2)}=W^{(2)}a^{(1)}+b^{(2)})
[0018] Where (z^{(2)}) is the input to the output layer, (W^{(2)}) is the weight matrix from the hidden layer to the output layer, (a^{(1)}) is the output of the hidden layer, and (b^{(2)}) is the bias term of the output layer.
[0019] Activation function of the output layer (selected according to the task type):
[0020] For regression problems (predicting continuous motor parameters), a linear activation function is used:
[0021] (y=z^{(2)})
[0022] For classification problems (such as determining whether the motor state is normal), use the softmax activation function:
[0023] (y=softmax(z^{(2)})).
[0024] As a preferred embodiment of the present invention, the specific steps of establishing the neural network model include the following: determining the number of layers of the neural network and the number of neurons in each layer; randomly initializing the weight matrices (W^{(1)}) and (W^{(2)}) and bias terms (b^{(1)}) and (b^{(2)}) of the neural network; selecting an appropriate loss function according to the task type, wherein for regression problems, the commonly used loss function is the mean square error (MSE); for classification problems, the commonly used loss function is the cross entropy loss; and selecting an optimization algorithm to update the parameters of the neural network.
[0025] As a preferred embodiment of the present invention, the specific steps of establishing the neural network model also include the following: using preprocessed data to train the model, calculating the predicted value through forward propagation, and then calculating the error between the predicted value and the true value according to the loss function, calculating the gradient of the error relative to the parameter through the back propagation algorithm, and using the optimization algorithm to update the parameters to reduce the error; using the validation set to evaluate the performance of the model, and adjusting the network structure, parameter initialization, and learning rate hyperparameters as needed to optimize the performance of the model.
[0026] As a preferred embodiment of the present invention, the specific steps of establishing the neural network model also include the following: during the training process, monitoring the performance of the model on the validation set and preventing overfitting by using early stopping; once the model training is completed and its performance is verified, it is deployed to the actual running electric motorcycle for real-time prediction and identification of motor parameters.
[0027] As a preferred embodiment of the present invention, the optimization algorithm includes but is not limited to a gradient descent algorithm, a stochastic gradient descent algorithm and an Adam algorithm.
[0028] As a preferred embodiment of the present invention, the fourth step specifically includes: inputting the collected data into a trained model; the model calculates the predicted value of the motor parameter based on the input data; comparing the predicted value with a preset threshold to determine whether the motor parameter is normal; if the predicted value is abnormal, taking corresponding measures to adjust and optimize.
[0029] Compared with the prior art, the present invention has the following beneficial effects:
[0030] The motor parameter identification method of the present invention utilizes artificial intelligence technology, particularly a neural network model, to establish a mapping relationship between motor parameters and operating data. This method can predict and identify motor parameters based on real-time motor operating data, thereby optimizing and controlling the motor. The motor parameter identification method of the present invention is efficient, accurate, and real-time, enabling effective monitoring and optimized control of electric motorcycle motors, thereby improving the performance and reliability of electric motorcycles.
[0031] 2. The present invention collects and analyzes the operating data of the motor in real time. These parameters are important indicators reflecting the performance of the motor. By monitoring and optimizing these parameters, it can ensure that the motor operates in the best state, thereby improving its performance and efficiency.
[0032] 3. The present invention optimizes the maintenance process of electric motorcycles. Traditional motor maintenance methods usually rely on experience and practice, lacking scientificity and accuracy. The method provided by the present invention can provide more accurate maintenance suggestions through data analysis and model prediction, such as predicting motor life and discovering potential problems in advance. This helps to optimize the maintenance process of electric motorcycles, reduce maintenance costs, and improve maintenance efficiency.
[0033] 4. During actual operation, the present invention collects motor operating data in real time and uses a trained model to predict and identify parameters. The collected data is input into the model, which then calculates predicted values for the motor parameters based on the input data. The predicted values are compared with preset thresholds to determine whether the motor parameters are normal. If the predicted values are abnormal, appropriate adjustments and optimization measures are taken to achieve effective control and optimized operation of the motor.
[0034] 5. The present invention also uses artificial intelligence technology and neural network models to identify motor parameters, which represents the technological innovation direction of the electric motorcycle industry. Through continuous research and optimization of this method, it can promote technological progress and industrial upgrading in the electric motorcycle industry and inject new vitality into the development of the industry.
[0035] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] In the attached figure:
[0037] Figure 1 This is a flow chart of the motor parameter identification method of the present invention. DETAILED DESCRIPTION
[0038] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The following embodiments are used to illustrate the present invention.
[0039] A motor parameter identification method, such as Figure 1 As shown, the following steps are included:
[0040] Step 1: Collect data from the electric motorcycle motor during operation; pre-process the data;
[0041] Step 2: Use artificial intelligence technology to establish a mapping relationship model between motor parameters and operating data;
[0042] Step 3: Train the model to accurately predict the motor parameters based on the motor’s operating data.
[0043] Step 4: During the actual operation of the electric motorcycle, the motor operating data is collected in real time and the trained model is used to predict and identify parameters.
[0044] Step 5: Optimize the motor parameters based on the identified parameters.
[0045] The data collection in the first step includes but is not limited to collecting the voltage, current, speed and temperature data of the motor, and the data preprocessing includes but is not limited to data cleaning, denoising and normalization.
[0046] The mapping relationship model adopts a neural network model, and the basic structure of the neural network model includes an input layer, a hidden layer, and an output layer; wherein the forward propagation process formula of the neural network is:
[0047] Linear transformation from input layer to hidden layer (z^{(1)}=W^{(1)}x+b^{(1)})
[0048] Where (z^{(1)}) is the input of the hidden layer (or called the activation value before), (W^{(1)}) is the weight matrix from the input layer to the hidden layer, (x) is the input data (i.e., the motor operation data), and (b^{(1)}) is the bias term of the hidden layer;
[0049] The activation function of the hidden layer is:
[0050] (a^{(1)}=f(z^{(1)}))
[0051] Where (a^{(1)}) is the output of the hidden layer (or called the activation value), (f) is the activation function;
[0052] Linear transformation from hidden layer to output layer (z^{(2)}=W^{(2)}a^{(1)}+b^{(2)})
[0053] Where (z^{(2)}) is the input to the output layer, (W^{(2)}) is the weight matrix from the hidden layer to the output layer, (a^{(1)}) is the output of the hidden layer, and (b^{(2)}) is the bias term of the output layer.
[0054] Activation function of the output layer (selected according to the task type):
[0055] For regression problems (predicting continuous motor parameters), a linear activation function is used:
[0056] (y=z^{(2)})
[0057] For classification problems (such as determining whether the motor state is normal), use the softmax activation function:
[0058] (y=softmax(z^{(2)}))
[0059] The specific steps of establishing the neural network model include the following: determining the number of layers of the neural network and the number of neurons in each layer; randomly initializing the weight matrices (W^{(1)}) and (W^{(2)}) and bias terms (b^{(1)}) and (b^{(2)}) of the neural network; selecting an appropriate loss function according to the task type, where for regression problems, the commonly used loss function is the mean square error (MSE); for classification problems, the commonly used loss function is the cross entropy loss; and selecting an optimization algorithm to update the parameters of the neural network.
[0060] Use preprocessed data for model training, calculate the predicted value through forward propagation, then calculate the error between the predicted value and the true value based on the loss function, calculate the gradient of the error relative to the parameter through the backpropagation algorithm, and use the optimization algorithm to update the parameters to reduce the error; use the validation set to evaluate the performance of the model, and adjust the network structure, parameter initialization, and learning rate hyperparameters as needed to optimize the performance of the model.
[0061] During the training process, the performance of the model on the validation set is monitored and overfitting is prevented by using early stopping. Once the model training is completed and its performance is verified, it is deployed to the actual running electric motorcycle for real-time prediction and identification of motor parameters.
[0062] The optimization algorithms include but are not limited to gradient descent algorithm, stochastic gradient descent algorithm and Adam algorithm.
[0063] The fourth step specifically includes: inputting the collected data into the trained model; the model calculates the predicted value of the motor parameter based on the input data; comparing the predicted value with the preset threshold to determine whether the motor parameter is normal; if the predicted value is abnormal, taking corresponding measures to adjust and optimize.
[0064] Specifically, regarding the threshold comparison, after obtaining the predicted motor parameters, it is necessary to compare them with the preset threshold to determine whether the motor parameters are normal. The following is the comparison code for the threshold input:
[0065]
[0066] If the predicted value exceeds the threshold, a warning message is output; otherwise, information indicating that the parameter is within the normal range is output.
[0067] First, the present invention collects data from the electric motorcycle motor during operation, including voltage, current, speed and temperature, and performs data preprocessing, such as cleaning, denoising and normalization. These preprocessing steps help improve the quality and accuracy of the data and provide a reliable foundation for subsequent model establishment.
[0068] Next, the present invention uses artificial intelligence technology to establish a mapping relationship model between motor parameters and operating data. This invention employs a neural network model whose basic structure includes an input layer, a hidden layer, and an output layer. Through a forward propagation process, the input data undergoes a linear transformation and an activation function to obtain a predicted value for the output layer. Simultaneously, an appropriate loss function and optimization algorithm are selected based on the task type to calculate the error between the predicted value and the true value. Backpropagation is then used to update the neural network parameters to minimize this error.
[0069] During the model building process, the present invention also requires model training, verification, and tuning. Model training is performed using preprocessed data, with the predicted value calculated through forward propagation. The error is then calculated based on the loss function, the gradient is calculated through a backpropagation algorithm, and the parameters are updated using an optimization algorithm. Simultaneously, the performance of the model is evaluated using a validation set, and the network structure, parameter initialization, and learning rate parameters are adjusted as needed to optimize the model's performance. During the training process, the performance of the model on the validation set needs to be monitored to prevent overfitting, and techniques such as early stopping are used to improve the model's generalization ability.
[0070] Finally, during actual operation, the present invention collects motor operating data in real time and uses a trained model to predict and identify motor parameters. The collected data is input into the model, which then calculates predicted motor parameter values based on the input data. The predicted values are then compared with preset thresholds to determine whether the motor parameters are normal. If the predicted values are abnormal, appropriate adjustments and optimization measures are taken to achieve effective control and optimized operation of the motor.
[0071] The motor parameter identification method of the present invention utilizes artificial intelligence technology, particularly a neural network model, to establish a mapping relationship between motor parameters and operating data. In this way, the motor's parameters can be predicted and identified based on its real-time operating data, thereby achieving optimization and control of the motor.
[0072] The present invention collects and analyzes the operating data of the motor in real time. These parameters are important indicators reflecting the performance of the motor. By monitoring and optimizing these parameters, it can ensure that the motor operates in the best state, thereby improving its performance and efficiency.
[0073] The present invention optimizes the maintenance process of electric motorcycles. Traditional motor maintenance methods usually rely on experience and practice, lacking scientificity and accuracy. The method provided by the present invention can provide more accurate maintenance suggestions through data analysis and model prediction, such as predicting motor life and discovering potential problems in advance. This helps to optimize the maintenance process of electric motorcycles, reduce maintenance costs, and improve maintenance efficiency.
[0074] During actual operation, the present invention collects motor operating data in real time and uses a trained model to predict and identify parameters. The collected data is input into the model, and the model calculates predicted values of the motor parameters based on the input data. The predicted values are compared with preset thresholds to determine whether the motor parameters are normal. If the predicted values are abnormal, appropriate measures are taken to adjust and optimize them to achieve effective control and optimized operation of the motor. The motor parameter identification method of the present invention is efficient, accurate, and real-time, and can achieve effective monitoring and optimized control of the electric motorcycle motor, thereby improving the performance and reliability of the electric motorcycle.
[0075] The present invention also uses artificial intelligence technology and neural network models to identify motor parameters, representing the technological innovation direction of the electric motorcycle industry. Through continuous research and optimization of this method, it can promote technological progress and industrial upgrading in the electric motorcycle industry and inject new vitality into the development of the industry.
[0076] It will be understood that the present invention is described by way of some embodiments, and it will be appreciated by those skilled in the art that various changes or equivalent substitutions may be made to these features and embodiments without departing from the spirit and scope of the present invention. In addition, under the teachings of the present invention, these features and embodiments may be modified to adapt to specific circumstances and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are intended to be protected by the present invention.
Claims
1. A method for identifying motor parameters, characterized in that: The following steps are included: Step 1: Collect data from the electric motorcycle motor during operation; pre-process the data; Step 2: Use artificial intelligence technology to establish a mapping relationship model between motor parameters and operating data; Step 3: Train the model to accurately predict the motor parameters based on the motor’s operating data. Step 4: During the actual operation of the electric motorcycle, the motor operating data is collected in real time and the trained model is used to predict and identify parameters. Step 5: Optimize the motor parameters based on the identified parameters.
2. The motor parameter identification method according to claim 1, characterized in that: The data collection in the first step includes but is not limited to collecting the voltage, current, speed and temperature data of the motor, and the data preprocessing includes but is not limited to data cleaning, denoising and normalization.
3. The motor parameter identification method according to claim 1, characterized in that: The mapping relationship model adopts a neural network model, and the basic structure of the neural network model includes an input layer, a hidden layer, and an output layer; wherein the forward propagation process formula of the neural network is: Linear transformation from input layer to hidden layer (z^{(1)}=W^{(1)}x+b^{(1)}) Where (z^{(1)}) is the input of the hidden layer (or called the activation value before), (W^{(1)}) is the weight matrix from the input layer to the hidden layer, (x) is the input data (i.e., the motor operation data), and (b^{(1)}) is the bias term of the hidden layer; The activation function of the hidden layer is: (a^{(1)}=f(z^{(1)})) Where (a^{(1)}) is the output of the hidden layer (or called the activation value), (f) is the activation function; Linear transformation from hidden layer to output layer (z^{(2)}=W^{(2)}a^{(1)}+b^{(2)}) Where (z^{(2)}) is the input to the output layer, (W^{(2)}) is the weight matrix from the hidden layer to the output layer, (a^{(1)}) is the output of the hidden layer, and (b^{(2)}) is the bias term of the output layer. Activation function of the output layer (selected according to the task type): For regression problems (predicting continuous motor parameters), a linear activation function is used: (y=z^{(2)}) For classification problems (such as determining whether the motor state is normal), use the softmax activation function: (y=softmax(z^{(2)})).
4. The motor parameter identification method according to claim 3, characterized in that: The specific steps of establishing the neural network model include the following: determining the number of layers of the neural network and the number of neurons in each layer; randomly initializing the weight matrices (W^{(1)}) and (W^{(2)}) and bias terms (b^{(1)}) and (b^{(2)}) of the neural network; selecting an appropriate loss function according to the task type, where for regression problems, the commonly used loss function is the mean square error (MSE); for classification problems, the commonly used loss function is the cross entropy loss; and selecting an optimization algorithm to update the parameters of the neural network.
5. The motor parameter identification method according to claim 3, characterized in that: The specific steps of establishing the neural network model also include the following: using preprocessed data to train the model, calculating the predicted value through forward propagation, then calculating the error between the predicted value and the true value based on the loss function, calculating the gradient of the error relative to the parameter through the back propagation algorithm, and using the optimization algorithm to update the parameters to reduce the error; using the validation set to evaluate the performance of the model, and adjusting the network structure and parameter initialization parameters as needed to optimize the performance of the model.
6. The motor parameter identification method according to claim 5, characterized in that: The specific steps of establishing the neural network model also include the following: during the training process, monitoring the performance of the model on the validation set and preventing overfitting by using early stopping; once the model training is completed and its performance is verified, it is deployed to the actual running electric motorcycle for real-time prediction and identification of motor parameters.
7. The motor parameter identification method according to claim 4, characterized in that: The optimization algorithms include but are not limited to gradient descent algorithm, stochastic gradient descent algorithm and Adam algorithm.
8. The motor parameter identification method according to claim 1, characterized in that: The fourth step specifically includes: inputting the collected data into the trained model; the model calculates the predicted value of the motor parameter based on the input data; comparing the predicted value with the preset threshold to determine whether the motor parameter is normal; if the predicted value is abnormal, taking corresponding measures to adjust and optimize.