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

Abstract

The application relates to a permanent magnet synchronous motor temperature prediction method and system based on temperature time sequence input, and the method comprises the following steps: acquiring parameter sequences in the operation process of a permanent magnet synchronous motor through a data acquisition system, and extracting normal mode input; selecting a moment in the operation process of the permanent magnet synchronous motor as an initial moment based on the normal mode input, adding an initial temperature sequence and a time difference sequence; acquiring a cooling water inlet and outlet temperature difference in the operation process of the permanent magnet synchronous motor as an augmented input of the model; performing feature selection according to the normal mode input, the initial temperature sequence, the time difference sequence and the augmented input, constructing a model input feature set for permanent magnet synchronous motor temperature prediction, loading the model input feature set into a permanent magnet synchronous motor temperature prediction model, and acquiring the internal temperature of the permanent magnet synchronous motor. Compared with the prior art, the application expands the use range of various machine learning algorithms in the field of permanent magnet synchronous motor temperature prediction, and improves the prediction accuracy and calculation speed of the model.

Description

Technical Field

[0001] This invention relates to the field of motor temperature prediction technology, and in particular to a method and system for predicting the temperature of a permanent magnet synchronous motor based on temperature timing input. Background Technology

[0002] Estimating the internal temperature field of a permanent magnet synchronous motor (PMSM) using predictive methods has significant research value and promising application prospects. It allows for the prediction of the internal temperature field of the PMSM using structural data or operating parameters without affecting motor operation, thus extending the motor's lifespan and increasing the reliability of the electric drive system. Since machine learning methods are dataset-driven, the determination of the model input is crucial to the prediction accuracy and computational speed of PMSM temperature prediction models based on machine learning algorithms.

[0003] Currently, determining the input set for permanent magnet synchronous motor temperature prediction models relies on mechanistic analysis and human experience. Conventional model inputs are obtained by analyzing the coupling relationship between the thermodynamic field and electric and magnetic fields, including electromagnetic physical quantities (such as voltage and current), cooling conditions (such as cooling water temperature and flow rate), and operating conditions (such as torque and speed). However, conventional model inputs do not include time-series temperature information, and are mostly only applicable to machine learning algorithms with time-series prediction capabilities. This also imposes limitations on machine learning algorithms used for temperature prediction modeling.

[0004] For example, the invention with publication number CN114117895A discloses a method for real-time prediction of rotor temperature of permanent magnet synchronous motor. The mathematical model based on the motion equation and current equation in the synchronous rotating coordinate system with rotor flux orientation selects eight variables as model inputs: d-axis voltage, q-axis voltage, d-axis current, q-axis current, mechanical angular velocity, load torque, ambient temperature of motor operation and coolant temperature, and establishes an LSTM-CNN network prediction model to predict the real-time rotor temperature of permanent magnet synchronous motor.

[0005] The invention disclosed in CN112395815A discloses a method for predicting the temperature of a permanent magnet synchronous motor. It constructs a PSNLSTMs model for predicting the temperature of a permanent magnet synchronous motor. The model input includes ambient temperature, coolant temperature, motor speed, motor torque, d-axis voltage, q-axis voltage, d-axis current, and q-axis current at continuous time intervals, for a total of 8 dimensions of model input.

[0006] The invention disclosed in CN112183835A discloses a method, device and system for early warning of water guide tile temperature trend based on machine learning. It determines the input factors affecting the temperature of water guide tile through multiple experiments and prior experience, and then establishes a water guide tile temperature prediction model based on neural network. The model inputs include: unit output, outlet oil temperature 1, outlet oil temperature 2, swing 1, and swing 2.

[0007] The invention disclosed in CN114444382A presents a fault diagnosis and analysis method for wind turbine gearboxes based on machine learning algorithms. It obtains 16 operational parameters from the wind farm as model inputs to predict temperature, including wind speed, wind power, gearbox oil temperature, wind direction, yaw angle, pitch angle, voltage, current, instantaneous generator speed, grid active power, gearbox intermediate shaft drive end bearing temperature, gearbox intermediate shaft non-drive end bearing temperature, motor-side gearbox high-speed shaft bearing temperature, gearbox lubricating oil sump temperature, gearbox lubricating oil inlet temperature, and gearbox oil pressure before the filter screen.

[0008] The above scheme determines the input set of the motor temperature prediction model based on physical mechanisms and prior experience. The model input includes variables related to electric field, magnetic field, thermal field, and operating conditions. However, the input set itself does not contain temperature time series data. Therefore, it is only suitable for machine learning algorithms with time-series prediction capabilities. For other non-time-series prediction machine learning algorithms, the above input set cannot compensate for the lack of time-series prediction capabilities. Furthermore, the above scheme directly uses motor operating parameters without input augmentation and feature selection, and may not necessarily obtain the optimal or even good model input set. Summary of the Invention

[0009] The purpose of this invention is to overcome the shortcomings of the prior art by providing a method and system for predicting the temperature of a permanent magnet synchronous motor based on temperature timing input, thus making up for the deficiencies of machine learning algorithms that do not have timing prediction capabilities.

[0010] The objective of this invention can be achieved through the following technical solutions:

[0011] A method for predicting the temperature of a permanent magnet synchronous motor based on temperature timing input includes the following steps:

[0012] The parameter sequence during the operation of the permanent magnet synchronous motor is obtained through a data acquisition system, and the input of the conventional model is extracted.

[0013] Based on the conventional model input, a certain moment in the operation of the permanent magnet synchronous motor is selected as the initial moment, and a temperature sequence and a time difference sequence are added. The temperature sequence is the motor temperature at the initial moment, and the time difference sequence is the timing difference between the initial moment and the subsequent operation time of the permanent magnet synchronous motor.

[0014] The temperature difference between the inlet and outlet of the cooling water during the operation of the permanent magnet synchronous motor is obtained and used as the augmented input of the model;

[0015] Based on the conventional model input, temperature sequence, time difference sequence, and augmented input, feature selection is performed to construct a model input feature set for temperature prediction of permanent magnet synchronous motors;

[0016] The model input feature set is loaded into a pre-established permanent magnet synchronous motor temperature prediction model to obtain the internal temperature of the permanent magnet synchronous motor.

[0017] Furthermore, the model input feature set includes time difference, initial temperature, cooling variables, electromagnetic variables, and operating conditions. The cooling variables include ambient temperature, cooling water inlet temperature, cooling water outlet temperature, and cooling water flow rate.

[0018] Furthermore, the electromagnetic variables include bus voltage, bus current, three-phase voltage, and three-phase current, and the operating conditions include speed and torque.

[0019] Furthermore, the feature selection is performed by combining recursive feature elimination and principal component analysis.

[0020] Furthermore, the temperature prediction model for the permanent magnet synchronous motor is a machine learning model, which is a Gaussian process regression model.

[0021] The present invention also provides a temperature prediction system for permanent magnet synchronous motors based on temperature timing input, comprising:

[0022] The conventional model input acquisition module is used to acquire the parameter sequence during the operation of the permanent magnet synchronous motor through the data acquisition system and extract the conventional model input.

[0023] The timing addition module is used to select a certain moment in the operation of the permanent magnet synchronous motor as the initial moment based on the conventional model input, and add a temperature sequence and a time difference sequence. The temperature sequence is the motor temperature at the initial moment, and the time difference sequence is the timing difference between the permanent magnet synchronous motor and the initial moment in subsequent operation moments.

[0024] The input augmentation module is used to obtain the temperature difference between the inlet and outlet of the cooling water during the operation of the permanent magnet synchronous motor, which is used as the augmented input of the model.

[0025] The feature selection module is used to perform feature selection based on the conventional model input, temperature sequence, time difference sequence and augmented input to construct a model input feature set for temperature prediction of permanent magnet synchronous motor;

[0026] The temperature prediction module is used to load the model input feature set into a pre-established permanent magnet synchronous motor temperature prediction model to obtain the internal temperature of the permanent magnet synchronous motor.

[0027] Furthermore, the model input feature set includes time difference, initial temperature, cooling variables, electromagnetic variables, and operating conditions. The cooling variables include ambient temperature, cooling water inlet temperature, cooling water outlet temperature, and cooling water flow rate.

[0028] Furthermore, the electromagnetic variables include bus voltage, bus current, three-phase voltage, and three-phase current, and the operating conditions include speed and torque.

[0029] Furthermore, the feature selection is performed by combining recursive feature elimination and principal component analysis.

[0030] Furthermore, the temperature prediction model for the permanent magnet synchronous motor is a machine learning model, which is a Gaussian process regression model.

[0031] Compared with the prior art, the present invention has the following advantages:

[0032] (1) Based on the operating mechanism of motors and machine learning knowledge, this invention obtains the input set of a permanent magnet synchronous motor temperature prediction model containing temperature time series through feature engineering. By obtaining these 15 parameter values ​​at the motor operating time, the internal temperature field information of the motor can be predicted quickly and accurately. The model input set is obtained through input augmentation and feature selection. On the one hand, it improves the prediction accuracy and calculation speed of the model and makes full use of appropriate data information. On the other hand, it also makes up for the deficiency of some machine learning algorithms that cannot perform time series prediction to a certain extent, so that more machine learning algorithms can be used for the prediction of the internal temperature field of permanent magnet synchronous motors.

[0033] (2) This invention adds an initial temperature sequence and a time difference sequence to construct a model input set containing temperature time series information, and adds temperature and time series attributes to the data samples in the training set and the test set, thus broadening the application scope of machine learning algorithms in the field of temperature prediction of permanent magnet synchronous motors.

[0034] (3) This invention combines different feature selection methods to construct the optimal model input set, which simplifies the number of model inputs, extracts data information that is beneficial to the model prediction accuracy and calculation speed, and improves the model performance.

[0035] (4) Based on the input set of the permanent magnet synchronous motor temperature prediction model with temperature time sequence designed according to the present invention, machine learning-based modeling and experimental verification were carried out, demonstrating its superiority in improving model performance.

[0036] The specific application process is as follows: Using motor operation data obtained in the laboratory, and after selecting a conventional model input from sensors based on human experience and physical knowledge, temperature and time series data were added, and the model input was augmented. Finally, based on the above model input, feature selection was performed using recursive feature elimination and principal component analysis methods. Ultimately, an input set for a permanent magnet synchronous motor temperature prediction model containing temperature and time series data was established. Comparison of prediction results shows that the model prediction accuracy was significantly improved after adding the initial temperature series and time difference series, and the model prediction performance and computation speed were also significantly improved after feature selection. Attached Figure Description

[0037] Figure 1 This is a flowchart illustrating a method for predicting the temperature of a permanent magnet synchronous motor based on temperature timing input, provided in an embodiment of the present invention.

[0038] Figure 2 This is a flowchart illustrating the construction process of a model input set provided in an embodiment of the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0040] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0041] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0042] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship in which the product of this invention is usually placed during use. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0043] It should be noted that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0044] Furthermore, terms such as "horizontal" and "vertical" do not imply that components must be absolutely horizontal or suspended, but rather that they can be slightly tilted. For example, "horizontal" simply means that its direction is more horizontal than "vertical," and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0045] Example 1

[0046] like Figure 1 As shown, this embodiment provides a method for predicting the temperature of a permanent magnet synchronous motor based on temperature time-series input. The determination of the model input set combines knowledge from the field of computer science with the physical mechanism of permanent magnet synchronous motors. Based on the raw data directly collected by sensors during motor operation, feature extraction and augmentation are performed, and finally, the final model input set is determined through feature selection, thereby performing temperature prediction.

[0047] Specifically, the following steps are included:

[0048] S1: Obtain the parameter sequence during the operation of the permanent magnet synchronous motor through the data acquisition system, and extract the input of the conventional model;

[0049] S2: Based on the input of the conventional model, select a certain moment in the operation of the permanent magnet synchronous motor as the initial moment, and add a temperature sequence and a time difference sequence. The temperature sequence is the motor temperature at the initial moment, and the time difference sequence is the timing difference between the permanent magnet synchronous motor and the initial moment in subsequent operation moments.

[0050] S3: Obtain the temperature difference between the inlet and outlet of the cooling water during the operation of the permanent magnet synchronous motor, and use it as an augmented input for the model;

[0051] S4: Based on the conventional model input, temperature sequence, time difference sequence, and augmented input, feature selection is performed by combining recursive feature elimination and principal component analysis to construct a model input feature set for temperature prediction of permanent magnet synchronous motors;

[0052] S5: Load the model input feature set into the pre-established permanent magnet synchronous motor temperature prediction model to obtain the internal temperature of the permanent magnet synchronous motor.

[0053] The model input feature set includes time difference, initial temperature, cooling variables, electromagnetic variables, and operating conditions. Cooling variables include ambient temperature, cooling water inlet temperature, cooling water outlet temperature, and cooling water flow rate. Electromagnetic variables include bus voltage, bus current, three-phase voltage, and three-phase current. Operating conditions include speed and torque. The model output is the internal temperature of the permanent magnet synchronous motor.

[0054] The above scheme will be described in detail below.

[0055] 1. Conventional input obtained directly from sensors

[0056] Based on literature review results and knowledge of motor physical structure, and considering the influence of factors such as cooling conditions and operating conditions, the raw laboratory data of parameters collected by sensors were initially selected as the model input, including ambient temperature, cooling water inlet temperature, cooling water outlet temperature, cooling water flow rate, bus voltage, bus current, three-phase voltage (3 voltage values), three-phase current (3 current values), speed, and torque.

[0057] 2. Add temperature and time information

[0058] Since some machine learning algorithms lack time-series attributes in their models, and the internal temperature of a permanent magnet synchronous motor is a series of variables that change continuously over time, this study supplements the model with time-series attributes for the data samples. Taking any point in time during motor operation as the initial moment, the time difference between the initial moment and subsequent moments of motor operation is included in the model input set. Simultaneously, the motor temperature at the initial moment is also used as model input, adding a time-series prediction attribute to the machine learning modeling process.

[0059] 3. Add new model inputs

[0060] Since constructing new model inputs by combining existing model inputs is considered one of the effective ways to construct the optimal feature set, model input augmentation is performed based on the above model inputs.

[0061] Since the internal temperature of the motor is greatly affected by the motor cooling conditions, the cooling factors during motor operation are supplemented by using the temperature difference between the inlet and outlet of the cooling water as an augmented input to the model.

[0062] 4. Combine feature selection methods to obtain the final model input.

[0063] The model input set obtained above includes conventional inputs directly collected by sensors, temperature and time series information, and augmented model inputs. To avoid the reduction in model accuracy due to overfitting, two feature selection methods, recursive feature elimination and principal component analysis, are combined.

[0064] The final model input set includes: ambient temperature, cooling water inlet temperature, cooling water outlet temperature, cooling water flow rate, bus voltage, bus current, three-phase voltage (3 voltage values), three-phase current (3 current values), rotational speed, time difference, and initial temperature. Among these, the time difference and initial temperature are used to provide temperature timing information.

[0065] During motor operation, these parameters can be obtained directly or indirectly, and machine learning models, such as Gaussian process regression models, can be established to quickly and accurately obtain the predicted internal temperature of the permanent magnet synchronous motor.

[0066] like Figure 2 As shown below, a specific implementation process of this solution is provided:

[0067] 1. First, the parameter sequence of the permanent magnet synchronous motor during operation is obtained using a data acquisition system.

[0068] 2. Next, extract the conventional model inputs directly collected by the sensors, including: ambient temperature, cooling water inlet temperature, cooling water outlet temperature, cooling water flow rate, bus voltage, bus current, three-phase voltage (3 voltage values), three-phase current (3 current values), speed, and torque.

[0069] 3. Next, based on the input of the conventional model, a certain moment in the operation of the permanent magnet synchronous motor is selected as the initial moment. Temperature sequence and time difference sequence are added to add time-series prediction attributes to the machine learning modeling process. Among them, the temperature sequence is the motor temperature at the initial moment obtained by direct or indirect means, and the time difference sequence is the timing difference between the initial moment and the subsequent operating moments of the motor.

[0070] 4. Then, based on the existing model input, the cooling factors during motor operation are supplemented by using the temperature difference between the inlet and outlet of the cooling water as the augmented input of the model.

[0071] 5. Finally, based on the above model input, feature selection is performed on the model input set by combining recursive feature elimination and principal component analysis, and a corresponding model is established. The final model input feature set for temperature prediction of permanent magnet synchronous motor includes time difference, initial temperature, cooling water temperature difference, ambient temperature, cooling water inlet temperature, cooling water outlet temperature, cooling water flow rate, bus voltage, bus current, three-phase voltage (3 voltage values), three-phase current (3 current values), speed, torque, etc.

[0072] The above is an introduction to the method embodiments. The following system embodiments will further illustrate the present invention.

[0073] This embodiment also provides a temperature prediction system for permanent magnet synchronous motors based on temperature timing input, including:

[0074] The conventional model input acquisition module is used to acquire the parameter sequence during the operation of the permanent magnet synchronous motor through the data acquisition system and extract the conventional model input.

[0075] The timing addition module is used to select a certain moment in the operation of the permanent magnet synchronous motor as the initial moment based on the input of the conventional model, and add a temperature sequence and a time difference sequence. The temperature sequence is the motor temperature at the initial moment, and the time difference sequence is the timing difference between the initial moment and the subsequent operation time of the permanent magnet synchronous motor.

[0076] The input augmentation module is used to obtain the temperature difference between the inlet and outlet of the cooling water during the operation of the permanent magnet synchronous motor, which is used as the augmented input of the model.

[0077] The feature selection module is used to select features based on the conventional model input, temperature sequence, time difference sequence, and augmented input to construct a model input feature set for temperature prediction of permanent magnet synchronous motors.

[0078] The temperature prediction module is used to load the model input feature set into a pre-established permanent magnet synchronous motor temperature prediction model to obtain the internal temperature of the permanent magnet synchronous motor.

[0079] It should be noted that the specific details and beneficial effects of the device in this application can be found in the above-described method embodiments, and will not be repeated here.

[0080] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A temperature prediction method for a permanent magnet synchronous motor based on temperature time series input, characterized in that, The method comprises the following steps: Obtain parameter sequences in the operation process of the permanent magnet synchronous motor through a data acquisition system, and extract a normal model input, which comprises an ambient temperature, a cooling water inlet temperature, a cooling water outlet temperature, a cooling water flow, a bus voltage, a bus current, a three-phase voltage, a three-phase current, a rotating speed and a torque; Select a time in the operation process of the permanent magnet synchronous motor as an initial time based on the normal model input, add a temperature sequence and a time difference sequence, the temperature sequence is a motor temperature at the initial time, and the time difference sequence is a time difference value between a subsequent operation time of the permanent magnet synchronous motor and the initial time; Obtain a cooling water outlet-inlet temperature difference in the operation process of the permanent magnet synchronous motor as an augmented input of the model; Perform feature selection according to the normal model input, the temperature sequence, the time difference sequence and the augmented input, and construct a model input feature set for temperature prediction of the permanent magnet synchronous motor; Load the model input feature set into a pre-established temperature prediction model of the permanent magnet synchronous motor, and obtain an internal temperature of the permanent magnet synchronous motor.

2. The temperature prediction method of a permanent magnet synchronous motor based on temperature time-series input according to claim 1, characterized in that, The model input feature set comprises a time difference, an initial time temperature, a cooling variable, an electromagnetism variable and an operation condition, and the cooling variable comprises an ambient temperature, a cooling water inlet temperature, a cooling water outlet temperature and a cooling water flow.

3. The temperature prediction method of a permanent magnet synchronous motor based on temperature time-series input according to claim 2, characterized in that, The electromagnetism variable comprises a bus voltage, a bus current, a three-phase voltage and a three-phase current, and the operation condition comprises a rotating speed and a torque.

4. The temperature prediction method of a permanent magnet synchronous motor based on temperature time-series input according to claim 1, characterized in that, The feature selection is performed in combination with recursive feature elimination and principal component analysis.

5. The temperature prediction method of a permanent magnet synchronous motor based on temperature time-series input according to claim 1, characterized in that, The temperature prediction model of the permanent magnet synchronous motor is a machine learning model, and the machine learning model is a Gaussian process regression model.

6. A temperature prediction system for a permanent magnet synchronous motor based on temperature time series input, characterized by, The method comprises the following steps: A normal model input acquisition module is configured to obtain parameter sequences in the operation process of the permanent magnet synchronous motor through a data acquisition system, and extract a normal model input, which comprises an ambient temperature, a cooling water inlet temperature, a cooling water outlet temperature, a cooling water flow, a bus voltage, a bus current, a three-phase voltage, a three-phase current, a rotating speed and a torque; A time sequence adding module is configured to select a time in the operation process of the permanent magnet synchronous motor as an initial time based on the normal model input, and add a temperature sequence and a time difference sequence, the temperature sequence is a motor temperature at the initial time, and the time difference sequence is a time difference value between a subsequent operation time of the permanent magnet synchronous motor and the initial time; An input augmenting module is configured to obtain a cooling water outlet-inlet temperature difference in the operation process of the permanent magnet synchronous motor as an augmented input of the model; A feature selection module is configured to perform feature selection according to the normal model input, the temperature sequence, the time difference sequence and the augmented input, and construct a model input feature set for temperature prediction of the permanent magnet synchronous motor; A temperature prediction module is configured to load the model input feature set into a pre-established temperature prediction model of the permanent magnet synchronous motor, and obtain an internal temperature of the permanent magnet synchronous motor.

7. The temperature prediction system for a permanent magnet synchronous motor based on temperature time-series input according to claim 6, characterized in that, The model input feature set comprises a time difference, an initial time temperature, a cooling variable, an electromagnetism variable and an operation condition, and the cooling variable comprises an ambient temperature, a cooling water inlet temperature, a cooling water outlet temperature and a cooling water flow.

8. The temperature prediction system for a permanent magnet synchronous motor based on temperature time-series input according to claim 7, characterized in that, The electromagnetic variables include bus voltage, bus current, three-phase voltage and three-phase current, and the operating conditions include rotating speed and torque.

9. The temperature prediction system for PMSM based on temperature time series input according to claim 6, wherein, The feature selection is combined with recursive feature elimination and principal component analysis.

10. The temperature prediction system for a permanent magnet synchronous motor based on temperature time-series input according to claim 6, characterized in that, The permanent magnet synchronous motor temperature prediction model is a machine learning model, and the machine learning model is a Gaussian process regression model.

Images