Permanent magnet synchronous motor temperature estimation method based on model data dual drive

Abstract

The invention discloses a permanent magnet synchronous motor temperature estimation method based on model data dual driving. A model driving module, a data driving module and a state estimation module are included. The model driving module models the thermal behavior of the motor so as to determine a thermal model of the motor; the data driving module identifies online parameters of the motor by using real-time data of a motor sensor to obtain real-time parameters of a motor thermal model; a self-adaptive state transition equation of the state estimation module is driven by a motor thermal model and real-time parameters of the motor thermal model at the same time, and meanwhile a state estimator achieves motor temperature estimation according to the self-adaptive state transition equation. The permanent magnet synchronous motor temperature estimation method has the beneficial effects that the permanent magnet synchronous motor temperature estimation method with high reliability is provided, the permanent magnet temperature and the stator winding temperature are accurately monitored, and the control performance and the safety of a permanent magnet synchronous motor system are improved.

Description

Technical Field

[0001] The present invention relates to temperature estimation, and in particular to a method for estimating the temperature of a permanent magnet and a stator winding of a permanent magnet synchronous motor. Background Art

[0002] In the past three decades, permanent magnet synchronous motors have been increasingly widely used due to their excellent characteristics such as high power, high torque density and high efficiency. However, the temperature of the permanent magnet has a significant impact on its performance. The increase in temperature weakens the magnetic field strength of the permanent magnet, thereby reducing the torque density of the motor. When the temperature exceeds a certain limit, this decline may become irreversible. Similarly, the temperature of the stator winding is critical to the safe operation of the permanent magnet synchronous motor. Excessive winding temperature increases resistance, affects control performance, and may even cause the winding to fuse, seriously threatening production safety. In order to solve the challenges caused by the increase in temperature during the operation of permanent magnet synchronous motors, it is crucial to accurately monitor the temperature of the permanent magnets and stator windings. However, traditional temperature measurement methods usually rely on wireless temperature sensors installed on the permanent magnets and stator windings. These devices are not only costly, but also reduce the robustness of the system due to direct contact with motor components.

[0003] In order to deal with a series of problems existing in the traditional permanent magnet synchronous motor temperature monitoring method and avoid the use of complex devices such as wireless temperature sensors, the industry has proposed some methods for estimating the permanent magnet temperature and stator winding temperature. The literature [AMEL-Refai, NC Harris, TM Jahns, and KM Rahman, "Thermal analysis of multibarrier interior PM synchronous machine using lumped parameter model," IEEE Trans. Energy Convers., vol. 14, no. 2, pp. 303–309, Jun. 2004.] proposed a lumped parameter thermal model of a multi-barrier interior permanent magnet synchronous motor. The model consists of 12 thermal resistance networks and can estimate the temperature of 9 critical points inside the machine, including the permanent magnet temperature and the stator winding temperature.

[0004] The above literature provides valuable research ideas for the study of permanent magnet synchronous motor temperature estimation. However, the above literature is only based on thermal models. When the thermal model parameters change due to aging or other factors, the estimated temperature becomes unreliable. Therefore, a high-reliability permanent magnet synchronous motor temperature estimation method is proposed, which is a key technical issue in the application of intelligent algorithms to permanent magnet synchronous motor temperature estimation. Summary of the invention

[0005] The technical problem to be solved by the present invention is that there is a lack of a high-reliability permanent magnet synchronous motor temperature estimation method in the field of permanent magnet synchronous motor temperature estimation, and a permanent magnet synchronous motor temperature estimation method based on data model dual drive is proposed.

[0006] In order to achieve the above objective, according to one aspect of a specific implementation of the present invention, a model driven module, a data driven module and a state estimation module are provided.

[0007] The model-driven module models the thermal behavior of the motor to determine the motor thermal model; the data-driven module uses the real-time data of the motor sensor to identify the online parameters of the motor to obtain the real-time parameters of the motor thermal model; the adaptive state transfer equation of the state estimation module is driven by the motor thermal model and its real-time parameters at the same time, and the state estimator realizes the motor temperature estimation according to the adaptive state transfer equation.

[0008] Specifically, the model driving module models the thermal behavior of the motor and obtains the motor thermal model. The motor thermal model includes a lumped parameter thermal network model and a motor thermal field simulation model.

[0009] Specifically, the data-driven module uses motor sensor data to perform online parameter identification on the motor thermal model through data-driven methods, including data fitting method, probability statistics method and deep learning method.

[0010] Specifically, the adaptive state transfer equation of the state estimation module is driven by the motor thermal model and its real-time parameters. At the same time, the state estimator realizes the motor temperature estimation according to the adaptive state transfer equation. The state estimator includes a Kalman filter and a particle filter.

[0011] Specifically, the permanent magnet temperature and the stator winding temperature are estimated in real time during the operation of the permanent magnet synchronous motor, the model driving module and the data driving module are input into the adaptive state transfer equation, and then the temperature of the permanent magnet synchronous motor is estimated in real time during the operation of the permanent magnet synchronous motor through the state estimator.

[0012] The beneficial effect of the present invention is that a highly reliable permanent magnet synchronous motor temperature estimation method is proposed, which can achieve accurate monitoring of permanent magnet temperature and stator winding temperature and improve the control performance and safety of the permanent magnet synchronous motor system.

[0013] The content of the present invention is further explained below in conjunction with the accompanying drawings and specific implementation methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a brief structural schematic diagram of the present invention; DETAILED DESCRIPTION

[0015] The following takes a permanent magnet synchronous motor temperature estimation as an example and describes the technical solution of the present invention in detail, clearly and completely in combination with the accompanying drawings to facilitate professionals and technicians in this field to better understand the present invention.

[0016] Example

[0017] like Figure 1 As shown in the schematic diagram of the structure of the present invention, the specific steps of the temperature estimation method of the permanent magnet synchronous motor based on dual drive of model data in this example include:

[0018] In this example, a model-driven module is first designed to model the thermal behavior of the motor, thereby determining the lumped parameter thermal network model of the motor; in the design of the data-driven module, the real-time data of the motor sensor is used to identify the online parameters of the motor, and the real-time parameters of the lumped parameter thermal network model are obtained; the adaptive state transfer equation of the state estimation module is driven simultaneously by the lumped parameter thermal network model and its real-time parameters, and a particle filter is used as a state estimator to realize the motor temperature estimation according to the adaptive state transfer equation.

[0019] The model driving module in this example describes the thermal behavior inside the motor by establishing a lumped parameter thermal network model of the permanent magnet synchronous motor. The model includes the following main thermal nodes: permanent magnet m, stator winding s, rotor r, and housing c. The temperature change of each thermal node is controlled by the relationship between heat capacity and heat conduction, and the thermal dynamic model is established through the energy balance equation. Taking the permanent magnet and stator winding as an example, its heat conduction equation can be described as:

[0020]

[0021] Where: C m , C s are the heat capacities of the permanent magnet and stator winding, respectively, T m , T s are the temperatures of the permanent magnet and stator winding, Q m , Q s are the heat powers absorbed by the permanent magnet and stator winding, Q m→s , Q s→c They are the heat conduction power from permanent magnet to stator winding and from stator winding to casing respectively.

[0022] In this way, the internal thermal behavior of the motor can be fully described and represented in the form of a lumped parameter thermal network model.

[0023] The data-driven module in this example collects the operating data of the permanent magnet synchronous motor in real time and uses the data fitting method to identify the parameters of the lumped parameter thermal network model online. These operating data include motor speed, three-phase current, cooling medium temperature, etc. Based on these real-time data, the module dynamically adjusts the thermal conductivity and heat capacity parameters of the model to improve the adaptability of the thermal model to the actual operating state. When the cooling conditions or load conditions change, the module can adjust the model parameters in real time to make it more suitable for the actual operating environment, thereby providing high-precision input for subsequent state estimation.

[0024] The state estimation module in this example is based on a particle filter, which realizes real-time estimation of the temperature of a permanent magnet synchronous motor by combining the lumped parameter thermal network model and the adaptive state transfer equation formed by the real-time identified model parameters. The specific workflow of the particle filter includes the following steps:

[0025] Initialize particle set: The particle filter represents the possible state of the system in the form of a particle set. During initialization, each particle represents a possible temperature distribution, and the initial distribution satisfies the system prior information:

[0026]

[0027] in, is the initial i-th particle, N is the number of particles, and p(x0) represents the initial state distribution of the system.

[0028] State prediction: Use the thermal adaptive state transfer equation to predict the particles. The state transfer formula is:

[0029]

[0030] in, is the ith particle at time t, is the state transfer function, u t is the system input at time t, w t The system noise input at time t obeys the zero-mean Gaussian distribution with a covariance of Q.

[0031] Weight update: According to the temperature data z collected by the sensor at time t in real time t , the weight of the particle at time t Update. The weight calculation formula is:

[0032]

[0033] in, is the observation likelihood of the particle.

[0034] Resampling: To prevent particle degradation (i.e., the weights are concentrated in a few particles), a resampling step is used to regenerate the particle set so that the particles are evenly distributed in high-probability areas.

[0035] Temperature estimation output: Finally, the temperature estimation result at time t is output based on the weighted average value of the particles in is the estimated temperature of the permanent magnet at time t, Estimated temperature of stator winding at time t:

[0036]

Claims

1. A temperature estimation method for a permanent magnet synchronous motor based on dual drive of model data, characterized in that , including the following steps: Step 1, design the model driving module; Step 2, design data-driven module; Step 3, design a state estimation module; Step 4, real-time estimation of permanent magnet synchronous motor temperature: input the model driving module and the data driving module into the adaptive state transfer equation, and then use the state estimator to estimate the temperature of the permanent magnet synchronous motor in real time during the operation of the permanent magnet synchronous motor.

2. The model driven module according to claim 1, characterized in that: The thermal behavior of the motor is modeled and used to obtain the motor thermal model.

3. The data driving module according to claim 1, characterized in that: Using motor sensor data, online parameter identification of the motor thermal model is performed through a data-driven method.

4. The state estimation module according to claim 1, characterized in that: The adaptive state transfer equations are driven by the motor thermal model and its real-time parameters. Meanwhile, the state estimator realizes real-time temperature estimation of the permanent magnet synchronous motor according to the adaptive state transfer equations.

5. The motor thermal model according to claim 2, characterized in that: Includes lumped parameter thermal network model and motor thermal field simulation model.

6. The data driven method according to claim 3, characterized in that: Including data fitting method, probability statistics method and deep learning method.

7. The state estimator according to claim 4, characterized in that Includes Kalman filter and particle filter.

Images