An on-line parameter identification method of asynchronous motor based on data generation model

Abstract

The invention discloses an asynchronous motor parameter online identification method based on a data generation model. This method combines the advantages of model-based identification and data-basedidentification, The overall idea is: (1), the probabilistic model between output torque and state is established by off-line data. (2), in the process of on-line identification, the model-based identification method is used as 'data generator'. Then, the on-line identification of motor parameters is carried out by using the data-based method combined with the torque probabilistic model in (1), andthe probabilistic model of motor parameter identification value is finally obtained. The method of the invention has the advantages that: in the identification process, the training data set is easily obtained; The identified parameters are not affected by the model error. The parameters identified make the motor keep the optimal torque-current ratio in operation. The parameters identified take into account the uncertainties in the operation of the motor and have high robustness.

Description

technical field [0001] The invention relates to the technical field of motor control, in particular to an on-line identification method of asynchronous motor parameters based on a data generation model. Background technique [0002] During the operation of the asynchronous motor, the output characteristics will change with the change of the internal parameters. Therefore, it is necessary to identify the parameters of the motor online during the running process. Most of the traditional parameter methods for asynchronous motors are based on the voltage-current dynamic model or equivalent circuit model of the motor, which are called model-based methods, such as model reference adaptation method (MRAS), sliding mode observer method (SMO) and linear Least Squares (LSM) and more. However, model-based methods have the following disadvantages: [0003] 1) The identification accuracy and identification convergence of parameters greatly depend on the equivalent model of the motor, ...

AI Technical Summary

Problems solved by technology

[0003] 1) The parameter identification accuracy and identification convergence greatly depend on the equivalent model of the motor, which is easily affected by noise and model stability

Therefore, in some extreme working conditions (such as extremely low speed, high speed field weakening, etc.), there are disadvantages such as low identification accuracy and poor robustness.

[0004] 2) Only the physical values ​​of the parameters can be identified. However, in some cases, the physical values ​​of these parameters cannot make the motor run in an optimal state

[0007] 1) Data-based methods require a large amount of labeled data for training, and it is difficult to obtain such data

[0008] 2) The training process is offline. Like the model-based method, it cannot eliminate the influence of various uncertain factors during online operation.

Therefore, it is only suitable for offline identification of motor parameters

Images