| Autor: |
Ye Li, Dazhi Wang, Shuai Zhou |
| Jazyk: |
angličtina |
| Rok vydání: |
2023 |
| Předmět: |
|
| Zdroj: |
Energy Reports, Vol 9, Iss , Pp 521-528 (2023) |
| Druh dokumentu: |
article |
| ISSN: |
2352-4847 |
| DOI: |
10.1016/j.egyr.2022.11.068 |
| Popis: |
As an important mechanical parameters of permanent magnet synchronous motor (PMSM), the moment of inertia determines the control optimization and condition monitoring performance of the controller. However, since inertia cannot be measured directly, it is necessary to choose an appropriate method to identify inertia. In this paper, an extended sliding mode observer (ESMO) is proposed by taking the motor speed and inertia as the extended state variable of the observer to identify the inertia. A deformation mode of a saturation function is used as the switching function to mitigate the chattering phenomenon of conventional SMO. For unknown disturbance torque in the state equation, an improved radial basis function neural network (RBFNN) is proposed to estimate it. The improved RBFNN determine the RBF center through the eigenvector of the input data and then trains the neural network, which improves the training accuracy. The effectiveness of the method is verified by simulation experiments. The experiments show that the ESMO combined with the improved RBFNN can effectively and accurately identify motor inertia which is a method worthy of reference. |
| Databáze: |
Directory of Open Access Journals |
| Externí odkaz: |
|
