Open AccessMTPA Control of IPMSM Drives Assisted by Deep Neural Network
Author(s) -
Bin Zhang,
Tianfu Sun,
Chengli Jia,
Riyang Yang,
Linghui Long,
Jianing Liang
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1585/1/012036
Subject(s) - artificial neural network , control theory (sociology) , control (management) , torque , signal (programming language) , computer science , scheme (mathematics) , mathematics , artificial intelligence , physics , mathematical analysis , thermodynamics , programming language
In this paper, a novel MTPA control scheme assisted by deep neural network is proposed based on a virtual signal injection concept. The deep neural network models the complex relationship between the electromagnetic torque and the d- and q-axis currents. The mathematical model in the conventional virtual signal injection MTPA control is substituted by the deep neural network. In this way, the MTPA control errors of conventional mathematical model based MTPA control schemes and conventional virtual signal injection based MTPA control schemes due to the neglect of the derivatives of machine parameters with respect to current angle or d-axis current can be avoided. The proposed control scheme was assessed by simulations under various operating conditions. Simulation results illustrate that the proposed MTPA control scheme could control the IPMSM operating on the MTPA points accurately and the errors caused by the neglect of the derivatives of machine parameters with respect to current angle or d-axis current were avoided.