Data-Driven-Based Vector Space Decomposition Modeling of Multiphase Induction Machines

For contemporary variable-speed electric drives, the accuracy of machine's mathematical model is critical for optimal control performance. Basically, phase variables multiphase machines are preferably decomposed into multiple orthogonal subspaces based on vector space decomposition (VSD). In available literature, identifying correlation between states governed by dynamic equations and parameter estimate different IM remains scarce, especially under unbalanced conditions, where effect secondary sounds influential. Most literature has relied simple RL circuit representation to these subspaces. To this end, paper presents an effective data-driven-based harmonic n-phase IMs using sparsity-promoting techniques machine learning with nonlinear dynamical systems discover governing equations. Moreover, proposed approach computationally efficient, it precisely identifies both electrical mechanical dynamics all a single transient startup run. Additionally, derived can be reformulated standard canonical form induction easily extract parameters online measurements. Eventually, modeling experimentally validated 1.5 Hp asymmetrical six-phase machine.