Online multi‐parameter estimation of interior permanent magnet motor drives with finite control set model predictive control

Abstract

This study presents an online multiparameter estimation scheme for interior permanent magnet motor drives that exploits the switching ripple of finite control set (FCS) model predictive control (MPC). The combinations consist of two, three, and four parameters are analysed for observability at different operating states. Most of the combinations are rank deficient without persistent excitation (PE) of the system, e.g. by signal injection. This study shows that high frequency current ripples by MPC with FCS are sufficient to create PE in the system. This study also analyses parameter coupling in estimation that results in wrong convergence and propose a decoupling technique. The observability conditions for all the combinations are experimentally validated. Finally, a full parameter estimation along with the decoupling technique is tested at different operating conditions.

An Effective Model-Free Predictive Current Control for Synchronous Reluctance Motor Drives

“…In model-based predictive current control schemes the current predictions are based on the state-space model of the system, conveniently written in a dq reference frame synchronous to the rotor flux. The prediction accuracy is strongly influenced by the knowledge of the parameters, for which several identification techniques are available [7]- [10].…”

Approach for classifying direct PCs applied to AC motor drives

“…In these experiments, sampling frequency is adjusted to 10 kHz, the nominal load is applied to the motor, and the motor rotates in the steady state. In each experiment, average switching frequency of the inverter is calculated using (20), where N is the total switching transitions during a fixed period, e.g. 200 ms, and K is total number of inverter switches which are six in two-level VSI…”

“…MPC is the other category of PCs that have been widely used in power electronics and AC motor drives [19,20]. In MPCs, to achieve high-performance control of a motor, a cost function is introduced in which various linear or non-linear terms can be considered.…”

Unified Predictive Current Control of PMSMs with Parameter Uncertainty

“…Advanced robust techniques are often introduced to design control law, such as the control-Lyapunov function [13], disturbance estimation [14], and differential flatness [15]. Another strategy lies in the parameter identification, such as a recursive inductance estimator [9], multi-parameter identification with a decoupling method [16], model reference adaptive technique [17], extended Kalman filter [18], modified particle swarm optimization [19], Adaline neural network [20], etc. All those methods aim to acquire accurate parameters, which is not necessary in the PCC schemes.…”