To achieve high robustness and performance, this paper proposes an enhanced observer based on a predictive error model for deadbeat predictive current control method (DPCC-PEMO). Firstly, the mathematical model of permanent magnet synchronous motor (PMSM), deadbeat predictive current control (DPCC) method and disturbance analysis are presented. Secondly, to estimate the trend of current variation in advance under non-linear disturbance effect of model mismatch, the predictive error model is designed based on recursive least squares (RLS) algorithm. Furthermore, the enhanced observer combined with the predictive error model (PEMO) is proposed and its stability and high convergence are proved. Finally, utilizing the PEMO, the predictive current model is established and the proposed DPCC method is developed. The experimental results validate the strong robustness and excellent dynamic tracking performance of proposed method in PMSM systems.

Due to the harsh operating environment and complex operating conditions of electric vehicles (EVs), the physical parameters of the motor undergo nonlinear changes, leading to mismatches in both the physical model and system inertia. To address this issue, an adaptive inertia observer-based model-free predictive current control (MF-PCC) strategy is proposed in this paper, and applied to the EV's permanent magnet synchronous motor (PMSM) driving system to adjust the system inertia online. The current load of the EV is converted to mass and load inertia. An adaptive inertia method is designed to achieve inertia matching between system inertia and load inertia based on online estimated load torque, reducing the influence of inertia mismatch. The implementation of the proposed method requires no physical parameters to eliminate their effects. The control performance is analyzed in principle using Bode diagrams and zero-pole maps with different sampling periods and inertia ratios. The effectiveness and correctness of the proposed method are demonstrated through experimental results compared with the MF-PCC with fixed inertia and conventional PCC strategies, as well as the advantages of better dynamics and current quality with enhanced robustness through adaptive inertia.

As the updating frequency decreases, the stagnation effect poses a challenge to the accuracy of model-free predictive control (MFPC) based on signal gradients. To mitigate this issue from the perspective of sampled data, a low-stagnation model-free predictive current control (MF-PCC) strategy is proposed in this article and applied in permanent magnet synchronous motor (PMSM) drives. Recognizing that some elements of the sampled data do not contribute to accurately depict the motion characteristics and operational states of the system, a frequency-converting double second-order generalized integrator (FC-DSOGI) structure is improved and utilized to extract these elements and reinject them by a random gain, thereby generating a coercive difference to reduce the stagnation effect. Furthermore, fuzzy logic is developed to determine the optimal gains for the control law, thereby enhancing the control performance. The generalized universal model (GUM) is chosen as an illustrative case. Through experimental validation, the effectiveness of the proposed method is demonstrated with enhancements in current quality and model accuracy, alongside enhanced robustness. Moreover, it showcases the superiority of the low-stagnation over conventional control strategies.

To improve the performance and the robustness of finite control set predictive current control (PCC) under the influence of current measurement error in the actual system, robust exploration based finite control set predictive current control (RE-PCC) is proposed in this paper. Firstly, to analyze directly transient and steady state of system operation, a voltage cost function without weighting factor is designed based on predicted voltage error vector, reference voltage vector and evaluated voltage vector. Secondly, the robust exploration is designed to overcome the influence of measurement error by combing the amplitude and phase attribute between the predicted voltage error vector and the ideal reference voltage vector. Thirdly, based on the robust exploration, the desired sector and the prior vector set is derived for cost function evaluation. It is great benefit to computational burden and switching losses by reducing the seven vectors to three. Finally, the RE-PCC method is experimentally compared with an improved PCC method in terms of robustness, current quality, torque ripple, dynamic characteristics and efficiency. The experimental results verify the high robustness and superior control performance of RE-PCC method in surface-mounted permanent magnet synchronous motor (SPMSM) drives.

A cascaded deadbeat predictive control strategy with online disturbance compensation is proposed for a three-phase interleaved boost converter in this paper. The topology of the three-phase interleaved converter is symmetric, so the inner loop controller is also designed symmetrically. For the purpose of realizing the error-free tracking of reference value, the deadbeat predictive control method is adopted for inner and outer loops with Luenberger observers, which are designed to estimate and compensate the disturbances of load variation in the power model as well as the unknown resistor of inductance in the current model. To eliminate the influence of a time delay, a two-step predictive control method is adopted in the predictive model. In the aspect of parameter design, the pole placement method is adopted to determine the gain of the observer. A series of simulations and experiments are carried out to test the proposed strategy under steady and dynamic conditions. It is shown that the proposed control strategy has faster dynamic response and stronger robustness against disturbance than the conventional model predictive control.

Model predictive control has been widely developed for electrical drive systems. The weighting factors are key parameters affecting the control performance of the motor. This paper proposes a Bayesian ascent and Gaussian process method for the optimized weighting factors calculation. The root mean square error of the current in the d-q axis is taken as a criterion. A data-driven probabilistic proxy model is formed based on the Gaussian process, and the optimal weighting factor is obtained by Bayesian ascent. Grid sampling, random sampling, and sampling in the maximum possible region are integrated to improve sampling reliability. Furthermore, the maximum torque per ampere calibration and motor parameter identification can be implemented by the proposed method as well. Finally, a new perspective of global parameter optimization based on data probability is proposed. The effectiveness of the proposed methods is verified by experimental results.