This letter proposes a data-driven iterative learning predictive control architecture for power converters. The main objectives of this letter are to enhance the robustness and remain the high performance of finite control-set model predictive control (FCS-MPC) under unmodeled dynamics and parameter mismatch conditions. More specifically, an iterative dynamic linearization technique is utilized to equivalently reformulate the nonlinear power converter system at each operating point. Based on this, a model-free adaptive control scheme is presented to iteratively determine the optimal control actions. Due to the incorporation of iterative learning control and data-driven concept into the FCS-MPC framework, the effect of parameter perturbations can be alleviated in the proposed method, while creating a positive effect on the tracking error. Finally, a convergence analysis is provided and experimental investigations on a three-level neutral-point-clamped (NPC) converter confirm the effectiveness of the proposed method.

In this article, a novel robust data-driven model-free predictive control framework based on the I/O data of the controlled plants, which is performed by incorporating the neural predictor-based model-free adaptive control and finite control-set model predictive control, is first proposed. The salient feature of the suggested framework is that the uncertainties, such as unmodeled dynamics and external disturbances, can be explicitly addressed in controlled systems. From a practical standpoint, however, the potential of this proposal is limited by a significantly increased online computational complexity, which makes it difficult to implement. To circumvent this limitation, a supervised imitation learning technique using data labeled is developed to imitate the known suggested controller, which the majority of the online computational burden can be transformed into offline computing by utilizing a trained artificial neural network subject to robustness characteristics. In particular, this development motivates a much simpler robust predictive control solution, which is convenient to implement in applications. Thus, by this proposal, the online implementation of much more complex predictive control strategies is made possible, and it explores a new possibility for future development of the complex control methodology. Finally, extensive simulative and experimental investigations for modular multilevel converter validate the interest and viability of the proposed design methodology.

The focus of this article is to introduce the concept of an online reinforcement learning (RL) solution and to propose a novel finite control-set model predictive control framework subject to system uncertainties, which possesses the excellent applicative potential for power converter systems with unknown perturbations. In this framework, the control task is performed by incorporating an adaptive neural network approximation-based RL and neural predictor-based predictive current control solution. To be more precise, a critic neural network is responsible for learning a strategic utility function online, and an actor network is developed to derive control behaviors by approximating the unknown model dynamics and optimizing the learned utility function obtained from the critic network. Compared to previous works, it not only attenuates the inherent issues of system uncertainties and unknown disturbances, but also provides a flexible framework and allows the enhancement of control property. Furthermore, by deploying the Lyapunov approach, it shows that all signals in the closed-loop system are uniformly ultimately bounded. Finally, numerical simulation and experiments validate our theoretical findings.

This letter investigates the possibility of deploying a novel finite control-set model predictive control solution for solving the ongoing research challenges in predictive control regulated modular multilevel converter, i.e., model parameter sensitiveness and excessive computational burden as well as weighting factors selection. Specifically, it is realized by cascading a predictor-based neural network design, which enables a smooth and fast identification of system dynamics, and a computationally efficient finite-set predictive control, which is responsible for simplifying the rolling optimization and reducing the computational complexity. The main contribution of the proposed methodology relies on the fact that no knowledge of any model parameters and weighting factors in whole control process are required, which leads to a significant enhancement in the robustness and reliability of the control system in the presence of parametric uncertainties, while remaining computationally feasible. Finally, the stability analysis is given, and the proposed methodology is experimentally assessed for modular multilevel converter, where steady-state and transient-state performance tests confirm the interest of the proposal.

This paper investigates a data-driven based predictive current control (DD-PCC) approach for a modular multilevel converter (MMC) to circumvent the sensitiveness to parameter variation and unmodeled dynamics of finite control-set model predictive control (FCS-MPC) method. By integrating a model-free adaptive control based data-driven solution into the FCS-MPC framework, the performance deterioration caused by model uncertainties is suppressed. The design of the suggested controller is only based on input-output measurement data, where neither the parameter information nor the knowledge of detailed MMC models are required, leading to improved robustness against parameter drifts and model uncertainness. Moreover, a simplified cost function formula that takes into account output current tracking and circulating current regulation is constructed to efficiently determine the optimal insertion index of each arm. Finally, simulation and experimental results are obtained to verify the steady-state, dynamics, and robustness performance of the proposed approach.

An event-triggered control technique has been developed recently. This technique explicitly reduced the signal transmission by introducing a flexible design of threshold inequalities. It was later extended to event-triggered model-predictive control for power converter systems. In this letter, by incorporating this control technique into an extended state-observer-based finite-control-set model-predictive control framework, we have developed a new model-predictive control architecture for power converter systems with parametric uncertainties. Meanwhile, a novel cost function with respect to the angle minimization term is embedded into this proposal. The novelty of our development lies not only in integrating the event-triggered mechanism with the suggested finite-control-set model-predictive control architecture for facilitating the alleviation of performance deterioration caused by parameter variations and model uncertainties, but also in a multiobjective optimization design that allows the switching frequency in a low value. Finally, extensive simulative and experimental investigations for a modular multilevel converter confirm the interest and the viability of the proposed design methodology.