Improved Performance for PMSM Sensorless Control Based on Robust-Type Controller, ESO-Type Observer, Multiple Neural Networks, and RL-TD3 Agent.

This paper summarizes a robust controller based on the fact that, in the operation of a permanent magnet synchronous motor (PMSM), a number of disturbance factors naturally occur, among which both changes in internal parameters (e.g., stator resistance Rs and combined inertia of rotor and load J) and changes in load torque TL can be mentioned. In this way, the performance of the control system can be maintained over a relatively wide range of variation in the types of parameters mentioned above. It also presents the synthesis of robust control, the implementation in MATLAB/Simulink, and an improved version using a reinforcement learning twin-delayed deep deterministic policy gradient (RL-TD3) agent, working in tandem with the robust controller to achieve superior performance of the PMSM sensored control system. The comparison of the proposed control systems, in the case of sensored control versus the classical field oriented control (FOC) structure, based on classical PI-type controllers, is made both in terms of the usual response time and error speed ripple, but also in terms of the fractal dimension (DF) of the rotor speed signal, by verifying the hypothesis that the use of a more efficient control system results in a higher DF of the controlled variable. Starting from a basic structure of an ESO-type observer which, by its structure, allows the estimation of both the PMSM rotor speed and a term incorporating the disturbances on the system (from which, in this case, an estimate of the PMSM load torque can be extracted), four variants of observers are proposed, obtained by combining the use of a multiple neural network (NN) load torque observer and an RL-TD3 agent. The numerical simulations performed in MATLAB/Simulink validate the superior performance obtained by using properly trained RL-TD3 agents, both in the case of sensored and sensorless control.

Comparative Performance Analysis of the DC-AC Converter Control System Based on Linear Robust or Nonlinear PCH Controllers and Reinforcement Learning Agent.

Starting from the general topology and the main elements that connect a microgrid represented by a DC power source to the main grid, this article presents the performance of the control system of a DC-AC converter. The main elements of this topology are the voltage source inverter represented by a DC-AC converter and the network filters. The active Insulated Gate Bipolar Transistor (IGBT) or Metal-Oxide-Semiconductor Field-Effect Transistor (MOSFET) elements of the DC-AC converter are controlled by robust linear or nonlinear Port Controlled Hamiltonian (PCH) controllers. The outputs of these controllers are modulation indices which are inputs to a Pulse-Width Modulation (PWM) system that provides the switching signals for the active elements of the DC-AC converter. The purpose of the DC-AC converter control system is to maintain ud and uq voltages to the prescribed reference values where there is a variation of the three-phase load, which may be of balanced/unbalanced or nonlinear type. The controllers are classic PI, robust or nonlinear PCH, and their performance is improved by the use of a properly trained Reinforcement Learning-Twin Delayed Deep Deterministic Policy Gradient (RL-TD3) agent. The performance of the DC-AC converter control systems is compared using performance indices such as steady-state error, error ripple and Total Harmonic Distortion (THD) current value. Numerical simulations are performed in Matlab/Simulink and conclude the superior performance of the nonlinear PCH controller and the improvement of the performance of each controller presented by using an RL-TD3 agent, which provides correction signals to improve the performance of the DC-AC converter control systems when it is properly trained.