Parallel Tracking Control for Beam Pumping Units Based on Adaptive Dynamic Programming

ABSTRACT

This paper develops a novel parallel tracking control method for beam pumping units (BPUs), broadening the application scope of parallel control. By integrating neural networks with parallel control, the proposed approach can ensure the robustness of BPUs, with a rigorous stability proof provided. First, the simulation model and problem transformation are introduced to formulate the tracking control problem and construct an augmented error system for the parallel control design. Subsequently, a neural network (NN) is employed to approximate the unknown BPU system dynamics. Furthermore, the NN‐based implementation of the proposed method is presented utilizing the residual of the Hamilton‐Jacobi‐Bellman (HJB) equation. By designing the control law in differential form, the controller compensates for unknown disturbances using historical states. Simulation results demonstrate the convergence of the parallel control algorithm, while comparative studies validate the superior performance of the proposed controller.