A Comparative Analysis of Linear and Nonlinear Control of Wave Energy Converter for a Force Control Application

Abstract

The aim of wave energy converters (WECs) is to harvest the energy from the ocean waves and convert into electricity. Optimizing the generator output is a vital point of research. A WEC behaves as a nonlinear system in real ocean waves and a control that approximates the behaviour of the system is required. In order to predict the behaviour of WEC, a controller is implemented with an aim to track the referenced trajectory for a force control application of the WEC. A neural model is implemented for the system identification and control of the nonlinear process with a neural nonlinear autoregressive moving average exogenous (NARMAX) model. The neural model updates the weights to reduce the error by using the Levenberg-Marquardt back-propagation algorithm for a single-input-single-output (SISO) nonlinear system. The performance of the system under the proposed scheme is compared to the same system under a PI-controller scheme, where the PI gains have been tuned accordingly, to verify the control capacity of the proposed controller. The results show a good tracking of dq (direct-quadrature) axes currents by regulating the stator currents, and hence a force control is achieved at different positions of the translator. The dynamic performance of the control is verified in a time domain analysis for the displacement of the translator.

Keywords

• current control, force control, neural NARMAX, permanent magnet linear generator (PMLG), wave energy converter.