Dynamic MPPT Controller Using Cascade Neural Network for a Wind Power Conversion System with Energy Management

Abstract

ABSTRACT This paper proposes a dynamic maximum power point tracking (MPPT) controller for a permanent magnet synchronous generator (PMSG)-based variable speed autonomous wind power conversion system (WPCS) with an energy storage system (ESS). The dynamic controller extracts the maximum power from the WPCS under variable speed conditions using a cascade neural network (CNN)-based MPPT algorithm. Also, the efficient power management between an ESS, load and WPCS is obtained using a conventional PI controller. The CNN is trained and tested using simulation and experimental data, respectively. The performance of CNN-based MPPT is also compared with feed-forward-based MPPT in terms of accuracy and complexity. This paper also discusses FPGA implementation aspects of CNN-based MPPT and hardware resources management for desired accuracy. The proposed method is first implemented on a simulation model 900-watt PMSG-based WECS and results are validated. Further, the enthusiastic results were finally evaluated in a hardware setup of 1 HP PMSG-based WECS. The proposed method is shown to extract maximum power with a simple structure and provide a better response to variable wind speed.