Experimental Analysis and Adaptive Intelligent Load Frequency Control (Case Study of Shiroro Hydroelectric Power Generation Station, Nigeria)

Abstract

The stabilization of electricity is depended on absolute load frequency control. Rather than complicated advanced control techniques, this paper presents an adaptive intelligent load frequency control (AI-LFC) scheme that can be applied to improve the control performance of hydroelectric power generation plant in the presence of environmental disturbances. The AI-LFC scheme has been proposed for the adaptive control of a well-established Shiroro hydroelectric power generation Station located in Niger State, Nigeria as the case study. This study begins with 3-year data acquisition and experimental data analysis of the key sixteen parameters from the Shiroro Hydroelectric Power Plc in Niger State, Nigeria. Five important output control parameters have been identified to ensure the smooth operation and efficient control of the HPGP to maintain stable grid frequency. Prescribed reference trajectory tracking of the output predictions of the five control parameters as well as the output prediction errors have been used to evaluate the performance of the proposed AI-LFC scheme against that of a well-tuned artificial neural network-based proportional-integral-derivative (NN-based PID) controller for performance comparison purposes. The simulation results show that the proposed AI-LFC scheme outperforms the NN-based PID controller in terms of absolute tracking of the prescribed reference trajectory with absolute zero output predictions errors. The NN-based PID controller exhibits larger output prediction errors, overshoots, non-minimum and oscillatory behaviours, and require significant amount of control efforts to track the desired reference trajectory with occasional inability to reach the desired prescribed reference trajectory. The proposed AI-LFC scheme has been successfully formulated, implemented and validated for the modeling and control of the Shiroro hydroelectric power generation plant as a case study. The proposed AI-LFC outperforms a well-tuned NN-based PID controller and offers promising optimal adaptive control potentials that could be adapted for direct modeling and adaptive control of renewable systems.