Particle Swarm Optimization and Gradient Descent Methods for Optimization of PI Controller for AGC of Multi-area Thermal-Wind-Hydro Power Plants

The automatic generation control (AGC) of three unequal interconnected Thermal, Wind and Hydro power plant has been designed with PI controller. Further computational intelligent technique Particle Swarm Optimization and conventional Gradient Descent technique have been used to improve the performance of Automatic Generation Control (AGC) system. The reheat turbine for thermal area and appropriate generation rate constraint (GRC) have been considered for thermal area. Particle swarm optimization (PSO) technique and Gradient Descent methods are used to simultaneously optimize the proportional gain (Kp), integral gains (Ki), speed regulation parameter (Ri) and frequency bias (Bi) parameter of different areas. Most of the literature for AGC used classical approach based on integral squared error (ISE) technique etc. for optimal selection of controller parameters. This is a trial and error method; extremely time consuming when several parameters have to be optimized simultaneously. The computational intelligence based technique like PSO is more efficient and fast technique for optimization of different gains in load frequency control. Further the performance of PSO is better than GD method for optimization of various parameters and the controller gives the improved dynamic performance for three area network with Thermal-Wind-Hydro power plants. MATLAB/SIMULINK is used as a simulation tool.