A Comparison of PI Tuning by Direct Search and Genetic Algorithm for Optimal Control of AGC in Continuous-Discrete Mode

This paper deals with Automatic generation control (AGC) of interconnected hydrothermal system in continuous-discrete mode using proportional-integral (PI) controller with different tuning approaches. Here the PI controller is initially tuned using local optimization technique such as Fminsearch (Existing MATLAB function) and optimal control strategies were taken as integral square error (ISE), integral time-absolute error (IATE) and integral time square error (ISTE). Then the same PI controller is tuned by using evolutionary algorithm i.e. Genetic Algorithm (GA). For the given system appropriate generation rate constraint (GRC) has been considered both for the thermal and hydro plants. System performances are examined considering 1 % Step Load Perturbation (SLP) in both thermal and hydro area with 1 second sampling period. Finally the performance of both the local and global optimization algorithms is compared in terms of the time domain specifi cations both for frequency deviation in each area, and tie line power.