ANN Created Real Time Load Pattern Base Frequency Normalization Studies of Linked Electric Power System

In most of the automatic generation control (AGC) studies proposed so far, area control error (ACE) signal is derived considering fixed step size load disturbance which does not represent the real time operating condition of power system adequately and may cause sometimes the over regulation of the power system. Therefore, a very short-term load forecasting (STLF) using artificial neural network (ANN) is proposed to obtain a load disturbance pattern to derive an effective AGC scheme. Further, real time load data of a particular month are collected from a 220 kV substation and are used to perform STLF. The predicted hourly load is used to determine future load estimates considering a 10 minute interval basis. The ACE signal is derived accordingly. The model predictive control (MPC) based AGC scheme is designed to counter the upcoming load variations very effectively. A two-area power system having thermal power plants and interconnected via parallel AC/DC tie-lines is considered for the investigations. Furthermore, the dynamic performance of the designed control strategy is also evaluated considering the governor dead-band and generation rate constraint (GRC).