Optimal Generation Scheduling under ABT using Forecasted Load and Frequency

The conventional Generation Scheduling (GS) is a day ahead activity, aiming at minimizing operating cost subject to system as well as device constraints. The anticipated load, generation and network availability are inputs to GS problem. The new attribute added to conventional GS problem is frequency. Under Availability Based Tariff (ABT), frequency plays important role in deciding the real time generation and dispatch. The conventional GS algorithm is modified to incorporate frequency dependant part of tariff. Thus, prior knowledge of frequency along with load will be contributory in the formation of more realistic schedule of generation. This paper reports estimation of hourly frequency and load using Statistical Analysis and Artificial Neural Network (ANN) respectively. Frequency variation is random and uncertain in nature. Accurate prediction of frequency is difficult. Using probability theory and statistical analysis, the hourly frequency is estimated which forms an important input to new generation scheduling algorithm along with hourly load forecasted using ANN. The schedule prepared considering estimated frequency and load at every hour will help operator in taking appropriate decision in real time dispatch to earn maximum benefits.