Short Term Load Forecasting Using In-House Algorithm Developed at ERLDC

India’s fast economic growth is being reflected in its rapidly growing power system. Renewable integration is also taking place at a very high rate matching with India’s commitment for achieving net zero goal [1]. Grid operation is becoming more challenging day by day as variability and uncertainty in the supply side increases. To ensure safe and reliable grid operation, a proper day ahead resource adequacy planning is being performed and made available to operator before actual real time operation. For ensuring resource adequacy on a day ahead basis accurate demand and generation forecasting is extremely essential. While for DISCOMs accurate demand forecasting is essential for portfolio management at the same time, it is equally important for system operators for proper reserve estimation. The electricity demand across various regions of the country exhibits a unique blend of household, agricultural, and industrial load. In the eastern region of the country demand is mainly dominated by domestic load and especially space cooling load [2]. Therefore, it is highly sensitive to weather parameters. This makes accurate forecasting extremely difficult for eastern regional states. Even the most advanced artificial intelligence-based algorithms can sometimes exhibit significant forecasting errors and tuning these complex algorithms in real time by system operators for intraday forecasting becomes difficult. To address this challenge a simple algorithm has been developed in-house at the eastern regional load despatch center. The proposed forecasting algorithm is blend of similar day forecast and heat index based incremental demand estimation. This algorithm exhibits better results and able to capture impact of weather disturbance more accurately than other commonly used methods. In this paper this simple yet effective method has been described for accurate short term load forecasting along with the challenges and way forward.