From forecasting to scheduling in short term power procurement using stochastic optimization

Short Term Load Forecasting is required for efficient operation of generation units and for making decisions on power trading in the short term markets. Presence of multiple factors affecting load forecasts leads to supply demand imbalance owing to forecasting errors. Excess/deficit power is balanced through real time markets and have associated costs. Presence of asymmetry in the costs due to excess/shortage of power leads distribution companies (DISCOMs) to make power procurement/scheduling decisions that differ from unbiased forecasts to mitigate overall losses. This leads to the stochastic scheduling problem that is described in this paper.