Operating Reserve forecasting in a wind integrated power system using Hybrid Support Vector Machine-Fuzzy Inference System

Abstract

— In a restructured power system, Ancillary services (AS) are required to balance load generation mismatches and to meet unforeseen contingencies. Operating Reserve is a major part of AS which is highly uncertain to forecast, mainly due to unpredictability of customer needs, over or under production of Energy and unpredictability in the integration of renewable energy sources. In this work wind integration is considered as a factor to forecast operating reserve. Increase of wind integration into power system, needs larger quantities of operating reserve. This demands an increase in the cost of generation and emissions. Forecasting the Operating Reserve Ancillary Service helps the system operators (SO) to plan scheduling of generators in advance and also in better bidding environment. Forecasting tools like feed-forward networks, Time series models were used to forecast load and Electricity price in the past. In this paper a hybrid method consisting of Support Vector Machines (SVM) and Fuzzy Interface System (FIS) is used to forecast Operating Reserve in Day-ahead market. Case studies using CAISO and ERCOT ISOs are presented. The SVM-FIS method is found to be better forecasting tool to predict the operating reserve Ancillary Service.