A Model Predictive Control for the Dynamical Forecast of Operating Reserves in Frequency Regulation Services

The intermittent and uncontrollable power output from the ever-increasing renewable energy sources, require large amounts of operating reserves to retain system frequency within its nominal range. Based on day-ahead load forecasts, many research works have proposed conventional stochastic approaches define their optimum margins for reliability enhancement at reasonable production cost. In this work, we aim delivering real-time forecasting lower operating-reserve requirements based intra-hour weather update predictors. critical predictors historical data, train an artificial model that is able forecast ahead with great accuracy. This a feed-forward neural network two hidden layers, which performs forecasts aid predictive control developed recommendations intra-hourly and, assessing impact significance target, it corrects imposed deviations. Performing daily simulations annual time-horizon, observe significant improvements exist in terms decreased reserve regulate violated frequency. fact, these can exceed 80% during specific months winter when compared robust formulations isolated systems.