Development of VRE forecast error models for evaluating impact on generation scheduling

As Thailand's renewable energy sector — particularly solar and wind power—expands under the government's Renewable and Alternative Energy Development Plan (AEDP), forecasting energy production becomes increasingly important due to the inherent variability and uncertainty of these renewable sources. This research develops a model to assess and analyze forecast errors in solar and wind power generation. The model development process is based on the concept of probabilistic forecasting, which presents prediction outcomes as a range of possible values, each associated with a specific probability. This approach utilizes probability distributions to represent the likelihood of various future events. The study employs a parametric approach to error modeling, which involves the use of mathematical models with explicitly defined parameters to describe the probability distribution of the variable of interest. This method entails selecting an appropriate distribution model and estimating its parameters based on the available data. The resulting model is expressed in the form of a probability density function (PDF). The study examines the distribution of forecast errors for solar and wind energy by dividing the data into four distinct time periods to enhance model accuracy. To apply forecast error model to evaluate the impact of forecast accuracy on generation scheduling, and to help determine the proper size of the operating reserve, the model applies probabilistic forecasting techniques. These include adjusting the Prediction Interval (PI) width through scale parameters to capture forecast uncertainty. The model's flexibility allows for adjustments to forecast accuracy levels based on varying time periods, providing insights into optimizing energy production planning and reducing forecast-related costs. This approach will enhance operational efficiency and support sustainable energy use in Thailand.