Wind Distribution Analysis and Prediction for a Risk-Aware Renewable Energy Market

The government is planning to switch the main energy sources from fuels to renewables in accordance with the global decarbonization goal. However, the way that the current power system operates does not take the stochasticity of renewable energy into consideration. This increases the public cost of electricity when renewable sources experience shortfalls due to their instability. To solve this issue, we designed a new risk-aware optimization problem in place of the old deterministic model to be used for managing the electricity production schedules and prices. This study focuses on wind energy. We analyzed wind distributions at the 33 existing NYISO wind farms and constructed predictive models to help with the important transition to renewable energy in society. The wind distribution prediction model achieved a classification accuracy of 81% and the covariance matrix prediction model obtained MAE = 0.53. They can be applied together to generate the uncertainty set for solving the risk-aware optimization problem