Ultra-short Term Wind Power Forecast for Offshore Energy System Based on EEMD-TCN

In recent years, a coordinated energy supply system of offshore generator set and onshore power grid has been established to improve the stability of energy supply for offshore oil and gas projects. The wind speed at sea is high and fluctuating, thus improving the accuracy of ultra-short-term wind power forecasts can effectively promote the optimal and coordinated operation of offshore generator sets and onshore power sources, while also lowering carbon emissions. This study proposes a ultra short term wind power forecasting model based on empirical mode decomposition. The ensembled empirical mode decomposition (EEMD) technique was applied to separate historical wind power generation data to different kinds, which are high frequency, medium frequency, and low frequency. Key meteorological variable data index were collected as the model train set and reconstruct the other high and medium frequency components. A temporal convolutional network is used to forecast the two reconstruction components. A real-world example shows how the suggested prediction model may increase the effect of wind power generation predictions in ultra-short-term and better follow trends in wind power variance.