Integrated and Fire Spiking Neuron Model for Improved Wind Speed Forecasting

The widespread integration of renewable energy resources (RERs) is needed for achieving sustainable development goals (SDGs) like affordable and clean energy, climate action and industry, innovation, and infrastructure. Wind energy is a type of RER with huge potential to fulfill the ever-increasing electricity demand of the world. However, the intermittent nature of wind hinders the large-scale integration of wind turbines into the existing power system. The main source of intermittency is due to wind speed (WS), and this intermittency can be overcome by implementing an accurate forecasting model. The traditional WS forecasting models require huge data for improved outcomes and have a high computational time. Therefore, in this proposed study, we present a novel approach that leverages the spiking neurons functionality for improved WS forecasting with reduced computational time. The spiking neurons with the configuration of integrated and fire (I&F) are used to propose a new architecture called the I&F neurons network (IF-NN). The datasets of four different geographical locations of Saudi Arabia are used for simulation purposes, and the performance of IF-NN is compared with state-of-the-art networks. Findings illustrate that for the datasets of Al Jouf and Turaif cities, the proposed model records the improvement of 74.3% and 68.8% in mean absolute percentage error (MAPE) as compared to the recurrent neural network-long short-term memory (RNN-LSTM) technique, which was found to be the second-best performing model for these datasets. Furthermore, the MAPE of IF-NN is 69.9% and 65.9% better than the MAPE of convolutional neural network-LSTM (CNN-LSTM), which gives the second-best forecasting performance among the models used for comparative analysis for Haffer Al Batin and Yanbu datasets. The comparative analysis also illustrates that IF-NN has better computational time as compared to RNN for each dataset because of spiking neurons.