Energy-saving Scheme of 5G Base Station Based on LSTM Neural Network

Abstract

Abstract As China’s new infrastructure,5G has received national and social attention. 5G promotes economic to grow rapidly. But, the high energy consumption caused by the massive deployment of 5G base stations cannot be ignored. The total annual power consumption is expected to reach 243 billion degrees when the 5G base station is fully built. In the tidal scene, some 5G base station in an idle state still power fully, which causes great power waste. The historical volume of base station business data is used to train LSTM model, and predict the future base station business. When the business is lower than the threshold, the base station will be closed to avoid unnecessary power waste. And the LSTM model prediction results fits the original data ideally. By implementing the power saving strategy, the energy consumption of the base station is reduced by 18.97 %. A single station can save 1174 degrees of electricity yearly. It can be seen that the energy saving effect is remarkable.