The power consumption costs of base stations (electricity fees) accounts for about 16% of the network operation costs. Therefore, energy saving and emission reduction for base stations are the major concerns of operators. According to statistics, the traffic volume on the network has obvious tidal effect. The difference in traffic volume between peak hours and off-peak hours reaches up to four times. However, a majority of base stations keep running (for 24 hours with all resources occupied). The power consumption does not dynamically change with traffic volume, which causes a waste of resources. In the power consumption composition of a conventional macro base station, the power consumption of the main equipment accounts for 50%, in which 80% derives from the power consumption of RF units. The power consumption of power amplifiers (PAs) contributes 79% to the power consumption of RF units. Distributed base stations (the power consumption of the main equipment accounts for 90%+) are becoming the mainstream, in which the power consumption ratio of the main equipment will continue to increase. Therefore, in terms of the main equipment, effective reduction in the power consumption of carrier PA modules is key to energy saving for base stations.
Artificial intelligence is introduced to the element network management system (EMS) of base stations to predict the cell traffic. In this way, the shutdown threshold of each cell can be accurately set. The power consumption of a base station can be reduced by 15% to 30% without sacrificing the coverage and network KPIs of the base station, which decreases the network operation costs.