Research On Load Forecasting And Energy Storage Scheduling System For 5G Base Station

5G base station construction,one of the new infrastructures,is steadily advancing the construction process.Due to the high frequency band of 5G communication,the cost of electricity is about 6-12 times of 4G base stations,and the high power consumption of 5G communication increases the burden of power grid during the peak electricity consumption period.And the energy storage configured in 5G base stations is idle for a long time,resulting in the waste of electrical energy resources,so how to effectively revitalize the idle energy storage resources in base stations and increase the base station energy storage revenue while reducing the regional load peak-to-valley difference is the focus of current research.To this end,this paper studies the following aspects:(1)Introducing the load composition of 5G base stations and the impact of user communication load on base station energy consumption,analyzing the spatial and temporal correlation of 5G base station load while further studying the characteristics of base station energy storage batteries and constructing an energy storage model to explore the scheduling potential of base station standby energy storage and provide a basis for base station energy storage scheduling in the later thesis.(2)5G base station load data prediction is the core of this thesis.In this thesis,a Complete Ensemble Empirical Mode Decomposition with Adaptive Noise(CEEMDAN)data processing method is adopted to decompose the complex load signal into multiple component signals and improve the fit of the neural network.To improve the prediction accuracy of the algorithm,Particle-Swarm-Optimization(PSO)algorithm is used to optimize the number of neurons and learning rate of the Gate Recurrent Unit(GRU).Combining CEEMDAN and PSO-GRU for 5G base station load data prediction,the accuracy of the combined prediction model is high after the analysis of arithmetic cases,which verifies the superiority of the combined prediction algorithm constructed in this thesis.(3)The optimal scheduling of 5G base station energy storage is the focus of the research in this thesis.To address the problems of poor particle search ability and over aggregation of particles in Multi-Objective Particle Swarm Optimization(MOPSO),we construct an improved Multi-Objective Particle Swarm Optimization(IMPROVE Multi-Objective Particle Swarm)by changing the weights,learning factors and adaptive dispersal of particles.Objective Particle Swarm Optimization(IMOPSO).Meanwhile,the thesis aims at minimizing the regional load variance and maximizing the base station energy storage revenue,calculating the base station energy storage scheduling threshold based on the dispatchable capacity and load prediction value,studying the 5G base station energy storage scheduling control strategy,building the energy storage scheduling model,and solving it with IMOPSO algorithm.Through case analysis,it is verified that the scheduling model and the solution algorithm can effectively reduce the peak-to-valley ratio of regional load and improve the energy storage revenue of base stations,thus realizing mutual benefits for both power grid companies and communication operators.(4)Designing the architecture of 5G base station energy storage management system and using Qt software to complete the interface design of 5G base station energy storage scheduling system,which realizes the "visualization" of energy storage management and reduces the difficulty of operation and maintenance for base station staff.