Study On The Modeling Method For Base Station Energy Consumption Based On Data-Driven

Base station is an important unit of Energy consumption in communication network node,accurate calculation of EPC(Energy Performance Contracting)mode of the base station energy saving becomes a bottleneck in the field of technology.Currently,the existing calculation methods of the artificial energy saving(such as contrast testing method,start-stop cycle test method and base station benchmarking method)can't meet the engineering requirements that the energy saving is calculated accurately in the Energy Performance Contracting mode because of its large error,poor real-time performance,high operation cost.So,studying relatively accurate calculation method of communication base station energy saving online and establishing a precise calculation of base station energy consumption model have obvious practical significance and application prospects on the implementation evaluation of base station energy-saving projects.Based on the comprehensive analysis of the inlfuence between static parameters(base station environment characteristics,the area of base station,building material,etc.)and the dynamic data(the total energy consumption,the key equipment energy consumption,air-condition energy consumption,indoor and outdoor temperature,etc.)of communication base station and energy consumption,this paper gets the main factor influencing the energy consumption of communication base station from the multi-dimensional relationships among time,space,energy consumption,environmental parameters,etc.And the paper puts forward the base station energy consumption model based on the Sliding Window CPSO-LSSVM modeling algorithm.Then,the chaotic particle swarm collaborative optimization algorithm is used to optimize least squares support vector machine model parameters and compared with some other optimization algorithms,the result verifies that it has a better ability of optimizing parameters.On this basis,the Sliding Window is used to establish dynamic model of energy consumption witch updates automatically as the new Sliding Window forms.Comparing several modeling algorithms,experimental results show that the algorithm proposed in this paper has better generalization ability;the result of experiment which is established based on the data of base station also shows that the predictive value based on the base station energy consumption model fits the actual energy consumption data.So the model can be used for early warning and forecast in base station energy consumption management,also it be used to provide technical support for base station energy saving assessment in EPC mode.