EER Parameters Identification Of Communication Base Station Air-condition Based On Particle Swarm Optimization

Communication industry becomes the major promoter of energysaving and emission reduction of the whole society because of its pillarindustry characteristic, within which base stations are the keyenergy-saving unit because of its huge energy consumption. Some energysaving technologies of base stations have been used in the energy savingof base stations such as intelligent soft turn-off, ventilation and heatexchanging mechanism and new energy and materials for energy-saving.But the evaluation of energy-saving efficiency is still heavily dependenton manual off-line analysis methods, which has become the bottleneck ofthe overall advancement of the energy saving technologies ofcommunication base stations. Therefore, the calculation method ofreal-time energy saving based on energy efficiency ratio (EER) hasbecome a hot research topic.Firstly, the current situation and strategies of energy saving forcommunication base stations are analyzed, followed by the review of the technical methods and assessment principles. Secondly, the defects ofmanual off-line evolution methods on energy saving efficiency forcommunication base stations are analyzed objectively, then the real-time,online and automatic evolution methods is pointed out as thedevelopment trend. Finally, a method based on Particle SwarmOptimization (PSO) to identify parameters on the EER model ofair-conditions is put forward. An adaptive hill-climbing particle swarmoptimization based on population entropy is presented in this paper,considering the problems of local optimum and slow convergence speedat later time caused by Particle Swarm Optimization (PSO). Then ahybrid of Particle Swarm Optimization and wavelet neural network basedon nonlinear conjugate gradient method is presented, which uses waveletneural network to fit the parameters of the model and an improved PSO tooptimize the weights of wavelet neural network. As a result, theeffectiveness and accuracy of the EER model is raised.On the basis of algorithm research, part-time energy consumptiondata and related parameters are collected from one communication operator’s12test base stations equipped with fresh air energy savingsystems and battery constant temperature cabinet. The model gotten bytraditional least squares is compared with that of the improved PSO, andthe results shows that the latter method performs much more accuratelythan the former one in the calculation of base station benchmark energyconsumption. Therefore, the improved PSO in this paper has high validityand feasibility in the application of system parameter identification.