民航机场用电短期能耗优化预测仿真  

Simulation of Short-Term Energy Consumption Optimization for Civil Aviation Airport

作　　者：郭晓静 张娇婷 万雅萍 GUO Xiao-jing;ZHANG Jiao-ting;WAN Ya-ping(School of Electronic Information and Automation,Civil Aviation University of China,Tianjin 300300,China;Quality Control Room of China Southern Airlines Shanghai Branch Aircraft Repair Shop,Shanghai 201103,China)

机构地区：[1]中国民航大学电子信息与自动化学院,天津300300 [2]南航上海分公司飞机维修厂质量管理室,上海201103

出　　处：《计算机仿真》2018年第9期31-36,共6页Computer Simulation

基　　金：国家自然科学基金委员会与中国民用航空局联合资助项目(U1333111);民航局科技创新引导资金项目(应用技术研究类)(20150227);中央高校基本科研业务费中国民航大学专项基金(3122013D021)

摘　　要：由于民航机场电能耗数据具有量少、非线性的特点,导致机场能耗预测模型建立困难。传统算法通过建立灰色神经网络模型进行机场用电短期能耗预测,但由于神经网络模型随机初始化权值以及可利用的数据不足依然带来了很大误差。为提高民航机场用电短期能耗模型预测精度,提出了改进的灰色深度信念网组合预测模型。首先利用机场电能耗历史数据建立改进的灰色预测模型;然后将灰色预测结果、多维历史用电量数据特征和影响机场电能耗的主要因素共同作为深度信念网的输入,确定网络结构;最后基于此预测模型对某机场东区用电量数据进行预测,并与传统的灰色神经网络预测结果进行对比,仿真结果表明,所提出的结合了灰色模型的深度学习网组合模型在机场用电量短期预测中具有较高的精度。Due to the characteristics that the data of airport energy consumption are less and nonlinear, it is diffi- cult to establish an airport energy consumption forecasting model. The traditional algorithm uses the grey neural net-work model to predict the short-term energy consumption of airport power, but it still brings a lot of errors caused by random initialization weights of traditional grey neural network and the insufficiency of the available data. In order to improve the prediction accuracy of short-term energy consumption model of civil aviation airport, an improved grey depth belief network combination forecasting model was proposed. Firstly, the gray prediction model was established by using the historical data of airport energy consumption. Then put the grey prediction results, the multidimensional historical data and the main factors influencing the energy consumption at the airport were taken as the inputs of the deep belief network to determine the network structure. Finally, this model was used to predict the electricity consumption data in a certain airport, and the results were compared with the traditional gray neural network prediction results. The simulation results show that the proposed forecasting model of gray depth belief network has higher pre- diction accuracy in short-term electricity at the airport.

关 键 词：灰色模型 深度信念网 对比散度算法 反向传播算法 能耗预测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]