基于GA-SVM的城市天然气中长期负荷预测

Study on Urban Natural Gas Load Middle-Long Term Forecasting Base on GA-SVM Model

随着天然气工业的发展,城市面临着储气设施建设、燃气管网规划等一系列问题。分析城市天然气中长期负荷预测影响指标,从内部环境、外部环境及用户消费3个角度出发,提取其中9个参数作为城市中长期负荷的影响因子,并采用遗传算法优化支持向量机的惩罚因子及核参数,建立了基于GA(遗传算法)-SVM(支持向量机)的城市天然气中长期负荷预测模型。利用该模型对北京燃气用气量进行中长期负荷预测,并与SVM回归预测相比。结果表明:GA-SVM模型有比较快速训练速度、较高的预测精度,所提出的GA-SVM优化模型在中长期天然气负荷预测上具有更优的泛化能力和学习能力,能够为城市燃气管网规划、储气设施建设等提供指导作用。

With the development of natural gas industry,cities are facing a series of problems,such as the construction of gas storage facilities,gas pipe network planning and so on.This paper analyzes the impact indicators of urban natural gas medium and long-term load forecasting,and extracts nine parameters as the impact factors of urban medium and long-term load from three aspects of internal environment,external environment and user consumption,and uses genetic algorithm to optimize the penalty factor and kernel parameter of support vector machine,A medium and long term load forecasting model of urban natural gas based on GA(genetic algorithm)-SVM(support vector machine)is established.The model is used to predict the medium and long-term load of gas consumption in Beijing and Shenyang,and compared with SVM regression prediction.The results show that GA-SVM model has fast training speed and high prediction accuracy,and the GA-SVM optimization model has better generalization ability and learning ability in medium and long term natural gas load forecasting,which can provide guidance for urban gas network planning and gas storage facilities construction.