摘要
提出了一种EMD与SVM的组合预测模型,对通用航空飞机燃油流量进行预测。首先对数据缺失值与异常值进行处理,应用经验模态分解算法对燃油流量数据进行分解,得到各分量IMF,然后采用支持向量机对每一个分量进行预测。在预测过程中,采用PSO算法对支持向量机的参数进行优化,最后叠加各分量得到预测数据。采用通航飞机实际飞行数据进行验证,结果表明该组合模型可以有效地预测燃油流量,准确率较高,其MSE可以达到0.254,高于传统的单一预测模型。
A combined prediction model based on EMD and SVM is proposed to predict the fuel flow of general aviation aircraft.First,the data missing values and outliers are delt with,the fuel flow data are decomposed to obtain each component IMF by using empirical mode decomposition algorithm.Then,SVM is adopted to predict for each IMF.In the process of prediction,PSO algorithm is proposed to optimize the parameters of SVM,and the superposition of all components is forecast.Finally,the prediction data is obtained by superimposing each component.The results show that the combined model can effectively predict fuel flow with high accuracy and the MSE can reach 0.254,which is higher than the traditional single prediction model.
作者
马玉猛
MA Yu-meng(College of Aeronautical Engineering,Binzhou University;Engineering Research Center of Aeronautical Materials and Devices,Binzhou University;Key laboratory of Aeronautical Optoelectronic Materials and Devices,Binzhou University,Binzhou 256603,China)
出处
《滨州学院学报》
2022年第4期20-24,共5页
Journal of Binzhou University
基金
滨州学院实验技术研究项目(BZXYSYXM201803)
滨州学院科研基金项目(BZXYQNLG201704)。
关键词
燃油流量
预测
经验模态分解
粒子群算法
支持向量机
fuel flow
prediction
empirical model decomposition
particle swarm optimization algorithm
support vector machine