基于LibSVM代用燃料有效功率增量预测方法的研究  

Research on the Method of Effective Power Increase of the Alternative Fuel Forecast Based on LibSVM

作　　者：朱培根[1] 梅卫江[1] 石秀锋[1] 边金英[1] 

机构地区：[1]石河子大学机械电气工程学院,石河子832003

出　　处：《石河子大学学报（自然科学版）》2012年第5期657-660,共4页Journal of Shihezi University(Natural Science)

基　　金：石河子大学重大科技攻关计划项目(GXJS2008-ZDGG03-01)

摘　　要：由于代用燃料燃烧系统的复杂性和非线性,对发动机有效功率有显著影响,而有效功率增量能有效反应发动机的动力性能。本文在归纳总结支持向量机的基本原理的基础上,采用正交试验,借助LibSVM软件包,利用支持向量机回归预测模型对燃烧系统的有效功率增量进行预测,经计算得出MSE=0.00212692,R=0.998038,表明支持向量机对小样本、非线性和高维回归预测有较高的预测精度、有效性及可行性。The complexity and nonlinearity of the alternative fuel combustion system have a significant influence on effective power of engines, and the effective power increase can effectively respond to the power performance of the engines. With the aid of the LibSVM software package, this paper applies support vector machine regression prediction model to the predicting the ef- fective power increase of the combustion system by orthogonal experiments. The calculated result is：MSE= O. 00212692,R= 0. 998038,indicating that support vector machine has a relative high prediction precision, validity and the feasibility for small sample,nonlinear and higher dimensional regression prediction.

关 键 词：LIBSVM 燃料 有效功率增量 预测 支持向量机 

分 类 号：TK428.9[动力工程及工程热物理—动力机械及工程]