基于支持向量机的船舶柴油机故障诊断的研究

通过建立VC维统计学习理论,利用数学建模,建立并划分最优超平面以取得支持向量来训练,选取并考虑相关的影响因子以此构建样本数据集,以期对柴油机故障实现精确的诊断;而对于非线性空间情况,可采用核函数的思想来转化为线性空间,以此降低算法的复杂度;通过与人工神经网络方法的比较,表明该方法具有运算速度快,泛化能力强等优点;支持向量机(SVM)可以较好解决柴油机故障诊断中的机器过学习、小样本、高维数、非线性等问题。

Feasible Research on Fault Diagnosis of Marine Diesel Engine Based on Support Vector Machine

A new machine learning method called SVM which will be applied in the fault diagnosis for marine diesel and its feasibility are introduced in the paper.By establishing VC dimension statistical learning theory and mathematical modeling,an optimized hyperplane can be set up so as to obtain support vector to train sample data,related influential factors can be chosen and considered to construct sample data table for realizing precise fault diagnosis.As for non-linear space,it is effective to be converted to a linear space by adopting kernel function for minimizing complexity of the algorithm.By comparison to the ANN method,SVM has the advantages of strong generalization and fast operation.The SVM can be the good solution for over-learning,small sample,high dimension and non-linear problem etc.