CAJ | 학술논문

针对柴油机的振动信号蕴含大量状态信息难以监测的问题，文章提出了利用小波包分解提取振动信号各频段内的能量值作为特征矩阵输入到支持向量机中进行模式识别与故障诊断。对于SVM参数选取困难及其对分类效果的影响，提出利用粒子群优化算法选取惩罚因子和径向基核函数参数，让SVM模型实现最佳分类。结果表明：该方法可以有效提高分类效率和准确率，从而实现对柴油机的故障诊断。
침대시유궤적진동신호온함대량상태신식난이감측적문제，문장제출료이용소파포분해제취진동신호각빈단내적능량치작위특정구진수입도지지향량궤중진행모식식별여고장진단。대우SVM삼수선취곤난급기대분류효과적영향，제출이용입자군우화산법선취징벌인자화경향기핵함수삼수，양SVM모형실현최가분류。결과표명：해방법가이유효제고분류효솔화준학솔，종이실현대시유궤적고장진단。
The vibration signal of the diesel contains plenty of state information, put forward the wavelet packet decomposition to extract the frequency vibration signal energy value as the characteristic matrix ,then put it into support vector machines for pattern recognition and fault diagnosis . For the SVM parameter selec-tion is difficult and have impactions on the classification results, proposing the particle swarm optimization algorithm to select the penalty factor and the radial basis kernel function parameter, let SVM model to a-chieve the best classification. The results show that: this method can effectively improve the classification accuracy and efficiency, so as to realize the fault diagnosis of diesel engine.