LS-SVM在烟气轮机振动故障诊断中应用研究

烟气轮机机组是利用余热发电原理回收高温热能再生电能的装置,由于烟气轮机机组的故障现象对企业经济和安全生产造成了很大的影响,所以准确判断机组故障发生点和降低故障现象具有十分重要的现实意义。通过对原始信号的三层小波分解提取信号的特征向量,再采用LS-SVM不同核函数及其对应不同参数的选择与实验进行分类研究,得到RBF核函数的分类效果最佳。

Research of Application of LS-SVM on Vibrancy Fault Diagnosis in the Gas Turbine

The use of gas turbine unit to achieve the principle of high-temperature heat recovery renewable energy,but because breaks down the failure of gas turbine unit,the production of economic and security caused huge losses.Thus failure timely to determine the suspicious points and to reduce the frequency of failure,while protecting the environment is of great practical significance.Research through the three layers of the original signal wavelet decomposition to extract the signal feature vector,and then use LS-SVM methods of choices and experiments of different kernel functions and corresponding parameters to classify,RBF kernel function classification is best.