V参数SVM在烟气轮机故障诊断中的应用

烟气轮机机组是利用余热发电的原理回收高温热能再生电能,但是由于多方面的原因,烟气轮机机组时常发生故障,对生产的经济性和安全性造成巨大损失。为了保障设备正常运行,对设备进行实时监测,及时判断故障发生的可疑点,减少故障发生的频率,同时保护环境,采用对原始信号的3层小波分解,提取信号的特征向量,再采用C-SVC和V参数支持向量机二种方法进行分类实验,得到V参数的支持向量机分类效果最佳。

Application of V parameter of SVM for fault diagnosis in gas turbine

The gas turbine unit usually uses the principle of high-temperature heat recovery renewable energy,but due to various reasons,the failure of gas turbine unit causes huge losses for economical efficiency and security of production.To ensure normal operation,a real-time monitoring was performed,the suspicious points were determined to reduce the frequency of failure and protect the environment.The signal feature vector was extracted through decomposing three layers wavelet of the original signal,a classified experiment was performed with C-SVC and V parameters,the best V parameters of support vector machine classification was obtained.