HA.01燃气-蒸汽联合循环机组共振故障自动监测方法

HA.01 Automatic Monitoring Method for Resonance Fault of Gas Steam Combined Cycle Units

为提高HA.01燃气-蒸汽联合循环机组共振故障自动监测能力,提出了基于广义S变换和IBA-SVM的HA.01燃气-蒸汽联合循环机组共振故障自动监测方法,。在HA.01燃气-蒸汽联合循环机组安装多个传感器采集机组振动信号,利用广义S变换转换采集的原始振动信号,提取共振故障特征向量,并组成特征向量集合,通过IBA算法优化SVM参数,并将机组共振故障特征向量集作为IBA-SVM的机组共振故障自动监测模型的输入,经过模型训练,输出训练结果,完成机组共振故障自动监测。实验结果表明:该方法在迭代次数为60次以后,最佳适应度为98%并趋于稳定,平均适应度平均适应度高于70%,鲁棒性较好,同时机组共振故障的自动监测准确率高达98.9%,分类效果最佳,提高对机组共振故障的自动监测能力。

Research on the automatic monitoring method for resonance faults of HA.01 gas-steam combined cycle units based on generalized S-transform and IBA-SVM,the automatic monitoring ability of resonance faults of HA.01 gas-steam combined cycle units are improved.Install multiple sensors in the HA.01 gas-steam combined cycle unit to collect unit vibration signals,use the generalized S-transform to transform the collected original vibration signals,extract resonance fault eigenvectors,and form a set of eigenvectors.Optimize SVM parameters through the IBA algorithm,and use the set of common vibration fault eigenvectors as the input of the IBA-SVM unit resonance fault automatic monitoring model.After model training,output the training results,Complete automatic monitoring of unit resonance faults.The experimental results show that after 60 iterations,the optimal fitness of this method is 98% and tends to be stable,with an average fitness of over 70% and good robustness.At the same time,the accuracy of automatic monitoring of unit resonance faults is as high as 98.9%,with the best classification effect,improving the automatic monitoring ability for unit resonance faults.