Intelligent Fault Diagnosis of Three-Trans Towers’ Bolts Based on EMD and LS-SVM

The looseness of bolts is one of the main reasons for the tower collapse of three-trans line. The vibration data are difficult to obtain and the collected signal is strongly nonlinear and non-stationary. Aiming at realize the accurate detection of the bolt status and ensure the stable operation of power transmission lines, an intelligent status detection method based on empirical mode decomposition (EMD) and least square support vector machine (LS-SVM) was put forward. The original vibration signal was decomposed into several intrinsic mode function (IMF) components based on EMD, different bolt status with the different frequency band energy of IMF. The frequency band energy of the first four IMFs were selected as the feature vectors, then these feature vectors were input into LS-SVM classifier for the recognition of the running status of bolts. Test results show that the proposed method has certain effectiveness and can accurately identify the different status with a small number of samples.