An internet of thing monitoring system with high-precision time reference for transmission lines and the fault recognition method

An Internet of Things (IoT) monitoring system with a high-precision time reference is developed in this research. The terminals of the system power themselves through self-induction from the high-voltage transmission lines. The IoT system is dispatched along the transmission lines and can monitor the running status, including the temperature and humidity of the environment, the rms line current, the lines galloping, and the ground height. Besides these parameters, the system can detect the fault current in time and record the fault current waveform. The time synchronization accuracy reaches 100 ns, which guarantees the fault recognition and location along the whole transmission lines. To accurately recognize the fault type, a fault recognition method based on time-domain statistical features and support-vector machine (SVM) is proposed. First, the signal is processed with an improved wavelet denoising algorithm. Next, the temporal feature is extracted with statistical parameters in the time domain. Finally, the fault pattern is recognized by a multi-class SVM model. Experimental tests show that the terminal of the system has a positioning accuracy of centimetre level and can effectively monitor the galloping for transmission lines and can detect and record the abnormal current waveforms for the fault type recognition. Simulation experiments show that the proposed recognition method have better recognition performance.