Features Characterizing the Surface State of HV Insulator Glass Model under Desert Pollution

This paper deals with the monitoring of the surface state of HV insulator model under discontinuous layer of pollution. Natural desert sand is used to simulate the pollution influence in Algerian Sahara regions. Experimental tests are carried out on a glass plane model of 1512L cap-pin insulator largely used in the said regions. Leakage current LC signal is recorded for different positions and widths of the polluted-band. First, the obtained results are discussed to describe the behavior of the insulator model. For each configuration, time-frequency decomposition is established using Discrete Wavelet Transform (DWT) for information extraction corresponding to the flashover process. Decomposition results indicate that the detail D3 is highly correlated with the measured LC signal. Recurrence Plot (RP) technique and Recurrence Quantification Analysis (RQA) are applied to quantify the temporal patterning in LC signal, giving information on the position and width of the polluted-band on the insulating surface. It was found that RQA indicators increase with the polluted-band width according to the position of this polluted-band. Based on the plan model, the findings demonstrate the capability of RQA indicators for the glass insulator monitoring, providing information on the width and position of the polluted-band on its surface.