Abstract--Leakage current frequency components are frequently used as a tool for surface condition monitoring on polymeric insulation material and their pollution severity. Fast Fourier Transform (FFT) is one of the methods that are applied for the analysis, but it has some limitation in non-stationary signal. This paper presents analysis of leakage current on polymeric insulation material for high voltage application in frequency domain and time-frequency representation. Tracking and erosion test (Inclined Plane Test (IPT)) complying with BS EN 60587-2007 is conducted on polymeric insulation to select a set of different leakage current patterns from capacitive, resistive and discharge. Then, the leakage current patterns are analysed in frequency domain and time-frequency representation using Fast Fourier Transform and spectrogram technique, respectively. It is found that the surface condition of polymeric insulation material state can be classified accurately by using spectrogram compared to Fast Fourier Transform. Index Terms--Inclined-Plane Test, Leakage Current, Fast Fourier Transform, Time-Frequency Representation, spectrogram I. INTRODUCTION N high voltage engineering or its applications, insulation is the most important part to prevent the flow of current to undesired paths. The dielectric strength and electrical field are a key properties of insulating material and they become the major factors that determine the failure of the insulation[1]. It is being widely accepted by many power utilities worldwide, including Malaysia. It is being widely accepted by many power utilities worldwide, including Malaysia. There are many advantages of polymeric composite material over conventional ceramic material with a detailed comparison between ceramic and polymer electrical insulation includes as explained in [2]. A good insulation system give better design, performance and life span of the electrical apparatus. Even though, it has been used widely, there are still many issues such as ageing performance, expected lifetime and their long-term reliability are not known and therefore are a concern to users. To overcome these, a large number of studies and research activities for improvement on their performance have been made. These include the development of new materials, the understanding of deterioration of chemical, electrical and mechanical over the stress, design and manufacturing process of material, and also development of practical testing, monitoring, reliability methods of measuring and service performance. This paper describes the reliability method to measure loss of hydrophobicity on polymeric insulation, which leads to tracking and erosion as well as to flashover under contaminated condition. One of the key indicators widely accepted to measuring the surface condition of polymeric insulation either in service or accelerated aging laboratory test is by investigating its leakage curent (LC) signal. For LC performance factor that should be controlled is flashover mechanism because of dry band formation on the surface. Basically, LC signal that leads to surface flashover can be divided into three events state [2] which are the early aging period (EAP), transition period (TP), and late aging period (LAP). They consists of capacitive current, resistive current, non- linear current and non-linear with discharge current [3]. When the sample is completely dry or in hydrophobic state, it is classified as EAP and only low capacitive LC flowed. In the second state, TP, the polymer insulation surface became hydrophilic and LC will change significantly from capacitive to resistively. Finally, during the LAP, LC is completely resistive and surface discharge took place. Measurement of the LC such as amplitude, pulse, accumulated charge and discharge duration has been used to provide information on degradation [2]. But onwards literatures show, LC harmonic component analysis give better information [2, 4-6]. G.P Bruce and S.M Rowland [7], Suwarno et al [8], A.H El Hag et al [9-11] and Hussein Ahmad [12] examined low harmonic components of LC as a diagnostic tool to study aging and surface condition. A correlation was found that frequency component correspond to surface discharge event as well as high pollution severity. In most of previous works, LC is normally analyzed in frequency domains, and the fast Fourier transform (FFT) is used for transformation technique. However, LC signals are usually in non-stationary pattern especially during dry-band condition and surface discharge. Thus, this technique is only suitable for stationary signals and does not provide temporal information [13]. Any abrupt signals localized in time require more than FFT. Another reason was reported by C.Muniraj and S.Chandrasekar [14] who found that the analysis of LC in polluted polymer insulator shows that FFT is fast in computation but possess limitations in resolution. In time or frequency domain plot, it just gives half of the information about the signal. A frequency domain plot will tell "what" the Leakage Current Analysis on Polymeric Surface Condition using Time-Frequency Distribution N. Q. Zainal Abidin, A. R. Abdullah, N. Norddin, A. Aman, and K. A. Ibrahim Faculty of Electrical Engineering Universiti Teknikal Malaysia Melaka (UTeM) I 2012 IEEE International Power Engineering and Optimization Conference (PEOCO2012), Melaka, Malaysia: 6-7 June 2012 978-1-4673-0662-1/12/$31.00 ©2012 IEEE 171