Volume, Issue 9, PP: - 9, FEB. WAVELET DE-NOISING AND ANALYSIS OF UHF PARTIAL DISCHARGES IN HIGH VOLTAGE POWER TRANSFORMER K V RAMPRASAD *. Professor, Dept of ECE, KALLAM HARANADHA REDDY INSTITUTE OF TECHNOLOGY, GUNTUR, AP, INDIA. Email: kvrp9@gmail.com Abstract- Partial Discharge (PD) in power transformers are known to be a common phenomenon. However, its value must not exceed a given standard limit. Most commonly occurring phenomena of PD relate to wedge type, float type and corona discharges. These need to be identified to determine the type of PD for controlling the peak magnitude. The method suggested in this paper is based on digital recording of signal from simulated model, FFT analysis of the signal and time ~ frequency resolution using Wavelet analysis. The Wavelet analysis is also helpful in reconstructing a given signal. This will be useful in eliminating any contribution due to spurious signal. Key Words: - Partial Discharge, corona, FFT, Wavelet, Transformer, UHF antenna. I. INTRODUCTION A power transformer is an important link in a power system and it has to endure the maximum system voltage and the total power transmitted. The insulation system is the key component of the transformer. Any failure in the insulation initially develops as a partial failure where the insulation can t withstand the local electrical stress. The failure leads to either a low energy discharge or a high current arcing. Such discharges are known as "Partial Discharges" (PD). The partial discharges slowly deteriorate the insulation by causing a local thermal breakdown. It is, therefore, necessary to measure PD in a power transformer right from manufacturing. However, in-situ PD measurement can t employ the factory method because of the high investment and electromagnetic interference due to site environment. There are a few non- electrical methods employed for this purpose. A PD is a natural phenomenon in high voltage equipment. These PD s can be generated due to several mechanism e.g. presence of floating metal particle, protrusion on the conductor, internal discharges in the paper or surface of insulation. As the transformer ages due to normal/abnormal overload or short circuit) operation, the deterioration occurs in the component. Hence, oil has to be changed or improved by oil filtration unit. Since the bulk of oil filled in a power transformer is very high, of order of hundreds of tons, the small metal particles enter the transformer, irrespective of the method of filtering. At times the shield material becomes a source of small metal particles. Further, bolted joints and end frames are also the potential source of metal particles. Thus in any transformer, the floating metal particles circulate and are available at various locations. While moving with oil, it keeps circulating along the winding, along the duct or along the core. They become one of the potential sources of PD. In addition, any high electric stress region exceeding local breakdown stress of oil gives rise to corona in terms of PD. In an actual transformer either individually or simultaneously all the phenomena may be occurring continuously[]. Thus PD is bound to occur during transformer operation within limit. For power transformer, the maximum level specified by Indian and International standard is pc (Pico Coulombs). Since the transformer is always a closed apparatus it is not possible to perceive which type of source (wedge, floating and corona) is acting in any transformer. Therefore a simulation experiment should be carried out on a modular basis. Each module should simulate one phenomenon at one instant. such PD signal in terms of ultra high frequency are recorded using appropriate instrument.the frequency analysis and Wavelet analysis of any signal may suggest the type of PD pattern. The method of detection and wavelet analysis for float type discharge has been reported by Sudarshanam et al. []. Not much of the work is known to have been reported in this area, where partial discharge is analyzed for its time domain component using Wavelet Transform (WT) technique []. II. PRESENT WORK The present work deals with model formulation, partial discharge simulation, high frequency measurement, and signal processing using Wavelet Transform. In order to carry out the analysis, the time domain signal of partial discharge is recorded on a digital oscilloscope. A typical signal is considered for Fourier Transform analysis. All dominant frequencies in the frequency response of the time domain signal are analyzed for their actual contribution in the signal using Gabor Wavelet. The contribution of each Vol. (9), ISSN: -9, FEB PP: - 9
Volume, Issue 9, PP: - 9, FEB. frequency in terms of its time signal can be used to predict the type of discharge occurring in a transformer. An additional work has been reported to reconstruct the signal by summing all time domain component values obtained from Wavelet Transform. The reconstructed signal is compared with the original signal after removal of noise present to evaluate the efficacy of Gabor Wavelet in obtaining component signals. The work also aims at reconstructing the signal to determine the effectiveness of Wavelet Transform in breaking in terms of accurate time domain signal [, ]. III. EXPERIMENTAL SETUP A system has been developed using an UHF antenna, suitable for partial Insulation structure known as Radome, RF amplifier, signal cables and measurement system. In order to simulate various PD phenomena, a tank filled with oil along with antenna is used. Various types of discharges have been simulated inside an oil medium and the signals due to the partial discharges are captured using the UHF antenna. The signals are analysed and chracterised for different types of discharges. The UHF antenna, is basically a duel arm Archimedean spiral type with an active frequency range of. - GHz. The nominal gain is db. A two stage RF amplifier is used to amplify the signals. The active frequency range of the amplifier is. to GHz with a single stage gain. The measuring system consists of bit, G samples/sec maximum sampling rate, GHz band width digital storage oscilloscope (DSO) IV. EXPERIMENTAL RESULT AND FFT ANALYSIS A few typical UHF signals generated from PD are shown from Fig.. to Fig.. The frequency responses of those measurements are shown in Fig. []....... -. -. -. -. -..........9 x - Fig.. A Typical PD signal of float type. x -. -. - -..........9 x - Fig.. A Typical PD signal of Corona type.. -. -. -. -..........9 x - Fig.. A Typical PD signal of wedge type..... x (-a) Vol. (9), ISSN: -9, FEB PP: - 9
Volume, Issue 9, PP: - 9, FEB. 9 In the present study, the Gabor wavelet equation has been used and is given by the following equation Ψ(t) = exp(-t /σ ) * Cos (t) () Where, σ is a constant and controls the band of frequencies to be identified. VI. APPLICATION OF WAVELET FOR ANALYSIS OF SIGNALS Fig.. Fig........ x.... (-b) (a) FFT of the UHF signal corresponding to Figure (b) FFT of the UHF signal corresponding to Figure..... x (-c) (c) FFT of the UHF signal corresponding to Figure V. METHOD OF WAVELET ANALYSIS The Wavelet transform W(b, a) of a function I(t) with respect to a given mother wavelet Ψ (t) is defined as [] W(b,a) = / a I(t) Ψ((t-b)/a) dt () - The aim of the present analysis is, to use wavelet technique to convert the signal from time domain to time ~ frequency domain and determine the contribution of each frequency in the main signal. This provides an insight into the dominant frequencies contributing to any type of discharge. In order to evaluate the time varying current waveform at a particular frequency, it is essential to determine the multiplication factor (k). This factor converts the wavelet transform of the time waveform at a particular scale parameter (a) into the transient current waveform for the corresponding frequency [] a = /πf The selection of σ value is based on the accuracy with which the output waveform (wavelet transform of the input waveform) at a particular scale parameter matches with the input/original waveform. Here the k value decreases with increase of σ and the value of k/f is constant for a particular value of σ. Fig.. and Fig.. Shows the Gabor Mother Wavelet for the frequency of MHz at σ = and respectively...... -. - - - - - Fig.. Gabor Mother Wavelet for σ = σ = a = /πf () Where, a = scale parameter b = translation parameter f = frequency Vol. (9), ISSN: -9, FEB PP: - 9
Volume, Issue 9, PP: - 9, FEB... σ =... MHz σ =... -.. -. -. -. -. -. -. -. - - - - - - Fig.. Gabor Mother Wavelet for σ = VII. RESULTS AND DISCUSSION Frequency spectrum characteristic of wedge type discharge depicted in Fig.. Does not adequately define the original signal in terms of contribution of each frequency. A dominant frequency may suggest a high magnitude with short duration or a low magnitude with long duration existence since the magnitude for a given frequency defines the area under the curve for magnitude ~ time curve of the component signal. However, the response aids in defining the presence of a particular signal. It is this characteristic that is utilized by Wavelet Transform to identify a particular type of signal. Typically from Fig. (a) to (d) show the time ~ magnitude value for all dominant frequencies. It is seen that higher contribution comes either from a few low frequencies (, MHz) component or a few very high frequency (, MHz) component. The remaining frequencies have very low contribution in the main signal. Such characteristic can be utilized for identifying the range of frequencies which contribute to a given type of discharge. A large databank of such measurement can be used to confirm the validity of the above hypothesis. Finally, all the component signals have been added with reference to time to obtain the reconstructed signal. The reconstructed signal is shown in Fig.. It is also seen that the Wavelet Transform can be used for eliminating any signal if it is considered noise. The component frequencies have been calculated and are shown in Fig.. (a) to (d) -..........9 x -.... -. -. -. (-a) -..........9 x - x - - - - (-b) MHz σ = MHz σ = -.........9 x - (-c) Vol. (9), ISSN: -9, FEB PP: - 9
Volume, Issue 9, PP: - 9, FEB..... -. -. -. -..........9 x - (-d) Figure. Component signal for various frequencies (a) MHz (b) MHz ( c) MHz (d) MHz -. -. -. -. -. -..........9 x - Fig.a. Reconstructed signal of wedge Type of PD signal.... MHz σ =..... x Fig. b. FFT of the reconstructed signal corresponding to Figure a The work reported in this paper deals with Partial Discharges in Power Transformers on the basis of experimental models. Various types of signals e.g. wedge type, float type and corona discharges have been recorded using UHF antenna & analyzed using FFT as well as Gabor Wavelet.The analysis suggests that time ~ frequency response can be used to identify the type of fault on the basis of frequency and contribution of an individual frequency in time domain. It is found that the lower frequencies (MHz, MHz) and high frequencies (MHz, MHz) contribute most to wedge type discharge. Medium frequencies have insignificant contribution in over all signals. The additional work reported in the paper deals with successful splitting of the signal in time domain and reconstruction of the signal. The analysis can be effectively used for eliminating any unwanted signal in the main. IX. REFERENCES [] R.V. Maheswari, P. Subburaj, B. Vigneshwaran and M. Willjuice Iruthayarajan Partial Discharge Signal Denoising using Adaptive Translation Invariant Wavelet Transform-Online Measurement, J Electr Eng Technol Vol.9, No.: 9-,. [] T.Sudarshanam, K.V.Ramprasad, H.S.N. Murthy, A. Govardhan and B.P. Singh, Application of Gabor wavelet for analysis of partial discharge signals in high voltage power equipment, in Proc. Intern. Conf. CCPE, Paper Id:, Chennai, July -9,. [] M. Rajeshwara Rao, B.P. Singh, Using Wavelet for the Detection and Localization of Interturn fault in the High Voltage Winding of a Power Transformer, IEEE Trans on Dielectrics and Electrical Insulation, Vol., pp., No., August [] N.R. Deshpande, S.V. Kulkarni, V.M. Gadre, and S.A. Khaparde, Recent trends in applications of Wavelet Transform to Power System Engineering, in Proc. th North American Power Symp., Aug.. [] I. Daubechies, Ten Lectures on Wavelets, nd Edition, SIAM, Philadelphia, 99. [] Digital Signal Processing, Fourth Edition (John G. Proakis, G. Manolakis) Prentice-Hall of India, New Delhi. [] Robi Polikar, The Wavelet Tutorial Part-I, II and III Multiresolution Analysis, the Continuous Wavelet Transform, Second Edn, Ames, Iowa, 99. [] M.M.Rao M.J. Thomas, and B.P.Singh Frequency Characteristics of Very Fast Transient Currents (VFTC) in a kv GIS, IEEE Trans. Power Del., vol., no., pp.-, Oct.. VIII. CONCLUSIONS Vol. (9), ISSN: -9, FEB PP: - 9