R.G. van Heeswijk

Parameter Estimation for the Weibull Distribution

The time to electric breakdown, and the electric field necessary to result in breakdown of solid insulation, seem to be best represented by a Weibull probability distribution. This tutorial paper reviews the graphical method of estimating the parameters of the Weibull distribution. It also presents the method of Maximum Likelihood as an objective technique, developed by statisticians for improving the estimates of the Weibull parameters, especially for censored experiments. Procedures for calculating confidence intervals are also given. It is shown in an example that the confidence intervals can be very wide for sample sizes commonly used in dielectric life tests.

Investigation of the effect of repetitive voltage surges on epoxy insulation

The aging of epoxy electrical insulation by unipolar, repetitive voltage surges has been investigated to determine whether such surges can precipitate premature failure of high-voltage motor windings and gas insulated switchgear spacers. A full factorial statistical test was devised and performed on 180 pure epoxy insulation specimens to examine the effect of voltage magnitude, polarity, and surge repetition rate on the life of unfilled epoxy specimens containing a needle-plane electrode geometry. All factors and interactions were determined to have a significant effect on the life, and life prediction equations were estimated. Increasing the voltage magnitude or repetition rate, decreased or increased the life, respectively. Aging was found to occur at as low a stress as 2 MV/cm. Gradual aging of the epoxy does occur under repetitive voltage surges. Based on a realistic number of surges with typical voltage amplitudes which can normally occur from vacuum switchgear during service, gradual deterioration of the epoxy turn insulation in motor stator windings is possible in some situations. Situations where surge aging may occur are identified. >

Electrical aging and electroluminescence in epoxy under repetitive voltage surges

An extensive statistical accelerated aging experiment, involving a total of 180 unfilled type epoxy specimens, has been conducted to determine the number of voltage surges required to initiate electrical trees under various magnitudes, surge repetition rates, and polarity conditions. The specimens were tested one at a time according to a full factorial experimental design. During many of the aging tests, the electroluminescence (EL) from the epoxy specimens was monitored. Conventional Weibull analysis techniques were used to investigate the experimental results. The analysis showed that life decreased with increasing applied voltage and was longer for negative polarity surges. As the surge repetition rate increased, the number of surges to initiate a tree increased. A modified inverse power model, with a voltage power law constant similar to that found in AC aging was determined to be appropriate. No correlation was observed between either the trend in EL over time or the total amount of EL and the number of surges required to initiate a tree. >

Lifting Field of Free Conducting Particles in Compressed SF6 with Dielectric Coated Electrodes

The lifting field of a free conducting particle (FCP) situated on a dielectric coated electrode in compressed SF6 gas is calculated based on two different particle charging mechanisms-partial discharges occurring in the gas surrounding the particle and charge conduction through the dielectric coating. Calculated values are compared with experimental results obtained for different thicknesses of anodized (Al203) coatings and at different gas pressures. It is suggested that at lower pressures a partial discharge mechanism is predominant whereas at higher pressures a charge conduction mechanism is predominant.

The Statistical Analysis of a High Voltage Endurance Test on an Epoxy

This paper is concerned with the accelerated aging of solid insulating materials, as a means of estimating their service lives and identifying superior dielectrics. The assumed failure mechanism is electric treeing. A constant-stress accelerated aging experiment on a transparent epoxy at three voltage levels is described. All factors thought to be important in the electric treeing mechanism were rigidly n-controlled, and the inception and growth intervals were measured separately. The results indicate that the times to inception and failure are three-parameter Weibull distributed, whereas the growth period may be either Lognormal or three-parameter Weibull distributed. The Inverse Power Law could not be n-confirmed. There was little correlation between the inception and growth intervals. Statistical analysis showed that there is much uncertainty in these results, and this is true for most published accelerated aging tests done to date. The general applicability of accelerated aging tests is questioned.

Modeling of the dynamic performance of transient recorders used for high voltage impulse tests

Digital recorders are becoming more commonly used in the recording of steep-front HV impulses. Although many digitizers exhibit a nonlinear deterioration in dynamic performance with increasing input signal steepness, there is no accepted method for predicting the magnitude of the resulting errors. The authors describe the development and implementation of a computer modeling technique for establishing the magnitude of these errors. The model is based on time-domain test data combined with a basic understanding of the digitizers operating principles. The model is general in nature and can be used for the prediction of errors generated by any recorder that has systematic errors. The model was used to predict the maximum error which can occur when a particular EBS (electron-bombarded semiconductor) recorder is used to monitor HV impulse tests on power apparatus. Its use allowed for the determination of error limits for records of steep-front impulse tests which will be standardized in the near future. >

Electroluminescence in epoxy insulation

Electroluminescence (EL) in transparent epoxy subjected to repetitive voltage surges has been investigated. The effect of surge voltage magnitude, repetition rate, and polarity on the magnitude and number of light pulses emitted was studied. Using a needle-plane electrode geometry, the light inception stress was determined to be approximately 200 kV/mm for both positive and negative voltage surges. As the peak voltage of a surge increased, the number of light pulses increased exponentially. More light pulses were emitted under negative surges than under positive ones, and more were detected on the leading edge than the trailing edge of a negative surge, whereas the reverse occurred under a positive surge. As the surge repetition rate was increased, the amount of EL per surge decreased. Finally, there was no change of the spectral composition of EL with either polarity or magnitude of the applied surge. The results suggest that the detected EL may be due to single photons. The observed EL response is attributed to the creation and depletion of trapped space charges. >

The effect of a metallic particle near a spacer on flashover phenomena in SF/sub 6/

A study of the flashover mechanism initiated by metallic particles in SF/sub 6/ gas is presented. Using a high-speed Imacon photographic camera, it was established that a metallic particle near the insulator surface plays an important role in flashover development. The influence of the particle on flashover development depends on the pressure of the SF/sub 6/ gas. At atmospheric pressure, the flashover was always initiated at a point on the particle farthest from the nearest electrode. At high pressures, it was found that breakdown develops from the electrode of relative positive polarity in the form of a leader. The velocity of the discharge crossing the gap increases exponentially when the distance between the particle and the insulator surface decreases. >

Spacer Involvement in Conducting-Particle-Initiated Breakdown in Compressed Gas Insulated Systems

This paper describes the results of an investigation into conducting particle initiated breakdown in a 76/250 mm coaxial electrode configuration, representative of a 145 kV gas insulated transmission line (GITL) or gas insulated substation. A free moving particle is known to initiate breakdown through a sequential two-stage process. The involvement of an epoxy disk spacer in such a breakdown, when the particle is at a distance from the spacer, is studied through specially designed experiments. For 6.4 mm long, 0.45 mm diameter aluminum particles it is observed that the discharge path may involve the spacer surface even when the particle is several centimeters away from the spacer surface. However, the 60 Hz ac breakdown voltages for discharge paths involving the spacer surface are not significantly different from those without spacer involvement, if the particle-to-spacer distance is greater than 3 mm.

Free particle-initiated 60 Hz breakdown at a spacer surface in a gas-insulated bus

The influence of freely moving wire particles on 60-Hz AC surface flashover of epoxy and acrylic disk spacers in a 76/250 mm coaxial system was examined in sulfur hexafluoride gas at pressures of 300 and 500 kPa. The experimental results show that 0.45-mm-diameter aluminium wire particles, 6.4 mm long, will lift and reach a nearby spacer at voltages comparable to system operating voltage. The combination of metallic particles and spacer does not reduce the breakdown voltage below that for free particles without a spacer unless a particle is lodged on the spacer very close ( >