Probing the uSE Of Green Insulating Oils in transformers based on their statistical breakdown data

The growing demand for improved fire safety, biodegradability and life extension of transformers and other oil filled high voltage-equipment has prompted the research and development efforts to produce ¿green insulating oil¿ based on natural esters. In the present investigation three seed based oils (corn, canola and palm olein oils) were compared with conventional mineral insulating oil by evaluating their breakdown characteristics. Results show that corn and canola oils exhibit much higher 60 Hz breakdown voltage than mineral oil. To evaluate compatibility of vegetable oils with existing transformer design it is essential not only to compare the mean breakdown voltages but the dispersion in the breakdown data has also to be taken into account to determine the minimum withstand voltage level. However, there exists controversy on the type of statistical model on which these data best fit. In this paper, breakdown voltage data of vegetable oils and mineral oil were subjected to Weibull and Gaussian models of statistical distributions and both were found to fit better on the Weibull model, however larger dispersion was shown in vegetable oils in the lower (¿ 5%) voltage breakdown probability range as compared to mineral oil. For a better statistical inference, a computer based Monte Carlo simulation was devised and used to precisely discriminate the two models. Analysis of the simulation results further confirm that the data of both types of insulating oils fit better on the Weibull model.