Design of Experiments Predictive Models as a Tool for Lifespan Prediction and Comparison for Enameled Wires Used in Low-Voltage Inverter-Fed Motors

Since the development of power electronic components, which allowed the manufacturing of reliable and efficient inverters, variable speed drives using inductive motors have become more and more popular. The pulsewidth modulation (PWM) technique has proven to be a very effective method of rotational speed control. However, the fast-changing voltage pulses, with very steep slopes (in the order of a few kilovolt/microsecond), have brought new hazards for the electrical insulation system of such motors. Very high frequency harmonic components of PWM voltage will result in significant overvoltage due to an impedance mismatch between the cable and the motor. As an effect, the voltage seen by some parts of the insulation system may exceed the partial discharge inception voltage stating the localized partial discharges activity. The aim of this article is to investigate and analyze the aging process of the enameled wire exposed to different factors and to propose a method allowing them to predict their lifespans in given conditions. This study introduces a prediction based on the design of experiments method and the statistical Weibull distribution. Thanks to the model obtained with short multistress aging tests, it is possible to predict the results of significantly longer ones. Moreover, the adopted methodology is proposed that allows us to predict the scatter of the long tests based on the short-time results dispersion. The same approach is used to compare different products between each other and rank them. All model predictions are compared with the experimental data in order to prove the model accuracy.