A Four Factor Lifetime Estimation Model of Polyimide (PI) Insulations Using Full Factorial Design of Experiment Method (DOE)

The high-frequency power transformer (HFPT) insulation system is subjected to significant electrical and thermal stresses, which makes it challenging to characterize the interplay between physical factors that affect the aging behavior of polyimide (PI) insulations. To address this challenge, we conducted electro-thermal aging tests on PI films, considering various physical factors, including voltage amplitude, frequency, temperature, and humidity. This article aimed to study the impact of these factors on the insulation life of PI films and their interactions. We used the full factorial design of experiment (DOE) method to analyze the relationship between multiphysical factors and insulation lifetime. An orthogonal design allows for the independent estimation of each main effect and interaction between factors. The lifetime data from experiments are well aligned with the Weibull distribution, and the goodness of fit is 0.95, which is higher than the critical value of 0.91 specified by IEEE 930. Pareto analysis was used to identify the most influential factors and their interactions affecting PI insulations. The results showed that the ball-to-plate electrode caused aging and erosion on the insulation film, leading to partial discharges (PDs) and complete breakdown. Among the examined factors, voltage amplitude, frequency, and temperature exhibited negative correlations with insulation life, while humidity (up to 60%) displayed a positive correlation. Of these factors, voltage amplitude, temperature, and their interactions had the most significant coupling effect, while the coupling of other factors had a lesser influence. The design of the experiment model significantly reduced the experiment cost by minimizing the number of experiments. We validated the design of the experiment model by employing response optimization analysis.