Criterion to determine the minimum sample size for load spectrum measurement and statistical extrapolation

Abstract

A criterion to determine the minimum sample size for load spectrum measurement and its statistical extrapolation was developed in this paper. The proposed criterion – based on the t-distribution – was employed to evaluate the minimum measurement volume of a measured load spectrum. The results demonstrated that this criterion performed well. The resultant volume of the load spectrum measurement was then used to calculate the minimum mileage of the measured load spectrum through multiplying the sample length. The calculated minimum mileage was larger than the distance of the measured load spectrum. This implied that the load spectrum measurement needed to continue because the sample size cannot represent the population. Therefore, a modified parametric extrapolation method was subsequently proposed to generate a full-design-life load spectrum based on the measured load spectrum. Load spectrum changes in the frequency domain before and after extrapolation were observed using the proposed modified parametric extrapolation method.