Distributed Load Identification for Uncertain Structure Based on LHS-GA and Improved L-Curve Method

Abstract

In order to identify the upper and lower bounds of distributed force exciting on an uncertain structure, a comprehensive approach combining genetic algorithm based on Latin hypercube sampling (LHS-GA) and improved L-curve method is built up in this paper. For uncertain parameter expressed by interval form, LHS-GA is presented to seeking the maximum and minimum amplitudes of distributed load in the interval. Aiming at distributed load reconstruction for a specific sampling point in the interval, which is the objective function of LHS-GA, the traditional L-curve method is improved to weaken the ill-posedness problem and obtain accurate distributed force amplitude. Numerical example results indicate the comprehensive method is able to acquire precise bounds of distributed load. Meanwhile, it possesses a strong anti-noise performance.