Data Analysis and Forecast for Aircraft Structure Impact Damage Based on GP Symbolic Fitting Method

Abstract

In order to research the aircraft impact damage, one symbolic fitting method for analyzing and forecasting the damage data is proposed based on genetic programming (GP). The method can be used to forecast the impact damage by recognizing the rule in some groups of actual data including impact parameters and damage hole size. The principle of GP symbolic fitting method is briefly introduced. The fitting model is created with some sample data respectively for training and testing from Sorenson experiential equation. The computation with Matlab program indicates the model has a good performance to fit and forecast the damage data with avoiding the noise. The application of GP symbolic fitting method can help to decrease the times of fire experiments. Since the method can recognize the complicated nonlinear relationship between the impact parameters and damage data, it is more applicable than theoretical analysis and experiential equation to forecast the aircraft impact damage.