摘要: Aircraft structural load analysis has great significance for individual aircraft life monitoring and structural health management. Multiple linear regression method is commonly used for flight data based in-service structural load... 展开 Aircraft structural load analysis has great significance for individual aircraft life monitoring and structural health management. Multiple linear regression method is commonly used for flight data based in-service structural load monitoring at critical locations. Based on multi-collinearity diagnosis, residual analysis and stepwise regression, an approach for synthetically building an optimal multiple linear regression load model is proposed in this paper. First of all, the multi-collinearity between independent variables should be diagnosed and attenuated; then, residual analysis is carried out to remove abnormal observations; finally, stepwise regression is used to find the optimal combination of independent variables. Furthermore, it's pointed out that traditional multi-collinearity diagnosis methods have some defects and two more feasible methods for reducing the multi-collinearity are put forward and compared. As a case study, load and stress samples of a critical wing attachment bulkhead location are used to illustrate the procedures of establishing an optimal regression equation. It is proved that the approach can not only ensure the accuracy of aircraft structural load analysis, but also select the parameters of great statistical influence, which has significant engineering practical value. 收起