Probabilistic design of uncertainty for aluminum alloy sheet in rubber fluid forming process
The inherent variations in aluminum alloy sheet properties and rubber fluid forming (RFF) process of aircraft make quality control extremely difficult. Conventional finite element method (FEM) simulation is certain deterministic, all the input parameters such as material properties and process parameters are fixed numbers, but rarely considers the variations in an actual production process. The inclusion of uncertainty in the design and optimization should lead to deeper understanding about the relationship between process variables uncertainties and the part quality characteristics.This thesis proposed a probabilistic design approach of uncertainty, considering the inherent variations of material properties and uncertainties of RFF process. The approach integrates FEM simulation, design of experiments technique, process uncertainty modeling and response surface method to build an explicit function to connect process inputs to process performance outputs. For RFF process of aircraft, the uncertainties inputs are the 2B06-O aluminum alloy sheet properties such as strain hardening coefficient “ n , strength coefficient “ K , Youngs modulus “ E and etc., the process conditions including forming forces and lubrication condition. Otherwise, the outputs are quality evaluation index such as relative thinning and springback value to predict fracture and springback. The optimal uncertainties inputs can be screened by Plackett-Burman method when the noise factors are too much. Through the formulation of probabilistic design and the robust reliability analysis methods, the robust optimization result can be found. Ultimately, the sensitivity of inputs variations to quality evaluation index could be reduced, defect rates could be minimized and product quality would be improved. The typical flanging parts of 2B06-O aluminum alloy sheet in rubber fluid forming procedure is used to verify the practicability and validity of the proposed approach by the comparison between the probabilistic design approaches data and the experiment data.
Wang Miao Li Dongsheng
School of Mechanical Engineering & Automation, Beihang University, Beijing 100191, China
国际会议
桂林
英文
1-12
2010-11-16（万方平台首次上网日期，不代表论文的发表时间）