Machining of aerospace structural components involves many thin-wall rib and flange sections. These thin-wall sections are dictated by design consideration to meet required strength and weight constraints. These components are either forged or cast to the approximate final shape and the end milling process is used to finish the parts. Alternatively, the component is machined from a solid block of material by end milling with roughing and finishing cuts. During machining, the cutting forces cause deflection of the thin-wall section, leading to dimensional form errors that cause the finished part to be out of specification. In this thesis, a new methodology for the prediction of wall deflection during machining of thin-wall feature is presented. The new methodology aims to increase the efficiency on modelling the deflection prediction in machining thin-wall component. The prediction methodology is based on a combination of finite element method and statistical analysis. It consists of a feature based approach of parts creation, finite element analysis of material removal and statistical regression analysis of deflection associated with cutting parameters and component attributes. The model is developed to take into account the tool-work geometries on material removal process during machining process. Mathematical models are developed for the wall deflection correlated with cutting parameters and component attributes. The prediction values have been validated by machining tests on titanium alloys parts and show good agreement between simulation model and experimental data. In addition, the cutter compensation method derived from the deflection prediction values can be used to reduce the magnitude of surface error, thus improving the component accuracy for machining thin-wall feature. By adopting the cutter compensation method, only one machining pass is required to machine the thin-wall feature. This compares favourably to the current practice in step method which requires many machining passes. All research results have been derived for four different cases of typical aerospace component, but it is shown that these results can be applicable for other component shape and materials. To assist commercial applications, a customized computer program has been developed for the hybrid model. The computer program is an integrated data exchanges between modules upon users input on the design information and machining parameter for automatically generate the solid model, material removal model and FEM analysis. The new method is able to reduce the analysis time from weeks to hours.
展开▼