Modeling of machining thin-walled aerospace structures

Thin-walled monolithic aerospace parts and rotor blades have high flexibilities causing severe static and dynamic deflections during machining. Poor dimensional accuracy and surface finish due to deflections scrap the costly parts. This thesis presents mathematical models to simulate thin-walled part machining in virtual environment considering the varying structural workpiece characteristics and process induced damping along complex toolpaths. Traditionally having full order finite element (FE) models at several toolpath locations is prohibitive in the machining of curved blades. New, computationally efficient, reduced and full order workpiece dynamics update models are developed. Removed materials between discrete locations are defined as substructures of the initial workpiece. First, in-process workpiece frequency response functions (FRFs) are directly updated by coupling fictitiously negative dynamic stiffness of the removed materials. The model is improved by introducing substructure decoupling in time domain. The workpiece structure is modified by coupling fictitious substructures having the negative mass and stiffness of removed volumes. Mode shapes of the in-process workpiece are perturbed, and mode frequencies and workpiece FRFs are updated. The computed FRFs of the thin-walled parts are used to predict the chatter stability, static deflections, forced vibrations, and their effects on tolerance violations along the toolpath. Unlike the conventional empirical process damping coefficients, a comprehensive analytical model to predict the machining process damping is proposed. The cutting edge is discretized in the chip width direction, and contact pressure between the edge element and workpiece surface is estimated using the tool geometry, vibration parameters, and work material properties. The specific process damping force of each element is evaluated by integrating the contact pressure. The damping force is linearized by representing it with equivalent viscous damper. A generalized five-axis ball-end milling dynamics model is developed in frequency domain by incorporating the dynamics update and process damping models for flexible parts. Relative tool-workpiece vibrations are projected into the local chip thickness direction and the dynamic milling equation is derived. Milling stability is assessed at discrete locations using Nyquist criterion, and chatter regions and frequencies are predicted along cutting. The proposed digital process models are experimentally verified and expected to guide engineers in process development.Applied Science, Faculty ofMechanical Engineering, Department ofGraduat