Predictive modeling for tool deflection and part distortion of large machined components

摘要

Machined monolithic components provide the foundation for modern aircraft structures requiring high performance designs in terms of weight, strength, and fatigue properties. Part distortions arising from machining-induced stresses and tool deflection resulting from high dynamic cutting forces frustrate manufacturing and assembly processes, necessitating expensive in-process and post-machining corrective measures aimed at eliminating their effects. Component weight and cost requirements are often compromised by adding thicker component section designs or off-machine part flipping processes as a distortion control mechanism. Similarly, small depths of cut are taken to minimize cutting forces and tool deflection. Additional spring passes are also taken to eliminate any undercut errors introduced by unanticipated deflection and distortion problems. The ability to accurately predict and minimize tool deflections and part distortion via simulations can significantly reduce manufacturing and assembly costs. This paper presents physics-based models for predicting tool deflection and part distortions by considering the appropriate physics for each problem. Dynamic cutting forces predicted by physics-based machining models, and tool compliance properties are incorporated into a detailed linear elastic deflection model in order to predict in-process deflections along a computer numerical control (CNC) machining toolpath. Similarly, bulk stress state and machining-induced stresses for large, monolithic part machining are taken into account for predicting part distortions. Sources of stresses may include heat treatment, quenching, forging, and machining operations. Bulk stress data from heat treatment predictive models can also be imported and mapped onto the workpiece finite element model. CNC part programs, along with stresses arising from corresponding tooling, are processed and analyzed. Results of a validation study for workpiece distortion predictions are presented for a number of monolithic, thin-walled components. Predictions for tool deflection are also compared against experimental measurements for multiple cutting configurations and tool diameters. Good correlation is found between predictions and measurements of deflection.