A framework for predicting the geometrical errors of thin- floor components in end milling using a flexible fixturing setup

Abstract

Thin-floor components are commonly used in aerospace industry such as spars, and bulkheads. Their superlative ratio of strength to weight makes them highly demanded by manufacturers. One of the inevitable problems in machining such products is the deformation which is induced by cutting forces during manufacturing and causes surface form errors. This issue becomes more important when a flexible fixturing system is involved to support the workpiece. In order to assure the machining accuracy in milling with this situation, a prediction model for surface dimensional errors is required to avoid costly compensation operations and reach high productivity. In this dissertation, a structured simulation procedure in milling suitable for part geometrical errors induced by axial cutting forces is proposed while the workpiece is fixed to (on) a flexible fixturing setup. The process is designed in an FE model as an implicit/Static analysis of material removal and deformation under the influence of applied axial cutting forces. A cutting force model is used to measure the average cutting forces in different positions of the cutter. An Abaqus Python API is applied to conduct numerous iterative procedures during creating the model in Abaqus including parametric study, creating repetitive geometry and managing multiple steps and forces. The theoretical model is based on the assumption that the workpiece is an elastic body. Because there is not much support on the back face of the plate, it will deform vertically and causes cutting deflections. The advantages of the proposed model over previous works are: 1. A framework is proposed to predict the workpiece non-linear behavior during machining due to its constant changing geometry. 2. It greatly speeds up the experimental loop. 3. It can help to develop an off-line error compensation model in Abaqus by manipulating and adjusting the depth of cut value through the trajectory in an iterative process. The first results show a gap between the theoretical and experimental models due to the simplification of a dynamic movement into a static behavior. Therefore, in the next step, the numerical model is developed in an attempt to reduce this gap.

Date	13 Aug 2021
Original language	American English
Awarding Institution	École de technologie supérieure
Supervisor	Jean-François Chatelain (Supervisor)