Machining error source diagnostics using a turning process simulator.

Abstract

This dissertation presents a systematic methodology to analyze interactions among the components of the machining system and their effects on the part accuracy and the process stability. One unique feature of this approach is its capability to simulate and analyze varying-time-delay problems efficiently. Based on the proposed methodology, a complete CNC turning system was modeled and calibrated. The corresponding experiments for model calibration and validation were performed. Then a turning process simulator, i.e., a new-generation CAM software tool, was developed to simulate machining inaccuracy and to predict machine-tool chatter. The part dimensions were reasonably well simulated, while the machine-tool chatter could not be reasonably predicted due to the lack of a predictive dynamic force model in three dimensional cutting. There is a need for a 3-D dynamic force model whose parameters can be predicted from steady state cutting data. The original contributions of this research are summarized as follows: (1) A turning process simulator has been proposed to simulate the interactions among the machine tool components and the cutting process on machining inaccuracy and instability. The same formulation strategy can also be applied to other machining processes, e.g., boring, milling, drilling, etc. (2) A formulation methodology has been proposed to simulate the system equations with time-varying delays efficiently. (3) A general and efficient method has been proposed for stability analysis against machine-tool chatter. (4) A cutting process model has been proposed to simulate the resultant force in contouring. The proposed model is also a possible candidate for nonlinear chatter simulation in three-dimensional cutting. (5) The interactions among the subunits of a CNC lathe and the cutting process have been found to be important. The coupling effects among the structural subunits were also found to be important during model calibration. Input signals for structural dynamics identification are required to be internal loads or excitations to the machine tool for accurate parameter estimation.