Volumetric error model for online machine tool compensation

RWTH Publications (RWTH Aachen)

Abstract

Achieving high-precision machining in large-scale manufacturing poses significant challenges, particularly due to thermal effects that induce structural deformations in machine tools. Addressing these errors is crucial both for precise part production and for enabling closer quality-control loops through On-Machine Measurements. Current mitigation strategies often rely on extensive climatization measures, which are both energy-intensive and costly. At the same time, rising pressure to lower energy consumption and costs threaten the competitiveness of precision machining, which underscores the necessity of accurate machine tool calibration and thermal compensation techniques for machine tools. In this thesis, a system for real-time static and thermal error compensation for machine tools will be presented contributing to the realization of “Virtual Climatization”. The approach includes research on a flexible machine tool calibration model that can incorporate heterogeneous measurement data. The model forms the basis for a significantly faster and automated machine tool calibration procedure using an “On-The-Fly” approach. It leverages laser tracker measurements captured during continuous machine movement. By acquiring data without frequent standstills, it allows for measuring static machine tool errors with negligible thermal drift in under ten minutes. The method thereby enables investigating transient thermal behavior of the machine tool with a high temporal resolution. An experimental validation involved testing this calibration method on four different machine tools covering medium to large working volumes. An “Abstracted Physical Body model” can subsequently be used to predict thermally induced deformations of the machine tool structure using efficiently solved spline-based mechanics. This model is designed to be applicable to retro-fit applications, as it does not require detailed structural information or training datasets, while still providing good predictions. A method for integrating these predictions into the existing machine control infrastructure using standard machine tool communication interfaces is investigated. By updating the machine’s geometric compensation data, real time compensation can be achieved without extensive hardware or control-system modifications. A formal uncertainty analysis of all components is performed. In alignment with established standards, the contributions of all system components are identified and analyzed. This analysis thereby enables using the compensation system for traceable On-Machine Measurements. A thorough experimental validation of the combined system is executed in two stages. First, the combined system can reduce deviations by up to 75 % in tests referenced to laser tracker data on a medium-sized machine tool. Secondly, the measurement uncertainty can be reduced by up to 50 % in On-Machine Measurement validation experiments based on the VDI/VDE 2617 standards for CMM qualification. This can ultimately lead to improved precision and accuracy in manufacturing and geometric inspection, resulting in higher product quality and reliability.

Authors 0

1. Author list not loaded yet.