Thermally induced errors play a critical role in controlling the level of machining accuracy. They can represent a significant proportion of dimensional errors in produced parts. Since thermal errors cannot totally be eliminated at the design stage, active errors compensation appears to be the most economical and realistic solution. Accurate and efficient modeling of the thermally induced errors is an indispensable part of the error compensation process. This paper presents an integrated and comprehensive modeling approach for real-time thermal error compensation. The modeling process is based on multiple temperature measurements, Taguchi’s orthogonal arrays, artificial neural networks and various statistical tools to provide cost effective selection of appropriate temperature variables and modeling conditions as well as to achieve robust and accurate thermal error models. The experimental results on a CNC turning center confirm the feasibility and efficiency of the proposed approach and show that the resultant model can accurately predict the time-variant spindle thermal drift errors under various operating conditions. After compensation, the thermally induced spindle errors were reduced from 19m to less than 1 m. The proposed modeling optimization strategy can be effectively and advantageously used for real-time error compensation since it presents the benefit of straightforward application, reduced modeling time and uncertainty.

No abstract

The aim of this study is to develop an effective on-line ANN-based approach for quality estimation in resistance spot welding. The proposed approach examines the welding parameters and conditions known to have an influence on weld quality, and builds a quality estimation model step by step. The modeling procedure begins by establishing relationships between welding parameters (welding time, welding current, electrode force and sheet metal thickness), welding conditions represented by typical characteristics of the dynamic resistance curve and welding quality indices (nugget diameter, nugget penetration, and indentation depth), and the sensitivity of these elements to the variation of the process conditions. Using these results and various statistical tools, three estimation models are developed. The first one is based exclusively on welding parameters. The second model is based on characteristics of the dynamic resistance curve. The third estimation model combines welding parameters and characteristics of dynamic resistance curves. In order to carry out the models building procedure, an extensive number of welding experiments are required. For this purpose, Taguchi’s efficient method of experimental planning is adopted. The results demonstrate that the developed models can provide an accurate on-line estimate of the weld quality, under different welding conditions.