This paper presents a comprehensive ANN based multisensor fusion approach designed to support the implementation of an adaptive error compensation of geometric, thermal and dynamic errors for enhancing the accuracy of CNC machine tools. | Journal of Automation and Control Engineering Vol. 1, No. 3, September 2013 A Multisensor Approach for Error Compensation in CNC Machine Tools Abderrazak El Ouafi and Noureddine Barka University of Quebec at Rimouski /Engineering Department, Rimouski, Canada Email: {abderrazak_elouafi, noureddine_barka}@ Abstract—This paper presents a comprehensive ANN based multisensor fusion approach designed to support the implementation of an adaptive error compensation of geometric, thermal and dynamic errors for enhancing the accuracy of CNC machine tools. Accurate and efficient model to perform on-line error prediction is an essential part of the compensation process. The proposed approach consists of the following major steps: (i) design of an integrated spatial-variant model describing the machine topology, (ii) measurement of path-dependent rigid-body errors according to the model, (iii) design of time-variant models for error components through sensors fusion, (iv) continuous monitoring of the machine conditions using position, force, speed and temperature sensors for one-line error components prediction and integration to produce a correction vector, (v) total positioning error synthesis and software compensation. Implemented on a turning center, the proposed approach led to a consistent model able to accurately and reliably provide an appropriate error identification and compensation under variable machine tool conditions. Index Terms—machining, machine tools, machine tool accuracy, error compensation, sensor fusion, neural network I. INTRODUCTION The quest for higher precision has stimulated the development of highly accurate CNC machine tools (MT). Unfortunately, the machining task is usually performed in the presence of disturbances, which can induce systematic and random errors that tend to adversely affect the MT accuracy. Indeed, the MT accuracy is affected by various errors related to geometric imperfections, thermal deformations, load effects .
