Geometric errors on CNC machine tools have reduced significantly over the past few decades and compensation either through the controller or with retrofit systems have helped improve accuracy even further. This move is driven by a trend of ever tightening tolerances on components manufactured on these machines, especially where simplified assembly is required such as the Aerospace industries. Thermal errors have affected the accuracy of production machines for a long time but now for many industries with large machines and small error budgets, they are the dominant source of inaccuracy and are often the most difficult to reduce. Many solutions exist for the machine tool builder to reduce the errors that can be applied at the design stage including symmetric structures, low coefficient materials, liquid cooling systems etc. Design effort and cooling systems can add significant cost and usually cannot eliminate the errors but only help reduce them. Active and pre-calibrated compensation can be an important and effective alternative with some machine tool builders incorporating temperature sensors on the machine at the build stage. These systems are invariably controller-based using a simple model to estimate error at the tool and they are usually specific to a machine or range of machines within the company. These systems can produce good results for linear repeatable thermal errors and they work best in conjunction with good design [1]. Generally most thermal errors measured between the tool and workpiece are caused by a complex interaction of structural distortions having different heat sources, thermal time constants and expansion coefficients. Because of these causes the errors are time-varying and non-linear and are therefore difficult to model accurately. Significant research in the field of thermal error compensation has produced a number of sophisticated retrofit techniques for modelling thermal errors that employ techniques such as Neural Networks and multiple linear regression analyses [2, 3, 4, 5]. These systems can produce excellent results but often require complicated hardware and software systems and testing regimes to train the models or optimise temperature sensor positions for the particular machine or application. This paper discusses new compensation capabilities and model development techniques that build on work by White [6, 7] to enhance the efficiency of application and machine accuracy. The philosophy relies on comprehensive measurement of temperature and a flexible compensation system that uses a novel programming language dedicated to modelling non-rigid behaviour of machine structural elements. The system practicality has been proven through industrial application. Machine downtime, although undesirable, is required for completion of most phases of a thermal compensation system particularly the measurement of machine thermal behaviour, hardware implementation of temperatures sensors and compensation system and finally validation testing. It has already been mentioned that many successful modelling techniques require long testing regimes to provide sufficient information for robust model creation. The techniques described in this paper are designed to minimise this requirement in all areas.
[1]
Derek G. Ford.
Laser Metrology and Machine Performance VI
,
2003
.
[2]
Jenq-Shyong Chen,et al.
Improving the machine accuracy through machine tool metrology and error correction
,
1996
.
[3]
Andrew P. Longstaff,et al.
Practical experience of thermal testing with reference to ISO 230 Part 3
,
2003
.
[4]
S. R. Postlethwaite,et al.
Measuring And Modelling Thermal Distortion On CNC Machine Tools
,
1970
.
[5]
Jun Ni,et al.
Dynamic neural network modeling for nonlinear, nonstationary machine tool thermally induced error
,
2005
.
[6]
S. R. Postlethwaite,et al.
A general purpose thermal error compensation system for CNC machine tools
,
1970
.
[7]
Aun-Neow Poo,et al.
Thermal error measurement and modelling in machine tools. Part II. Hybrid Bayesian Network—support vector machine model
,
2003
.
[8]
Simon Fletcher,et al.
Volumetric compensation through the machine controller
,
1970
.
[9]
Andrew P. Longstaff,et al.
Volumetric error compensation through a Siemenscontroller
,
2005
.
[10]
Jun Ni,et al.
Adaptive model estimation of machine-tool thermal errors based on recursive dynamic modeling strategy
,
2005
.
[11]
Aun-Neow Poo,et al.
Error compensation in machine tools — a review: Part II: thermal errors
,
2000
.