Efficient Offline Thermal Modelling for Accurate Assessment of Machine Tool Thermal Behaviour

Thermal gradients from internal and external heat sources cause instabilities which affect the machine tool positional accuracy. Positioning error results from deformation of the machine structure due to linear thermal expansions of some machine parts combined with the thermal behaviour of associated complex discrete structures producing non linear thermal distortions. Thermal gradients due to internally generated heat and varying environmental conditions pass through structural linkages and mechanical joints where the roughness and form of the contacting surfaces act as resistance to thermal flow and affect the heat transfer coefficients. Measurement of long term thermal behaviour and associated thermal deformations in the machine structure is a time consuming procedure and most often requires machine downtime and is therefore considered a dominant issue for this type of activity, whether for characterisation or correction. This paper presents a novel offline technique using Finite Element Analysis (FEA) to simulate the combined effects of the internal and external heat sources on a small vertical milling machine (VMC). Detailed long term experimental testing of the effects of temperature distribution in the machine structure and in-depth heat transfer work to obtain accurate values of heat transfer coefficients across joints is reported. Simplified models have been created offline using FEA software and the evaluated experimental results applied for offline simulation of the thermal behaviour of the machine structure. The FEA simulated results obtained are in close correlation with the obtained experimental results. FEA simulation enables quick and efficient offline assessments of temperature distribution and displacement in the machine tool structures along with characterisation of the machine under variable environmental conditions. This results in a significant reduction in machine non productive downtime and can provide significantly more thermal data for the creation and validation of robust long term error compensation models.