Thermal effects in uncontrolled factory environments are often the largest source of uncertainty in large volume dimensional metrology. As the standard temperature for metrology of 20˚C cannot be achieved practically or economically in many manufacturing facilities, the characterisation and modelling of temperature offers a solution for improving the uncertainty of dimensional measurement and quantifying thermal variability in large assemblies.Technologies that currently exist for temperature measurement in the range of 0-50˚C have been presented alongside discussion of these temperature measurement technologies' usefulness for monitoring temperatures in a manufacturing context. Particular aspects of production where the technology could play a role are highlighted as well as practical considerations for deployment.Contact sensors such as platinum resistance thermometers can produce accuracy closest to the desired accuracy given the most challenging measurement conditions calculated to be ~0.02˚C. Non-contact solutions would be most practical in the light controlled factory (LCF) and semi-invasive appear least useful but all technologies can play some role during the initial development of thermal variability models.
Main textIn large volume metrology, thermal effects make a significant contribution to the uncertainty of measurements. Large structures that are often 20 m in length or greater are currently being assembled in factory environments where there can be thermal gradients of around 3-5˚C from floor to ceiling at any given time. Over a 24 hour cycle, the variation in ambient temperature can be as much as 15˚C.Whilst dimensional measurements in industry are sometimes taken alongside ambient temperature measurements which are used for linear scaling within the metrology software, this is often a sole measurement at one point in space, at one instance.One potential solution to the problem of thermally induced uncertainty is to model the thermal characteristics of the measurand. This can be achieved by monitoring the temperature and using this data to update a computational model that can more accurately predict the thermal and gravitational effects of the environment. The computational model makes use of the nominal CAD geometry of the measurand, alongside finite element analysis and tolerancing software.