Digital Twins are becoming increasingly popular for process modelling because they facilitate experimentation without upsetting the manufacturing process. A Digital Twin is constructed using a simulation of a process. It allows virtual exploration of the effect of changing conditions even when the real process is very steady. In other words, Digital Twins are a valuable asset to the analyst’s toolbox.
Statistically designed experiments are being used to interrogate an approximating “digital twin” constructed using simulation of a real-life complex process. Starting with a factorial orthogonal experiment, the design space is explored using Bayesian adaptive design. The non-linear responses are modelled using Gaussian Processes to emulate the underlying process and predict outcomes in what-if scenarios. This is all part of a move towards optimising the underlying process.
The work is a joint study by NU Solve in the School of Maths, Statistics and Physics and the School of Engineering at Newcastle University and is supported by the Centre for Process Analytics and Control Technology (CPACT) through its feasibility study scheme.
CPACT is an industry-academic club founded in 1997 and is Europe’s longest running community of practice in process analysis and control. The member organisations are manufacturing companies from all sectors of the Chemistry-using industries, technology vendor companies, and University and Research Centre organisations.
Besides feasibility studies and research projects, CPACT provides its members with a programme of webinars, courses, data analysis software, and events like the APACT conference which will be held on-line on 19-23 April 2021. On-line APACT is free and open to all.
CPACT is also supporting and presenting at the forthcoming on-line ENBIS conference on Data Science in Process Industries on 17 and 18 May 2021. ENBIS 2021 Online Spring Meeting: Data Science in Process Industries – Events – ENBIS
This article is by by Shirley Coleman, Technical Director of NU Solve.
For information on joining CPACT or registering for On-line APACT contact email@example.com
By Newcastle University