Usually industrial processes are monitored by many sensors, which typically generate huge volumes of non-standardised multi-dimensional data, both numeric and images. In practice a large proportion of this data is not used to the fullest. This project will use historic and on-line process data to develop predictive process models for real-time optimisation of production processes. This optimisation takes place along multiple competing objectives, most of them being quality criteria.
The benefit for the companies involved is to enable them to produce higher quality end products with less downtime, thereby minimising waste and loss of productivity. With Tata Steel and BMW, the process chain represents steel production and steel forming, from a producer's as well as a consumer's perspective. LIACS, CWI and MonetDB contribute the big data storage and processing, data mining and data driven modelling, and optimization and multiple criteria decision making expertise.
The principal investigators involved are Prof. Dr. T.H.W. (Thomas) Bäck at LIACS, Dr. S. (Stefan) Manegold at CWI, Kees Jonker M.Sc. at Tata Steel, Arnulf Lipp at BMW and Prof. Dr. M.L. (Martin) Kersten at MonetDB.
Read more (in Dutch)