Authors: Vladimir B. Bokov
Addresses: 'NPP Automatica' JSC, 77 B. Nizhegorodskaya St., City of Vladimir, Russia
Abstract: Process or product characterisation involves the determination of quantitative estimates of physical quantities from experiment, along with estimates of their associated uncertainties. Herewith a comprehensive model is the key to extracting information from the experimental data. The information obtained depends directly on model performance. With this concern a novel response surface modelling technique has been fashioned by using jointly original explanatory variables and physics-based transformed explanatory variable. It lets us approximate adequately the true unknown model on the considered region of interest for the original explanatory variables using first-order design only. This technique has allowed attaining the advanced level of model precision. Pneumatic gauge candidate models| building, solving, and validation reviled that the first-order response surface model with physics-based transformed explanatory variable permits to attain adequacy and maximum precision.
Keywords: curvilinear relationship; first-order experimental design; explanatory variables transformation; pneumatic gauge modelling; response surface modelling; information extraction; information retrieval; experimental data.
International Journal of Experimental Design and Process Optimisation, 2010 Vol.1 No.4, pp.327 - 347
Published online: 31 Aug 2010 *Full-text access for editors Access for subscribers Purchase this article Comment on this article