Direct process estimation from tomographic data using artificial neural systems

The paper deals with the goal of component fraction estimation in multicomp onent flows, a critical measurement in many processes. Electrical capacitan ce tomography (ECT) is a well-researched sensing technique for this task, d ue to its low-cost, non-intrusion, and fast response. However, typical syst ems, which include practicable real-time reconstruction algorithms, give in accurate results, and existing approaches to direct component fraction meas urement are flow-regime dependent. In the investigation described, an artif icial neural network approach is used to directly estimate the component fr actions in gas-oil, gas-water, and gas-oil-water flows from ECT measurement s. A two-dimensional finite-element electric field model of a 12-electrode ECT sensor is used to simulate ECT measurements of various flow conditions. The raw measurements are reduced to a mutually independent set using princ ipal components analysis and used with their corresponding component fracti ons to train multilayer feed-forward neural networks (MLFFNNs). The trained MLFFNNs are tested with patterns consisting of unlearned ECT simulated and plant measurements. Results included in the paper have a mean absolute err or of less than 1% for the estimation of various multicomponent fractions o f the permittivity distribution. They are also shown to give improved compo nent fraction estimation compared to a well known direct ECT method. (C) 20 01 SPIE and IS&T.