A BAYESIAN-GAUSSIAN NEURAL-NETWORK AND ITS APPLICATIONS IN-PROCESS ENGINEERING

Recently, artificial neural networks have been widely applied in proce ss engineering, where the back-propagation neural networks are most fr equently used, while the recurrent neural networks and radial basis fu nction neural networks are sometimes used. However, the intrinsic vuln erable points of these networks in long training time, local minima an d lack of self-tuning ability impair their further, specifically on-li ne, applications. To this end, a Bayesian-Gaussian neural network is i ntroduced in this paper. Simulation studies on its application to the dynamic behaviour prediction of a nonlinear single-input single-output system, as well as to the static performance and dynamic behaviour pr edictions of circulating fluidized bed boilers, are provided to assess the advantages of this network, the results of which indicate that th e BGNN could be a good alternative in neural network model based appli cations in process engineering. (C) 1998 Elsevier Science S.A. All rig hts reserved.