Consider neural networks for process identification

Abstract

An industrial application of neural networks (NNs) to process identification appears to confirm the potential of this technology for this application. The multivariable process has 7 dependent variables, 11 independent variables and is believed to contain significant nonlinearities. The input-output data required for modeling were collected over 10 days. The objective in this article is to determine the applicability of NNs for modeling an industrial multivariable problem involving 11 inputs and 7 outputs. The inputs and outputs are coded for proprietary reasons, but all other relevant information is provided.