Modeling via Artificial Neural Network
In this chapter, artificial neural network technology is applied to predict the basis weight and moisture content to improve the paper product quality. Historical data from a paper production company in Canada are analyzed and applied to train a multilayer feedforward backpropagation network. Considering that generalized descent method, which is a typical optimization algorithm in backpropagation, has some major drawbacks, a conjugated gradient method is proposed for training neural networks. The results have shown that the neural network gives accurate paper quality prediction. The application of artificial neural network helps us to gain a better understanding of dependence of quality variables on the operating conditions and to overcome large time-delay in paper machine control systems.
KeywordsNeural Network Artificial Neural Network Neural Network Model Basis Weight Steam Pressure
Unable to display preview. Download preview PDF.
- Behnam B. Introduction to neural networks for intelligent control. ITT Control System Magazine 1988; 4: 3–7Google Scholar
- Bhat N. and McAvoy T. Use of neural nets for dynamic modeling and control of chemical processes systems. Computers Chem. Engng. 1990; 14: 573–583Google Scholar
- Donat J.S., Bhat N. and McAvoy T.J. Optimizing neural net based predictive control. Proc of American Control Conference, San Diego, 1990, pp 2466–2471Google Scholar
- Kim H.C., Shen X., Rao M., Mcintosh A. and Mahalec V. Refinery product volatility prediction using neural network. Proc 42nd Canadian Chemical Engineering Conference, Toronto, Canada, 1992, pp 243–245Google Scholar
- Minsky M. and Papert S. Perceptrons. Cambridge, MA, MIT Press, 1969Google Scholar
- Rumelhart D.E., Hinton G.E. and Williams R.J. Learning internal representations by error propagation. In: Parallel Distributed Processing, Vol 1. MIT Press, Cambridge, 1986, pp 318–62Google Scholar
- Stein R. Selecting data for neural networks. AI Expert 1993; 2: 42–47Google Scholar
- Wasserman P.D. Neural computing: theory and practice. Van Nostrand Reinhold, New York, 1989Google Scholar