Modeling a scrubber using feed-forward neural networks

Mathematical modeling in the pulp and paper industry is largely based on physical models that contain many empirical correlations and assumptions. A technique that can complement these physical models, and that does not require simplifying assumptions, is a neural network. Neural networks can be used to model systems where physical models are not available or to improve existing physical models with various techniques of hybrid modeling. Such models have an advantage of the robustness of physical models combined with the higher accuracies offered by empirical techniques such as neural networks. Most systems in paper drying are quite nonlinear and are often too complicated to be accurately described with physical models. Neural networks are powerful tools that can solve a variety of nonlinear modeling problems. Using an example of a scrubber, this paper illustrates the versatility of this technique. Neural networks were used to predict the optimum operational conditions in designing a scrubber for heat recovery from paper machines. In this study, feed-forward neural networks are applied to estimate the outlet water temperature of the scrubber process. The networks are trained with data obtained from experiments carried out on two pilot scrubbers. The accuracy of the neural network model is significantly higher than that of the physical model. Application : the neural network is very useful when the physical phenomena of the process are unknown or too complex to be solved with a physical model.