A comparison of dynamic models for an evaporation process

This paper presents the development of three dynamic models of a multi-effe ct, falling-film evaporator and compares their performance. The models deve loped are: an analytically-derived model, an artificial neural network and a linear regression model with an ARX (Auto-Regressive with eXogenous input s) structure. The development of the analytical model follows a systems app roach to analysing the process. The paper focuses on the development of the neural network, in particular developing techniques to improve model flexi bility and the use of prior knowledge. The neural network was formed by com bining submodels, each modelling a specific element of the overall system, resulting in a modular-structured model. The elements to be modelled were s elected using prior knowledge of the system. The linear ARX model was struc tured in a similar manner. It was found that the empirical models had a sup erior predictive performance over the analytical model. The modular models also provide benefits in terms of model development effort, flexibility and simple implementation within model-based control strategies.