No abstract

Developing a dynamic model of the process is the fnst step toward implementing a modern controller. Because of the complexity of chemical process, most process control models are identified, that is, developed by input a known sequence of inputs, recording the response, and fitting a model to describe the dynamic behavior. These models are usually linear time invariant models. This research focuses on the application of neural networks to the development of dynamic models. In particular, this paper presents a modification of the common layered structure used for backward error propagation by the addition of direct linear connections between the input and output layers. This creates a number of advantageous compared to traditional structures including easy initialization, better evaluation of the learning algorithm, and better extrapolation outside of the learning sets.

No abstract