| Summary: | Recent years has seen the emergence of a new paradigm in system’s identification known as Artificial Neural Network (ANN). ANN is a methodology inspired from the structure and mechanism of human brain. Similar to the human brain (albeit in simplistic scale), ANN has the learning capability that enables it to remember past information. ANN is also known to be proficient in approximating nonlinear function to arbitrary accuracy in a black-box manner. As such, if adequate training over a sufficiently rich data is provided, the network will be able to capture the information contained within the data and store the in the form of model which can be utilize to predict future characteristics.
This paper describes the basic mechanism of ANN and its application in the identification of polymerization process. The results obtained highlights the proficiency of ANN models in predicting the reactor product concentration, thus recommending its application in other model-related process engineering tasks.
|
|---|