Interactive modeling software for supervisory control of industrial processes with increased complexity

Engineering systems are now becoming more and more complex and increasingly dependent on automation, giving rise to a fear that even the smallest error in the system may cause a disaster. Among various factors that contribute to the increased system complexity are the intrinsic non-linearity within the system and the scale enlargement. It is understood that even simple non-linearity may lead to complex phenomena such as non-linear oscillation, bifurcation and chaos that are hard to model, predict or control. Scale enlargement leads to well-known problems in supervisory control like complicated operators´ situation awareness, data overload during system malfunction, etc. In this paper, an interactive modeling method would be proposed to support the extraction of meaningful patterns from large amount of plant data using data mining techniques and the latest grey-box modeling approach. This method would be demonstrated over a simple CSTR system.