| Summary: | Chemical processes are systems that include complicated network of material, energy and process flow. As time passes, the performance of chemical process gradually degrades due to the deterioration of process equipments and components. The early detection and diagnosis of faults in chemical processes is very important both from the viewpoint of plant safety as well as reduced manufacturing costs. The conventional way used in fault detection and diagnosis is through the use of models of the process, which is not easy to be achieved in many cases. In recent years, an artificial intelligence technique such as neural network has been successfully used for pattern recognition and as such it can be suitable for use in fault diagnosis of processes [1]. The application of neural network methods in process fault detection and diagnosis is demonstrated in this work in two case studies using simulated chemical plant systems. Both systems were successfully diagnosed of the faults introduced in them. The neural networks were able to generalise to successfully diagnosed fault combinations it was not explicitly trained upon. Thus, neural network can be fully applied in industries as it has shown several advantages over the conventional way in fault diagnosis.
|
|---|