Expert failure detection technique for distillation column

In this paper we describe an expert system for diagnosis of faults in a distillation column which are in the incipient stage but that might lead to a serious situations in the near future. Since the distillation column is a large-scale system, it is therefore natural to assume that there will be a huge number of faults due to multiple causes. Hence, in such complicated systems, it becomes very difficult to handle the problem of failure detection without the aid of expert systems. When a failure takes place in the distillation column, an error will occur in the pressure and both the top and bottom product compositions. These errors are used by the expert system to detect the faults. To demonstrate the feasibility of applying expert systems in fault detection and diagnosis, we illustrate a numerical example. The results demonstrate the effectiveness of expert systems for real time fault diagnosis