An advisory system using a rule–based approach has been developed in which the knowledge required to perform hazard identification is divided into process–specific and process–general components. Hazard identification has been carried out using the modified Hazard and Operability Study (HAZOP) method. In the proposed modified HAZOP, the two study nodes are connected in one mode of analysis. The process–specific knowledge, which consists of a conventional HAZOP study result, has been stored in the database. This process–general knowledge consists of rule–based which has been developed from the result of process simulation. The combination of hazard identification technique with process simulation result is important, in order to analyse the causes and consequences of the deviation in the process. For hazard identification, the process deviations selected are flow rate, temperature, and pressure. An inference engine for this advisory system has been developed using Visual Basic programming language, for appropriate interaction between knowledge–based components, in order to identify process–specific of causes and consequences for each process deviation specified. The procedure is based on the proposal HAZOP algorithm modified from a conventional HAZOP. The case study used is a packed column of an oleo chemical plant. The study has contributed to an improvement of hazard identification technique, which proposed a modified HAZOP algorithm by considering the consequences of the operation for each process deviation. The modified HAZOP algorithm has been proposed in a generic manner, however, the advisory system developed in this study is limited to the application for packed column of oleo chemical plant only.
Key words: Hazard identification, advisory system, ruled based, process simulation, packed column
