Application of multivariable process monitoring techniques to HAZOP studies of complex processes

Abstract

Multivariable process monitoring (MPM) algorithms are very popular for early detection of abnormalities and diagnosis of process conditions to ensure plant safety, reliability, and production quality during operation. Hazard and operability (HAZOP) study is a systematic brainstorming session to identify the credible causes and consequences of the process upsets. Then the process safeguards in place are reviewed to evaluate risk of HAZOP scenarios. According to IEC61882 HAZOP studies have proved to be very useful in a variety of different process industries. However, the method has some limitations to be considered in potential applications and particularly for complex processes. This paper aims at improving the HAZOP study by applying MPM algorithms to perform a holistic and more accurate consequence analysis and effective evaluation of the process safeguarding strategy. The proposed monitoring algorithm consists of two layers of decision boundaries. The 1st layer is based on normal operating condition and the 2nd layer (i.e. suggested in this paper) is based on safe operating condition of plant. The effectiveness of the proposed method is verified by investigations on an operating complex polymerization plant.