Fuzzy logic model for deodorizer troubleshooting in palm oil refining

Abstract

Deodorizer failure roots from many sources that disrupting plant operation. A correct action for remedies during troubleshooting can save company time and profit. In this study based on real plant data, a fuzzy logic model was developed to troubleshoot the root cause of problem for deodorizer failure. An abnormality of deodorizer root cause of problem was monitored through normal condition baseline. In order to build the fuzzy troubleshooting system, the linguistic data was gathered by interviewing palm oil refining plant workers that included experienced plant operators, supervisors and executives. Numerical data were collected by reviewing Standard Operating Procedure (SOP), Process and Instrumentation Diagram (P&ID), plant manual, etc. Root cause of deodorizer failure was separated into seven categories: Fatty Acid Distillate (FAD), cooling tower, hotwell, sparging steam, High Pressure Boiler (HPB), other vacuum parameter and others. The proposed model was successfully able to predict the action needed for deodorizer failure. The model can be employed for training to inexperience workers and operators.