Fault Prediction Of Pharmaceutical Air Compressor Using The Intelligent Model Based On The Bayesian Network

This paper presents a new approach of diagnosis and prognostic in real-time of strategic equipment of pharmaceutical industry. This approach is developed using Bayesian network (BN) which consider industrial data and feedback experience. The objective is to detect, locate and prevent any malfunction of the air compressor (oil-free) without air contamination, dedicated to pharmaceutical industry, BEKER Laboratories (Dar El Beida-Algeria). The study is based on the functional analysis of the air compressor to obtain the fault tree (FT). This FT is transformed into BN to diagnose automatically the compressor and prevent any malfunctioning.