Diagnosis of gasoline-fuelled engine exhaust fume related faults using electronic nose

Abstract

Fault diagnosis, isolation and restoration from failure are crucial for maintenance and reliability of equipment. In this paper, a condition monitoring approach that uses the sense of smell was investigated to diagnose ignition and loss of compression faults in gasoline-fuelled engine. An electronic nose based condition monitoring system was used to obtain smell print of the exhaust fumes of an automobile gasoline engine in different normal and faulty operating conditions. The data were analyzed with fuzzy c-means, hybrid principal component analysis and artificial neural network. Fuzzy C- means clustering was used to ascertain the extent to which the smell prints can characterize the selected engine faulty and normal conditions. Silhouette diagrams and silhouette width figures were used to validate the clusters. The faults considered were all correctly classified by hybrid principal component analysis and artificial neural network algorithm with 100% accuracy.