Health & Environmental Research Online (HERO)


Print Feedback Export to File
2663169 
Journal Article 
GASOLINE ENGINE INTAKE MANIFOLD LEAKAGE DIAGNOSIS/PROGNOSIS USING HIDDEN MARKOV MODEL 
Ahmed, Q; Iqbal, A; Taj, I; Ahmed, K 
2012 
Yes 
International Journal of Innovative Computing, Information and Control
ISSN: 1349-4198 
8 
7A 
4661-4674 
WOS:000306679500016 
Leakages in the air intake system (AIS) of gasoline engine can deteriorate its performance causing poor fuel efficiency, air pollution and retarded driving performance. The diagnosis and prognosis of air leakage becomes inevitable in order to enhance reliability and improve fuel consumption. Currently, air leakages detection is hardly found in any of the available On Board Diagnostic version-II (OBD-II) scanners. In this paper, a challenging task of detecting manifold air leakage at early stage has been resolved by employing discrete Hidden Markov Model (HMM). Discrete HMM is an stochastic classifier that has been exploited for the first time to generate useful information about AIS condition based maintenance. The proposed fault diagnosis and prognosis (FDP) scheme can detect air leaks and consequently the severity of air leakage is explored to update the schedule of maintenance prior to any mishap. The validation of the proposed algorithm is carried out on 1.3L production vehicle engine. The experimental results demonstrate that HMM based FDP scheme accurately detects air leakage at early stage and informs about its approximate severity. The suggested scheme for leakage diagnosis is cheaper, does not require any extra hardware installations and it remains valid for all OBD-II compliant commercial vehicles.