Early Fault Diagnosis of Vehicle EGR System Based on Support Vector Machine Technique

Abstract

It has been a challenge to the automotive world to manufacture the smart vehicle components that contribute to the performance of vehicle operations to adhere due to strict vehicle emission norms. In this context, engine operations need to carry systematically, which helps to maintain the function of the vehicle emission system properly. Now a day’s electronic sensors are playing a vital role in smarter vehicle component operation. In-vehicle exhaust gas recirculation system is responsible to control the emission in diesel and petrol engines vehicle. Many times due to fault in vehicle exhaust gas recirculation system, vehicle engine gets heated drastically that causes to seize the engine and also lose the control on the emission of gases. At present faults in the exhaust gas, recirculation system is diagnosed mostly after the failure of exhaust gas recirculation, at the vehicle repair center, with the help of a knowledge base and manual observations. It is essential to predict earlier exhaust gas recirculation failure possibilities to avoid vehicle engine impact and emission operations. This paper discusses the use of machine learning techniques to predict the exhaust gas recirculation failure possibilities with a support vector machine classifier. To predict exhaust gas recirculation status effectively its correlated three parameters like coolant level, exhausted particle temperature present in the cold exhaust gas recirculation pipe, and boost temperature sensor is considered. Overall 94.11% prediction accuracy using the support vectors machine is achieved.