Article Preview
TopLiterature Review
There are diverse techniques available for misfire detection each having a specific advantage but the main challenge is the non availability of a retrofit for engines that do not have sophisticated controls and embedded sensors for monitoring vehicle performance. The various possible techniques with a detailed analysis are performed to design and develop the best possible artificial intelligence model.
Extensive studies have been done using measurement of instantaneous crank angle speed (Tinaut, Andres, Laget, & Jose, 2007) and diverse other techniques have been developed on similar lines to predict misfire (Lee & Rizzoni, 1995). These methods call for a high resolution crank angle encoder and associated infrastructure capable of identifying minor changes in angular velocity due to misfire. The application of these techniques becomes more challenging due to continuously varying operating conditions involving random variation in acceleration coupled with the effect of flywheel, which tries to smoothen out minor variations in angular velocity at higher speeds. Fluctuating torque experienced by the crankshaft through the drive train poses additional hurdles in decoding the misfire signals. In-cylinder pressure monitoring is very reliable and accurate as individual cylinder instantaneous mean effective pressure could be calculated in real time. However, the cost of fitting each cylinder with a pressure transducer is prohibitively high. This initiated the quest for identifying a low cost model capable of competing with the existing solutions.
The use of Harr and Daubechies wavelets for signal processing by approximation (Yajnik & Mohan.S, 2009) is a workable idea but the use of signal approximation techniques could lead to loss of information, hampering the possibility of growing this model in to a full vehicle monitoring system. The implementation of wavelet based clustering techniques (Palanisamy & Selvan, 2009) for handling high dimension data is encouraging. The use of pattern recognition techniques including wavelets for structural health monitoring reported by Navarro and Mejia (2010) can be reliably extended for misfire detection. A detailed work is reported by Jinseok (Chang, Kim, & Min, 2002) using a combination of engine block vibration and wavelet transform to detect engine misfire and knock in a spark ignition engine. The use of engine block vibration is encouraging since it requires minimum instrumentation. The use of wavelets in all the above mentioned work has one common challenge; the increased computational complexity due to the additional load induced into the model by the wavelets.