Intelligent Diagnosis for Fuel Line Fault of Diesel Engine based on Vibration Signatures

Early fault detection is vital in maintaining system stability and to decrease the cost associated with maintenance. This paper presents an approach to identify the fuel line failure in a diesel engine based on its vibration signals and machine learning. Vibration signals were acquired on the fuel line of the engine for both normal and faulty conditions for engine ramp up condition. After acquiring the time domain vibration signals, various features were extracted and have been analyzed in time domain and time-frequency domain. After selecting the most effective feature, a machine learning algorithm (i.e., support vector machine (SVM)) based diagnosis model has been developed. Results showed that the proposed SVM based method can identify the fuel line fault correctly. This study can be useful for early detection of this critical fault in diesel engine and take useful decision before any catastrophic failure happens because of this fault.