Elamin, Fathi (2013) Fault Detection and Diagnosis in Heavy Duty Diesel Engines Using Acoustic Emission. Doctoral thesis, University of Huddersfield.

A condition monitoring program applied to diesel engines, improves safety, productivity, increases serviceability and reduces maintenance costs. Investigation of a novel condition monitoring systems for diesel engine is attracting considerable attention due to both the increasing demands placed upon engine components and the limitations of conventional techniques. This thesis documents research conducted to assess the monitoring capabilities used acoustic emission (AE) analysis. It focuses on the possibility of using AE signals to monitor the fuel injector and oil condition.

A series of experiments were performed on a JCB, four-stroke diesel engine. Tests under healthy operating conditions developed a detailed understanding of typical acoustic emission generation in terms of both the source mechanisms and the characteristics of the resulting activity. This was supplemented by specific tests to investigate possible acoustic emission generation due to the piston slap and friction.

The effect of faults on the injector waveform was investigated using the injection system and at one sensor location. To overcome the reflections and injection system configuration effects the method of acoustic emission impedance was used. This enabled the injector signal to be successfully extracted and clearly shows its capability for detecting even minor combustion deviations between engine cylinders.

Comparison between signals and measurement of the oil condition showed both provided useful information about the lubrication processes. Simulation and experimental work have demonstrated the capability of this technique to detect lubrication related faults and irregular lubrication variability between the engine's cylinders.

A review of the AE sources in diesel engines and how to represent the AE signals generated is presented. Three analysis methods were used: time-domain analysis using parameters such as Root Mean Square (RMS), variance, mean and kurtosis; frequency-domain analysis which relied on the amplitudes of the frequency components of the measured signals; and time-frequency domain analysis extracting features so that the energy content of the signals and the frequency components were localized simultaneously.

In this work, data has been obtained from tests on a diesel engine, where the engine load, speed, temperature and the oil lubrication type were changed. The monitored signal and its difference from that obtained for normal engine conditions was noted as a fault signature that could be used for fault detection and diagnosis.

Final_Thesis_November_2013.pdf - Accepted Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (4MB) | Preview


Downloads per month over past year

Add to AnyAdd to TwitterAdd to FacebookAdd to LinkedinAdd to PinterestAdd to Email