Mechanical faults and failures are a common occurrence within all mechanical systems. Over time a minor mechanical fault can turn into a major failure if not detected and serviced with proper maintenance. The detection and classification of minor faults is essential in reducing the occurrence of major faults, which will result in an increased lifetime of the mechanical system. Additionally, a good fault detection method should lead to increased efficiency, increased safety, improved performance, and reduced total lifetime costs of the system. One of the most commonly used mechanical systems is the internal combustion engine. Internal combustion engines dominate the automotive industry, and have numerous other applications in generation, transportation, etc. This thesis presents the development of a fault detection and diagnosis (FDD) system for use with an internal combustion engine valve train.

A FDD system was developed with a focus on the valve impact amplitudes. Engine cycle averaging and band-pass filtering methods were tuned and utilized for improving the signal to noise ratio. A novel feature extraction method was developed that included a local RMS sliding window method and an adaptive threshold. Faults were seeded in the form of deformed valve springs, as well as abnormal valve clearances. The engine’s manufacturer specifies that a valve spring with 3 mm or more of deformation should be replaced. This thesis investigated the detection of a relatively small 0.5mm spring deformation. Valve clearance values were adjusted 0.1mm above and below the nominal clearance value (0.15mm) to test large clearance faults (0.25mm) and small clearance faults (0.05mm). The performance of the FDD system was tested using an instrumented diesel engine test bed. A comparison of numerous signal processing techniques and classification methods was performed.

The FDD system implementing the Naïve-Bayes classification method produced a worst case detection accuracy (DA) of 99.95% and worst case classification accuracy (CA) of 99.95% for spring faults of 0.5mm deformation, tested on multiple valves with a training size of 40 engine cycles. The total FDD execution time including feature extraction, training, and testing over 11,000 engine cycles was 4.5s. Alternative classification methods also worked well with the FDD system, with decision trees and linear discriminant analysis producing worst case CAs of 98.96% and 97.77%, respectively. Further experimental investigations were done where fault scenarios were varied, including simultaneous fault scenarios, and numerous parameter values were altered. The proposed FDD method gave reliable and accurate classification results for many different cases, demonstrating the generality and robustness of the proposed method.