Kernel methods are able to exploit high-dimensional spaces for representational advantage, while only operating implicitly in such spaces, thus incurring none of the computational cost of doing so. They appear to have the potential to advance the state of the art in control and signal processing applications and are increasingly seeing adoption across these domains. Applications of kernel methods to fault detection and isolation (FDI) have been reported, but few in aerospace research, though they offer a promising way to perform or enhance fault detection. It is mostly in process monitoring, in the chemical processing industry for example, that these techniques have found broader application. This research work explores the use of kernel-based solutions in model-based fault diagnosis for aerospace systems. Specifically, it investigates the application of these techniques to the detection and isolation of IMU/INS sensor faults – a canonical open problem in the aerospace field. Kernel PCA, a kernelised non-linear extension of the well-known principal component analysis (PCA) algorithm, is implemented to tackle IMU fault monitoring. An isolation scheme is extrapolated based on the strong duality known to exist between probably the most widely practiced method of FDI in the aerospace domain – the parity space technique – and linear principal component analysis. The algorithm, termed partial kernel PCA, benefits from the isolation properties of the parity space method as well as the non-linear approximation ability of kernel PCA. Further, a number of unscented non-linear filters for FDI are implemented, equipped with data-driven transition models based on Gaussian processes - a non-parametric Bayesian kernel method. A distributed estimation architecture is proposed, which besides fault diagnosis can contemporaneously perform sensor fusion. It also allows for decoupling faulty sensors from the navigation solution.

Rotating machinery is a critical asset of industrial plants worldwide. Bearings and gearboxes are two of the most common components found in rotating machinery of industrial plants. The malfunction of bearings and gearboxes lead the machine to fail and often these failures occur catastrophically leading to personnel injuries. Therefore it is of high importance to identify the deterioration at an early stage. Among the techniques applied to detect damage in rotating machinery, acoustic emission has been a prevalent field of research for its potential to detect defects at an earlier stage than other more established techniques such as vibration analysis and oil analysis. However, to reliably detect the fault at an early stage de-noising techniques often must be applied to reduce the AE noise generated by neighbouring components and normal component operation. For this purpose a novel signal processing algorithm has been developed combining Wavelet Packets as a pre-processor, Hilbert Transform, Autocorrelation function and Fast Fourier transform. The combination of these techniques allows identification of g repetitive patterns in the AE signal that are attributable to bearing and gear damage. The enhancement for early stage defect detection in bearings and gears provided by this method is beneficial in planning maintenance in advance, reducing machinery down-time and consequently reducing the costs associated with bearing breakdown. The effectiveness of the proposed method has been investigated experimentally using seeded and naturally developed defects in gears and bearings. In addition, research into the optimal Wavelet Packet node that offers the best de-noising results has been performed showing that the 250-750 kHz band gives the best SNR results. The detection of shaft angular misalignment using Acoustic Emission has been investigated and compared with acceleration spectra. The results obtained show enhancements of AE in detection shaft angular misalignment over vibration analysis in SNR and stability with varying operational conditions.