Advances in machine condition monitoring technologies are driven by the rise in complexity of modern machines and the increased demand for product reliability. Condition monitoring research tends to focus on the development of signal processing algorithms that are sensitive to machine faults, robust under time-varying operating conditions, and informative regarding the nature and extent of machine faults. A significant challenge remains for monitoring the condition of machines that are subject to time-varying operating conditions. The here presented work is concerned with the development of cost effective condition monitoring algorithms. It is investigated how empirical models (including probability density distributions and regression functions) may be used to extract diagnostic information from machine response signals that have been generated under fluctuating operating conditions. The proposed methodology is investigated on a number of case studies, including gearboxes, alternator end windings, and haul roads. It is shown how empirical models for machine condition monitoring may generally be implemented according to one of two basic approaches. The two approaches are referred to as discrepancy analysis and waveform reconstruction. Discrepancy analysis is concerned with the comparison of a novel signal to a reference model. The reference model is sufficiently expressive to represent vibration response as measured on a healthy machine over a range of operating conditions. The novel signal is compared to the reference model in such a manner that a discrepancy signal transform is obtained. A discrepancy signal is sensitive to faults, robust to time-varying operating conditions, and inherently simple. As such it may further beWaveform reconstruction implements a regression function to model machine response as a function of different state space variables. The regression function may subsequently be exploited to extract diagnostic information. The machine response may for instance be reconstructed at a specified steady state operating condition. This renders the signal wide-sense stationary so that Fourier analysis may be applied. analysed in order to extract periodicities and magnitudes as diagnostic markers. / Dissertation (MEng)--University of Pretoria, 2013. / gm2014 / Electrical, Electronic and Computer Engineering / unrestricted

Vibration-based condition monitoring is a key and crucial element for asset longevity and to avoid unexpected financial compromise. Currently, data-driven methodologies often require significant investments into data acquisition and a large amount of operational data for both healthy and unhealthy cases. The acquisition of unhealthy fault data is often financially infeasible and the result is that most methods detailed in literature are not suitable for critical industrial applications. In this work, unsupervised latent variable models negate the requirement for asset fault data. These models operate by learning the representation of healthy data and utilise health indicators to track deviance from this representation. A variety of latent variable models are compared, namely: Principal Component Analysis, Variational Auto-Encoders and Generative Adversarial Network-based methods. This research investigated the relationship between time-series data and latent variable model design under the sensible notion of data interpretation, the influence of model complexity on result performance on different datasets and shows that the latent manifold, when untangled and traversed in a sensible manner, is indicative of damage. Three latent health indicators are proposed in this work and utilised in conjunction with a proposed temporal preservation approach. The performance is compared over the different models. It was found that these latent health indicators can augment standard health indicators and benefit model performance. This allows one to compare the performance of different latent variable models, an approach that has not been realised in previous work as the interpretation of the latent manifold and the manifold response to anomalous instances had not been explored. If all aspects of a latent variable model are systematically investigated and compared, different models can be analysed on a consistent platform. In the model analysis step, a latent variable model is used to evaluate the available data such that the health indicators used to infer the health state of an asset, are available for analysis and comparison. The datasets investigated in this work consist of stationary and time-varying operating conditions. The objective was to determine whether deep learning is comparable or on par with state-of-the-art signal processing techniques. The results showed that damage is detectable in both the input space and the latent space and can be trended to identify clear condition deviance points. This highlights that both spaces are indicative of damage when analysed in a sensible manner. A key take away from this work is that for data that contains impulsive components that manifest naturally and not due to the presence of a fault, the anomaly detection procedure may be limited by inherent assumptions made in model formulations concerning Gaussianity. This work illustrates how the latent manifold is useful for the detection of anomalous instances, how one must consider a variety of latent-variable model types and how subtle changes to data processing can benefit model performance analysis substantially. For vibration-based condition monitoring, latent variable models offer significant improvements in fault diagnostics and reduce the requirement for expert knowledge. This can ultimately improve asset longevity and the investment required from businesses in asset maintenance. / Dissertation (MEng (Mechanical Engineering))--University of Pretoria, 2020. / Eskom Power Plant Engineering Institute (EPPEI) / UP Postgraduate Bursary / Mechanical and Aeronautical Engineering / MEng (Mechanical Engineering) / Unrestricted