The condition monitoring of multi-component systems utilizes multiple sensors to capture the functional condition of the systems and allows the sensor information to be used to reason about the health information of the systems or components. Chapter 3 considers the situation when sensor signals capture unknown mixtures of component signals and proposes a two-stage vibration-based methodology to identify component degradation signals from mixed sensor signals in order to predict component-level residual lives. Specifically, we are interested in modeling the degradation of systems that consist of two or more identical components operating under similar conditions. Chapter 4 focuses on the interactive relationship between tool wear (component degradation) and product quality degradation (sensor information) that widely exists in multistage manufacturing processes and proposes a high-dimensional stochastic differential equation model to capture the interaction relationship. Then, real-time quality measurements are incorporated to online predict the residual life of the system. Chapter 5 develops a strategy of dynamic workload adjustment for parallel multi-component systems in order to control the degradation processes and failure times of individual components, for the purpose of preventing the overlap of component failures. This chapter opens a new research direction that focuses on the active control of degradation rather than only the modeling part.

This research develops an equipment failure prognostics model to predict the equipment’s chance of survival, using LAD. LAD benefits from not relying on any statistical theory, which enables it to overcome the problems concerning the statistical properties of the datasets. Its main advantage is its straightforward process and self-explanatory results. Herein, our main objective is to develop models to calculate equipment’s survival probability at a certain future moment, using LAD. We employ the LAD’s pattern generation procedure. Then, we introduce a guideline to employ generated patterns to estimate the equipment’s survival probability. The models are applied on a condition monitoring dataset. Performance analysis reveals that they provide comprehensible results that are greatly beneficial to maintenance practitioners. Results are compared with PHM’s results. The comparison reveals that the LAD models compare favorably to the PHM. Since they are at their beginning phase, some future directions are presented to improve their performances.

The aim of this study is to propose an alternative approach in prognostics for airborne avionics system in order to enhance maintenance process and aircraft availability. The objectives are to analyse the dependency of avionic systems for fault propagation behaviour degradation, research and develop methods to predict the remaining useful life of avionics Line Replaceable Units (LRU), research and develop methods to evaluate and predict the degradation performances of avionic systems, and lastly to develop software simulation systems to evaluate methods developed. One of the many stakeholders in the aircraft lifecycle includes the Maintenance, Repair and Overhaul (MRO) industry. The predictable logistics process to some degree as an outcome of IVHM gives benefit to the MRO industry. In this thesis, a new integrated numerical methodology called ‘System Level Airborne Avionic Prognostics’ or SLAAP is developed; looking at a top level solution in prognostics. Overall, this research consists of two main elements. One is to thoroughly understand and analyse data that could be utilised. Secondly, is to apply the developed methodology using the enhanced prognostic methodology. Readily available fault tree data is used to analyse the dependencies of each component within the LRUs, and performance were simulated using the linear Markov Model to estimate the time to failure. A hybrid approach prognostics model is then integrated with the prognostics measures that include environmental factors that contribute to the failure of a system, such as temperature. This research attempts to use data that is closest to the data available in the maintenance repair and overhaul industry. Based on a case study on Enhanced Ground Proximity Warning System (EGPWS), the prognostics methodology developed showed a sufficiently close approximation to the Mean Time Before Failure (MTBF) data supplied by the Original Equipment Manufacturer (OEM). This validation gives confidence that the proposed methodology will achieve its objectives and it should be further developed for use in the systems design process.

A large volume of data has become commonplace in many domains these days. Machine learning algorithms can be trained to look for any useful hidden patterns in such data. Sometimes, these big data might need to be summarized to make them into a manageable size, for example by using histograms, for various reasons. Traditionally, machine learning algorithms can be trained on data expressed as real numbers and/or categories but not on a complex structure such as histogram. Since machine learning algorithms that can learn from data with histograms have not been explored to a major extent, this thesis intends to further explore this domain. This thesis has been limited to classification algorithms, tree-based classifiers such as decision trees, and random forest in particular. Decision trees are one of the simplest and most intuitive algorithms to train. A single decision tree might not be the best algorithm in term of its predictive performance, but it can be largely enhanced by considering an ensemble of many diverse trees as a random forest. This is the reason why both algorithms were considered. So, the objective of this thesis is to investigate how one can adapt these algorithms to make them learn better on histogram data. Our proposed approach considers the use of multiple bins of a histogram simultaneously to split a node during the tree induction process. Treating bins simultaneously is expected to capture dependencies among them, which could be useful. Experimental evaluation of the proposed approaches was carried out by comparing them with the standard approach of growing a tree where a single bin is used to split a node. Accuracy and the area under the receiver operating characteristic (ROC) curve (AUC) metrics along with the average time taken to train a model were used for comparison. For experimental purposes, real-world data from a large fleet of heavy duty trucks were used to build a component-failure prediction model. These data contain information about the operation of trucks over the years, where most operational features are summarized as histograms. Experiments were performed further on the synthetically generated dataset. From the results of the experiments, it was observed that the proposed approach outperforms the standard approach in performance and compactness of the model but lags behind in terms of training time. This thesis was motivated by a real-life problem encountered in the operation of heavy duty trucks in the automotive industry while building a data driven failure-prediction model. So, all the details about collecting and cleansing the data and the challenges encountered while making the data ready for training the algorithm have been presented in detail.

No description available.

Yes / Diagnostics and prognostics have significant roles in the reliability enhancement of systems and are focused topics of active research. Engineered systems are becoming more complex and are subjected to miscellaneous failure modes that impact adversely their performability. This everincreasing complexity makes fault diagnostics and prognostics challenging for the system-level functions. A significant number of successes have been achieved and acknowledged in some review papers; however, these reviews rarely focused on the application of complex engineered systems nor provided a systematic review of diverse techniques and approaches to address the related challenges. To bridge the gap, this paper firstly presents a review to systematically cover the general concepts and recent development of various diagnostics and prognostics approaches, along with their strengths and shortcomings for the application of diverse engineered systems. Afterward, given the characteristics of complex systems, the applicability of different techniques and methods that are capable to address the features of complex systems are reviewed and discussed, and some of the recent achievements in the literature are introduced. Finally, the unaddressed challenges are discussed by taking into account the characteristics of automotive systems as an example of complex systems. In addition, future development and potential research trends are offered to address those challenges. Consequently, this review provides a systematic view of the state of the art and case studies with a reference value for scholars and practitioners.

Industry 4.0 push forward the development of concepts such as artificial intelligence, big data, and Industrial Internet of Things, which claims an evolution of the monitoring systems design in terms of the accessibility to the information. In this project, the author describes the design of a condition monitoring system to monitor the state of different components of an extrusion line and propose a system that allows predictive maintenance in industry, specifically an extrusion system. In this framework, Condition Based Maintenance, CBM, the health state of the component is continuously monitored. In some cases, the monitoring can be periodical. The goal is to make repairs in the opportune moment, by receiving data of malfunction, so the efficiency is maxed. The aim is to develop a system to monitor the state of different components of an extrusion line and to propose a system that allows predictive maintenance in industry and that the project can be used as a guideline to a complete condition monitoring system implementation in an industrial environment. To be able to achieve this, some first steps must be accomplished. These are a Taxonomy, or a breakdown of the system into individual elements and how they are related; and carrying out a Failure Modes Effect and Criticality Analysis, known as FMECA. With these studies, the author displays the failure modes that are critical for the operating of the system and thus, which have the most likelihood to occur while having a big impact. From the information extracted the author presents a model based in accelerometer, temperature and MCSA (Motor Current Spectral Analysis) sensors. Furthermore, the data obtained will need to be analysed. With that in mind, the operating frequencies as well as the failure modes frequencies must be studied, which will allow correct identification while analysing the data. This analysis will be done by characterizing data and applying analysing techniques as FFT or Hilbert.

No description available.

No description available.

No description available.