Reliability engineering advancements to enhance fleet asset management in service-oriented business models
- DIRECTORS: Adolfo Crespo and Óscar Salgado
- UNIVERSITY: Universidad de Sevilla
This thesis is framed in the context of asset management, more specifically it is focused on contributing to the development and advance of reliability engineering through a value proposition that considers multi-disciplinary aspects. The contributions in this thesis aim to improve decision-making in the fleet asset management through an innovative approach that combines the capabilities of automatic learning algorithms with the advantages and benefits of service-oriented business models.
The present manuscript is divided into five chapters containing the different aspects of the research. The first chapter introduces the subject and context in which this thesis is developed along with an explanation of the main foundations of the research. This first chapter also explores the problem to be solved and the research questions that will guide the lines of evolution of the thesis’ work. Then chapter 2 gathers all the related works and concepts that constitute the basis of the research. The second chapter is divided into a literature review section that includes works by authors in the disciplines of asset management, servitization and PSS, maintenance management, and reliability engineering. On the other part, chapter 2 includes the theoretical background containing the mathematical developments and algorithms that constitute the basis of the developments in the thesis. Chapter 3 presents the methodology that has been followed and the phases that compose it: exploratory, developmental, verification and validation, and results. Additionally, this chapter presents the contributions of the thesis, two developments that propose important advances in reliability engineering, and a proposal that considers multiple methodologies, methods, and tools to realize value in different industries from the two developments. This third chapter also includes a summary of the different publications considered within the scope of this thesis. The results derived from the application of the contributions to different industries are presented in the fourth chapter. More specifically, the chapter considers the application of the developments in the railway industry, in the aerospace industry, and in wind energy projects. Finally, the last chapter contains the conclusions of the thesis, which consider general aspects as well as research considerations. This last section also considers future lines of research.
The advances in reliability engineering presented in this thesis aim at reducing the level of uncertainty currently involved in fleet asset management. By the proposal of an innovative dynamic reliability model based on artificial neural networks, it is possible to model the impact of the operational context on the behaviour of the assets. By the proposal of this model, it is possible to advance towards the customization of maintenance plans according to the operational context, and therefore towards service scenarios in which the offer of the OEMs is a bundle of the asset and the services known as PSS. In addition, to improve fleet asset management in this paradigm, a clustering approach based on assets’ reliability is proposed. Through this approach, it is possible to segment the fleet into groups in which the assets have similar failure behaviour and therefore it allows managing the maintenance needs of the fleet in a structured and systematic approach. This thesis also contributes to the body of knowledge with methodologies and tools to translate these developments into added value. The thesis contributions are important considerations regarding the necessary information, data gathering and quality level to apply the latest advances in reliability engineering. Besides, important advances are proposed for maintenance management, among others to be able to integrate reliability-based approaches with condition-based maintenance approaches and to be able to evaluate the feasibility and profitability of implementing advanced reliability and maintenance management models. This thesis also explores the need for technological capacitation in order to adopt these approaches in a cost-effective and efficient way, and contributes with a tool that facilitates the process of technological capacitation that OEMs have to undergo to benefit from a business model based on maintenance servitization.
The application of the developments through case studies in the railway sector, in the aerospace sector, and in the wind energy sector, shows that through the proposals of this thesis it is possible to reduce the level of uncertainty involved in the management of fleets of assets. The proposed models demonstrate that they are capable of identifying the impact that operating conditions have on the failure frequency and that they make it possible to estimate the reliability of the assets based on how they are operating; even considering changes in the operational context and other possible phenomena. In addition, the clustering approach facilitates fleet management and has demonstrated that integrated with the dynamic reliability model based on neural networks, it is capable of supporting maintenance strategies that lead to higher levels of availability at lower costs than the current ones.
The research in the thesis has important practical implications for fleet management, especially in OEMs that are adopting service-oriented business models. The developments show important improvements with respect to current technologies and the methods and tools with which they are accompanied allow OEMs to offer services of greater added value. However, despite the potential shown by the value proposals of this thesis, it is necessary to deepen the research along the lines described in the last part of the thesis through future research work.