Search

In recent years, a considerable amount of effort has been devoted, both in industry and academia, to improving maintenance. Time is a critical factor in maintenance, and efforts are placed to monitor, analyze, and visualize machine or asset data in order to anticipate to any possible failure, prevent damage, and save costs. The MANTIS Book aims to highlight the underpinning fundamentals of Condition-Based Maintenance related conceptual ideas, an overall idea of preventive maintenance, the economic impact and technical solution. The core content of this book describes the outcome of the Cyber-Physical System based Proactive Collaborative Maintenance project, also known as MANTIS, and funded by EU ECSEL Joint Undertaking under Grant Agreement nº 662189. The ambition has been to support the creation of a maintenance-oriented reference architecture that support the maintenance data lifecycle, to enable the use of novel kinds of maintenance strategies for industrial machinery. The key enabler has been the fine blend of collecting data through Cyber-Physical Systems, and the usage of machine learning techniques and advanced visualization for the enhanced monitoring of the machines. Topics discussed include, in the context of maintenance: Cyber-Physical Systems, Communication Middleware, Machine Learning, Advanced Visualization, Business Models, Future Trends. An important focus of the book is the application of the techniques in real world context, and in fact all the work is driven by the pilots, all of them centered on real machines and factories. This book is suitable for industrial and maintenance managers that want to implement a new strategy for maintenance in their companies. It should give readers a basic idea on the first steps to implementing a maintenance-oriented platform or information system.

Proactive and collaborative decision support is an integral part of a proactive maintenance strategy. Condition monitoring is only one part of the equation, since the actual assets’ condition needs to be combined with other relevant information, such as customer needs, ambient conditions, business models and service contracts in order for a service team to make the right decision and take the appropriate actions. The goal of the MANTIS project is to provide the service team with an intuitive, proactive and context-aware system for industrial maintenance that supports maintenance by proactively pushing relevant information to the right people at the right time, by intelligently filtering and summarizing information to prevent information overload through context awareness, by automatically and dynamically scheduling and adapting maintenance plans, thereby keeping the human in the loop at all times. The chapter is organized as follows: • Section 5.2 focuses on root cause analysis providing a short introduction to the theoretical background and a catalogue of techniques that have been demonstrated to be useful in this kind of analysis. Such techniques are mainly based on statistical and machine learning approaches. Moreover, a set of real-world applications are shortly introduced; • Section 5.3 deals with the identification of the remaining useful life of components. The section provides a short theoretical background, a catalogue of the useful approaches, and an analysis of some use cases to demonstrate the applicability of the described techniques analysing different modelling approaches; • Section 5.4 investigates how to alert and predict failures. The section provides an extensive catalogue of useful techniques. The presented techniques are also mainly based on statistical and machine learning approaches; • Section 5.5 provides some real examples that come from different application domains where most of the techniques previously presented have been applied with valuable results.

The elevated complexity and costs of production assets combined with the requirements for high-quality manufactured products necessitate novel design and condition-based maintenance approaches that are able to provide the required levels of availability, maintainability, quality and safety while decreasing the cost of the system as a whole and throughout the production lifecycle.

In recent years, a considerable amount of effort has been devoted, both in industry and academia, to improving maintenance. Time is a critical factor in maintenance, and efforts are placed to monitor, analyze, and visualize machine or asset data in order to anticipate to any possible failure, prevent damage, and save costs. The MANTIS Book aims to highlight the underpinning fundamentals of Condition-Based Maintenance related conceptual ideas, an overall idea of preventive maintenance, the economic impact and technical solution. The core content of this book describes the outcome of the Cyber-Physical System based Proactive Collaborative Maintenance project, also known as MANTIS, and funded by EU ECSEL Joint Undertaking under Grant Agreement n� 662189. The ambition has been to support the creation of a maintenance-oriented reference architecture that support the maintenance data lifecycle, to enable the use of novel kinds of maintenance strategies for industrial machinery. The key enabler has been the fine blend of collecting data through Cyber-Physical Systems, and the usage of machine learning techniques and advanced visualization for the enhanced monitoring of the machines. Topics discussed include, in the context of maintenance: Cyber-Physical Systems, Communication Middleware, Machine Learning, Advanced Visualization, Business Models, Future Trends. An important focus of the book is the application of the techniques in real world context, and in fact all the work is driven by the pilots, all of them centered on real machines and factories. This book is suitable for industrial and maintenance managers that want to implement a new strategy for maintenance in their companies. It should give readers a basic idea on the first steps to implementing a maintenance-oriented platform or information system.

The elevated complexity and costs of production assets combined with the requirements for high-quality manufactured products necessitate novel design and condition-based maintenance approaches that are able to provide the required levels of availability, maintainability, quality and safety while decreasing the cost of the system as a whole and throughout the production lifecycle.

The purpose of this chapter is to describe the MANTIS reference architecture. The generic focus here is on an architecture that enables service-based business models and improved asset availability at lower costs through continuous process and equipment monitoring, aided by big data analysis. This architecture takes into account needs of various industries in the forefront of service-based business and operations models. It also takes into account less mature industrial domains, where improvements in maintenance can be only achieved gradually and consistently. The higher level requirements for the whole project are described by [The MANTIS Consortium, 2018] and further tuned for the architecture in [Jantunen et al., 2016].

The most important goal of a power plant operator is to maximize the availability and performance of their power plant and to minimize the generating costs of energy. To reach this goal, the maintenance of the power plant is crucial and maintenance can account up to 30 % of the cost of the produced energy.Most modern power plants use a combination of scheduled maintenance and corrective maintenance. In this strategy, critical components, such as turbines and pumps, are serviced based on statistical trend data and less critical components are left to a run-to-failure strategy, where maintenance is not done until the machinery fails. These maintenance strategies are not cost optimized since scheduled maintenance usually results in either unnecessary repairs if servicing is done too frequently or potentially catastrophic failures if service is neglected. Ineffective maintenance not only wastes materials and resources, but lost operating time resulting from equipment failure is a significant expenditure to a power plant operator.Effective maintenance planning and scheduling is essential for shortening revision and downtime periods. Most of the revision work is done by external contractors and work delays can be expensive. By maintaining components based on their condition, revision work and cost can be optimized.Current condition based maintenance strategies usually rely on periodical measurements done by experts, because the power plant staff do not have the necessary skills to analyse the measurement data. Most power plants collect information that could be utilized in condition based maintenance, but the data is not used because it is difficult to interpret or because it lacks parameters for fault prediction. Implementing new sensors in power plant equipment can be difficult.

In recent years, a considerable amount of effort has been devoted, both in industry and academia, to improving maintenance. Time is a critical factor in maintenance, and efforts are placed to monitor, analyze, and visualize machine or asset data in order to anticipate to any possible failure, prevent damage, and save costs. The MANTIS Book aims to highlight the underpinning fundamentals of Condition-Based Maintenance related conceptual ideas, an overall idea of preventive maintenance, the economic impact and technical solution. The core content of this book describes the outcome of the Cyber-Physical System based Proactive Collaborative Maintenance project, also known as MANTIS, and funded by EU ECSEL Joint Undertaking under Grant Agreement nº 662189. The ambition has been to support the creation of a maintenance-oriented reference architecture that support the maintenance data lifecycle, to enable the use of novel kinds of maintenance strategies for industrial machinery. The key enabler has been the fine blend of collecting data through Cyber-Physical Systems, and the usage of machine learning techniques and advanced visualization for the enhanced monitoring of the machines. Topics discussed include, in the context of maintenance: Cyber-Physical Systems, Communication Middleware, Machine Learning, Advanced Visualization, Business Models, Future Trends. An important focus of the book is the application of the techniques in real world context, and in fact all the work is driven by the pilots, all of them centered on real machines and factories. This book is suitable for industrial and maintenance managers that want to implement a new strategy for maintenance in their companies. It should give readers a basic idea on the first steps to implementing a maintenance-oriented platform or information system.

This book aims to highlight the fundamental underpinnings of Condition-Based Maintenance and related conceptual ideas. It also presents an overall view of i) proactive maintenance, and ii) condition-based maintenance, including its potential economic impact and latest technical solutions.