From Prognostics and Health Systems Management to Predictive Maintenance 2

Knowledge, Reliability and Decision

Brigitte Chebel-Morello, Jean-Marc Nicod, Christophe Varnier

• Publisher: Wiley-ISTE
• Year: 2017
• ISBN:
  9781119436850
• Edition: 1
• Topic: Technology & Engineering
• Subtopic: Aeronautic & Astronautic Engineering
• Index: Technology & Engineering

This book is the second volume in a set of books dealing with the evolution of technology, IT and organizational approaches and what this means for industrial equipment.

The authors address this increasing complexity in two parts, focusing specifically on the field of Prognostics and Health Management (PHM). Having tackled the PHM cycle in the first volume, the purpose of this book is to tackle the other phases of PHM, including the traceability of data, information and knowledge, and the ability to make decisions accordingly.

The book concludes with a summary analysis and perspectives regarding this emerging domain, since without traceability, knowledge and decision, any prediction of the health state of a system cannot be exploited.

PART 1
Traceability of Information and Knowledge Management

1
Intelligent Traceability of Equipment

1.1. Introduction

Over the last few decades, the business awareness of companies has been based on their intangible capital, the knowledge and expertise of their employees. Many works have been dedicated to creating, sharing and capitalizing on this expert knowledge.

However, just like expertise and practical skills, data also contributes to the intangible capital of companies and it is carefully recorded. Banks jealously guard their databases. Companies distrust the cloud due to security issues related to data, which may fall into the wrong hands and be exploited by competitors or harm data owners. In the age of connected objects, of easy data harvesting and instantaneous remote access, it is evident that well-exploited data represents a gold mine. Indeed, data is becoming a resource which can be exploited by the economy. Data must be secured and exploited and this is the basis of informed decision-making within a domain.

The process of knowledge capitalization corresponds to the notion of knowledge management, which was defined by Davenport as the process of collecting, distributing and effectively using knowledge [DAV 94], an approach developed by [DUH 98]. Traceability is an essential element of the capitalization of knowledge related to different stages of a product’s evolution [BIS 08]. Several works have highlighted the notion of the growth of products’ lifecycle management as one of knowledge management [AME 05] and [STA 15]. In fact, the research consortium of the project FP6 IP 507 100 PROMISE (PROduct lifecycle Management and Information tracking using Smart Embedded systems) has remarked that traditional PLM systems lack product knowledge and visibility in the two MOL and EOL phases, and has recommended developing traceability and knowledge capitalization during the lifecycle.

The aim of our work is to design and develop a solution for processing and capitalizing knowledge related to industrial equipment throughout its lifecycle and for making it available for operators to easily access this equipment in an understandable way and at the required moment.

New PLM possibilities [RAN 11] are being introduced, thanks to continuous developments in the domain of information systems regarding radio-frequency identification (RFID), sensor network technology and, more generally, in product embedded information devices (PEID). A new generation of products called smart or intelligent products is being developed [KIR 11]. According to [YAN 09], it makes the information easily accessible for designers, users or disassemblers of the product or equipment. However, although these intelligent products are capable of gathering data during their lifecycle, they lack the means of extracting information and acquiring knowledge from this data. To reach this goal, we have tackled three challenges:

• – Creation of a so-called intelligent product which allows users access to reliable information capable of being read or manipulated, as well as an available deduction related to the current health state of the product;
• – Transformation of data in knowledge in a memory that stores all the information concerning the product during its lifecycle and which can be accessed from the product;
• – Proposition for decision support services, online prognostics and monitoring of the health state of equipment and support services for maintenance and recycling of products. These services should be available via an information system which can be easily accessed through the product.

We will address these challenges in three stages: (i) after having defined what intelligent equipment is and having considered the work in this domain, we will orient ourselves toward the data exchange infrastructure CL2M, featuring RFID tags connected to an e-maintenance platform and equipped with deduction tools. (ii) We have developed a knowledge capitalization process that stores the knowledge in an operating memory which is distributed on the equipment and the e-maintenance platform. The knowledge is formalized according to the maintenance ontology IMAMO_RFID and made is available by means of an intelligent product infrastructure that ensures knowledge sharing. (iii) We have developed web services that require the availability of information regarding the state of (mal)functioning of the equipment along its lifecycle. In this way, we propose different decision support services:

• – a support service for recycling components (products), in which data is indispensable in order to provide such a service [SIM 00];
• – a support service for the monitoring and prognostic processes of the health state of the component;
• – a support service for maintenance action planning.

As a result, this chapter will begin by outlining some state-of-the-art intelligent products followed in section 1.3 by a presentation of a knowledge capitalization process that monitors a component’s health state along its lifecycle, and the proposition of an ontology called IMAMO_RFID, defined for intelligent products. Section 1.4 will be dedicated to the infrastructure of an intelligent product with the exchange of data and information and the implementation of decision support services.

1.2. State-of-the-art intelligent products

1.2.1. Definition of intelligent products

In order to monitor a product during its MOL phase, this product has to be intelligent in McFarlane’s sense. [MCF 03] defines the product via a physical and informational representation stored in a database, which is associated with an intelligence provided by a decision support agent. An intelligent product is characterized by five main properties: possession of a unique identify, a capability to communicate effectively with its environment, ability to retain or store data about itself, and a potential to participating in or making decisions relevant to its own destiny. Other definitions of intelligent products exist [MEY 09, MCF 03] as listed [KIR 11] by Kiritsis, who synthesized them by defining an intelligent product as a system containing sensing, memory, data processing, reasoning and communication capabilities. He listed four intelligence levels ranging from a physical product without any embedded system to products with product embedded information devices (PIED).

1.2.2. Research on intelligent products

Research on intelligent products suggests integrating the product within an infrastructure of processed data sharing and lifecycle management. Ranasinghe et al. [RAN 11] have studied three Product Lifecycle Information Management architectures (PLIM) for collecting and accessing product data, both industrial and academic ones: (i) the EPC network architecture (Electronic Product Code), (ii) the DIALOG system (Distributed Information Architectures for collaborative logistics) and (iii) the WWAI network (World Wide Article Information). These architectures show some weaknesses regarding data synchronization when network disturbances occur. In other words, without a network connection there is no access to databases and the update cannot take place. Xiaoyu [YAN 09] remarks that the intelligent product should have some kind of mechanism that could use the collected lifecycle data to provide services. He suggests that an intelligent product should contain at least three fundamental elements: (i) an intelligent data unit (IDU), (ii) an access service and (iii) a communication infrastructure (CSI).

Z.Y. Wu et al. [WU 14] provide a software platform for lifecycle knowledge management that allows systematic assistance for the development of a product, in particular for new heating valves in nuclear power plants. For this purpose, record data has been exploited in order to support the design of the products (fault types, causes and recommendations regarding the actions to be undertaken).

Kiritsis proposes a closed-loop infrastructure for product lifecycle management, the CL2M (Closed-Loop Product Lifecycle Management), which was developed during the European project PROMISE [KIR 11]. T...

Table of contents

• Cover
• Table of Contents
• Title
• Copyright
• Introduction
• PART 1: Traceability of Information and Knowledge Management
• PART 2: Post-prognostic Decision
• Conclusion
• Bibliography
• Index
• End User License Agreement