Deliverable D4.1 Z-BRE4K strategies and policies · Z-BRE4K Project Grant Agreement nº 768869 –...

Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017

D4.1 V1.0 Page 1/ 55

Grant agreement nº: 768869

Call identifier: H2020-FOF-2017

Strategies and Predictive Maintenance models wrapped around physical

systems for Zero-unexpected-Breakdowns and increased operating life

of Factories

Z-BRE4K

Deliverable D4.1 Z-BRE4K strategies and policies

Work Package 4

Design of Strategies &

Integration of Intelligence

Document type : Report

Version : V1.0

Date of issue : 20th March 2019 (M18)

Dissemination level : Public

Lead beneficiary : 13 – EPFL

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement nº 768869. The dissemination of results herein reflects only the author’s view and the European Commission is not responsible for any use that may be made of the information it contains.

The information contained in this report is subject to change without notice and should not be construed as a commitment by any members of the Z-BRE4K Consortium. The information is provided without any warranty of any kind. This document may not be copied, reproduced, or modified in whole or in part for any purpose without written permission from the Z-BRE4K Consortium. In addition to such written permission to copy, acknowledgement of the authors of the document and all applicable portions of the copyright notice must be clearly referenced. © COPYRIGHT 2017 The Z-BRE4K Consortium.

All rights reserved.


D4.1 V1.0 Page 2/ 55

Executive Summary

Abstract

This document reports the results and deliverable of Task 4.1,

i.e. Z-BRE4K strategies and policies. Accordingly, taking into

account existing plant strategies and polices from the pilots

(i.e. T1.1 & T1.4), this Task T4.1 and the associated

deliverable D4.1 focus on updating the existing and

development of new strategies to improve maintainability

and operating life of production systems. Our approach

follows a method to translate optimization objectives defined

at production and factory levels, into optimized maintenance

policies at asset/production process levels. Starting with the

existing (i.e. corrective and time-based preventive)

maintenance policies and their particular strategies of Z-

BRE4K end-users SACMI, GESTAMP, and PHILIPS, the

deliverable highlights the update of respective policies and

processes, instantiated through novel Z-BRE4K strategies to

cope with the offerings and findings of predictive tools: Z-

PREVENT/PREDICT/DIAGNOSE/REMEDIATE failures, Z-

ESTIMATE RUL of assets, Z-MANAGE alarms and mitigation

actions, and Z-SYNCHRONISE with shop floor operations and

plant management systems, while ensuring the Z-SAFETY of

workers.

Keywords Strategy implementation, Maintenance policies, Predictive

maintenance, Z-Strategies, Zero-breakdown, Manufacturing


D4.1 V1.0 Page 3/ 55

Revision history

Version Author(s) Changes Date

V0.1 Gökan MAY (EPFL) Deliverable outline 25/04/2018

V0.2 Gökan MAY (EPFL) Updated deliverable outline 14/05/2018

V0.3 Jovana Milenkovic

(ATLANTIS) Updated deliverable outline 01/06/2018

V0.4 – V0.5 Polivios Raxis (ATLANTIS) Added Content in Section 4.1 – 4.2 08/06/2018

12/06/2018

V0.6 – V0.9 Dimitrios Daskalakis

(ATLANTIS)

Added Content in Section 4.2 – 4.3

– 4.4

27/06/2018 -

09/07/2018

V0.10 Katerina Tsinari

(ATLANTIS) Section 3 12/07/2018

V0.11 Daniel Caljouw (PHILIPS) Section 5.3 12/07/2018

V0.12 Gökan MAY (EPFL) Added Section 2 Content & revised,

updated and formatted Section 3 18-20/07/2018

V0.13 Gökan MAY (EPFL) Added Section 1 Content & revised,

updated and formatted Section 4 02-03/08/2018

V0.14 Davide Baldisseri

(SACMI) Section 5.1 28/09/2018

V0.15 Gökan MAY (EPFL) References, Formatting and

Revision 05/11/2018

V0.16 Joaquín Piccini

(GESTAMP) Sections 5.2 03/12/2018

V0.17 SACMI, GESTAMP,

PHILIPS, EPFL Section 6 20/02/2019

V0.18 Gökan MAY (EPFL) Final edit and formatting 11/03/2019

V0.19 Jovan Milenkovic

(ATLANTIS) Peer-Review 20/03/2019

V0.20 Daniel Gesto Rodriguez

(AIMEN) Peer-Review 20/03/2019

V1.0 Gökan MAY (EPFL) Addressed Peer-Reviews & Final

edit and formatting 20/03/2019


D4.1 V1.0 Page 4/ 55

Table of Contents

EXECUTIVE SUMMARY .............................................................................................. 2

ABBREVIATIONS ....................................................................................................... 6

LIST OF FIGURES ....................................................................................................... 8

LIST OF TABLES ......................................................................................................... 9

1 INTRODUCTION ............................................................................................... 10

2 Z-BRE4K STRATEGIES OVERVIEW ...................................................................... 11

3 EMBEDDED INTELLIGENCE ............................................................................... 14

3.1 AS-IS Status in Embedded intelligence ............................................................................ 15

3.1.1 Biologically Inspired Embedded Systems ..................................................... 15

3.1.2 Multi-agent Systems .................................................................................... 15

3.2 TO-BE Status in Embedded intelligence........................................................................... 16

4 INDUSTRIAL MAINTENANCE STRATEGIES ......................................................... 18

4.1 Risk-based maintenance ................................................................................................. 19

4.2 Predictive Maintenance ................................................................................................... 22

4.3 Condition-based maintenance ........................................................................................ 25

4.4 Preventive Maintenance ................................................................................................. 28

4.5 Corrective Maintenance .................................................................................................. 31

5 AS-IS MAINTENANCE STRATEGIES AND POLICIES OF Z-BRE4K END-USERS ......... 34

5.1 SACMI .............................................................................................................................. 34

5.1.1 Production System ....................................................................................... 34

5.1.2 Plant Strategy and Policies ........................................................................... 35

5.2 GESTAMP ......................................................................................................................... 36



5.3 PHILIPS ............................................................................................................................. 38



6 TO-BE MAINTENANCE SCENARIOS OF THE END-USERS AFTER Z-BRE4K SOLUTION 40

6.1 SACMI’s Plant Maintenance Plan (TO-BE SCENARIO) ..................................................... 40

6.1.1 Scope ............................................................................................................ 40


D4.1 V1.0 Page 5/ 55

6.1.2 Data collection and analysis ......................................................................... 40

6.1.3 IoT (Sensor & Automation) Gateway and Maintenance Reports HMI ........ 41

6.1.4 Machine Simulators for Preventive, Predictive and Prescriptive

Maintenance through Machine Learning and physical model retrofitting ................ 42

6.1.5 Interface for Operations Management and coordination with MES ........... 43

6.2 GESTAMP’s Plant Maintenance Plan (TO-BE SCENARIO) ................................................ 43

6.2.1 Scope ............................................................................................................ 43


6.2.3 IoT (Sensor & Automation) Gateway and Maintenance Reports HMI ........ 45



6.2.5 Interface for Operations Management and coordination with MES ........... 49

6.3 PHILIPS’ Plant Maintenance Plan (TO-BE SCENARIO)...................................................... 50

6.3.1 Scope ............................................................................................................ 50


6.3.3 IoT (Sensor & Automation) Gateway ........................................................... 51



6.3.5 Interface for Operations and Maintenance ................................................. 52

7 CONCLUSION ................................................................................................... 53

REFERENCES ........................................................................................................... 54


D4.1 V1.0 Page 6/ 55

Abbreviations

Abbreviation Name

AI Artificial Intelligence

APM Asset Performance Management

CAD Computer Aided Design

CBM Condition Based Monitoring

CCM Continuous Compression Moulding

CECM Cognitive Embedded Condition Monitoring

CMMS Computerized Maintenance Management System

CoF Consequences of Failure

CTQ Critical to Quality

DCS Distributed Control System

EAM Enterprise Asset Management

ERP Enterprise Resource Planning

FMEA Failure Mode and Effects Analysis

FMECA Failure Mode, Effects, and Criticality Analysis

FTA Fault Tree Analysis

GA Genetic Algorithm

GUI Graphical User Interface

HMI Human Machine Interface

HW Hardware

ICT Information and Communication Technology

IDS Industrial Data Space

IIOT Industrial Internet of Things


D4.1 V1.0 Page 7/ 55

KPI Key Performance Indicator

KRI Key Risk Indicator

MAS Multi-agent System

MES Manufacturing Execution System

ML Machine Learning

NN Neural Network

OEM Original Equipment Manufacturer

OPC-UA OPC Unified Architecture

PCA Principal Component Analysis

PdM Predictive Maintenance

PLC Programmable Logic Controller

PM Preventive Maintenance

PoF Probability of Failure

RBM Risk Based Maintenance

RCA Root Cause Analysis

RCM Reliability Centred Maintenance

RUL Remaining Useful Life

SW Software

XML Extensible Mark-up Language


D4.1 V1.0 Page 8/ 55

List of Figures

Figure 1. Synergies and interactions between the eight Z-Strategies ........................................ 12

Figure 2. The maintenance maturity pyramid (Pathak, 2018) .................................................... 18

Figure 3. The maintenance plan organized by task prioritization ............................................... 19

Figure 4. Risk based Maintenance Framework (Fiix, 2018) ........................................................ 20

Figure 5. Risk Matrix and risk-based decision making ............................................................... 21

Figure 6. Benefits of Predictive Maintenance ............................................................................. 22

Figure 7. Predictive Maintenance within Industrial revolution ................................................... 24

Figure 8. PdM Maturity Matrix ................................................................................................... 25

Figure 9. CBM optimises costs between preventive and corrective maintenance (Toms, 1995) 26

Figure 10. Preventive Maintenance Philosophy .......................................................................... 28

Figure 11. Maintenance task creation ........................................................................................ 30

Figure 12. Preventive Maintenance Task .................................................................................... 31

Figure 13. Scheduling view and calendar .................................................................................... 31

Figure 14. Corrective Maintenance Processing ........................................................................... 33

Figure 15. Enriched FMECA file with sensor & alarm information associated (SACMI) .............. 41

Figure 16. Machine simulators module and Predictive Maintenance module ............................ 42

Figure 17. Stamping line data collection scheme. ....................................................................... 45

Figure 18. Press data XML structure ........................................................................................... 46

Figure 19. OPC-UA SERVER Information model. .......................................................................... 46

Figure 20. Data sharing scheme with IDS ecosystem. ................................................................. 47

Figure 21. CECM system based on IR imaging for arc-welding monitoring ................................ 47

Figure 22. Preliminary structure of the OPC-UA information model ........................................... 48

Figure 23. FMEA for Forming Operations .................................................................................... 48

Figure 24. Machine simulators module developed for GESTAMP’s modules .............................. 49

Figure 25. Data Stream ............................................................................................................... 51


D4.1 V1.0 Page 9/ 55

List of Tables

Table 1. Example of typical utility system application for PdM (Liggan and Lyons, 2011) ......... 23

Table 2. CBM benefits and obstacles........................................................................................... 27

Table 3. CBM related standards (Shin and Jun, 2015) ................................................................. 27

Table 4. Preventive maintenance benefits and obstacles ........................................................... 29


D4.1 V1.0 Page 10/ 55

1 INTRODUCTION

This document presents the results of Task 4.1 defining and describing maintenance strategies

to improve maintainability and increase operating life of production. Taking into account

existing plant strategies and polices from the industrial pilots of Z-BRE4K end-users (i.e. SACMI,

GESTAMP, PHILIPS) as also described in the deliverables D1.1 and D1.4, D4.1 focuses on the

update of existing and development of new strategies based on real data to improve

maintainability and operating life of production systems. Thus, starting with the AS-IS

maintenance policies and particular strategies of Z-BRE4K industrial end-users, the deliverable

also reports on the TO-BE maintenance scenarios of the end-users after the implementation of

the Z-BRE4K solution. Accordingly, the Task T4.1 and the associated deliverable D4.1 propose

specific methodology to change the orientation of the plant’s maintenance plan from

reactive/preventive to predictive via adaptation of Z-Strategies at each pilot use-case.

Following the logic of the industrial maintenance policies and strategies, the deliverable is

organized as follows:

Section 2 explains the initial conception of Z-BRE4K strategies as well as their implementation,

and Section 3 defines and describes AS-IS and TO-BE embedded intelligence systems. Section 4

highlights the industrial maintenance strategies and policies implemented in manufacturing.

Accordingly, Sections 4.1-to-4.5 analyses the state-of-the-art on risk-based maintenance

(Section 4.1), predictive maintenance (Section 4.2), condition-based maintenance (Section 4.3),

preventive maintenance (Section 4.4), and corrective maintenance (Section 4.5). Subsequently,

Section 5 illustrates AS-IS maintenance strategies and policies of Z-BRE4K end-users while

Section 6 provides TO-BE maintenance scenarios of the end-users after implementation of the

Z-BRE4K solution. Next, Section 7 explains the relevance and adaptation of Z-Strategies at each

pilot use-case. Finally, Section 8 concludes the document by highlighting the main results

achieved and the connections with future activities.


D4.1 V1.0 Page 11/ 55

2 Z-BRE4K STRATEGIES OVERVIEW

In this Section of the D4.1, for the completeness of the deliverable document, we provide a

summary and overview of the Z-BRE4K strategies. More information and details concerning the

initial conception of Z-Strategies and the way to implement these strategies in Z-BRE4K are

provided in the Z-BRE4K deliverable D1.5. The innovative synergies between online data

gathering systems, real-time simulation models, data-based models and the knowledge

management system form the main strategies which contribute to achieve zero breakdowns in

manufacturing. In this context, the proposed solution comprises the introduction of eight (8)

scalable strategies at component, machine and system level, all of which can be applied in the

existing manufacturing plants with minimum interventions, targeting (1) the prediction

occurrence of failure (Z-PREDICT), (2) the early detection of current or emerging failure (Z-

DIAGNOSE), (3) the prevention of failure occurrence, building up, or even propagation in the

production system (Z-PREVENT), (4) the estimation of the RUL of assets (Z-ESTIMATE), (5) the

management of the aforementioned strategies through event modelling, KPI monitoring and

real-time decision support (Z-MANAGE), (6) the replacement, reconfiguration, re-use,

retirement, and recycling of components/assets (Z-REMEDIATE), (7) synchronizing remedy

actions, production planning and logistics (Z-SYNCHRONISE), (8) preserving the safety, health,

and comfort of the workers (Z-SAFETY). Each of the developed strategies are triggered based on

predicting, detecting and assessing the impact of system level events that cause low

performances, generate failures, and increase the costs. Figure 1 highlights the synergies and

interactions between the eight Z-Strategies for building a novel predictive maintenance platform

and the role of each strategy is further explained below.

Z-PREDICT: The events detected from the physical layer of the system are engineered into high

value data that stipulates new and more accurate process models. Such an unbiased systems

behaviour monitoring and analysis provides the basis for enriching the existing knowledge of the

system (experience) learning new patterns, raising attention towards behaviour that cause

operational and functional discrepancies (e.g. alarms for predicted failures) and the general

trends in the shop-floor. The more the data pool is being increased the more precise

(repeatability) and accurate the predictions will be. The estimations for the future states involve

the whole production line – network of machines and components. The system can thus predict

with high confidence the expected performance of components and their maintenance needs,

predicting current or emerging failures, allowing better production planning and decision

making on their RUL. Hence, the ability to optimise the manufacturing processes according to

the RUL, production needs, and the maintenance operations is the key innovation to fulfil the

industrial requirements.

Z-PREVENT: The prevention of failure occurrence strategy is based on the prediction strategy

(i.e. degraded performance of assets or failure) realised across the shop-floor for condition

monitoring of machinery and respective produced quality. The Z-PREDICT is predecessor of Z-

PREVENT. The initial estimation of the future states is based on the simulation and modelling of

the parameters. For each predicted failure or low performance (e.g. due to fatigue, wear), the

responsible factors are identified and flagged through the FMEA system. The system analyses

these factors based on an initial estimation, which after the simulation these are updated


D4.1 V1.0 Page 12/ 55

recursively. The result of this process is to avoid the building up or even propagation of a failure

that leads to breakdown based on each recorded event both from previous and current states.

The strategy thus prevents multiple alarm activations on similar failures.

Figure 1. Synergies and interactions between the eight Z-Strategies

Z-DIAGNOSE: This strategy is invoked when a current or an emerging failure is detected

considering the condition at all three levels – machine, product, shop-floor. In such a scenario,

an alarm is being triggered to flag the events that resulted in a failure or system performance

degradation. By mapping the true reasons, the system is then able to avoid generating the failure

or its emergence by weighting the system model. The strategy also involves more actions and

processes to deal both with the generation of the diagnosed failure, and its severity increase to

the next iterations as well as its impact to the production line. Depending on the criticality of

the generated failure, the system can either adapt its parameters to prolong the RUL until the

next maintenance, or plan to the production for maintenance. The final decision on the actions

is based on the Z-MANAGE strategy.

Z-ESTIMATE: This strategy combines the information from the Z-DIAGNOSE and Z-PREDICT

estimating the RUL of the assets. The estimated values are also combined with the information

from the maintenance operations (physical examination from operators) as well as from the

specifications provided from the manufacturer. The latter is used as the starting point for the

estimation process, which after each iteration the deviation of the real-model from the physical

model is reduced having an accurate virtual-model wrapped around the actual state of each

machine and its components. The trends for the fatigue and wear rates provide a confident RUL

estimation.

Z-MANAGE: This strategy is executing the overall supervision and optimisation of the system.

The failures are processed with the Decision Support System (DSS) tools and are interfaced with

Manufacturing Execution Systems (MES). False positives and false negatives are clustered within

the Z-PREDICT and Z-PREVENT Strategies. To achieve so, the previous acquired knowledge and

incidents are also processed to fine tune the system’s performance. Additionally, the production

is optimised by better scheduling (Z-SYNCHRONISE), taking into account the impact of each

failure. The optimised scheduling and adaptability of the manufacturing improves the overall

flexibility, placing a premium on the production systems, extending their operating life, while


D4.1 V1.0 Page 13/ 55

preserve increased machinery availability.

Z-REMEDIATE: This strategy involves the decision making in the event of a failure, which

classifies and categorises the input in terms of criticality, type, etc. Based on the

component/asset types (repairable-non repairable) and their RUL the strategy decides for the

following: (1) replace, (2) reconfigure and/or re-use, (3) retire, and (4) recycle. This strategy

triggers the Z-SYNCHRONISE and Z-SAFETY strategies from which the maintenance actions can

be planned and organized.

Z-SYNCHRONISE: The predecessor Z-REMEDIATE strategy identifies the type of action required

for diagnosed failures which are then fused with the Z-MANAGE output. This strategy

synchronises all the remedy actions with internal and external supply-chain tiers, as well as with

production planning and logistics. It is therefore responsible to shift the production from one

machine to another due to failure or deteriorated condition/performance, acting as the “end-

effector” thus leading to optimised scheduling and reduced costs by carrying out maintenance

activities on time.

Z-SAFETY: This strategy is invoked to increased Health & Safety during Z-BRE4K shop-floor

operations. Since most of the accidents occur during maintenance actions, the Z-SAFETY

prevents any activation to the machine that is under investigation or repair. The “Safety-Mode”

lifts any unauthorised control from the personnel for the whole duration of the maintenance.

Apart from reducing the accidents Z-SAFETY also takes into account the comfort of the human

personnel on the shop-floor, e.g. extreme heat or noise may be tolerable for the machines but

not for humans. Therefore, the health & safety procedures are also taken into account towards

the operation feedback of the whole production line.


D4.1 V1.0 Page 14/ 55

3 EMBEDDED INTELLIGENCE

When we are talking regarding the intelligent system, we are referring to a system that is able

to react appropriately in order to change various situations without the users input. However,

the main challenges for intelligent solutions in the embedded systems actually come from

dependability and real-time requirements as well as from constraints on cost, size, and power

consumption. There are two intelligent methods for embedded systems (Elmenreich, 2003),

biologically inspired and multi-agent systems presented respectively within the sub-section

3.1.1 and 3.1.2.

In generally, "intelligence" means the complete efficiency of an individual’s mental processes,

particularly their comprehension, learning/recall and reasoning capacities used to identify

"intelligent" solutions in engineering. Also, the usage of intelligent algorithms provides the

ability to solve problems that stem from changing situations (Elmenreich, 2003). Furthermore,

the embedded intelligence is considered to overcome the gap between sensor networks and

applications in smart environments (e.g. autonomous systems, assistant living systems, personal

robots) while the research on extracting embedded intelligence from the digital traces of

human-IoT interaction is still at the beginning (Guo et al., 2013).

During the recent years, the awareness was raised concerning the asset Remaining Useful Life

(RUL) optimisation and how to maintain the optimal system level performance while assets age

and at times with growing and dynamic loading demands, a transition to predictive maintenance

(section 4.2) from reactive (section 4.5) and traditional condition-based monitoring and

maintenance (section 4.3) is required in order to achieve return of investment (ROI) and

performance targets (Miguelañez-Martin and Flynn, 2015).

According to Wilfried Elmenreich (2003), at least five potential reasons exist in order to employ

an intelligent solution:

▪ Dependability: Applications for harsh environments such as process control applications

call for a solution that adapts to changing situations like performance loss or break-

down of a component. For such applications, intelligent solutions enable graceful

degradation or self-stability properties.

▪ Efficiency: An intelligent solution might be able to increase efficiency of the given

resources.Autonomy: An intelligent solution might be able to perform the same task as

a traditional system without or with reduced requirement for human supervision or

interaction.Easy Modelling: An intelligent generic self-organizing solution liberates the

system de-signer from modelling and implementation issues. This reduces the chance

of human error and reduces cost and time in the design phase.Maintenance costs: An

intelligent system might require less frequent service iterations since it is able to run for

long durations without human interaction.

▪ Insufficient alternatives: Sometimes there is no traditional approach to solve a given

problem satisfyingly, which forces the application of an intelligent solution. For example,

in data analysis, the application of neural networks solves the problem of nonlinear

correlations, which is not supported by traditional approaches.


D4.1 V1.0 Page 15/ 55

3.1 AS-IS Status in Embedded intelligence

One of the possible intelligent methods for embedded systems (Guo et al., 2013) are biologically

inspired, such are neural networks, genetic algorithms and Neuro-Fuzzy Systems respectively

presented within the section 3.1.1.

3.1.1 Biologically Inspired Embedded Systems

One of the most common examples for biologically inspired computing is the neural networks

(NNs) that consist of interconnected neurons where each is set with the input and output

connections. While the concept of a neuron cell is very simple since it contains a simple add-

and-compare mechanism that sums up the input signals and generates an output signal, the

entire NN shows emergent properties such as learning and reasoning.

A derivative-free and stochastic optimisation method that builds on ideas from the natural

selection and the evolutionary process is genetic algorithm (GA) that needs a minimum

information about the problem to be solved and therefore makes it quite easily to be applied.

The GA needs an initial population of “genes”, an algorithm that allows to cross-mix these genes,

and a fitness function that produces a comparable value on the quality of an actual solution.

After recombination and mutation of genes the GA uses the fitness function to select the best

genes for the new population by making multiple iterations, the GA approaches to the solution

that is equal or better than the value from the beginning.

Digital rules and imprecise information are bridged by Fuzzy Logic forms where the inference

method of the logic is similar to the human brain while supporting the implementation of control

algorithms for imprecise sensors that perform better than traditional control methods.

However, one of the disadvantages of the Fuzzy Logic is the lack in an effective learning

mechanism and auto-tuning. The combination of Fuzzy systems with neural networks

overcomes some problems of NNs and Fuzzy Logic, by providing an adapting system with a rule-

based model.

3.1.2 Multi-agent Systems

The idea of a multi-agent system (MAS) came as the interconnection between several widely

independent agents enabling the collaboration to function beyond the capabilities of a single

agent of the set-up. A MAS is defined by Wooldridge and Jennings (1995) as a hardware or

software-based computer system that provides the following properties where:

1. Autonomy are agents operate without the direct intervention of humans or others and have

some kind of control over their actions and internal state (Castelfranchi, 1994).

2. Social ability or agents that interact with other agents (and possibly humans) via some kind

of agent-communication language (Genesereth and Ketchpel, 1994).

3. Reactivity are agents that perceive their environment, (which may be the physical world, a

user via a graphical user interface, a collection of other agents, the internet, or perhaps all

of these combined), and respond in a timely fashion to changes that occur in it.


D4.1 V1.0 Page 16/ 55

4. Pro-activeness: Agents do not simply act in response to their environment, they are able to

exhibit goal-directed behaviour by taking the initiative.

In general, MASs may enhance speed (due to parallelism), reliability (due to redundancy),

efficiency, flexibility while the recent research has also shown the applicability to the embedded

systems domain (Guo et al., 2013).

3.2 TO-BE Status in Embedded intelligence

The predictive maintenance regime requires the access to the condition of the assets, data and

the knowledge that can be extracted from these data. Embedded decision-making agents that

contain reasoning algorithms in order to optimise the long-term management of heterogeneous

assets, provide fast dynamic response to events by autonomously coupling resource capabilities

with alarms in real time. The main objective of a predictive maintenance process is to advance

equipment reliability by identifying possible problems before they actually cause failures,

further damage as well as to increase the cost of the asset. Secondly, its objective is to provide

advance warning of problems that are developing before this equipment fail catastrophically

during a production run. More information of proactive maintenance action and its benefits on

reducing the possibility of a fault can be find within the section 4.2.

The input for the prediction and diagnostic task to produce optimum fault detection are the

results from the embedded tools and annotated sensor data. The output from diagnostic and

prediction is the input to the planning task involving sub-tasks such fault recovery and on-line

learning, if it is adequate. The different tasks and their decomposition into subtasks can be used

as the basis for constructing the model. If a list of knowledge roles, which serve as input/output

in these tasks, is formulated, the most important ones, which can be taken by different

knowledge types (Miguelañez-Martin and Flynn, 2015) are:

▪ Parameter: a measured or calculated quantity whose value can detect abnormal

behaviour;

▪ Source: Something that can be observed or detected;

▪ Symptom: A negative source;

▪ Norm: Expected values of a parameter for normal condition;

▪ Discrepancy. A quantified difference to the norm;

▪ Fault: Cause of symptom;

▪ Location: Where a symptom or fault is found;

▪ Action: An activity to eliminate a fault or to improve situation.

These knowledge roles could represent the meta-concepts in the knowledge-based system

while expressing the relation task-domain. Several domain knowledge models can be

constructed for the maintenance scenarios that are defined as domain models representing the

knowledge of the domain independently of their use. However, the application of the predictive

maintenance knowledge-based model will employ existing domain models using concepts and

relations from these models to optimise the knowledge transfer (Miguelañez-Martin and Flynn,

2015).


D4.1 V1.0 Page 17/ 55

Today, an engine’s maintenance is no longer just a traditional event of a but it is a matter of how

to detect the first sign from the engine and to know it before there is a need for preventing the

problem. Engineers can properly analyse the equipment failures and forecast the probability of

the same equipment failing in the same asset or other units, or undertake the processes, such

as data collection, data clustering, testing, fault or defect diagnosis, planning spare parts, making

recommendations, reporting major factors affecting a system’ s life, all in a technical and timely

manner. All layers are very important and can be used when a system observer participant in a

particular communication. Whether a maintenance engineer can exploit in elliptical or

anaphoric resolution is depending in part on the role that the engineer has most recently played

in the communication in the physics-based infrastructure (Miguelañez-Martin and Flynn, 2015).


D4.1 V1.0 Page 18/ 55

4 INDUSTRIAL MAINTENANCE STRATEGIES

Nowadays, industries are quite pressured to continuously improve asset performance and their

reliability while at the same time putting the effort to minimise the costs and to ensure safety

(Pathak, 2018). The new technologies, Industrial Internet of Things (IIoT) and Industry 4.0 are

the once that offers to users the possibility to prepare the strategical plan while forecasting and

optimising the maintenance that is beyond the traditional reactive maintenance. This future of

maintenance, operations and asset management, namely Asset Performance Management

(APM) 4.0 is presented at Figure 2, providing the insight within the maturity pyramid, represents

the journey toward more proactive and optimised maintenance execution.

Figure 2. The maintenance maturity pyramid (Pathak, 2018)

The most basic approach is the reactive maintenance also called as corrective maintenance

presented within the sub-section 4.5. It is suitable for the non-critical assets that have non or

very little immediate impact on safety or plant availability while having the minimal repair or

replacement costs. The second level is the preventive maintenance (PM) described in sub-

section 4.4. This strategy follows the maintenance to be followed on a fixed time schedule or

based on operational statistics and manufacturer/industry recommendations of good practice.

While the sub-section 4.3 focuses on the physical condition of equipment, how the Condition

Based Maintenance is operating and when the measurable parameters are good indicators of

impending problems, the Predictive Maintenance (PdM) strategy presented within the sub-

section 4.2 is used for more complex and critical assets offering analytics solutions to learn an

asset’s unique operating profile during all loading, ambient and operational process conditions.

Finally, the implementation of risk-based maintenance presented within the Figure 1, involves

the comprehensive maintenance strategy described in detail within the sub-section 4.1, that

leverages existing data, advanced analytics and simulations while forecasting in order to

understand the true issues driving asset performance and reliability.


D4.1 V1.0 Page 19/ 55

4.1 Risk-based maintenance

Shinsuke Sakai (2010) reported that the Risk Based Maintenance (RBM) was introduced first in

the chemical engineering and petroleum refining fields, as well as that it is expanding to a broad

range of industrial fields e.g. shipbuilding, gas, electric power, steelmaking and rocket ground

facilities. In general, the RBM (SKF, 2018) is a quantitative and financially-based analysis

technique that can increase the profitability of the operation while optimizing the total life cycle

cost without compromising safety or environmental issues within various industries (Khan and

Haddara, 2003). It defines the opportunities for incremental improvement by removing the low-

value tasks while presenting the tasks that address high commercial risk areas further analysing

the costs and benefits of steps to mitigate failures. This suitable strategy provides a systematic

approach to determine the most appropriate asset maintenance plan (Figure 3) and while

implementing this maintenance plan, the risk of asset failure will be low.

Figure 3. The maintenance plan organized by task prioritization

The RBM approach assists in designing an alternative strategy to minimise the risk resulting from

failures and breakdowns while its adaptation is crucial for the development of cost-effective

maintenance policies (Krishnasamy et al., 2005). The risk information and its general

consequences as well as the general methods used to mitigate and predict the risk, needs to be

collected, evaluated in the context of the facility under consideration and ranked either as

acceptable or unacceptable risks in order to determine the plan to inspect the system using a

condition monitoring approach. At this stage the proposal for mitigating the risk is prepared

followed by the evaluation and the reassessment against various factors e.g. legal and regulatory

requirements. The risk-based maintenance framework is applied to each system in a facility and

it is presented at Figure 4.


D4.1 V1.0 Page 20/ 55

Figure 4. Risk based Maintenance Framework (Fiix, 2018)

When we are talking regarding the risk-based maintenance solutions (Pathak, 2018), two key

benefits need to be mentioned. Essentially, the first one permits the prioritisation of asset

management by focusing on the assets that need attention within the company. It is important

to ensure, the most important assets, receive priority and more thorough analysis in order to

reach the optimal maintenance. There are several different techniques of risk-based

maintenance known so far and some of them are provide below.

Reliability Centred Maintenance (RCM) is a process that ensures the systems to continue doing

the users requests in their present operating context (Moubray, 1997). This technique is

generally used to achieve enhancements in fields such as the establishment of safe minimum

levels of maintenance. The successful implementation of RCM will lead to increase of cost

effectiveness, reliability, machine uptime, and the level of risk that the organization is managing

is going to be understand better.

Failure Mode and Effects Analysis (FMEA) (i.e. failure modes) is the first step of a system

reliability study that involves reviewing as many components, assemblies as well as subsystems

to identify the failure modes and their causes and effects. It is a qualitative analysis (Rausand

and Høyland, 2004), but may be also considered as a quantitative basis when mathematical

failure rate models (Tay and Lim, 2008) are combined with a statistical failure mode ratio

database.

Fault Tree Analysis (FTA) is an analysis method mainly used in the fields of safety and reliability

engineering trying to identify the best ways in reducing the risk or determining the event rates

of a safety accident or a particular system level (functional) failure. Specifically, it is used in the

aerospace (Goldberg et al., 1994), nuclear power, chemical and process (CCPS, 2008; CCPS,

1999; OSHA, 1994), pharmaceutical (ICH, 2005), petrochemical and other high-hazard industries

but is also used in fields as diverse as risk factor identification relating to social service system

failure (Lacey, 2011).

Failure Mode, Effects and Criticality Analysis (FMECA) is an extension of failure mode (FM) and

effects analysis (FMEA) to include a means of ranking the severity of the FMs to allow

prioritization of countermeasures (IEC, 2006). This is done by combining the severity measure


D4.1 V1.0 Page 21/ 55

and frequency of occurrence to produce a metric called criticality (i.e. considering criticality

combined with severity as a measure of risk).

Root Cause Analysis (RCA) is actually the problem-solving methodology that allows users to

quickly diagnose the cause when asset failures occur and take action to eliminate reoccurring

incidents (Wilson, 1993). The second key advantage of the risk-based maintenance is actually

the management strategy (Nag et al., 2007) that provides to users the detailed analysis and

simulations they can use to visualise the effects of deploying different asset management

strategies and follow the impact of differing asset management approaches resulting in an

aligned strategic approach to operations and asset management. The goal is to achieve the

short-term efficiencies as well as long-term sustainability (Pathak, 2018).

In general, there are many different methodologies and various approaches that have been

established to undertake a risk analysis within industry facility. According to J. Tixier et al., (2002)

more than sixty methodologies have been identified and divided into identification, evaluation

and hierarchisation phases which leads us to the conclusion that there is no one standard

method for assessing the risks. Furthermore, three different approaches can be used to

determine the possible risks that exist, the qualitative, the semi-quantitative and the

quantitative approach while including the deterministic and probabilistic approaches that can

estimate the probability of these risks. Also, risk matrix1, evaluates the impact of the

maintenance task is used to obtain the application of risk management principles to

maintenance tasks by comparing the assets probability of failure (PoF) and the assets

consequences of failure (CoF). On the other hand, the maintenance is essentially treated as a

risk control process where the owners and managers decide whether to spend more time on

managing each side of the prevention (i.e. the probability of failure through preservation) and

prevention or recovery (i.e. the consequences of failure through recovery, repairs and renewals).

Note that the main reason we conduct maintenance is that we need to understand the

application of risk principles to maintenance and using practices and systems necessary to

support decision making. Risk-based decision-making is at the heart of asset management and

this requires mindful consideration of the relationship between the probability of failure (PoF)

and the consequences of failure (CoF). The complexities of these correlations can be captured

on a risk matrix (Figure 5) where the risk events are arranged into four categories.

Low-Impact-High-Probability (LI-HP)

High-Impact-High-Probability (HI-HP)

P R

O B

A B

I L

IT Y

of

failu

re (

Co

F)

Low-Impact-Low-Probability (LI-LP)

High-Impact-Low-Probability (HI-LP)

C O N S E Q U E N C E S of failure (PoF)

Figure 5. Risk Matrix and risk-based decision making

1 http://www.assetinsights.net/Glossary/G_Risk-Based_Maintenance_%28RBM%29.html

http://www.assetinsights.net/Glossary/G_Risk-Based_Maintenance_%28RBM%29.html


D4.1 V1.0 Page 22/ 55

4.2 Predictive Maintenance

Predictive maintenance (PdM) is considered to be one of the main forces for improving

productivity as well as being the way how to achieve "just-in-time" and no delays in

manufacturing (Amruthnath and Gupta, 2018a). Predictive maintenance approach allows the

convenient scheduling of corrective maintenance preventing the unexpected equipment

failures. This technique is actually designed to support the condition of in-service equipment in

order to predict when maintenance must be performed.

Figure 6. Benefits of Predictive Maintenance2

It is important to know which equipment requests maintenance in order to plan better the

maintenance work (e.g. spare parts, people, etc.) and to avoid the breakdowns. The idea is to

minimise these errors; to reduce the time they consume thus increasing plant availability.

Furthermore, increased equipment lifetime, increased plant safety, fewer accidents with

negative impact on environment, optimised spare parts handling, etc., are also some of the

potential advantages that are offered through the PdM (Figure 6). Note that predictive

maintenance relies on the actual condition of equipment, rather than average or expected life

statistics, to predict when maintenance will be necessary. Furthermore, data collection and pre-

processing, fault detection, maintenance scheduling and resource optimisation, early detection

fault, failure time prediction are some of the main components necessary for implementing

predictive maintenance (Amruthnath and Gupta, 2018b).

2 https://www.elp.com/articles/print/volume-93/issue-3/sections/generation/optimizing-electric-utility-o-m-with-predictive-analytics.html

https://www.elp.com/articles/print/volume-93/issue-3/sections/generation/optimizing-electric-utility-o-m-with-predictive-analytics.html

https://www.elp.com/articles/print/volume-93/issue-3/sections/generation/optimizing-electric-utility-o-m-with-predictive-analytics.html


D4.1 V1.0 Page 23/ 55

There are various effective predicting failure techniques that provide sufficient warning time for

upcoming maintenance. These approaches are part of condition-based monitoring

considerations that are best employed in consultation with equipment manufacturers and

condition monitoring experts3. Choosing the correct condition-monitoring technique depends

on the need of a company and the type of assets an organisation employs while the chosen tool

should be highly effective providing the sufficient warning time for upcoming maintenance.

While estimation the equipment condition, predictive maintenance employs the testing

technologies (e.g. infrared, acoustic, corona detection, vibration analysis, sound level

measurements, oil analysis etc.) and other specific online tests. There are many predictive

maintenance tools that can be employed that require machines to be running at normal capacity

and do not interfere with the production schedules. The most common condition-monitoring

tools used in predictive maintenance4,5,6, act on the analytics collected by the devices and

sensors where various Condition Monitoring tools (e.g. eMaint CMMS, Maintenance

Connection, IBM Maximo, Fluke Condition Monitoring, etc.) can help companies to develop

accurate predictions when a piece of equipment will require maintenance or replacement

(Robin, 2006; Kennedy, 2006; Yung, 2006). A variety of technologies (Table 1) are used to help

diagnose the condition of assets using non-destructive techniques such as:

▪ Vibration analyses - mainly used in performance for equipment such pumps and motors

to detect misalignment, imbalance, mechanical looseness or wear on pumps or motors.

▪ Infrared thermography - identifies unusually high temperature conditions in

transmissions, gearboxes, bearings and many more with infrared cameras.

▪ Oil analysis - measures an asset’s number and size of particles, a lubricant’s health and

if it has been contaminated.

▪ Ultrasound analyses - are used to detect mechanical malfunctions of movable parts and

faults in electrical equipment, e.g. leak in pipe systems, tanks.

▪ Current analysis – measure the current and voltage of electricity supplied to an electric

motor.

▪ Acoustic analysis – are used to detect liquid, gas or vacuum leaks.

▪ Electrical analysis – measure the motor current readings using clamp on ammeters.

▪ Operational performance – are using the sensors throughout a system in order to

measure pressure, temperature, flow etc.

▪ Other condition-monitoring techniques - shock pulse, fluid analysis, performance

trending, stereoscopic photography and material (non-destructive) testing, e.g.

ultrasonic, eddy current, borescope inspections.

Table 1. Example of typical utility system application for PdM (Liggan and Lyons, 2011)

PdM Technique Applications

3 https://www.fiixsoftware.com/condition-based-maintenance/ 4 https://www.emaint.com/what-is-predictive-maintenance/ 5 https://www.lce.com/Predictive-Maintenance-Strategy-84.html 6 https://whatis.techtarget.com/definition/predictive-maintenance-PdM

https://www.fiixsoftware.com/condition-based-maintenance/

https://www.emaint.com/what-is-predictive-maintenance/

https://www.lce.com/Predictive-Maintenance-Strategy-84.html

https://whatis.techtarget.com/definition/predictive-maintenance-PdM


D4.1 V1.0 Page 24/ 55

Vibration

analyses

Rotating

Equipment/Drive

System

Structural

Vibration

Motors Fan Balancing

Oil analysis Component Wear

and Tear

Oil Degradation Water Ingress in

Oil

Equipment

Overheating

Ultrasound

analyses

Stream Trap

Testing

Leak Detection Electrical Arcing Valve Integrity

Going through the industrial revolution (Figure 7), Industry 4.0 or the Information and

communication technology (ICT) industry, is the latest industrial revolution and it is affecting all

areas of life (Wang, 2016). It uses the artificial intelligence (AI) which requires a minor human

involvement, transitioning from an input and output approach to a smooth conversation

between humans and robots. Thus, the actual machines are able to:

▪ make decisions

▪ provide technical assistance

▪ calculate and to determine risk factors and

▪ improve the work environment that brings the incensement in return due to maximised

efficiency

This process, Predictive Maintenance in Industry 4.0 (PdM 4.0), is fundamentally changing the

manufacturing industry.

Figure 7. Predictive Maintenance within Industrial revolution7

Furthermore, the PdM 4.0 refers not only to the representation of the fourth level of maturity

in predictive maintenance but also on its application of big data analytics (Figure 8). While, the

visual inspections refer to periodic physical inspections where the conclusions are based solely

on inspector’s expertise, the instrument inspections (or periodic inspections) are decisions

7 https://www.engineering.com/AdvancedManufacturing/ArticleID/15798/How-Predictive-Maintenance-Fits-into-Industry-40.aspx

https://www.engineering.com/AdvancedManufacturing/ArticleID/15798/How-Predictive-Maintenance-Fits-into-Industry-40.aspx

https://www.engineering.com/AdvancedManufacturing/ArticleID/15798/How-Predictive-Maintenance-Fits-into-Industry-40.aspx


D4.1 V1.0 Page 25/ 55

based on a combination of inspector’s expertise and instrument read-outs. Furthermore, the

real-time condition monitoring is continuous real-time monitoring of assets, with alerts given

based on pre-established rules or critical levels. Finally, the Level 4 (i.e. PdM 4.0) is continuous

real-time monitoring of assets, with alerts sent based on predictive techniques, such as

regression analysis that has become possible using smart, connected technologies that unite

digital and physical assets. This concept even not new, is the massive investments in technology

since it is necessary to overcome the massive volumes of data required often limited

deployment to only the largest organisations (Coleman et al., 2017). The results reported within

the Predictive Maintenance 4.08 specify that a new level of the predictive maintenance is

reached by only a few companies and that the predictive maintenance strategies are facing

significant challenges that are dealing with the evolution of the equipment, instrumentation and

manufacturing processes they are actually supporting.

Figure 8. PdM Maturity Matrix9

4.3 Condition-based maintenance

The detection of deterioration processes in its early stage, through the retrieval and

interpretation of equipment measurements, may provide a significant reduction in maintenance

costs and minimise the risk of the occurrences of undesired failures. Condition Based Monitoring

(CBM) is supported by mature technologies holding a dominant position in the mix of

maintenance strategy of every company seeking fir excellence in maintenance10.

One type of the Predictive Maintenance is actually the CBM that involves sensors which measure

the status of an asset over time while it is in operation11. The collected (sensor) data are used to

predict failures as well as to establish trends and to calculate the remaining life of asset. Note

must be made that the CBM maintenance is only performed when the data (indicators) shows

8 https://www.pwc.nl/nl/assets/documents/pwc-predictive-maintenance-4-0.pdf 9 https://www.pwc.nl/nl/assets/documents/pwc-predictive-maintenance-4-0.pdf 10 https://abe.gr/en/condition-based-maintenance/ 11 https://inspectioneering.com/tag/condition+based+monitoring

https://www.pwc.nl/nl/assets/documents/pwc-predictive-maintenance-4-0.pdf

https://www.pwc.nl/nl/assets/documents/pwc-predictive-maintenance-4-0.pdf

https://abe.gr/en/condition-based-maintenance/

https://inspectioneering.com/tag/condition+based+monitoring


D4.1 V1.0 Page 26/ 55

that performance is decreasing, failure is very likely to occur. The indicators (non-invasive

measurements, visual inspection, performance data, scheduled tests, etc.) can be gathered at

certain or continuous intervals and apply condition-based maintenance to mission critical or

non-mission critical assets12. Basically, the CBM is a maintenance strategy that monitors and

track the actual asset condition and decides what maintenance needs to be followed. Its main

goal is to spot upcoming equipment failure and to schedule the maintenance when it is needed

and necessary.

Figure 9. CBM optimises costs between preventive and corrective maintenance (Toms, 1995)

In general, there are two common maintenance thinking that are being usually employed, the

preventive and the corrective maintenances that are presented in detail within sections 4.3 and

4.4 respectively. However, corrective i.e. reactive maintenance can have severe performance

costs while the preventive i.e. scheduled maintenance replaces parts before the end of their

useful life13. The CBM philosophy, presented at Figure 9, optimises the transaction between

maintenance costs and performance costs by increasing both availability and reliability while

eliminating unnecessary maintenance activities allowing the preventive and corrective actions

12 https://www.fiixsoftware.com/condition-based-maintenance/ 13 https://www.swri.org/sites/default/files/brochures/condition-based-maintenance.pdf


https://www.swri.org/sites/default/files/brochures/condition-based-maintenance.pdf


D4.1 V1.0 Page 27/ 55

to be scheduled at the optimal time14. Despite being useful (e.g. improving the system reliability,

minimising the maintenance costs, etc.) on the other hand there are several challenges (e.g. the

initial cost of CBM can be high) of CBM exploitation. Furthermore, introducing CBM will raise a

major question in how the maintenance is performed and potentially to the entire maintenance

company organisation. Also, the technical side of the CBM is not always simple and the Table 2

summarizes both advantage and disadvantages of the CBM philosophy.

Table 2. CBM benefits and obstacles

While the asset is working, the CBM is performed

which leads to minimisation of disruptions to

normal operations.

The test equipment for condition monitoring is

expensive to install and databases are cost

consuming while analysing.

Reductions of the asset failures costs. Cost consuming to train the stuff since it is

necessary to engage a professional with know-

how to analyse the data and perform the work.

Improvement of equipment reliability. Difficulties in detection the fatigue or uniform

wear failures with CBM measurements.

Minimisation of unscheduled downtime due to

catastrophic failure.

Condition sensors may not survive in the operating

environment.

Minimisation of time spent on maintenance.

May require asset modifications to retrofit the

system with sensors

Minimisation of overtime costs by scheduling the

activities.

Unpredictable maintenance periods

Minimizes requirement for emergency spare parts

More optimal optimisation than manufacturer

recommendations (i.e. optimisation of

maintenance intervals).

Improvement of workers safety.

Reduction of the possibility for collateral damage

to the system.

Furthermore, there are several international standards related to CBM approach and their

details are summarised in Table 3. Some of them are the condition monitoring and diagnostics

standards for machinery industry (e.g. ISO 13372, ISO 13373, etc.) while there are standards as

well related to the issues of integration and data sharing among manufacturing facilities for CBM

(e.g. ISO 18435) (MIMOSA OSA-EAI).

Table 3. CBM related standards (Shin and Jun, 2015)

Standards Subject

IEEE 1451 Smart transducer interface for sensors and actuators

IEEE 1232 Artificial Intelligence Exchange and Service Tie to All Test Environment

ISO 13372 Condition monitoring and diagnostics of machines—Vocabulary

14 https://www.fiixsoftware.com/condition-based-maintenance/



D4.1 V1.0 Page 28/ 55

ISO 13373-1 Condition monitoring and diagnostics of machines – Vibration condition

monitoring—Part 1. General procedures

ISO 13373-2 Condition monitoring and diagnostics of machines—Vibration condition

monitoring – Part 2. Processing, analysis and presentation of vibration data

ISO 13374 MIMOSA OSA-CBM formats and methods for communicating, presenting and

displaying relevant information and data

ISO 13380 Condition monitoring and diagnostics of machines—General guidelines on using

performance parameters

ISO 13381-1 Condition monitoring and diagnostics of machines—Prognostics, general

guidelines

ISO 14224 Petroleum, petrochemical and natural gas industries-collection and exchange of

reliability and maintenance data for equipment

ISO 17359 Condition monitoring and diagnostics of machines—General guidelines

ISO 18435 MIMOSA OSA-EAI diagnostic and maintenance applications integration

ISO 55000 Asset management

It must be noted that the CBM can bring value to your organisation in many ways improving the

equipment reliability, decreases maintenance costs, eliminating the unplanned downtime

resulting from equipment failure and prevent major failures that lead to health, safety, and

environmental risks15.

4.4 Preventive Maintenance

Preventive maintenance (Figure 10) is service, that is being processed by responsible staffs for

the purpose of maintaining equipment in adequate operating condition, providing correction of

initial failure before it occurs or before it develops into major defects16.

Figure 10. Preventive Maintenance Philosophy17

15 http://www.ashcomtech.com/cbm-strategy 16 https://reliabilityweb.com/articles/entry/the-importance-of-preventive-maintenance 17 http://sundaybizsys.com/preventive-maintenance/

http://www.ashcomtech.com/cbm-strategy

https://reliabilityweb.com/articles/entry/the-importance-of-preventive-maintenance

http://sundaybizsys.com/preventive-maintenance/


D4.1 V1.0 Page 29/ 55

This work is a regular and continuous action engaged on equipment in order to prevent and to

avoid its breakdowns. Furthermore, maintenance (e.g. tests, measurements, adjustments, parts

replacement and cleaning) is performed specifically to prevent faults from happening, following

the philosophy of evading or mitigating the consequences of equipment failure by replacing

worn components before they actually fail. Maintenance activities are designed to preserve and

restore equipment reliability and they include partial or complete repairs at defined periods, oil

changes, lubrication, some adjustments, etc., while the workers can report the equipment fall

in order to replace/repair damaged parts before they cause the complete system failure.

Unfortunately, the implementation of a preventive maintenance program can be both time and

cost consuming which creates constant discussions if it is worth installing18. Ideally, a preventive

maintenance (PM) will prevent all equipment failure before it occurs while the same time it

saves time, reduces costs and runs the operation efficiently and productively. Preventive

maintenance offers a number of key benefits but also some disadvantages as well (Table 4). It

manages maintenance tasks in order for maintenance operations to be ran smoothly.

Furthermore, it saves on maintenance costs, prioritising maintenance tasks based on operations,

and minimising the disruption to the work schedule when maintenance is performed19. Finally,

this kind of maintenance:

▪ manages all maintenance tasks,

▪ saves on maintenance costs,

▪ prolongs life of company equipment,

▪ less unplanned downtime caused by equipment failure,

▪ less unnecessary maintenance and inspections,

▪ fewer errors in day-to-day operations,

▪ improves reliability of equipment,

▪ fewer expensive repairs caused by unexpected equipment failure that must be fixed

quickly,

▪ reduces risks of injury.

Table 4. Preventive maintenance benefits and obstacles

Very simple maintenance strategy to implement. Need for investment in time and resources.

No additional cost for condition monitoring

technology.

Slight inspection into the actual condition of

equipment.

Improvements in compliance and safety. Requires the training of employees.

It may be said that preventive maintenance (PM) is a must since it is a repetitive maintenance

that is performed in order to ensure asset reliability and to abolish any possible equipment

failures/downtime that could be occurring20. This maintenance can be observed as a proactive

18 http://ableserve.com/issue-1/the-benefits-of-preventive-maintenance/ 19 https://www.micromain.com/what-is-preventive-maintenance/ 20 https://reliabilityweb.com/articles/entry/the-importance-of-preventive-maintenance

http://ableserve.com/issue-1/the-benefits-of-preventive-maintenance/

https://www.micromain.com/what-is-preventive-maintenance/

https://reliabilityweb.com/articles/entry/the-importance-of-preventive-maintenance


D4.1 V1.0 Page 30/ 55

approach that establishes a scheduled inspection of assets in order to verify the dependency

and to prolong the asset durability. The problems (e.g. maintenance costs, errors in day-to-day

operations, equipment that is not reliable, etc.) can be avoided with a usage of a computerised

maintenance management software (CMMS) system that actually offers preventative

maintenance as one of its main functions. The CMMS offers the “top” overview to companies of

the entire facility and various locations within, in order to ensure the effective preventative

maintenance, schedule a part of all standard operating procedures. Therefore, preventative

maintenance software provides tools such are automatic triggers, email integration, reminders,

equipment information and auto-assigned task which can lead the maintenance process21.

Several CMMS software exist today on the market, e.g. GP MaTe, UpKeep, EZOfficeInventory,

ManWinWin, Fixd, SIVECO, I2S, ORBIS, SAP, etc., that provide help in order to schedule, plan,

manage, and track maintenance activities, offering non-stop support for an organization’s

preventive maintenance (PM) program. The general concept of creating the effective preventive

maintenance plan is described within the simple example of AIMMS software and respective

steps are presented through the Figure 11 - Figure 13.

Within a New Task tab, the task is programmed (Figure 11) where the asset, the category of the

task, the task operator, the task’s criticality and other information vital to effectively performing

the work, are defined by the User which controls a PM calendar and overviews the task

catalogue. Based on the created tasks over a certain period, the PM can be easily designed and

all personnel and failures during this period are known.

Figure 11. Maintenance task creation

In order to develop an effective preventive maintenance program (Figure 12), scheduling plays

the important role. The preventive maintenance tasks provide help on a shop floor automatically

generating the PM tasks based on a daily, weekly or monthly basis. PM tasks contain description,

assets, which prototype should be used, as well as the category of each task and their

instructions. The estimated duration of the task is the value that allows users to schedule all

21 https://www.hippocmms.com/blog/preventive-maintenance-program-in-six-steps

http://gpmate.com/

https://onupkeep.com/

https://www.ezofficeinventory.com/

https://manwinwin.com/en/homepage/

https://fixd.io/en

http://www.siveco.ro/en

http://www.i2s.gr/en/

https://gbr.orbis.org/en

https://www.sap.com/products/supply-chain-iot/asset-management-eam.html

https://abe.gr/en/aimms-functionality/

https://www.hippocmms.com/blog/preventive-maintenance-program-in-six-steps


D4.1 V1.0 Page 31/ 55

preventive maintenance tasks, slot them in the calendar and allocate suitable personnel in the

schedule. Besides, the PM tasks contain the elements of a simple task and the conditions on

which the prevention is based, it also covers the condition-based maintenance tasks, calendar

or counter based tasks and a combination of the above all.

Figure 12. Preventive Maintenance Task

Based on the tasks created, a calendar (Figure 13) is created in which the daily tasks are shown

as well as who are the responsible workers for the tasks, the hours need for the completion of

the task etc. Depending on this AIMMS view, the user can assign PMs to those who work at the

moment and schedules for the next maintenance tasks.

Figure 13. Scheduling view and calendar

4.5 Corrective Maintenance

Corrective maintenance is a maintenance task or operation that is performed in order to

identify, distinct and correct a particular fault. This maintenance is performed in order to restore

the failed machine, equipment or system to an operational condition within the tolerances or


D4.1 V1.0 Page 32/ 55

limits established for in-service operations22. Corrective maintenance can be either planned or

unplanned and it can be subdivided as:

▪ Immediate corrective maintenance – where the work starts as soon as a failure occurs

▪ Deferred corrective maintenance - in which work is delayed in conformance to a given

set of maintenance rules

The technical standards concerning corrective maintenance are set by the International

Electrotechnical Commission International Standards for all electrical, electronic and related

technologies within the IEC 60050 chapter 191 the “Dependability and quality of service” as well

as the IEC 60050-191 (IEC, 1990). According to Pintelon et al. (2006), with the usage of the

correct maintenance strategy, the downtime and the maintenance cost can be radically

decreased since sometimes it can be impossible to predict or prevent a failure. Corrective

maintenance, also called break down maintenance, can be in some cases the only option to be

applied, however it cannot be scheduled and as such it makes it harder to plan it and it costs

more to perform. On the other hand, the costs associated with corrective maintenance include

repair costs, lost production and lost sales. This retroactive maintenance and strategy involves

the following steps to be taken:

1. failure following,

2. failure diagnosis in order to eliminate such a part,

3. failure cause,

4. replacement order,

5. part replacement,

6. test of function and

7. corrective maintenance continuation usage.

The process of corrective maintenance begins with a diagnosis of the failure to determine why

it has occurred which can include a physical inspection of a system, use of a diagnostic computer

to evaluate the system, interviews with users and other numerous steps. It is important to

determine what caused the problem in order to take appropriate action and to be aware of

multiple component or system failures that may have occurred simultaneously. This step by step

elementary procedure is followed while the failure is the one that activates the steps. Within

the Industry 4.0, the modern technologies reduce the inherent drawbacks of corrective

maintenance23 by providing device history, fault patterns, repair advice or availability of spare

parts. At the example of SAP, the corrective maintenance processing presented within the Figure

22 “Department of Defense Standard Practice; Reliability-Centered Maintenance (RCM) Process.” MIL-STD-3034. Department of Defense. Jan 21 2011. http://www.everyspec.com/MIL-STD/MIL-STD-3000-9999/MIL-STD-3034_30534/

23 http://www.controlengeurope.com/article/159477/Mobile-maintenance--proof-of-Industry-4-0-payback.aspx

https://www.sap.com/index.html

http://www.everyspec.com/MIL-STD/MIL-STD-3000-9999/MIL-STD-3034_30534/

http://www.everyspec.com/MIL-STD/MIL-STD-3000-9999/MIL-STD-3034_30534/

http://www.controlengeurope.com/article/159477/Mobile-maintenance--proof-of-Industry-4-0-payback.aspx


D4.1 V1.0 Page 33/ 55

14, is involved in preventive and regular maintenance processes. Also, the plant maintenance

involves the definition of following steps24:

▪ The plant maintenance operator enters a notification in SAP System requesting the

maintenance in order to repair defective equipment.

▪ The maintenance designer creates, plans and schedules a maintenance work order

within the system.

▪ The work order is received by the technician.

▪ An authorised person in the preventive maintenance (PM) system approves and

completes the work as per the work order.

Figure 14. Corrective Maintenance Processing25

24 https://www.tutorialspoint.com/sap_pm/sap_pm_corrective_maintenance.htm 25 https://www.slideshare.net/AlhadiAkbarNibel/sap-plant-maintenance-overview-alhadi-a-nibel-52002333

https://www.tutorialspoint.com/sap_pm/sap_pm_corrective_maintenance.htm

https://www.slideshare.net/AlhadiAkbarNibel/sap-plant-maintenance-overview-alhadi-a-nibel-52002333


D4.1 V1.0 Page 34/ 55

5 AS-IS MAINTENANCE STRATEGIES AND POLICIES OF Z-BRE4K END-USERS

Each end user of Z-BRE4K conducted a research and examined the current maintenance

practices employed to achieve the company’s goals and various strategies that are being

incorporated into these practices. The respective sub-sections 5.1-to-5.3 including the AS-IS

scenario per each end user present these details.

5.1 SACMI

The packaging sector for food and beverages features extremely high productivity requirements,

thus the manufacturing systems and machinery utilized for the production of recipients and

closures has to be not only very fast, but also highly reliable. In this regard, maintenance actions

play a key role in guaranteeing the production requirements (e.g. up to 50,000 closures/hour),

where the economic loss of an unexpected failure would result in a 24/36-hour intervention,

which may translate into a loss of circa one hundred thousand Euro.

In this context, the correct execution and timing of maintenance activities is of great importance

also in assuring that equipment is switched off as little as possible (minimize downtimes due to

maintenance interventions) and components are substituted when their residual life is almost

expired (minimize costs due to wasting of operative components).

5.1.1 Production System

In the case of SACMI’s use case, the production system is represented by the Continuous

Compression Moulding (CCM) machine, which performs a hydraulic rotary press carrousel in

order to manufacture plastic closures staring from a hot polymer pellet which is compressed

and cooled down in a whole carousel ride.

As presented in D1.4 of Z-BRE4K project, the main components of SACMI’s production system

are:

▪ Plastic extruder;

▪ Plastic dose (pellet) cutting carrousel (revolver);

▪ Compression moulding carrousel;

▪ Product extraction and evacuation system;

▪ Hydraulic system;

▪ Cooling system;

▪ Pneumatic system;

▪ Electric System.

These subsystems are subject to very diverse mechanical failures, which contemporaneously

present different levels of probability to happen and severity in case of occurrence. To this

extent, the knowledge generated from the design, manufacturing and operation activities of

CCM machines has translated into several useful instruments for SACMI’s current maintenance

strategy:

▪ FMECA analysis;


D4.1 V1.0 Page 35/ 55

▪ Data-logging of KPIs and dedicated sensors;

▪ Maintenance guide and e-learning modules.

As presented in D1.4 of Z-BRE4K, SACMI’s current strategy to avoid unexpected breakdowns is

Preventive Maintenance, which directly translates into programmed breakdowns to apply

routine activities, which start from checking certain KPI values to the substitution of pieces (i.e.

bearings) and/or consumables (i.e. oil).

The Failure Modes of CCM’s subsystems are assessed by means of the FMECA technique and an

exhaustive, updated version of the failure modes of the three subsystems analysed within Z-

BRE4K (Plastic Extruder, Hydraulic System and Cooling System) has been already reported in

T1.4.

Accordingly, data is already logged in order to support the assessment of CCM’s performance

and State of Health. The data available within the production system comes from either process

(level 0) or sensors (level 1) or industrial automation platform (level 2). To this regard, SACMI

has already identified a number of sensors that can complete their current sensor array solution

in order to improve the condition monitoring of the CCM machine towards a real Predictive

Maintenance strategy.

5.1.2 Plant Strategy and Policies

Apart from the best-selling CCM machine, SACMI develops integral solutions to manage

complete automation on CCM-based lines and it is aiming to integrate machine automation

together with execution systems resource planning infrastructures, in a digital, connected, smart

factory. Thus, SACMI is putting much effort on digitisation processes and on integrated

supervision and control systems of machine and production lines.

SACMI’s experience has made possible the creation of Human Expertise for Reactive Engineering

(H.E.R.E), a system of solutions to the smart factory service and its production processes.

It defines not only the simultaneity of the presence, but also the experience that SACMI owns in

designing reactive systems, which respond to requests in real time, with actions and data. The

system makes sure that everything is just a click away, or "tap", even from mobile devices.

SACMI’s HERE architecture allows to complete the flow from machines in plant to ERP: the data

collected from machines will be valorised and transmitted to DCS, supervisors dedicated to

different production phases. Finally, this information will be stored at the customer’s ERP, who

may check and monitor the machine state.


D4.1 V1.0 Page 36/ 55

5.2 GESTAMP


The GESTAMP demonstrator will be linked to the demonstration of a Lighthouse manufacturing

process: FRAMETOP is a multi-stage zero defect manufacturing system of next generation

automotive chassis. The demonstrator will take place at the GESTAMP Chassis Headquarters in

the Automotive Intelligence Centre (AIC) in the Basque Country.

The demonstrator considers a multistage zero-defect manufacturing cell for the frame

components of light aluminium and steel components of next automotive models. The

manufacturing process integrates both stamping, robotized welding and inline quality control in

a multi-stage and simulation supported manufacturing process. The process should provide

2,300,000 parts a year that will be fed into OEM car manufacturers to assemble their particular

models. This process provides a critical part in the automotive assembly and one with high

throughput. Any production break due to unscheduled maintenance will have huge impacts not

simply at GESTAMP level but will propagate towards car OEMs with a potential scale in cost

impacts of several orders of magnitude. This new process that will be the reference process for

global manufacturing of this component in future car references incorporate new elements such

as inline quality control equipment towards zero defect manufacturing. Z-BRE4K should ensure

that such lines are not just zero defect but zero unexpected breakdown, supported by a common

and integrated information framework with multi-purpose application.

The demonstrator is particularly well suited to demonstrate how zero-defect manufacturing

processes and quality control information when intelligently combined with condition

monitoring and machine and component models, can lead to cognitive solutions and

prescriptive maintenance solutions that adapt machine and production operations to ensure

maximum zero-defect throughput with no unscheduled breakdowns.


In GESTAMP facilities around the world there are different competence departments.

Production strategy for chassis parts is not different from body in white strategy. As it was

previously said, manufacturing of chassis products is related to two main transformation

operations: Forming and Welding. Cold forming department makes an intermediate product

from sheet metal (broad material). This product is transferred to the Welding department.

Within the welding department the product will be assembled, which results in a finished metal

part. Finally, the assembly department combines subassemblies into a final product. Despite

forming, welding and assembly are the main departments, along the manufacturing process also

participate Quality, Maintenance and Logistics departments.

The demonstrator is particularly well suited to demonstrate how zero-defect manufacturing

processes and quality control information when intelligently combined with condition

monitoring and machine and component models can lead to cognitive solutions and prescriptive

maintenance solutions that adapt machine and production operations to ensure maximum zero-

defect throughput with no unscheduled breakdowns. The FRAMETOP process incorporates a (i)

stamping cell, (ii) robotic welding cell, (iii) intelligent fixtures system (supported by in-process


D4.1 V1.0 Page 37/ 55

simulation tools) and (iv) inline multi-sensor quality control equipment. The manufacturing

system is supported by manufacturing plant information management system (CAPTOR MES),

smart maintenance PRISMA GMAO, PROMIND (process optimization) and M3 platform for

dealing with quality control information.

Nowadays, at GESTAMP facilities, the maintenance methodology is time-based maintenance

before breakdown maintenance or maintenance activities when breakdowns occur. A

breakdown maintenance can either be a tool failure like a stroke breakage, tool wear, scrap,

stuck wire, etc. or a CTQ outside tolerances.

The Pirana system is a core system in GESTAMP´s business. It is used to manage their Assets,

Planned Preventative Maintenance Schedules, Task Schedules, Interventions and Work Order

System in each area.

Below are a few bullet points of things the system delivers

▪ Breakdown analysis of Facilities and Equipment.

▪ Numerous Automated emailed 24hr reports by Department.

▪ Automated Met-Lab job notifications.

▪ Automated emails for jobs left open.

▪ Automated emails for jobs re-assigned.

▪ Different types of work now added to Assembly.

▪ First Time Through / Weld Checks for Programmers.

▪ TPM / Clean and Check for Zone leaders.

▪ Ease to export data straight into Excel.

▪ Ability to add images, documents to Work Orders & Tasks.

▪ Pirana Reporting by Site and by Department.

When a breakdown occurs, the specific die(s) has to be exchanged and brought to the die

workshop.

▪ Visual inspection of the product and strip and tool (part) for signs of damage or wear.

▪ Analysing Asset (maintenance) history (Pirana System).

▪ Analysing Tool part life (Pirana system).

▪ Analysing Product measurements data (CTQ’s).

The outcome is used to determine the necessary type of maintenance. The type of maintenance

may be:

▪ Profiling or sharpening the profile of the worn parts.

▪ Exchange of the damaged or worn parts.

▪ Height adjustment of wear parts by means of shims.

It has to be taken into account that accuracy and the diversity of the wear parts, together with

their interactions between each other during processing is a big challenge for maintenance. On

the other hand, the quality of corrective activities depends on skills and craftsmanship of the

mechanics. In many cases highly skilled support (Engineering department) is needed in case of

(non-standard) problem solving. When Engineering department is requested for solving the


D4.1 V1.0 Page 38/ 55

problem, many management root-cause analysis tools are used, such as: A3, 5 why´s, Ishikawa

diagram, etc.

Welding

When a breakdown occurs, the specific welding equipment has to be exchanged and fixed (if

possible) in the maintenance workshop. This scenario is really stressful due to the continuous

necessity to have in stock a great amount of spare parts that might solve the breakdowns when

they occur. Such situation represents a huge budget that is not used and there is no guarantee

that will be fully used. On the other hand, when breakdowns occur in a welding cell, the mean

time to repair is critical due to the low cycle times requested. In many cases highly skilled, 2nd

line support (Engineering department) is needed in case of (non-standard) problem solving.

Based on the results achieved in the previous analysis, new internal standards and procedures

are written in order to avoid the rout cause to appear again in a production welding cell.

5.3 PHILIPS


Traditional there are different competence departments within PHILIPS Drachten. Department

Cold forming and hardening makes an intermediate product from sheet metal, this product is

transferred to department metal finishing. Within the metal finishing department, the product

will be machined, which results in a finished metal part. Finally, the assembly department

combines subassemblies into a final product.

Within the value stream map of the Cutter there was an opportunity to combine the different

competences to one production line. This is done within the cutter flow line. Sheet metal is the

start product and the product which comes out of the line is an assembled cutter placed in a

shaving cap.

The cutter flow line was a part of a cost down project. Because of this cost down project the

production strategy on this line is to keep it in full production 24/7. If the demand on products

will decrease, older production lines will be stopped to keep this new line running. The only

planned moments to do a full line stop are the preventive maintenance moments every 2 million

products. This is a full stop every 4 weeks, which takes 8 hours. Every 2 weeks there is a full stop

of 1 hour where only 1 module is exchanged (this module is exchangeable with a module from

another production line).


To guarantee the quality of the end product all critical wear parts have wear indicators (Just like

a tire thread wear indicator) and are fitted with stroke counters. The preferred maintenance

methodology is ‘intermittent condition-based maintenance’ or ‘time-based maintenance’ above

breakdown maintenance.

A breakdown maintenance can either be a tool failure like a punch breakage, tool wear, loose

scrap, etc. or a CTQ outside control limits.


D4.1 V1.0 Page 39/ 55

Current maintenance methodology when breakdowns occur:

When a breakdown occurs, the specific die(s) has to be exchanged and brought to the die

workshop. Here all the available data necessary for proper work preparation is being collected.

Preparation consists of:

a) Visual inspection of the product and strip and tool (part) for signs of damage or wear.

b) Analysing Asset (maintenance) history (Enterprise asset management (EAM) system).

c) Analysing Tool part life (EAM system).

d) Analysing Product measurements data (CTQ’s).

The outcome is being used to determine the necessary type of maintenance. The type of

maintenance can be:

a) Profiling or sharpening the profile of the worn parts.

b) Exchange of the damaged or worn parts.

c) Height adjustment of wear parts by means of shims.

The maintenance activities are supported by ‘technical out of control action plan or TOCAP and

‘tailor made, die specific maintenance manuals’.

The following shall be noted:

▪ Accuracy and the diversity of the wear parts, together with the interactions between

these parts during processing, is a big challenge for maintenance.

▪ Quality of work depends on skills and craftsmanship of the mechanics.

▪ In many cases highly skilled, 2nd line support is needed in case of (non-standard)

problem solving.

▪ Breakdowns results in high maintenance costs and a large safety stock level.


D4.1 V1.0 Page 40/ 55

6 TO-BE MAINTENANCE SCENARIOS OF THE END-USERS AFTER Z-BRE4K

SOLUTION

After the research and examination of the current maintenance status, Z-BRE4K end users have

recognized the level of smart and intelligence maintenance technologies involved within their

plants. With respect to the preventative maintenance idea, TO-BE scenarios have been prepared

and respective sub-sections 6.1-to-6.3 provide all the details.

6.1 SACMI’s Plant Maintenance Plan (TO-BE SCENARIO)

The SACMI/CDS TO-BE SCENARIO lies within the digital transformation strategy that will lead

SACMI to enhance its range of machinery with new capabilities and to respond its customers’

demands with new, improved services.

In particular, CDS will take advantage of SACMI’s enhanced CCM machine, which is being

refurbished with a HW/SW platform to go beyond the current preventive, programmed

maintenance strategies and the monitoring of critical parameters.

These new maintenance strategies will take advantage of several Z-BRE4K’s platform

components so that the risk-based and predictive maintenance features of the Z-Strategies can

be used in the CCM machine.

6.1.1 Scope

Given the complexity of SACMI’s CCM machine, the project will cover a limited number of

mechanical subsystems, as it has been already cited in other tasks so far: T1.1 User

Requirements, T1.4 Use Cases, T2.3 Machine Simulators. These three subcomponents are:

▪ Plastic Extruder (EX),

▪ Hydraulic Unit (HU),

▪ Thermal regulator (TH).

All these mechanical subsystems have been further analysed in order to enhance technology

developers with the required information for delivering a solution. Firstly, a detailed FMECA of

the mechanical components (i.e. electric motors, pumps, valves, etc.) analysis has been

reviewed and updated for each of the three use cases, which have been reported in dedicated

spreadsheets.

Moreover, a detailed analysis of all the automation and sensors has been carried out, together

with a guess match between the failure modes and the sensors that may register deterioration

trends and anomalies. Contemporaneously, the alarms of the system have been analysed, thus

being filtered out the ones not related with the failures or anomalies of the tree use-case

subsystems.

6.1.2 Data collection and analysis

For each of the three use cases reported above, SACMI, together with other Z-BRE4K partners,


D4.1 V1.0 Page 41/ 55

has collected and analysed all the technical information available and has elaborated dedicated

reports so technical partners can optimize their software components. This information can be

summarized in the following:

▪ FMECA Spreadsheet files describing mechanical components one by one, their different

failure modes, effects and criticalities for each subsystem;

▪ Sensor spreadsheet files, reporting the main technical specifications: type of signal,

frequency of measurement, interface, need for signal, postprocessing, etc.; and

mechanical components related to each of the sensors.

▪ Alarms Spreadsheet files, reporting only the alarms related to failures, filtering out all

the other generics, which may refer to non-critical issues such as the start-up routine or

related to the change in parameter configuration.

▪ Enriched FMECA Spreadsheet files, which looks to put together FMECA, sensors and

alarms for each of the failure modes described. These files describe the ontology of the

CCM machines so that the different software modules addressing the Z-Strategies can

be effectively developed.

Figure 15 illustrates the enriched FMECA file with sensor and associated alarm information for

the SACMI use case.

Figure 15. Enriched FMECA file with sensor & alarm information associated (SACMI)

6.1.3 IoT (Sensor & Automation) Gateway and Maintenance Reports HMI

SACMI’s CCMs machines have been retrofitted with HOLONIX’s iLike Machines so that condition

monitoring information, which is enhanced with the semantics of the machine, can be analysed

by the suite of ML algorithms. iLike Machines features an IoT Gateway, HW/SW infrastructure,

that provides the Predictive Maintenance module with a continuous data stream of operations

data (sensors and automation) and events (alarms) happening in the system.


D4.1 V1.0 Page 42/ 55

Simultaneously, maintenance operations information will be gathered through a tablet PC with

a dedicated application so that maintenance personnel can provide information after a

breakdown happens.

Starting from the three analysed modules (EX, HU, TH), an interactive questionnaire will gather

feedback of the maintenance operations carried out. A multiple-choice questionnaire will ask

for the subsystem that requires maintenance, and in particular which critical component (i.e.

plastic extruder screw) has to undergo maintenance. For that component, the operator will be

asked to provide feedback with regards to the failure mode occurred. Shall this failure mode not

be present in the list, the HMI will permit to add new failures not previously reported, as well as

to include comments. The gathered information will be used for the refinement of the Machine

Learning algorithms for predictive maintenance, which will make use of the machine ontology

(decision trees), operations data (sensors and automation) and alarms.

6.1.4 Machine Simulators for Preventive, Predictive and Prescriptive Maintenance through Machine Learning and physical model retrofitting

Both the operations (sensor) data and the event-based information (alarms, maintenance report

feedback) will be used by the Predictive Maintenance Module in order to deliver predictive and

prescriptive maintenance strategies. On the one hand, historical sensor and alarm data will be

combined with maintenance operations feedback so that the implementation of ML algorithms

can be optimized: rule-based methods (decision trees), regression algorithms, statistical

methods such as neural networks among others.

Additionally, this data-driven analysis presented above will be retrofitted with a reduced order

model of the physical assets, one for each of the three subsystems analysed in Z-BRE4K, in order

to improve the reliability of the ML algorithms that will both prescribe the failures and will

estimate the Remaining Useful Life of components and/or modules. Figure 16 highlights SACMI’s

machine simulators module and predictive maintenance module.

Figure 16. Machine simulators module and Predictive Maintenance module

Once the Machine Simulator of the CCM module has been tuned by the iterative integration of

the data-driven and physics-based model, the continuous stream of operations data will be

plugged into a real time Predictive Maintenance module that will look for anomalies/

abnormalities and prescribe failures (Z-PREDICT, Z-PREVENT, Z-DIAGNOSE) and will estimate


D4.1 V1.0 Page 43/ 55

Remaining Useful Life of both the critical parts and subsystems (Z-ESTIMATE).

6.1.5 Interface for Operations Management and coordination with MES

The information derived from the Predictive Maintenance Module will be combined with MES

information in order to provide additional services and a graphical user interface (GUI) for

Operations Managers, showing the main KPI and KRI of the shop floor. This decision support

system will equally improve the management of both manufacturing and maintenance

operations (Z-MANAGE) by permitting an optimization of maintenance scheduling, spare parts

order placement and stock management (Z-SYNCHRONIZE). Additionally, for each of the three

modules an analysis of the incumbent failure modes will be carried out, which will thus enhance

the temporary change of production parameters for imminent failure avoidance (Z-REMEDIATE)

in coordination with the MES.

6.2 GESTAMP’s Plant Maintenance Plan (TO-BE SCENARIO)

The GESTAMP TO-BE SCENARIO lies within the digital transformation strategy that will lead

GESTAMP to improve its level of manufacturing processes quality and performance with new

capabilities and to respond its customers’ demands with improved products and processes.

In this regard, GESTAMP has launched this strategy by upgrading and modifying the main devices

of the chassis manufacturing flow. Thus, an 800t servo hydraulic press, a laser measuring device

and the welding machines have been modified in order to improve real time process knowledge.

Therefore, these devices can now give process information in real time to be communicated to

a common platform to go beyond the current preventive, programmed maintenance strategies

and the monitoring of critical parameters.

These new maintenance strategies will take advantage of several Z-BRE4K’s platform

components so that the risk-based and predictive maintenance features of the Z-Strategies can

be used in the chassis manufacturing machines.

6.2.1 Scope

One of the historical problems of the cold stamping sector lies in the difficulty of predicting the

breaking of the tools used in cold stamping machines. Most often, preventive maintenance is

still applied, and components are replaced after a fixed number of strokes. Tonnage monitoring

equipment has been around for decades. Several models can automatically set high and low

limits around a desirable tonnage reference, if these limits are violated, a fault occurs.

Additionally, tonnage monitor can help to adjust the tonnage level to produce quality parts using

less energy and reducing the wear and tear on the press and die. In this regard, the cognitive

embedded condition monitoring (CECM) component of Z-BRE4K will explore novel techniques

of data analysis and pattern recognition to automatically extract predictive features than can be

used to optimize the maintenance, while reducing the dimensionality of the data to be

transmitted to the cloud.

On the other hand, a novel IR vision system have been developed for monitoring the arc welding


D4.1 V1.0 Page 44/ 55

process. Infrared imaging provides information of the melt pool and surrounding areas, such as

geometry and temperature distribution, the contact tip temperature or gap. This information

can be exploited for process monitoring with two aims: predictive maintenance and quality

check.

Given the complexity of GESTAMP’s manufacturing machines, the project will cover a limited

number of mechanical subsystems, as it has been already cited in other tasks so far: T1.1 User

Requirements, T1.4 Use Cases, T2.3 Machine Simulators. These three subcomponents are:

▪ 800t Servo Hydraulic Press,

▪ Welding Machines,

▪ Laser Measuring device.

All these mechanical subsystems have been further analysed in order to enhance technology

developers with the required information for delivering a solution. Firstly, a detailed FMECA of

the main manufacturing processes (i.e. stamping, welding and dimensional inspection) analysis

has been reviewed and updated for each of the three use cases, which have been reported in

dedicated spreadsheets.

Moreover, a detailed analysis of the new installed sensors has been carried out, together with a

guess match between the failure modes and the sensors that may register deterioration trends

and anomalies. Nowadays, nominal manufacturing parameters evolution is being studied to

then be able to identify deviations from these nominal curves. Afterwards, differences in curves

shape would be correlated with systems wear and NOK manufactured chassis products.


For each of the three use cases reported above, GESTAMP, together with other Z-BRE4K

partners, has collected and analysed all the technical information available and has elaborated

dedicated reports so technical partners can optimize their software components. This

information can be summarized in the following:

▪ FMECA Spreadsheet files describing mechanical components one by one, their different

failure modes, effects and criticalities for each subsystem;

▪ Sensor spreadsheet files, reporting the main technical specifications: type of signal,

frequency of measurement, interface, need for signal, post processing, etc.; and

mechanical components related to each of the sensors.

▪ Alarms Spreadsheet files, reporting all alarms related to failures. In future activities, this

alarm logs will be modified by filtering out all the other generics, which may refer to

non-critical issues such as the start-up routine or related to the change in parameter

configuration.

▪ Enriched FMECA Spreadsheet files will be developed, in order to put together FMECA,

sensors and alarms for each of the failure modes described. These files will describe the

ontology of the CCM machines so that the different software modules addressing the Z-

Strategies can be effectively developed.


D4.1 V1.0 Page 45/ 55

6.2.3 IoT (Sensor & Automation) Gateway and Maintenance Reports HMI

GESTAMP’s machines have been retrofitted with AIMEN’s Machines so that condition

monitoring information, which is enhanced with the semantics of the machine, can be analysed

by the suite of ML algorithms. AIMEN’s Machines features an IoT Gateway, that provides the

Predictive Maintenance module with a continuous data stream of operations data (sensors and

automation) and events (alarms) happening in the system. Figure 17 shows the data collection

scheme for GESTAMP’s stamping line.

Figure 17. Stamping line data collection scheme.

Two strain gage sensors have been installed in the two connecting rods of the press. These

locations have been selected to produce the most significant strain curve for press diagnostics.

The distribution of the load over the connecting rods causes inertia forces producing cyclic axial

force and stress, bending moment and stress perpendicular and parallel to the eccentric shaft

axis. The tonnage signature provides important information that can be used to make

statements about the load, change in stock thickness and hardness, part lubrication, tooling

wear, stuck scrap in the die, and misfeeding.

The collection of all the raw data from sensors is driven by the press PLC and HMI. For each

stroke an XML file is created with the following content, including force, lubrication alarms,

motor’ torque and temperature, and temperature warnings. Figure 18 presents the XML

structure for GESTAMP’s press data.


D4.1 V1.0 Page 46/ 55

Figure 18. Press data XML structure

The Press HMI has been setup to generate and transmit an XML file with the sensor data to a

common path in the central server. The CECM data acquisition module is triggered each time a

new XML file is created on the server. The acquisition module parses the XML file and structures

the raw data following the Information Model designed for the OPC-UA server.

An OPC-UA Server has been implemented to publish the raw data, relevant predictive features

and labels, and to control the CECM Software. The information model implemented in the OPC-

UA server is shown in the next figure. The server includes a CECM object that has two methods,

start () and stop () to control the embedded processing, and three vectors with the results of the

algorithms: PCA features, Classification labels and Score. Besides, the information model also

includes all the raw data coming from the sensor structured in a more intuitive way. Figure 19

illustrates GESTAMP’s OPC-UA server information model.

Figure 19. OPC-UA SERVER Information model.


D4.1 V1.0 Page 47/ 55

The integration of the CECM with the other modules of Z-BRE4K will be done through an IDS

connector implemented with FIWARE components. A system adapter based on a FIWARE IoT

Agent will be able to acquire the packed information from the OPC-UA server and convert it to

NGSI data format to feed the IDS connectors. Figure 20 provides GESTAMP’s data sharing

scheme with the IDS ecosystem.

Figure 20. Data sharing scheme with IDS ecosystem.

Two different cameras working in the long-wave infrared (LWIR 8-14um) and mid-wave infrared

(MWIR 1-5um) range have been considered and tested to develop the IR arc-welding monitoring

system. The main layout of the component is shown in Figure 21.

Figure 21. CECM system based on IR imaging for arc-welding monitoring

An OPC-UA server has been implemented in the embedded system to publish the packed

features extracted from the raw video streams. The preliminary structure of the OPC-UA

information model is shown in Figure 22, including the quality check, e.g. OK/NOT_OK part,

compressed features and current raw image.

Simultaneously, maintenance operations information will be gathered through a tablet PC with

a dedicated application so that maintenance personnel can provide information after

breakdown happen.

Starting from the three analysed modules (stamping, welding and measuring device), an

interactive questionnaire will gather feedback of the maintenance operations carried out. A

multiple-choice questionnaire will ask for the subsystem that requires maintenance, and in


D4.1 V1.0 Page 48/ 55

particular which critical component (i.e. wire feeder circuit) has to undergo maintenance. For

that component, the operator will be asked to provide feedback with regards to the failure mode

occurred. Shall this failure mode not be present in the list, the HMI will permit to add new

failures not previously reported, as well as to include comments. The gathered information will

be used for the refinement of the Machine Learning algorithms for predictive maintenance,

which will make use of the machine ontology (decision trees), operations data (sensors and

automation) and alarms.

Figure 22. Preliminary structure of the OPC-UA information model


Sensor data and the event-based information (alarms, maintenance report feedback) will be

used by the Predictive Maintenance Module in order to deliver predictive maintenance

strategies (Figure 23). Historical sensor and alarm data will be combined with maintenance

operations feedback so that the implementation of ML algorithms can be optimized: rule-based

methods (decision trees), regression algorithms, statistical methods such as neural networks

among others.

Figure 23. FMEA for Forming Operations


D4.1 V1.0 Page 49/ 55

In this regard, this data-driven analysis presented above will be retrofitted with a reduced order

model of the physical assets, one for each of the three subsystems analysed in Z-BRE4K, in order

to improve the reliability of the ML algorithms that will both prescribe the failures and will

estimate the Remaining Useful Life of components and/or modules.

Once the Machine Simulator (Figure 24) has been tuned by the iterative integration of the data-

driven and physics-based model, the continuous stream of operations data will be plugged into

a real time Predictive Maintenance module that will look for anomalies/abnormalities and

prescribe failures (Z-PREDICT, Z-PREVENT, Z-DIAGNOSE) and will estimate Remaining Useful Life

of both the critical parts and subsystems (Z-ESTIMATE).

Figure 24. Machine simulators module developed for GESTAMP’s modules

6.2.5 Interface for Operations Management and coordination with MES

The information derived from the Predictive Maintenance Module will be combined with MES

information in order to provide additional services and a graphical user interface (GUI) for

Operations Managers, showing the main KPI and KRI of the shop floor. This decision support

system will equally improve the management of both manufacturing and maintenance

operations (Z-MANAGE) by permitting an optimization of maintenance scheduling, spare parts

order placement and stock management (Z-SYNCHRONIZE). Additionally, for each of the three

modules an analysis of the incumbent failure modes will be carried out, which will thus enhance

the temporary change of production parameters for imminent failure avoidance (Z-REMEDIATE)

in coordination with the MES.


D4.1 V1.0 Page 50/ 55

6.3 PHILIPS’ Plant Maintenance Plan (TO-BE SCENARIO)

With the solutions provided by Z-BRE4K the maintenance will transform from time-based

maintenance to predictive maintenance as mentioned in 4.2 or even to prescriptive

maintenance. Predictive maintenance will lead to several big benefits. Some examples of the

benefits are: no more unnecessary maintenance stops and shorter lead times for maintenance

due to prescription.

6.3.1 Scope

As described in earlier tasks the scope of the PHILIPS use case lies within the tooling or dies

which are placed within the cold forming press. To make sure the Z-BRE4K components are able

to perform to its maximum there are several sensors placed within these dies, as well as before

the machine to collect data from the incoming material. At the end of the cold forming process

there is a measuring machine which does a measurement on 100% of the products. This data is

also provided to the Z-BRE4K components. All of the data from the cold forming press itself as

well as the die counters and maintenance logs will also be input for the various Z-BRE4K

components.

Especially tools number 1, 2, and 6 are interesting and used as three separate use cases within

the project:

▪ Tool number 1 has several punching steps in the process. This step cuts out the rough

shape of the cutting elements.

▪ Tool number 2 is a flattening step in the process. This flattening is used to get the right

thickness on the end of the cutter tips.

▪ Tool number 6 has again some punching steps in the process. Which results directly in

a critical to quality parameter within the measuring machine.


As described in deliverable 1.4 and in the scope above there are several data collection points.

▪ Data for the steel strip coming in is collected as a reel and for every product individually

there is the thickness and the temperature at the time of measuring thickness.

▪ Within the dies there are 6 acoustic emission sensors which are continuously monitoring

the system but only collecting a signal once every minute.

▪ The press itself has some status and error codes and is equipped with an OPCUA Data

collector which can collect data from the press every cycle. E.g. press force, oil

temperature, motor amperes. Product counters etc.

▪ After the press there is a measuring machine which does a measurement on 100% of

the cutting elements.

▪ A separate measurement is done based on statistical process control to make sure the

cutters have the right thickness.

During analysis of these data points it became clear that it is hard to synchronize all the data

points. This because of the different frequencies, product related or process related timestamps.


D4.1 V1.0 Page 51/ 55

A benefit is that all the product related parameters are coupled by the dotmatrixcode (DMC) on

the product. With the DMC we are able to couple a lot of data. Unfortunately, not yet all the

data.

PHILIPS is currently busy to get the last part of synchronisation between the products, the press

and the acoustic emission sensors in place. This will be done by using product counters and

resetting them on frequent bases.

Above data is all real-time data. To get up to speed with the different analysis both the FMECA’s

and the CAD models were shared as well.

With the FMECA’s Z-BRE4K is able to get faster results in the predictive models and where

possible already in prescriptive. Because of the known or probable effects to failures which can

be coupled in the historical data.

With the CAD models Z-BRE4K is able to simulate production and start to see breakdowns

coming even if these breakdowns are not yet in the dataset of the real-time machine data.

6.3.3 IoT (Sensor & Automation) Gateway

All real-time data is collected by sensors and machines coupled to the factory network of

PHILIPS. Within this Factory network there are edge devices which collect the data. Some of the

data are already in readable formats, other are still raw data. After processing the data, the edge

devices send the data to the different database solutions within the PHILIPS factory.

From the different databases the relevant data is collected and send through the industrial data

space connector to the Orion context broker (Figure 25).

Figure 25. Data Stream


When all the data from the above mentioned paragraphs arrives in the Z-BRE4K system there

are eight different strategies on handling the data.

Z-Predict will start up using the data and errors from the historical data and the simulations from

the physical models. With machine learning algorithms Z-predict will be able to see the

upcoming errors. Using this Information together with expert knowledge from the FMECA’s Z-

Prevent will be able to give data to Z-ESTIMATE to make an estimation on remaining useful life

(RUL). Z-DIAGNOSE will help to make this loop smarter when the Z-BRE4K system is working real-


D4.1 V1.0 Page 52/ 55

time.

The remaining useful life from Z-ESTIMATE will make sure there is no unnecessary maintenance.

With the current time-based maintenance, it is possible to have maintenance to soon which

leads to more maintenance and more production stops. On the other hand, there is a risk on

doing maintenance too late. This will lead to breakdowns and corrective maintenance. This also

leads to production stops. Especially in the weekends, when there are no mechanics on the

premises, these breakdowns take a long time. Mechanics have to be warned and they have to

come to the factory to fix these breakdowns.

The Z-MANAGE, Z-REMEDIATE and Z-SYNCHRONISE strategies will make the predictive

maintenance into prescriptive maintenance by giving advice on how to handle the upcoming

maintenance moments. This means the work to be done will be clear for a mechanic and it will

shorten the lead-time. As described in 5.3.2, preparation is a great part of the time-based

maintenance which is current in place. If this can be done during normal operation, maintenance

time will be shortened.

6.3.5 Interface for Operations and Maintenance

The information from the different Z-BRE4K modules should be presented to the different

stakeholders. For the operators there should be a graphical user interface which states the

condition of the different tools on the press. The decision support system should help the

operator in the choice to either change or repair a tool or to do full maintenance on the set of

tools.

The maintenance department should have the information of the parts which need to be

repaired/replaced and the mean time to repair for the maintenance tasks coming up. With this

information all the maintenance tasks and maintenance technicians can be planned in order to

get the tasks done as efficient as possible.

A high level information system should be available for production management to give a quick

overview on availability of coming period.

The complete package of the Z-BRE4K system with the right integration in the production and

maintenance system will make sure there are no unexpected breakdowns and will decrease

maintenance time significant.


D4.1 V1.0 Page 53/ 55

7 CONCLUSION

The deliverable D4.1 reported on the activities and results of the Task T4.1, i.e. strategies to

improve maintainability and increase operating life of production. Accordingly, initial

conception of Z-BRE4K strategies as well as their implementation, AS-IS and TO-BE embedded

intelligence systems, along with the state of the art industrial maintenance strategies and

policies implemented in manufacturing have been addressed. Besides, the main emphasis has

been highlighting the orientation of the plants’ maintenance plan from reactive/preventive to

predictive in order to shift from AS-IS maintenance strategies and policies of Z-BRE4K end-users

towards TO-BE maintenance scenarios after implementation of the Z-BRE4K solution. Based on

the insights from the deliverable, T4.2 will then develop the algorithms to optimize maintenance

vs. production and decide on the optimal combination of Z-Strategies to deploy.


D4.1 V1.0 Page 54/ 55

REFERENCES

▪ Amruthnath, N. and Gupta, T., 2018a, April. A research study on unsupervised machine learning algorithms for early fault detection in predictive maintenance. In 2018 5th International Conference on Industrial Engineering and Applications (ICIEA)(pp. 355-361). IEEE.Amruthnath, N. and Gupta, T., 2018b, March. Fault class prediction in unsupervised learning using model-based clustering approach. In Information and Computer Technologies (ICICT), 2018 International Conference on (pp. 5-12). IEEE.

▪ Castelfranchi, C., 1994, August. Guarantees for autonomy in cognitive agent architecture. In International Workshop on Agent Theories, Architectures, and Languages (pp. 56-70). Springer, Berlin, Heidelberg.

▪ CCPS (2008). The Public Version of Dow’s Reactive Chemicals Program document was provided on CD to those purchasers of Guidelines for Hazard Evaluation Procedures, 3rd Edition, Center for Chemical Process Safety (CCPS), ISBN: 978-0-471-97815-2.

▪ CCPS (1999). Guidelines for Chemical Process Quantitative Risk Analysis. New York, American Institute of Chemical Engineering, Center of Chemical Process Safety (CCPS), ISBN 978-0-8169-0720-5.

▪ Coleman, C., Damodaran, S., Chandramouli, M., Deuel, E., 2017. Making maintenance smarter - Predictive maintenance and the digital supply network, May 2017, [online] Available: https://www2.deloitte.com/insights/us/en/focus/industry-4-0/using-predictive-technologies-for-asset-maintenance.html.

▪ Elmenreich, W., 2003, June. Intelligent Methods in Embedded Systems. In WISES (pp. 3-11).

▪ Goldberg, B.E., Everhart, K., Stevens, R., Babbitt III, N., Clemens, P. and Stout, L., 1994. System engineering toolbox for design-oriented engineers.

▪ Fiix (2018). Risk-based maintenance. Available at https://www.fiixsoftware.com/maintenance-strategies/risk-based-maintenance/

▪ Genesereth, M.R., Ketchpel, S.P., 1994. Software agents. Communications of the ACM, 37(7), pp.48-53.

▪ Guo, B., Zhang, D., Yu, Z., Liang, Y., Wang, Z. and Zhou, X., 2013. From the internet of things to embedded intelligence. World Wide Web, 16(4), pp.399-420.

▪ ICH (2005). International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (2005). ICH Harmonised Tripartite Guidelines – Quality risk management Q9.

▪ IEC (2006). IEC 60812 Technical Committee, 2006. IEC 60812. Analysis Techniques for System Reliability-Procedure for Failure Mode and Effects Analysis (FMEA).

▪ IEC (1990). International Electrotechnical Commission, 1990. IEC 60050 (191) International Electrotechnical Vocabulary, Chapter 191: Dependability and quality of service. Geneva: Author.

▪ Kennedy, S. (2006). New tools for PdM. Plant Services. ▪ Khan, F.I. and Haddara, M.M., 2003. Risk-based maintenance (RBM): a quantitative

approach for maintenance/inspection scheduling and planning. Journal of loss prevention in the process industries, 16(6), pp.561-573.

▪ Krishnasamy, L., Khan, F. and Haddara, M., 2005. Development of a risk-based maintenance (RBM) strategy for a power-generating plant. Journal of Loss Prevention in the Process Industries, 18(2), pp.69-81.

▪ Lacey, P., 2011. An application of fault tree analysis to the identification and management of risks in government funded human service delivery.

https://www2.deloitte.com/insights/us/en/focus/industry-4-0/using-predictive-technologies-for-asset-maintenance.html

https://www2.deloitte.com/insights/us/en/focus/industry-4-0/using-predictive-technologies-for-asset-maintenance.html

https://www.fiixsoftware.com/maintenance-strategies/risk-based-maintenance/


D4.1 V1.0 Page 55/ 55

▪ Liggan, P. and Lyons, D., 2011. Applying predictive maintenance techniques to utility systems. Pharmaceutical Engineering.

▪ Miguelañez-Martin, E. and Flynn, D., 2015. Embedded intelligence supporting predictive asset management in the energy sector.

▪ Moubray, J., 1997. Reliability-centered maintenance. Industrial Press Inc. ▪ Nag, R., Hambrick, D.C. and Chen, M.J., 2007. What is strategic management, really?

Inductive derivation of a consensus definition of the field. Strategic management journal, 28(9), pp.935-955.

▪ OSHA (1994). Process safety management guidelines for compliance. US Department of Labor-Occupational Safety and Health Administration.

▪ Pathak, A., 2018. Asset performance management 4.0 and beyond. Available at https://reliabilityweb.com/articles/entry/asset-performance-management-4.0-and-beyond

▪ Pintelon, L., Pinjala, S.K. and Vereecke, A., 2006. Evaluating the effectiveness of maintenance strategies. Journal of Quality in Maintenance Engineering, 12(1), pp.7-20.

▪ Rausand, M. and Høyland, A., 2004. System reliability theory: models, statistical methods, and applications (Vol. 396). John Wiley & Sons.

▪ Robin, L. (2006) "Slick tricks in oil analysis" Plant Services. ▪ Sakai, S., 2010. Risk-based maintenance. JR East Technical Review, 17. ▪ Shin, J.H. and Jun, H.B., 2015. On condition-based maintenance policy. Journal of

Computational Design and Engineering, 2(2), pp.119-127. ▪ SKF (2018). Risk Based Maintenance. Available at http://www.skf.com/caribbean/

services/asset-management-services/asset-efficiency-optimisation/strategize/maintenance-strategy-review/risk-based-maintenance/index.html

▪ Tay, K.M. and Lim, C.P., 2008. On the use of fuzzy inference techniques in assessment models: part II: industrial applications. Fuzzy Optimization and Decision Making, 7(3), pp.283-302.

▪ Tixier, J., Dusserre, G., Salvi, O. and Gaston, D., 2002. Review of 62 risk analysis methodologies of industrial plants. Journal of Loss Prevention in the process industries, 15(4), pp.291-303.

▪ Toms, L.A., 1995. Machinery Oil Analysis: Methods, Automation&benefits: A Guide for Maintenance Managers & Supervisors. Larry A. Toms.

▪ Wang, K., 2016. Intelligent predictive maintenance (IPdM) system–Industry 4.0 scenario. WIT Transactions on Engineering Sciences, 113, pp.259-268.

▪ Wilson, P.F., 1993. Root cause analysis: A tool for total quality management. ASQ Quality Press.

▪ Wooldridge, M. and Jennings, N.R., 1995. Intelligent agents: Theory and practice. The knowledge engineering review, 10(2), pp.115-152.

▪ Yung, C. (2006). Vibration analysis: what does it mean? Plant Services. ▪ [Z-BRE4K 2017] Z-BRE4K Project. “Grant Agreement – 768869”, July 2017. ▪ [Z-BRE4K 2018] Z-BRE4K Project. “D1.1: Z-BRE4K user requirements”, January 2018. ▪ [Z-BRE4K 2018] Z-BRE4K Project. “D1.3: Z-BRE4K system architecture”, January 2018. ▪ [Z-BRE4K 2018] Z-BRE4K Project. “D1.4: Z-BRE4K use cases”, January 2018. ▪ [Z-BRE4K 2019] Z-BRE4K Project. “D1.5: Z-BRE4K strategy implementation and risk

analysis”, 2019.

https://reliabilityweb.com/articles/entry/asset-performance-management-4.0-and-beyond

https://reliabilityweb.com/articles/entry/asset-performance-management-4.0-and-beyond

http://www.skf.com/caribbean/%20services/asset-management-services/asset-efficiency-optimisation/strategize/maintenance-strategy-review/risk-based-maintenance/index.html




Date post:	11-Jul-2020
Category:	Documents
Upload:	others
View:	6 times
Download:	0 times

Deliverable D4.1 Z-BRE4K strategies and policies · Z-BRE4K Project Grant Agreement nº 768869 –...

Documents