Upload
others
View
7
Download
0
Embed Size (px)
Citation preview
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
5.2.1 Production System ....................................................................................... 36
5.2.2 Plant Strategy and Policies ........................................................................... 36
5.3 PHILIPS ............................................................................................................................. 38
5.3.1 Production System ....................................................................................... 38
5.3.2 Plant Strategy and Policies ........................................................................... 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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
6.2.2 Data collection and analysis ......................................................................... 44
6.2.3 IoT (Sensor & Automation) Gateway and Maintenance Reports HMI ........ 45
6.2.4 Machine Simulators for Preventive, Predictive and Prescriptive
Maintenance through Machine Learning and physical model retrofitting ................ 48
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
6.3.2 Data collection and analysis ......................................................................... 50
6.3.3 IoT (Sensor & Automation) Gateway ........................................................... 51
6.3.4 Machine Simulators for Preventive, Predictive and Prescriptive
Maintenance through Machine Learning and physical model retrofitting ................ 51
6.3.5 Interface for Operations and Maintenance ................................................. 52
7 CONCLUSION ................................................................................................... 53
REFERENCES ........................................................................................................... 54
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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).
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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).
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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/
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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/
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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;
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
D4.1 V1.0 Page 36/ 55
5.2 GESTAMP
5.2.1 Production System
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.
5.2.2 Plant Strategy and Policies
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
5.3.1 Production System
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).
5.3.2 Plant Strategy and Policies
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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,
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
6.2.2 Data collection and analysis
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
6.2.4 Machine Simulators for Preventive, Predictive and Prescriptive Maintenance through Machine Learning and physical model retrofitting
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
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
6.3.2 Data collection and analysis
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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
6.3.4 Machine Simulators for Preventive, Predictive and Prescriptive Maintenance through Machine Learning and physical model retrofitting
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-
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.
Z-BRE4K Project Grant Agreement nº 768869 – H2020-FOF-2017
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.