deliverable d 6.3 normative input for standardisation of
TRANSCRIPT
G A 7 7 7 5 1 3 P a g e 1 | 51
Indicator Monitoring for a new railway Paradigm in seamlessly
integrated Cross modal Transport chains – Phase 2
Deliverable D 6.3
Normative input for standardisation of CBM data
structures
Reviewed: (yes/no)
This projecthas received funding from the EuropeanUnion'sHorizon2020 Programme
ResearchandInnovationactionundergrantagreementNo 777513.
Thisdocumentreflectstheviewsoftheauthor(s)anddoesnotnecessarilyreflecttheviewsor
policyoftheEuropeanCommission.Whilsteffortshavebeenmade toensuretheaccuracyand
completenessofthisdocument, the IMPACT-2 consortiumshallnotbeliableforanyerrorsor
omissions,howevercaused.
Project acronym: IMPACT-2
Starting date: Month 16
Duration (in months): 12
Call (part) identifier: H2020-S2RJU-CFM-2017
Grant agreement no: 777513
Due date of deliverable: Month 28
Actual submission date: 15-11-2019 Responsible/Author: Harald Ackermann, Georg Ermer, Tadeusz Szczepaniak;
DB Systemtechnik Dissemination level: PU Status: Draft
Ref. Ares(2020)580129 - 30/01/2020
G A 7 7 7 5 1 3 P a g e 2 | 51
Document history
Revision Date Description
1 16/09/2019 First version of deliverable
2 25/10/2019 Second version of deliverable
3 15/11/2019 Third version of deliverable
Report contributors Name Beneficiary Short Name Details of contribution
Claudio Cavalletti Hitachi Rail STS Deliverable 6.2
Neto Helder CP/ EMEF Additional information and comments
Benoit Guyot SNCF Additional information and comments
G A 7 7 7 5 1 3 P a g e 3 | 51
Table of contents
Abbreviations and acronyms ................................................................................................................... 7
1 Executive summary ......................................................................................................................... 9
2 Background .................................................................................................................................... 10
3 Einleitung ....................................................................................................................................... 11
4 Input for Standardising CBM Data ................................................................................................. 13
4.1 Definition of CBM data ..............................................................................................13
4.2 Overview of CBM data ...............................................................................................14
4.3 Overview regarding input from WP6 ...........................................................................15
4.4 Existing standards ongoing standardization activities ...................................................16
4.5 Digital data model (cyber-physical model)/ Deliverable 6.5............................................17
4.6 Collaboration with other Shift2Rail partners ................................................................18
5 Selection of Standardisation Areas ............................................................................................... 21
5.1 Basic approach to selection ........................................................................................21
5.2 Data flow and format ................................................................................................22
5.2.1 Data flow ....................................................................................................................... 22
5.2.2 Data format ................................................................................................................... 23
5.3 Frequency of data extraction and transfer ...................................................................24
5.4 Physical data transmission from vehicles to ground ......................................................25
5.5 On-board rolling stock software and databases ............................................................26
5.6 Landside data collection and storage ..........................................................................27
5.7 Access and transfer to CBM visualization and application systems ..................................28
5.8 Common data semantics for alignment and interoperability of vehicle condition .............30
6 Standardisation Proposals Regarding Data Structure and Data Designation ............................... 31
6.1 Data structure model ................................................................................................31
6.1.1 Background and objectives ........................................................................................... 31
6.1.2 Basic structure of the "generic data model" ................................................................. 32
6.1.3 Product group structure ................................................................................................ 32
6.1.4 Function group .............................................................................................................. 34
6.1.5 Component part condition and error description ......................................................... 35
6.1.6 Location structure ......................................................................................................... 36
6.1.7 Class-specific codes / examples ..................................................................................... 38
6.1.8 Summary relating to the generic data model ............................................................... 39
G A 7 7 7 5 1 3 P a g e 4 | 51
6.2 Data designation model .............................................................................................40
6.2.1 Background, problem and objectives ............................................................................ 40
6.2.2 Basic structure and function.......................................................................................... 40
6.2.3 Examples of uniform data designation .......................................................................... 42
6.2.4 Summary of the data annotation .................................................................................. 44
6.3 Benefit of standardisation for CBM usage in the rail sector ............................................45
7 Conclusions and data requirements for CBM ............................................................................... 46
8 Future Development ..................................................................................................................... 47
9 Further steps towards standardization ......................................................................................... 48
10 References ................................................................................................................................. 49
11 Annexes ..................................................................................................................................... 50
12 Antitrust statement ................................................................................................................... 51
G A 7 7 7 5 1 3 P a g e 5 | 51
List of tables
Table 1: Summary of the relevant data for the CBM from Deliverable 6.2 .......................................... 15
Table 2: Overview of the existing standards and ongoing standardization activities ........................... 16
Table 3: Contact to other Shift2Rail projects ........................................................................................ 19
Table 4: Product group structure in the generic data model ................................................................ 33
Table 5: Extract from the main product groups in EN 15380-2:2006-06 .............................................. 33
Table 6: Excerpt from subproduct groups in EN 15380-2:2006-06 ....................................................... 33
Table 7: Excerpt from EN 15380-2:2006-06 with examples to which a number is assigned to use them
as sub-subproduct group....................................................................................................................... 33
Table 8: Extract from the ISI component part catalogue ...................................................................... 34
Table 9: Extract from the ISI component part catalogue ...................................................................... 34
Table 10: Precise component part type description according to the generic data model, consisting of
product group and function group ........................................................................................................ 35
Table 11: Component part condition and error description in the generic data model....................... 35
Table 12: Component part conditions as per DB condition catalogue ................................................. 36
Table 13: Location group in the generic data model ............................................................................ 37
Table 14: Mounting position group, mounting positions and installation position based on the generic
data model ............................................................................................................................................ 37
Table 15: Detailed explanation of the example for data designation ................................................... 42
Table 16: The data set "Environment" of the product "Dust sensor" contains among other things the
attribute "DustLevelConcentration" with the data type Float with the unit "g / m³". ......................... 44
Table 17: A sensor of the product "Soft" contains the attribute "IsSensorFailure" with the data type
Boolean - Without unit in the set "Error message" because only "True" or "False" is returned. ......... 44
Table 18: The product "temperature sensor" has among other things the attributes of the "maximum
and minimum room temperature" in the unit "Kelvin" in the data type Float in the data set
"Parameter"........................................................................................................................................... 44
G A 7 7 7 5 1 3 P a g e 6 | 51
List of illustration
Figure 1: Four quadrant model in maintenance ................................................................................... 11
Figure 2: Cyber-physical model taking the "Vehicle monitors vehicle" quadrant by way of example . 12
Figure 3: Cyber-physical model for Vehicle monitors itself .................................................................. 18
Figure 4: cyber physical modell with input from the partners .............................................................. 20
Figure 5: Cyber-physical model added with Support tool and Arrow to it ........................................... 28
Figure 6: Basic structure of the generic data model ............................................................................. 32
Figure 7: Assignment of locations within the generic data model ........................................................ 37
Figure 8: Class code based on the generic data model ......................................................................... 38
Figure 9: Fault codes Class 430 translated into error codes based on the generic data model ........... 39
Figure 10: Operating data Class 430 translated into the generic data model ...................................... 39
G A 7 7 7 5 1 3 P a g e 7 | 51
Abbreviations and acronyms
Abbreviation / acronym
Description
BT Bombardier Transportation
BUS System for data transmission
CCA Condition Based Maintenance
CBM Condition Based Maintenance
CCA Cross Cutting Activities
CCS Control Command and Signaling
CDM Conceptual Data Model
CEN/ CENELEC Comité Européen de Normalisation - European Committee for Standardization
CLC CENELEC
DB Deutsche Bahn
DIN Deutsches Institut für Normung -
EDP electronic data processing
ECM Entity in Charge of Maintenance
EMU Electrical Multiple Unit
EN Europäische Norm – European Standard
ETB Ethernet Train Backbone
EDP Electronic data processing
EuroSpec European Specifications (for railway vehicles) - initiative of several European railway
companies
ISI Integriertes System Instandhaltung - Integrated system of maintenance of the DB
JSON Javascript Object Notation – a programming language
JU Joint Undertaking -
HTTPS Hypertext Transfer Protocol Secure – a Standard for data exchange
IMPACT-2 Indicator Monitoring for a new railway PAradigm in seamlessly integrated Cross modal
Transport chains – Phase 2
IT Information technology
LCC Life Circle Costs
G A 7 7 7 5 1 3 P a g e 8 | 51
Abbreviation / acronym
Description
LRU Last Replaceable Unit
MQTT Message Queuing Telemetry Transport - a Standard for data exchange
MPG Main Product Group – Assignment within the norm EN 15380-2
MVB Main vehicle bus
OPC UA Open Platform Communications Unified Architecture – a Standard for data exchange
S2R Shift2Rail
SPG Sub Product Group – Assignment within the norm EN 15380-2
TBD To Be Done
TCN train communication network
TCMS Train Control & Management System
TD Technical Demonstrator
TRL Technology Rediness Level
UIC International Union of Railways (UIC, French: Union internationale des chemins de fer)
UL Underwriters Laboratories – a global safety certification company
USB Universal Serial Bus
WA Work Area
WP Work Package
WTB Wire Train Bus
XML eXstensible Markup Language – a programming language
G A 7 7 7 5 1 3 P a g e 9 | 51
1 Executive summary
Deliverable 6.4 provides a summary of the work in Task 6.7 Standardisation within the Shift2Rail
project from Work Package 6 – Smart Maintenance.
On the basis of the range of work within WP6, and in collaboration with the specified partner working
groups it was found that standardisation for data-based maintenance, such as CBM or predictive
maintenance, is crucial. Necessary standardisation areas and standardisation solutions for CBM data
are illustrated in the following deliverable.
It must, however, also be stated that it is not worthwhile for individual areas to implement
standardisation, since this kind of standardisation could entail restrictions in the future work, or
standardisation would not bring any benefits. These "negative examples" are also described and
justified.
The data used for maintenance essentially follows the same processing sequence; as such, it is no
longer necessary to differentiate between the four quadrants of CBM. The main problems with
processing were identified in this respect in the area of the structure, the designation and the
classification of the data, as well as with just a single interface to all data.
To this end, two standardisation approaches were presented. These entailed standardising the way in
which the volume of data is structured – data structure model – and standardising the data designation
– data designation model.
Finally, the advantages of standardisation are highlighted, the connections with other deliverables
from WP6, and suggestions provided for the further treatment of this topic.
G A 7 7 7 5 1 3 P a g e 10 | 51
2 Background
This document "Normative input for standardisation of CBM data structures" illustrates the problems
associated with maintenance when processing data and provides initial solutions in this respect.
It examines problems associated with maintaining infrastructure and vehicles in the overall rail system.
From the insights regarding the existing data from Deliverable 6.6 CBM Data Structures, problems,
which occur when processing data for maintenance, were discussed in consultation with projects
outside and within WP6.
In connection with Deliverable 6.5 Smart maintenance concept for infrastructure, rolling stock and
CCS, the basic approaches to data processing were found.
In connection with Deliverable 6.4 Results for rail vehicles, problems with data processing and data
recognition were found.
In connection with the partner working groups [TD3.6 (Dynamic Railway Information Management
System Demonstrator), WA 4.2 (Integrated Mobility), TD 2.9 (Traffic Management System), TD 1.1
(Traction), TD 1.2 (Train Control and Monitoring System), TD 5.1 (Freight Electrification, Brake and
Telematics) and TD 5.3 (Wagon Design)], contradictions with these were excluded and the respective
solution strategies discussed.
In connection with WA 3.2 (Standardisation), ongoing standardisation topics were analysed.
G A 7 7 7 5 1 3 P a g e 11 | 51
3 Einleitung
Ongoing digitalisation and the associated Industry 4.0 are being used in an increasing number of
industrial segments. As such, this also influences the rail sector with associated maintenance of
vehicles and infrastructure. Increasingly, the rail system has access to new public opportunities for
using data and new data analysis systems. These opportunities can be used to gradually transition from
scheduled preventive maintenance to data-based maintenance. As a result, condition-based
maintenance (CBM) can be introduced in an initial step. In a further step, predictive maintenance is
also conceivable, which analyses and processes historic or data available in real time that is relevant
to maintenance, to predict in advance when faults or problems will arise in a system. Data-based
maintenance should save time and resources, and guarantee or even improve safety. The keys to doing
so are, however, the control and reliability of the data.
Irrespective of which area generates data in the rail system and its technical characteristics, the data
follows a similar, shared logic, whereby the same basic requirements are placed on the data.
The data generation process involves four quadrants: vehicle monitors itself, vehicle monitors
infrastructure, infrastructure monitors vehicle, and infrastructure monitors itself - See Figure 1 below.
Figure 1: Four quadrant model in maintenance
To use data for data-based maintenance, the data must, irrespective of in which quadrant the data is
generated, then be transferred to a uniform structure so it can be analysed subsequently. Interim
storage is possible following generation, structuring and/or after the analysis. Subsequently the data
is provided to the maintenance operator by means of a maintenance tool in the form of a work order.
G A 7 7 7 5 1 3 P a g e 12 | 51
This data flow has already been considered in Deliverable 6.5 for the flow of information as a cyber-
physical model and is again copied for data in this deliverable. Figure 2 below shows the information
flow from Deliverable 6.5; more detailed information follows in section 4.5:
Figure 2: Cyber-physical model taking the "Vehicle monitors vehicle" quadrant by way of example
Although the process is very similar in all rail areas, but also within a quadrant, there are only very few
end-to-end harmonised solutions. As such, urgent harmonisation is required here to achieve
compatibility between various solutions and to prevent additional work as a result. The main problem
in this respect is finding the right data, understanding that data and assigning it accordingly.
As a partial solution to these problems, this deliverable helps find a solution for standardising data in
all rail areas and to manage, understand and process data better in future.
G A 7 7 7 5 1 3 P a g e 13 | 51
4 Input for Standardising CBM Data
4.1 Definition of CBM data At the time of publication of this document, there is still no official definition of the term CBM data
(CBM = condition-based maintenance). Since this term, however, is constantly being used in relation
to maintenance and, in turn, to maintenance data, a definition initially follows within this deliverable:
CBM data is taken to mean hereinafter all available data (e.g. raw data, error messages, process data,
condition logs, sensor data) of a vehicle or other asset, which can be used to assess the condition of
those assets. In addition, data is also included that has already been processed, e.g. from the
associated analysis program, i.e. all data that reflects the condition of an asset across the entire process
chain.
The following are examples of CBM data from Deliverable 6.4:
• raw sensor signals (analogue and digital) from individual measuring devices and actuators
• process & state variables incl. message and events data from TCN - train communication
network (incl. MVB, WTB, ETB and periphery BUS)
• operational parameters displayed or/and recorded on the vehicle (central and local displays)
• diagnostic messages displayed or/and recorded on the vehicle (central and local displays)
• records of diagnostic messages and parameters transmitted on the landside (i.e. on-board
diagnostic and data transmission)
G A 7 7 7 5 1 3 P a g e 14 | 51
4.2 Overview of CBM data In addition to the hitherto classic methods of assessing the condition of component parts, components
and whole assets using conventional monitoring (regular measurements, testing and visual inspection,
etc.), the targeted monitoring of CBM data provides another option for permanent monitoring of
assets and derivation of individual maintenance requirements.
Since to date there are no adequate recording and processing standards, most projects (e.g.
information on a vehicle class, a classification yard or a barrier system) were programmed and created
independently of each other. As a result, existing data provided and available on the assets have
different data formats, structure etc. in the entire rail sector. The same data is also read out, stored,
transmitted, analysed and processed further individually in each case.
The interfaces of the various assets for data processing are in some cases incompatible or prove very
time-consuming to design so that the same processing and analysis software can process them.
Moreover, the processing systems, in their current configuration, are structured differently in each
company with their interface to the database for the maintenance tasks and accordingly fitted with
different interfaces.
The structure of the data or the encoding with each asset is also different.
Collaboration between various companies or various projects proves difficult with the aforementioned
different basic requirements and in any case entails a huge amount of additional work.
The goal must be to find common solutions, based on the current partial solutions from all the project
partners and on the existing overarching standards in IT. For this reason it is important to first examine
these standards.
G A 7 7 7 5 1 3 P a g e 15 | 51
4.3 Overview regarding input from WP6 The analyses regarding Deliverable 6.3 should be based on the results of Deliverable 6.2.
Table 1 from the summary of Deliverable 6.2 is shown below, which provides a brief overview of the
results. Deliverable 6.2 analysed over 100 different use cases within Shift2Rail. As a result, it was
recognised that the data and how that data is processed differ greatly.
Table 1: Summary of the relevant data for the CBM from Deliverable 6.2
Quadrant Units Frequency Format Protocol
Infrastructure/
infrastructure
35 TBD/ Configurable XML or
JSON
OPC UA, TCP/IP
Infrastructure
monitors vehicle
15 After every train run XML,
JSON, jpg/ TBD
TBD/
OPC UA, TCP/IP
Vehicle monitors
infrastructure
8 Each day/ After every
train run
Binary, TBD G4, LTE, rsync,
TBD
Vehicle monitors
vehicle
50 Permanent, 60,000 km,
every 15 min.
JSON, MVB TBD,
dat, JSON,
TBD/ Mobile
Network
Further, input from WP6 was provided with the aid of a survey and discussed during the work package
meetings; even more differences were found in this respect, such as in the data structure and data
processing. This helped find the suitable approaches to standardisation, as described in section 6.
G A 7 7 7 5 1 3 P a g e 16 | 51
4.4 Existing standards ongoing standardization activities There are already numerous standards and ongoing standardisation activities relating to CBM data.
Table 2 below illustrates selected important standards and standardisation activities. Standardisation
approaches from the results of the work within Shift2Rail are illustrated in WA 3.2.
Table 2: Overview of the existing standards and ongoing standardization activities
Standard Beschreibung Aktivität
IEC61375 series - Electronic railway equipment - Train communication network (TCN)
The standard shows the architecture of data communication systems (architecture of a communication system for the data communication between vehicles of the open trains, the data communication within the vehicles and the data communication from train to the ground) in trains.
active
UIC 438-3, Identification marking for tractive stockyes
Illustrates the system used to identify rail vehicles. In this way, the vehicle can be assigned uniquely to a class, a country and a company, and it can be identified which type of vehicle is involved.
active
UIC 556 - Information transmission in the train (train bus)
The leaflet shows the general conditions that the data transmission equipment of RIC-compatible passenger trains must fulfil.
active
UIC 557 – Diagnostic on passenger rolling stock
The leaflet describes the functionality of the various diagnosis systems in rail vehicles.
active
UIC 559 - Specification "Diagnostic Data Transmission" from railway vehicles
The present document specifies a common content and format for "diagnostic data transmission" from railway vehicles, as well as basic processing structures of data in XML.
active
UL 4600 - Standard for Safety for the Evaluation of Autonomous Products
"UL 4600 evaluates whether autonomous systems can safely perform their intended functions without human intervention, based on their current state and sensing of their operating environment. The standard also covers the reliability of hardware and software necessary for machine learning, sensing the operating environment, and other safety aspects of autonomy."
In progress
EN 15380 - Railway applications - Designation system for railway vehicles
The standard series shows various identification systems according to which rail vehicle components are structured. This deliverable refers to Part 2 from 2006 (Product groups).
active
LinX4Rail - Conceputal data model
The conceptual data model will be developed in a future Shift2Rail group and cover a data model for all areas of the rail system. It takes place in the S2R project LinX4Rail.
In progress
railML - Railway Markup Language
A data structure for rail transport based on XML. The current version 3.1 was released in February 2019. The standard is licensed under CC-BY-NC-ND.
active
EN 15016-4 - Technical drawings - Railway applications - Part 4: Data exchange; German version
The standard illustrates how data relating to bills of materials and technical drawings must be transferred so that this data is generally understandable. Security during transmission is also examined.
Revision planned
G A 7 7 7 5 1 3 P a g e 17 | 51
EN 17023 - Railway applications – Railway vehicle maintenance – Creation and modification of maintenance plan
The standard shows the processes for creating and modifying maintenance plans for rail vehicles.
active
EULYNX As part of the ERA-funded program EULYNX a uniform standard for modular interlocking technology should be created.
In progress
ITS (Intelligent Transport Systems)
The ITS project run by ISO/TC 204 – WG8 is a standardisation of the information, communication and control systems in the transportation sector (excluding vehicle control systems).
In progress
4.5 Digital data model (cyber-physical model)/ Deliverable 6.5 The steps within the entire CBM process from data acquisition on the asset to deriving rules for
maintenance should apply in the entire rail system. As a result, this paves the way for cross-system
solutions for CBM applications for all assets. The steps were developed in Deliverable 6.5 "Smart
Maintenance System" and based on the necessary steps to provide practical maintenance in future for
all assets.
It should be shown in this deliverable how cross-system solutions can be supported through
standardisation relating to CBM data.
Figure 3 below shows the cyber-physical model for the quadrant "Vehicle monitors itself". This model
is, however, equally applicable to all four quadrants. This process essentially includes the following
steps:
1. Data acquisition on the asset
2. Handover interface
3. Transfer
4. Storage
5. Data analysis with pattern recognition
6. Rules for maintenance
G A 7 7 7 5 1 3 P a g e 18 | 51
Figure 3: Cyber-physical model for Vehicle monitors itself
This cyber-physical model initially shows the data flow for the vehicle asset. It is, however, basically
applicable to all assets in the rail sector in all four quadrants with the aforementioned process steps 1
to 6. This results in the need to develop uniform, harmonised solutions for the technical
implementation of these process steps.
The current situation is, however, characterised by the existence at present of various, custom
solutions for maintenance to cover the subareas of the cyber-physical model. Often these solutions
have been developed directly for a certain use case, making them incompatible with other solutions.
This makes it more difficult to develop an overall system for maintenance, since there are already
major differences even with similar subsystems, such as two different vehicle classes. This issue arose
with the work relating to Deliverable 6.4 and is explained in greater detail there.
4.6 Collaboration with other Shift2Rail partners The deliverable requires close cross-Shift2Rail collaboration with various partner IPs/ TDs.
The collaboration and the key results/statements are summarised in Table 3.
G A 7 7 7 5 1 3 P a g e 19 | 51
Table 3: Contact to other Shift2Rail projects
Ref
ere
nce
to
D
eliv
erab
le/
Dra
ft
- The
mo
st D
eliv
erab
les
of
this
w
ork
pac
kage
- Stan
dar
dis
atio
n
Ro
llin
g D
evel
op
men
t P
lan
D7
.1 O
pen
dat
a: a
rev
iew
of
the
stat
e-o
f-th
e-
art
D7
.2 R
equ
irem
ents
an
alys
is a
nd
Def
init
ion
of
stan
dar
d g
uid
elin
e D
7.3
Op
en s
tan
dar
d in
terf
ace
pro
toty
pe
D7
.4 P
roto
typ
e V
alid
atio
n
- - Del
iver
able
6.2
– C
BM
Dat
a st
ruct
ure
s
- -
Co
nta
ct P
erso
n
Ecka
rt M
eesm
ann
(D
B)
Stef
an T
esar
DB
,
Ro
lan
d K
uh
n B
T
Lau
ren
t Sc
hm
itt
SNC
F, Y
ing
Lösc
hel
DB
Gu
lia V
ign
ola
H
itac
hi,
Mar
io
Mik
usc
hek
D
B
Mar
io
Mik
usc
hek
D
B
Cla
ud
io
Cav
alet
ti
Hit
ach
i
Ro
lan
d K
uh
n B
T
Ren
ato
R
od
rige
s D
B
Stat
emen
ts, r
esu
lts,
inp
ut
for
the
del
iver
able
Del
iver
able
ap
pro
ach
es a
re
sup
po
rted
Su
pp
ort
s d
eliv
erab
le a
pp
roac
hes
. P
rovi
des
pla
tfo
rm, f
or
dat
a
tran
sfer
No
te c
olla
bo
rati
on
wit
h C
DM
! U
niq
ue
list
of
req
uir
emen
ts
Act
ivit
y fr
om
oth
er
stan
dar
dis
atio
n t
op
ics.
No
sim
ilar
app
roac
hes
els
ewh
ere
Sup
po
rts
app
roac
hes
, th
ey u
se
the
sam
e tr
ansf
er p
roto
col a
s in
sect
ion
6.2
Sup
po
rts
del
iver
able
ap
pro
ach
es
Sup
po
rts
del
iver
able
ap
pro
ach
es
Hig
h d
ive
rsit
y o
f th
e e
xist
ing
CB
M d
ata
No
te c
olla
bo
rati
on
wit
h C
DM
! U
niq
ue
list
of
req
uir
emen
ts
Inp
ut
of
the
del
iver
able
to
CD
M
Inh
alt
TD/W
A
Trac
tio
n (
PIN
TA)
Trai
n c
on
tro
l an
d
mo
nit
ori
ng
syst
em
(CO
NN
ECTA
)
Traf
fic
Man
agem
ent
Syst
em
Stan
dar
dis
atio
n
dyn
amic
rai
lway
in
form
atio
n
man
agem
ent
syst
em
dem
on
stra
tor
Frei
ght
Elec
trif
icat
ion
, B
rake
an
d T
ele
mat
ics
Wag
gon
des
ign
Info
rmat
ion
id
enti
fica
tio
n
Inte
grat
ed m
ob
ility
Co
nce
ptu
al d
ata
mo
del
Co
op
erat
ion
TD 1
.1
TD: 1
.2
TD 2
.9
WA
3.2
TD: 3
.6
TD 5
.1
TD 5
.3
TD 6
.6
WA
4.2
CD
M
Essentially all project partners have agreed to the deliverable ideas. Some supplements were also
highlighted. No partner has rejected the content, most of them believe the approaches (see section 6)
are important and agree to the standardisation proposals.
G A 7 7 7 5 1 3 P a g e 20 | 51
The following Figure 4 shows on the basis of the cyber-physical model where interfaces between this
deliverable and the various partner IPs/ TDs were identified as part of the collaboration.
Figu
re 4
: cyb
er p
hys
ical
mo
del
l wit
h in
pu
t fr
om
th
e p
artn
ers
G A 7 7 7 5 1 3 P a g e 21 | 51
5 Selection of Standardisation Areas
5.1 Basic approach to selection As the aforementioned actual state highlights, there is a major need for action to harmonise and
standardise CBM data along the cyber-physical model.
Based on the project requirements in WP6, standardisation regarding CBM data is required in
particular for industry-wide, uniform data semantics and interoperability and the following areas must
at least be investigated as part of this further analysis:
· data flow and format
· frequency of data extraction and transfer
· physical data transmission from vehicles to ground
· on-board rolling stock software and databases
· landside data collection and storage
· access and transfer to CBM visualisation and application systems
Initial information on further processing, data analysis and other data processing is highlighted in this
deliverable. Since, however, there is no working overall system at present, which uses the entire data
flow process, this can only be illustrated here on the basis of examples.
When determining the areas to be standardised, it must also be borne in mind that there are areas in
which standardisation is not worthwhile or possible. This is also illustrated below.
An initial selection of the standardisation areas has already been made as part of the "Scope Definition
WP6 (Deliverable 6.1)" with various experts. To this end, agreement was reached with:
• experts within the working group WP6
• partner working groups within S2R (as per project requirements specification)
• experts outside S2R
As a result of this consultation between experts, it was worked out which of the aforementioned areas
should be standardised for CBM data. At the same time, however, it was also shown in which areas
standardisation is not necessary or even undesirable.
The expert opinions for the individual areas of the CBM data are illustrated below.
G A 7 7 7 5 1 3 P a g e 22 | 51
5.2 Data flow and format
5.2.1 Data flow
The data flow governs the transfer of data from one EDP device to another. The transfer is completed
via various interfaces (see section 5.3) such as USB interface, Wi-Fi or the mobile internet.
The standards regarding these transfers are stipulated outside the rail sector, such as by the
telecommunications companies. Within these transfers, the communication is made between the EDP
devices using transfer protocols; the transfers via the respective interface, however, with the
respective interface protocol.
The customary transfer protocols are HTTPS, MQTT and OPC UA, apart from other protocols. HTTPS is
better suited to transferring small volumes of data for real-time monitoring, the MQTT or OPC UA, by
contrast, can be used more effectively to transfer large volumes of data, to transfer for example, the
data storage unit containing the "historic data" from the day during the vehicle idle times.
The data flow is therefore defined outside the rail sector. The type of transfer is, in turn, dependent
on the use case (mobile data or USB; lots of data at once or only what is absolutely necessary); as a
result there should be no restriction and, in turn, no standardisation to one type of data flow in this
area.
Conclusion: standardisation not desirable
G A 7 7 7 5 1 3 P a g e 23 | 51
5.2.2 Data format
The analysis of the actual situation has shown that there are generally valid formats that are
independent of the rail system (e.g. XML or JSON). These formats are designed outside the rail sector
and, among other things, within the rail sector in wide-ranging ways.
The analyses as part of Deliverable 6.2 did not result in any uniform application of these standards.
Furthermore, it was found in the expert discussions that standardisation in this area is not worthwhile,
since there have hitherto been no problems with the file format.
On the contrary, any standardisation of the available standards would restrict the usability for various
use cases. Moreover, it must be assumed that these standards will continue to be developed very
quickly so that current file formats could be deemed outmoded in a few years.
The analysis has shown that the following formats were predominantly used:
XML: eXtensible Markup Language: a static format which can be adapted simply to the particular use
case. It allows for platform-independent data exchange. XML separates content from its structure and
presentation; as a result, information such as text, graphics and audio recordings can be easily
transferred. [1]
JSON: (Javascript Object Notation): is a text-based, language-independent, human-readable data
interchange format used for representing simple data structures and objects in web browser-based
code. It is easy to parse (analysing of data structures) and generate for machines. JSON is simply a way
to represent data structures, as opposed to a full markup language. JSON documents are relatively
lightweight and are rapidly executed on a web server. [2]
A format is, however, often not defined (see Table 1). If other activities are analysed, for instance in
the CDM group, it can be found that Rail-ML (an XML format modified for rail) is also often represented,
but which is barely used, if at all, as part of maintenance or at those points where data is generated
for maintenance.
The various file formats have so far not caused any problems in data processing, since they were always
programmed openly to date and any other format for further processing the data can be created
quickly using a conversion tool. It is, however, important that the formats do not entail any encrypted
formats, but open formats, as otherwise conversion or simple readability is not ensured. As such, a
standardisation approach is not required within this deliverable. The use of open formats is all that is
important.
Conclusion: standardisation not desirable
G A 7 7 7 5 1 3 P a g e 24 | 51
5.3 Frequency of data extraction and transfer A distinction must be drawn between the frequency of the data extraction and the frequency of the
data transfer. While the data transfer depends on the type of monitoring and the further processing
of the CBM data, the data extraction is dependent on the component part being monitored in each
case. In any case it therefore depends on the relevant use case.
Two examples in this respect:
Monitoring the rotational speed requires a high sampling rate, irrespective of how this information is
further processed, since the wheel may rotate several times a second. Short-term changes in the
rotational speed should also be expected (say through braking or acceleration).
By contrast, the monitoring of room temperature requires a low sampling rate. Even with a high heat
output (e.g. by stabling in the sun or an electric heater) it is unlikely that the room temperature will
fluctuate wildly within a second. Here a sampling rate of a few seconds to a few minutes would be
sufficient.
The frequency of the data transfer depends on the particular user's boundary conditions. These include
the storage capacity on the vehicle, the transfer possibilities during operation, the type of data to be
transferred, the type of analysis and the type of monitoring. With regard to the monitoring type, these
include real-time monitoring, monitoring only in critical moments, i.e. seamless monitoring overall or
partial monitoring.
Since every user must decide for themselves which extraction data they require in the particular use
case, it is neither worthwhile to impose a restriction by standardising the frequency of data extraction,
nor to impose standardisation for the frequency of the data transfer.
Conclusion: standardisation not desirable
G A 7 7 7 5 1 3 P a g e 25 | 51
5.4 Physical data transmission from vehicles to ground In connection with CBM data you need to analyse in a single analysis unit the transfer from vehicles to
the landside in the case of physical data transmission, as well as the transmission of the data of an
infrastructure asset, such as decentralised fitted switch.
Basically there is the option of transferring data wirelessly (e.g. Wi-Fi, mobile network), wired (e.g. via
Ethernet) or using a storage medium (e.g. USB stick or laptop). This, however, depends in turn on the
company and intended purpose. The standards for transferring and the associated transfer protocols
are stipulated by the respective interface operator, e.g. the mobile communications network
companies, and as a result do not need to be standardised at this point.
In the context of the data transfer, however, it is also important to note the provisioning of the data.
At present the situation is such that the central MVB (Main Vehicle Bus) can pick off lots of data. This,
however, does not apply to all data; for instance you can only read out on the MVB whether the doors
have been released in a local transport DB double-decker carriage so the passenger can open them or
not. The information relevant to maintenance, such as door opening and closure times, and number
of the load cycles are only output by the door control unit to the door and are not normally readable
as such. (Bombardier wrote a script for Deliverable 6.4 specifically to read out this data; this script can
then be used to read out this data via the MVB).
It must therefore be determined that there are different sources for CBM data with a different
structure and architecture in the 'vehicle' asset alone.
The necessary centralisation of the data as well as the creation of a uniform data architecture for
vehicles and the provisioning of data on just one standardised interface, is implemented by the project
TD 1.2 CONNECTA and does not need to be pursued further in WP6. This standardisation implemented
through TD 1.2 CONNECTA is a key requirement for the further standardisation of CBM data.
There is currently no known project that defines a central data interface for all four quadrants.
Conclusion: standardisation not desirable
G A 7 7 7 5 1 3 P a g e 26 | 51
5.5 On-board rolling stock software and databases The on-board rolling stock software and databases of a vehicle do not constitute a substantial issue for
maintenance, since these are mainly only created and processed by the vehicle manufacturer. The
vehicle software is usually the property of the manufacturer. It is subject to constant change, especially
if this involves procuring new vehicles (as opposed to older vehicles), these often have state-of-the-art
software.
More important than the on-board software for the user is the interface on which the user is provided
with their data in a suitable form. The underlying structure must be able to supply this data suitably,
depending on how the user would like to access this data. The structure (database) that underlies the
storage and management, and management and storage of data itself is of secondary interest. The
provisioning of the data should in future run over a centralised interface only. This provides simple
access to the data. This is developed, as shown in section 5.3 by TD 1.2 "CONNECTA" and deemed
necessary for future CBM measures.
Conclusion: standardisation not desirable.
G A 7 7 7 5 1 3 P a g e 27 | 51
5.6 Landside data collection and storage Data acquisition, storage and access
To clarify the issue of data storage, you first need to clarify who the data belongs to. The vehicle
manufacturer produces the software that generates the data. The data helps the manufacturer
develop future vehicles and improve current vehicles.
The vehicle keeper generally owns the vehicle or the rights to use this vehicle. The data, which the
vehicle creates, originated based on its use cases associated with its ownership. The keeper requires
this data for maintenance to rectify or prevent problems.
There is currently no general arrangement regarding ownership of the data. It must be regulated
separately each time in the contract or governed centrally in future by law.
If user-based data collection (regardless of whether manufacturer or vehicle keeper) is assumed, a
different format is required for the collection and storage of data depending on the use case. The data
tends to be acquired via sensors that generate an analogue or digital signal. A device then processes
this signal. This data is then either processed or sent for further processing to a suitable device
(automatically or also manually). Depending on the sensor, different data is created during the
acquisition and various processing steps exist depending on the analysis method. As such,
standardisation is not possible here from a current standpoint.
The data tends to be stored on data servers or cloud solutions, but can also be stored on a USB stick
or computer in the individual instance.
Another possible future storage method could be Europe-wide centralised storage. This would mean
that the data would be available to everybody and created by everybody. So it be could be possible for
each manufacturer and user to consult straightaway problems affecting a class to prevent such
problems in future.
Centralised data management for storing and sorting data within a work area then constitutes a
necessary measure. It must also be possible to set up a link for third parties to a storage location, where
necessary with partial anonymisation of data.
These different technical solutions, whether centrally or company-specific or also storage quantity and
transfer mode, obviously depend on the respective use case. Hence standardisation is not worthwhile
at present, since this would result in unnecessary restrictions.
Conclusion: standardisation not desirable
G A 7 7 7 5 1 3 P a g e 28 | 51
5.7 Access and transfer to CBM visualization and application systems
Figure 5: Cyber-physical model added with Support tool and Arrow to it
CBM visualisation and application systems are systems for monitoring individual components of the
asset, maintenance databases and maintenance management tools (e.g. SAP) or also systems for
monitoring the entire cyber-physical model process. This therefore involves maintenance
management systems or support tools.
The information for these systems comes from a storage system (database) where the data is buffered
or also stored long term, or from direct inputs from the Maintenance Operator itself or from analysis
systems.
There are disparate end users that use the maintenance management systems and support tools. Apart
from maintenance, these end users may also be involved in engineering for the assets, in parts
procurement or also in operating the assets. In the maintenance alone, a distinction is essentially
drawn between the end users based on their task definition in the maintenance process. For instance,
the end user "ECM 2" (responsible for maintenance specifications and fleet monitoring) requires
different information from the CBM process than the end user ECM 4 (responsible for maintenance
implementation). These different requirements of end users call for custom visualisation solutions.
According to the particular requirements relating to the particular task in maintenance, the
requirements from the particular company in relation to maintenance and the different maintenance
management tools in the companies, the data is then accessed differently or sent straight to the
systems.
G A 7 7 7 5 1 3 P a g e 29 | 51
To illustrate these tasks accordingly in the cyber-physical model, the model needs to be extended to
include the support tool that indicates the particular information to the particular maintenance
functions and the arrow that shows that information flows from the memory to the maintenance
management system and support tool, see Figure 5.
The different systems in the maintenance functions and the different requirements placed on these
mean that at present it does not make sense to define a uniform standard for access and transfer.
Conclusion: standardisation not desirable
G A 7 7 7 5 1 3 P a g e 30 | 51
5.8 Common data semantics for alignment and interoperability of vehicle condition Basically the processing of data and information, as described at the start, always follows a similar
pattern (see section 4.5). Data is collected, transferred, processed, stored and provided for the user.
This is done at the moment in respective custom solutions so that each solution, regardless of whether
infrastructure or vehicle and regardless of whether from the same manufacturer or multiple
manufacturers, constitutes a standalone solution. As a result, everyone faces the same problem here.
So a solution needs to be found where all work and task areas associated with maintenance-related
areas are also involved, or can get involved:
• TD 1.1 (Traction),
• TD 1.2 (Train control and monitoring system),
• TD 2.9 (Traffic Management System),
• TD 3.6 (Dynamic railway information management system demonstrator),
• WA 4.2 (Integrated mobility),
• TD 5.1 (Freight Electrification, Brake and Telematics),
• TD 5.3 (Wagon design),
• WP6 partners (see appendix "Survey").
The key problems relating to the data are currently associated with access to the data on an interface,
the readability of the data, the assignment of the data to a corresponding component and the
interpretation of the data relating to a condition or problem.
As already described in section 5.4 the necessary centralisation of the data as well as the creation of a
uniform data architecture for vehicles and the provisioning of data on just one standardised interface
is done by the TD 1.2 (Connecta) project. This resolves the problem of uniform access to the data.
The standardisation of the data itself is not dealt with by the TD 1.2 CONNECTA project. This is,
however, essential for the efficient, effective use of the CBM data in maintenance and effects in
particular readability, traceability and assignment to components and errors as well as the
interpretation of data.
This statement is also confirmed by the experience in the subprojects of Deliverable 6.4 CBM results
for rail vehicles.
Conclusion: standardisation of the data structure and the data designation including semantics.
G A 7 7 7 5 1 3 P a g e 31 | 51
6 Standardisation Proposals Regarding Data Structure and
Data Designation
As a result of the analysis of the standardisation areas required in the individual tasks in WP 6 Task 6.7, an actual standardisation requirement exists primarily in the following areas:
- Data structure (structuring of large volumes of data)
- Data designation, content and semantics of the individual data set
The following takes a closer look at this issue and two possible solutions to these problems are illustrated.
6.1 Data structure model
6.1.1 Background and objectives
In the current state, there are numerous different types of data with virtually all assets. This will be
explained below taking the vehicle by way of example. Here, data can be picked off directly by
components or subsystems via the MVB (Main Vehicle Bus). In the case of existing vehicles, this data
is normally not designed or only designed to a small degree for use in maintenance. Apart from the
actual access to the data (technical and legal), above all the understanding of the data and its clear
assignment to individual components, component parts and conditions of the affected assets is
therefore particularly important.
An initial step to solving this problem is a predefined, standardised structure of data with clear
reference to the structure of the considered assets.
Given the particular importance of this issue for vehicles, the structure model to be standardised
(hereinafter referred to as "generic data model") is basically presented taking vehicles by way of
example and explained in detail. Transferring this approach and structure logic to other assets in the
rail system is worthwhile.
Railway undertakings normally operate vehicles from various manufacturers, in different expansion
stages and with different hardware and software versions.
Each vehicle has its own manufacturer-specific codes and signals for conditions and error messages as
well as commands and data structures. In addition, changes are also often made within a class with
different as-delivered versions (e.g. series/ software versions). This makes it difficult to deal with
existing data or to compare it.
To implement data-based maintenance, such as CBM or predictive maintenance, the access to,
allocation of and the understanding of the available data already poses a major problem for the
aforementioned reasons.
The objective must therefore be to create a uniform structure model of the available data. This uniform
data structure should meet the following requirements:
• uniform mapping of the particular vehicle irrespective of the manufacturer
• simple structure for supporting end users (maintenance, engineering and operation)
• consideration of existing structure standards for assets (vehicles)
G A 7 7 7 5 1 3 P a g e 32 | 51
The information in chapter 6.1 are based on the DB Systemtechnik Project for the maintenance of the
DB Class 430. It is currently at the DB work in progress. Contact person is Jan Eichhorn from DB
Systemtechnik.
6.1.2 Basic structure of the "generic data model"
The task of the generic data model is to assign the existing signals to the components, conditions and
mounting positions. This structuring is designed from a user perspective as simply and understandably
as possible for maintenance purposes on the basis of a well-known structure model as per EN 15380-
2.
The vehicle structuring illustrated in EN 15380-2 is used as the basis for the further necessary detailing
of the structure model for use in maintenance. To clearly assign the data and signals, further
subdivision levels with the following information are necessary:
• Further subdivision of the vehicle down to the smallest maintenance unit to be analysed
• Condition description of component part/ function
• Position of the individual component part
The basic structure is shown below in Figure 6
Figure 6: Basic structure of the generic data model
For rail vehicles the product group structure in EN 15380-2 (2006-06) is illustrated. For the remaining
subdivision, a new standard must be drafted.
To be able to precisely assign a product in accordance with the generic data model, subareas must first
be analysed.
6.1.3 Product group structure
The product group structure as per EN 15380-2 forms the existing standard in the generic data model
and so constitutes the central support. The following areas of the data model should form the new
standard. The existing central support and the new standards are again illustrated by There are already
numerous standards and ongoing standardisation activities relating to CBM data. Table 2 below
illustrates selected important standards and standardisation activities. Standardisation approaches
from the results of the work within Shift2Rail are illustrated in WA 3.2.
Table 2.
G A 7 7 7 5 1 3 P a g e 33 | 51
Table 4: Product group structure in the generic data model
EN 15380-2 divides component parts into two different groups to specify them. There are different
main product groups (MPG) to which subproduct groups (SPG) are assigned respectively. In the
aforementioned standard, examples are also assigned to the subproduct group. They are also copied
for the generic data model as sub-subproduct group (SSPG). This is again illustrated below with extracts
from EN 15380-2 in Table 5, Table 6 and Table 7.
Table 5: Extract from the main product groups in EN 15380-2:2006-06
MPG Name of MPG
F Power system, drive unit
L Air conditioning
M Ancillary operating equipment
N Doors, entrances
Table 6: Excerpt from subproduct groups in EN 15380-2:2006-06
MPG MPG Name SPG Unterproduktgruppe Name
M Ancillary operating equipment A Ancillary operating equipment
M Ancillary operating equipment B Sanding equipment
M Ancillary operating equipment C Lubricating equipment
N Doors, entrances A Doors, entrances
N Doors, entrances B External doors
N Doors, entrances D Entrances, steps (not inside)
Table 7: Excerpt from EN 15380-2:2006-06 with examples to which a number is assigned to use them as sub-subproduct group.
MPG SBG Number Example
N A 001 Emergency switch
N A 002 Lock layout plan
N B 001 Connecting door
G A 7 7 7 5 1 3 P a g e 34 | 51
N B 005 Door panelling
N B 011 Entrance door
N B 015 Pivot bearing
N B 018 Sliding door
6.1.4 Function group
Since the product group structure in EN 15380-2 is not specific enough to determine a component part
type unmistakably, the generic data model needs to be supplemented with a functional group. In the
example shown this is done on the basis of the DB ISI catalogue. As part of future standardisation this
component part catalogue must be harmonised and updated centrally.
Table 8: Extract from the ISI component part catalogue
DB has been using the ISI component part catalogue since 2002 as the function catalogue. It contains
a list of general component part designations and categorisation of the component parts, thus
providing the basis for standardising maintenance work and is always kept up to date. An extract is
shown below.
Table 9: Extract from the ISI component part catalogue
Category code
Category name Component part type code
Component part type name
Component part code
Component part name
M M parts general BE Control element ZMD Door actuation inside
E E parts general EC Electrical device ZWS Door blocking control unit
M M parts general VM Fastener M ZWT Door handle plug connection
N M parts special TE Door element ZHD Driver's cab entry door
S Special categories SE System/device FMJ Locking system/unit
F E parts special SL Switch BRV Locking control switch
N M parts special SS Lock DHR External emergency unlocking device
M M parts general BE Control element YQN Emergency release outside
N M parts special SS Lock JJE Pneumatic locking
F E parts special TA Button GPT Stop call button
G A 7 7 7 5 1 3 P a g e 35 | 51
M part = Mechanical component part; N part = Special mechanical component part; E part = Electrical
component part; F part = Special electrical component part; S part = Special category
Apart from DB's ISI catalogue there are also other catalogues that map the component part functions,
such as EN 15380-4:2013-05 "Railway applications – Classification system for railway vehicles – Part 4:
Function groups". Compared with the ISI catalogue this is, however, very general and is not
permanently updated, meaning that newer component parts are not listed.
The product group as per EN 15380-2 and the function group as per the ISI component part group are
required together to develop a code for just one component part type within the generic data model.
This is described below in Table 10.
Table 10: Precise component part type description according to the generic data model, consisting of product group and function group
6.1.5 Component part condition and error description
Apart from the component part, the condition or an existing problem must also be assigned to the
generic data model code. This happens in the next step.
Table 11: Component part condition and error description in the generic data model
The analysis of the error signals, bus signals and standardised damage conditions gave rise at DB to a
condition catalogue, which includes the most common errors. With the aid of the condition catalogue,
damage and damage conditions can be assigned clearly. In this way, damage warnings can already be
G A 7 7 7 5 1 3 P a g e 36 | 51
sent to the maintenance depot. The conditions were already categorised and extended to include an
error description.
They are required in the generic data model to display the condition to Maintenance, in addition to
the specific component part. A few errors and conditions with assigned condition descriptions are
shown below by way of example (see Table 12).
Since each condition code illustrates just one defined condition, it is particularly important here that
the same notation and same assignment of a condition to a code are used on a cross-manufacturer
basis. Otherwise, you would again have huge problems in maintenance if two manufacturers use the
same code for different conditions.
Table 12: Component part conditions as per DB condition catalogue
Signal type Condition code Condition description/ unit
Damage condition 10-004 Deformed
Damage condition 10-005 Flat spots
Damage condition 10-006 NA
Subjective symptoms 11-001 Noise
Subjective symptoms 11-002 Temperature too hot/too cold (damage
warning, subjective feel)
Error message 13-002 Reduced function/no function
Error message 13-004 Does not open/does not close (mechanical)
Error message 13-005 Phase-to-phase short circuit
Error message 13-006 Capacity limit reached (full)
Control command 14-001 Switch on/switch off
Control command 14-002 Couple/separate vehicles
Configuration parameter 15-001 Threshold, maximum
Configuration parameter 15-003 Condition fulfilled yes / no
Functional condition 22-001 Switched on/switched off
Functional condition 22-006 Direction driver's cab 1/driver's cab 2 (FS1/FS2)
Parameter 23-001 Number
Parameter 23-005 Temperature
Parameter 23-006 Speed
External environment parameter 24-001 Air pressure
6.1.6 Location structure
Typically multiple instances of the same assemblies can be installed in a vehicle. As a result, a
component part cannot be clearly assigned across the product group and the function group within a
vehicle. For instance, the DB Class 423 (an electric power car) has 24 external passenger doors, 2 drive
systems, 8 traction motors and 4 roof-mounted air conditioning systems. To resolve this problem, a
location group is also assigned to the "generic data model". This location group enables the
components/ component parts in need of maintenance to be located clearly (Table 13).
G A 7 7 7 5 1 3 P a g e 37 | 51
Table 13: Location group in the generic data model
If you divide the mounting positions into different groups, specific errors can also be found during
maintenance (e.g. failure of the doors on the right side of the vehicle). Therefore it makes sense to
classify the mounting positions by mounting position group (e.g. left side of the vehicle), mounting
position (e.g. side wall) and the specific installation position (e.g. door 22). A few mounting positions
are shown below by way of example.
Table 14: Mounting position group, mounting positions and installation position based on the generic data model
OGR (location group) Location
EPL (installation position) OGR text Mounting position Designation of installation position
2 8 0 Side wall R Right side of veh., door All doors
2 8 10 Side wall R Right side of veh., door Door 1R
2 8 120 Side wall R Right side of veh., door Door 12
2 8 220 Side wall R Right side of veh., door Door 22
2 8 430 Side wall R Right side of veh., door Door 43 (right side of veh.)
3 8 0 Side wall L Left side of veh., door All doors
3 8 10 Side wall L Left side of veh., door Door 1L
3 8 130 Side wall L Left side of veh., door Door 13 (left side of veh.)
Figure 7: Assignment of locations within the generic data model
G A 7 7 7 5 1 3 P a g e 38 | 51
6.1.7 Class-specific codes / examples
With the aid of the aforementioned elements, each component part can now be determined clearly.
Available codes can also be copied to this model so that in Maintenance the condition or the problem
of each component part can be identified quickly. As a result, the company saves time with
troubleshooting and can in some cases already rectify errors early on, which could occur in the
foreseeable future. A component part code is now made up of the product group as per EN 15380-2,
the function group as per the ISI component part catalogue, the component part condition and the
mounting position.
A component part code might look like this:
NB006_NSSCRV_14-022_2-8-0
Figure 8 shows the translation of this error sorted by the respective groups of the generic data model.
Further, other possibilities for the same component part are also shown, in some cases at other
mounting positions.
Figure 8: Class code based on the generic data model
Existing error codes can also be shown in the generic data model. One advantage in this respect is that
available error codes and conditions can be reformatted by an interface. Figure 9 and Figure 10 below
show this.
G A 7 7 7 5 1 3 P a g e 39 | 51
Figure 9: Fault codes Class 430 translated into error codes based on the generic data model
Figure 10: Operating data Class 430 translated into the generic data model
6.1.8 Summary relating to the generic data model
The generic data model provides the option of reflecting exactly all component parts in the form of
data.
The "generic data model" structures the various signals available on the vehicle to a degree of detail
required for the user in maintenance. The model also includes, apart from the familiar structure
elements of the product group structure as per EN 15380-2, information on further component part
subdivisions, information on the condition and on the individual mounting position of the elements to
be analysed.
The structure levels beyond EN 15380-2 must be standardised and subsequently continuously updated
centrally and extended.
G A 7 7 7 5 1 3 P a g e 40 | 51
The "generic data model" was presented here as an example just for vehicles. Transfer of the
structuring logic to all parts of the rail system is recommended.
6.2 Data designation model
6.2.1 Background, problem and objectives
While the data structure model in section 0 initially involves the structural classification of the total
volume of the data, in other words the assignment of the data to the component parts, errors and
events, the necessary standardisation of the designation and notation of the individual data is
analysed below. This is then explained in detail taking an example from the area of infrastructure.
At present there are major differences with the designation of components, component parts and
functions as well as with the designation of the different conditions including error description. To
further standardise the data to promote the use of data-based maintenance and for cross-
manufacturer analysis it is therefore necessary to also conduct harmonisation in this area. Such
standardisation of the designation and notation supports a cross-system analysis and evaluation of the
CBM data irrespective of manufacturer, operator or maintenance operator. Furthermore, this kind of
harmonisation is advantageous for a subsequent automation process.
As the example below shows, different error messages are generated with the same component part
and with the same error at different manufacturers. The maintenance operator receives an error
message, which may look like one of the following examples, depending on the manufacturer:
Manufacturer 1: "Track module fault detected"
Manufacturer 2: "Lamp status 1/2 (Lz1 / Lz2) red (in series) channel de-energised"
As shown in the example, different notations or conditions are defined for similar errors. This
constitutes a substantial obstacle for the necessary data analysis for the purpose of CBM.
Standardisation therefore aims to harmonise data designation and data notation. This harmonisation
must be manufacturer-, operator- and maintenance operator-independent. In addition, it must also
be ensured of course that this harmonisation also remains unaffected in the case of software
modifications.
When standardising the designation and notation of the data, the details necessary for the particular
asset down to the Last Replaceable Unit (LRU) as well as the description of all relevant conditions must
be taken into account.
For the purpose of future predictive maintenance, the extension of the conditions to include suitable
graduations (e.g. "half full" or wear reserve) is also possible.
The information in chapter 6.2 are based on the DB Netze Project DIANA. It is already used by DB.
6.2.2 Basic structure and function
The model for designating the data is explained below using an infrastructure model as an example.
A data set basically consists of four elements:
1. Product type – designation of the asset to be analysed (e.g. switch, signal, barrier system)
2. Product sets – collector for information categories (e.g. manufacturer information, condition
information)
G A 7 7 7 5 1 3 P a g e 41 | 51
3. Attributes - specific designation of the information in the product sets (e.g. lamp faulty,
manufacturer: Siemens, use up to maximum 50°C)
4. Assignments relating to the attribute – collection of other necessary information (e.g. data
type "Double", measurement unit "mm")
To guarantee comparability of similar component parts, the same component parts must be
designated uniformly irrespective of the manufacturer. Irrespective of where the component part was
installed (e.g. a level crossing in Germany or in France). The same applies to the conditions on the
component part. They are always similar irrespective of location and thus comparable (e.g. "level
crossing closed").
Both are information that must be processed in a defined way.
This information is stored in the form of attributes. The annotation of the attributes must always be
the same to aid understanding. For instance you could specify uniformly the attribute "IsName" for the
name of a component part:
IsName = "[Name of the component part]"
After it was stipulated one-time centrally that the component part name is stored under the attribute
"IsName", anyone wanting to give a name to a component part knows that they have to do this under
the attribute "IsName" or also that the name of the component part can be found under this attribute.
In addition to the component part name, there are, however, also other attributes that can be assigned
to a component part. For instance, this could also be the manufacturer name or the serial number.
Example:
Manufacturer = "[component part manufacturer]"
SerialNumber = "[serial number of the component part]"
So that any involved party knows which information must be stated under which attribute, attribute
tables must be maintained centrally.
To be able to better classify these attributes overall, they are summarised in groups, known as data
sets. In this way, all information that is associated with the manufacturer is written in the DeviceSet;
all information that is associated with the component function itself is written to the FunctionSet or
information on subcomponents to ProductGroupSets.
In addition to the clear data notation, in other words the attributes, it is also necessary in IT to assign
a data type to the data. Depending on what kind of data an attribute should include, another data type
is required. Otherwise, the data may under certain circumstances not be readable, analysable or
combinable with each other. As such, an assignment to a corresponding file type is required
immediately when defining the attribute. Depending on the use case, various data types are used. Data
types exist, for instance, for:
• Strings
• True/false logic (Boolean)
• Integers
• Floating-point number with higher accuracy (double)
G A 7 7 7 5 1 3 P a g e 42 | 51
An attribute that only has two states (mostly "true" or "false") includes a different data type than an
attribute that has multiple states, for instance a millilitre-based fill level.
It must also be stipulated which unit is used for which attribute so the unit can be interpreted clearly
during data processing. The unit is not stored within the attribute, but is required as known. The unit
should therefore be noted in an attribute table.
A reservoir for lubricating oil could contain 2,000 ml of oil. This could include, for instance, the
following information as file type "double"
IsOil = "2,000"
If another party involved has programmed their software to litres because they would like to work with
decimal places, they would understand here that the reservoir contains 2,000 l of oil.
During the further processing, a limit value of say 500 ml could be stated, which is constantly checked
and issues a warning as soon as this value is not reached:
If IsOil < "500" then WarningOil = "true"
If the attribute "IsOil" were to be stored in an unsuitable data type for this information (e.g. string),
the system could not process the data in the way shown above, since the data type string is understood
as a string of characters.
Conversely, for the aforementioned example "IsName" would not work as "double", but only as string,
since a string cannot be stored in the format "double".
In addition to the aforementioned comments, there are other IT reasons for assigning data types,
which are, however, not decisive for the data model used here and which will not be described further
at this point.
6.2.3 Examples of uniform data designation
The introductory example from section 6.2.1 is, as described, very difficult to process using electronic
data processing since the faults from the two sample manufacturers look different.
Manufacturer 1: "Track module fault detected"
Manufacturer 2: "Lamp status 1/2 (Lz1 / Lz2) red (in series) channel de-energised"
Adapted to the described data model with the clear data designation, you get a classification which is
understandable on a cross-manufacturer basis. A brief description of the example is provided in Table
15 below:
• ProductGroup<Signal>: Lamp status 2 (Lz2)
o ProductGroup<LightPoint>: red
▪ LampStatus: Failed
Table 15: Detailed explanation of the example for data designation
Type of assignment Name of assignment Assignment content Description
G A 7 7 7 5 1 3 P a g e 43 | 51
Product ProductGroup <signal>: Lamp status 2 (Lz2)
Assignment to the component part
Set ProductGroup <LightPoint>: red Assignment to the component of the component part
Attribute LampStatus Failed Status of the component part component
So that each involved party knows which signals and information are assigned to which attributes and
sets, an attribute table needs to be maintained centrally for each of the individual products. This must
be accessible to everyone involved. The attribute table must state at least the set with the respective
attributes, their importance, the data type of the attributes and the particular unit which is assigned
to the attribute. Three relevant examples follow.
G A 7 7 7 5 1 3 P a g e 44 | 51
Table 16: The data set "Environment" of the product "Dust sensor" contains among other things the attribute "DustLevelConcentration" with the data type Float with the unit "g / m³".
Name Meaning Data typ Unit
2.2.1.1 DustlevelConcentration Dust Concentration for dust sensor models that measure concentration. The data type is only available if the dust sensor has been created with the parameter "DustSensorMeasurementType.Concentration".
Float g/m³
2.2.1.2 DustLevelPPM Dust concentration for dust sensor models that measure the amount of dust (particles). The data type is only available if the dust sensor has been created with the parameter "DustSensorMeasurementType.PPM".
Float ppm
Table 17: A sensor of the product "Soft" contains the attribute "IsSensorFailure" with the data type Boolean - Without unit in the set "Error message" because only "True" or "False" is returned.
Name Meaning Data typ Unit
3.5.2.1 IsSensorFailure True: The sensor for measuring the direction of rotation is disturbed. For example, 0 mA for a 4mA-20mA current interface
Boolean -
Table 18: The product "temperature sensor" has among other things the attributes of the "maximum and minimum room temperature" in the unit "Kelvin" in the data type Float in the data set "Parameter".
Name Meaning Data typ Unit
2.2.2.1 pTemperatureRoomMaxFailure
Maximum room temperature Float K
2.2.2.2 pTemperatureRoomMinFailure
Minimum room temperature Float K
2.2.2.3 pPortId PortId of the input to which the sensor is connected. A list of all input PortIDs must be provided to the operator in advance. The identifier of the Input PorIds from the list must be physically attached by means of stickers or signs above the connector.
String -
6.2.4 Summary of the data annotation
To process CBM data uniformly in future in the rail sector, the content of the data must be annotated
uniformly and follow a common structure.
The structure and designation the data must be controlled centrally to be able to respond regularly to
changes. Everyone involved must follow and comply with this structure. Changes should only be
organised from the central entity so that here too all involved companies can implement the same
changes.
G A 7 7 7 5 1 3 P a g e 45 | 51
In addition to the centrally specified data structure and designation, it must be shown for which data
this structure is valid and how these features must be handled.
6.3 Benefit of standardisation for CBM usage in the rail sector Maintenance of vehicles and infrastructure still tends to be preventive today. This means component
parts are swapped out regardless of their condition. In some cases, perfectly intact component parts
are removed. The replacement intervals tended to be derived from experience of operating these
vehicles or similar classes.
Unexpected component failures are only identified once these failures occur and cannot be predicted
or are very difficult to predict. This requires an immediate operational response and disrupts
operations.
This should be changed in future. Through a uniform structure of data and its classification, the user
should be able at any time to find out the current condition of a vehicle.
By standardising the data structure and data designation it will be possible immediately in future to
include a new class or a new component of the infrastructure into the maintenance system and
analysis system.
In this way, most vehicle or infrastructure problems will be able to be detected in future by means of
continuous monitoring (tailored to the particular user) before such problems occur. Further
information in this respect is also included in Deliverable 6.5.
Automatic identification will allow conventional maintenance tasks to be shortened and, in some
cases, even omitted entirely. This reduces LCC costs of a vehicle or an infrastructure system
substantially in some cases, reducing in turn the expenditure of the companies.
G A 7 7 7 5 1 3 P a g e 46 | 51
7 Conclusions and data requirements for CBM As part of the work on this deliverable, it was found that standardisation of the following elements is
not worthwhile based on the insights from Deliverable 6.2:
• Data format
• Transfer protocols
• Data and sampling frequency
• On-board software for rail vehicles
• The type of physical data transfer from the vehicle to the landside
• The type of data storage
• CBM visualisation systems and data transfer to these
Within the working group WP6 it was found and confirmed by all participants that there are major
problems with data processing, which can be rectified by standardising the following elements:
• Single data structure, by classifying the data into a model that clearly reflects the vehicle
structure and the function of the component parts and can be assigned clearly to just one
component part on the basis of the data name.
• Unique data designation, which is structured so that a component part, which is used by
multiple manufacturers or operators, always gets the same designation, as well as the
designation of the conditions so that a condition that a component part can adopt, always has
the same conditions irrespective of the manufacturer and operator.
• Data classification that specifies precisely in what format data must be classified with which
subgroupings.
• Table of data meaning that shows in which unit the data has (e.g. liter or milliliters) and
explains its exact meaning.
• Data interface where all CBM data can be picked off so that it is no longer necessary to build
interfaces at multiple locations. Irrespective of the interface, the data can then be further
processed.
• Clarification of the legal situation regarding who the data belongs to.
• Centralised management of the data that updates the specifications for structure, designation
and classification of the data.
These insights serve as input for the smart maintenance concept in Deliverable 6.5 in relation to the
Guideline for Manufacturers/ Operators/ Maintenance Operators. They are equally interesting for the
partner TDs since similar problems were found here.
G A 7 7 7 5 1 3 P a g e 47 | 51
8 Future Development In addition to the illustrated need for standardisation, the following areas in particular will need to be analysed in greater detail and regulated on an overarching basis in future
• Quality/ reliability of data (including necessary continuity of the data provisioning and transfer, as well as the data truth – Deliverable 6.4)
• Data security (protection against tampering, hacking or similar)
• Data safety (reliability – Deliverable 6.2)
• Security plan for special events (data loss, wireless dead zone, transfer interruption, ethical issue – Who decides whether a component part is bad, and an accident happens?)
• Data management (centrally/remotely) and accessibility to the data with various access options
G A 7 7 7 5 1 3 P a g e 48 | 51
9 Further steps towards standardization
The standardisation approaches of deliverable 6.3 are incorporated into the summary of the standardisation approaches S2R (tabular overview WP5 – Deliverable). They are also submitted as input for the CDM group (Shift2Rail project LinX4Rail) to provide guidance here for the joint data model.
The overall list of standardisation WP5 must be agreed between Joint Undertaking and European/international standardisation committees (CEN/ CENELEC) to place the standardisation projects accordingly. There a decision will be taken in a further step regarding which committee is commissioned with the further development regarding standardisation of the CBM data.
For the standardisation approaches devised as part of this deliverable, various additional steps for further development are conceivable. These steps may also be performed in parallel. These are:
1. ISO TC 204 – WG8 to further process the topic to devise ISO standardisation
2. CEN/ TC 256 / WG48 or the IEC/TC 9/WG 43 to further process the topic to devise standardisation for rail vehicle maintenance
3. CEN/ TC 256-SC1 to further process the topic as part of devising standardisation for rail infrastructure-infrastructure - track
4. CLC/ SC 9XA to further process the topic to devise signal and data exchange in the infrastructure
5. New standard by CEN/ TC 256 / WG48 to standardise the data
6. CEN/ TC 256 / WG19 to undertake revision to provide a greater level of detail of 15380-2
7. Creation and development of an EuroSpec for data standardisation
G A 7 7 7 5 1 3 P a g e 49 | 51
10 References
No References
G A 7 7 7 5 1 3 P a g e 50 | 51
11 Annexes
No Annexes
G A 7 7 7 5 1 3 P a g e 51 | 51
12 Antitrust statement
While some activities among competitors are both legal and beneficial to the industry, group activities of competitors are inherently suspect under the antitrust/ competition laws of the countries in which our companies do business. Agreements between or among competitors need not be formal to raise questions under antitrust laws. They may include any kind of understanding, formal or informal, secretive or public, under which each of the participants can reasonably expect that another will follow a particular course of action or conduct. Each of the participants in this initiative is responsible for seeing that topics which may give an appearance of an agreement that would violate the antitrust laws are not discussed. It is the responsibility of each participant in the first instance to avoid raising improper subjects for discussion, notably such as those identified below. It is the sole purpose of any meeting of this initiative to provide a forum for expression of various points of view on topics
▪ (i) that are strictly related to the purpose or the execution of the initiative,
▪ (ii) that need to be discussed among the participants of the initiative,
▪ (iii) that are duly mentioned in the agenda of this meeting and
▪ (iv) that are extensively described in the minutes of the meeting.
Participants are strongly encouraged to adhere to the agenda. Under no circumstances shall this meeting be used as a means for competing companies to reach any understanding, expressed or implied, which restricts or tends to restrict competition, or in any way impairs or tends to impair the ability of members to exercise independent business judgment regarding matters affecting competition. As a general rule, participants may not exchange any information about any business secret of their respective companies. In particular, participants must avoid any agreement or exchange of information on topics on the following non-exhaustive list:
1. Prices, including calculation methodologies, surcharges, fees, rebates, conditions, freight rates, marketing terms, and pricing policies in general;
2. any kind of market allocation, such as the allocation of territories, routes, product markets, customers, suppliers, and tenders;
3. production planning; marketing or investment plans; capacities; levels of production or sales; customer base; customer relationships; margins; costs in general; product development; specific R&D projects;
4. standards setting (when its purpose is to limit the availability and selection of products, limit competition, restrict entry into an industry, inhibit innovation or inhibit the ability of competitors to compete);
5. codes of ethics administered in a way that could inhibit or restrict competition; 6. group boycotts; 7. validity of patents; 8. ongoing litigations.