Tunnels and Underground Cities: Engineering and Innovation meet Archaeology,
Architecture and Art – Peila, Viggiani & Celestino (Eds)
© 2019 Taylor & Francis Group, London, ISBN 978-1-138-38865-9
Building information management for tunneling
F. Robert & A. Rallu
CETU (Tunnel Study Centre), Bron, France
C. Dumoulin
French national research project MINnD, Paris, France
N. Delrieu
ANDRA (French National Radioactive Waste Management Agency), Châtenay-Malabry, France
M. Rives
Vianova Systems, Boulogne-Billancourt, France
M. Beaufils
BRGM (French Geological Survey), Orléans, France
ABSTRACT: MINnD (Modeling Interoperable Information for sustainable Infrastructures)
is a French national research project dedicated to BIM for infrastructures. Many working
groups are in progress, treating roadways, bridges, environment, etc. On the initiative of
ANDRA, the French National Radioactive Waste Management Agency, a major effort has
been undertaken for underground works. In 2017, a first step was to build the Data Diction-
ary, proposing an organic decomposition of the systems and subsystems, as well as the neces-
sary data flow between the different stakeholders (Information Delivery Manual). This paper
presents the main results of this work. This concerns the definition of relevant data concerning
the geometry, geology, hydrogeology, geotechnics, design, construction of the structure itself,
including ventilation, electrical and management equipment.
1 INTRODUCTION
The French national research project MINnD (French acronym meaning Modeling Interoper-
able Information for sustainable Infrastructures) was born from the following observation:
the digital model that was developed for industrial products (cars, aircrafts, etc.) that have the
distinction of being modeled independently their environment, is not satisfactory for infra-
structures that interact with the ground, and whose construction phases must be able to be
modeled (earthworks, pouring, construction tools, etc.). If things are changing for the build-
ing, the MINnD national research project was launched to deal with the problem of infrastruc-
tures. Several use cases have been identified. For example, use case No. 3 “Bridge” for
bridges, was developed in MINnD.
The use case No. 8 relating to the underground infrastructures is emblematic of the inter-
action between the structure and its environment insofar as the structure is in permanent inter-
action with it (soil, networks, and neighboring works) as well during construction phase to the
operation phase. In 2017, under the leadership of ANDRA (French National Radioactive
Waste Management Agency), an important work was led to work on the modeling of data
relating to underground structures. This work was carried out on two levels: on the one hand
the modeling of the civil engineering structure and its equipment (led by the working group
named GC – cf. Section 5) and on the other hand the modeling of the environment of the
2950
structure (led by the working group named GT – cf. Section 6), namely the soil and its various
components. This paper presents the main results achieved and future prospects.
2 MINND PROJECT
MINnD is a French national research project involving 71 partners and gathering stakeholders
of the construction sector including owners, engineering companies, contractors, industrial
companies, public and private laboratories, universities, Grandes Ecoles. Cf. www.minnd.fr.
Controlling and sharing information are key issues for the construction industry which
must nowadays cope with major changes in its activity, such as project complexity, eco-design
development, new types of partnership between actors (Public Private Partnerships, Conces-
sions), obligation to manage risks (anticipation, evaluation, distribution) or development of
Building Information Modeling (BIM). Drawings, notes and records, files have shown their
limits. Moving to information modeling applied to infrastructure is already in progress. The
first challenge is therefore to move to the item that is the finest information by establishing a
structure and a recognized internationally standard for information exchange as well as to use
adapted tools, either transversal, such as digital models, or specialized and developed intern-
ally by each actor.
The digital mockup of a manufactured product focuses on a “new” product. Regarding
buildings, the BIM describes only the building itself, often neglecting the surrounding environ-
ment. Foundations and connections to the external networks are modeled, but the surround-
ing ground is often omitted, or described only at its final stage. The focus is clearly put on the
definition of the delivered product, mainly on the design. The construction steps including
stop pouring concrete, the necessary tools, logistics and temporary elements are not modeled.
Regarding tunnels, the focus cannot be limited to the supporting structure and functional
systems. On one hand, the geometric design of the supported road or railway is a key driver of
the tunnel design. And on the other hand, the surrounding ground has to be taken into
account as well as the impacted and impacting environment.
Some considerations imposed by the road must be properly addressed during the design
process, including design speed, design traffic volume, number of lanes, level of service, sight
distance, alignment, super-elevation and grades, cross section, lane width, horizontal and ver-
tical clearance. These considerations are also mandatory for bridge design. MINnD is cur-
rently finishing a research work regarding BIM applied to bridges. The results of this research
have been shared with the Infrastructure Room of buildingSMART International (bSI) and
will contribute to the extension of the international standard “ISO 16739:2013 Industry Foun-
dation Classes (IFC) for data sharing in the construction and facility management industries”
for application to bridges. This work will provide a good input for supporting road geometry
applied to tunneling.
What is really new regarding tunneling is geo-modeling Instead of designing what has to be
built and how to do it, only observations can provide some information locally about what is
in place and interpretations try to fill the gap and imagine the missing information. When the
tunnel is bored or the ground cut, the characteristics of the extracted ground are finally
known, but only at the end when analyzing the muck. But the influence of the surrounding
environment and the impact of the tunnel construction on the environment will remain
estimations and therefore induce risk management.
Except the geotechnical aspect, a tunnel looks like a “long” building following the align-
ment of the hosted road or railway. Several functional systems describe the operational sys-
tems, such as drainage, ventilation, electricity. . . But risk management, construction
management, ground boring including replacement of the ground in place by the built struc-
ture, temporary works, logistics require a 4D modeling, which means the capability of model-
ing the progress of the construction versus time.
Finally, Building Information Modeling should offer a 3D “architectural” model defining
the tunnel project as it will be delivered to the client, including the tunnel structure, the oper-
ational systems, the environment (ground, terrain, external networks) modified by the project.
2951
This model will be associated to a 4D model able to manage the construction sequences, and
to a structural analysis model. In addition, all the knowledge related to the existing environ-
ment that will not be modified directly by the project but could impact or be impacted by the
project will be described in a geo-model. Links between the geo-model and the project models
will be established to manage the changes that can occur in the different models during the
project lifecycle.
All this activity is developed in a specific sub-project called MINnD UC8 (Use Case n° 8)
Tunnel.
3 THE ORIGIN OF THE USE CASE N°8 DEDICATED TO UNDERGROUND
INFRASTRUCTURES
3.1 Context
ANDRA (French National Radioactive Waste Management Agency) is a public industrial
and commercial establishment (EPIC), in charge of long-term radioactive waste management
operations.
In order to manage the ultimate radioactive waste that cannot be stored on the surface or
on sub-surface (at a shallow depth), ANDRA is designing a deep geological underground
radioactive waste storage facility: Geologic Industrial Storage Center (Cigéo). This project
combines all difficulties related to data management and collaborative work of complex pro-
jects, such as (i) industrial installations, (ii) large infrastructures, (iii) complex underground
work and (iv) basic nuclear installations. This kind of project involves a large panel of profes-
sions, each of them using different methods of communications. So data exchange can be
hard.
Consequently, ANDRA identified quickly the challenge of a single model shared database
and making easier collaborative work and data management. This latter must be accessible by
all stakeholders and during all phases of the life cycle of a work via an interoperable format.
Figure 1. Life cycle of a building project (source: buildipedia.com).
2952
3.2 PLM and BIM for information sharing and data management
This project involves many works of different nature:
• large-scale ground-level infrastructure integrated in its surrounding environment;
• complex underground structures with shafts and galleries;
• industrial equipment’s and facilities;
• multiple working and living quarters for several hundred people;
• operating sites.
Indeed, a key point is to share the information between the various stakeholders throughout
the project lifecycle, which is designed for secular times. The challenges in terms of coherence
and technical, geometric and geographical coordination are numerous.
A tunnel project is therefore part of an approach of more than a century of multi-sector,
multi-stakeholder cooperation and co-activity, working in parallel across various stages of the
project: (i) design, (ii) construction, (iii) operation and maintenance, and (iv) post-closure
monitoring.
This therefore represents a challenge in terms of secure data management and traceability,
and one that ANDRA is anticipating through the use of two data management tools that are
tried and tested in both industry and construction: PLM and BIM.
A first tool is essential to manage in real time the integral memory of the project: the Prod-
uct Lifecycle Management (PLM), i.e. the management of the life cycle of the infrastructure
project. It supports, structures and archives all project information in a secure database. This
includes a digital biography that incorporates supplementary data to BIM (knowledge of
waste producers’ packages, operating history, procedures, etc.) and which facilitates the moni-
toring, traceability, exchange and archiving of any information. PLM is configured to provide
tailored solutions to the specific needs of each stakeholder through industry-oriented views.
Responsible for all technical data, PLM also supplies the corporate information system.
The second data and project management tool is BIM: Building Information Model/Model-
ling/Management, all in one. This is a collaborative working method for generating and work-
ing with construction data across the project lifecycle. BIM draws on information exchanged
via multi-sector and multi-stakeholder digital models. It is shared by design offices, technical
staff, service providers and operators, who all interact around this same dynamic representa-
tion and model. Each change is reflected throughout the overall model, which enables an
accurate assessment of the future design and makes it possible to carry out simulations. BIM
thus provides a shared vision of the project and allows access to the basic data for each com-
ponent: geometry, materials, mass, costs, timeframes, suppliers, etc.
For his infrastructure project, ANDRA wants to meet the challenge of marrying
• BIM, which manages 3D geometry and basic information for the project’s components;
• and PLM, which organizes and combines this information with the records of all project
data.
It is a digital innovation that combines the benefits of the two data management tools and
presents the complete information of PLM in BIM’s highly visual digital model. This synergy
has to be able to ensure more streamlined collaboration between the various sectors, driving
operational performance. To achieve this goal, ANDRA’s teams work with the MINnD UC8
working group to identify and harmonize the standards and requirements of each stakeholder
to integrate them into this new collaborative interface, able to evolve over time.
4 BIM STANDARDIZATION PANORAMA, MINND POSITIONING
4.1 Standardization through IFC
All what is created or modified by the project has to be defined in a 3D “architectural” model
for approval by all project stakeholders. The model description could be based on the inter-
national standard “ISO 16739:2013 Industry Foundation Classes (IFC) for data sharing in the
2953
construction and facility management industries” including the infrastructure extension fore-
seen by buildingSMART International. Standardized description of the environment is also
critical for infrastructure design, especially for engineering structure such as tunnel. This will
include the terrain model and the ground layers included in the project limits.
To cover this need, one approach is to extend IFC and define new classes. The China Rail-
way BIM Alliance, for example, proposes definition of concepts such as IfcRockSoilMass, Ifc-
DrillHole and IfcDrillHoleLayer to respectively describe geological units, boreholes, and
layers identified in the geology core. The project has to manage and track the construction of
the tunnel, and therefore the associated earthworks and tunnel boring activities.
But the aim is not to address geosciences features, but to identify existing standards outside
of the BIM/IFC community and define links with them. For a given project, geosciences have
to cover a larger space than the project limits in order to identify the ground characteristics
related to the tunnel.
4.2 Alternatives to IFC
In terms of standardization of Geographic Information Systems (GIS) data in general, several
relevant initiatives and standards can be mentioned.
In 2007, INSPIRE, the European Directive introduce several themes and models for data
that should be shared in the European Community. Among those topics, most relevant to
mention are Building (BU), Geology (GE), Soil (SO), Environmental Monitoring Facility
(EF), Natural Risk Zones (NZ).
Most models from INSPIRE rely on standards developed by the Open Geospatial Consor-
tium (OGC), an international non for profit organization dedicated to support and develop-
ment of standards for geographical information.
One regularly mentioned standard from OGC for bridging BIM and GIS is CityGML that
aims at describing city models. Reciprocal conversion from IFC to CityGML is a topic
addressed by several working groups, especially one dedicated joint working group between
bSI and OGC named Integrated Data Built Environment Domain Working Group
(IDBE DWG).
Concerning geology, OGC launched in 2017 a Domain Working Group (DWG) to cover
this topic: the GeoScienceDWG. Main objective of this group is to harmonize practices, sup-
port and enhance existing standards in that domain inside and outside OGC. This task par-
ticularly influences the development and refinements of OGC geoscience standards
GeoSciML and GroundWaterML2, which have been designed to describe both geological and
hydrogeological features, introducing concepts such as GeologicUnit, Faults, Boreholes,
HydrogeoUnits, Voids. . . but also their associated properties such as EarthMaterial. Those
standards have been designed by geoscientists, and are widely implemented by providers of
this kind of data, especially geological surveys.
5 MAIN RESULTS OF GC WORKING GROUP
The GC (Civil Engineering and Equipments) working group of the MINnD-UC8 National
Project was set up to gather the required expertise in order to be able to address the various
domains that compose an underground infrastructure, being its civil works parts or its
equipments.
This comprehensive knowledge was carried out, after a formal request for proposal process,
with more than 25 experts from underground infrastructure owners, consultants and contrac-
tors, as well as with IFC experts; all with a solid experience of underground infrastructure
projects, in France and internationally.
The goal of the GC working group is to provide specifications to the independent bSI
organization for the production of extensions to the current IFC4.x (x being the latest version)
format in order to keep fruitful exchange mechanisms of design/build/operate information
2954
representing underground infrastructures data; such information being organic, functional or
spatial related.
Ultimately, such IFC extensions for underground infrastructures are to be proposed to the
ISO organization for deployment as internationally applicable industry standards.
As a first step, the GC working group scoped what underground infrastructures might be:
tubes, shafts and storing cells, with a view to produce a breakdown of all of their civil works
structures and equipments, their interdependences (relationships) and their characteristics
(properties).
The approach used to conduct this work followed a ‘why’/’what’/’how’ systemic analysis to
outline an exhaustive perspective of the programmatic/functional/organic aspects of the infra-
structures concerned. This led us to identify a series of sub-systems, as featured below:
Figure 2. Decomposition into sub-systems.
In parallel, we focused on where data exchanges take place in a design process (between
industry domains experts/at what phases), in a construction process (including between detailed
design experts and constructive methods experts) as well as in an operate/maintain context.
As a second step, GC working group started an analysis of the possible implementation of the
new objects classes (in IFC terms) needed to provide the functional, organic and spatial represen-
tations of the various components of the sub-systems, their relationships and their properties.
This work required first to leverage the existing features classes and their hierarchies avail-
able in IFC4.1, and second to propose an enrichment of these, while aiming at introducing as
minimum extra complexity as possible into the existing IFC conceptual model.
Then, the GC working group will develop a proposition for an extended conceptual model
and the corresponding IFC schema (based on the IFC4.1), with a view to provide definitions
for the IFC5 development roadmap, in capitalizing the IFC Alignment 1.1 and IFC Overall
Architecture projects led by bSI. Ultimately, this work shall help the bSI organization in:
Figure 3. IFC Tunnel conceptual model (simplified view).
2955
• The development of a semantic description of underground infrastructures in a language
implementing the concepts and logic used by underground construction experts;
• The set up and use a common data dictionary for underground infrastructures;
• The development of subsoil modeling (based upon industry practices and input
from OGC);
• The resolution of the appropriate objects breakdown into the global IFC schema.
This initiative will also be fed by other national projects shooting for open, international
standards and identify use-cases for underground infrastructures.
6 MAIN RESULTS OF GT WORKING GROUP
Environmental modeling follows different rules than Building Information Modeling. While
BIM aims at defining the system to build and the actions to do to have it, geomodeling is
about exposing what has been observed and possible interpretation that can be built from
those observations. In that context, geomodeling is always associated to uncertainties. Then
to cope with them, the project coordinator relies on a risk driven management approach,
focusing on the risk of delay and extra cost that each possible interpretation could lead to.
Based on that method, the way that geotechnical standardization is addressed in that pro-
ject does not only consist in providing standards to describe each geoscience feature, but also
in being able to capture the information lineage, thus retrieve the observations and assump-
tions that have been used to build the geomodels and associated interpretations. For this
reason, the first action of the group was to map how geotechnical knowledge is built, and
describe the path that is followed to build the interpretation and risk assessment.
6.1 Definition of geotechnical work
In France, the geotechnical activities are described by the standard NF P 94–500 from the
French Association of Normalization (AFNOR). This standard defines the goals that should
be reached at each time of the project in several domains (geology, hydrogeology. . .). In add-
ition, the French Tunneling and Underground Space Association (AFTES) also propose
guidelines for interpretation and illustration of the standard to facilitate its application
(GT43R1F1).
Lecture and analysis of that standard lead to the proposition of dividing geotechnical activ-
ity in ten main topics: RECO deals with the collection of observations, measures and informa-
tion on the field, test, laboratory analysis and survey. GEOL, HYDR and GTCH respectively
deals with the creation of geological, hydrogeological and geotechnical model, including the
research of coherence between them. CALC and MECO respectively deals with infrastructure
sizing, determination of interaction with the soil, definition of the geotechnical influenced
zone (ZIG) and the proposition of appropriate constructions. ENVI and AVOI respectively
discuss the management of the impacts on both environment and built areas. Finally, RISK is
about the assessment and management of risks and uncertainties associated to the data and
proposed models and MRCH addresses the collection of useful data and deliverable that shall
be available for tenders.
6.2 Information Delivery Manual
For this project, ten Information Delivery Manuals (IDM) were developed, one per topic.
Each IDM consists in a process map and a glossary. The process map provides a graphical
representation of the sequence of activities in the topic and the involved data. The glossary
acts as a legend for the process map, and enable to provide more detailed description of both
data and actions.
Process maps from UC8-GT all have the same design: in a middle part, actors and processes
sequence. On the left, data used in the processes. On the right, data updated or created by the
2956
processes. Each process and datum have an identifier in order to ensure that same data does
not have different names in different IDMs and facilitate links between IDM (eg. CALC.D3
stands for the datum “D3” of the “CALC” IDM). Such IDM design also enables to identify
easily the key concepts and properties to consider in the Conceptual Model.
6.3 Conceptual Model inspiration and first propositions
The IDM clearly highlight that geotechnical activities consist in a chain of tasks including col-
lection of information and formulation of hypothesis that once linked, enable to model the
subsurface and determine important information for construction. Discussions in the group
also emphasize that results from modeling can be very different from a project to another.
The context of the project often constrains and centers the modeling tasks, and necessitates
use of specific methods and tools. This lead to a wide range of possible model outputs: from
3D voxels to 2D cross-sections for geometrical modeling, to more numerous and original
results for analytical modeling.
Instead of trying to provide standard description for all of those possible outputs, the
MINnD UC8-GT decided to focus on two objectives: 1- describing and associating properties
to identified “components” of the subsurface, 2 – describing the geotechnical processes activ-
ities and context to facilitate data reuse.
Regarding the first objective, OGC standards GeoSciML and GroundWaterML2 already
introduce concepts to describe components of the subsurface. This includes geologic units, geo-
logic structures, aquifers, water bodies, voids. As several international organizations, especially
geological surveys adopted them and ensure the maintenance of those standards, relying on them,
in opposition to extending IFC for example, appears to be a sustainable and long term solution.
Regarding the second objective, the MINnD UC8-GT focuses on the ISO: 19156 standards,
also known as OGC Observation & Measurements. Basic purpose of this standard is to facili-
tate exchange of information describing observation acts and their results. The key concept is
called “Observation” that aims at determining the value of a property (cf. Figure 4). The
observation result – the value – describes a phenomenon or property of a feature, the feature-
of-interest of the observation. Observations can be realized by an instrument or a sensor, but
also by a process chain, human observer, an algorithm, a computation or a simulator. Obser-
vation properties provide context or metadata to support evaluation, interpretation and use of
the observation results. Finally observation can be linked to another through the relatedOb-
servation association.
In that context, Observations & Measurements (O&M) can be used both for describing acts
relative to data collection (observed by humans or sensors), modeling (human interpretation
Figure 4. Observation schema from ISO19156/OGC Observations & Measurements standard.
2957
assisted by software), but could also by extension be used for preconisation (proposition
based on model run). The relatedObservation association can be used for lineage and data
discovery from results. Finally, the use of such standard would also facilitate data provision
and access, as OGC also proposes several interfaces to handle O&M structured data: e.g.
Sensor Observation Service (SOS) and SensorThings API. Model results and associated data
would then be retrieval through standardized protocols.
7 BUSINESS APPROACH, CONSIDERATION OF AFTES’S GUIDELINES
The work to be carried out had to take into account the rules of the art and the uses recom-
mended by the profession of the underground works. This is why the CETU, as a government
institution, joined the partners of the MINnD project in trying to bring its neutral vision of
things and by eliciting the AFTES (The French Tunneling and Underground Space Associ-
ation, member of ITA) recommendations in the process of data exchange between the differ-
ent actors, and the formalism of the usual deliverables.
The French Tunneling and Underground Space Association (AFTES) aims to unite and
mobilize all players in the profession: investors, project owners, project managers, design
firms and inspection bodies, research and training institutes, contractors, consultants, archi-
tects, town planners, equipment manufacturers, academics. . .
Thus, for the constitution of the data dictionaries and the organic decomposition of the
constituent elements of an underground infrastructure, the bilingual French-English diction-
ary, used in the framework of the CETU international actions (PIARC, ITA) was used to dis-
pose of an unambiguous basis for working directly in English in order to have a bilingual
taxonomy of systems and subsystems.
From this organic decomposition into systems and subsystems, the information delivery
manuals (IDM) were produced using specific schemes from the AFTES recommendation
GT32R2F1 dealing with the characterization of geological, hydrogeological and geotechnical
uncertainties and risks, and the recommendation GT32R3F1 dealing with the consideration
of technical risks in projects. Moreover, some parts of the Information Document on the Tun-
nels Prices, published by CETU in 2016, were used in order to structure some IDM according
to a representative decomposition in underground works.
Figure 5. Risk Management Methodology Summary Flowchart (AFTES recommendation GT32R2F1).
2958
These recommendations describe the iterative process of risk management to be implemented
during the different stages of the design process (cf. Figure 5), as well as the documents to be
produced (the deliverables) for tendering. We can quote the following deliverables:
• Book A1, which includes geotechnical input raw data;
• Book A2, which contains input raw data for nearby buildings;
• Book A3, which contains input raw data for the natural and human environment;
• Book B1, which is the geotechnical summary report;
• Book B2, which is the sensitivity report of neighboring works;
• Book B3, which contains comments on environmental constraints;
• Book C, which is the design report;
• Notice of respect of the environment;
• Risk management plan and the residual risk register.
Thus, the different deliverables, their descriptive content and the flow of information between
the various project stakeholders were described in accordance with AFTES recommendations,
and therefore according to the business approach advocated by the entire profession.
8 CONCLUSION
The work of both working groups (GC and GT) goes on in order to have a panorama cover-
ing the entire life cycle. The next step is to manage the information related to the operation,
maintenance, evolution, and, should they occur, damages, repairs, renovations. Data related
to the neighboring structures may also be added in so far as they are available. In order to
ultimately lead to ISO standardization, the work continues on with the extension of standard
proposals from bSI (building SMART International) and OGC (Open Geospatial Consor-
tium), such as IFC-Tunnel, GeoSciML and GroundWaterML. The MINnD project welcomes
any partner who would like to get involved.
REFERENCES
www.minnd.fr
Guide d’application au domaine des ouvrages souterrains de la norme NF P 94-500 (version 2013) relative
aux missions d’ingénierie géotechnique – AFTES Recommendation n°GT43R1F1 - November/Decem-
ber 2015 – n°252.
Caractérisation des incertitudes et des risques géologiques, hydrogéologiques et géotechniques – AFTES
Recommendation n°GT32R2F1-2012 - n°232.
Prise en compte des risques techniques dans les projets d’ouvrages souterrains en vue de la consultation des
entreprises – AFTES Recommendation n°GT32R3F1 - November/December 2016 – n°258.
CETU - Document d’information – Prix des tunnels – March 2016 - (http://www.cetu.developpement-dur
able.gouv.fr/genie-civil-a572.html).
ISO 16739:2013 Industry Foundation Classes (IFC) for data sharing in the construction and facility man-
agement industries.
AFNOR Standard NF P 94-500 Missions d’ingénierie géotechnique – classification et spécifications -
November 2013.
Stascheit J., Meschke G., Koch C., Hegeman F. & König M., 2013, Process oriented numerical simula-
tion of mechanized tunneling using an IFC-based tunnel product model, 13th International Conference
on Construction Applications of Virtual Reality, 30–31 October 2013, London.
Mao S., Lebrun JL, Doukari O., Aguejdad R. & Yuan Y, Modélisation 3D BIM multi-échelle d’un
projet BTP Tunnel, Spatial Analysis and GEOmatics conference, SAGEO 2015.
Yabuki N., Representation of caves in a shield tunnel product model. In book: eWork and eBusiness in
Architecture, Engineering and Construction, pp.545–550, September 2008.
2959
