Final Report Summary - NETTUN (New Technologies for Tunnelling and Underground Works)
Executive Summary:
NeTTUN was initiated in answer to the European Commission’s expectations stated in the NMP.2011.4.0-2 call. The project was designed in a way to enable groundbreaking change in the construction, management, and maintenance of tunnels. This was to be realised via a set of focussed WPs addressing key scientific and technical challenges: (i) ground prediction ahead of TBMs; (ii) unmanned exchange of drag bits and disc cutters; (iii) increase of cutter tools lifetime; (iv) software modelling the global cost and time risks on tunnel projects; (v) a suite of systems for modelling the ground movements created by EPB tunnelling and for designing measures to protect the surrounding structures; (vi) a Decision Support system for optimised tunnel maintenance management.
The NeTTUN Consortium was assembled to address this challenging set of research themes and objectives. Two beneficiaries joined the consortium of initially 21 partners and brought their unique scientific expertise and experience of the tunnelling sector. An Advisory Committee of six renowned experts completed the task force needed to achieve the ambitious goals of NeTTUN.
The management team made of Ecole Centrale de Lyon, a French top-level engineering school involved in international research (project coordinator) and NFM, the French Tunnel Boring Machine manufacturer (scientific and technical management) allowed a successful control of the project throughout its 54 months duration, starting 1st September 2012.
Exploitation was a key driver from the start of the project, taking the 6 Key Exploitable Results to a stage where exploitation can become reality. Six new patents were applied for, protecting the foreground generated during the project. In the course of the project, the following exploitable results were obtained:
TULIPS: A ground prediction system for TBMs combining new radar antennas, electronics, and data processing software, an S-wave seismic source with unprecedented performance associated with an innovative retractable seismic sensor, software for the automatic detection of features in the radar and seismic images and for the tracking of geotechnical events located ahead of the TBM, an integrated web-based end-user interface, and control software handling the complete process at the system level. Each subsystem was successfully tested on a specific test site.
HECTOR: A heavy-duty hydraulic robotic system comprising a robotic arm on which dedicated and interchangeable grippers can be mounted to handle disc cutters or drag bits, taking care of the complete tool replacement process in full reliability. New locking devices were developed for disc cutters and drag bits that allow handling by the “single handed” robot. The system is controlled by a high-level software handling the trajectories and communicating in real time with a low-level software managing the positioning of the individual actuators. The system was tested in-house and is ready for integration in a TBM.
New materials for drag bits: Twenty new materials were developed, based on a combination of chemical compositions and sintering processes. New innovative tribometers were designed and built to assess the behaviour of each produced materials, out of which six materials have demonstrated a 20% improvement on wear resistance. A set of drag bits was built with each of these materials and sent to a tunnel jobsite for a final assessment of the performance in the field (in progress).
EDO: A software programme that detects over-excavation situations at the earliest possible stage, raising an alarm in such conditions thus providing ample time to act, prevent further evolution of the cause, and avoid the occurrence of adverse consequences on structures and utilities. EDO is based on advanced statistical processing of real time data from the belt conveyor weight scales. The software was tested successfully against real data and simulated data, and will be integrated as a prototype in a tunnel project in the UK, in 2017.
Upgraded DAT: The Decision Aids for Tunnelling (DAT) is a simulation software programme that computes probabilistically the distribution of construction time and cost of tunnel projects in function of geological and construction related uncertainties. A new version of the DAT was developed in the course of NeTTUN, with the objective to reduce the influence of the human factor on the simulation results. To this aim dedicated databases for risk factors and tunnelling performance were added, and an expert system implemented in order to help the user generate unbiased input data. Through new added functionalities, the whole programmed was extended to a complete decision-support tool and is available as a licence or through simulation services.
PADTUN: A software programme supporting the decision-making process concerning tunnel diagnosis, and assisting tunnel maintenance experts in defining the appropriate strategy in terms of inspections and maintenance operations. This programme is now in operation at SNCF.
The NeTTUN beneficiary have planned to pursue their work together through further testing, industrialisation, and commercialisation of the developed products and services, which will benefit the whole sector of tunnelling.
Project Context and Objectives:
4.1.2.1 Context
Significant progress has been made in the recent years in the execution of tunnel projects, reaching new challenges in terms of length, difficult ground or rock characteristics, location in dense urban areas and the vicinity of other tunnels and structures, depth below water and related hydrostatic pressure. However there were in 2010 when the Call for Proposals was issued by the European Commission – and still exist today – remaining real and concrete problems that consistently affect mechanised tunnelling, which in turn is seen as lacking control over risks and costs.
Tunnelling is the area of the construction industry that has been most affected by major losses. Besides property losses often in the two-digit million range, third-party liability losses have also been high, and numerous people have lost their lives.
Achieving and demonstrating full control over tunnel projects in terms of risks and costs is a key driver to the expansion of world-wide infrastructure construction works that would contribute in solving a number of major concerns: improve the safety of lives of citizens, reduce air and soil pollution, give access to potable water, etc. Many European companies are or could be involved in such infrastructure projects and developing their expertise further is needed.
4.1.2.2 Main objectives
The NeTTUN project covers the 4 research themes listed in the Work Programme of Call “NMP.2011.4.0-2 Advanced underground technologies for intelligent mining and for inspection, maintenance and excavation”, and targets its three technical objectives (i), (ii) and (iii), as summarised in tables a) and b):
nr. Research theme
1 Ground prediction ahead of the excavation front
2 Reduction of risks (on health, safety and the environment), through the automation of tasks
3 Assessment of risks on surrounding structures, through modelling and simulation; costs and risks models; monitoring of tunnel and/or surrounding structures
4 Inspection and maintenance of underground structures
nr. Objectives
i Productivity increase by 20% minimum
ii Underground operations with zero impact on surroundings
iii Safe underground working and operating environments
iv Strengthened competitiveness of European industry
v Increased sustainable access to underground resources
Table a) and b) – Research themes and project objectives
Objective (iv) is a global objective that is achieved as described in §XX3.
Objective (v) is specific to the mining industry and not relevant for tunnels
The project will achieve industrial integration and implementation of solutions with proven feasibility.
The main scientific and technical objectives of NeTTUN were:
• Develop an advanced ground prediction system for TBM, integrating concurrent approaches and performing an automated diagnosis, that is applicable to soft grounds and lined tunnels
• Robotise the most risky maintenance operations in a TBM (exchange of disc cutters and drag bits in the cutter head)
• Manufacture improved TBM drag bits with extended lifetime
• Improve an existing Decision Aid Software in the modelling of global risks on tunnel projects
• Enhance the performance of modelling ground movements due to mechanised tunnelling in soft grounds
• Extend the monitoring capabilities in the vicinity of a tunnel work site
• Develop advanced and practical method to design and implement measures protecting surrounding structures from the adverse effects of TBM tunnelling
• Develop advanced TBM on-board monitoring in order to detect over-excavation situations
• Develop an expert system supporting human experts in the strategy of tunnel maintenance
The project was designed so that the technology developments cooperate in achieving the objectives defined by the NMP work programme. The four research themes interact logically in forming a complete project, and are not parallel developments led independently.
The project is architectured in 12 Work Packages.
Project Results:
1 TULIPS Ground prediction system
TBMs progress blind in the ground. Geotechnical accidents and more generally “obstacles” that can negatively impact the TBM progress are anticipated based on preliminary site investigations such as boreholes from the surface. In many instances, such as densely populated and built areas, this information is far from being sufficient. Encountering such obstacles can result in major technical problems such as TBM damage but also variable degrees of ground settlements ranging up to open sky sinkholes, with consequences such as the damage to surrounding structures (buildings, other tunnels, etc.) or even collapse of buildings.
The standard way for obtaining information about the ground ahead is based on boreholes made from the TBM shield; this approach is very slow and severely impacts the TBM availability. Alternative means for predicting the ground ahead of the excavation face have been expected by contractors and project owners. Although several systems have been designed, most of them are focused on tunnels in rock and perform poorly in soft ground lined tunnels. This is the target of TULIPS, the system developed within the NeTTUN project.
The requirements for TULIPS can be summarised as follows:
• Applicable to soft grounds and lined tunnels
• Operate the system in hidden time vs. the TBM operation
• Integrated inside TBM
• Fully automatic process providing readily usable data
• Fast, frequent and effective look ahead during boring
The approach consists in
• Using complementary physical approaches to capture information on the ground ahead
• Combining the data in the most efficient way through data fusion techniques
In its first implementation, and as a target for NeTTUN, data is provided by a radar subsystem and a seismic subsystem. Radar acquisition is done as a circular scan (continuous rotation for at least one or complete turn), with as many antennas as possible along the cutter head radius. Seismic acquisition is done as fixed images, by positioning the cutter head at given angular positions, for example every 60°, using one (or possibly several) sources and several sensors aligned along a same cutter head diameter. Implementing a 0.5 m pitch between sensors allows reaching the desired resolution.
The process is shown in Figure 2 and is described here:
• Data is captured during ring erection, and is associated to a position of the TBM
• Seismic data is inverted (Full Wave Inversion) automatically , which provides a focused image ready for further processing
• Both the radar data and the inverted seismic image are processed in order to detect features, i.e. areas of the image that could correspond to geotechnical events; these expected events are referenced to a position is the ground ahead
• The system relies on the geological context and the knowledge database to indicate the probable nature of each expected event; this output is also submitted to a geo-expert who can confirm or reject the results of the automatic interpretation process
• When the TBM has progressed to the next ring, a new acquisition/feature detection cycle is executed; through a process of tracking, the system confirms the position and characteristics of each expected event
• Once the TBM has passed through the position of an expected event, the geo-expert can confirm or modify the information related to the event; this is stored in the knowledge database.
Figure 2 – TULIPS data flow and process
A complete radar subsystem has been designed, starting with the design of 2 antennas, one for short range detection, one for long range. Although preliminary tests were quite promising it was eventually found during the prototype field tests that these designs would not reach the targeted performance. Therefore a new design was launched, which proved successful and applied to both the short and long range antennas through their implementation in a Planar Antenna Module. A complete electronic system was concurrently designed and built, and the related control software developed and integrated. Subsystem tests were performed in the lab, in a controlled (known) environment and gave results that met the requirements.
A complete seismic subsystem has been designed, based on S-waves, which are best suited to soft grounds. The hardware was developed, comprising a new seismic source (vibrator) reaching down to low frequencies and a sensor, both designed to be integrated in a TBM cutter head, i.e. deployed against the excavation face during the measurements (cutter head stopped) and retracted at any time the cutter head is rotating (mainly during excavation). Both these components were tested in the lab and showed that they met the requirements in terms of frequency spectrum but also robustness and reliability.
A test site was designed and built in Eindhoven (the Netherlands) in order to assess the performance of each subsystem, i.e. the capability of detecting various geotechnical scenarios and buried objects that were buried in known positions. A “TBM cutter head mockup” was designed and built, allowing a simple radar circular scan. After each complete acquisition of radar and seismic data, a layer of ground was removed from the surface, simulating the advance of the TBM, and a new data acquisition cycle performed.
An automatic Full Wave Inversion software was developed, delivering focused images from the data without any human intervention. This process requires that low frequencies be present in the data, hence the special characteristics of the source and sensors. The software was first tested and validated against simulated data and then optimised in terms of execution time, in order to meet the “real time” constraints of a TBM advancing from one ring to the next. It was eventually applied to the Eindhoven data. The target of a 30 minute process was proven to be achievable using a high-end available multi-core computer, devices that are currently offered at a reasonable price.
The complete TULIPS system software was developed, comprising:
• The data Analysis and Fusion software – for automatic interpretation and fusion of multiple sets of imaging data e.g. from different modalities (radar and seismic with different configurations)
• The Visualisation software – for viewing and interacting with raw and processed imaging data
All of these functionalities are available through a single end-user interface, implemented as a web-based application allowing users to be working anywhere on the tunnel jobsite or from a remote office. The sensor data itself can be viewed in 2D or 3D, onto which the detected features can be added. In addition the interface allows the user to display the longitudinal profile and geology of the tunnel together with the corresponding TBM position and the position and characteristics of predicted and past events, as well as TBM parameters that are representative of the current geological conditions. The user can navigate along the already excavated portion of the tunnel and display the stored images and associated features and events data. Altogether this forms a complete workstation with all information at hand for the geo-expert to perform a diagnosis of the situation ahead of the TBM.
The complete system was integrated as shown in Figure 3. A dedicated rugged “TULIPS box” was designed and built in order to house all the sensing system components that must fit inside the cutter head and ensure the adequate protection against water, pressure, vibrations and shocks. Communication between the TULIPS box and the TBM is done by a network through a slip ring also providing power, while the rotary joint allows to operate the hydraulic actuators for deploying./retracting the seismic components. This network also reaches the surface offices where the inversion computer and the end-user interface can be installed. The TULIPS-TBM controller communication was developed and tested, providing a “press button” start of a complete acquisition and data processing sequence.
Figure 3 – TULIPS system architecture with radar devices (in orange), seismic devices (light blue), and common devices (green)
Feasibility studies were made for the implementation of a full size TULIPS on various types and diameters of TBMs, which were successful although it is as expected more difficult on small TBM diameters, i.e. in the 6 to 7 m range.
The NFM contract for the delivery of the Lyon-Turin TBM (hard rock single shield, 11.25 m diameter) gave the opportunity to install a radar-only TULIPS. The antenna module was redesigned based on the prototype design in order to incorporate all the mechanical constraints, while the cutter head was specifically designed to accommodate two such antenna modules and all the related cabling and protective ducts (see Figure 4). The system was fully tested in the factory and on the jobsite after rebuilding the TBM. However the extreme conditions encountered at the very beginning of the drive (unstable excavation front with rock falling onto the antennas) led to damaging the antennas and ultimately to their destruction – unfortunately no real world image was obtained before this happened. However this experience has taught us a lot on the way to reinforce the system and achieve a reliable operation in the very tough environment of a TBM at work.
Figure 4 – The Lyon-Turin TBM at the NFM factory.
The housings for the two antenna modules can be seen.
2 Maintenance robot
Drag bits and disc cutters, which equip a TBM cutter head, are submitted to intense abrasive wear and their replacement is required along the progress of the TBM. This is done by operators who must enter the excavation chamber and the cutter head. In the EPB or Slurry machines used in soft grounds, access to the front part of the machine is complex and dangerous because this area is pressurised at several bars. This implies hyperbaric work with compression and decompression of the workers, with consequential health problems such as Decompression Illness. Such procedures are time consuming although the duration of work in the pressurised area is very limited, similarly to divers timetables. Operations are slow, in the order of several hours each time, during which the excavation is stopped.
Alternatives to hyperbaric interventions have been developed but they exhibit significant drawbacks because of their impact on the TBM design and cost. Therefore there is a high expectation from the contractors and the tunnelling industry for a safe and efficient approach for the replacement operations of disc cutters and drag bits. In NeTTUN we have decided to address the problem using a robot – HECTOR – as described below.
The concept is similar to the human interventions, except that a robot is used, which can operate at the excavation chamber counter-pressure, thereby eliminating all idle time corresponding to compression/decompression. During excavation, HECTOR is stored in a chamber located at the top of the TBM. When an intervention is ordered: a) the excavation chamber is partly emptied of its muck, which is replaced by compressed air; b) the cutter head is positioned so that the tool to be changed is in the upper vertical radius; c) the front gate opens and the robot extends itself in order to reach the cutter head and tool to replace; d) the robot gripper hols and releases the tool; e) the robot folds back into its storage chamber and positions the worn tool in a buffer; f) the robot picks a new tool from the buffer and operates in the inverse manner to mount it and secure it on the cutter head.
The HECTOR robot was designed completely from scratch in order to achieve the smallest possible volume and footprint, implementing only the movements that are required for the operations while ensuring the largest possible range of action. Safety in operation was a key driver which resulted in keeping the tool either securely fixed on the cutter head or held by the gripper. New disc cutter and drag bit locking mechanisms were designed to comply with this requirement and allow mounting and dismounting with a single “hand”, i.e. a dedicated gripper.
The robotic arm was designed and built together with the 8 hydraulic actuators and control circuitry. The hydraulic devices are themselves driven by a National Instruments controller hosting the low-level software.
Both a disc cutter and a drag bit grippers were designed. According to the cycle of operations planned, the appropriate gripper can be mounted on a standard interface at the end of the arm. Each gripper includes the necessary hydraulic wrenches to operate the locking mechanism while holding the disc cutter or drag bit. Each gripper also contains the electronic board that was purposely designed to host the corresponding local low-level software.
A test bench was designed and manufactured, including a mock-up of a cutter head section with several housings for disc cutters and drag bits. The system under tests at NFM is shown in Figure 5.
Figure 5 – Robot system under tests
The software architecture is shown in Figure 6 . It comprises a low-level part, which directly controls the actuators of the arm and gripper, and a high-level part that calculates the trajectory in real-time and sends orders to the low-level software so that the trajectory is effectively followed. The High-level part is equipped with the Human-Machine Interface using a 3D mouse, keyboard and screen, allowing for various modes of operation such as “manual” in the TBM frame of reference (used to position the gripper close to the tool to handle), “manual” in the gripper frame of reference (used once the gripper is in the vicinity of the tool to handle); or “automatic”, with predefined standard positions (e.g. “home”) that the robot then reaches automatically.
Figure 6 – Architecture of the robot control system
Integration of hardware and software was performed successfully at the NFM facilities in Le Creusot, with the participation of all concerned partners.
Feasibility studies were performed for the integration of the robot in various types and diameters of TBMs and resulted in a decision matrix. This confirmed that the robot is suited for TBMs above 8.75 m (with some impact on the TBM design) and can be integrated with close to zero impact in TBMs of above 10.5 m in diameter.
There was however no opportunity during the course of the NeTTUN project to design and deliver a TBM equipped with the robot, in spite of all efforts made (dissemination through papers, conferences) and direct contact with end-users e.g. those who participated in the market survey (Exploitation).
At the end of NeTTUN we have identified the technical points that need improvement and those that were not completed, such as the integration and tests with the drag bit gripper, the video feedback to the operator, and the cleaning system using high-pressure water jets.
3 Improved drag bits
TBM tunnelling in soft or mixed ground relies on the efficiency of “drag bits”, which scrape the soil off the excavation face. These tools suffer from different types of wear (abrasion, impact, fatigue...) which vary according to the ground abrasivity and hardness and TBM operating conditions. Replacing these tools generates heavy work for the TBM maintenance crew, with associated risks when working in hyperbaric conditions and costs – of personnel, replacement parts, and idle time. It is therefore of great interest to improve the lifetime of drag bits in order to reduce these problems (human risks; jobsite cost and tunnelling works duration). This is what NeTTUN has addressed in WP7.
Two aspects of the problem were addressed simultaneously as follows: a) improved hard materials, implemented as inserts that constitute the “active” parts in contact with the ground; b) optimised shape of the drag bit including the position of the hard material relative to the standard steel body. Work started with an assessment of the situation on wear, with the analysis of worn drag bits from TBMs on various jobsites: this phase is common to both tasks a) and b) as it provides information on the type of wear and a “map” of the wear intensity. This study showed that the wear of the carbide inserts located on the drag bits is due to loss of tungsten carbide grains, scratches, and fractures.
• Improved hard materials
Work on the development of new materials was targeted at obtaining an improvement of 20% minimum of the wear resistance as compared to a reference material (used for the current NFM drag bits). This development followed parallel approaches in terms of process and composition. The initial step was however to define the development and test plan, which led to the design and construction of tribometers allowing the comparison of wear between different materials and the reference. An abrasive wear tribometer was therefore made available at ECL-LTDS, while the one at TU Tallinn was designed to operate in a combined abrasion+impact mode.
Three families of new materials were developed as defined in the roadmap: family #1, based on conventional sintering with improved chemical composition in order to increase hardness and toughness; family #2 based on reactive sintering, producing bi-composites or double doped cermets; family #3 was decided based on the results obtained with materials #1 and #2, as a combination of the approaches giving the best results with materials #1 and #2 (optimisation of chemical composition (additives type and concentration) and production process, in conventional sintering or reactive sintering (straight route and mixed route)). At each step, the results of the previous step were used to orient the development of the next material.
The results show that the test method strongly influences the results – some materials behave very well in pure abrasive tests and poorly in situations with impact. This led to further research on the relation between wear resistance and mechanical properties (mainly hardness and toughness), in order to define a law giving a prediction of a material’s wear resistance.
Three materials behaved better than the target in both test conditions and were selected for field tests. Two materials were selected as reaching the target in impact conditions while having relatively poor performance in pure abrasion; one was selected as reaching the target in pure abrasion while having poor performance in impact conditions. In total, 6 new promising materials were selected for field tests.
Figure 7 – Prototype inserts of new materials
Inserts made of these 6 materials were produced at TU Tallinn (see Figure 7) and used to prepare a set of 48 drag bits that were sent to a jobsite in Singapore for field tests. These drag bits are identical in shape and configuration to the standard ones initially mounted on the TBM – only the change of insert material is tested.
• Optimised shape of drag bit
The analysis of the forces induced on the global and local levels of a drag bit is a necessary step to understand the wear phenomena observed on the tools and to elaborate more effective tool geometries and more resistant materials.
This objective was addressed by modelling the material flow around a drag bit moving through the ground: Finite Element modelling (continuous media), Discrete Element modelling (granular media), and a comparison with analytical models were used – see Figure 8.
Figure 8 – Discrete Element Modelling of ground flow in the vicinity of a moving drag bit
In addition, a small scale physical model was developed to validate the numerical modelling results. Through this process, the stresses on the drag bits were determined; these results were used to produce a report giving "design rules" for an optimal drag bit geometry and for an optimal configuration of the cutter head.
• Field tests of new drag bits
The prototype drag bits were installed on one of the 6.6 m diameter slurry TBMs at work on the T216 stretch of the Thomson East Coast Line in Singapore, as shown in Figure 9.
Figure 9 – Prototype drag bits mounted on the T216 TBM
Excavation started in November 2016 and was still going on at the end of the project in February 2017. It was not possible to obtain from the contractor that intermediate inspections be performed before the end of the drive, therefore the wear analysis will be performed by the partners after the end of the project. Results will be available only after the end of NeTTUN. However the results will be communicated to the project partners and the EC.
4 Decision Aids for Tunnelling
Tunnelling is one specific field of civil engineering that is more than others inherently risky. Very costly infrastructure projects are planned but the conditions of their execution remains affected by a high degree of uncertainty. The first and most important factor is related to the geotechnical conditions: whatever the extent of site surveys, uncertainties remain as to the nature and structure of the ground in between the samples taken whether by core drilling or geophysical imaging. The second risk factor concerns the construction method, the actual performance of the equipment used, the skills of the people working on the project, etc. When having to decide in launching a project, decision-makers need to know what risk they take in terms of costs and delivery time. A significant step was made in this field with the development of the Decision Aids for Tunnelling software (DAT) in the 1980’s by Prof H. Einstein (MIT) and the EPFL. This software is based on a probabilistic approach and simulates the construction of the whole tunnel network, including shafts, adits, etc. and produces cost vs. time graphs (scattergrams). These results are however related to the quality and reliability of the input data provided by the user: it has been considered that DAT overly relied on human estimation of geological and constructive risks. Reducing the human factor in the DAT results was taken as a goal for the NeTTUN partners, in WP8.
NeTTUN addressed the problem by introducing actual feedback from previous experience; this approach was implemented as a set of two databases:
• A database of tunnelling risk factors that contains all possible events related with the geology and construction method such as “in limestone, the TBM may encounter a large karst”,
• A database of tunnelling performances that gives ranges of costs per meter and advance rates depending on the ground conditions and the TBM type when tunnelling in normal conditions.
In addition to the databases, an expert system was developed that assists the DAT user in creating realistic input files, by checking and/or suggesting data through the database, by calculating TBM performance based on rock parameters, or by applying expert rules (see Figure 10). At the end of NeTTUN, the following rules were implemented and available within the DAT: the "Penalty factors" model, three analytical models (RME, QTBM and FPI), a rough estimation for preliminary studies, and the link to the databases, including the handling of non-perfect matches in the TBM Performance Database.
A significant effort went into the filling of the databases, collecting – through the NeTTUN partners, other companies, literature survey, and tunnelling associations – information about past projects in terms of geology, ground characteristics, advance rate, geotechnical events found, etc.
Figure 10 – General architecture of the upgraded DAT and interactions with expert system
A new user interface was developed for more user-friendliness. Through other added functionalities, the DAT was extended to a complete decision-support tool dedicated to construction management, allowing the user to update the expected construction time and cost of different variants (see Figure 11), the impact of performing additional geological surveys, of switching to a different TBM maintenance scenario, etc. The software also gives possibilities to enter in more details such as analysing the influence of a particular zone, of specific ground parameters, etc. on the cost vs. time results.
Figure 11 – “Include New Data”, one of the added functionalities. This allows the user compare the time-cost scattergrams of 2 options of the same project
The NeTTUN “end-users” (construction companies OHL, Razel, Astaldi and engineering companies Systra and BG) performed tests based on real tunnels. Based on the results of these tests, proposals were made for corrections and improvement; after discussions a decision was made to implement or not the proposed changes. The final version of the DAT software released at the end of NeTTUN contains these last changes.
Although it had been planned to test DAT at various stages of actual tunnel projects, i.e. both at design phase and in construction phase, this was finally not achieved by lack of suitable projects within the consortium. Still a comparison was made between the pre-NeTTUN and final NeTTUN versions of the DAT. This was performed on the Lyon-Turin base tunnel, which had been simulated in 2013 and was simulated again at the end of 2016 – this time taking into account a new starting point for the drives in Italy, but also the back-analysis of the exploratory tunnels of Maddalena and, partially, of St-Martin-La-Porte. The NeTTUN version provided results in line with the time-position diagrams resulting from the deterministic design. It helped in underlining the main expected difficulties and the remaining uncertainties, as well as the positive effects (i.e. lower variability) due to the exploratory/access tunnels and the possibility of new drives.
5 Controlling the impact of tunnelling on surrounding structures
The growing demand for public and sustainable transport requires the construction of an increasing number of underground infrastructures. The many constraints and technical challenges associated to underground construction in urbanised areas often lead to high costs and long completion times. This is particularly true for cities characterised by a high density of population, significant archaeological heritage, and the presence of masonry structures of historical and monumental value, which are particularly sensitive to subsidence induced by excavation. It is therefore often necessary to adopt complex control systems of the excavation process, in order to achieve the maximum limitation of deformations, to devise intense monitoring schemes, and, where necessary, to implement techniques for the mitigation of the potential damage and the protection of the structures affected by excavation, with a significant increase in the construction costs. In this context, WP9 focused on the following objectives: (i) to enhance the performance of physical and numerical modelling of ground movements due to mechanised tunnelling in soft grounds, (ii) to study the impact of EPBS tunnel excavation on different types of structures, (iii) to extend the monitoring capabilities in the vicinity of a tunnel work site, (iv) to develop a method to provide early detection of over-excavations in TBM tunnelling, and (v) to evaluate effectiveness and provide a rational basis for the design and implementation of mitigation measures. These objectives were addressed with complementary approaches, using a combination of numerical analysis, laboratory-scale experiments, and in-situ data. The work was organised in 6 Tasks; the bulk of the work that was carried out falls under Tasks 9.1 (physical modelling), 9.3 and 9.6 (numerical modelling of tunnel excavation and of mitigation measures), and 9.4 (early detection of over-excavation), which substantially met all the target objectives of the NeTTUN project. On the other hand, although it was originally intended that the laboratory, numerical, and theoretical developments should be tested against the monitoring results collected on the work sites of Contract T3 of Line C of Roma Metro, due to a number of reasons independent of partners (financial, political and archaeological constraints) the construction works were delayed. This meant that it was not possible to complete all the field work originally planned under Tasks 9.2 (field monitoring) and 9.5 (field trials of mitigation measures). Although the bulk of the work that was promised was carried out, there are many areas in which further work is required, either to fully meet the research objectives that were set in the original DOW, or because activities that were not originally planned have emerged as worth of exploring, beyond the original scope of the project, as summarised below, together with the main achievements of WP9.
• Physical modelling of TBM tunnelling
An extensive experimental campaign was conducted by researchers from ENTPE, using a 1g reduced-scale model of an EPB shield, to investigate the impact of tunnelling on neighbouring structures founded on piles and the efficiency of physical barriers on the limitation of surface displacements. Significant modifications of the original experimental set-up were undertaken in order to upgrade its geometric dimensions and its control system so that the experimental campaign could be carried out. Considerable financial and work efforts were also devoted to the design and construction of model structures (piles, groups of piles, and barriers) and of ancillary equipment to fill and empty the sand box, load and install the piles and monitor machine parameters, ground and structural response during TBM advancement. A total of 14 tests were carried out under the NeTTUN project, including one preliminary tests to verify the good functioning of the new equipment, two tests in virgin soil under different tunnelling conditions to be used as a reference, eight tests of the interaction of TBM tunnelling with piled foundations, and three tests of the interaction of TBM tunnelling with barriers. Tests carried out with the 1g EPBS reduced scale model of ENTPE have shown the ability of this device in its new configuration to simulate closely the principal operations of a real EPBS and to reproduce its 3D excavation process. The tests performed with a group of four (2x2) piles installed in the vicinity of the tunnel have shown that the soil-piles-TBM interaction is characterised by a redistribution of the working load in the piles of the group during TBM advance, which is partly due to the variations of ground stresses around the piles caused by the TBM advance. Additional tests should be carried out beyond the end of the NeTTUN project to study the influence of different factors. In particular, the effects of the distance between the piles and the influence of the conditions of contact between the piles head and the pile cap should be explored.
The three tests carried out using various models of diaphragm walls (see Figure 12) permitted to study the efficiency of these physical barriers to limit surface displacements. During these tests, only the wall depth relative to the tunnel position was varied.
Figure 12 – Model of diaphragm wall used during the tests – shown in position during container filling
The results obtained show that the three barriers used strongly limit the vertical displacements of the soil outside the barrier. In particular, the tests show that the magnitude of the soil settlements near the barrier decreases with the length of the wall. Additional tests will be carried out beyond the end of the NeTTUN project to analyse the influence of the weight of the barrier, explore the relevance of lightened barriers, and adjacent pile barriers, which represent, from a practical point of view, a non-negligible economic gain. The NeTTUN project permitted to collect a large number of experimental results obtained under various conditions of excavation and with different neighbouring structures (single piles, group of piles, mitigation barriers). These can provide a reliable support to calibrate numerical simulation models dealing with tunnel-soil-piles interactions.
• Numerical modelling
The numerical modelling working team, consisting of researchers from URO2 and NTUA, cooperated to develop a reliable Lagrangian 3D FE procedure to model TBM tunnelling. This included the realistic simulation of the main physical phenomena occurring during EPB shield tunnelling, such as e.g. the application of a variable support pressure at the face, the presence of a physical gap between the excavation boundary and the permanent lining (due to over-cut at the cutter-head, shield tapering, and installation of the segmental lining within the shield) and the execution of tail void grouting. The interaction between the shield and the ground was simulated introducing appropriate normal and tangential contact laws, while tail void grouting was modelled using a time-dependent setting law for the grout and an initial stress state in the grout elements to represent the injection pressure. Since a significant proportion of ground deformation due to tunnelling in fine-grained soils is associated with consolidation settlements, fully coupled consolidation analyses were carried out, with realistic rates of advancement of the shield.
Figure 13 – Finite element mesh employed for greenfield simulations
Alternative approaches for the simulation of mechanised tunnelling based on an Arbitrary Lagrangian Eulerian (ALE) description of the tunnelling process were also explored. The main aim of this work was the prediction of the volume loss, for given soil properties, geometry, and operational parameters of the TBM. In the ALE approach the advance of the shield is simulated as a continuous process and the support pressure applied at the tunnel face is a result of the analysis, depending on the ratio of the mass of excavated soil in unit time to the rate of extraction of the muck from the pressure chamber. At the moment, the applicability of the ALE technique has only been assessed against the results of the small scale physical model of mechanised tunnelling carried out at ENTPE. Also, the degree of detail and the number of the analyses carried out so far with the ALE approach were limited due to the huge computational effort demanded by the specific implementation of the ALE method. Many extensions of the numerical work are planned or currently under way, such as adopting the distribution of face support pressure resulting from the application of the ALE technique in Lagrangian analyses, and conversely, to apply some of the features of the Lagrangian analyses, such as e.g. the over-cut at the cutter-head, shield tapering, and tail void grouting, in the Eulerian ones. Numerical analyses of the results of physical models (virgin soil, piles and barriers) are currently being improved through the adoption of more advanced constitutive models for the soil, including yielding for small stress increments, hardening and evolving dilatancy. In principle, the Lagrangian numerical strategy that was developed allows modelling the complete system of ground/TBM/monument, from which the impact of the tunnelling works can be determined. However, transposition of the model to real size was achieved only with reference to the passage of the TBMs at the AMA site, both in greenfield and in the vicinity of the trial barrier. Due to the delays on the Line C works, the back analysis of the interaction of tunnelling with the Aurelian Wall will be carried out after the end of NeTTUN, because the details of the construction work are not entirely defined at this time. Although numerical results are very promising at this stage, the comparison with real field data is crucial for the ultimate validation of the proposed methods.
• Field work
The responsible partner for field work was METC, who designed the monitoring schemes, and acquired and installed the field instrumentation at the AMA site and on the Mura Aureliane at Porta Metronia, to monitor the ground response during passage of the TBMs. Background reading of the instrumentation is currently carried out on a regular basis. METC also performed all the necessary preliminary actions to carry out field trials of mitigation measures, including both compensation grouting and barriers. A barrier consisting of 48 adjacent concrete piles was constructed in June 2016 in a dedicated section of the AMA site. Monitoring of ground response to passage of the TBM in the vicinity of the barrier will include the measurement of vertical and horizontal displacements of a number of points both at surface and subsurface by precision levelling, extensometers and inclinometers, whose installation was completed by July 2016. The field trial of compensation grouting includes a dummy foundation and a monitoring system of surface and subsurface displacements (liquid levelling points, reference points for precision levelling, inclinometers and extensometers). Because the made ground contains potentially important artefacts/structures, a temporary grouting shaft, consisting of 26 concrete piles and 26+26 bentonite piles, was constructed from which the grout sleeve tubes will be inserted.
Figure 14 – Plan view and section of the compensation grouting field trials at Porta Metronia
Construction of the piles supporting the shaft was completed in November 2016 and, at present, excavation with archaeological methods has reached a depth of about 7 m below ground level. Due to the delays on the construction of the line, all the construction dates have had to be revised numerous times during the project and it has been clear from some time that the TBMs will pass neither through the AMA site nor in the vicinity of Porta Metronia within the scope of the NeTTUN project. Therefore, permission was asked to, and granted by, the EC to proceed with the acquisition and installation of the instrumentation, even if the relevant data would be collected after the end of the project. It is intended that METC will collect the data from the passage of the TBMs through the control sections, conduct the field trials and make the results available to the numerical work group and to all other partners of the NeTTUN consortium.
• Early detection of over-excavation (EDO) system
NTUA were in charge of designing and testing a system for the Early Detection of Over-Excavation (EDO) during TBM tunnelling, which can reduce the adverse consequences of significant over-excavations in an urban environment, including damage to surface structures and utilities due to excessive ground settlements. The system aims to detect over-excavations when they occur, rather than detecting the effects of over-excavations (typically ground surface settlements) as is currently done in classical monitoring. Thus, the proposed system can detect over-excavations early, before they evolve to a degree that can cause appreciable settlements and damage to structures. Another advantage of the proposed system is its simplicity, as it uses muck weight measurements on the conveyor belt (which are performed in practically all TBMs) and processes them analytically by software. Appreciable over-excavations may be caused by ground failure ahead of the TBM face (due to poor ground conditions, too low face pressure of the EPB system, excessive groundwater gradient, or a combination of the above), providing more excavated material than the anticipated mass. The EDO system incorporates a Muck Weight Measurement and Assessment Sub-System that exploits muck weight measurements, typically measured by digital scales on the conveyor belt and recorded by the TBM control system in real time. The coefficient of variation of the measurements depends on the tunnel diameter, the TBM type, the conveyor belt, the weight scale, and of course the ground itself. Thus, the coefficient of variation is a parameter which is calculated on the basis of initial measurements of muck weights for a TBM advance of a few segments and subsequently used in future assessments. In order to assess the effectiveness of the EDO system, a thorough investigation was conducted using actual (recorded) datasets of weight measurements from several TBM projects – Figure 15). An extensive numerical investigation was also carried out to quantify the expected face take and the resulting theoretical over-excavation, resulting in charts that are integrated into the EDO system.
Figure 15 – EDO system warnings for fictitious over-excavations added to real data
The next step in the development of the method is the implementation of the EDO system on a real TBM during operation, which will provide valuable experience and shed new light to various sources of uncertainty that appear when stored data are examined, mainly related to the origin of the variability that the weight measurements exhibit and how this is related to various operations taking place.
NTUA remains fully committed to the objective of integrating and testing the method in real time in the near future by investing the necessary man-power even after the completion of the NeTTUN programme. In fact, in close collaboration with NFM it is planned to integrate the system in two EPB machines currently manufactured by NFM and expected to be launched soon (Spring 2017) in the extension of the Northern Line of the London Tube in the UK.
6 Tunnel maintenance strategy
Tunnel maintenance requires complex decision making, which involves pathology diagnosis and risk assessment, to ensure full safety while optimising the maintenance and repair costs. In the digital era, a plethora of data about the tunnel conditions, including both periodic tunnel inspections and contextual factors affecting the condition of tunnels, is becoming available. Inspecting this data to identify notable combinations of disorders and influencing factors, to detect possible pathologies, and to assess the possible risk of further deterioration is highly complex and requires extensive expertise. The complexity is further aggravated by a range of factors, such as infrastructure aging, climate change, and rapidly increasing urbanisation. Moreover, this complex decision process with high societal and economic impact is often prone to subjectivity and poorly scales across use cases and domains.
These challenges were addressed in NeTTUN Work Package 10 by an interdisciplinary team of tunnel experts and knowledge engineers. A decision support system that assists with pathology diagnosis and assessment of tunnels (PADTUN) was developed. It involves two components: (1) Pathology assessment and (2) Urgency finder (see Figure 16).
Figure 16 – PADTUN system architecture
1st component: Pathology assessment assists tunnel experts in making decisions about a tunnel’s condition with respect to its disorders and diagnosis influencing factors. It uses semantic web technologies for knowledge capture, representation, and reasoning. Ontologies, which represent the main concepts and relations associated with pathology assessment, were developed to capture the existing decision process concerning the maintenance of tunnels, which provides a knowledge model for automated decision support. The PADTUN ontologies, which are represented in the web ontology language (OWL) that follows internationally established standards for knowledge reuse, are the first ever ontologies developed for the domain of tunnel diagnosis and maintenance. Pathology detection is based on tunnel disorders and influencing factors. The ontological knowledge (which includes 125 concepts, 49 relations, 590 individuals, and 3981 axioms) was provided by domain experts at SNCF, further extended by following tunnel maintenance guides, and validated by a domain expert from Swiss Rail. Extracts from the PADTUN ontologies are given in Figure 17.
Tunnel inspection data are linked to the PADTUN ontologies which enables applying the knowledge to identify tunnel regions where specific pathologies are detected (see Figure 18) for an example of a tunnel diagnosis with PADTUN).
2nd component: Urgency finder is crucial for tunnel maintenance because repairs are expensive and affect the operational efficiency of the tunnel. To predict the risk factor and the resulting repair urgency in the tunnel maintenance planning, PADTUN utilises supervised machine learning methods that learn from ground truth data of past human decisions. The data on the basis of which the PADTUN urgency finder component is developed consists of 47 tunnels of varying lengths, a total of approximately 60 tunnel kilometres. The data included periodic survey inspections for multiple years, as well as other data such as the method of construction of the tunnel. Expert annotations were collected for each 10m tunnel segment for each survey as to the degree of repair urgency. These annotations provided the ground truth for both model training and evaluation. Several machine learning methods, including Decision Trees, Random Forests, Support Vector Machines, were implemented and assessed. Decision Trees were selected as the best performing method with sufficient transparency to examine the system’s suggestions for the urgency of repair for each tunnel portion.
Figure 17 – Extract from the PADTUN ontologies: associations between a pathology and disorders (top) and association between pathologies and influencing factors (bottom).
Figure 18 – PADTUN web interface for tunnel diagnosis: the tunnel is divided by portions of 10 meters which are numbered (column headings); for each portion pathologies are detected (main rectangle); and the overall urgency repair indicated (first row of the main rectangle).
PADTUN was evaluated on a database of SNCF tunnels, and an example tunnel from Swiss Rail. The results were very good with only 3% of false negatives (actual pathologies that are not detected). Machine learning from tunnel data was applied to investigate the reasons for misdiagnosis. It was found that the misdiagnosed pathologies were rare occurrences of pathologies, with the exception of “creep” which had false positives due to an overlooked “overburden” constraint. PADTUN was deployed on a virtual machine on a computer at SNCF for routine tests with tunnel data. Subsequently, with the aid of a developer’s guide, which was provided to SNCF and guidance from personnel from Leeds, a native installation on an SNCF computer has been effected.
In a broader context, PADTUN illustrates a step change in combining knowledge engineering and machine learning to support complex decision making related to sustainable urban infrastructures. The approach is applicable not only to tunnels, but also to other transport constructions (e.g. bridges, roads, tracks). Possible extensions in this direction are being explored considering both research and innovation aspects.
Potential Impact:
1 Impact
The impact of the NeTTUN project is at several levels:
• NeTTUN was mainly aimed at improving the safety of specialised workers and, beyond that, of all European citizen users of underground structures, especially in urban areas.
• NeTTUN represents a significant step forward in the goal to fully automate TBMs and tunnelling works.
• It contributed to improving European cohesion. Nine European countries have participated in the project: Estonia, France, Germany, Greece, Italy, Spain, Switzerland, The Netherlands, and United Kingdom.
• It promotes the collaboration between universities, research centres, and industries (systems manufacturing and tunnel construction and maintenance) thereby combining the needs of the industries with the knowledge of researchers.
• It included the participation of one organisation from one most recent Member State: Estonia.
• The consortium included the participation of five SMEs: Sialtec, CISTEME, MI Partners, Geo2X, and INCAS Partners.
• As per the EC Strategy 2020, the project has already been of benefit to Education as it was presented at courses such as the Specializing Master Course on Tunnelling and Tunnel Boring Machines (from the Politecnico di Torino).
• The project indeed contributes to all the expected impacts listed in the Work Programme as further detailed below.
1.1 A project with enlarged business perspectives, delivering productivity increase
• NeTTUN extends the range of mechanised tunnelling, which is reported as the safest tunnelling method: at least 55% of recorded collapses occur with conventional methods.
• The project provides directly applicable answers to the real and concrete problems that consistently affect tunnelling with TBMs. TBMs currently operate in a “blind” environment without clear knowledge of what lies ahead of them at the face. With the development of the NeTTUN ground prediction system, combining radar, seismic, and data fusion expertise in a way that has never been done before, mechanised tunnelling will be able to improve project delivery times in a safer environment with real images and understanding of what lies ahead of the TBM.
• Project Risk Modelling has a highly positive consequence on the economics of tunnel projects. The improvements brought to the pre-existing DAT software will help decision-makers in planning projects and allow looking for alternatives with time and costs savings. Delivering projects on time and within budget will of course increase the competitiveness of the European tunnelling industry on the world stage, see below.
1.2 A project that delivers underground operations with zero impact on surroundings
• Nothing will have greater significance than demonstrating on the NeTTUN project the successful tunnelling under the Coliseum of Rome with zero impact on this UNESCO world heritage ancient site. All the knowledge is now available for a successful construction of the risky stretches o Line C. Dissemination of this result will carry great impact for Exploitation of this knowledge beyond the end of NeTTUN.
• The future of Tunnelling will demand more complex, longer and deeper tunnels and this is where the NeTTUN results will be most efficiently implemented. In such projects, government agencies are becoming more concerned with improving the tunnelling construction Health & Safety environment for workers as well as public safety and surrounding structures’ integrity. Extremely dense areas such as Singapore and Hong Kong have plans for large projects in which the level of technical difficulty is exacerbated by the complex ground conditions. For such projects the NeTTUN results in Ground Prediction, Robotised Maintenance, and Control of Impact of Tunnelling on Surrounding Structures are now ready to contribute in meeting this demand.
1.3 A project answering societal needs significantly by improving safety
This is the major concern with tunnelling activities and operational use.
• The objective of the NeTTUN project was to ensure a safer underground working and operating environment. The project, with the development of a TBM maintenance robot, better drag bits requiring less maintenance interventions, ground prediction system providing information about geotechnical issues ahead, prediction and monitoring of ground movements, has delivered key elements to achieve this goal.
• In a study of Notable Tunnel Failure Case Histories of 39 overseas cases which occurred from 1964 to 2014 in the last decade, at least 36 people lost their lives as a result of tunnel collapses, building evacuations took place involving hundreds of people, many buildings and major roads were damaged and closed, and tens of thousands of people were affected. A large part of these accidents would have been avoided by using advanced modelling of ground movements and protection measures as developed within NeTTUN and applied to the metro Line C construction in Rome. Implementing reliable ground prediction systems and improved use of the excavated material measurement such as the EDO, both developed within NeTTUN, would also have contributed in avoiding such accidents.
• Underground workers are exposed to personal risks, which is why a significant effort was put in the development of an efficient robotised system that performs maintenance operations in the most dangerous area of the TBM instead of humans, in more than 80% of the cases.
1.4 A project that strengthens competitiveness of European industry
1.4.1 Through Science, Research and Education
• Bringing competing academic laboratories together to share ideas and work on a common goal, such as:
o IRCTR at TU Delft and XLIM at the University of Limoges, teaming in the development of the best possible radar for ground prediction;
o ENTPE in France and the Universities of Roma and Athens working together on the management of geotechnical risks;
o LTDS at Ecole Centrale de Lyon and TU Tallinn developing advanced excavation tools and materials in a common effort.
• Improving partnership between industry and research for example:
o NFM will draw on its existing regional relationships with research organisations such as ENTPE, LTDS, INSA, ENISE, and expand it to a European level
o IDS will benefit from the best academic expertise in radar from TU Delft and XLIM for the creation of new advanced products
o OHL, Razel and the Metro C consortium partners such as Astaldi will develop partnerships with research laboratories that can be sustained on a longer term in applied science fields such as the wear of materials, geotechnical modelling or ground prediction.
o SNCF will be introduced to advanced mathematics and data fusion by working with University of Leeds
• Expanding research activities from a national level to a European level
o University of Leeds participation in NeTTUN will include building on the existing “Mapping the Underworld” projects and extending the techniques in a new domain with other European research beneficiaries (TU Delft, XLIM)
o Produce experimental lab equipment and software models that can be used for subsequent research works concerning geotechnics, tunnelling and material science.
o Attract new students to pursue graduate studies through NeTTUN’s Dissemination activities. There is a particular focus by the NeTTUN beneficiaries in Lyon – NFM, ENTPE, ECL – to attract new engineering students into the Tunnelling industry. This has proven to be very effective (and rewarding): 2 of ECL former students who worked with NFM during their last year project (wear of drag bits) pursued their career in the tunnelling area. They have been hired by major companies and currently work on major underground projects (Montreal metro and Grand Paris Express).
1.4.2 In Industry
• Challenging tunnelling works are currently a special expertise of European civil engineering companies. European civil engineers are commonly called in for their excellence, for projects all around the world (recent examples are Miami harbour, Hong-Kong high-speed rail link). The state of the art advances achieved by the NeTTUN project in Ground Prediction, Robotics for Automated Maintenance, Impact on Surrounding Structures and Project Risk Modelling offers the European Tunnelling industry a real advantage competitively.
• Tunnelling is a fast growing market place. As an example, in the United States, employment in construction is expected to rise 19% between 2010 and 2018; demand for commercial construction and an increase in road, bridge, and tunnel construction will account for the bulk of the job growth. European tunnelling companies need to take advantage of this growth, and NeTTUN advances will be a positive differentiator increasing European competitiveness.
• Improving European companies’ positions for winning developing world infrastructure projects. In particular underground public transportation networks in the very densely populated large cities in developing countries, and many water supply projects dependent on the feasibility of constructing river diversion tunnels, both would benefit from advances made in securing the global risks and costs of tunnelling.
• Reducing the risk of tunnel projects will foster the construction of infrastructure. Such projects rely on the expertise of European companies (e.g. metro project in Guatemala City), but competition is tough e.g. with Japan (Jakarta metro project) and NeTTUN results can assist in this area.
• Create jobs in European countries. NeTTUN has improved competitiveness for the industrial companies such as MI-Partners (creation of the Seismic Mechatronics spinoff selling innovative high performance seismic sources), NFM (improved TBMs and associated services), SNCF (optimised maintenance of their tunnel legacy), IDS (possible new business unit for a Ground Prediction system for Tunnelling), which will contribute to more activity and need for more personnel.
• More than 50% of the NeTTUN partners estimate that participating in the project has generated an increase in employment in their own organisation or at least safeguarded their resources.
1.5 A project that increases sustainable access to underground resources
• TULIPS – the ground prediction system – can be applied to other fields than TBM tunnelling. MI Partners have started commercialising the NeTTUN seismic source to clients in the geotechnical survey business, in particular those working in the oil and gas industry.
• The Ground Penetrating Radar technology developed by NeTTUN deals with issues that are currently limiting the performance of this equipment. On NeTTUN the engineering solutions for the ground penetrating radar (GPR) for TBM incorporating novel system design, antennas, electronics, data processing and information extraction algorithms deliver a breakthrough in short-range radar imaging technology. This technology will be extended to other applications which require intelligent subsurface sensing, including subsurface utility mapping, pipe detection, archaeological and geophysical surveys, construction works, security, and rescue operations particularly for mining and earthquake scenarios.
1.6 A project participating in quality of life improvement throughout the world
• Develop water supply systems for areas of the world demanding urgent access to potable water and irrigation water: many projects are dependent on the feasibility of constructing river diversion tunnels, which would benefit from advances made in securing the global risks and costs of tunnelling.
• Support international health programmes by implementing sewage water networks in the large cities of developing countries, thereby greatly reducing the risk of medical diseases and improving the environment.
2 Main dissemination activities
Dissemination activities were started with a lot of enthusiasm from the very beginning of the project. However, disseminating information on the project raises recurring questions on confidentiality. This is the main reason for which we created the Dissemination Committee, in order to decide on the paper quality but mainly on its confidentiality level. The main role of the Committee was to make the decision on submitting the abstract/paper or not. It also periodically informed the WP Leaders and Beneficiaries of the coming conferences and encouraged their preparation of papers and attendance for scientific and technology watch. The Dissemination Committee was active from Month 4 to Month 54.
2.1 Target Audiences
• The General Public and the Tunnelling Industry
NeTTUN partners were active at participating in Tunnelling fairs, most often associated with conferences throughout the whole duration of the project.
The most important event of the sector is the World Tunnel Congress that is organised by the ITA and is held every year. Altogether 15 papers were presented / will be presented at this congress (2013 to 2017 incl.) and included of the proceedings.
• The Educational and Academic Arenas
Out of the proposed activities to reach these audiences we have focused on the participation in scientific and technical conferences and publication of papers.
In this respect we have covered conferences on radar (AES, GPR2016, and ERC), ceramics (IWAC), materials and wear (ITC), geotechnics (3vX and JNGGI), artificial intelligence (ECAI).
Papers were published in recognised high-level journals such as Wear.
2.2 General Conferences on underground works
NeTTUN partners were active at participating in Tunnelling fairs, most often associated with conferences throughout the whole duration of the project.
The most important event of the sector is the World Tunnel Congress that is organised by the ITA (International Tunnel Association) and is held every year. Altogether 15 papers were presented / will be presented at this congress (2013 to 2017 included) and included in the proceedings.
AFTES (the French chapter of the International Tunnel Association) organises a congress every 3 years. The 2014 edition (Lyon) occurred in the course of NeTTUN, and submitting papers for the 2017 edition (Paris) was also done during the project. 3 papers have been prepared overall.
Other participations of NeTTUN in conferences on tunnelling were:
o Eastern European Tunnelling Conference (Athens, 2014): 6 papers
o TBMs in difficult grounds (DiGs) (Singapore, 2015): 2 papers
2.3 Scientific Conference Papers and Scientific Publications
The Dissemination Committee has been active at reviewing scientific papers proposed by partners. Acceptance was based on two criteria: preservation of IP and paper quality. Changes were most often requested until the papers were accepted, although a few never made their way to publication.
A total of 34 papers were accepted and published by NeTTUN beneficiaries at the following conferences, covering Radar science, Geoscience, Robotics, Tribology, Material science, Geotechnics, and Artificial intelligence: URSI, ECSMGE, ECAP, SEG, TUST, IWAGPR, EAGE, IET, JIFT, IAGIG, ESWC, AES, ICGPR, ERC, IWAC, JNGGI, ITC, Wear, 3vX, ECAI, GJI, EUROFUGE, Tribologia, EURO TUN, Rivista Italiana di Geotecnica, Near Surface Geophysics.
The paper on PADTUN (Leeds/SNCF) presented at ESWC2015 was elected “Best paper” at the conference.
Eight papers have been submitted for presentation beyond the end of NeTTUN, it is expected that they will all be accepted and presented.
2.4 General media
The work done on NeTTUN was also disseminated in general media:
• “Le Journal de Spirou” is a Belgian weekly magazine focused on the young population. It was founded in 1938 and is published in French (over 30,000 copies sold per week in France). Since 2009, some issues contain a special section called “The Lab”, which presents in a cartoon format scientific information, usually in the form of a visit to a research laboratory and interviews with researchers. The issue of 25 February 2015 was dedicated to Tunnel Boring Machines, through a visit made to the NFM factory in France where a 10.2 m EPB (Riyadh metro) was in its final stage of assembly.
• “Sciences et Avenir” is a monthly French magazine for popularisation of science created in 1947. Their April 2016 issue was focussed on the long tunnel projects through the Alps (Gotthard base tunnel completion and first step of Lyon-Turin). The NeTTUN project manager was interviewed for this article, in which our EC-funded research project was promoted. The journalist was invited to the NFM factory in Le Creusot for a presentation of the TBMs being assembled. A special (different) edition of this is available online at https://www.sciencesetavenir.fr/high-tech/transports/tunnel-ferroviaire-du-saint-gothard-la-percee-record-inauguree-le-1er-juin_103811 in which NeTTUN is also mentioned.
2.5 Dissemination and Marketing Materials
Marketing materials are crucial for representation of the identity of the NeTTUN Consortium.
We used the NeTTUN logo on all our public presentations material (e.g. at conferences) and when possible installed the NeTTUN banner in such events.
The leaflet prepared for the World Tunnel Congress 2013 was kept available throughout the project and was distributed during all trade fairs and conferences (distributed on each of the beneficiaries’ booth) and also handed out to all prospects and clients.
A specific brochure was developed by NFM, presenting TULIPS and HECTOR (robot); it was distributed at the WTC 2016 and is continuously used to promote these systems.
A specific brochure was developed by Seismic Mechatronics, the spin-off from MI Parthers, covering the “Lightning” source designed for TULIPS within the NeTTUN project.
2.6 Conclusion
Overall the dissemination activities were successful throughout the project in spite of some restrictions due to the necessary protection of the inventions, in order to allow for the possibility of patenting. This in particular applies to WP7 (new materials) for which there were purposely very few papers but a patent was applied for.
Obviously the beginning of the project is not favourable to publications. The volume rises each year, although there is some negative impact from the periodicity of events, such as the AFTES congress that takes place every 3 years only. It is very rewarding to see that 2017 will be a very active year with 17 publications already planned – and more to come – all beyond the end of NeTTUN.
3 Exploitation of results
Exploitation was addressed from the beginning of the project through a dedicated Committee but was not very successful until the EC provided support through ESIC at the beginning of the second period, which proved very efficient. This led to a better understanding of the objective and proposed methods to achieve the results, thus improving confidence in our capability to succeed. Appropriate contact persons were defined for each entity, making more sense than involving the researchers and engineers only.
A significant step ahead was the hiring at NFM of a dedicated intern for 6 months full time, who took the lead of the Exploitation Committee. Her contract was later extended until October 2016, thereby allowing in order to take the exploitation plan to a point where it can be executed.
After identifying the Key Exploitable Results (KER) with ESIC, a market analysis was conducted for several of them. Questionnaires were prepared, sent to and discussed with potential clients (contractors, tunnel engineering companies). In addition to valuable information on the potential, this approach also provided positive feedback on the viability of the chosen technical concepts and allowed the partners to develop their business network.
Once the KER situation was fixed and acknowledge by all parties, work was focused on the elaboration of a plan for the marketing strategy of each KER. Alternatives were discussed and, for some KERs, decided on. For others this still needs to be finalised.
At this stage, work was initiated on the IPR and focussed on drafting co-ownership agreements between partners in cooperation with the dedicated ESIC consultant. Such an agreement was finalised and signed between ECL-LTDS, TU Tallinn, and NFM for the new materials developed under WP7. This agreement also covers the application for a patent through TU Tallinn. An agreement was also found for the upgraded DAT, with licensing to an SME that provides services based on the use of DAT.
A spin-off company was created from MI-Partners (Seismic Mechatronics) that already delivers seismic sources based on the NeTTUN R&D work. Clients are geotechnical engineering firms working in particular in the oil and gas industry.
Altogether six patents were applied for to cover the NeTTUN results, not including their extensions to foreign countries within Europe and abroad. The concerned results are: TULIPS (the concept, the seismic source); HECTOR (disc cutter locking, drag bit locking, robotic arm architecture); new hard materials (composition and process).
Although a lot has been achieved in terms of Exploitation, there is still a lot to be carried out. The most important success factor is that for each KER one partner takes the lead and organises the work to be done, i.e. finalise the co-ownership agreements, complete the technical work where needed so that the systems are ready any time for demonstration to clients.
List of Websites:
The project website was launched on Month 6. It has been regularly updated by the Website Management Committee and MI Partners in the role of NeTTUN Webmaster.
The url is: www.nettun.org
Contact: pso.nettun@nfm-technologies.com
NeTTUN was initiated in answer to the European Commission’s expectations stated in the NMP.2011.4.0-2 call. The project was designed in a way to enable groundbreaking change in the construction, management, and maintenance of tunnels. This was to be realised via a set of focussed WPs addressing key scientific and technical challenges: (i) ground prediction ahead of TBMs; (ii) unmanned exchange of drag bits and disc cutters; (iii) increase of cutter tools lifetime; (iv) software modelling the global cost and time risks on tunnel projects; (v) a suite of systems for modelling the ground movements created by EPB tunnelling and for designing measures to protect the surrounding structures; (vi) a Decision Support system for optimised tunnel maintenance management.
The NeTTUN Consortium was assembled to address this challenging set of research themes and objectives. Two beneficiaries joined the consortium of initially 21 partners and brought their unique scientific expertise and experience of the tunnelling sector. An Advisory Committee of six renowned experts completed the task force needed to achieve the ambitious goals of NeTTUN.
The management team made of Ecole Centrale de Lyon, a French top-level engineering school involved in international research (project coordinator) and NFM, the French Tunnel Boring Machine manufacturer (scientific and technical management) allowed a successful control of the project throughout its 54 months duration, starting 1st September 2012.
Exploitation was a key driver from the start of the project, taking the 6 Key Exploitable Results to a stage where exploitation can become reality. Six new patents were applied for, protecting the foreground generated during the project. In the course of the project, the following exploitable results were obtained:
TULIPS: A ground prediction system for TBMs combining new radar antennas, electronics, and data processing software, an S-wave seismic source with unprecedented performance associated with an innovative retractable seismic sensor, software for the automatic detection of features in the radar and seismic images and for the tracking of geotechnical events located ahead of the TBM, an integrated web-based end-user interface, and control software handling the complete process at the system level. Each subsystem was successfully tested on a specific test site.
HECTOR: A heavy-duty hydraulic robotic system comprising a robotic arm on which dedicated and interchangeable grippers can be mounted to handle disc cutters or drag bits, taking care of the complete tool replacement process in full reliability. New locking devices were developed for disc cutters and drag bits that allow handling by the “single handed” robot. The system is controlled by a high-level software handling the trajectories and communicating in real time with a low-level software managing the positioning of the individual actuators. The system was tested in-house and is ready for integration in a TBM.
New materials for drag bits: Twenty new materials were developed, based on a combination of chemical compositions and sintering processes. New innovative tribometers were designed and built to assess the behaviour of each produced materials, out of which six materials have demonstrated a 20% improvement on wear resistance. A set of drag bits was built with each of these materials and sent to a tunnel jobsite for a final assessment of the performance in the field (in progress).
EDO: A software programme that detects over-excavation situations at the earliest possible stage, raising an alarm in such conditions thus providing ample time to act, prevent further evolution of the cause, and avoid the occurrence of adverse consequences on structures and utilities. EDO is based on advanced statistical processing of real time data from the belt conveyor weight scales. The software was tested successfully against real data and simulated data, and will be integrated as a prototype in a tunnel project in the UK, in 2017.
Upgraded DAT: The Decision Aids for Tunnelling (DAT) is a simulation software programme that computes probabilistically the distribution of construction time and cost of tunnel projects in function of geological and construction related uncertainties. A new version of the DAT was developed in the course of NeTTUN, with the objective to reduce the influence of the human factor on the simulation results. To this aim dedicated databases for risk factors and tunnelling performance were added, and an expert system implemented in order to help the user generate unbiased input data. Through new added functionalities, the whole programmed was extended to a complete decision-support tool and is available as a licence or through simulation services.
PADTUN: A software programme supporting the decision-making process concerning tunnel diagnosis, and assisting tunnel maintenance experts in defining the appropriate strategy in terms of inspections and maintenance operations. This programme is now in operation at SNCF.
The NeTTUN beneficiary have planned to pursue their work together through further testing, industrialisation, and commercialisation of the developed products and services, which will benefit the whole sector of tunnelling.
Project Context and Objectives:
4.1.2.1 Context
Significant progress has been made in the recent years in the execution of tunnel projects, reaching new challenges in terms of length, difficult ground or rock characteristics, location in dense urban areas and the vicinity of other tunnels and structures, depth below water and related hydrostatic pressure. However there were in 2010 when the Call for Proposals was issued by the European Commission – and still exist today – remaining real and concrete problems that consistently affect mechanised tunnelling, which in turn is seen as lacking control over risks and costs.
Tunnelling is the area of the construction industry that has been most affected by major losses. Besides property losses often in the two-digit million range, third-party liability losses have also been high, and numerous people have lost their lives.
Achieving and demonstrating full control over tunnel projects in terms of risks and costs is a key driver to the expansion of world-wide infrastructure construction works that would contribute in solving a number of major concerns: improve the safety of lives of citizens, reduce air and soil pollution, give access to potable water, etc. Many European companies are or could be involved in such infrastructure projects and developing their expertise further is needed.
4.1.2.2 Main objectives
The NeTTUN project covers the 4 research themes listed in the Work Programme of Call “NMP.2011.4.0-2 Advanced underground technologies for intelligent mining and for inspection, maintenance and excavation”, and targets its three technical objectives (i), (ii) and (iii), as summarised in tables a) and b):
nr. Research theme
1 Ground prediction ahead of the excavation front
2 Reduction of risks (on health, safety and the environment), through the automation of tasks
3 Assessment of risks on surrounding structures, through modelling and simulation; costs and risks models; monitoring of tunnel and/or surrounding structures
4 Inspection and maintenance of underground structures
nr. Objectives
i Productivity increase by 20% minimum
ii Underground operations with zero impact on surroundings
iii Safe underground working and operating environments
iv Strengthened competitiveness of European industry
v Increased sustainable access to underground resources
Table a) and b) – Research themes and project objectives
Objective (iv) is a global objective that is achieved as described in §XX3.
Objective (v) is specific to the mining industry and not relevant for tunnels
The project will achieve industrial integration and implementation of solutions with proven feasibility.
The main scientific and technical objectives of NeTTUN were:
• Develop an advanced ground prediction system for TBM, integrating concurrent approaches and performing an automated diagnosis, that is applicable to soft grounds and lined tunnels
• Robotise the most risky maintenance operations in a TBM (exchange of disc cutters and drag bits in the cutter head)
• Manufacture improved TBM drag bits with extended lifetime
• Improve an existing Decision Aid Software in the modelling of global risks on tunnel projects
• Enhance the performance of modelling ground movements due to mechanised tunnelling in soft grounds
• Extend the monitoring capabilities in the vicinity of a tunnel work site
• Develop advanced and practical method to design and implement measures protecting surrounding structures from the adverse effects of TBM tunnelling
• Develop advanced TBM on-board monitoring in order to detect over-excavation situations
• Develop an expert system supporting human experts in the strategy of tunnel maintenance
The project was designed so that the technology developments cooperate in achieving the objectives defined by the NMP work programme. The four research themes interact logically in forming a complete project, and are not parallel developments led independently.
The project is architectured in 12 Work Packages.
Project Results:
1 TULIPS Ground prediction system
TBMs progress blind in the ground. Geotechnical accidents and more generally “obstacles” that can negatively impact the TBM progress are anticipated based on preliminary site investigations such as boreholes from the surface. In many instances, such as densely populated and built areas, this information is far from being sufficient. Encountering such obstacles can result in major technical problems such as TBM damage but also variable degrees of ground settlements ranging up to open sky sinkholes, with consequences such as the damage to surrounding structures (buildings, other tunnels, etc.) or even collapse of buildings.
The standard way for obtaining information about the ground ahead is based on boreholes made from the TBM shield; this approach is very slow and severely impacts the TBM availability. Alternative means for predicting the ground ahead of the excavation face have been expected by contractors and project owners. Although several systems have been designed, most of them are focused on tunnels in rock and perform poorly in soft ground lined tunnels. This is the target of TULIPS, the system developed within the NeTTUN project.
The requirements for TULIPS can be summarised as follows:
• Applicable to soft grounds and lined tunnels
• Operate the system in hidden time vs. the TBM operation
• Integrated inside TBM
• Fully automatic process providing readily usable data
• Fast, frequent and effective look ahead during boring
The approach consists in
• Using complementary physical approaches to capture information on the ground ahead
• Combining the data in the most efficient way through data fusion techniques
In its first implementation, and as a target for NeTTUN, data is provided by a radar subsystem and a seismic subsystem. Radar acquisition is done as a circular scan (continuous rotation for at least one or complete turn), with as many antennas as possible along the cutter head radius. Seismic acquisition is done as fixed images, by positioning the cutter head at given angular positions, for example every 60°, using one (or possibly several) sources and several sensors aligned along a same cutter head diameter. Implementing a 0.5 m pitch between sensors allows reaching the desired resolution.
The process is shown in Figure 2 and is described here:
• Data is captured during ring erection, and is associated to a position of the TBM
• Seismic data is inverted (Full Wave Inversion) automatically , which provides a focused image ready for further processing
• Both the radar data and the inverted seismic image are processed in order to detect features, i.e. areas of the image that could correspond to geotechnical events; these expected events are referenced to a position is the ground ahead
• The system relies on the geological context and the knowledge database to indicate the probable nature of each expected event; this output is also submitted to a geo-expert who can confirm or reject the results of the automatic interpretation process
• When the TBM has progressed to the next ring, a new acquisition/feature detection cycle is executed; through a process of tracking, the system confirms the position and characteristics of each expected event
• Once the TBM has passed through the position of an expected event, the geo-expert can confirm or modify the information related to the event; this is stored in the knowledge database.
Figure 2 – TULIPS data flow and process
A complete radar subsystem has been designed, starting with the design of 2 antennas, one for short range detection, one for long range. Although preliminary tests were quite promising it was eventually found during the prototype field tests that these designs would not reach the targeted performance. Therefore a new design was launched, which proved successful and applied to both the short and long range antennas through their implementation in a Planar Antenna Module. A complete electronic system was concurrently designed and built, and the related control software developed and integrated. Subsystem tests were performed in the lab, in a controlled (known) environment and gave results that met the requirements.
A complete seismic subsystem has been designed, based on S-waves, which are best suited to soft grounds. The hardware was developed, comprising a new seismic source (vibrator) reaching down to low frequencies and a sensor, both designed to be integrated in a TBM cutter head, i.e. deployed against the excavation face during the measurements (cutter head stopped) and retracted at any time the cutter head is rotating (mainly during excavation). Both these components were tested in the lab and showed that they met the requirements in terms of frequency spectrum but also robustness and reliability.
A test site was designed and built in Eindhoven (the Netherlands) in order to assess the performance of each subsystem, i.e. the capability of detecting various geotechnical scenarios and buried objects that were buried in known positions. A “TBM cutter head mockup” was designed and built, allowing a simple radar circular scan. After each complete acquisition of radar and seismic data, a layer of ground was removed from the surface, simulating the advance of the TBM, and a new data acquisition cycle performed.
An automatic Full Wave Inversion software was developed, delivering focused images from the data without any human intervention. This process requires that low frequencies be present in the data, hence the special characteristics of the source and sensors. The software was first tested and validated against simulated data and then optimised in terms of execution time, in order to meet the “real time” constraints of a TBM advancing from one ring to the next. It was eventually applied to the Eindhoven data. The target of a 30 minute process was proven to be achievable using a high-end available multi-core computer, devices that are currently offered at a reasonable price.
The complete TULIPS system software was developed, comprising:
• The data Analysis and Fusion software – for automatic interpretation and fusion of multiple sets of imaging data e.g. from different modalities (radar and seismic with different configurations)
• The Visualisation software – for viewing and interacting with raw and processed imaging data
All of these functionalities are available through a single end-user interface, implemented as a web-based application allowing users to be working anywhere on the tunnel jobsite or from a remote office. The sensor data itself can be viewed in 2D or 3D, onto which the detected features can be added. In addition the interface allows the user to display the longitudinal profile and geology of the tunnel together with the corresponding TBM position and the position and characteristics of predicted and past events, as well as TBM parameters that are representative of the current geological conditions. The user can navigate along the already excavated portion of the tunnel and display the stored images and associated features and events data. Altogether this forms a complete workstation with all information at hand for the geo-expert to perform a diagnosis of the situation ahead of the TBM.
The complete system was integrated as shown in Figure 3. A dedicated rugged “TULIPS box” was designed and built in order to house all the sensing system components that must fit inside the cutter head and ensure the adequate protection against water, pressure, vibrations and shocks. Communication between the TULIPS box and the TBM is done by a network through a slip ring also providing power, while the rotary joint allows to operate the hydraulic actuators for deploying./retracting the seismic components. This network also reaches the surface offices where the inversion computer and the end-user interface can be installed. The TULIPS-TBM controller communication was developed and tested, providing a “press button” start of a complete acquisition and data processing sequence.
Figure 3 – TULIPS system architecture with radar devices (in orange), seismic devices (light blue), and common devices (green)
Feasibility studies were made for the implementation of a full size TULIPS on various types and diameters of TBMs, which were successful although it is as expected more difficult on small TBM diameters, i.e. in the 6 to 7 m range.
The NFM contract for the delivery of the Lyon-Turin TBM (hard rock single shield, 11.25 m diameter) gave the opportunity to install a radar-only TULIPS. The antenna module was redesigned based on the prototype design in order to incorporate all the mechanical constraints, while the cutter head was specifically designed to accommodate two such antenna modules and all the related cabling and protective ducts (see Figure 4). The system was fully tested in the factory and on the jobsite after rebuilding the TBM. However the extreme conditions encountered at the very beginning of the drive (unstable excavation front with rock falling onto the antennas) led to damaging the antennas and ultimately to their destruction – unfortunately no real world image was obtained before this happened. However this experience has taught us a lot on the way to reinforce the system and achieve a reliable operation in the very tough environment of a TBM at work.
Figure 4 – The Lyon-Turin TBM at the NFM factory.
The housings for the two antenna modules can be seen.
2 Maintenance robot
Drag bits and disc cutters, which equip a TBM cutter head, are submitted to intense abrasive wear and their replacement is required along the progress of the TBM. This is done by operators who must enter the excavation chamber and the cutter head. In the EPB or Slurry machines used in soft grounds, access to the front part of the machine is complex and dangerous because this area is pressurised at several bars. This implies hyperbaric work with compression and decompression of the workers, with consequential health problems such as Decompression Illness. Such procedures are time consuming although the duration of work in the pressurised area is very limited, similarly to divers timetables. Operations are slow, in the order of several hours each time, during which the excavation is stopped.
Alternatives to hyperbaric interventions have been developed but they exhibit significant drawbacks because of their impact on the TBM design and cost. Therefore there is a high expectation from the contractors and the tunnelling industry for a safe and efficient approach for the replacement operations of disc cutters and drag bits. In NeTTUN we have decided to address the problem using a robot – HECTOR – as described below.
The concept is similar to the human interventions, except that a robot is used, which can operate at the excavation chamber counter-pressure, thereby eliminating all idle time corresponding to compression/decompression. During excavation, HECTOR is stored in a chamber located at the top of the TBM. When an intervention is ordered: a) the excavation chamber is partly emptied of its muck, which is replaced by compressed air; b) the cutter head is positioned so that the tool to be changed is in the upper vertical radius; c) the front gate opens and the robot extends itself in order to reach the cutter head and tool to replace; d) the robot gripper hols and releases the tool; e) the robot folds back into its storage chamber and positions the worn tool in a buffer; f) the robot picks a new tool from the buffer and operates in the inverse manner to mount it and secure it on the cutter head.
The HECTOR robot was designed completely from scratch in order to achieve the smallest possible volume and footprint, implementing only the movements that are required for the operations while ensuring the largest possible range of action. Safety in operation was a key driver which resulted in keeping the tool either securely fixed on the cutter head or held by the gripper. New disc cutter and drag bit locking mechanisms were designed to comply with this requirement and allow mounting and dismounting with a single “hand”, i.e. a dedicated gripper.
The robotic arm was designed and built together with the 8 hydraulic actuators and control circuitry. The hydraulic devices are themselves driven by a National Instruments controller hosting the low-level software.
Both a disc cutter and a drag bit grippers were designed. According to the cycle of operations planned, the appropriate gripper can be mounted on a standard interface at the end of the arm. Each gripper includes the necessary hydraulic wrenches to operate the locking mechanism while holding the disc cutter or drag bit. Each gripper also contains the electronic board that was purposely designed to host the corresponding local low-level software.
A test bench was designed and manufactured, including a mock-up of a cutter head section with several housings for disc cutters and drag bits. The system under tests at NFM is shown in Figure 5.
Figure 5 – Robot system under tests
The software architecture is shown in Figure 6 . It comprises a low-level part, which directly controls the actuators of the arm and gripper, and a high-level part that calculates the trajectory in real-time and sends orders to the low-level software so that the trajectory is effectively followed. The High-level part is equipped with the Human-Machine Interface using a 3D mouse, keyboard and screen, allowing for various modes of operation such as “manual” in the TBM frame of reference (used to position the gripper close to the tool to handle), “manual” in the gripper frame of reference (used once the gripper is in the vicinity of the tool to handle); or “automatic”, with predefined standard positions (e.g. “home”) that the robot then reaches automatically.
Figure 6 – Architecture of the robot control system
Integration of hardware and software was performed successfully at the NFM facilities in Le Creusot, with the participation of all concerned partners.
Feasibility studies were performed for the integration of the robot in various types and diameters of TBMs and resulted in a decision matrix. This confirmed that the robot is suited for TBMs above 8.75 m (with some impact on the TBM design) and can be integrated with close to zero impact in TBMs of above 10.5 m in diameter.
There was however no opportunity during the course of the NeTTUN project to design and deliver a TBM equipped with the robot, in spite of all efforts made (dissemination through papers, conferences) and direct contact with end-users e.g. those who participated in the market survey (Exploitation).
At the end of NeTTUN we have identified the technical points that need improvement and those that were not completed, such as the integration and tests with the drag bit gripper, the video feedback to the operator, and the cleaning system using high-pressure water jets.
3 Improved drag bits
TBM tunnelling in soft or mixed ground relies on the efficiency of “drag bits”, which scrape the soil off the excavation face. These tools suffer from different types of wear (abrasion, impact, fatigue...) which vary according to the ground abrasivity and hardness and TBM operating conditions. Replacing these tools generates heavy work for the TBM maintenance crew, with associated risks when working in hyperbaric conditions and costs – of personnel, replacement parts, and idle time. It is therefore of great interest to improve the lifetime of drag bits in order to reduce these problems (human risks; jobsite cost and tunnelling works duration). This is what NeTTUN has addressed in WP7.
Two aspects of the problem were addressed simultaneously as follows: a) improved hard materials, implemented as inserts that constitute the “active” parts in contact with the ground; b) optimised shape of the drag bit including the position of the hard material relative to the standard steel body. Work started with an assessment of the situation on wear, with the analysis of worn drag bits from TBMs on various jobsites: this phase is common to both tasks a) and b) as it provides information on the type of wear and a “map” of the wear intensity. This study showed that the wear of the carbide inserts located on the drag bits is due to loss of tungsten carbide grains, scratches, and fractures.
• Improved hard materials
Work on the development of new materials was targeted at obtaining an improvement of 20% minimum of the wear resistance as compared to a reference material (used for the current NFM drag bits). This development followed parallel approaches in terms of process and composition. The initial step was however to define the development and test plan, which led to the design and construction of tribometers allowing the comparison of wear between different materials and the reference. An abrasive wear tribometer was therefore made available at ECL-LTDS, while the one at TU Tallinn was designed to operate in a combined abrasion+impact mode.
Three families of new materials were developed as defined in the roadmap: family #1, based on conventional sintering with improved chemical composition in order to increase hardness and toughness; family #2 based on reactive sintering, producing bi-composites or double doped cermets; family #3 was decided based on the results obtained with materials #1 and #2, as a combination of the approaches giving the best results with materials #1 and #2 (optimisation of chemical composition (additives type and concentration) and production process, in conventional sintering or reactive sintering (straight route and mixed route)). At each step, the results of the previous step were used to orient the development of the next material.
The results show that the test method strongly influences the results – some materials behave very well in pure abrasive tests and poorly in situations with impact. This led to further research on the relation between wear resistance and mechanical properties (mainly hardness and toughness), in order to define a law giving a prediction of a material’s wear resistance.
Three materials behaved better than the target in both test conditions and were selected for field tests. Two materials were selected as reaching the target in impact conditions while having relatively poor performance in pure abrasion; one was selected as reaching the target in pure abrasion while having poor performance in impact conditions. In total, 6 new promising materials were selected for field tests.
Figure 7 – Prototype inserts of new materials
Inserts made of these 6 materials were produced at TU Tallinn (see Figure 7) and used to prepare a set of 48 drag bits that were sent to a jobsite in Singapore for field tests. These drag bits are identical in shape and configuration to the standard ones initially mounted on the TBM – only the change of insert material is tested.
• Optimised shape of drag bit
The analysis of the forces induced on the global and local levels of a drag bit is a necessary step to understand the wear phenomena observed on the tools and to elaborate more effective tool geometries and more resistant materials.
This objective was addressed by modelling the material flow around a drag bit moving through the ground: Finite Element modelling (continuous media), Discrete Element modelling (granular media), and a comparison with analytical models were used – see Figure 8.
Figure 8 – Discrete Element Modelling of ground flow in the vicinity of a moving drag bit
In addition, a small scale physical model was developed to validate the numerical modelling results. Through this process, the stresses on the drag bits were determined; these results were used to produce a report giving "design rules" for an optimal drag bit geometry and for an optimal configuration of the cutter head.
• Field tests of new drag bits
The prototype drag bits were installed on one of the 6.6 m diameter slurry TBMs at work on the T216 stretch of the Thomson East Coast Line in Singapore, as shown in Figure 9.
Figure 9 – Prototype drag bits mounted on the T216 TBM
Excavation started in November 2016 and was still going on at the end of the project in February 2017. It was not possible to obtain from the contractor that intermediate inspections be performed before the end of the drive, therefore the wear analysis will be performed by the partners after the end of the project. Results will be available only after the end of NeTTUN. However the results will be communicated to the project partners and the EC.
4 Decision Aids for Tunnelling
Tunnelling is one specific field of civil engineering that is more than others inherently risky. Very costly infrastructure projects are planned but the conditions of their execution remains affected by a high degree of uncertainty. The first and most important factor is related to the geotechnical conditions: whatever the extent of site surveys, uncertainties remain as to the nature and structure of the ground in between the samples taken whether by core drilling or geophysical imaging. The second risk factor concerns the construction method, the actual performance of the equipment used, the skills of the people working on the project, etc. When having to decide in launching a project, decision-makers need to know what risk they take in terms of costs and delivery time. A significant step was made in this field with the development of the Decision Aids for Tunnelling software (DAT) in the 1980’s by Prof H. Einstein (MIT) and the EPFL. This software is based on a probabilistic approach and simulates the construction of the whole tunnel network, including shafts, adits, etc. and produces cost vs. time graphs (scattergrams). These results are however related to the quality and reliability of the input data provided by the user: it has been considered that DAT overly relied on human estimation of geological and constructive risks. Reducing the human factor in the DAT results was taken as a goal for the NeTTUN partners, in WP8.
NeTTUN addressed the problem by introducing actual feedback from previous experience; this approach was implemented as a set of two databases:
• A database of tunnelling risk factors that contains all possible events related with the geology and construction method such as “in limestone, the TBM may encounter a large karst”,
• A database of tunnelling performances that gives ranges of costs per meter and advance rates depending on the ground conditions and the TBM type when tunnelling in normal conditions.
In addition to the databases, an expert system was developed that assists the DAT user in creating realistic input files, by checking and/or suggesting data through the database, by calculating TBM performance based on rock parameters, or by applying expert rules (see Figure 10). At the end of NeTTUN, the following rules were implemented and available within the DAT: the "Penalty factors" model, three analytical models (RME, QTBM and FPI), a rough estimation for preliminary studies, and the link to the databases, including the handling of non-perfect matches in the TBM Performance Database.
A significant effort went into the filling of the databases, collecting – through the NeTTUN partners, other companies, literature survey, and tunnelling associations – information about past projects in terms of geology, ground characteristics, advance rate, geotechnical events found, etc.
Figure 10 – General architecture of the upgraded DAT and interactions with expert system
A new user interface was developed for more user-friendliness. Through other added functionalities, the DAT was extended to a complete decision-support tool dedicated to construction management, allowing the user to update the expected construction time and cost of different variants (see Figure 11), the impact of performing additional geological surveys, of switching to a different TBM maintenance scenario, etc. The software also gives possibilities to enter in more details such as analysing the influence of a particular zone, of specific ground parameters, etc. on the cost vs. time results.
Figure 11 – “Include New Data”, one of the added functionalities. This allows the user compare the time-cost scattergrams of 2 options of the same project
The NeTTUN “end-users” (construction companies OHL, Razel, Astaldi and engineering companies Systra and BG) performed tests based on real tunnels. Based on the results of these tests, proposals were made for corrections and improvement; after discussions a decision was made to implement or not the proposed changes. The final version of the DAT software released at the end of NeTTUN contains these last changes.
Although it had been planned to test DAT at various stages of actual tunnel projects, i.e. both at design phase and in construction phase, this was finally not achieved by lack of suitable projects within the consortium. Still a comparison was made between the pre-NeTTUN and final NeTTUN versions of the DAT. This was performed on the Lyon-Turin base tunnel, which had been simulated in 2013 and was simulated again at the end of 2016 – this time taking into account a new starting point for the drives in Italy, but also the back-analysis of the exploratory tunnels of Maddalena and, partially, of St-Martin-La-Porte. The NeTTUN version provided results in line with the time-position diagrams resulting from the deterministic design. It helped in underlining the main expected difficulties and the remaining uncertainties, as well as the positive effects (i.e. lower variability) due to the exploratory/access tunnels and the possibility of new drives.
5 Controlling the impact of tunnelling on surrounding structures
The growing demand for public and sustainable transport requires the construction of an increasing number of underground infrastructures. The many constraints and technical challenges associated to underground construction in urbanised areas often lead to high costs and long completion times. This is particularly true for cities characterised by a high density of population, significant archaeological heritage, and the presence of masonry structures of historical and monumental value, which are particularly sensitive to subsidence induced by excavation. It is therefore often necessary to adopt complex control systems of the excavation process, in order to achieve the maximum limitation of deformations, to devise intense monitoring schemes, and, where necessary, to implement techniques for the mitigation of the potential damage and the protection of the structures affected by excavation, with a significant increase in the construction costs. In this context, WP9 focused on the following objectives: (i) to enhance the performance of physical and numerical modelling of ground movements due to mechanised tunnelling in soft grounds, (ii) to study the impact of EPBS tunnel excavation on different types of structures, (iii) to extend the monitoring capabilities in the vicinity of a tunnel work site, (iv) to develop a method to provide early detection of over-excavations in TBM tunnelling, and (v) to evaluate effectiveness and provide a rational basis for the design and implementation of mitigation measures. These objectives were addressed with complementary approaches, using a combination of numerical analysis, laboratory-scale experiments, and in-situ data. The work was organised in 6 Tasks; the bulk of the work that was carried out falls under Tasks 9.1 (physical modelling), 9.3 and 9.6 (numerical modelling of tunnel excavation and of mitigation measures), and 9.4 (early detection of over-excavation), which substantially met all the target objectives of the NeTTUN project. On the other hand, although it was originally intended that the laboratory, numerical, and theoretical developments should be tested against the monitoring results collected on the work sites of Contract T3 of Line C of Roma Metro, due to a number of reasons independent of partners (financial, political and archaeological constraints) the construction works were delayed. This meant that it was not possible to complete all the field work originally planned under Tasks 9.2 (field monitoring) and 9.5 (field trials of mitigation measures). Although the bulk of the work that was promised was carried out, there are many areas in which further work is required, either to fully meet the research objectives that were set in the original DOW, or because activities that were not originally planned have emerged as worth of exploring, beyond the original scope of the project, as summarised below, together with the main achievements of WP9.
• Physical modelling of TBM tunnelling
An extensive experimental campaign was conducted by researchers from ENTPE, using a 1g reduced-scale model of an EPB shield, to investigate the impact of tunnelling on neighbouring structures founded on piles and the efficiency of physical barriers on the limitation of surface displacements. Significant modifications of the original experimental set-up were undertaken in order to upgrade its geometric dimensions and its control system so that the experimental campaign could be carried out. Considerable financial and work efforts were also devoted to the design and construction of model structures (piles, groups of piles, and barriers) and of ancillary equipment to fill and empty the sand box, load and install the piles and monitor machine parameters, ground and structural response during TBM advancement. A total of 14 tests were carried out under the NeTTUN project, including one preliminary tests to verify the good functioning of the new equipment, two tests in virgin soil under different tunnelling conditions to be used as a reference, eight tests of the interaction of TBM tunnelling with piled foundations, and three tests of the interaction of TBM tunnelling with barriers. Tests carried out with the 1g EPBS reduced scale model of ENTPE have shown the ability of this device in its new configuration to simulate closely the principal operations of a real EPBS and to reproduce its 3D excavation process. The tests performed with a group of four (2x2) piles installed in the vicinity of the tunnel have shown that the soil-piles-TBM interaction is characterised by a redistribution of the working load in the piles of the group during TBM advance, which is partly due to the variations of ground stresses around the piles caused by the TBM advance. Additional tests should be carried out beyond the end of the NeTTUN project to study the influence of different factors. In particular, the effects of the distance between the piles and the influence of the conditions of contact between the piles head and the pile cap should be explored.
The three tests carried out using various models of diaphragm walls (see Figure 12) permitted to study the efficiency of these physical barriers to limit surface displacements. During these tests, only the wall depth relative to the tunnel position was varied.
Figure 12 – Model of diaphragm wall used during the tests – shown in position during container filling
The results obtained show that the three barriers used strongly limit the vertical displacements of the soil outside the barrier. In particular, the tests show that the magnitude of the soil settlements near the barrier decreases with the length of the wall. Additional tests will be carried out beyond the end of the NeTTUN project to analyse the influence of the weight of the barrier, explore the relevance of lightened barriers, and adjacent pile barriers, which represent, from a practical point of view, a non-negligible economic gain. The NeTTUN project permitted to collect a large number of experimental results obtained under various conditions of excavation and with different neighbouring structures (single piles, group of piles, mitigation barriers). These can provide a reliable support to calibrate numerical simulation models dealing with tunnel-soil-piles interactions.
• Numerical modelling
The numerical modelling working team, consisting of researchers from URO2 and NTUA, cooperated to develop a reliable Lagrangian 3D FE procedure to model TBM tunnelling. This included the realistic simulation of the main physical phenomena occurring during EPB shield tunnelling, such as e.g. the application of a variable support pressure at the face, the presence of a physical gap between the excavation boundary and the permanent lining (due to over-cut at the cutter-head, shield tapering, and installation of the segmental lining within the shield) and the execution of tail void grouting. The interaction between the shield and the ground was simulated introducing appropriate normal and tangential contact laws, while tail void grouting was modelled using a time-dependent setting law for the grout and an initial stress state in the grout elements to represent the injection pressure. Since a significant proportion of ground deformation due to tunnelling in fine-grained soils is associated with consolidation settlements, fully coupled consolidation analyses were carried out, with realistic rates of advancement of the shield.
Figure 13 – Finite element mesh employed for greenfield simulations
Alternative approaches for the simulation of mechanised tunnelling based on an Arbitrary Lagrangian Eulerian (ALE) description of the tunnelling process were also explored. The main aim of this work was the prediction of the volume loss, for given soil properties, geometry, and operational parameters of the TBM. In the ALE approach the advance of the shield is simulated as a continuous process and the support pressure applied at the tunnel face is a result of the analysis, depending on the ratio of the mass of excavated soil in unit time to the rate of extraction of the muck from the pressure chamber. At the moment, the applicability of the ALE technique has only been assessed against the results of the small scale physical model of mechanised tunnelling carried out at ENTPE. Also, the degree of detail and the number of the analyses carried out so far with the ALE approach were limited due to the huge computational effort demanded by the specific implementation of the ALE method. Many extensions of the numerical work are planned or currently under way, such as adopting the distribution of face support pressure resulting from the application of the ALE technique in Lagrangian analyses, and conversely, to apply some of the features of the Lagrangian analyses, such as e.g. the over-cut at the cutter-head, shield tapering, and tail void grouting, in the Eulerian ones. Numerical analyses of the results of physical models (virgin soil, piles and barriers) are currently being improved through the adoption of more advanced constitutive models for the soil, including yielding for small stress increments, hardening and evolving dilatancy. In principle, the Lagrangian numerical strategy that was developed allows modelling the complete system of ground/TBM/monument, from which the impact of the tunnelling works can be determined. However, transposition of the model to real size was achieved only with reference to the passage of the TBMs at the AMA site, both in greenfield and in the vicinity of the trial barrier. Due to the delays on the Line C works, the back analysis of the interaction of tunnelling with the Aurelian Wall will be carried out after the end of NeTTUN, because the details of the construction work are not entirely defined at this time. Although numerical results are very promising at this stage, the comparison with real field data is crucial for the ultimate validation of the proposed methods.
• Field work
The responsible partner for field work was METC, who designed the monitoring schemes, and acquired and installed the field instrumentation at the AMA site and on the Mura Aureliane at Porta Metronia, to monitor the ground response during passage of the TBMs. Background reading of the instrumentation is currently carried out on a regular basis. METC also performed all the necessary preliminary actions to carry out field trials of mitigation measures, including both compensation grouting and barriers. A barrier consisting of 48 adjacent concrete piles was constructed in June 2016 in a dedicated section of the AMA site. Monitoring of ground response to passage of the TBM in the vicinity of the barrier will include the measurement of vertical and horizontal displacements of a number of points both at surface and subsurface by precision levelling, extensometers and inclinometers, whose installation was completed by July 2016. The field trial of compensation grouting includes a dummy foundation and a monitoring system of surface and subsurface displacements (liquid levelling points, reference points for precision levelling, inclinometers and extensometers). Because the made ground contains potentially important artefacts/structures, a temporary grouting shaft, consisting of 26 concrete piles and 26+26 bentonite piles, was constructed from which the grout sleeve tubes will be inserted.
Figure 14 – Plan view and section of the compensation grouting field trials at Porta Metronia
Construction of the piles supporting the shaft was completed in November 2016 and, at present, excavation with archaeological methods has reached a depth of about 7 m below ground level. Due to the delays on the construction of the line, all the construction dates have had to be revised numerous times during the project and it has been clear from some time that the TBMs will pass neither through the AMA site nor in the vicinity of Porta Metronia within the scope of the NeTTUN project. Therefore, permission was asked to, and granted by, the EC to proceed with the acquisition and installation of the instrumentation, even if the relevant data would be collected after the end of the project. It is intended that METC will collect the data from the passage of the TBMs through the control sections, conduct the field trials and make the results available to the numerical work group and to all other partners of the NeTTUN consortium.
• Early detection of over-excavation (EDO) system
NTUA were in charge of designing and testing a system for the Early Detection of Over-Excavation (EDO) during TBM tunnelling, which can reduce the adverse consequences of significant over-excavations in an urban environment, including damage to surface structures and utilities due to excessive ground settlements. The system aims to detect over-excavations when they occur, rather than detecting the effects of over-excavations (typically ground surface settlements) as is currently done in classical monitoring. Thus, the proposed system can detect over-excavations early, before they evolve to a degree that can cause appreciable settlements and damage to structures. Another advantage of the proposed system is its simplicity, as it uses muck weight measurements on the conveyor belt (which are performed in practically all TBMs) and processes them analytically by software. Appreciable over-excavations may be caused by ground failure ahead of the TBM face (due to poor ground conditions, too low face pressure of the EPB system, excessive groundwater gradient, or a combination of the above), providing more excavated material than the anticipated mass. The EDO system incorporates a Muck Weight Measurement and Assessment Sub-System that exploits muck weight measurements, typically measured by digital scales on the conveyor belt and recorded by the TBM control system in real time. The coefficient of variation of the measurements depends on the tunnel diameter, the TBM type, the conveyor belt, the weight scale, and of course the ground itself. Thus, the coefficient of variation is a parameter which is calculated on the basis of initial measurements of muck weights for a TBM advance of a few segments and subsequently used in future assessments. In order to assess the effectiveness of the EDO system, a thorough investigation was conducted using actual (recorded) datasets of weight measurements from several TBM projects – Figure 15). An extensive numerical investigation was also carried out to quantify the expected face take and the resulting theoretical over-excavation, resulting in charts that are integrated into the EDO system.
Figure 15 – EDO system warnings for fictitious over-excavations added to real data
The next step in the development of the method is the implementation of the EDO system on a real TBM during operation, which will provide valuable experience and shed new light to various sources of uncertainty that appear when stored data are examined, mainly related to the origin of the variability that the weight measurements exhibit and how this is related to various operations taking place.
NTUA remains fully committed to the objective of integrating and testing the method in real time in the near future by investing the necessary man-power even after the completion of the NeTTUN programme. In fact, in close collaboration with NFM it is planned to integrate the system in two EPB machines currently manufactured by NFM and expected to be launched soon (Spring 2017) in the extension of the Northern Line of the London Tube in the UK.
6 Tunnel maintenance strategy
Tunnel maintenance requires complex decision making, which involves pathology diagnosis and risk assessment, to ensure full safety while optimising the maintenance and repair costs. In the digital era, a plethora of data about the tunnel conditions, including both periodic tunnel inspections and contextual factors affecting the condition of tunnels, is becoming available. Inspecting this data to identify notable combinations of disorders and influencing factors, to detect possible pathologies, and to assess the possible risk of further deterioration is highly complex and requires extensive expertise. The complexity is further aggravated by a range of factors, such as infrastructure aging, climate change, and rapidly increasing urbanisation. Moreover, this complex decision process with high societal and economic impact is often prone to subjectivity and poorly scales across use cases and domains.
These challenges were addressed in NeTTUN Work Package 10 by an interdisciplinary team of tunnel experts and knowledge engineers. A decision support system that assists with pathology diagnosis and assessment of tunnels (PADTUN) was developed. It involves two components: (1) Pathology assessment and (2) Urgency finder (see Figure 16).
Figure 16 – PADTUN system architecture
1st component: Pathology assessment assists tunnel experts in making decisions about a tunnel’s condition with respect to its disorders and diagnosis influencing factors. It uses semantic web technologies for knowledge capture, representation, and reasoning. Ontologies, which represent the main concepts and relations associated with pathology assessment, were developed to capture the existing decision process concerning the maintenance of tunnels, which provides a knowledge model for automated decision support. The PADTUN ontologies, which are represented in the web ontology language (OWL) that follows internationally established standards for knowledge reuse, are the first ever ontologies developed for the domain of tunnel diagnosis and maintenance. Pathology detection is based on tunnel disorders and influencing factors. The ontological knowledge (which includes 125 concepts, 49 relations, 590 individuals, and 3981 axioms) was provided by domain experts at SNCF, further extended by following tunnel maintenance guides, and validated by a domain expert from Swiss Rail. Extracts from the PADTUN ontologies are given in Figure 17.
Tunnel inspection data are linked to the PADTUN ontologies which enables applying the knowledge to identify tunnel regions where specific pathologies are detected (see Figure 18) for an example of a tunnel diagnosis with PADTUN).
2nd component: Urgency finder is crucial for tunnel maintenance because repairs are expensive and affect the operational efficiency of the tunnel. To predict the risk factor and the resulting repair urgency in the tunnel maintenance planning, PADTUN utilises supervised machine learning methods that learn from ground truth data of past human decisions. The data on the basis of which the PADTUN urgency finder component is developed consists of 47 tunnels of varying lengths, a total of approximately 60 tunnel kilometres. The data included periodic survey inspections for multiple years, as well as other data such as the method of construction of the tunnel. Expert annotations were collected for each 10m tunnel segment for each survey as to the degree of repair urgency. These annotations provided the ground truth for both model training and evaluation. Several machine learning methods, including Decision Trees, Random Forests, Support Vector Machines, were implemented and assessed. Decision Trees were selected as the best performing method with sufficient transparency to examine the system’s suggestions for the urgency of repair for each tunnel portion.
Figure 17 – Extract from the PADTUN ontologies: associations between a pathology and disorders (top) and association between pathologies and influencing factors (bottom).
Figure 18 – PADTUN web interface for tunnel diagnosis: the tunnel is divided by portions of 10 meters which are numbered (column headings); for each portion pathologies are detected (main rectangle); and the overall urgency repair indicated (first row of the main rectangle).
PADTUN was evaluated on a database of SNCF tunnels, and an example tunnel from Swiss Rail. The results were very good with only 3% of false negatives (actual pathologies that are not detected). Machine learning from tunnel data was applied to investigate the reasons for misdiagnosis. It was found that the misdiagnosed pathologies were rare occurrences of pathologies, with the exception of “creep” which had false positives due to an overlooked “overburden” constraint. PADTUN was deployed on a virtual machine on a computer at SNCF for routine tests with tunnel data. Subsequently, with the aid of a developer’s guide, which was provided to SNCF and guidance from personnel from Leeds, a native installation on an SNCF computer has been effected.
In a broader context, PADTUN illustrates a step change in combining knowledge engineering and machine learning to support complex decision making related to sustainable urban infrastructures. The approach is applicable not only to tunnels, but also to other transport constructions (e.g. bridges, roads, tracks). Possible extensions in this direction are being explored considering both research and innovation aspects.
Potential Impact:
1 Impact
The impact of the NeTTUN project is at several levels:
• NeTTUN was mainly aimed at improving the safety of specialised workers and, beyond that, of all European citizen users of underground structures, especially in urban areas.
• NeTTUN represents a significant step forward in the goal to fully automate TBMs and tunnelling works.
• It contributed to improving European cohesion. Nine European countries have participated in the project: Estonia, France, Germany, Greece, Italy, Spain, Switzerland, The Netherlands, and United Kingdom.
• It promotes the collaboration between universities, research centres, and industries (systems manufacturing and tunnel construction and maintenance) thereby combining the needs of the industries with the knowledge of researchers.
• It included the participation of one organisation from one most recent Member State: Estonia.
• The consortium included the participation of five SMEs: Sialtec, CISTEME, MI Partners, Geo2X, and INCAS Partners.
• As per the EC Strategy 2020, the project has already been of benefit to Education as it was presented at courses such as the Specializing Master Course on Tunnelling and Tunnel Boring Machines (from the Politecnico di Torino).
• The project indeed contributes to all the expected impacts listed in the Work Programme as further detailed below.
1.1 A project with enlarged business perspectives, delivering productivity increase
• NeTTUN extends the range of mechanised tunnelling, which is reported as the safest tunnelling method: at least 55% of recorded collapses occur with conventional methods.
• The project provides directly applicable answers to the real and concrete problems that consistently affect tunnelling with TBMs. TBMs currently operate in a “blind” environment without clear knowledge of what lies ahead of them at the face. With the development of the NeTTUN ground prediction system, combining radar, seismic, and data fusion expertise in a way that has never been done before, mechanised tunnelling will be able to improve project delivery times in a safer environment with real images and understanding of what lies ahead of the TBM.
• Project Risk Modelling has a highly positive consequence on the economics of tunnel projects. The improvements brought to the pre-existing DAT software will help decision-makers in planning projects and allow looking for alternatives with time and costs savings. Delivering projects on time and within budget will of course increase the competitiveness of the European tunnelling industry on the world stage, see below.
1.2 A project that delivers underground operations with zero impact on surroundings
• Nothing will have greater significance than demonstrating on the NeTTUN project the successful tunnelling under the Coliseum of Rome with zero impact on this UNESCO world heritage ancient site. All the knowledge is now available for a successful construction of the risky stretches o Line C. Dissemination of this result will carry great impact for Exploitation of this knowledge beyond the end of NeTTUN.
• The future of Tunnelling will demand more complex, longer and deeper tunnels and this is where the NeTTUN results will be most efficiently implemented. In such projects, government agencies are becoming more concerned with improving the tunnelling construction Health & Safety environment for workers as well as public safety and surrounding structures’ integrity. Extremely dense areas such as Singapore and Hong Kong have plans for large projects in which the level of technical difficulty is exacerbated by the complex ground conditions. For such projects the NeTTUN results in Ground Prediction, Robotised Maintenance, and Control of Impact of Tunnelling on Surrounding Structures are now ready to contribute in meeting this demand.
1.3 A project answering societal needs significantly by improving safety
This is the major concern with tunnelling activities and operational use.
• The objective of the NeTTUN project was to ensure a safer underground working and operating environment. The project, with the development of a TBM maintenance robot, better drag bits requiring less maintenance interventions, ground prediction system providing information about geotechnical issues ahead, prediction and monitoring of ground movements, has delivered key elements to achieve this goal.
• In a study of Notable Tunnel Failure Case Histories of 39 overseas cases which occurred from 1964 to 2014 in the last decade, at least 36 people lost their lives as a result of tunnel collapses, building evacuations took place involving hundreds of people, many buildings and major roads were damaged and closed, and tens of thousands of people were affected. A large part of these accidents would have been avoided by using advanced modelling of ground movements and protection measures as developed within NeTTUN and applied to the metro Line C construction in Rome. Implementing reliable ground prediction systems and improved use of the excavated material measurement such as the EDO, both developed within NeTTUN, would also have contributed in avoiding such accidents.
• Underground workers are exposed to personal risks, which is why a significant effort was put in the development of an efficient robotised system that performs maintenance operations in the most dangerous area of the TBM instead of humans, in more than 80% of the cases.
1.4 A project that strengthens competitiveness of European industry
1.4.1 Through Science, Research and Education
• Bringing competing academic laboratories together to share ideas and work on a common goal, such as:
o IRCTR at TU Delft and XLIM at the University of Limoges, teaming in the development of the best possible radar for ground prediction;
o ENTPE in France and the Universities of Roma and Athens working together on the management of geotechnical risks;
o LTDS at Ecole Centrale de Lyon and TU Tallinn developing advanced excavation tools and materials in a common effort.
• Improving partnership between industry and research for example:
o NFM will draw on its existing regional relationships with research organisations such as ENTPE, LTDS, INSA, ENISE, and expand it to a European level
o IDS will benefit from the best academic expertise in radar from TU Delft and XLIM for the creation of new advanced products
o OHL, Razel and the Metro C consortium partners such as Astaldi will develop partnerships with research laboratories that can be sustained on a longer term in applied science fields such as the wear of materials, geotechnical modelling or ground prediction.
o SNCF will be introduced to advanced mathematics and data fusion by working with University of Leeds
• Expanding research activities from a national level to a European level
o University of Leeds participation in NeTTUN will include building on the existing “Mapping the Underworld” projects and extending the techniques in a new domain with other European research beneficiaries (TU Delft, XLIM)
o Produce experimental lab equipment and software models that can be used for subsequent research works concerning geotechnics, tunnelling and material science.
o Attract new students to pursue graduate studies through NeTTUN’s Dissemination activities. There is a particular focus by the NeTTUN beneficiaries in Lyon – NFM, ENTPE, ECL – to attract new engineering students into the Tunnelling industry. This has proven to be very effective (and rewarding): 2 of ECL former students who worked with NFM during their last year project (wear of drag bits) pursued their career in the tunnelling area. They have been hired by major companies and currently work on major underground projects (Montreal metro and Grand Paris Express).
1.4.2 In Industry
• Challenging tunnelling works are currently a special expertise of European civil engineering companies. European civil engineers are commonly called in for their excellence, for projects all around the world (recent examples are Miami harbour, Hong-Kong high-speed rail link). The state of the art advances achieved by the NeTTUN project in Ground Prediction, Robotics for Automated Maintenance, Impact on Surrounding Structures and Project Risk Modelling offers the European Tunnelling industry a real advantage competitively.
• Tunnelling is a fast growing market place. As an example, in the United States, employment in construction is expected to rise 19% between 2010 and 2018; demand for commercial construction and an increase in road, bridge, and tunnel construction will account for the bulk of the job growth. European tunnelling companies need to take advantage of this growth, and NeTTUN advances will be a positive differentiator increasing European competitiveness.
• Improving European companies’ positions for winning developing world infrastructure projects. In particular underground public transportation networks in the very densely populated large cities in developing countries, and many water supply projects dependent on the feasibility of constructing river diversion tunnels, both would benefit from advances made in securing the global risks and costs of tunnelling.
• Reducing the risk of tunnel projects will foster the construction of infrastructure. Such projects rely on the expertise of European companies (e.g. metro project in Guatemala City), but competition is tough e.g. with Japan (Jakarta metro project) and NeTTUN results can assist in this area.
• Create jobs in European countries. NeTTUN has improved competitiveness for the industrial companies such as MI-Partners (creation of the Seismic Mechatronics spinoff selling innovative high performance seismic sources), NFM (improved TBMs and associated services), SNCF (optimised maintenance of their tunnel legacy), IDS (possible new business unit for a Ground Prediction system for Tunnelling), which will contribute to more activity and need for more personnel.
• More than 50% of the NeTTUN partners estimate that participating in the project has generated an increase in employment in their own organisation or at least safeguarded their resources.
1.5 A project that increases sustainable access to underground resources
• TULIPS – the ground prediction system – can be applied to other fields than TBM tunnelling. MI Partners have started commercialising the NeTTUN seismic source to clients in the geotechnical survey business, in particular those working in the oil and gas industry.
• The Ground Penetrating Radar technology developed by NeTTUN deals with issues that are currently limiting the performance of this equipment. On NeTTUN the engineering solutions for the ground penetrating radar (GPR) for TBM incorporating novel system design, antennas, electronics, data processing and information extraction algorithms deliver a breakthrough in short-range radar imaging technology. This technology will be extended to other applications which require intelligent subsurface sensing, including subsurface utility mapping, pipe detection, archaeological and geophysical surveys, construction works, security, and rescue operations particularly for mining and earthquake scenarios.
1.6 A project participating in quality of life improvement throughout the world
• Develop water supply systems for areas of the world demanding urgent access to potable water and irrigation water: many projects are dependent on the feasibility of constructing river diversion tunnels, which would benefit from advances made in securing the global risks and costs of tunnelling.
• Support international health programmes by implementing sewage water networks in the large cities of developing countries, thereby greatly reducing the risk of medical diseases and improving the environment.
2 Main dissemination activities
Dissemination activities were started with a lot of enthusiasm from the very beginning of the project. However, disseminating information on the project raises recurring questions on confidentiality. This is the main reason for which we created the Dissemination Committee, in order to decide on the paper quality but mainly on its confidentiality level. The main role of the Committee was to make the decision on submitting the abstract/paper or not. It also periodically informed the WP Leaders and Beneficiaries of the coming conferences and encouraged their preparation of papers and attendance for scientific and technology watch. The Dissemination Committee was active from Month 4 to Month 54.
2.1 Target Audiences
• The General Public and the Tunnelling Industry
NeTTUN partners were active at participating in Tunnelling fairs, most often associated with conferences throughout the whole duration of the project.
The most important event of the sector is the World Tunnel Congress that is organised by the ITA and is held every year. Altogether 15 papers were presented / will be presented at this congress (2013 to 2017 incl.) and included of the proceedings.
• The Educational and Academic Arenas
Out of the proposed activities to reach these audiences we have focused on the participation in scientific and technical conferences and publication of papers.
In this respect we have covered conferences on radar (AES, GPR2016, and ERC), ceramics (IWAC), materials and wear (ITC), geotechnics (3vX and JNGGI), artificial intelligence (ECAI).
Papers were published in recognised high-level journals such as Wear.
2.2 General Conferences on underground works
NeTTUN partners were active at participating in Tunnelling fairs, most often associated with conferences throughout the whole duration of the project.
The most important event of the sector is the World Tunnel Congress that is organised by the ITA (International Tunnel Association) and is held every year. Altogether 15 papers were presented / will be presented at this congress (2013 to 2017 included) and included in the proceedings.
AFTES (the French chapter of the International Tunnel Association) organises a congress every 3 years. The 2014 edition (Lyon) occurred in the course of NeTTUN, and submitting papers for the 2017 edition (Paris) was also done during the project. 3 papers have been prepared overall.
Other participations of NeTTUN in conferences on tunnelling were:
o Eastern European Tunnelling Conference (Athens, 2014): 6 papers
o TBMs in difficult grounds (DiGs) (Singapore, 2015): 2 papers
2.3 Scientific Conference Papers and Scientific Publications
The Dissemination Committee has been active at reviewing scientific papers proposed by partners. Acceptance was based on two criteria: preservation of IP and paper quality. Changes were most often requested until the papers were accepted, although a few never made their way to publication.
A total of 34 papers were accepted and published by NeTTUN beneficiaries at the following conferences, covering Radar science, Geoscience, Robotics, Tribology, Material science, Geotechnics, and Artificial intelligence: URSI, ECSMGE, ECAP, SEG, TUST, IWAGPR, EAGE, IET, JIFT, IAGIG, ESWC, AES, ICGPR, ERC, IWAC, JNGGI, ITC, Wear, 3vX, ECAI, GJI, EUROFUGE, Tribologia, EURO TUN, Rivista Italiana di Geotecnica, Near Surface Geophysics.
The paper on PADTUN (Leeds/SNCF) presented at ESWC2015 was elected “Best paper” at the conference.
Eight papers have been submitted for presentation beyond the end of NeTTUN, it is expected that they will all be accepted and presented.
2.4 General media
The work done on NeTTUN was also disseminated in general media:
• “Le Journal de Spirou” is a Belgian weekly magazine focused on the young population. It was founded in 1938 and is published in French (over 30,000 copies sold per week in France). Since 2009, some issues contain a special section called “The Lab”, which presents in a cartoon format scientific information, usually in the form of a visit to a research laboratory and interviews with researchers. The issue of 25 February 2015 was dedicated to Tunnel Boring Machines, through a visit made to the NFM factory in France where a 10.2 m EPB (Riyadh metro) was in its final stage of assembly.
• “Sciences et Avenir” is a monthly French magazine for popularisation of science created in 1947. Their April 2016 issue was focussed on the long tunnel projects through the Alps (Gotthard base tunnel completion and first step of Lyon-Turin). The NeTTUN project manager was interviewed for this article, in which our EC-funded research project was promoted. The journalist was invited to the NFM factory in Le Creusot for a presentation of the TBMs being assembled. A special (different) edition of this is available online at https://www.sciencesetavenir.fr/high-tech/transports/tunnel-ferroviaire-du-saint-gothard-la-percee-record-inauguree-le-1er-juin_103811 in which NeTTUN is also mentioned.
2.5 Dissemination and Marketing Materials
Marketing materials are crucial for representation of the identity of the NeTTUN Consortium.
We used the NeTTUN logo on all our public presentations material (e.g. at conferences) and when possible installed the NeTTUN banner in such events.
The leaflet prepared for the World Tunnel Congress 2013 was kept available throughout the project and was distributed during all trade fairs and conferences (distributed on each of the beneficiaries’ booth) and also handed out to all prospects and clients.
A specific brochure was developed by NFM, presenting TULIPS and HECTOR (robot); it was distributed at the WTC 2016 and is continuously used to promote these systems.
A specific brochure was developed by Seismic Mechatronics, the spin-off from MI Parthers, covering the “Lightning” source designed for TULIPS within the NeTTUN project.
2.6 Conclusion
Overall the dissemination activities were successful throughout the project in spite of some restrictions due to the necessary protection of the inventions, in order to allow for the possibility of patenting. This in particular applies to WP7 (new materials) for which there were purposely very few papers but a patent was applied for.
Obviously the beginning of the project is not favourable to publications. The volume rises each year, although there is some negative impact from the periodicity of events, such as the AFTES congress that takes place every 3 years only. It is very rewarding to see that 2017 will be a very active year with 17 publications already planned – and more to come – all beyond the end of NeTTUN.
3 Exploitation of results
Exploitation was addressed from the beginning of the project through a dedicated Committee but was not very successful until the EC provided support through ESIC at the beginning of the second period, which proved very efficient. This led to a better understanding of the objective and proposed methods to achieve the results, thus improving confidence in our capability to succeed. Appropriate contact persons were defined for each entity, making more sense than involving the researchers and engineers only.
A significant step ahead was the hiring at NFM of a dedicated intern for 6 months full time, who took the lead of the Exploitation Committee. Her contract was later extended until October 2016, thereby allowing in order to take the exploitation plan to a point where it can be executed.
After identifying the Key Exploitable Results (KER) with ESIC, a market analysis was conducted for several of them. Questionnaires were prepared, sent to and discussed with potential clients (contractors, tunnel engineering companies). In addition to valuable information on the potential, this approach also provided positive feedback on the viability of the chosen technical concepts and allowed the partners to develop their business network.
Once the KER situation was fixed and acknowledge by all parties, work was focused on the elaboration of a plan for the marketing strategy of each KER. Alternatives were discussed and, for some KERs, decided on. For others this still needs to be finalised.
At this stage, work was initiated on the IPR and focussed on drafting co-ownership agreements between partners in cooperation with the dedicated ESIC consultant. Such an agreement was finalised and signed between ECL-LTDS, TU Tallinn, and NFM for the new materials developed under WP7. This agreement also covers the application for a patent through TU Tallinn. An agreement was also found for the upgraded DAT, with licensing to an SME that provides services based on the use of DAT.
A spin-off company was created from MI-Partners (Seismic Mechatronics) that already delivers seismic sources based on the NeTTUN R&D work. Clients are geotechnical engineering firms working in particular in the oil and gas industry.
Altogether six patents were applied for to cover the NeTTUN results, not including their extensions to foreign countries within Europe and abroad. The concerned results are: TULIPS (the concept, the seismic source); HECTOR (disc cutter locking, drag bit locking, robotic arm architecture); new hard materials (composition and process).
Although a lot has been achieved in terms of Exploitation, there is still a lot to be carried out. The most important success factor is that for each KER one partner takes the lead and organises the work to be done, i.e. finalise the co-ownership agreements, complete the technical work where needed so that the systems are ready any time for demonstration to clients.
List of Websites:
The project website was launched on Month 6. It has been regularly updated by the Website Management Committee and MI Partners in the role of NeTTUN Webmaster.
The url is: www.nettun.org
Contact: pso.nettun@nfm-technologies.com