Equipping TBMs for tough conditions 28 Nov 2019

Shani Wallis, TunnelTalk

Predicting trouble ahead; preventing EPB clogging; mixed conditions in hard rock face; machine learning for autonomous TBM operation; disc cutter design, and much more were topics discussed during two informative days of conferencing at the 2019 TBM DiGs conference in Denver. Convened by a team of the Colorado School of Mines, under the chairmanship of Jamal Rostami, the small number but highly-knowledgeable delegates of engineers and professionals from more than 20 countries, including China, India, Austria, Norway, Germany, Spain, UK, USA, Japan and Korea, profited by the specialised programme of plenary keynote sessions and break-out into two tracks of topic-specific presentations. Within a crowded calendar of industry conferences, these specialised examinations of single topic themes are high value within the schedule of wider scope conferences such as the annual World Tunnel Congresses.

Herbert Einstein of MIT opened proceedings
Herbert Einstein of MIT opened proceedings

Professor Herbert Einstein of MIT (Massachusetts Institute of Technology) opened proceedings with an invited keynote on the assessment, decision-making and management of risk involved in tunnel construction and TBM operation. Professor Einstein demonstrated that much can be predicted about the conditions to be expected from information available from historic data, alignment geological site investigation and ahead of facing both natural and human error risks. By illustration, a project was cited where predictions called for a TBM to move from open to closed mode operation. The change did not take place and face loss and surface settlement were the consequence.

Among several questions, TunnelTalk asked: given that the consequence as described occurred within the TBM advance of one ring, could the operation of the TBM be programmed to react automatically to the predicted or detected changes in geology. The answer came from another delegate in the audience affirming that the possibility is already available on TBMs. Richard McLane of Traylor Bros explained how interlocks, as they are called, are programmed into the TBM programmable logic controller (PLC) to bring TBM operations to a stop automatically if, for example, pressure in the excavation chamber falls below a trigger value or if continuous backfill of the annulus behind the segmental lining stops. The concept has been used on project TBMs for several years, he added, but that it is not applied as industry best practice. Professor Einstein concurred that artificial intelligence or machine learning is the way forward and will become more sophisticated as technology and big data collection and management improves with time, research and development.

Artificial intelligence and machine learning

Much discussion about the possibility of automatic TBM operation carried on from the TBM DiGs conference in Denver to the 2019 ITA Awards and Cutting Edge conference in Miami the following week. It was the system developed by the MMC-Gamuda team in Malaysia for autonomous operation of the Herrenknecht variable density TBMs engaged currently on Line 2 of the KVMRT metro in Kuala Lumpur that won the category for Technical Innovation of the Year.

While the contributions during the final Discussion Panel of the Cutting Edge conference, which hosted the ITA Awards event, ranged from those who fully embrace the technological possibilities, and those who warn be careful what you wish for, the team from Malaysia confirmed that it is precisely during high risk reaches that operation of the machines are taken away from the human and taken over by the intelligent machine. “The human operator takes too long to appreciate something has changed and too long again to decide the solution and react. Intelligent systems for autonomous machines, operating with all the data available, can compute the situation in seconds or milliseconds and react equally as fast.” It was suggested also that this will solve the issues of training skilled TBM operators and addressing the critical shortage of skilled TBM operators in the global industry today.


To advance the discussion, Ulrich Meidl of Germany, in his invited keynote, explained how much information can be gathered by current sensors and monitoring systems. “This is now truly impressive, and especially with the added capacity of cloud computing. Systems that monitor TBM operating parameters along can deliver more than 1,000 items of data every two minutes. Add to this the information from settlement monitoring networks and logistics management systems, the collection can run into the several thousand pieces of data. From all of this data we need to combine big data with human engineering to arrive at the analysed parameters that are of importance to the daily operation and management of a TBM project. This will be more than using the technology as a black-box recording of data for back analysis as it is today.”

Recent development of TBMs, particularly for multi-mode machines capable of dealing with wider ranges of soil types and converting from open to closed modes, and modes in between, was discussed in several papers included a paper by Karen Bäppler of Herrenknecht on the design and operation of the variable density TBM concept.

Difficult conditions presented by hard rock geology gave rise to interesting discussions.

Mixed face intrusion in a hard rock mass
Mixed face intrusion in a hard rock mass

This started with a suggestion by Björn Nilsen and Pal Jakobsen of the NTNU University of Science in Trondheim, Norway, that there can be mixed face conditions in a hard rock TBM heading, conditions where rock of different compressive strengths and qualities can be present across the same face, conditions that can cause damaging impact for cutters during excavation.

Don Del Nero of the USA followed up and opened his address about abrasive conditions with a quote by TBM consultant Richard Lewis that “wear can add the same price as the TBM again to a project”. Del Nero went on to explain that quartz is harder than most manmade steels and that:

  • the quantity of quartz on a TBM job should be included in the geotechnical baseline report (GBR);
  • the composition of cobbles and boulders should be a baseline item, not just their predicted number;
  • there should be a licensed geologist on every tunnelling job;
  • international bodies such as the ITA or the ISRM should work on developing universal standards for testing and reporting of abrasivity;
  • that while carbides are the best anti-wear metals at present, new materials such as cermats are being developed, and that
  • research and development of cutter tool life x ground abrasivity is coming up with some far-reaching disc cutter design ideas to assist also disc operation on mixed face, multi-mode TBMs.
Mega machine experiences from Seattle

A full afternoon session at TBM DiGs, plus the keynote of another session, was given over to presentations about the experience of the mega Hitachi Zosen EPBM for the SR99 Alaskan Way viaduct replacement highway tunnel in Seattle – not anything about the cause of the breakdown or two year repair process which remains under a court case subjudice, but from the restart in January 2016 through to breakthrough on 7 April 2017, achieving up to 14 rings per day and completing 8,186ft of 17.48m o.d. excavation and segmental lining to 15.54m i.d. in 452 calendar days and 195 excavation days.

Mega TBM advance to breakthrough

Invited keynote speaker Juan Luis Magro, Construction Manager for Dragados and for the Seattle Tunnel Partnership construction JV, quantified some of the remarkable statistics associated with a TBM of 17.5m o.d. including an excavation volume of 1,000 tons or 675yd3/shove for the 2.1m wide ring of segmental lining; that 7,000 tags were feeding to the PLC data logger every three seconds from the TBM monitors and more than 2,000 instruments on the alignment; that the cutterhead was dressed with 743 cutting tools, 106 of them able to be accessed and changed in hollow atmospheric arms of the cutterhead; that there were 272 dives during a total of five hyperbaric interventions; that settlement for passage under the viaduct was predicted at 0.25-0.95cm and was actually limited to 0.08-0.25cm and that, at the end of the process, it was proven that super-sized TBM dives under urban areas and high-rise city centres is possible, “it can be done”.

There were many questions in follow up to Magro and to Brian Russell of HNTB, engineer to the construction JV, who gave the earlier paper, discussing various points about settlement control, soil conditioning, the selection of EPB technology over a slurry TBM alternative, and the management of fewer than expected boulders. It was an enlightening exposé into the operation of mega EPB TBMs and the many parameters to be controlled for success.

Sindre Log of The Robbins Company added to the cutter discussion and answered a TunnelTalk question about the regularity of changing worn or damaged cutters, either one at a time as necessary, or several during a planned maintenance period. He confirmed that there are advantages of changing all cutters, for a complete new set, but that other considerations, such as extended downtime, can cancel or neutralise the benefit.

The performance of the disc cutters also affects the possibility of recycling excavated material as concrete aggregate, as discussed by Robert Galler of Leoben University and the ZAB Underground Research Centre of Austria. Too much grinding into fines reduces the value of the excavated material and the efficiency of processing excavated material into aggregate.

Staying with hard rock TBM operations, the realities of high-volume, high-pressure ground water ingress into a TBM heading, and the efforts of using grout injections to stem the flow, was the topic of an impressive presentation by Knut Garshol from Norway. Probing and the cycles of pre-excavation grouting were discussed with the recommendation that shorter arrays with more frequent overlaps for pre-excavation grouting is much more effective than trying for 100m long grouting arrays.

Taking a large diameter soft ground TBM through a tight S curve alignment was the paper by Professor Mistutaka Sugimoto of the Nagaoka University in Japan. Special liner plate segments of about 25cm wide were used to take the curves and when asked by TunnelTalk if special needs were required of the TBM tail seals the answer was no, that the four wire brushes and lots of tail seal grease manage the seal through the curves.

Fig 1. Quick sketch during conference discussions
Fig 1. Quick sketch during conference discussions

The discussion went on to the sealing of the articulation joint, and while Brian Khalighi for Robbins and Matthew Greger of Robbins described systems applied by Robbins to seal the standard straight overlap articulation joint, Professor Sugimoto described a knuckle type articulation joint which, he said, is easier to seal (Fig 1).

Of particular interest to the delegates, during the Greger presentation about operation of the mega Robbins slurry TBM for the Hiroshima highway project in Japan, was the information that vacuum segment erectors are not permitted by regulation in Japan. Instead, Greger detailed the design and operation of the lock-in mechanical pin erector system for lifting and placing the project’s mega-sized precast segments.

Hyperbaric intervention was the focus of the keynote address by Werner Burger, Chief Engineer for Herrenknecht. In describing the high pressures acting on the underwater drives for the Lake Mead intake tunnel in Nevada, the Eurasia subsea highway under the Bosphorus in Turkey and for the Tuen Mun-Chek Lap Kok subsea highway link in Hong Kong, Burger explained how certified hyperbaric divers lived for up to 28 days in compressed air conditions to be shuttled down the tunnel to the TBMs to complete their maintenance shifts in the excavation chambers. During the coffee break, TunnelTalk asked if an operation similar to the Project Moses on the Storebaelt railway project under the sea in Denmark in the 1990s that installed a dewatering system on the seabed to lower the ground water pressure ahead of the TBM drives, had been considered to lower the hyperbaric pressures on these extreme TBM drives. The answer was that the ground conditions on the recent projects was too permeable and a seabed dewatering system prohibitively expensive in scope and operation.

Guides and a hydraulic pin that inserts and twists is the key to the mechanical segment erector

An appropriate wrap up of proceedings was presented by Brian Fulcher of the USA who, as Animateur of the ITA Working Group 14 on Mechanised Tunnelling, explained how the Working Groups of the International Tunnelling and Underground Space Association (ITA) provides the international forum in which many of the suggestions for investigation are being progressed and through which the reports of the Groups are freely available for pdf download. In response to a question from TunnelTalk, Fulcher confirmed that the Working Groups – 23 in total and covering a wide range of specific topics – are open to all who are interested to participate. “Being a Member of the ITA or a Member of a Member Nation delegation is not a prerequisite,” he said. “Anyone in the industry is welcome to become involved in the activities of the groups. Ours is one of the most active of the Groups and we had 45 in the Working Group sessions at the WTC in Naples in May (2019). It is by expanding the pool of experience and contribution that the best for the industry as a whole and internationally is achieved.” More information about the ITA Working Groups and their free download publications is available on the ITA website.

In closing the event, Conference Chair Rostami announced that the next TBM DiGs conference will be organised and staged under the Chairmanship of Professor Galler at the University of Leoben in Austria in November 2021. All were invited to join the event to continue the discussions.


Add your comment

Thank you for taking the time to share your thoughts and comments. You share in the wider tunnelling community, so please keep your comments smart and civil. Don't attack other readers personally, and keep your language professional.
In case of an error submitting Feedback, copy and send the text to
Name :

Date :

Email :

Phone No :

   Security Image Refresh
Enter the security code :
No spaces, case-sensitive