The paper that follows is the 2010 Sir Harold Harding Memorial Lecture. This lecture is the latest in the series of lectures given by eminent speakers on tunnelling and related subjects at the British Tunnelling Society.
Introduction
Chairman, ladies and gentleman. It’s a great privilege to be invited to give this year’s Harding Lecture. Harold Harding, as many will know, was a founder member and founding chairman of our society and also a great engineer and tunneller. He was one of the best and funniest after dinner speakers I have ever heard. One of his bon mots for engineers was “be prepared to be surprised but never appalled”.
I’m very pleased to see some of my more mature colleagues sitting in their normal front row places. Most of you here tonight are the generation with broad minds and narrow hips, whereas we are the generation where the two have changed places.
My lecture will mainly cover the period from the start of what might be called modern tunnelling, the 1950s onwards, which coincides with the period of my own involvement in underground works.
I’ll start with some history as I think it is important to understand history in order to understand where we are today and of course to help us to avoid the mistakes of the past. So historically I will look briefly at the period from when the great Brunels carried out their work on the Thames tunnel and then go on from there to the development of shields and tunnel boring machines and the early use of compressed air.
I will talk mostly about the use and development of closed face machines, which is the title of the lecture, and mention a number of the associated subjects such as the development of tunnel linings to fit in with the sophistication of the TBMs, the development of grouting techniques, both annular and ahead of the face, the better understanding of settlement, the use of conditioners as an aid to closed face tunnelling and the way that compressed air has been largely replaced by the use of these machines and finally the improvement in tunnel logistics.
Now it is conventional wisdom that in any form of civil engineering construction there are 5 important criteria that we must all satisfy.
¦ Cost and budget, bring the project in under budget
¦ Schedule, bring the project in on time.
¦ Safety, minimum accidents.
¦ Quality, satisfy the clients desire for fulfilling the design intent and provide a good quality finish.
¦ Protection of the environment from a visual, noise and pollution point of view.
Now these criteria are important, whether you’re a contractor, an engineer or an owner. They are also largely interrelated. So my main proposition is that the use of closed face tunnelling machines, if properly designed, controlled, monitored and managed will help he achievement of all 5 of these criteria. This is especially so considering the current state of the art and the sophistication of these machines and techniques.
History of Modern Tunnelling – 1825 to Today
I will start of course with the great Brunel’s shield used for the construction of the Thames tunnel between 1825 to 1843, see Figure 1. Considering the state of technology at that time it was a marvellously ingenious design. It was made in cast iron and each individual vertical part could move forward independently.
During the same period came the introduction of compressed air by Cochrane’s patent of 1830. At first they didn’t use a shield with compressed air, but soon discovered that this was dangerous. Cochrane was in fact a very interesting man. He became 10th Earl of Dundonald and commanded the Chilean navy at a time when they defeated the Spanish in a sea battle. He was appointed at that time by the then Chilean President, Bernardo O’Higgins – and that’s not the start of an Irish joke. I should also mention that the first use of compressed air in our immediate area here in London was for the first Blackwall tunnel built between 1892 and 1897.
Then came Barlow’s patent for a circular shield in 1864. Then in 1869 Greathead, using Barlow’s shield patent, designed a shield that was used in the Tower subway, see Figure 2. This was the first use of a circular shield together with a bolted cast iron lining.
Interestingly enough Barlow got his idea for the shield patent when he was working on the caissons for the construction of the first Lambeth Bridge in London in 1862. My father was working on the present replacement bridge when he was married in 1929.
Over in America the first use of compressed air was for a 5.5m diameter tunnel constructed in 1874 under the Hudson River, Then shortly after in 1879 the Kattendyk tunnel was constructed in Antwerp, also using compressed air.
Then come Beaumont and English, who designed the tunnel boring machine for the first attempt at the Channel Tunnel in 1880, see Figure 3. They achieved 2km of tunnel on the UK side using this machine, which was operated by compressed air.
Meantime over on the Canadian/ American border in 1890 a tunnel shield was used, together with compressed air, for the construction of the first St Clair River railway tunnel, see Figure 4. This was notable, amongst other things, for the first use of an erector. You can see from this picture that there was a little bit of the old and the new involved in the construction.
Then coming into the 20th century came the Price Excavator in 1901, see Figure 5. Forty of these machines were made by Markhams of Chesterfield, who also made 2 of the Channel Tunnel machines in 1998, but who sadly no longer make tunnelling machines. 18km of the Northern and Piccadilly lines in London were constructed using Price Excavators. Note that the cutterhead was mounted on a centre shaft.
Then came the Whittaker Machine in 1923. These machines, which look very old fashioned, were manufactured by Sir William Arrol of Glasgow, the contractor who built the Forth Railway Bridge which was opened in 1890. One of these machines was used for an experiment for the Channel Tunnel in 1925. When the Channel Tunnel was constructed in 1989 one of the entrepreneurial geologists, Paul Varley, organised a team and dug out the machine. It was refurbished and is now an exhibit in the Science Museum. The idea for these machines was originally encouraged by the army’s desire to mine under the German lines in World War 1. Another interesting point, from a closed face machine point of view, is that this machine used a screw conveyor in order to convey the spoil from the front to the back of the machine.
In the early 1960s came the development of machines for cohesive ground. Figure 6 shows two of those machines. On the left is the McAlpine Digger using a centre shaft for mounting the cutterhead. On the right is the Kinnear and Moodie Drum Digger. The drum which held the cutter head rotated in between fixed rollers and the main axial thrust was taken by a crude thrust bearing which consisted of a greased soft metal plain thrust bearing, that wore away very rapidly. These machines were bought by the London Transport Board at the time of constructing the Victoria line in London in the early 1960s and they represented the first use of what is now called the Owner Procurement Process (OPP). As a matter of interest the contractors who were working on the Victoria line through the 1960s included names, some of which have become defunct, such as Cochrane, Lilley, Marples Ridgeway, Mitchell Brothers, Kinnear and Moodie, Mowlem, Waddington, Charles Brand, Balfour Beatty, and Kier. Some of them still exist and some of them were absorbed into bigger companies.
The Priestly 100” TBM arrived on the scene in the late 1960s, see Figure 7. This was one of the first uses of a cross roller bearing for the mounting of the cutter head. Robbins had initiated this technology for rock TBMs. So they had discarded the use of a centre shaft, which was comparatively crude and they had turned to the use of a full size roller bearing. One of the early Priestley TBMs was used in 1970 by Edmund Nuttall to construct the Ely Ouse water tunnel in East Anglia. This set up a world record at the time of 435m (1,427 feet) of lined tunnel constructed in one week. That’s actually 7m more than the record on the Channel Tunnel marine drive in 1991, 21 years later. It should, of course, be appreciated that the Channel Tunnel was a much larger diameter than the Ely Ouse tunnel and was of the order of 12km out under the sea at the time the record was achieved.
So at that time, in the early 1970s, unless the ground was cohesive or hard, it was still necessary to use a shield or a TBM with compressed air in order to control the ingress of water, with all the various dangers of the bends, bone necrosis, blowouts, etc.
Slurry Tunelling Machines (STMs)
At the start it was just slurry machines, which arose out of the 1964 patent by John Bartlett of Mott, Hay and Anderson for the Bentonite Tunnelling Machine. They were developed for use in soft ground and mainly used in granular materials below the water table. The face was supported by bentonite slurry and the excavated material, mixed with the slurry was transported out of the tunnel in a slurry pipeline. A separation and slurry cleaning plant was required on the surface. In the top right hand corner of the first page of John Bartlett’s patent it says the patent was for a “rotary mechanical digging mechanism in front of a bulkhead in which a liquid thixotropic suspension is delivered under pressure to the space in front of the bulkhead so as to contact (and support) the working face on which the digging mechanism works, and the spoil excavated by the digging mechanism is removed together with a proportion of the liquid suspension.” Anyone who has used STMs will see that this is an exact description of how slurry machines still function today, albeit the mechanisms have become considerably more sophisticated.
In the UK we started an experimental tunnel in 1972 at New Cross, London using a 4.12m diameter closed face TBM. This became known as the New Cross experiment. It took 8 years, from 1964 to 1972, to find the funding for that experiment which was a great shame. The experiment, which consisted of just 140m of tunnel, did prove the concept of Bentonite Tunnelling Machines as they were then called. There was a slurry circuit which was not exactly the same, but very similar to the type of slurry circuits that are used today, see Figure 8.
In 1975 came the first commercial contract which was for a sewer project in Warrington, north west England. The tunnel was 1.3km long and used a 2.44m diameter Slurry Tunnelling Machine (STM), and was constructed by Edmund Nuttall. It had twin blow up tail seals that were patented at that time, see Figures 9. The first 240m of that tunnel was constructed in sandstone using a conveyor and an open face. In effect it was the first use of what nowadays is called a Mixshield. A patent for a wire brush seal was taken out by the National Research and Development Corporation, NRDC, in 1971. They were one of the funders, along with the London Transport Board, of the original experiment for the Bentonite Tunnelling Machine. It is for a patent using a wire brush, which is very similar to the type of tail seals that are used nowadays, except that there are now multiple seals with fibrous grease being fed in between them.
At the same time as we were carrying out the experiment in New Cross the Germans had started in Hamburg with what they then called a hydro shield, see Figure 10. I believe they still call it a hydro shield. It’s a very elegant concept using twin bulkheads. There is the main bulkhead, in front of which the slurry is held under pressure. Then there is a three quarter bulkhead ahead of the main bulkhead and sandwiched between the two is a large air bubble held under a pre-determined pressure of compressed air. It is the pressure of that compressed air that is used to control the pressure of the slurry in the face. It is a very good system and is still used in Herrenknecht STMs and maybe some others. Just returning to the original Hamburg STM they had to use compressed air in the tunnel because at the time they were carrying out the original experiment they hadn’t developed a tail seal that could hold the pressure.
In Japan they started using STMs in 1964, the same year as the Bartlett patent. The Japanese machines were originally called mud shields. Figure 11 shows a typical one made by Okumura. They used them in very soft material and the process was that soft material squeezed through the very narrow slot at the front. I never inquired about the amount of settlement they created, but I imagine it was quite high.
By the end of the 1970s over 1,000 STMs had been used worldwide, but a vast majority of those were in Japan, where they were catching up on a poorly developed infrastructure in Tokyo and elsewhere.
Today the different makes of STMs control pressure in different ways.
First of all, the hydro shield with the compressed air bubble. With this I suggest that it is possible to control the pressure at the face within +/- 0.05 bar.
Then there are the other machines which use differential pumping in and out of the face to control the pressure. Typical of this are the Okumura Markham 6m diameter STMs used in Cairo in the mid 1980’s for the Wastewater project. They can control the pressure within +/- 0.1 bar. These machines were still using a centre shaft. In the opinion of the Japanese designers at the time it is not advisable to use a centre shaft on a tunnel boring machine much above 6m diameter. This limitation arises from the differential eccentric loads caused by either mixed soft and hard ground or boulders.
I’d like to pause briefly there and mention what one can call “the tunnelling system”, it’s all obvious in a sense. The tunnelling system is in 3 parts:
¦ Firstly the ground,
¦ Secondly the tunnel boring machine
¦ Thirdly the permanent lining
It’s vital, of course, to have compatibility between all 3. You must also suit the tunnel boring machine to the ground and not the other way round, as sometimes happens. You must have co-operation between the tunnel boring machine manufacturer and the lining manufacturer. In addition I suggest that owners and engineers have a responsibility to check that the contractor is proposing the right equipment for the project, as it is usually the owner who ends up paying the final bill.
Earth Pressure Balance Machines, EPBMs
A summary is shown on Figure 12. The Japanese recognised that STMs were very limited in the range of ground to which they were applicable. EPBMs were developed by the Japanese in the mid 1970’s. They only used them within Japan during the first 5 or 6 years. These machines effectively broadened the range of applicable ground conditions. They are much simpler than the STM, with the face supported by conditioned excavated spoil and with the excavated spoil removed from the face using a screw conveyor and transported by train or conveyor to the surface. There is no requirement for the complex slurry treatment plant on the surface. They have to some extent replaced the STM, but certainly not entirely. There is still a variety of ground conditions that require an STM. There are a number of main parts of an EPBM. There is the main bulkhead, with the cutterhead mounted in front, and the screw conveyor for extracting the material, which is part of the control mechanism for achieving the desired earth pressure. The full earth pressure is present at the front of the screw, but by the time the spoil arrives at the outlet end of the screw it must be at atmospheric pressure, in order that it can be ejected on to the belt conveyor in free air. The tail seal at the rear end is important. The airlock for personnel to enter the face to repair the cutterhead and replace the cutter tools is important.
A very early use of this type of machine in the USA was one used under the Anacostia River on the Washington DC Metro. It is a Hitachi Zozen EPBM. It can be seen from Figure 13 that the screw conveyor is very short compared to what is used today. Also nowadays we put the entrance to the screw in the invert. The machine worked reasonably well, but found difficulty in controlling the earth pressure. It can be seen that it had active articulation between the front and mid body.
As a result of the development of closed face tunneling machines new horizons have been opened up to tunnellers. We are now able to tunnel in ground using closed face machines where it was not possible to tunnel previously. For example: The 8.7m EPBMs that were used on the French side of the Channel Tunnel, shown in Figure 14: The 8.75m EPBMs used on the Storebaelt Railway tunnel in Denmark under extremely difficult sub-sea conditions: The very large 14.87m diameter Groene Hart machine and The 6m STMs that were used in Cairo. On all these projects it would have been very difficult, if not impossible, to construct tunnels without closed face TBMs.
So the question to ask ourselves is “what is this all about?” It’s about making sure that when you’re tunneling through the ground, the ground doesn’t know you’re there, or doesn’t feel any impact. Or to quote Randy Essex of Hatch Mott MacDonald in the US, “We are still in the business of trying to fool Mother Nature”.
I refer to the report produced by the British Tunnelling Society and the Institution of Civil Engineers in 2005, published by Thomas Telford and titled, “Closed Face Tunnelling Machines and Ground Stability”. It represents a guideline for best practice and is an excellent reference guide.
Choosing Between STMs and EPBMs
A very important subject is how you choose between an STM and an EPBM. Obviously ground conditions is the most important one. Figure 15 shows the grading of ground suitable for each type of machine. Generally an STM is suitable for the more granular soils whereas the EPBM is more suitable for silty, clayey soils. This is because the STM’s separation plant will struggle to keep the slurry clean if there is too much fine material below 60 microns. Whereas the EPBM will find difficulty in controlling the pressure using the screw conveyor if the percentage of fines (again below 60 microns) is less than say 10-15%. However this distinction is more blurred these days because:
¦ Firstly the Slurry treatment plant in STMs has become very much more sophisticated and can therefore remove much finer material from the slurry. There are very sophisticated vibrating screens nowadays which of themselves can separate material down to the 60 micron size: then varying sizes of hydrocyclone for getting it even cleaner: finally hydraulic filter presses or centrifuges for removing the extremely fine material.
¦ Secondly for EPBMs the use of long screws, double screws, double piston pumps and the use of more sophisticated conditioners has helped considerably with controlling the pressure when there is a lack of fines. Figure 16 shows a double screw and it can be seen from the diagram that if the screws are rotated at different rates of rotation, you can build up a sand plug between the screws to aid in the pressure holding capacity.
¦ Finally, conditioning for EPBMs which is now a very significant part of the echnology. Up to the mid 1980s it was a choice between bentonite, polymer or water. In the mid 1980s the Japanese introduced surfactant foams which produce a better viscosity, give help in binding the fluids and solids together and finally produce a significant reduction in cutterhead torque.
Another criterion to consider is the permeability of the ground. A distinction can be made that STMs are more appropriate if permeability is above 1×10-5m/s and EPBMs are more appropriate for ground below that figure. With added fines, such as pulverised limestone, EPBMs can perform in even more permeable ground.
Hydrostatic heads must also be considered, especially if a high hydrostatic head is combined with high permeability. In this case an STM may be more suitable. Settlement is a key factor. However with the introduction of foam conditioners to EPBMs there is not a lot to choose between the two types in terms of controlling ground movement. Both types, properly managed, have the capability of controlling volume loss to less than 1%. I would suggest that 0.5% is now a figure that engineers and contractors should be aiming at today. There is quite a lot of case history, as shown on Figure 35, to support this figure. The Channel Tunnel Rail Link is a good example. It achieved an overall average of approximately 0.6% volume loss. On the St Clair River Tunnel in Canada, using a 9.5m EPBM it was less than 1%. Interestingly enough on the Warrington Sewer project, which was carried out in the mid 1970’s, and used a prototype STM, Nuttalls managed to keep the volume loss down to 1.37%. On the Heathrow tunnel where they used compressed air in the face of an EPBM instead of a full face of conditioned spoil, which is a departure from normal, they kept volume loss to 0.5%.
Other factors to consider, that would not normally lead to distinguishing between STMs and EPBMs, are the presence of boulders and the ability to measure excavated quantities accurately. But the overriding need is to select a TBM that can control ground movement and volume loss.
To conclude on the subject of choosing the type of closed face TBM, it should be said that the choice is sometimes extremely difficult.
Control of Settlement
There are a number of key factors when using a closed face TBM which are important in terms of controlling settlement.
¦ The earth or slurry pressure at the face must be kept at a calculated level. Earth or slurry pressure should be calculated for the full length of the tunnel. It is a mistake to take one figure and try to keep it at that throughout the tunnel length, as the required pressure will vary due to the ground conditions, the tunnel depth, the hydrostatic pressure etc.
¦ Use a minimum overcut around the shield. This can be in the form of a bead or a tapered shield resulting in not too large an annular gap around the shield.
¦ Fill the gap around the shield continuously with a bentonitic paste. Interlock the pumping of this paste with the forward movement of the TBM. This can be considered to be a fairly recent innovation.
¦ Use tail skin grouting to fill the annular gap between the excavated profile and the extrados of the permanent lining. This is now an accepted method and in my view must be used. Interlock the pumping of the annular grout with the forward movement of the TBM
¦ It is also vital that the TBMs are supplied to the site with both adequate thrust and torque. There must be no limitation on excavation rate due to lack of either thrust or torque.
¦ Another important subject, which is mentioned below, is the open area of the cutter head.
¦ The ability of the machine to steer properly within preset limits is also most important. If TBMs go off line, a large percentage of available thrust will be absorbed in making steering corrections.
Specifying and Designing Closed Face TBMs
There are some important details that must be considered when specifying or designing closed face tunnel boring machines. It is vital that all the necessary facilities are fitted to a TBM before it leaves the factory. It must leave the factory as it were “all singing and all dancing” and ready for any eventuality below ground. It’s too late if you are trying to modify the machine below ground- it will cause delay and very large expense:
¦ The Permanent Lining. The technology has had to keep pace with the development of these sophisticated machines. On the Victoria Line on the London Tube in the 1960s the transition happened almost overnight between the use of cast iron and the use of concrete, which is virtually uniform nowadays, except for special requirements such as the junctions between running tunnels and cross passages. Figure 17 shows the 9.5m internal diameter trial rings on the Hallandsaas Project. Each of these rings is 2.25m long and uses tapered dowels on the circle joints. This has enabled the building of an extremely accurate ring below ground.
¦ It is vital to have the correct open area on the face of the cutter head for EPBMs. It is not so important for an STM. I suggest that 35% open area on the cutterhead should be the aim. However if there are special reasons, this can be reduced to 30% but no lower. On the Storebaelt EPBMs in Denmark, it was lower at approximately 25% and this gave severe problems in terms of spoil jamming in the cutterhead.
¦ The ability to be able to carry out pre-treatment ahead of the face is essential. Figure 18 shows drilling through both the skin of the TBM and through the cutterhead. Many engineers say don’t go through the cutterhead because the rods might break and that would cause problems for the cutterhead and cutting tools. However on the Hallandsaas project in Sweden they are currently drilling ahead and grouting through the cutterhead in order to reduce the large amounts of water. This has not caused problems.
¦ Conditioners have already been mentioned with surfactant foams described as revolutionary. So the use of foams, bentonite, polymers, water, and added fines are all standard nowadays. Therefore the facilities for delivering them to the excavation chamber must be designed into the TBM.
¦ Risk management and safety are important subjects but they are subjects for another paper.
¦ I must also mention the use of the Owner Procurement Process (OPP) on various projects. It’s been well used internationally. The first use was on the Victoria Line, which is referred to earlier with the use of the McAlpine Centre Shaft TBM and the Kinnear and Moody Drum Digger in 1963. This was followed by the Melbourne rail loop in 1972. Then in other places such as the Sheppard Line in Toronto. The main reasons for using the OPP is that first of all you obtain a risk reduction because there’s more time than normal to think about and develop the design. The machine will not be a second hand machine from contractors who happen to have it in their yard and would like to use it, but the machine may not be exactly suitable, so there would be compromises. Very importantly there is usually a considerable time saving if tunnel boring is on the critical path.
Figure 1: Brunel’s shield for the construction of the Thames tunnel Figure 2: Greathead’s shield, used in the Tower subway Figure 4: St. Clair River Tunnel – 1890 Figure 3: Beaumont & English TBM-1880 Figure 5: The Price Rotary Tunnelling Machine – 1901 Figure 6: Drum Digger & McAlpine Digger c. 1963 Figure 7: Priestley 100″ TBM – 1967 Figure 8: Bentonite Tunnelling Machine Circuit Figure 9: Tailseal Patent – 1971 Figure 10: Hydroshield – 1971 Figure 11: Japanese Slurry Machine 1966 Figure 12: Typical Earth Pressure Balance TBM Figure 13: Anacostia EPBM-1985 Figure 14: New Horizons for Tunnelling – the 8.7m EPBMs that were used on the French side of the Channel Tunnel Figure 16: Double Screw Figure 15: Soil Grading Curve Figure 17: Tunnel Linings – Hallandsaas Figure 18: Pre-treatment ahead of the Face
