From the 1950s to the 1970s the use of tunnel boring machines (TBMs) for hydro power projects steadily increased. How-ever, the high capital cost of these machines — partly arising from the fact that they were custom-designed for each project — limited their spread. In the early days, the total cost of a TBM was charged to a project and many machines were used for only one tunnel.

The 1970s saw steady increases in the power and weight of TBMs and a corresponding increase in the rock strength that could efficiently be excavated. Later TBMs are generally very robust in design and there are many examples of machines which, over their lifetimes, have worked on many projects and bored a total of 25-40km of tunnel. Because of their long lives and the added stock of new machines produced annually, there is a growing worldwide inventory of second-hand TBMs.

As a result of these developments, today most TBMs used for hydro projects are second-hand, and the machines are generally modified as required for the project. The diameter can be changed and rock support or tunnel lining equipment fitted as required for the specific geology. An example is the reconditioned machine used to complete the headrace tunnel at the Takisato hydro project in Japan.

The Takisato project

The Takisato hydroelectric project is located in the centre of Hokkaido island, which lies just north of the Japanese main island of Honshu. The project owner is Hokkaido Electric Power.

The headrace tunnel excavation and lining contract was awarded to a joint venture composed of four Japanese companies: Taisei, Kumagai, Mitsui and Itou Hokudenkogyo. Taisei is the lead partner, and the project manager is Tezuka Hiroshi.

A dam is being built on a narrow stretch of the Sorachi river. Below the dam the valley opens into a wide area of farmland, while above the proposed reservoir lies the world-class ski resort village of Furano — site of the 1972 Winter Olympics. These two limitations mean the dam can be built to a height which only allows a hydraulic head of 22m. The headrace tunnel provides an additional 36m of hydraulic head. The final power station will have a peak electrical production of 57MW from a maximum flow of 150m3/min.

The 2.8km-long headrace tunnel was excavated at 8.3m diameter, and when lined it will be 6.9m internal diameter. Of the total, 2.65km was bored by TBM and the remainder is being driven by NATM. The TBM had, in addition, to bore a 110m-long access tunnel before converging with the design line of the headrace tunnel.

The tunnel geology consists primarily of mudstone of the Neogene period and shale and limestone of the Mesozoic period. The mudstone is very weak with a UCS of only 3-15MPa and there are two fault zones, one of which is quite long. The shale and sandstone UCS ranges from 30MPa to 70MPa. There are also several fault zones in the shale and sandstone, but they did not prove particularly difficult to negotiate.

The TBM began the tunnel by boring the 110m access tunnel through a 200m radius turn and through the major fault zone. This was followed by 400m through very weak mudstone before contin-uing through the shale and sandstone.

Due to the extremely mixed geological conditions anticipated, Taisei elected to use a double shield TBM for the excavation. The original tunnel design called for full concrete prefabricated segmental lining to be employed through the mudstone and fault areas. In the remainder of the tunnel it was anticipated that a wide range of ground support methods could be used, from ring beams only to various combinations of invert segments, roof bolts, ring beams and shotcrete.

Due to the short length of the tunnel, Hokkaido Electric Power and Taisei sought various means by which to reduce the cost for capital equipment for the project. One possible method suggested was using a second-hand tunnel boring machine. This was a new departure for the Japanese industry, where the use of second-hand TBMs is still rare, but the contractor Taisei was well aware that worldwide the use of second-hand machines had become more common, and Hokkaido Electric Power was also willing to take a new approach. Taisei contacted The Robbins Company (TRC) with an inquiry for available second-hand TBMs of the required double shield type. Due to the absence of a suitable second-hand double shield unit, TRC proposed to modify an existing open type TBM for the project.

The TBM proposed was a typical Robbins-style mainbeam TBM which had been used previously on the Grand Maison hydroelectric project in France.

On the French project the TBM had bored 5.9km of 7.7m diameter tunnel through granitic rocks. For the Takisato project several major changes were required, including:

•Convert the TBM to a double shield design.

•Increase the diameter from 7.7m to 8.3m.

•Provide a new backloading cutter head for use in the weaker rock types.

•Increase the power and provide two cutter head speeds and reverse rotation.

•Design the machine to excavate a 200m radius turn.

•Design for a maximum component size of 3.5×3.5×6.0m with a maximum weight of 60 tons (to meet transport restrictions).

The final design included a very short cutter head with minimum periphery exposure to minimise cutter head friction in caving ground in weak rock or fault zones.

The cutter head is a six-piece design with the typical four outer head sections and a two-piece inner head. This configuration is necessary to stay within the transport size restrictions. The cutter head is designed for use with Robbins’ backloading 432mm (17in) diameter wedge-lock cutters.

The original symmetric two-row, tapered roller bearing was replaced with a higher capacity asymmetric two-row tapered roller bearing to withstand the increased thrust and large diameter.

The shield design concept is different from Robbins’ usual design in two areas. First, the articulation joints (at either end of the telescoping shield) are sealed with a natural rubber seal, of a Japanese design, as is typically used on Japanese earth-pressure-balance (EPB) and slurry type TBMs. Second, the various shields (forward shield, telescoping shield, gripper shield and tail shield) are all concentric with the tunnel centreline. Because the shields are of successively smaller diameters from front to rear, this means that the annulus area between the shield and the tunnel wall gets successively larger. Again, this is typical of Japanese EPB and slurry machines.

The ever-increasing area between the TBM and the tunnel wall allows some time for the ground to converge without squeezing the TBM. It also allows an area under the shields for fines to pass with-out contaminating the articulation joints. The undesirable side effect is, of course, that a large volume of backfill must be used behind the segmental lining.

The original mainbeam and gripper assemblies are maintained in the final design and still provide the steering mechanism as well as the thrusting mechanism. In competent geology it is possible to use the gripping system, which will allow the machine to advance without the use of the segment lining.

The TBM is fitted with an Atlas Copco COP 1238 probe drill which can drill in 13 different positions over the top 120° of the tunnel. There are many other small changes to the design which are too numerous to detail in this article.

TRC also supplied a complete double track backup system for the project.

Some changes helped make the TBM simpler in use: Robbins used a Mitsubishi PLC for the electrical control system, which later allowed the Japanese to install a Japanese language program. New controls were designed using labels in international symbols and Japanese language.

During the preparation process two Japanese engineers were in residence at the Robbins factory to audit the design and manufacture process.

The existing Grand Maison TBM was transported from its storage place in France to Tacoma, Washington, US for the refurbishment and modification.

The completed Takisato TBM was shipped from Tacoma to Japan in early 1997. Unfortunately, the project got off to a rough start when the transport ship met with extreme weather conditions in the North Pacific. In severe storms at sea, two complete containers were lost over the side of the vessel and a third container was broken in two over the ship’s rail, with half of the container going overboard. Many components were damaged inside their containers, when their lashing was broken in the tossing sea. The results included lost hydraulic power packs, a damaged electrical system, and damaged major hydraulic cylinders.

Damaged parts were sent by air to TRC’s factory for repair and new components were manufactured to replace the components lost at sea. All work was performed under a very short schedule and the replacements were shipped by air to Hokkaido, so there was only a minimal delay in the startup of the equipment.

The TBM breakthrough was on 2 December 1997, on schedule. Maintaining the original schedule was not easy as the geology proved much more unstable than originally anticipated. The original design required only rock bolts for support in most of the tunnel, with full segments used in only perhaps 10% of the tunnel and various combinations of ring beams, bolts and wire mesh in the remainder.

In fact all areas of the tunnel required greater support than the original design. Full segmental lining was used on a great deal of the tunnel. In all areas where full segment lining was not required, ring beams with shotcrete, rock bolts and wire-mesh or sheet-metal lining were employed in various combinations. It was found that rock bolts alone were not sufficient in any place.

The access tunnel (110m) and the first 400m of the headrace tunnel were particularly difficult to excavate. The mudstone was weaker than anticipated and required full segmental lining throughout. In addition, the mudstone was exceptionally ‘sticky’ and blocked cutter head muck buckets and the muck chute unless a great deal of water was added at the face. These were the conditions at the beginning of the tunnel, when the learning curve is the steepest. To make matters worse, the 200m radius curve in the access tunnel had to be negotiated in this same period.

In spite of these difficulties, the JV maintained the construction schedule and, indeed, set a Japanese excavation record for the most material excavated in a month ( 20,560 m3). The JV achieved a best single day of 28.4 linear metres excavated and a best calendar month of 380 linear metres.

The work of this long-lived TBM is not yet over. It is now in storage in Japan, and Taisei plans to employ it on another project before the year is out.