The low lying south-west region of Japan’s Yokohama city has been suffering from severe flooding in times of heavy rainfall for many years now. To alleviate this serious problem a 3.5m i.d pipe was installed to divert the storm-waters using the shield tunnelling method over a length of 10km. This paper reports the final section of this challenging project, which was the construction of the final section of pipe to the River Sakai. This short 194.5m long stretch, at some 90o to the TBM driven tunnel, was installed using the new super-sized pipe jacking method for the first time in Japan (figure 1).

This section of drainage pipe also needed to be 3.5m i.d., to interface with the rest of the main line. If the i.d. had been less than 3m, standard pipe jacking would have been the method of choice. However, being 3.5m i.d, the project posed an unusual problem as the Japanese Road Law limits the maximum height of a trailer truck to less than 3m, and therefore the maximum diameter of a transportable jacking pipe to 3m.

To overcome this, a supersized pipe jacking method using segmental pipes and a segmental pipe-jacking machine was developed. The work was completed between July and October 2005 and was carried out for Yokohama City, Japan. The world’s first segmental (half-split) pipe was developed by Nippon Zenith pipe and installed by Toda Corporation.

Why not open cut or shield?

The road above the planned pipe alignment was only 2.5m wide. There would have been several major problems to address if the open-cut method had been adopted. Firstly, all of the land would have to have been acquired from private owners making the exercise very costly. Another serious issue was the fact that the groundwater level was high and the ground cover shallow (figure 2). In order to minimise the influence of the construction on the surrounding ground, chemical grout injection would have been necessary. However, this was not an option as the site was adjacent to agricultural land raising fears over the chemical grout’s detrimental effect on the local environment.

The possible use of the shield tunnelling method was also disregarded as the area required for the launch shaft and associated construction site would have been far too large. Additionally, the construction cost per meter would be too expensive and the construction period too long.

Problems with the pipe jacking

It was decided that pipe jacking was the way to go, but for the fore mentioned reason (i.e the limited transportable pipe size), a new type of jacking pipe had to be developed. Not only would an assembled pipe have to have the same physical strength as a single unit pipe but it must also provide the same water tightness and be easily and quickly assembled on site.

Additionally, the mud pressure type pipe-jacking machine had to be highly functional and also easy to assemble on site.

Apart from the challenges associated with the equipment required, behind the thrust wall of the launch pit were main gas, water and telecommunication pipes (figure 3). The results from the FEM analysis showed that if the jacking force exceeded 13,000kN, the existing pipes would be affected.

Although not a technical problem, one of the biggest challenges to the project was the lack of pioneering spirit usually seen among Japanese municipalities. Initially Yokohama City was reluctant to use a method that had not been successfully tested in the past.

The solution

After careful consideration, Yokohama City decided to use this new pipe jacking method after it was confirmed that the assembled pipe could show the same physical performance as a single piece pipe.

To help achieve this, a cotter type pin was used to connect the two halves of the pipe. In order to use such pins, the project accuracy had to be less than plus or minus 1.5mm. Using the match casting method helped achieve this. A fast-curing acrylate resin was then applied to the joint surface of the halved pipes. The curing time for this was approximately 2 hours and the whole assembling process took just 3 hours. The combined effect of the pins and the acrylate resin made the physical strength of the assembled type pipe as strong as an un-jointed pipe. A water pressure test was carried out to test the joint between the two halves of the pipe and the joint between two pipe lengths. This test proved that the pipe could withstand pressures of up to 0.3MPa.

The main part of the pipe-jacking machine was divided into 10 pieces with the weight of each piece less than 10t. Bolted connections were used to assemble all of the parts. A spigot-type joint was then placed between the segments improving the waterproofing of the joint and making assembly and disassembly easier.

To reduce the jacking force, lubricant was injected in two stages. A two-part mixed type plasticizer was used for the first stage to seal the tailvoid avoiding dilution by groundwater. For the second stage, a highly viscous lubricant was added to complement the first stage lubricant.


On the whole, the project was a success with no delays or accidents. The lubricant worked well and the jacking force was reduced from 26,216kN to 6,700kN. As a result, the jacking was undertaken without effecting the existing pipes behind the jacking pit.

The only difficulty which has to be mentioned was due to the coarse grained soils. A pump was not sufficient to take out the excavated soil, so a belt conveyer was used instead. Because of the small pit size, only a 2 cu m container could be used which meant stopping every 140mm of excavation. The cycle time is shown in figure 4. For the continuous operation, each pipe had to be set before the completion of the previous pipe jack which took 3 hours and 30 minutes. Nevertheless, the whole process went smoothly and installation was completed in almost two months.


The success of this 3.5m diameter pipe jack was important news in Japan. This project clearly demonstrated that pipe installations of over 3m diameter can be constructed using pipe jacking and not just by the conventional shield tunnelling method.

If this project had been carried out using the shield tunnelling method, it is thought that it would have taken more than 15 months (from development of the machine to completion of the tunnel lining) whereas this project was completed in a total 6 months including the development of the jacking machine.

The Japan Microtunnelling Association is currently working on standardisation of this supersized pipe jacking method, with an i.d. of up to 5m.

In recent years, especially in Japan’s urban areas, floods are becoming more and more frequent from the increasingly heavy rainfalls. Using this method to install rain storage or conveyance pipes would be a suitable way to manage this problem.