Bangkok, the capital and most populous city of Thailand, has a population of around 10.5m (as of 2020), which is about 15% of the country’s total. Taking into account the entire Bangkok Metropolitan Area, there are more than 14m people living or working there, which makes this area one of the most congested in the world in terms of traffic and people.

The Asian investment boom seen in the 1980s and 1990s led many multinational corporations to locate their regional headquarters in Bangkok, turning the city into a regional force in terms of finance and business, and making the city among the world’s most visited tourist destinations.

Bangkok’s rapid growth, coupled with very poor urban planning, has resulted in inadequate infrastructure and a chaotic cityscape, leading to chronic and crippling traffic jams which caused severe air pollution in the 1990s. The city has since turned to public transport in an attempt to solve the problem, with five rapid transit lines in operation and others under construction.

Proposals for the development of rapid transit were put forward in the 1970s, but many of those plans failed to materialise. It was only in 1999 that Thailand’s first rapid transit system, the so-called BTS Skytrain, went into operation.

The underground MRT (Metropolitan Rapid Transit) subsequently opened in 2004, and has been expanded ever since. As a mass rapid transit system, the MRT comprises two rail lines, with a further three currently under construction and due to open in 2021. The MRT Blue Line was the first to open in 2004, followed by the Purple Line, opened in August 2016.

Both Blue and Purple lines are operated under a concession granted by the Mass Rapid Transit Authority of Thailand (MRTA), and together, comprise 45 operational stations and a combined route length of 60km, with a daily carrying capacity of 470,000 passengers,. The Orange Line is the newest development in the mass transport network of the city.

The city’s rapid growth has also led to the phenomenon of ‘verticalisation’, with thousands of new commercial and residential buildings being erected in just a few decades. This has created the rich resource of the city’s geotechnical parameters.

BANGKOK’S GEOLOGY

Bangkok is located in the delta of the Chao Phraya River, which meanders through the city in a southerly direction towards the Gulf of Thailand – around 25km south of downtown. Average elevation is only around 1.5m above sea level. The area has been gradually drained, and the course of the river has been modified over decades by the construction of several canals.

The Bangkok underground sedimentary series of clays is characterised by a top layer of soft marine clay, known as ‘Bangkok Clay’, averaging 15m in thickness, and which overlies an aquifer system. This feature has contributed to the effects of subsidence caused by extensive ground water pumping and is currently one of the biggest challenges for the underground construction sector.

In the early 1980s, subsidence reached a rate of 120mm/year, but actions taken since then have lessened the severity of the situation, although subsidence is still occurring at a rate of 10–30mm per year. Indeed, parts of the city are now 1m below sea level, highlighting fears that by 2050, the city may be submerged.

The succession of the named Bangkok Clay is described as a horizontal arrangement of clay layers in three main units:

  • Soft clay, from the surface down to a depth of around 10m.
  • Medium clay, from 10m down to around 15–20m deep, with a variable thickness of 5–10m.
  • Stiff clay, between 15–25m deep, with a variable thickness of 5–10m.
  • Sandy formation at depths of over 25–30m.

Figure 1 shows a typical section through the geology. The soil composition, together with the aforementioned geology, makes excavating a tunnel by EPB TBM, choosing the ideal soil conditioning foams, and controlling all TBM parameters during excavation, a challenge to overcome.

PHYSICAL PROPERTIES OF BANGKOK CLAYS

The physical properties of Bangkok Clays can be defined and are presented in Table 1, which is readily available in most common bibliographies relating to Bangkok’s underground geotechnology.

Water content (as usual) decreases going ‘top-down’ to a more consolidated clay and logically, the same applies to the void index. The implications of this for the TBM and the generic behaviour of soil conditioners, is:

  • The deeper the material, the lower should be the FER of the conditioning foam, which may require to be used with some additional and limited injections of water (WIR);
  • Plasticity and liquidity limits show the change in the degree of consolidation.

CHOICE OF SOIL CONDITIONING AGENTS

A wide range of soil conditioning admixtures are available. These are typically water-based solutions of different tensides, normally with readily biodegradable properties, suitable for injection at the front of excavation, in the excavation chamber and along the screw conveyor of the TBM.

As a general classification, different tensides are suitable and recommended in relation to the physical properties of the soils to be excavated. A simple, but not exhaustive criterion is to define the use of foaming agents based on the particle-size distribution (PSD) of the soil.

Regarding Bangkok clays, the analysis made at the supplier’s lab resulted in an average content of particles below 0.004mm of 30–70%, a silt content of 20–40% and a sand content of 7–25%.

The average granulometric characterisation of several samples of Bangkok clays are shown in Table 2.

A chart showing particle size distribution (PSD) of Bangkok clays relating to tensides provided by Sika Thailand is shown in Figure 2.

PRE-TRIALS IN THE LAB

An extensive trial phase was conducted at the labs of Sika Thailand on samples provided by the site and taken from depths that were comparable with the tunnel alignment.

After measuring the main physical properties of the soil, soil conditioning campaigns were undertaken to estimate and predict the possible FER, FIR, WIR to be adopted during the actual tunnelling operations to provide the contractor with first guidelines on the conditioning parameters.

As a basic assumption, the following parameters were adopted for the foam generation:

  • C% of foaming agent in the foaming solution: 2%
  • Two values of FER targeted: 1:5 and 1:10

Foam was generated with a high-speed lab stirrer (4,000rpm). It should be mentioned that this system does not provide efficient foam generation as does a real foam generator; the stirrer creates bigger bubbles and shorter half-life of the foam.

Consistency testing after conditioning was undertaken with a standard shock table test to limit the consumption of soil samples, thus allowing the repetition of the conditioning test if needed.

The shock table test was done with 15 shocks in three steps of five shocks, recording the spread value at each step. The expected value of spread to be reached after three shock steps is between 140mm–160mm.

Figure 4 correlates the values of spread given by a shock table on small samples with the values of soil consistency provided by the slump test with an Abrams cone. This latter can be used only if big enough samples are provided by the jobsite, but the good correlation between both methods demonstrates that reliable results can be achieved with smaller samples and using a scaled test.

By the end of the pre-testing campaign, four main types of soils could be identified based on the physical parameters and related to specific soil conditioning parameters as obtained in the lab:

  • Face 1: a mix of 50% soft clay and 50% medium density clay.
  • Face 2: the face of the tunnel is completely made of medium clay at 100%.
  • Face 3: the face of the tunnel is completely made of stiff clay at 100% with a water content as low as 20- 40%.
  • Face 4: a mix of 50% stiff clay and 50% sand.

For Faces 1, 2 and 3, the foaming agent (SikaFoam 501 LS–TH) could provide a good consistency at FER ranging from 1:5 to 1:10, having lower values for very consolidated clays, a FIR of 20%–60% as maximum values detected during the lab trials; and WIR of 10–15% being on 15% when highly consolidated clays were used. The use of the foaming agent for granular soil at Face 4 conditions could improve the conditioning of the soil but also the foaming agent for clays reached quite good performances in clay percentages in the range of 50%.

In Table 3 detailed information is provided.

THE TBM COCKPIT SOFTWARE

The use of a software tool, developed jointly by Sika Services and Sika Thailand, allowed continuous monitoring of the soil conditioning parameters forecast versus the actual parameters – consolidating long lists of field data into one screen with clear statements on average foaming agents consumption compared with excavation speeds of the machine, for the specific dimension of the tunnel, duration and length of drive.

This also allows us to consolidate parameters of many rings into one cockpit view for the estimation of average parameters along the section of tunnel under consideration.

In summary, the analysis was undertaken considering the following points:

  • To compare the forecast data with post-excavation data.
  • Parameters considered were:
  • FIR
  • FER
  • WIR
  • Real speed of excavation vs target speed
  • Total chemicals consumption (foaming agent) per ring
  • General assessment of excavation operations.

Figure 5 is a screenshot of the TBM cockpit tool.

COMPARISON OF LAB TEST WITH SITE RESULTS

Monitoring took place during the excavation of 871 rings; the forecast data provided by Sika Thailand after lab screening met – in quite a good way – the parameters achieved by the TBMs, especially in cohesive grounds which have:

  • Acceptable correlations between FER forecast and actual FER.
  • Actual FIR has been much lower than forecast (with a much lower consumption of foaming agent, compared with forecasts from the lab pre-screening).
  • Actual WIR has a random path compared with forecast WIR, being the values in the magnitude range.
  • In relation to the concentration of foaming agent into the foaming solution, the contractor actualised the value suggested by the provider, being between 1.8 and 2%.

Figure 6 shows a graph for the comparison between actual value and forecast values.

CONCLUSIONS

Collecting excavation parameters along almost 1,200m of tunnel has shown that the change in the degree of consolidation of Bangkok Clays is only slightly affecting the conditioning parameters of ground at different geological faces.

An exact fitting between post excavation actualised parameters and forecast parameters is a difficult target because lab assessments on soil samples collected before excavation may be influenced by the quality of the samples (humidity, contamination, representativeness, etc).

Lab trials have a tendency to overestimate the needed FIR and at the same time to work with FER values close to 1:8–1:10. FIR values at the site can even be 50% lower, while maintaining a good level of conditioning, good extraction rates of spoil at screw conveyor; at the same time, targets of advancing speed, torque, etc. are well achieved. A definition of conditioning parameters in the lab only gives indicative results which can be used as a generic and preliminary reference.

Moreover, TBM foam generators are much more effective than other manual systems or stirrers used in lab to generate foams. This makes the site soil conditioning much more efficient than with lab samples. Pre-screening with adequate lab foam generators can improve the above aspects but the quality of the ground sample still remains uncertain. Figure 7 shows a good equipment set-up for foam generation in a lab.