THE SOUTHEAST Collector (SeC) Gravity Trunk Sewer is a 15km, 3m-diameter tunnel currently under construction to augment the existing York Durham Sewage System (YDSS). This system is required for the anticipated population growth in Ontario’s York Region. The tunnel is being constructed with four EPBMs and the project has a complex system of 18 shafts, of which 16 and three braced excavations were constructed in advance of the tunnel drives. The overburden soils at the site consist of preconsolidated glacial till and glaciofluvial gravel, sand, silt and clay deposits underlying Ordovician-age bedrock.
Active dewatering was not acceptable for the project due to concerns with affecting the existing groundwater table. Several different approaches to constructing the shaft were used including secant piles, slurry diaphragm wall and shotcrete lining, and in some cases, combinations of these approaches. One of the critical work shafts required on the project was Shaft 11, which was a TBM access and maintenance shaft. This was a 4.65m outside diameter shaft, extending to a depth of 42m. A profile section of the shaft is illustrated in Figure 1.
The proposed excavation support system shown in plan in Figure 1 was a slurry wall panel approach. The intial plan was to construct the slurry panels as shown, excavate the circular shaft and line the shaft with shotcrete as the excavation progressed.
During the excavation of the slurry panels, a combination of very hard soils and and unanticipated boulders delayed and came very close to stopping the slurry wall installation. At this time, the contractor terminated the slurry panel construction at 20m and evaluated other options to provided the temporary earth support and groundwater control to complete the excavation and shaft construction. The slurry panel depth of 20m would prevent inflows and instability from the upper aquifer in the soil profile. Ground freezing was proposed as a method to not only provide lateral support and groundwater control but also provide bottom stability by penetrating deep enough and ‘keying into’ the underlying Sunnybrook impermeable formation. Preliminary evaluation of the soil profile and comparison with frozen soil parameters from other projects indicated that in these soils the frozen earth wall could provide excavation support for the entire depth without any need for additional lining.
The ground freezing process would include mobilization, drilling and installation of freeze pipes as well as an estimated six to eight weeks of freezing. Of particular concern was the six- to eight-week freezing time, when there would be no shaft or tunnel production.
To minimise delays, the contractor decided to proceed with the excavation of the plastic till to a depth of 28m during the formation of the freeze. To avoid any delay in construction schedule, without waiting for the formation of the freezing, it was decided to proceed with excavation and shotcrete from the surface to a depth of 28m, which was located entirely in plastic till with the exception of a 3.2m layer of sand to sand and gravel (stratum 2), which had been already cut off by executed slurry panels. Just before starting of the excavation of the shaft, required freezing pipes were installed and connected to a mobile refrigeration plant and the freezing formation process was initiated.
As the frozen barrier continued to form to secure the deeper parts of the shaft, excavation proceeded in the top 27m with shotcrete being applied concurrently with mining. The shotcrete had been originally designed to be resistant to ground and water pressure after de-activation of the freezing. Because of the freezing application an extra 50mm of shotcrete was applied to provide an insulating layer to protect the freeze.
Another issue, unique to the freezing process was the development of stresses between the existing panels and the shotcrete line. During the excavation, the soil in that would be unfrozen. At some point after the application of the shotcrete, the soil would freeze, expand and generate stresses against the shotcrete liner and possible damage this temporary lining.
This is different from the deeper phase where the shotcrete would be applied after freezing and expansion. A detailed analysis was required to evaluate the soil structure interaction during this phase.
The first step in most freeze designs is the structural analysis. In this phase the thickness of the frozen earth wall surround the excavation is determined. The thickness of the frozen earth wall is governed by several components, the lateral conservative and do not adequately address the deformation or pressures against the shotcrete wall.
For this reason a PLAXIS model was developed to evaluate the stresses and potential deformation and cracking of the applied shotcrete. This model not only computed the internal stresses of the frozen and unfrozen soil, but also allowed for the expansion of the pore water when frozen.
The PLAXIS model shown in Figure 2 is a simplification of the slurry panels. While the panels were actually rectangular, a cylindrical model was used to take advantage of symmetry. Pressure created against existing structures is frequently an issue in ground freezing and often very hard to evaluate. In this case, it was a critical issue as the shotcrete was not reinforced and while cracking could be tolerated, total failure was certainly not acceptable.
Prior to the shaft construction, several geotechnical borings were taken and relatively undisturbed samples were retrieved. Several sealed samples were still available for further testing. Simple tests to evaluate frost expansion were conducted in a geotechnical laboratory. Precise measurements of the sample lengths and perimeter were taken upon extrusion from the sampling tubes.
The samples were then frozen, and the measurements taken again. From this data, it was possible to estimate volumetric expansion of the soil upon freezing. For the plastic till, it was determined that the average expansion was approximately 1.1 per cent.
This expansion was included in the PLAXIS model, and showed that cracking would occur, but not to the extent that there would be excessive damage to the shotcrete. The model also indicated that there was the potential for cracking in the shotcrete lining at depths below the slurry panels. This cracking would be the result of freezing expansion pressure in the deeper soils as well as minimal deformation of the frozen soil upon excavation. While these deformations were also considered acceptable to the overall project, it was deemed necessary to monitor the stresses and deformation in the lining to alert the contractor of deformations greater than anticipated by the modeling. A series of pressure meters between the soil and shotcrete were installed as shown in Figures 3a and 3b. Figure 4 shows the measured deformation in the interior of the shaft. The deformations that were measured were similar to what was indicated in the PLAXIS analysis. An unusual aspect of this data is that the sand material seemed to deform at a greater rate than the fine-grained till. This is contrary to conventional thinking, however the author cautions that generalized conclusions should not be reached, absent frozen soil laboratory compression tests.
Of important consideration of this project is the useful information gained by the pressure cells. Referring to Figure 5, it can be observed that the pressures increase with time. It should be noted that there was an equipment issue on the Radial South instrument that resulted in data loss.
Figure 5 demonstrates that the pressure continued to build. By day 51, the contractor was concerned that the pressures were too high and could damage the lining. In an effort to reduce this pressure, the circulating calcium chloride coolant was warmed from -35 to -10 degree Celsius. Pressures immediately began to decrease. This information provides two very important considerations that can be applied to future freezing projects.
Perhaps the most important consideration is that the data shows that it is possible to decrease frost pressures against structures once a frozen earth wall is formed stresses can be reduced by simply increasing the brine temperature. This process can be applied to many potential future projects.
Another important aspect of this data is the interpretation. A preliminary observation of the data would indicate that the pressures in the lining increased because of the longterm weakening of the frozen earth. However, the fact that the pressures decreased with increasing brine temperatures show that the pattern is reversible and more attributed to the increasing growth of the soil mass. This means that on future projects, ice growth can be controlled somewhat with temperature and possibly flow adjustments. What should be most apparent on this project, however, was that control of the pressures was only possible due to the presence of data obtained by the pressure meters and extensometers. This data provided the information to observe situations, make adjustments and ensure a safe, successful project
