Ted Williams Tunnel

Approximately 7 kilometers (4.3 miles) of Central Artery/ Tunnel (CA/T) slurry walls will ultimately become the permanent walls of the tunnel structure. In the majority of the alignment, vertical loading as opposed to lateral stability governs wall depths. Costly excavation of rock will be necessary to reach the required depths, which commonly exceed 36.6 m (120 feet).

The selection of appropriate values for end bearing and side resistance is critical to keeping project costs within budget, given the considerable volume of wall construction, because these values dictate the depths to which the walls must be embedded into bedrock. The Massachusetts Highway Department authorized the B/PB team to determine these values
ultimately (as opposed to the Area Geotechnical Consultant or Section Design Consultant) because similar slurry construction is common to multiple design sections and B/PB is responsible for maintaining design consistency.

We found that developing these design parameters highlighted several challenges that are frequently encountered in geotechnical engineering:

• Establishing design parameters in highly variable ground conditions
• Contending with dissimilar classifications (by different consultants) of the same core samples
• Assessing the risk of various “failure mechanisms” of a given structure to establish a corresponding level of conservatism in the selection of geotechnical design values
• Maintaining a design that falls within budgetary constraints.

Earlier Design Recommendations

Several years before the detailed design of the slurry walls was underway, the Area Geotechnical Consultant (AGC) for the CA/T Central Area—the region where slurry walls must resist vertical loads—developed the recommended parameters for design of the slurry walls. The recommendations included values for allowable end bearing and side resistance for glacial soils and various classifications of bedrock. Of the two sets of design values for allowable
end bearing-settlement controlled and strength controlled-settlement controlled values were recommended.

Settlement Controlled Values. The settlement controlled values were derived from an assumed tolerable settlement
of the slurry walls of 25 mm (1 inch), which applies to the condition where the slurry walls are acting as underpinning elements that support the existing elevated highway structure. The tolerable total and differential settlements that the elevated structure could withstand were represented as typical values for building construction. Allowable end bearing pressures of the B2 and B3 rock (severely to moderately weathered Argillite and moderately to slightly weathered Argillite, respectively) were back-calculated using elastic parameters and an assumed 25.4 mm of compression of the rock mass. To derive the corresponding value for B1 rock (completely to severely weathered Argillite), the AGC assumed the rock had similar engineering properties to those of a hard clay. A one-dimensional consolidation analysis was performed, assuming a recompression ratio of 0.02 and 25.4 mm of settlement to back calculate an end bearing pressure.

Strength Controlled Values. Strength controlled values for allowable end bearing were derived typically from results of unconfined compression tests of core samples. Allowable end bearing pressures were calculated primarily by classical methods of determining bearing capacity. The shear strength of the rock for these calculations was assumed to be one-half the unconfined compressive strength.

The AGC’s recommended design values appeared reasonable at the time they were made. As final design progressed, however, they led to calculated wall embedments that seemed excessive.

CLTP Used to Verify Design Parameters

We took advantage of a caisson load test program (CLTP) that was slated for another CA/T design area (with similar geologic features to those of the Central Area) to verify the slurry wall design parameters. The CLTP consisted of the installation of a series of drilled shafts that were then tested with Osterberg load cells. A test boring performed at each shaft location was logged by the SDC for the design section targeted by the CLTP. Pressuremeter tests (PMT) were performed in each boring. Shafts were socketed in zones representative of the full range of weathering and quality of the Cambridge Argillite, the bedrock formation that underlies the entire project alignment. The PMT results were then compared to the boring logs.

The comparison of logs was the source of complications. As mentioned, the AGC had used a modified rock description system and divided the bedrock into the classifications of B1, B2, and B3., but the SDC used the standard project-wide classification system and did not subdivide the bedrock. To correlate the CLTP results to the slurry walls, the boring logs needed to be “translated” and the Central Area AGC was requested to log the CLTP borings.

In our comparison of the two sets of logs, inconsistencies in descriptions of weathering, which resulted in shifts in the bedrock classifications, were noted-not surprisingly. We knew we had to examine the core samples and reconcile the differences in classifications.

Once the rock was classified on a consistent basis, CLTP results were transferable to the Central Area. It was not
possible to correlate the PMT data to rock classification (and corresponding design values), however. The PMT is sensitive to borehole conditions and instrument placement and the data obtained exhibited wide variations in similar strata.

Development of Project Design Parameters

The design parameters developed subsequently by B/PB apply only to the Central Area slurry walls. Load transfer mechanisms in the sockets of the slurry walls may vary from those of drilled piers as a result of differing geometries, installation methods, and the consequences of exceeding theoretical load carrying capacity, which is discussed below in greater detail. Additionally, drilled pier capacities need to be assessed more critically because structural loads are resisted by discrete elements that often act essentially independently. The slurry wall has the ability to transfer loads longitudinally from zones of lower bearing capacity to zones of greater vertical bearing capacity.

Modeling of Tunnel Behavior. A key step in determining foundation design parameters for a given structure is to identify the potential modes of failure. Once these are established, appropriate parameters and design values
for soil and rock capacity can be specified. The following discussion applies only to the case where the tunnel invert bears on the marine clay stratum.

The maximum loading condition of the Soldier Pile Tremie Concrete (SPTC) walls occurs when the tunnel structure is completed and backfilled. At this stage the SPTC walls are integrated with the tunnel structure through continuous structural connections to both the invert and roof slabs. In such a configuration, deflection of the wall becomes a function of the stiffness of the combined wall/roof/invert section. The resultant flexural stiffness in the longitudinal direction will enable the structure to safely span considerable distances. Detrimental settlement of the tunnel structure should not be anticipated unless rock quality is considerably more adverse than assumed in design, on a continuous basis, for a longitudinal distance of 30 m (100 feet) or more. The probability of all variations in rock properties and stratification between existing borings becoming significantly worse than presently shown on the subsurface profiles is remote.

A closer look at the mode of “failure” or deformation of the tunnel provides further evidence of the unlikelihood of detrimental settlement. If the actual bearing capacity of the bearing stratum of the slurry wall is exceeded along the length of the wall, the walls will begin to settle. As the walls move downward, the invert will be drawn down as well, and contact pressures between the invert slab and subgrade will increase, resulting in a simultaneous decrease in wall loads. The
proportion of load that is transferred to the invert is a function of subgrade modulus. Additionally, if the slurry wall settles, side resistance along the entire height of the wall is mobilized. The side resistance of only the strata located below tunnel invert are taken into account in calculating wall depths.

In terms of the present structural model, as the bearing capacity is approached and exceeded, the stiffness of the wall support will considerably decrease. considerably. The vertical pressure thus becomes more uniformly distributed among the wall and invert. Note that a “stable” design can be achieved if the tunnel bears entirely on the clay. However, because the slurry walls are installed to act as underpinning elements and because it was elected to implement an integrated design, the walls bear in the rock and thus attract loads in the completed tunnel configuration.

Regardless, the fact remains that the SPTC walls will attract much higher vertical pressures in service conditions as a result of the variation of stiffness between the rock and marine clay. The SPTC walls need to extend to appropriate depths in the rock to resist these loads. Thus, allowable bearing pressures should be based on strength, or bearing capacity, as opposed to control of deformations.

Basis for selection of design values. Design values for slurry walls in the Central Area (Table 1 on page 45) were derived from results of the CLTP, analysis of data obtained by the AGC, and data and design parameters associated with other projects in the Boston area. The CLTP results could not be used as the sole source of design values because the entire range of rock properties identified in the Central Area were not tested. Additionally, three of the CLTP borings were located at a distance greater than 20 feet from the test location, and thus the stratification in the core samples may not correspond to those of the test sockets.

The revised values, particularly those associated with end bearing, exceed those recommended several years earlier.
As stated, the original recommendations represent settlement controlled values relating to underpinning of the existing highway structure. Underpinning does not govern wall design, however, because the most extreme loading occurs in the completed tunnel condition when settlement becomes a secondary design consideration and, therefore, strength controlled values were used in determining the revised values.

The CLTP results indicate that allowable end bearing values greater than those shown above could have been selected, particularly in the cases of B2 and B3 rock. The selection of lesser values was deemed prudent, however, because the stratification along the slurry walls will not be verified on a continuous basis.

Design values for the till were derived from CLTP results, the geotechnical reports, and previous projects. There exists considerable design and performance experience in the till in the Boston area. To develop values for the B1 rock, its measured engineering properties were compared to those of the till. It was concluded that the slurry wall design values for the till and B1 rock should be similar. Note that these values are typical of those used in foundation design for similar rock quality in the Boston area. Values for B2 and B3 rock were derived primarily from CLTP results.

Conclusion

Implementation of the revised design values resulted in decreases in the quantity of costly slurry wall excavation in rock. The action taken by B/PB represented one example of the fulfillment of our obligation to control project costs while maintaining quality in the constructed product.