The los angeles county metropolitan Transportation Authority’s (Metro) Regional Connector Transit Corridor (RCTC) project is a USD 1.18bn design-build contract awarded in 2014 to Regional Connector Constructors (RCC), a joint venture of Skanska and Traylor Brothers. Mott MacDonald is RCC’s principal designer. The project totals 3.1km of new light rail infrastructure with 1.6km of twin bored tunnels, 1.1km of cut-and-cover tunnels, three cut-and-cover stations, 0.3km of at-grade alignment, and a crossover cavern. Project construction is underway, with expected construction completion in 2020. A location map showing the alignment, stations, and crossover cavern is shown in Figure 1.
The cavern is located 15m below the centerline of 2nd Street and will be constructed from 2nd/Broadway Station. Once the cavern excavation is complete, permanent lining and internal structures will be built to accommodate crossover infrastructure, emergency egress, and an overhead ventilation plenum. A rendered perspective from within the cavern showing the finished lining, rail infrastructure (excluding crossover), and plenum structure is shown in Figure 2.
GEOLOGIC AND GEOTECHNICAL CONDITIONS
The geologic setting in the SEM cavern area consists of 2 to 3m of artificial fill (Af), underlain by about 3 to 5m of coarse-grained alluvial deposits (Qal2). The coarse-grained alluvium is underlain by the Fernando Formation bedrock, an extremely weak to weak massive clayey siltstone, that comprises an upper layer of moderately to highly weathered material, with a slightly weathered to fresh layer underneath. The Fernando Formation has relatively low hydraulic conductivity (average value of 1×10-5 mm/sec), so groundwater infiltration during excavation is expected to be limited to seeps that originate primarily from rock discontinuities.
Short-term-loading geotechnical parameters were selected for analysing a relatively rapid loading of the Tf during SEM excavation and seismic events, while drained parameters were considered for the overlying soils and long-term static response of the Tf. The Mohr-Coulomb strength failure criterion and a linear elastic-perfectly plastic deformational behavior were selected for design for all geologic units.
The design team had the opportunity to observe the massive and impermeable nature of the Fernando formation, as well as its standup time, during the construction of a nearby Tieback Removal Shaft and Adit (TRSA). This was a temporary structure designed by MM and constructed by RCC in the last quarter of 2016 to remove existing tiebacks in the path of the TBM. Through participation in the daily review meetings during the construction of the shaft and adit, lessons learned from the TRSA construction were incorporated into the SEM cavern risk register and design.
EXISTING SITE INFRASTRUCTURE
The cavern will be built below portions of three adjacent buildings and directly below a large LA County storm drain and operating roadway. A priority of the SEM evaluation was to determine if the Metromandated settlement limits of 13mm and an angular distortion limit of 1/600 could be met along the length of the cavern. A critical cross-section has the cavern between the Higgins Building, a 10-story Los Angeles Historical-Cultural Monument that was built in 1910, and the 10-story headquarters of the Los Angeles Police Department (LAPD) that was built in 2009. The Higgins building has two basement levels and slightly overlies the south side of the cavern. A one-level underground parking structure, adjacent to the LAPD headquarters, lies to the north of the cavern. The storm drain is a reinforced concrete structure that measures 3×3m and is located 5m above the vertical centerline of the cavern. A cross-section showing the existing site constraints is provided below in Figure 5.
DRIFT CONFIGURATION AND CONSTRUCTION SEQUENCE
Close collaboration of the designer and contractor is a key feature for SEM construction success that is clearly enhanced by a design-build delivery scheme. Additionally, in transportation projects of the scale of RCTC, global project schedule modifications can impact the design of structures and require adaptation. In the case of the crossover cavern, the initial design called for a two-drift sequence with a single sidewall and excavation advance lengths of 0.9m in top headings and 1.5m in benches. Two major accommodations that led to this configuration were that the left track bored tunnel would be completed prior to the SEM cavern excavation and the hammerhead at 2nd/Broadway station, both of which essentially precluded the possibility of a three-drift configuration. This design option is described in detail in Herranz et al. (2016) and is shown in Figure 6.
Subsequent to the development of this concept, other project constraints required an optimisation of the overall project schedule. To reduce RCTC duration, it was agreed that both bored tunnels would need to be excavated prior to SEM construction. This resequencing required adaptation of the SEM configuration and bored tunnels alignment optimisation. A three-drift scheme with two sidewalls (see Figure 7) was selected as the lowest-risk option, which also eliminated some challenges of the two-drift configuration including the cruciform joint connections between drifts that attracted significant bending moments and shear.
The first 18m of cavern excavation will be protected with a 53-unit pipe canopy system. The pipe canopy will be installed from the 2nd/Broadway station excavation prior to the commencement of the SEM cavern excavation. The pipes will be installed on 30cm centers along an arc of 90 degrees centered on the tunnel centerline and will have an OD of 140mm and a 13mm wall thickness.
INITIAL LINING DESIGN
The initial lining comprises 305mm of 34.5 MPa (5,000 psi) -fiber-reinforced shotcrete, lattice girders, supplemental local reinforcement in corners, and mesh in specified locations in the temporary sidewalls.
The design of the SEM cavern initial lining required a detailed simulation of the excavation sequence, geometry, ground characteristics, and existing structures. A three-dimensional multi-staged FLAC3D model was developed to consider all these features (Figure 8).
The use of the 3D model was essential to accurately assess ground relaxation in context with the excavation face. The Convergence-Confinement method typically used in 2D modeling can determine ground/support interaction in a two-dimensional plain strain analyses (AFTES 2001). The key issue is how to obtain the confinement loss or relaxation of the ground that occurs before placement of the initial lining. Several approximations are presented in technical literature (Bernaud et al, 1991; Bernaud and Rousset, 1992; Nguyen et al, 1993; Carranza and Fairhust, 2000 or Vlachopoulos and Diederichs, 2009) but they all consider full–face circular excavations with homogenous ground parameters, which differs significantly from the oval, multistage, shallow excavation of the SEM crossover cavern within elastic– plastic ground.
The curing of shotcrete over time was also considered in the models based on a correlation developed by Chang (1994) and reproduced in Thomas (2008), and the estimated duration of each construction stage was agreed with RCC (Figure 8, right). The structural design of the initial lining was performed following Metro Rail Design Criteria (2013) and codes and standards referenced therein, including ACI 318 (2014). FLAC3D’s flexibility allowed for customised capacity verifications embedded in the model (Figure 9).
Face stability was analysed with empirical formulations (Vermeer et al 2002; Kavvadas et al 2009) and with continuum-based analyses in FLAC3D as well as block stability in Unwedge 4.0 (Rocscience), all of which indicated that each heading, with the expected ground conditions, would be inherently stable without the need for continuous pre-support or face bolting.
If ground conditions differ from anticipated conditions, several supplemental support or “tool box” items were defined for construction, including spiling, fiberglass face dowels, dewatering and face stabilisation wedges. The use of “tool box” measures will be implemented as-needed to control displacements and provide stability for the excavation and existing structures.
SETTLEMENT ASSESSMENT
The settlement assessment for the two-drift base option included an extensive sensitivity analysis to evaluate the impact of various conditions, including the alluvium depth, Young’s modulus of the Fernando Formation, soil constitutive models, in-situ horizontal to vertical stress ratio (ko), excavation drift configuration, initial lining stiffness, adjacent building and foundation stiffness and pre-support measures. At the final stage of construction, the three-drift configuration showed very slight differences with the two-drift Base Option. This fact is attributed to the relatively favorable mining behavior of the Fernando formation and its limited yield zones predicted around the SEM excavation.
The equivalent ‘volume–loss’ with the final design configuration and expected ground parameters in the models was under 0.3% and the maximum displacement at surface level was approximately 16mm, with expected maximum building foundation settlements less than the 13mm contract limit.
The magnitude of the impacts to the adjacent building associated with the ground movements was determined to range from negligible to slight, corresponding to induced tensile strains under 0.15 per cent, when assessed in accordance with Boscardin and Cording (1989) and Burland et al. (2001) criteria. Impacts to the LA County storm drain were analysed using FLAC3D to assess the potential impacts to the structure when subjected to longitudinal settlements during SEM construction. The angular cavern, ground response evaluation, and the performance of the initial shotcrete lining. Instrumentation systems will include arrays of convergence monitoring points, ground surface settlement points, multi–point borehole extensometers, inclinometers, piezometers, tiltmeters and monitoring points on buildings, and utility monitoring points on the storm drain. Based on the numerical modeling results, action and maximum levels at each construction stage will be established. Displacements measured in the field during construction will be compared with those limits and those obtained in the numerical models. When necessary, SEM construction will be modified in the field with the ‘tool box’ items to keep displacements within allowable limits and maintain stability of all structures.
GAS AND WATERPROOFING
The RCTC alignment lies within the Methane Zone and Methane Buffer Zone, as defined by the City of Los Angeles Ordinance 175790. Correspondingly, the possibility of encountering methane gas and/or hydrogen sulfide during construction or subsequent operation of the LRT system is a known risk. The underground work has been classified as ‘potentially gassy’ by Cal/OSHA, however the concentration, pressure, and volume of these gases is expected to be sufficiently low to be mitigated within the excavation through adequate ventilation. As mitigation against long term operating risk, Metro technical specifications prescribe the use of methane and water resistant barriers.
A hydrocarbon resistant (HCR) high-density polyethylene (HDPE) membrane will be provided between the initial support and final lining over the full extent of the cavern. The HCR membrane will include a compartmentalisation system, with the combined use of water barriers and contact grout pipes and remedial grout tubes to ensure there are no gaps between the initial and final lining and seal potential leakages. Also, waterproofing was included at all interfaces with the bored tunnels and station to avoid water inflow at these locations.
PERMANENT LINING DESIGN AND ANALYSES
The final lining will be 28MPa (4000 psi) cast-in-place concrete with a minimum thickness of 460mm (18 in) in the crown and walls and a maximum thickness of approximately 1.8m (5 ft 9 in) in the invert. Due to the different stiffness of the cavern, station structure and the bored tunnels, flexible joints are provided at the interfaces between these structures.
The Metro design criteria for the final lining design requires a 100-year service life, neglecting any beneficial contributions of the initial lining to resist ground loads, hydrostatic pressures, live loads within the tunnel and on surfaces above the tunnel, and seismic loads. Ground-structure interaction analysis for the final lining design was performed with numerical models in FLAC and SAP2000.
The seismic design was initially performed using a pseudo-static analysis. Given the complexity of the cavern geometry and internal structure – including wall and plenum slab, a dynamic analysis of the cavern was performed. Lianides et at (2017) discusses the model set up and preliminary results obtained, the final findings of this analysis are currently being finalised. The final cavern structure includes a 600mm-thick headwall at the interface with the bored tunnels at the east side of the cavern. The analysis of the headwall, and its interaction with the perimeter final lining and the center wall (that acts as an internal buttress), was analysed with a 3D soil-structure interaction model in SAP2000, using static loads and seismically induced displacements.
CONCLUSIONS
The SEM crossover cavern in the Regional Connector Transit Corridor project in downtown Los Angeles, once completed, will be the largest tunneled excavation in Los Angeles. The cavern was initially designed with a two-drift configuration with top heading and bench, but a project schedule optimisation led RCC and MM to select a three-drift configuration while also limiting risks and improving constructability.
Design-build contracts promote close collaboration of contractors and designers, which was a key advantage in planning a SEM tunnel of this magnitude. The project exemplified its importance in large transportation projects, where design flexibility is crucial to accommodate global project changes. Also, for a project of similar size to RCTC, there are great benefits to have the designer present during construction of similar underground structures. Valuable information and lessons learned from observations of similar ground and its behavior during construction will only improve the SEM design process.
