Technical Review of the Lined Rock Cavern Concept and Design ...

Technical Review of the Lined Rock Cavern (LRC) Concept and Design Methodology 32
5.0 REVIEW OF THE LRC DESIGN METHODOLOGY
5.1 General Comments
A brief summary of the LRC design methodology was given in Section 3.3. In general,
the review documents convey a methodology that seems to recognize the need for, and
value of, a phased design approach that builds site-specific geophysical understanding
through incremental site characterization and analysis. Also reflected in the methodology
is the use of the Observational Approach during cavern excavation to treat unexpected
conditions that may require changes to the final LRC design. Thus, the methodology
shows recognition of the unique and uncertain characteristics of a rock mass. This is
further reflected, in a seemingly disciplined manner, through the use of a probabilistic
design approach.
Two key LRC design criteria are expressed in the review documents: (1) safety against
ground uplift; and (2) maximum induced strain range in the steel liner. Both criteria,
which are closely tied to the mechanical response of the rock mass, address the essential
LRC design aspect of limiting the strain induced in the steel liner — that is, the rock mass
overburden must be sufficient to provide effective resistance to the maximum cavern
pressure.
The LRC design methodology oversimplifies the evaluation of ground uplift by its use of
the rigid-cone and log-spiral limit-equilibrium models. Independent calculations using a
numerical model (refer to Sections 4.1.2 and 4.1.3) show that the soil-anchor analogy
used in the development of the rigid-cone and log-spiral models is not well-suited to describe
the more complex loading and material response of LRC conditions. Both the
rigid-cone and log-spiral models are exceedingly conservative, primarily because they
oversimplify the mechanistic aspects of ground uplift associated with a pressurized, lined
rock cavern. While uplift response can be evaluated mechanistically with relative ease
using a numerical model, there seems to be an objection in the review documents to using
a numerical model for this purpose. The documents suggest that such models are unable
to provide a reliability measure and have difficulties in handling large deformations. For
most available finite-element or finite-difference numerical codes in rock mechanics
(e.g., FLAC, in these analyses), neither of these arguments is accurate.
The two key LRC design criteria mentioned are central to the FLRC1 and FLRC2 models
used as part of the design methodology to evaluate the feasibility and the design of lined
rock caverns. These models represent streamlined calculation procedures that rely on
rock index properties and empirical relations to estimate the rock-mass mechanical properties
(i.e., stiffness and strength), and limit-equilibrium, finite-element and analytical
(homogeneous and isotropic) models to estimate cavern location (i.e., depth), maximum
gas pressure, cavern deformations, and steel-liner strain. Thus, these models represent
knowledge-based expert systems for LRC siting and design.