Dam Stability Assessment – Limit Equilibrium Analysis (LEA) or Finite Difference Analysis (FDA)?

By Red Earth Engineering In Geotechnical Engineering, Tailings and Water Dams No comments

Limit equilibrium analysis (LEA) for slope stability assessment became popular since Bishop’s method of slices was first introduced in 1955, followed by a number of variations. Although advanced numerical methods such as finite element analysis (FEA) or finite difference analysis (FDA) have been actively developed since personal computers became available, engineers keep holding on to LEA for various – more or less valid – reasons:

  • because it is ‘industry standard’
  • because everyone understands it
  • because numerical models are too hard to develop / require too many input parameters…

At REE we believe that numerical methods have reached maturity and robustness that no pro-LEA arguments can oppose. Software solutions focused on the factor of safety approach using the Strength Reduction Method (SSR), such as Plaxis, Optum G2, and the recently released FLAC/Slope, now allow rapid and accurate assessment of dam stability under various loading configurations. Automatic geometry and meshing tools, graphical definition of phreatic surfaces, and straightforward calculation routines mean that stability assessments can be completed as fast – if not faster – than using traditional limit equilibrium packages.

Some other strong arguments in favour of numerical methods include:

  1. Failure mechanisms (slip surfaces) develop naturally following mechanical principles, not using arbitrarily selected slip circles.
  2. The average factor of safety approach used in the method of slices can be unsafe in certain conditions where sensitive/brittle materials are present.
  3. Full stress-strain analysis can be undertaken in FEA/FDA using advanced material models (e.g. NORSAND).
  4. The deformation field at the “failure state” is a kinematically admissible deformation, whereas some artificially imposed failure planes (e.g. some deep wedge failures in soils) in LEA may be incorrect.

We present below a few comparative analyses that we carried out using Geostudio Slope/W (LEA) and Itasca FLAC 8 (FDA). Both the failure surfaces and estimated factors of safety are very close. Our goal is to build a database of case studies that will help bring confidence to our clients and peers that numerical methods are the way to go!

Contact us now if you want to know more about our modelling capabilities.

Case 1 – Short term, high groundwater level

Case 2 – Long term, low groundwater level

Case 3 – Post-seismic