A total of four activities are defined:

#### A3.1 Probabilistic assessment of penetration resistance of a bucket foundation

In 2014 and 2015 GRAM A3.1 combined achievable penetration depth with failure probability of suction installed caissons in sand addressing the following question: What adjustment factor should be used to achieve the same probability of failure if the relative density changes?

The work in 2016 will include developing a method for combining failure probability and for installation and capacity design, use installation information to update the failure probability of the as installed caisson capacity, use installation information to update the foundation stiffness.

#### A3.2 TBT concentration in the Drammen fjord

The research in 2014 and 2015 included developing methods of probabilistic analyzes of environmental problems with a focus on the following questions: What is the probability that the TBT concentration will fall under EQS- limit within an X number of years? A probabilistic approach of the hazard was also included. The case of natural decrease of TBT concentration in the Drammen fjord was investigated as a test case.

In 2016 a methodology to communicate the results of a probabilistic approach will be developed.

#### A3.3 Excavation - Uncertainty related to characteristic shear strength

The case study chosen for land foundations was related to the interpretation of undrained shear strength profile for an excavation problem. For this exercise, a location south of Oslo, near Lysaker, where large amount of samples were taken and special laboratory tests were carried out, was chosen. The available data were sent to a number of geotechnical engineers in NGI-Oslo and NGI-Trondheim, both in onshore and offshore divisions. Additionally anonymous data were sent to an "expert-panel" (a NIFS project including Vidar Gjelsvik from NGI), which is working on making guidelines on how to choose a design su-profile from different data sources.

The available data are also being processed statistically to derive the best estimate and standard deviation of the undrained shear strength profile. One of the main lessons from this exercise is compare the characteristic profile(s) developed by the "experts" with the calculated mean values and standard deviations. The results will be used to do a probabilistic design for a sheet pile wall at the excavation location.

#### A3.4 Probabilistic assessment for a quick clay slide

The example application chosen for natural hazards (landslide) was to assess the impact of uncertainties in shear strength parameters on the reliability of the performance prediction for a quick clay slope. A multivariate normal model among undrained shear strength, liquidity index, sensitivity and overconsolidation ratio of sensitive clays was developed. This model can reduce the uncertainty (COV) in undrained shear strength effectively using a Bayesian framework.

LOGIN

^{→}COOKIES | PRIVACY STATEMENT