Random set Finite element method in excavation

1
Study Evaluation of Random Set Method on Results
from Reliability analysis of Finite Element in
Deep Excavation
Article Code 443
Presenter Mehdi Poormousavian Other Authors Ali
Fakher
2
Outline

• Basic concepts
• Random set Finite Element
• Application to deep excavation
• Comparison with Point estimate
• Conclusion

3
Factor of safety and probability of failure
Concept
RS-FEM
Case Study
Result
Point Estimate
Conclusion
4
Reliability methods
Concept
There arent enough point data, instead interval
information are applied.
Wide range of point input data from soil tests
are at engineer disposal. more accurate but needs
more data.
RS-FEM
Case Study
Result
Point Estimate
Conclusion
5
Random Set Finite Element Method
Concept
RS-FEM

• Random set method
• Introduce by Dempster 1967,
• Random set method combine Finite element
  method(RS-FEM)
• Peschl 2004 and Schweiger 2007 illustrated RS-FEM
• Nasekhian 2011 used RS-FEM for tunnel
• Ghazian and Fakher 2014 used RS-FEM for excavation

Case Study
Result
Point Estimate
Conclusion
6
Concept
RS-FEM
Case Study
Result
Point Estimate
Conclusion
7
Saba Project
Concept
RS-FEM

• Northern wall particulars
• Depth 27 m
• Anchor 6 in height, 6 m bond length
• Soil touched soil(0-2)m, sand layer(2-10)m, sand
  layer(10-27)m

Case Study
Result
Point Estimate
Conclusion
8
First step of RS-FEM
Concept
RS-FEM

• Definition of Geometry or geometrics
• Software Plaxis V8.5 2D
• Soil behavior model Hardening-soil

Case Study
Result
Point Estimate
Conclusion
9
Concept
RS-FEM

• Software Plaxis V8.5 2D
• Soil behavior model Hardening-soil

Case Study
Soil Depth(m) water ? (KN/m3) c (KN/m2) ? () ? () E50 (KN/m2) Eur (KN/m2) m K0nc (KN/m2)
Touched 0-2 Drained 16.5 5 27 0 15000 45000 0.5 0.546
Sand 2-10 Drained 19 35 34 4 68000 204000 0.78 0.441
sand 10-27 drained 20.7 65 38 8 130000 390000 0.5 0.384
Result
Point Estimate
Conclusion
10
Second step of RS-FEM
Concept
RS-FEM

• Selection of input parameters

Case Study
soil soil probability c (KPa) ? degree E (MPa) ? (KN/m3) ? degree m anchor degree loading (KPa)
Touched Set 1 0.5 6-11 24-32 10-22 13.5-16.5 1-8 0.44-0.64 7-27 6-16
Touched Set 2 0.5 12-22 27-37 14-28 15.5-18.5 4-11 0.56-0.76 13-33 9-19
sand Set 1 0.5 26-36 26-33 56-92 18-21.6 1-8 0.45-0.75 7-27 6-16
sand Set 2 0.5 30-41 29-37 79-110 20-23.5 4-11 0.65-0.95 13-33 9-19
sand Set 1 0.5 48-71 31-41 105-135 17-20.7 1-8 0.35-0.65 7-27 6-16
sand Set 2 0.5 73-94 40-45 124-155 19.4-23.3 4-11 0.55-0.85 13-33 9-19
Result
Point Estimate
Conclusion
11
Third and fourth steps of RS-FEM
Concept
RS-FEM

• Reduce uncertainty via Variance reduction
  technique by Vanmarcke
• Sensitivity analysis

Case Study
Result
Point Estimate
Conclusion
12
Result from Sensitivity analysis
Concept
RS-FEM
Case Study
Result
Point Estimate
Conclusion
13
Main compound result from RS-FEM
Concept
RS-FEM
Case Study
18-40
RS-FEM
Result
Designed
Observed
Point Estimate
Conclusion
14
Main compound result from RS-FEM
Concept
RS-FEM
RS-FEM
Designed
Best fitted Rice distribution PF1.7E-06
Case Study
Best fitted Lognormal distribution PF1.19E-15
1.52-2.08
Result
Point Estimate
Conclusion
15
Compare RS-FEM and Point estimate method
Concept
RS-FEM
Case Study
Result
1.8s
Observed
PEM
Point Estimate
Conclusion
16
conclusion
Concept
RS-FEM

• The number of finite element analysis by RS-FEM
  is 64 which is much lower than a Monte Carlo
  method.
• horizontal displacement of excavated wall is
  gained between 18 to 40 mm, which versatile with
  field measurements (15 to 30 mm).
• Using RS-FEM before starting excavation leading
  to assurance for employers to be aware of
  displacements so that designs are optimized.

Case Study
Result
Point Estimate
Conclusion
17
doubt is an uncomfortable condition, but
certainty is a ridiculous one.
Introduction
Literature review
Purpose
Voltaire (1694-1778)
Material and sample preparation
Results and discussion
Conclusion