0% found this document useful (0 votes)
18 views

Applications of Artificial Intelligence and Machine Learning

Uploaded by

ziad.mostafa901
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
18 views

Applications of Artificial Intelligence and Machine Learning

Uploaded by

ziad.mostafa901
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 34

Supporting Partner

Applications of Artificial Intelligence


and Machine Learning in Geotechnical
Engineering
Rodaina ABOUL HOSN
Associate Professor at ESTP-Paris

Sustainable Geotechnical Engineering Summit / 5th December 2022


About this Presentation

1. Overview of Artificial intelligence (AI) and Machine Learning (ML)

2. AI in Geotechnical engineering - Examples of applications :

- Scale of Construction sites


- Laboratory scale
- Micro - scale

3. Benefits and Limitations of AI

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
Overview of AI and ML

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
Overview of AI and ML
What exactly is AI?

AI is a computational technique that attempts to mimic, in a very simplistic way, the


human cognition ability (e.g. brain, genes, nerve system) to think and learn on its own.

It is a simulation of human intelligence into machines to do tasks that we would


normally rely on humans to perform.

©digitalhealth
Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
Overview of AI and ML
What exactly is AI?

AI models are data-driven models which means that they rely on the data alone to
determine the structure of a phenomenon (or system) without the need for
assumptions or simplifications about that system, which is in contrast to most
physically-based modelling techniques.

Organized by ©moidev
#THEBIG5EXHIBITION
www.thebig5.ae
Overview of AI and ML
AI and ML

AI refers to the capability of a machine to


perform a certain behavior; while ML is the
algorithm that learns patterns from datasets to
predict future outcomes, recognize patterns, or
suggest different classes to the data.

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
Overview of AI and ML
Major AI Techniques

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
AI in Geotechnical Engineering

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
AI in Geotechnical Engineering

The use of different AI techniques in some applications in geotechnical engineering

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
Beghbani, A., Choudhury, T., Costa, S., Reiner, J. (2022). “Application of artificial intelligence in geotechnical engineering: A state-of-the-art review” Earth-Science Reviews 228,
https://doi.org/10.1016/j.earscirev.2022.103991.
AI in Geotechnical Engineering
Artificial Neural Networks

Concept

©book: Memristor and Memristive Neural Networks

©Medium

Organized©TheEconomicTimes
by
#THEBIG5EXHIBITION
www.thebig5.ae
AI in Geotechnical Engineering
Artificial Neural Networks (ANN)

Concept

©book: Memristor and Memristive Neural Networks

Perceptron
𝑥1
𝑤1
𝑛
𝑥2 𝑤2 σ𝑛𝑖=1 𝑥𝑖 𝑤𝑖 𝑦
+θj 𝑓 ෍ 𝑥𝑖 𝑤𝑖 +θj
Organized by
...

𝑖=1
#THEBIG5EXHIBITION
𝑤𝑛 𝑤𝑖 : weights www.thebig5.ae
𝑥𝑛 Θj : bias
Examples of Applications:
Scale of Construction Sites

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
AI in Geotechnical Engineering
Examples of Applications

Dam Safety
The Diablo Dam was built in 1936 along the Skagit River in
©hdrinc.com northwestern Washington.

To ensure the dam’s safety, a digital twin (virtual replica of the


dam) was created merging real data obtained with embedded
sensor network and incorporating machine learning and
artificial intelligence to perform predictive analysis and
determine how asset and geotechnical conditions would
https://www.hdrinc.com/insights/digital-twin-diablo-dam-comes-life
change with time (such as natural shifting, erosion of the
surrounding soil, automatically identify cracks and spalls)
allowing operators to take corrective actions to immediately
improve their targeted maintenance scheduling and help
ensure safe operations. Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
AI in Geotechnical Engineering
Examples of Applications

Slope Stability Safety on Construction Sites


©blog.strayos.com

©cherrystdwfi.com.hk

- Real time monitoring of the site


- Plan ahead for optimum slopes
Organized by
- Identify any misuse of personal protective equipment (e.g. safety helmet or vest) #THEBIG5EXHIBITION
- When a dangerous act is predicted, the system will alert the safety officer and inspector to
www.thebig5.ae
avoid injuries
AI in Geotechnical Engineering
Examples of Applications

Prediction of the load-settlement curve of Driven Piles

Several in-situ full-scale pile load tests, as well as


cone penetration test (CPT) data were used to
develop AI model.

The tests were conducted on sites of different soil


types and geotechnical conditions, ranging from
cohesive clays to cohesionless sands including
layered soils

Organized by
#THEBIG5EXHIBITION
Shahin, M. A. (2014). “Load-settlement modelling of axially loaded drilled shafts using CPT-based recurrent neural networks.” International Journal
www.thebig5.ae
of Geomechanics, ASCE, 14(6),
06014012(1-7).
Shahin, M. A. (2014). “Load-settlement modelling of axially loaded steel driven piles using CPT-based recurrent neural networks.” Soils and Foundations, 54(3), 515-522.
AI in Geotechnical Engineering
Examples of Applications

Prediction of the load-settlement curve of Driven Piles

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
AI in Geotechnical Engineering
Examples of Applications

Prediction of the load-settlement curve of Driven Piles

Performance measure Training Validation


Coefficient of correlation, r 0.997 0.994
Organized by Coefficient of Determination, R2 0.993 0.974
#THEBIG5EXHIBITION
Shahin, M. A. (2014). “Load-settlement modelling of axially loaded drilled shafts using CPT-based recurrent neural networks.” International Journal
www.thebig5.ae
of Geomechanics, ASCE, 14(6),
06014012(1-7).
Shahin, M. A. (2014). “Load-settlement modelling of axially loaded steel driven piles using CPT-based recurrent neural networks.” Soils and Foundations, 54(3), 515-522.
AI in Geotechnical Engineering
Examples of Applications
Prediction of the load-settlement curve of Driven Piles
How engineers can use the developed AI model?
AI models can be used to establish a relationship between inputs and outputs which can be implemented in a
simple user-friendly application.

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
Examples of Applications:
Laboratory Scale

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
AI in Geotechnical Engineering
Examples of Applications
Prediction of the Mechanical behavior of Railway Ballast

AI Model Features:

D50 : diameter at which 50% of the specimens pass


through the sieve
Cu : coefficient of uniformity
Diameter = 300 mm Cc : coefficient of curvature
Large-scale triaxial
apparatus Height = 600 mm e: void ratio
ϒ : bulk unit weight (kN/m3)
𝜎3 : confining pressure
εi : axial strain (%)
Δεi : axial strain increment (%)
qi : deviator stress (kPa)
εv,i : volumetric strain (%)
Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
Shahin, M. A., Indraratna, B. (2006). “Modeling the mechanical behaviour of railway ballast using artificial neural networks”, Canadian Geotechnical Journal, 43: 1144-1152.
AI in Geotechnical Engineering
Examples of Applications
Prediction of the Mechanical behavior of Railway Ballast

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
Shahin, M. A., Indraratna, B. (2006). “Modeling the mechanical behaviour of railway ballast using artificial neural networks”,
Canadian Geotechnical Journal, 43: 1144-1152.
Examples of Applications:
Micro - Scale

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
AI in Geotechnical Engineering
Examples of Applications
Fabric and Effective Stress Distribution in Internally Unstable Soils

The collapse of Teton dam (Wikipedia)


Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
Illustration of the suffusion process (Sibille (2016)).
AI in Geotechnical Engineering
Examples of Applications
Fabric and Effective Stress Distribution in Internally Unstable Soils

Skempton and Brogan (1994) proposed that a stress-


reduction factor, α, can be defined as the proportion
of the overburden acting on the loose fraction in the
no flow condition:

𝜎𝑓𝑖𝑛𝑒 = 𝛼 𝜎′

where σ’fine is the effective stress transferred by the


finer fraction; σ’ overburden effective stress; and α is
stress-reduction factor.

Organized by
#THEBIG5EXHIBITION
Different fabric cases (Shire et al. (2014)) www.thebig5.ae
AI in Geotechnical Engineering
Examples of Applications
Fabric and Effective Stress Distribution in Internally Unstable Soils

Discrete element modeling AI model


𝑫′𝟏𝟓
𝟏
𝒅′𝟖𝟓
𝟓
𝑭𝑪 𝟐
𝟕 𝜶
𝒆 𝟑 𝟔 Output Layer

𝑮 𝟒 Hidden Layer

Input Layer

Model features:

D’15/d’85: D’ represents the coarse fraction and


Organized by
d’ represents the finer fraction
#THEBIG5EXHIBITION
FC : Fines content (%) www.thebig5.ae
G : Gap ratio
e : Void ratio
AI in Geotechnical Engineering
Examples of Applications
Fabric and Effective Stress Distribution in Internally Unstable Soils

Organized by
Performance measures Training Validation
Coefficient of correlation, r 0.924 0.917 #THEBIG5EXHIBITION
www.thebig5.ae
Coefficient of Determination, R2 0.853 0.852
AI in Geotechnical Engineering
Examples of Applications
Fabric and Effective Stress Distribution in Internally Unstable Soils

𝑫′𝟏𝟓
𝟏
𝒅′𝟖𝟓 𝜶 = 𝟓. 𝟒𝟔 + 𝟒. 𝟒𝟖𝒕𝒂𝒏𝒉 𝑯𝟏 + 𝟑. 𝟖𝟐𝒕𝒂𝒏𝒉 𝑯𝟐
𝟓
𝑭𝑪 𝟐 𝑫′𝟏𝟓
𝟕 𝜶 𝑯𝟏 = −𝟎. 𝟎𝟖 − 𝟗. 𝟔 + 𝟎. 𝟕𝟏𝑭𝑪 − 𝟏. 𝟎𝟕𝒆 + 𝟔. 𝟏𝟔𝑮
𝒅′𝟖𝟓
𝒆 𝟑 𝟔 Output Layer
𝑫′𝟏𝟓
𝑯𝟐 = −𝟏. 𝟖𝟏 − 𝟑. 𝟔𝟐 ′ + 𝟕. 𝟕𝟑𝑭𝑪 − 𝟖. 𝟑𝟔𝒆 + 𝟐. 𝟎𝟓𝑮
Hidden Layer
𝒅𝟖𝟓
𝑮 𝟒

Input Layer

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
AI in Geotechnical Engineering
Examples of Applications
Fabric and Effective Stress Distribution in Internally Unstable Soils

Sensitivity analysis

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
AI in Geotechnical Engineering
Examples of Applications
Fabric and Effective Stress Distribution in Internally Unstable Soils
Sensitivity analysis

Organized by
The sensitivity analysis shows that the model is robust and able to reflect the role of important #THEBIG5EXHIBITION
parameters compared to the available geotechnical knowledge www.thebig5.ae
Benefits and Limitations of AI

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
Benefits and Limitations of AI
Benefits of AI

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
Benefits and Challenges of AI
Limitations and Challenges

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
Acknowledgements

Prof. Mohamed Shahin (Curtin University, Australia)

Organized by
#THEBIG5EXHIBITION
www.thebig5.ae
THANK YOU

You might also like