Spectral-Analysis-of-Surface-Waves (SASW) inversion by machine learning algorithm for shallow site investigation

Sadia Mannan Mitu (P96961)

Please use this identifier to cite or link to this item: https://ptsldigital.ukm.my/jspui/handle/123456789/457856

Full metadata record

DC Field	Value	Language
dc.contributor.advisor	Norinah Abd. Rahman, Dr.
dc.contributor.author	Sadia Mannan Mitu (P96961)
dc.date.accessioned	2023-09-12T09:14:25Z	-
dc.date.available	2023-09-12T09:14:25Z	-
dc.date.issued	2021-05-17
dc.identifier.other	ukmvital:124278
dc.identifier.uri	https://ptsldigital.ukm.my/jspui/handle/123456789/457856	-
dc.description	The Spectral Analysis of Surface Waves (SASW) method is an in-situ and nondestructive seismic method for determining the shear wave velocity profiles of geotechnical, pavement, and structural systems. Shear wave velocity is one of the important geotechnical parameters that have an empirical relationship with other soil properties such as bearing capacity, dry density, void ratio and confining pressure. One of the complex processes of the SASW data analysis is the inversion procedure. An initial soil profile needs to be assumed at the beginning of the inversion analysis which involves the calculation of the theoretical dispersion curve using forward modeling. The mechanism of inversion may not converge or take too long to converge if the initial trial is not relatively close. This is where expert judgment and input is required. Automating the inversion procedure will allow us to evaluate the soil parameters using the SASW method conveniently and rapidly. In this study, the feasibility of complete automation of the inversion process with machine learning algorithm was explored. Four types of algorithms have been evaluated: Multi-layer Perception (MLP), Random Forest (RF), Support Vector Machine (SVM) and Linear Regression (LR). For this purpose, 50 SASW field tests were conducted at various locations in Peninsular Malaysia and the dispersion curves obtained were used as input data in all the algorithms. 40 tests have been taken as training data and the rest of them as testing data. Results obtained from the four machine learning algorithms were compared with the inversion analysis process up to six meters of depth. Among all the algorithms, SVR shows the best correlation with the shear wave velocity (Vs) profile obtained from the inversion process. The range of root-mean-square (RMS) errors up to three meters depth vary from 6.79 to 29.60 for the SVR algorithm whereas it varies from 9.05 to 36.84 for the MLP algorithm, 10.13 to 29.30 for the RF algorithm and 10.66 to 31.20 for the LR algorithm. The confidence level of SVR associated with the RMS error is greater than 90% which is acceptable for foundation design purposes. The results illustrated that Support Vector Regression algorithms could potentially be used to estimate shear wave velocity of soil.,Master of Science
dc.language.iso	eng
dc.publisher	UKM, Bangi
dc.relation	Faculty of Engineering and Built Environment / Fakulti Kejuruteraan dan Alam Bina
dc.rights	UKM
dc.subject	Universiti Kebangsaan Malaysia -- Dissertations
dc.subject	Dissertations, Academic -- Malaysia
dc.subject	Shear wave velocity
dc.subject	Soil
dc.subject	Machine learning
dc.subject	Algorithm
dc.title	Spectral-Analysis-of-Surface-Waves (SASW) inversion by machine learning algorithm for shallow site investigation
dc.type	theses
dc.format.pages	243
dc.identifier.barcode	005804 (2021)(PL2)
Appears in Collections:	Faculty of Engineering and Built Environment / Fakulti Kejuruteraan dan Alam Bina

Files in This Item:

File	Description	Size	Format
ukmvital_124278+SOURCE1+SOURCE1.0.PDF Restricted Access		5.04 MB	Adobe PDF	View/Open

Show simple item record Recommend this item