CPTu-SPT correlation analyses based on pairwise data in Southwestern Taiwan. Issue 4 (2nd October 2022)

• Record Type: Journal Article
• Authors: Han, Xiao; Gong, Wenping; Juang, C. Hsein; Bowa, Victor Mwango; Khoshnevisan, Sara
• Is Part Of: Georisk. Volume 16:Issue 4(2022)
• Page Start: 622
• Page End: 639
• Publication Date: 2022-10-02
• Subjects: Piezocone penetration test (CPTu) -- standard penetration test (SPT) -- correlation analyses -- multivariate linear regression (MLR) -- model uncertainty; Environmental risk assessment -- Periodicals; Hazardous geographic environments -- Risk assessment -- Periodicals; Reliability (Engineering) -- Periodicals; Risk management -- Periodicals; 624.151
• DOI: 10.1080/17499518.2022.2028848
• Languages: English
• ISSNs: 1749-9518

Abstract

ABSTRACT: Although the piezocone penetration test (CPTu) is becoming increasingly popular in site investigation, the standard penetration test (SPT) is still widely adopted in engineering practices. Additionally, a lot of empirical formulas in geotechnical engineering are based on the SPT data, rather than CPTu data. Hence, analysis of CPTu-SPT correlations is a great issue in geotechnical engineering. In this paper, CPTu-SPT correlations are analyzed based on a high-quality side-by-side CPTu-SPT database established in Southwestern Taiwan. The conventional CPT-SPT correlations are first validated with the database collected, through which the importance of formulating a new transformation model is highlighted. Then, CPTu-SPT correlations are studied using multivariate linear regression (MLR) analysis based on this database, and the best transformation model (for predicting SPT N 60 -value from CPTu data) is identified by the best subset regression method. In comparison to traditional transformation models, this new model considers the pore water pressure, soil behaviour type index, and effective overburden stress explicitly; as an outcome, this new model could yield higher accuracy in mapping CPTu-SPT correlations. Finally, this new model is validated by the other set of pairwise data collected in Turkey, and the advantages of this new model over artificial neural network (ANN) models are discussed.