Data overload: Does more data really resolve uncertainty?

Authors: Duran, A Download Paper Open access courtesy of:

DOI https://doi.org/10.36487/ACG_repo/2335_34

Cite As:
Duran, A 2023, 'Data overload: Does more data really resolve uncertainty?', in PM Dight (ed.), SSIM 2023: Third International Slope Stability in Mining Conference, Australian Centre for Geomechanics, Perth, pp. 525-534, https://doi.org/10.36487/ACG_repo/2335_34



Abstract:
The author presents a case study where ongoing drilling campaigns coupled with laboratory testing were utilised to assess the geotechnical characteristics along strike in an area of complex geology. While ‘more’ geotechnical logging is paramount in providing a spatial understanding and resolving uncertainties in geology and structure, the author’s experience with the laboratory testing is somewhat contrary. For the case study, laboratory testing had been carried out on a campaign basis over an 11-year period, resulting in over 200 direct shear (DS) tests, 1,200 unconfined compressive strength (UCS) tests and 250 triaxial (TXL) tests. This paper highlights issues that are encountered with such extensive laboratory data. UCS testing, using point values, has well-defined methodologies in statistical treatment and assessing uncertainty and reliability. However, tractable methods to deal with DS and TXL are not as well-defined. The paper highlights the issues and presents what appears to be increasing uncertainty with more data, contrary to the commonly held view that more testing is better.

Keywords: uncertainty, laboratory testing, UCS testing, direct shear testing

References: