Abstract:
The Hidden Lake deposit at the Yellowknife Lithium Project features relatively competent hard rock, with spodumene-bearing pegmatite systems intruding felsic and quartz-rich metasediments. The greenfield site has limited geotechnical data where only basic logging parameters, such as Rock Quality Designation (RQD) and core recovery, are recorded in exploration drillholes.
SRK Consulting implemented the Core Damage Index (CDI), a machine learning-driven (ML) automated core photo classification tool, to enhance geotechnical characterization within the current drilling database. Over 9,900 dry core photographs from 278 drillholes (equivalent to 46,580 meters) were used to train a ML model for precise core row extraction and linearization, overcoming variability in photo quality and camera setup configurations. The CDI workflow ranks core fragments into four classes based on degree of brokenness. The results were composited and mapped for intensity at depth intervals, allowing for identification of discrete zones of increased fracturing and weaker breaks, particularly at pegmatite contacts, within meta-sediment country rock, and along brittle fault zones.
The logged CDI dataset provided consistent, high-resolution input for defining geotechnical domain boundaries based on observed damage trends, supporting the modeling of brittle structures, faults, and associated damage zones. The CDI workflow removed biases present in some legacy RQD datasets, unifying downhole interpretation and avoiding unfair penalization of solid rock intervals.
Although CDI is focused on brokenness and does not distinguish between mechanical and natural breaks or quantify rock strength directly, its integration with lithology and RQD data strengthens geotechnical modeling. Overall, the CDI approach advances the development of a robust brittle-structural model and improves risk management and design in the Hidden Lake deposit.
Author and Presenter:
- Aadi Sigdel | Rock Mechanics Consultant | SRK Canada
Click here to view other presentations at this event.
