An automatic interpretation tool, LogTrans, for geophysical borehole logs was developed during a recent CMTE/AMIRA Project, The Application of Geophysics to Mine Planning and Operations. LogTrans performs rapid analysis of multi-parameter logs and expedites presentation of interpreted results in a form meaningful to mining engineers and geologists.The algorithm exploits the contrasts in petrophysical signatures between different "classes" of rock. The classes may be distinguished by lithology, grade, mechanical properties, or any combination of characteristics. The LogTrans procedure entails two stages:• statistical characterisation, involving determination of the centroids (means or medians) and ranges (standard deviations or spreads) of the distributions of each petrophysical parameter for each class, based on a representative control data set; • discrimination, in which data points are assigned to the "nearest" control class in multi-parameter space.The LogTrans discrimination algorithm is simple, intuitive, fast and flexible. However, it is not a Black Box. Its performance relies fundamentally on the integrity of the control information: the geological, geochemical, and/or geomechanical data from the control holes, as well as the geophysical logs. In particular, drilling depths must be carefully reconciled with borehole logging depths. Interpretational performance is enhanced when parameter distributions are uni-modal and compact, and when high petrophysical contrasts exist between classes. Critical assessment of the control information can prompt introduction of new classes (by sub-division of existing classes or otherwise) and derivation of new parameters (by mathematical combination of original parameters) to better achieve the interpretational objective.LogTrans has been applied successfully to logging data sets collected from a number of Australian and overseas mines. Prediction accuracy in excess of 90% has been achieved in several cases. Its application for orebody delineation at Century and for grade range interpretation at Rosebery is illustrated in this paper. 269 Min. Res. Eng. 1999.08:269-284. Downloaded from www.worldscientific.com by UNIVERSITY OF QUEENSLAND on 08/16/15. For personal use only.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.