New Approaches for Rock Strength Estimation from Geophysical Logs Binzhong Zhou 1 , Stephen Fraser 1 , Mihai Borsaru 1 , Takao Aizawa 2 , Renate Sliwa 1 , and Tsutomu Hashimoto 2 1 CSIRO Exploration and Mining PO Box 883, Kenmore, QLD 4069 2 Suncoh Consultants Co. Limited 1-8-9, Kameido, Koto-Ku Tokyo, 136-8522, Japan ABSTRACT There are various ways to estimate rock strength from geophysical borehole data. Most previous approaches require an understanding of the properties of the intact rock and of the defects within it. Geophysical logs can be used to estimate both of these properties. In this paper, we use two approaches based on the Radial Basis Function (RBF) and Self Organizing Maps (SOM) methods to estimate rock strength from the specialist nuclear SIROLOG (spectrometric natural gamma, Prompt Gamma Neutron Activation) and conventional geophysical logs in anticipation that better performance can be achieved. The RBF and SOM approaches do not depend on any pre-existing assumptions or models, but estimate the rock strength based on parameters and relationships derived from the internal structure and relationships within the geophysical logging data set. Our RBF and SOM methods can readily accommodate variations in rock characteristics; but their performance largely depends on the completeness of the range of lithologic variation in an available control data base. Both the specialist SIROLOG and conventional geophysical logging data from the Newlands Mine (Collinsville) were used to demonstrate the effectiveness of the SOM and RBF algorithms to estimate the measured sonic log, and the UCS. Good results were achieved from both the RBF and SOM algorithms, which indicates the viability of these new methods for estimating rock strength from geophysical logs. Introduction Geophysical borehole logging is routinely conducted at Australian coal mines for various applications such as strata correlation from borehole to borehole. One of the most attractive applications would be to estimate the strength of the rocks as it is critically important to have a proper understanding and accurate estimation of the