Singh (16, 2001, 115-131)
drew attention to the difficulties of interpreting the results of plate jacking and plate loading tests. Although one may have reservations, the article is a useful contribution to the literature. The authors compared some of the earlier and more recent classification methods e.g. RMR, Q, and RMi (the latter developed by PalmstrΓΈm) and the degree to which they correlated with the measured results from the reviewed loading tests. A potential weakness of course is the correctness of the rock mass characterization at each test site, but collection by mostly one organization may have minimized this source of error.
Two of the older correlations between deformation modulus (which we can refer to as M) and RMR and Q, date from Bieniawski (1978) and Barton et al. (1980). These were specifically for rockmasses at the higher end of the quality scale, namely RMR)50 and Q)1. Naturally their development was limited to the database that was used at that time. The error introduced if attempts are made to use such correlations outside the intended range of the data base are clear, and hardly need to be emphasized. The authors' Table , showing the effects of varying uniaxial strength from 4 to 200 MPa was a clear example of the inadequacy of the 20-yearold Q-relation as a general formulation for all strengths of rock. This is because uniaxial strength did not appear explicitly in the 1980 formulation, as it was only designed to estimate moduli for rock masses with Q) 1. As a basis for distinct element modeling of medium strength rock masses, it has proved very successful.
The Norwegian first author is aware of the published improvements and generalizations made between the Qvalue, deformation modulus and seismic velocity. He was a guest and seminar participant at NGI in the same period as their development. Unfortunately, these widely published developments, including two ISRM congresses and an international symposium in India, were not