Professional Documents
Culture Documents
Technical Note
A Suggested Improvement to the Schmidt Rebound
Hardness ISRM Suggested Method with Particular Reference
to Rock Machineability
R. M. G O K T A N t
C. AYDAYI"
Table 1. Sample size vs p-values [4] Since the observation 37 is to the left of the mean, this
n p n p n p n p ratio correctly has a minus sign in front of it, although
2 1.15 7 1.80 15 2.13 50 2.58 its absolute value is of interest for the analysis. For 20
3 1.38 8 1.86 20 2.24 100 2.81 observations the tabulated value o f p in Table 1 is 2.24.
4 1.54 9 1.91 25 2.33 250 3.09
5 1.65 10 1.96 30 2.40 500 3.29 Since this is smaller than the ratio of the deviation to
6 1.73 12 2.04 35 2.45 1000 3.48 the standard deviation, the observation 37 should be
rejected. Therefore, the best value for the test would then
be the mean of the remaining 19 observations.
Chauvenet's criterion, a special case of the t-distribution,
is used for choosing the observations which belong to a
population, and rejecting the outlier observations. The CONCLUDING REMARKS
criterion states, "An observation in a sample of size N In the general field of rock excavation there is a need
is rejected if it has a deviation from the mean greater for establishing more exact relations between mechanical
than that corresponding to a 1/2 N probability. The properties of the rock and performance of cutting
criterion is discussed in some detail elsewhere [4], and machines. This goal can be best achieved if adequate
rules for the use of the criterion are given below: rock testing methods and data recording techniques are
1. Compute the mean and the standard deviation of used.
all observations. The Schmidt hammer recording technique as de-
2. Determine the ratio of the "suspiciously" large scribed by ISRM has gained widespread acceptance as
deviation divided by the standard deviation. Deter- a strength index for rock materials and as a means of
mine the limiting value p of this ratio from Table estimating the uniaxial compressive strength. However,
1, for the corresponding number of determinations since only the upper rebound values are considered in
N. this technique, it is felt that such a recording technique
3. If the observed ratio is greater than the value found might not accurately reflect the strength variations on
in the table, the observation may be rejected. rock faces.
In this Technical Note, a statistical method which
Example. Let us consider a set of actual Schmidt allows the integration of scattered data obtained during
hammer test results obtained by a calibrated L-hammer. Schmidt hammer testing is brought to the attention of
The test surface was inspected for microscopic defects
those researchers and engineers who work in the area of
and was made sufficiently smooth to take reliable read- rock excavation. It is thought that such an integration
ings with the hammer. Twenty single impacts separated method will have the advantage of objectively recording
by at least 30 mm approximately were taken. Arranging the characteristic variations in rock texture and micro-
the recorded data in rank order, we have: structure which predominantly affect the performance of
37,40,40,41,41,42,42,43,44,44,44,45,45,45,46,46,47. excavating machinery. However, it is emphasized that
there is a need for further research work involving case
It can be noticed that the distribution of these rebound studies in order to check its practical validity.
values is normal about a mean value of 43.15. In order
to arrive to an average value reflecting the hardness Accepted for publication 21 December 1992.
characteristics of this population, we have to discard
those observations which do not belong to the popu-
lation. Due to the aforementioned reasons, we are REFERENCES
particularly interested with the rejection of lower re- 1. Poole R. W. and Farmer I. W. Consistency and repeatability of
bound values. Therefore, in this case we want to know Schmidt hammer rebound data during field testing. Int. J. Rock
Mech. Min. Sci. & Geomech. Abstr. 17, 167-171 (1980).
if the rebound value 37 is representative of this popu- 2. ISRM. Suggested Methods for determining hardness and abrasive-
lation or not. ness of rocks. Int. J. Rock Mech. Min. Sci. & Geomech. Abstr. 15,
The mean of 20 rebound values is 43.15 and the 89-98 (1978).
3. Poole R. W. and Farmer I. W. Oeotechnicat factors affecting
standard deviation of all values is 2.41. The ratio of the tunnelling machine performance in Coal Measures rocks. Tunnels
suspected deviation to the standard deviation is: Tunnelling 10, 27-30 (1978).
4. Mann L. Applied Engineering Statistics for Practicing Engineers,
(37 - 43.15)/2.41 = -2.55. p. 175. Barnes & Noble, Inc., New York (1970).