Manager, Engineering Materials Office, Ministry of Transportation, Downsview, Ont.
(Received 2 October 1998; accepted 14 July 1999)
The accelerated mortar bar test, developed by the National Building Research Institute (NBRI) in South Africa, has been adopted in North America for the rapid identification of potentially alkali-silica reactive aggregates and may also be used for assessing the effectiveness of supplementary cementing materials. The purpose of the multi-laboratory studies, involving 46 and 32 laboratories, was to obtain data to develop a multi-laboratory precision statement for the test. The aggregates used in this study were Spratt siliceous limestone from a quarry in Ottawa, and a natural glacio-fluvial sand from north of Toronto in Ontario.
The natural sand gave an average expansion of 0.164% at 14 days and the quarried stone gave an average expansion of 0.417%. The multi-laboratory standard deviation after removal of outliers was found to be proportional to the amount of expansion. The average coefficient of variation at 14 days was found to be about 15% in the two studies.
The use of a single cement by all laboratories did not give a major improvement in precision. Autoclave expansion, and the alkali, sulfate and magnesium oxide (MgO) contents of cement were not related to mortar bar expansion at 80°C. A relationship was found between MgO content and autoclave expansion.
It is possible to prepare a precision statement as follows: For mortars giving average expansions after 14 days in solution of greater than 0.1% the multi-laboratory coefficient of variation (1s% of ASTM C 670) has been found to be 15.2%. Therefore, the results of two properly conducted tests in different laboratories on specimens of a sample of aggregate should not differ by more than 43% (d2s% of ASTM C 670) of the mean expansion.
Paper ID: CCA10432J