ISSN: 0149-6123
CODEN: CCAGPD
Page Count: 10

A New Statistical Method for Prediction of Concrete Strength from In-Place Tests
Reeve, CP
Research structural engineer and mathematical statistician, National Bureau of Standards, MD

Stone, WC
Research structural engineer and mathematical statistician, National Bureau of Standards, MD

Abstract

A method is presented for determining the functional relationship (correlation curve) between an in-place parameter and the compressive strength of concrete, based on laboratory-conducted tests. The method accounts for the error in both X (in-place) and Y (compressive strength). A statistical procedure is developed for predicting the characteristic strength of concrete in the field, which accounts for the variability of the field-conducted in-place tests, the difference between the coefficients of variation associated with the in-place parameter and the compressive strength, and the error in the laboratory-derived correlation curve. It is concluded that a normal or lognormal distribution can be assumed for the compressive and in-place strength of concrete when fitting the correlation curve, but for prediction of the characteristic strength in the field the lognormal assumption is necessary unless the associated coefficient of variation is less than 15%.

Keywords:
compressive strength, construction, statistical analysis, data processing, concretes, characteristic strength, correlation curve, in-place strength, strength prediction

Paper ID: CCA10048J
DOI: 10.1520/CCA10048J
ASTM International is a member of CrossRef.