La gammadensimetrie est une methode de mesure de la densite de differents materiaux ainsi de la porosite et de la teneur en eau d' eprouvettes. Cette etude analyse les causes d' erreur et determine les corrections puis les incertitudes de mesure de la densite, en condition saturee ou non. On montre ici que les causes d' erreur preponderantes sur la mesure de la densite sont, pour chaque hauteur de mesure, la repetabilite, le caractere aleatoire de l' emission radioactive et la precision du coefficient d' absorption. En outre, le caractere heterogene du materiau est une cause d' erreur majeure sur la valeur moyenne de la densite. On evalue ensuite les incertitudes sur diverses mesures : variation relative de densite entre deux determinations successives, porosite et taux de saturation d' une eprouvette de beton auscultee a une hauteur donnee. Pour determiner l' incertitude sur ces dernieres caracteristiques, il est necessaire de tenir compte des correlations entre les mesures successives par gammadensimetrie. Les causes d' erreur preponderantes sont alors la repetabilite et le caractere aleatoire de l' emission radioactive. (A). (Titre en anglais : Uncertainty of gammadensimetry measurements on concrete specimens). UNCERTAINTY OF GAMMADENSIMETRY MEASUREMENTS ON CONCRETE SPECIMENS. Gammadensimetry is a method for measuring the density of various materials as well as the porosity and water content of specimens. This study will analyze the causes of error and determine appropriate corrections, along with density measurement uncertainties under either saturated or unsaturated conditions. It will be shown that the predominant causes of density measurement error are, for each measurement height, repeatability, the random nature of radioactive emissions, and the accuracy of the absorption coefficient. Moreover, the heterogeneous characteristic of the material is as a major cause of error on the average density value. This article will then evaluate uncertainties on a number of distinct measurements: relative density variation between two successive determinations; porosity and saturation degree of a concrete specimen examined at a given height. In order to determine the level of uncertainty on such characteristics, it is necessary to take into account correlations between successive measurements by means of gammadensimetry. The most significant error causes thus prove to be repeatability and the random nature of radioactive emissions.
Samenvatting