Etude de la diagraphie neutron du granite de Beauvoir. Effet neutron des altérations et de la matrice du granite. Calibration granite. Porosité totale à l'eau et porosité neutron Analysis of the Beauvoir Granite Neutron Log. Neutron Effect of Alterations and of the Granite Matrix. Granite Calibration. Total Water Porosity and Neutron Porosity

Abstract

Cet article rend compte des travaux effectués sur la porosité du granite de Beauvoir (Sondage GPF 1 d'Echassières, Massif Central français). L'objectif de notre étude est de pouvoir obtenir des valeurs représentatives de la saturation en eau (porosité totale à l'eau n) du granite de Beauvoir à partir des mesures de porosité neutron PorositéN (diagraphie neutron BRGM) sans avoir recours aux mesures sur carottes. Notre démarche est expérimentale et nous avons tenté d'approfondir certains problèmes liés à l'utilisation de la diagraphie neutron dans une roche granitique. Deux facteurs principaux conditionnent la réponse neutron : la concentration en hydrogène de la formation (eau libre et eau de constitution de certains minéraux) et la présence d'éléments absorbeurs à forte section de capture comme le gadolinium, le cadmium, le bore, . . . et dans le cas du granite de Beauvoir, le lithium. A partir des mesures de porosité totale à l'eau n sur carottes, des essais de pertes au feu sur poudre qui nous permettent de déterminer la porosité neutron liée à l'eau de constitution PorositéN(OH-) et des analyses chimiques avec lesquelles nous évaluons la porosité neutron thermique PorositéN(ox) (Programme SNUPAR, Schlumberger) liée à la capture neutronique, nous reconstituons la porosité neutron totale PorositéNR du granite de Beauvoir. Pour 7 échantillons caractéristiques du granite de Beauvoir, nous réalisons grâce à ces résultats une nouvelle calibration du taux de comptage neutron initial corrigé du gradient thermique et de l'effet de trou. Grâce à cette opération, il est possible de déterminer, pour les échantillons traités, la porosité neutron du granite avec une calibration granite (PorositéNg) et non calcaire (PorositéNc). La connaissance de l'effet neutron de la matrice nous permet enfin d'évaluer la teneur en eau du granite (porosité totale à l'eau) et de comparer celle-ci avec la porosité mesurée sur carottes (n). Nous montrons que, pour le granite de Beauvoir, l'effet neutron de la matrice est important (en moyenne proche de 7%) et ne peut être négligé lorsque l'on mesure des porosités voisines de 0,5% sur carottes. La calibration de l'outil neutron dans le granite et non pas dans des calcaires est d'autre part capitale quant à la précision quantitative des résultats. This article describes the research done on the Beauvoir granite (Echassières GPF 1 borehole, French Massif Central range). The aim of this project was to obtain representative values of the water saturation (n total free water porosity) of the Beauvoir granite from PorosityN neutron porosity (BRGM neutron log). The exact knowledge of the porosity of a crystalline block is effectively fundamental to determine its possibilities for being used as a waste storage site. With this goal, neutron logging provides indispensable information concerning the characterization of a porous medium. Our procedure was experimental, and we tried to go more deeply into various problems linked to the use of neutron logging in a granitic rock. Two main factors governed the neutron response : (i) the hydrogen concentration of the formation (free water and combined water of various minerals) and (ii) the presence of absorber elements with a large capture cross-section such as gadolinium, cadmium, boron as well as lithium for the Beauvoir granite. After measuring the Beauvoir granite n total (free) water porosity on core samples, we evaluated the combined water content of each sample tested on the basis of fire loss tests on rock powder at 900°C. From the hydrogen atoms volumic concentration, we determined a hydrogen index that we directly converted into the PorosityN(OH-) neutron porosity, (by definition, pure water at 20°C has a hydrogen index of 1 which is equivalent to a 100% porosity). For the Beauvoir granite, the matrix combined water represents an average neutron porosity (Table 1) of about 4%. In the second phase, we used chemical analysis to evaluate the PorosityN(ox) thermal neutron porosity linked to neutron capture (Schlumberger's Nuclear Parameter Code, SNUPAR). A calibration curve (Fig. 1) between the (Sigma)mac macroscopic capture cross-section and the PorosityN neutron porosity enabled us to determine the PorosityN(ox) neutron capture porosity for all samples. The macroscopic capture cross-section of the Beauvoir granite, compared to other rocks (Table 2), is very high, about 86 cu. For the Beauvoir granite, the neutron capture porosity was estimated at about 2. 7% (Table 4). The lithium, with Li2O contents varying from 0. 3 to 1. 7%, is the one element which accounts for 85% of this effect (Table 3). Although the response of a neutron tool is not linear for low porosities (especially lower than 5%) and although in some cases the neutron effect of the matrix highly depends on the hydrogen index (close imbrication of neutron slowing and capture phenomena), we restored the PorosityNR total neutron porosity of the Beauvoir granite by stacking n, PorosityN(OH-) and PorosityN(ox) linearly. This porosity is 9% on the average. For this granite, the PorosityNma neutron matrix effect (PorosityNma = PorosityN(OH-) + PorosityN(ox)) is significant and accounts for 75% of the PorosityNR total neutron porosity corresponding to about 7%. This porosity thus cannot be neglected if the objective is to obtain representative water content values of the granite from neutron porosity log. This is why the second part of our project took up the problem of calibrating neutron tool for analyzing a granitic formation. For the Beauvoir granite, the neutron porosity data were obtained from standard calibration in limestone blocks. As the neutron effect of the granite matrix was not negligible, we performed our own calibration using seven granite samples with a perfectly well-known total neutron porosity (free water content and neutron matrix effect). We determined a PorosityNg granitecalibration neutron porosity. For this, the relationship between the N neutron counting rate and the PorosityNc standard neutron porosity (limestone calibration) was analysed (Fig. 2 and Fig. 3). For neutron porosities ranging from 0. 5 to 100%, we also examined the behaviour of N as a function of the borehole diameter (Fig. 4). To avoid the problem of correcting the hole effect and to be able to compare the porosities determined by neutron logging and measurements performed on core samples no matter what the depth was, the neutron counting rate was recalculated (N63) for a constant diameter of 63 mm (chosen arbitrarily), (Fig. 5). Then, a calibration curve between the N63 counting rate and the PorosityNR total reconstituted neutron porosity was determined for all the granite samples. From this relationship, we calculated a PorosityNg neutron porosity, called granitecalibration neutron porosity. Then PorosityNR, PorosityNg and PorosityNc were compared (Fig. 7, Fig. 9 and Fig. 10). By substracting the PorosityNma neutron matrix effect from PorosityNg and PorosityNc, we obtained the ng and nc water content values that we compared to the n water content measured on core samples (Fig. 8 and Fig. 11). PorosityNR, PorosityNg and neutron porosities and n, ng and nc water contents (free water saturation), were compared for the different alteration facies of the Beauvoir granite (Fig. 12). On the whole, there was a better (qualitative and quantitative) relationship between the PorosityNR total neutron porosity and the PorosityNg granitecalibration neutron porosity (which are about 9%), than with PorosityNc. The PorosityNc neutron porosity issuing from the limestone calibration was underestimated because the rock neutron matrix effect was not taken into account (around 7% for the Beauvoir granite). Concerning water contents, limestone calibration led to nc water content values that, in 60% of the cases, were less than 0% and hence unusable. On the other hand, the ng water content obtained from the granitecalibration neutron porosity was more satisfactory since it gave an average water content value in agreement with the total water porosity measured on core samples (2%). However, it can be noted that the average variations between PorosityNg and PorosityNc as well as between ng and nc for the different Beauvoir granite facies are not always coherent (Table 5). This phenomenon can be partly explained by the averaging effect of the values as well as the singular behaviour of some samples. Another reason for this is, above all, the lack of points in the low and high porosity ranges for plotting the granitecalibration curve. Considering the numerous error factors encountered during this type of experimental research, (eccentricity of the tool, non-existant granite calibration, difficult assessment of the total water porosity on core samples and of1he neutron matrix effect, . . . ) these results may be considered satisfactory. To conclude, we can say that it is possible to evaluate the total free water porosity of an unaltered or hardly altered granite not containing any chemical elements having a large neutron capture cross-section, by using neutron porosity. Two conditions must imperatively be respected :a) The neutron tool must be calibrated in granite blocks for which we know the total neutron effect, the water content (total free water porosity) and the neutron effect of the matrix, so as to obtain a more accurate evaluation of the total neutron porosity of the rock. b) A fairly clear idea must be obtained of the average combined water content of the matrix (e. g. from cuttings or another log) in order to evaluate the neutron matrix effect so as to be subsequently able to correct the porosity issuing from the neutron counting.