Comparing Porosity and Permeability Measurements With Advanced Apatial Image Analysis Data
Al-Mudhafar, A., Ghulami, M., and Viana Lugo, A.
Engineering Honours Degree, 2010
University of Adelaide
Porosity and permeability data are essential to any reservoir characterisation. These data are at the moment produced mainly by conventional analytical techniques such as Helium-porosity and Nitrogen-permeability measurements. This study evaluates the potential of advanced image data analysis of synthetic rock and/or standard core plugs to model key petro-physical parameters such as porosity and permeability. If advanced image analysis solutions prove successful to generate reliable information on the texture and petro-physical properties of rocks, it would prove a cost-effective, non-destructive tool for reservoir characterisation.
In the present study, 3 different core plugs consisting of 2 synthetic samples (glass beads and beach samd) and a natural core (Berea sandstone) were initially tested for their porosity and permeability using conventional methods like He-porosimeter and N-permeameter. The sample selection was based primarily on the cores’ simplicity. The same samples were later examined under an x-ray micro-tomographer with capable resolutions of approximately 9 microns to obtain 2D cross-sectional images of the sample at different depths. Image analysis was then performed on these 2D images to obtain porosity and grain size distribution of the sample. Although porosity was obtained successfully, a model of the grain size distribution was not. The reason was attributed to a lack of resolution but most importantly an understanding that 2D images do not give a clear view of the actual structure of the grains. It became to realisation that an equipment capable of 3D characterization of the grains was required and so a particle counter was used to do the job. Using an average of the 2D porosity values obtained from image analysis and grain size distribution data from the particle counter, permeability was calculated based on models primarily derived from that of Kozeny-Carman equation.
Porosity values obtained from the He-porosimeter were 19.9, 25.2 and 25.9% for the Berea sandstone, beach sand and glass beads respectively. It was believed that these values were underestimating the samples true property and so Archimedes’ method was used to confirm this calculation. Archimedes’ method gave porosity values of 20.4, 36.1 and 35.6% for the same samples respectively. For the He-porosimeter it was concluded that the low porosity values obtained for unconsolidated samples was most likely due to the irregular shape of the encapsulating container. The porosity values derived from image analysis were 21.9, 35.9 and 37.2% for the same samples and these were in good agreement with those obtained from Archimedes’ method.
Permeability values obtained from the N-permeameter were 133, 801 and 815 mD for the Berea sandstone, beach sand and glass beads respectively. The results for the Berea sample when compared to historic data were in fair harmony. However, with the unconsolidated samples the expected permeability was in the order of a couple of Darcies. A liquid permeameter was used for confirmation of these results and the outcome was a more reasonable set of permeability values. Permeabilities this time were 78, 4650 and 4789 mD for the same samples respectively. The anomaly in the N-permeameter readings was attributed to the blockage of one of the flow streams or to a faulty pressure transducer.
Permeability values calculated from the algorithms were all in general overestimating when compared to those derived from the liquid permeameter. However, Krumbein & Monk’s and Berg’s model were in fairly good agreement. According to the results achieved in this work, advanced image analysis of synthetic rock and/or standard core plugs to model key petro-physical parameters such as porosity and permeability seems like a reasonable tool depending on the chosen algorithm.