Fluid Distribution Effects on P-wave Velocity of CO2/brine Saturated Rocks: a Comparison Study and Implications for CO2 Storage Monitoring Using Seismic Method

Yi Zhang,Hyuck Park,Osamu Nishizawa,Tamotsu Kiyama,Ziqiu Xue

DOI: https://doi.org/10.1016/j.egypro.2017.03.1509

2017-01-01

Abstract:The evaluation of fluid saturation using P-wave velocity is affected by the fluid distribution at mesoscale significantly. Here we present the experimental results of two tests for measuring P-wave velocity under ultrasonic frequencies in two different rocks (one sandstone and one tuff, respectively) during CO2 displacing brine. Both tests were performed under a high injection pressure of 10 MPa in order to control CO2 in a supercritical state. The two rock samples display differences in rock heterogeneity. The sandstone sample shows obvious layered structures with high and low porosity layers, whereas the tuff sample shows a near homogeneous structure. During the dynamic fluid displacements tests, we used X-ray CT technique to image rock interior frequently and then calculated average CO2 saturation near the local wave propagation path. The results show very different saturation-velocity relationships for the two tests. The sandstone sample gives a relationship that is close to Gassmann-Hill limit, whereas the tuff sample gives a relationship located near the Gassmann-Wood limit. Meanwhile, X-ray CT images show the distribution of CO2 in the sandstone sample is more heterogeneous and larger fluid patches are developed in it during CO2 drainage. In contrast, the tuff sample shows the development of very small and dispersed CO2 patches and relatively well mixing state of CO2 and brine in the rock. We then use the random patchy saturation model to interpret the results theoretically. The results indicate that the characteristic patch size is in an order of tens of millimeters in the experiment using sandstone sample, whereas it is smaller than 0.1mm in the experiment using tuff sample. Whether the scale of the strong-contrast heterogeneity created by patchy CO2 saturation is significant to wave velocity is determined by the ratio of patch size and fluid diffusion length (A/L ratio), which is a function of wave frequency and hydraulic diffusion coefficient of the rock. If the scale of the heterogeneity is larger than the fluid diffusion length at a wave frequency, the strain field created by wave propagation cannot be relaxed during a half period and the media shows the stiffening effect–a higher velocity at macroscale. These suggest that one should pay attenuation to the effect of fluid distribution and degree of phase mixing when evaluating fluid saturation from P-wave velocity using seismic methods, if the reservoir develops significant large-scale heterogeneity. Furthermore, the velocity becomes insensitive to saturation changes when CO2 saturation becomes higher than approximately 0.2, if its changes follow the Gassmann-Wood relationship (with small A/L ratio at lower frequencies). In contrast, the velocity remains changing with CO2 saturation at higher CO2 saturation if the changes follow the Gassmann-Hill relationship (with large A/L ratio at higher frequencies). The changes suggest the selection of wave frequency is critical to validly monitor the saturation changes at higher CO2 saturations.