Using Geochemistry to Address H2S Production Risk due to Steam Injection in Oil Sands

Abstract

Abstract Steam injection for enhanced oil recovery induces chemical reactions within the reservoir, called aquathermolysis, which can lead to in-situ H2S generation and to H2S production at the wellhead. H2S production risk is particularly acute in oil sand reservoirs because they contain sulfur-rich bitumens. To forecast H2S production risk in these conditions, a workflow based on geochemical investigation and reservoir simulation has been developed. It relies on (1) a quick estimate, using a dedicated technique, of sulfur content and thermal reactivity of a large number of reservoir samples to map the H2S production risk over the field; (2) carrying out more time-consuming aquathermolysis experiments on oil sand samples selected from step 1, to define a kinetic model for H2S generation based on atomic sulfur thermal reactivity; (3) transforming this sulfur-based kinetic model into a molecular SARA components-based kinetic model, usable in a compositional and thermal reservoir simulator; (4) simulating the EOR process with the reservoir simulator to calculate H2S/oil ratio at the wellhead. The geochemical methodology has been applied to four oil sand samples from Athabasca. The results have underlined that sulfur content and sulfur thermal reactivity of oil sands measured with the quick estimation technique are well correlated with the amount of H2S produced from the more lengthy aquathermolysis experiments. Moreover, it was shown that the sulfur in the oil sand, when distributed among Saturates, Aromatics, Resins, Asphaltenes, Solid matrix and H2S, as a function of time and temperature of aquathermolysis, can be interpreted in terms of sulfur-based kinetic model. This model can be used for a calculation of H2S generation upon aquathermolysis at field production temperature and time scale, thus for estimating the H2S production potential. As detailed in another SPE paper (Barroux et al., 2013), reservoir simulation has been used to simulate a SAGD process in a generic 2D model of an Athabasca oil sand. One main finding of this study has been that the H2S/oil ratio at the wellhead appears to depend mainly on the stoichiometry of the kinetic model.