Interpretation of Induction Time and Nonstandard Spontaneous Imbibition Trends Utilizing In-situ Measurements – Identification of No-Flow Regions and Wettability Alteration

Summary This work aims to analyse and explain non-standard imbibition, observed here and frequently in the literature. Previously conducted spontaneous imbibition tests in fully oil-saturated and strongly water-wet Bentheimersandstone core plugs, using OEO (One End Open) and TEOFSI (Two Ends Open Free Spontaneous Imbibition) revealed a significant delay at start of imbibition (induction time) before standard theoretical recovery vs time behaviour was established. The radial corefaces had been sealed with epoxy glue to define no-flow boundaries and yield imbibition corresponding to one dimensional (1D) solutions. However; in-situ imaging revealed that flow occurred in a two-dimensional (2D) manner. Particularly, in-situ imaging showedthat the water saturation at the end of imbibition was much higher in the core center than close to the no-flow boundaries. The tests were simulated numerically to interpret possible causes for the non-standard behaviour. First, the core scale model was parameterizedby matching AFO (All Faces Open) experiments (same experimental conditions, but not applying epoxy) and some of the TEOFSI tests that seemed able to be corrected for induction time. The predicted behaviour of the remaining tests was in agreement in terms ofimbibition rate if an induction time correction was made, however much lower recovery was observed than predicted. Introducing no-flow regions in the model near the epoxy layers and an initially weakly oil-wet state centrally in the core were both necessary mechanisms to fully interpret the tests.The no-flow regions explained the difference in end recovery, but also impacted the imbibition rate (it was reduced). The initial weakly oil-wet state explained the low, but not zero imbibition rate in the induction period. A wettability alteration towardsstrongly water-wet then explained the resulting behaviour. It was found that this event was more likely triggered than gradual. It was however challenging to determine the triggering event. This work demonstrates that spontaneous imbibition tests are very sensitive to the flow properties near the no-flow boundaries and can potentially affect the interpretation of end pointsaturations and flow functions. In-situ imaging by PET-CT improved the interpretation of the results by direct implementation of no-flow regions in the model. Accurate spontaneous imbibition behaviour must be achieved in the laboratory before upscaling tothe field.