Modeling forced versus spontaneous capillary imbibition processes commonlyoccurring in porous sediments

The extraction of ganglia of non-wetting petroleum fluid from underground r eservoirs can occur whenever they are displaced and replaced by an invading wetting (e.g. aqueous) fluid phase. The problem is how to best specify, qu antify, observe, and beneficially control such petroleum recovery processes . Cases that are considered are those that result form the fact that there may be two separate but superimposed imbibition mechanisms involved that in herently can reinforce and supplement each other. For example, when the dis placing fluid phase more-or-less flows colinearly with the one being replac ed, this will happen whenever it is only the imposition of mechanical energ y gradients that causes it to happen. These kinds of processes can be thoug ht of as being due to a forced imbibition mechanism. At the same time, howe ver, a parallel spontaneous imbibition mechanism also can come into play wh enever the local free interfacial surface energy of the system simultaneous ly is being diminished with time. The latter happens, of course, whenever t he higher energy oil-rock interfacial tensions spontaneously become replace d by lower energy water-rock interfacial tensions. In effect, it is as thou gh the entering wetting fluids push forward in ways somewhat akin to why a ball will roll down an unobstructed hill path by itself. To obtain a clear picture of what is involved, consideration should be give n to the following three self-evident (but easily overlooked) propositions. First, it is easy to show that petroleum recovery due to the spontaneous i mbibition effects can occur even when externally imposed forcing effects ar e totally absent. Second, it turns out to be clearly indicated that even in cases when superficially the magnitudes of the spontaneous imbibition effe cts appear to be minuscule compared to those caused by prevailing superimpo sed forced imbibition mechanisms, the influence of the former effects may n ot themselves prove to be necessarily inconsequential. Third, whenever rati onally based algorithms are employed (viz. rather than simplified ones whic h are intended to mitigate and alleviate some of the computational problems ), this will be the best way to guarantee that quantitatively believable re servoir performance simulation outcomes will be achieved. This latter propo sition implicitly is affirmed by the work of Rose [Math. Geol. 22 (1990) 64 1] where reference was made to the apparent way simulation outputs for idea lized systems were different depending on either or not spontaneous imbibit ion effects were incorporated in the analyses. (C) 2001 Elsevier Science B. V. All rights reserved.