Migration of Alkaline Pulses in Reservoir Sands

Abstract

Abstract Establishing the amount of alkali loss by rock reactions is critical because successful application of most alkaline flooding techniques requires that hydroxide propagate through a large portion of the reservoir. This paper presents a mathematical analysis of the chromatographic movement of alkaline pulses when they are scaled to reservoir flow rates and distances. Using only this analysis and laboratory data, we show how to estimate the distance an alkaline pulse traverses under field conditions before its concentration diminishes to ineffective levels. Laboratory core tests and X-ray analyses identify the various mineral reactions and their rates. For clayey sands a fast, reversible, sodium/hydrogen ion exchange retards alkali concentration velocities. Fine silica and quartz are suggested as important dissolving minerals, with slower-dissolving clays and clay minerals releasing soluble aluminum, which may redeposit with soluble silica as new aluminosilicate minerals. While new mineral formation influences the aqueous aluminum and silica concentrations, hydroxide consumption appears to be controlled mainly by the dissolution reaction. First-order kinetics most closely represent the dissolution behavior, lumped-parameter rate constants are reported for Huntington Beach and Wilmington sands and for a Berea sandstone. Introduction Recent studies of alkaline interactions with reservoir rock, both in steamflooding and in caustic flooding, report the importance of slow mineral dissolution. The question arises: Can alkaline pulses propagate completely through a reservoir, or do hydroxide concentrations diminish to ineffective levels shortly after injection? We address this question by outlining a mathematical analysis for the migration of a pulse of alkali, which simultaneously ion exchanges and dissolves reservoir minerals. Since the computational results are presented in a nondimensional form, a new tool is now available to estimate how long an alkaline pulse remains active for reservoir flow rates and well spacings, with data obtained only from laboratory studies. When ascertaining potential chemical loss by rock consumption, it is tempting to screen various reservoirs with either dynamic column or static batch experiments (i.e., beaker or jar tests) and to report a single number for chemical loss (e.g.. in meq/100 g of rock). Such a procedure has validity when the rock reactions go to completion during laboratory time scales. SPEJ P. 998^