A New Comprehensive Approach for Predicting Injectivity Decline during Waterflooding

Abstract

Injectivity decline during sea waterflooding or produced water re-injection is widely observed in North Sea, Gulf of Mexico and Campos Basin fields. The formation damage occurs mainly due to the deposition of suspended solids around injectors and the build-up the external filter cakes in the well bores. The ability to predict injectivity decline accurately is of great importance for project designs and water management. A comprehensive model that incorporates a variety of factors influencing the process is desirable for the prediction. In this paper, a new comprehensive approach for predicting injectivity decline during water flooding is proposed. The deep bed filtration is described by novel stochastic random walk equations. The injectivity decline model takes into account the reservoir heterogeneity and the distribution of solid particles by sizes. It also accounts for the later formation of the external filter cake and its erosion. A piece of software SNY is developed with the proposed model. The model is able to capture the behaviors of the injectors in the field: the initial slow injectivity decline due to the deep bed filtration of suspended particles, the later faster decline due to the build-up of the external cake, and the temporary steady state due to the cake erosion. Stronger normal dispersion or median heterogeneity close to the injector leads to farther penetration of the particles and slower impedance increase. Neglecting the particle population heterogeneity may lead to the underestimation of formation damage and predicts late transition to external cake formation. The impedance at the steady state and the starting time are highly influenced by the cake properties. The impedance and the external cake thickness at the steady state are likely to be higher in horizontal wells than those in vertical wells.

Stabilization of External Filter Cake by Colloidal Forces in a “Well–Reservoir” System

“…Several other works presented the lever ratio of the order of magnitude that significantly deviates from the above-mentioned intervals: Paiva et al calculated l = 0.05 by comparison with experimental data; ,, the parameters used by Al-Abduwani et al, Zinati et al., and Yuan et al correspond to l = 5.0, l = 5.0, and 2.0, respectively. The reason for this deviation is ignoring the electrostatic force and assuming the unit value for the permeate force correction factor Φ H .…”

Type Curves for Injectivity Decline

“…The equilibrium condition contains an empirical constant -lever arm ratio l. Treatment of nine laboratory tests on cross flow filtration provides the lever ratios varying from 162 to 693. Several other works present the lever ratio of the order of magnitude that significantly deviate from the above mentioned interval (see, Al-Abduwani et al 2005, Paiva et al 2006, Zinati et al 2009, Yuan et al 2012. The reason for this deviation is ignoring the electrostatic force and assuming the unit value for the permeate force correction factor Φ H .…”