Although hydrochloric acid has been used for more than 60 years to matrix acidize carbonate reservoirs, the process is still as much art as science. Historically, design engineers have used a "rule-of-thumb" based on field experience to quantify the design parameters for a candidate well. However, each well is different and this "generic" approach for treatment design can result in under- or over-treating. This paper is directed at matrix acidizing of vertical water injection wells completed in a heterogeneous carbonate reservoir (Arab-D) in Saudi Arabia with focus on field validation of a commercial matrix acidizing design simulator. The "validation" process allows the production engineer to "calibrate" the simulator for his specific field, and use it with confidence to improve treatment designs.
In this field study three matrix-acidized injection wells were used for validation of the carbonate acidizing model. The wells were completed in 150 to 200 feet openhole intervals in a reservoir composed of a variety of limestone - dolomite mixtures. Treatments were performed using hydrochloric acid along with particulate diverter stages, which was bullheaded down casing. The post-treatment model validation process consisted of simulation of the actual treatment using various values of "acidizing efficiency" to yield an acceptable match with the final skin. Subsequently, a comparison of the simulated and actual surface treating pressure, and the prepost-treatment wellhead pressure during seawater injection was made. A detailed description of the validation process and the supporting well data are presented.
A good match of predicted surface pressure and actual wellhead pressure obtained from the treating report adds credibility to the simulator. Normally, the difference in actual and simulated pressure was less than 10%. Pressure data indicate that the benzoic acid/rock salt diverter systems were inefficient. The simulator indicates diverter inefficiency causes the high permeability zones to become "thief zones" resulting in non-uniform acid distribution. In all cases the pressure dropped 25 to 50% within ten minutes following injection of acid into the formation. An internal acidizing efficiency factor was increased to allow the predicted skin of the simulator to approximate the actual skin determined from a falloff test. It was discovered that the optimum efficiency factor was a function of the amount of dolomite in the reservoir. Guidelines for the determination of the efficiency factor based upon mineralogy are presented along with an explanation for this and other observations.