Underbalanced drilling (UBD) has the potential to add value by maximizing productivity and ultimate recovery by reducing formation damage. The benefits of UBD are of course dependent on the ability to maintain underbalanced conditions throughout the entire life of a well, especially during the drilling phase. The ability to maintain underbalanced conditions is complicated because real UBD wells are rarely, if ever, in a "steady-state" condition and are subject to constantly changing or transient flow behavior. Improved understanding of the transient flow behavior will increase the ability of the rigsite engineer to maintain the desired bottomhole pressure and thus minimize potential formation damaging, overbalanced periods. This paper illustrates how detailed transient analysis provides a rigorous engineering basis for selection of the appropriate methods, to maintain optimum downhole conditions by minimizing bottomhole pressure instability.
Methods to mitigate destabilizing transient effects with drill pipe injection are relatively well understood, due to extensive case histories, but less experience is available with concentric casing (CC) injection and thus less is known on options to control pressure instability.
The UbitTS transient flow simulator is used to optimize design and operational parameters to minimize well slugging tendency and pressure instability when the concentric casing injection technique is employed. Reference is made to a generic test well, but the methodology has application to all UBD operations considering CC injection.
The primary purpose of this paper is to provide insight for achieving stable underbalanced conditions during underbalanced drilling operations utilizing a concentric casing injection method. Since pressure instability is inherently a time dependant, or transient effect, steady state hydraulic flow models provide little assistance in mitigating against this behavior. Therefore, the use of transient flow simulations during UBD well design and implementation is discussed, to enable wellsite personnel to make real time decisions in order to achieve desired downhole conditions.
At the time this paper was prepared three projects utilizing the CC injection method were being performed by Shell in the Middle East. A recurrent problem in the initial wells in these projects was transient slugging behavior and resultant well instability associated with CC injection of gas. The ability to properly analyze this problem, and develop practicable methods to obtain pressure stability was limited by the use of steady state flow simulations. Steady state modeling provides a level of understanding whether a project is technically viable, and gives clues as to the transient controllability of the well, but does not model the true well behavior after flow disturbances, such as choke manipulation, are introduced. This deficiency is normally addressed in the design process thru the use of safety factors incorporated into the steady state design of maximum pressures and flow rates.
This paper details design considerations, such as gas to liquid ratio, critical gas injection rate, concentric casing volume, injection port restriction, and fluid density, to minimize pressure instability. Also discussed are operational techniques that in combination with "real time" transient modeling will assist the well site engineer make more informed decisions to efficiently stabilize pressures. It will also expose methods that actually amplify the problem. Realtime transient modeling may prove to be a more cost-effective approach to address pressure stability problems, when compared to the commonly adopted time-consuming and expensive "trial and error" approach.
Since detailed transient modeling is required for both the design and operational methods to minimize pressure instability, it is first required to validate the transient flow model. The model was validated against real well data collected from Shell's recent wells during periods of well slugging using the concentric casing injection method. An acceptable level of confidence was established with the model results, after actual data trending was properly matched. Thereafter, detailed sensitivity analyses on various control parameters were conducted to determine the most effective approach to reduce pressure instability. The validation exercise results are outside the scope of this paper but will be detailed in a subsequent paper.