Sand management/control

Simulation of Sand Production Caused by Water-Hammer Events

This paper proposes a new work flow to simulate water-hammer events, the resulting wellbore failure, and sand production in water injectors.

jpt-2018-10-189568heroto.jpg

A pressure pulse, known as a water hammer, can occur immediately after water-injection wells are shut in for emergency or operational reasons. Large pressure pulses may cause wellbore-integrity problems such as sandface failure and sand production. This paper proposes a new work flow to simulate water-hammer events, the resulting wellbore failure, and sand production in water injectors.

Introduction

For water-injection wells handling high injection rates, water-hammer signatures are observed when water injection is stopped. Designing water injectors and deciding how quickly or slowly to shut in wells requires careful attention. Prediction of sand failure caused by water-hammer events can help design shut-in protocols for water injectors. The new work flow developed in this paper integrates water-hammer simulations with sand-stability and -production predictions. The water-hammer simulation shows that rate changes during shut-in affect water-hammer amplitudes and attenuations significantly. Large pressure fluctuations, or large amplitudes in the water-hammer signature after a quick shut-in, are shown to result in significant sand failure, and a slow shut-in procedure can minimize sand production. Sizing of well-completion components and location of subsurface valves are key in the design of injection wells and can be optimized to soften the effect of water-hammer events and associated sand production.

The authors evaluate the effect of water hammer on sand failure with numerical-simulation methods. Two numerical simulators—a water-hammer model and a sand-production-prediction model—are integrated.

Results and Discussion

Performance of an injection well can be affected by many factors. Suspended solids or oil droplets in injection water can plug the matrix near the sandface and form a filter cake on the sand at the wellbore. Research has shown that the injectivity-decline process is affected mainly by filter cakes and crossflow between layers; this influence can be heightened by water-hammer pressure pulses during shut-in. The existence of water hammer, and its period, amplitude, and decay rate, varies depending on injection-well design and shut-in procedure.

Continuity and momentum-balance equations are solved for the fluid in the wellbore after shut-in of water injectors. Simulated bottomhole pressures (BHPs) containing water-hammer signatures are used as a boundary condition for the sand-production simulation. The input parameters for a vertical-well base case and a horizontal-well base case are summarized in Tables 1 and 2, respectively, of the complete paper. The wellbore trajectory of the horizontal well is assumed to have four sections.

The subsequent sand-failure responses caused by water-hammer BHP are simulated:

  • Near-wellbore frictional pressure drop (skin)
  • Location of surface-controlled subsurface safety valve (SCSSV) in the wellbore
  • Shut-in procedure

The values of parameters are then changed.
Effect of the Near-Wellbore Frictional Pressure Drop (Skin Factor). The presence of water hammer at the sandface affects sand-failure behavior significantly; the skin of the well is a crucial parameter controlling water-hammer pulses. In the water-hammer simulation, the near-wellbore frictional pressure drop is represented as well skin. Higher pressure drop represents a thicker and low-permeability filter cake. With a large resistance value, the high near-well pressure drop as well as the high skin is simulated. The skin of injectors also adversely affects the injectivity of the well.

On the basis of water-hammer simulation responses, the high skin (high near-well pressure drop, high resistance, and low injectivity index) results in a quicker attenuation and smaller initial amplitude of the water hammer. With lower pressure drops, water-hammer amplitudes and duration increase.

With regard to sand-failure behavior for the three cases, with low pressure change caused by small skin, a smooth plastic strain distribution is obtained that propagates along the minimum horizontal stress direction around the well. As the skin-induced pressure change increases, the plastic strain distribution becomes more localized, resulting in the development of shear bands. On the other hand, a steady pressure drop with no water-hammer signature allows the strain to be concentrated gradually.

In summary, a case with a significant water-hammer event resulting from small near-wellbore frictional pressure drop (small skin factor) gives rise to more sand failure than a case with no water-hammer event.

Location of SCSSV. The emergency shutdown of a water-injection well can be controlled by an SCSSV. When the SCSSV is closed, water-hammer ­pulses are transferred through the wellbore section from the SCSSV to the sandface. When the SCSSV is located at a deeper location in the wellbore, the distance the water-hammer pulse propagates becomes shorter and the wavelength of the water-hammer fluctuation becomes shorter. Because the fluid volume is smaller in the wellbore when the SCSSV is located deeper, the water-hammer signature decays more quickly in the deeper-SCSSV case.

The effect of these three different water-hammer events on sand-failure behavior is similar, but the cumulative sand-failure region depends on the combination of the magnitude of the pressure pulse and the duration of each cycle. Each potential location of the SCSSV in a well, therefore, will have to be modeled to determine the best location from a sand-control perspective.

Shut-In Procedure. The procedure of closing the valve determines how quickly or slowly the transient injection rate is changed from the initial injection rate to the complete shut-in state. This procedure of valve closure and ­associated rate changes is used for the water-hammer simulation to show the ­response of BHP. Three different shut-in procedures are tested: a quick shut-in in 5 seconds, a gradual shut-in over 50 seconds, and a stepwise shut-in in 20 seconds. In the quick shut-in, the BHP drops more than 450 psi in less than 10 seconds, and the water hammer attenuates in more than 50 seconds. In contrast, the slow shut-in process results in a gradual pressure decay without a water-hammer signature. The stepwise shut-in creates water-hammer pulses, but the maximum amplitude is approximately 200 psi. Each of these rate changes during stepwise shut-in creates water-hammer pulses. Sand-failure simulations are conducted with the in-situ stresses and initial pore pressure.

Because no water-hammer event occurs with the slow shut-in case, the development of sand failure is more localized. The results demonstrate clearly that a quick shut-in produces the largest failure area. However, the difference between slow shut-in and stepwise shut-in is negligible for this particular case. Thus, on the basis of the authors’ simulation work flow, operators are recommended to avoid a quick shut-in to prevent massive sand failure.

Permeability Reduction by Sand Failure During Water-Hammer Events. Injectivity loss caused by sand failure and sand production is a common issue related to shut-in and restart of injection wells. In unconsolidated formations, a severe injectivity loss can be triggered by several phenomena related to sand failure:

  • Aggregation of degraded sands (Fig. 1a)
  • Reinjection of accumulated fines and produced sands during water-hammer events (Fig. 1b)
  • Accumulation of produced sands in the wellbore and perforation tunnel
  • Invasion of fines (from accumulated filter cakes on the formation face) into the failed sand regions
jpt-2018-10-189568f1.jpg
Fig. 1—(a) Degraded sand aggregation in the formation. (b) Fines and sand reinjection into the formation caused by water hammer.

 

In order to take injectivity loss into account, the authors assume that, once sands fail, the permeability will be dramatically reduced. Simulations including wellbore skin factor, location of SCSSVs, and shut-in protocol have been re-­evaluated with the injectivity-loss assumption in order to analyze their effect on sand failure.

Results show that, with higher near-wellbore friction, a formation tends to be more stable. However, the difference between the failed regions with and without water hammer is significant when the injectivity loss is included. As formation permeability within the damage zone is reduced, pore-pressure distribution varies dramatically. A large pressure gradient is generated because of fines reinjection, thereby increasing sand failure. With water hammer, injectivity loss increases the extent of sand failure dramatically. This indicates that a decrease in near-wellbore permeability increases water-hammer-induced sand failure significantly.

Multiple Shut-In Events. Next, the severity of sand failure with and without water-hammer events for multiple shut-in events was compared. Results indicate that, with water hammer, the extent of the sand-failure region increases continually with the number of shut-in events. However, without water hammer, sand failure occurs in the second and third shut-in events, while, for the other events, the sand-failure rate is relatively small.

Conclusions

  • A large water-hammer pressure pulse can be generated as a result of a quick shut-in of a valve in a water injector. The pressure variation at the sandface has been simulated by water-hammer simulation and integrated with a geomechanical sand-failure model.
  • For unconsolidated sands, the sand-failure zone is much larger for cases with a water-hammer event.
  • The simulations can be used to optimize the location of SCSSVs.
  • Rapid valve closures always result in more sand failure and should be avoided.
  • A permeability reduction caused by crossflow and fines reinjection can influence near-wellbore pressure gradients dramatically and enhance the effect of water-hammer events on sand failure.
  • The sand-failure zone grows with each shut-in. This effect is magnified in the presence of water-hammer events.
  • Strain-softening behavior of the rock can dramatically increase sanding risks and severity. This is particularly true when water-hammer events are accounted for.
  • For low unconfined compressive strength, poorly consolidated sands, well shut-ins, and restarts must be controlled carefully to avoid water-hammer events.

This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 189568, “Sand Production Caused by Water-Hammer Events: Implications for Shut-In Protocols and Design of Water-Injection Wells,” by Haotian Wang, SPE, Jongsoo Hwang, SPE, and Mukul M. Sharma, SPE, The University of Texas at Austin, prepared for the 2018 SPE International Conference and Exhibition on Formation Damage Control, Lafayette, Louisiana, USA, 7–9 February. The paper has not been peer reviewed.