The objective of this Sand Failure Test was to determine whether initial sand control is necessary on a poorly consolidated gas field, or whether it can be deferred, and if so for how long. It was feared that sand production could come from the ever-increasing stress near the well bore resulting from reservoir depletion. The economic impact of the result is significant both in terms of completion costs (implementing sand control) and the number of development wells (because of the lower productivity associated with sand control).
Prediction of the maximum possible depletion before sand production requires up-scaling laboratory cavity failure tests to field conditions. For the reservoir considered, the impact of uncertainty in the scaling laws is large. A decision based on an excessively conservative estimate could lead to unnecessary sand control.
The objective of the Sand Failure Test was to remove this uncertainty. This test can be compared to a field scale cavity failure experiment. Productivity simulations were made to calculate the number of perforations which should be shot so as to create transient sand production: only four perforations were finally shot in the well. The validity of these calculations was confirmed by the field test. Sand was produced, and was followed then by a period of sand-free production at a lower production rate.
Operational aspects such as the use of sand detection devices and a trap to collect produced sand are discussed, as are problems related to the interpretation of the test.
Bottom hole flowing pressure was monitored. Dynamic skin was calculated from pressure and flow rate. This skin was interpreted in terms of gradually-increasing perforation cavity diameter.
The fraction of the amount of produced sand brought to surface is discussed.
The increased volume of the perforations corresponds to the measured sand production.
The critical draw down at which sand production occurs was determined from overall well behaviour, because in the detailed well performance, the early sand corresponding to the weakened crushed zone surrounding the perforations and the later sand production coming from the undamaged reservoir could not be distinguished.
An intriguing feature is that the cavity growth takes place at a fairly constant fluid velocity at the cavity wall. This velocity is of the same order of magnitude as the free falling velocity of sand particles. This could be evidence for the following two-step sand production process: shear failure followed by removal of sand grains by drag forces. In a homogeneous reservoir, such a process would not, as it is commonly thought, support the continuous growth of the most productive cavity. On the contrary, smaller perforations would tend to catch up on their larger neighbours so that in fact perforation cavities would tend both to become cylindrical in shape and also to grow simultaneously. The Sand Failure Test was followed by a micro fracturation test so that both in situ minimum reservoir stress and Poisson's ratio were estimated.
The total sand-free depletion that can be confidently predicted for this reservoir after the Sand Failure Test is 150 bars higher than that obtained from a conservative interpretation of laboratory data alone. The results of a Sand Failure Test can therefore have a great impact on field development, by reducing both initial completion costs and the number of development wells.