The X-1 well in a gas field in Trinidad was designed to be a high-rate gas producer from a 65° deviated well through the S1U, S1L and S2U sands at ∼9200 ft TVD. After a pre-drill sand prediction, the well was cased and perforated without sand control, but the perforations were oriented in the vertical plane (ie, topside and bottomside perfs) to limit sand production. Perforations were shot at 4 spf and 180º phasing, with ∼1,000 psi underbalance. The X-1 well produced up to ∼150 MMCFD, and was taken to depletion without any sand being produced. A production log showed that all three payzones were open. The well did not produce any water. A more thorough analysis has been made of the onset of sanding in the X-1 well, to understand the benefits of oriented perforations, and to benchmark the sand prediction against field observation. The method estimated sand strength from logs in the X-1 well. A correlation between log-predicted UCS and lab-measured thick-walled cylinder strength (TWC) in a neighboring well and adjacent sand provides the tie between log strength and actual strength measurements. The far-field stress was quantified using models and minifrac and fracpack data, and transformed to the local stresses around the perforations. The sanding potential of the well was then estimated by computing the critical bottomhole flowing pressure (CBHFP) for perforation shear-failure in the weakest zone, as a function of reservoir depletion. No sanding was predicted throughout the production history of X-1, ie, topside perfs are predicted to be stable for any drawdown and depletion to abandonment. The predictions reveal that topside perforations are superior (in terms of predicted sand stability) to phased perforated completions. The field observations of no sanding are consistent with (but do not prove) the model analysis for a 65º well with topside and bottomside perforations. However, the model suggests the perfs would have been unstable if they had been unoriented, phased perfs. The high-angle well in X-1, with oriented (topside) perfs, is suggested as a way to avoid sanding due to water influx or pressure shocks, simply because shear failure is not predicted. In the subject field, water influx appears to be a key to sand production.
In an adjacent well, X-2, also completed without sand control in the S2U sand, the well is deviated at 41º but perforations are normally phased. The horizontal perfs are predicted to fail in shear as soon as any drawdown is applied to the well. However, sanding is delayed past the prediction, and so the sanding prediction is conservative (although it may be accurate for shear failure of the perfs). The most likely explanation is that failed sand is held back by capillary cohesion due to connate water. Some other gas wells (both intermediate and deep) also demonstrate the shear-failure model can be conservative in predicting sand. In four HPHT wells, from other parts of the world, sanding is delayed substantially. In two of the wells, failed sand is held back by capillary cohesion. In the other two wells, the explanation given is arching effects acting to reduce the stresses. More work is required to understand the reasons for delayed sanding in gas wells.
The X-1 well in the subject field, Trinidad, is a gas well cased and perforated through the S1U, S1U, and S2U sands. Wells in the subject field produce mostly from S1, S2, and one other adjacent sand. The sands are weak in general: some wells in the field have sand control, while others do not. The X-1 well is oriented almost south at 169° azimuth, and deviated 65° from vertical at the S1 sands. The sands are located between 9709 and 10167 ft TVDss, measuring 11290 to 13566 ft along the wellbore. Water depth here is 230 ft. The perforations were oriented in the vertical plane (ie, topside and bottomside perfs) to limit sand production. Perforations were shot at 4 spf and 180º phasing, with ∼1,000 psi underbalance. Figure 1 shows that the X-1 well produced at a peak rate of ∼150 MMCFD, but then declined quicker than expected because the pressure compartment was smaller than expected. The well was taken to depletion without any sand being produced, as monitored by an acoustic detector (except for normal noise after shutins). A production log showed that all three payzones were open. The well did not produce any water until very late in its producing life.