There is nothing new about the risks of managing sand production in the oil industry. Nevertheless, new insights are often gained when new wells are drilled. A case in point is Shell’s experience developing the K gas field offshore The Netherlands with two sub-horizontal underbalanced development wells drilled from a monotower satellite. The first well was completed with sand control, while downhole sand control was installed underbalanced in the second well. Both acoustic and intrusive sand detection were installed to monitor and control sand production. Shell managed its sand production risks by (1) comparing predicted sand volume and sand erosion against limits for surface erosion and sand handling; (2) deciding on downhole sand control via a sand test; (3) installing downhole sand control underbalanced; (4) calibrating surface sand monitoring; and (5) implementing startup procedures.

Shell’s K well design features 5,578 to 6,562 ft (1,700 to 2,000 m) sub-horizontal drainholes

Figure 1. Cumulative sand volume prediction for K development wells as function of size effect (SE). (Graphics courtesy of Shell E&P Europe)
that target a 98- to 131-ft (30- to 40-m) reservoir interval. The main concern with subsurface sand is the formation of bridges across the drainhole or tubing that would block or choke production. The K field facilities consist of an unmanned, marine-access satellite platform, or monotower, that produces unprocessed wet gas via a 16-in. pipeline to the processing platform. The main sand concerns at surface are erosion of the K topsides and sand fill of pipeline and remote processing facilities.

The K sandstone gas reservoir is located at a true vertical depth of about 9,187 ft (2,800 m) and is characterized by low permeability (0.1 to 1.0 mD) but relatively high porosity (12 to 22%). In view of the low permeability, long sub-horizontal underbalanced drainholes were planned to maximize and ensure inflow performance and ultimate recovery. However, the high porosity translates into low rock strength and a real risk of significant sand production. Shell had to develop a robust technique to install sand control underbalanced, since none was available at the time.

Sand volume and flowline erosion
A prediction of the expected sand production volumes was carried out using the methodology outlined in Figure 1, which shows the cumulative sand volume as a function of drawdown pressure plus depletion, assuming a porosity distribution for the two development wells based on offset appraisal wells and a so-called size effect (SE) varying between 1.0 and 2.0. The predicted ultimate sand volume roughly varies between 71 and 7,063 cf (2 and 200 cu m). In view of the large remaining uncertainty, a sand test was planned at the end of the underbalanced drilling phase to achieve representative values of drawdown pressure plus depletion in the actual well. As illustrated in Figure 1, the sand volume produced during the sand test could then be used to decide which sand volume curve is representative of the field situation.

The maximum acceptable sand volume produced to surface was based on:
• the maximum sand volume that operations are capable of processing and
• the maximum sand rate and volume that will not result in excessive flowline erosion.
The worst case flow line erosion was predicted using an erosion spreadsheet for various gas and sand production scenarios. Based on a maximum flow line erosion of 5 mm, and taking processing constraints into account, the maximum acceptable sand volume had been set at 20 MT per well.

Deciding on sand control via test
After reaching total depth in the first well, a sand test was carried out with the drill string still in hole. No sand was produced to surface during the test and the decision was made to complete the hole without sand control.

In hindsight, the sand test must be considered ambiguous at best for two reasons. Firstly,
Figure 2. Cumulative erosion senario for K development wells for constant sand rate of 12 kg per 10E6 cu m.
significant washouts were observed while drilling the reservoir section. The solids volume recovered at surface while drilling well K-101 translates into an average hole diameter of about 9-in. as compared to the 6-in. bit size (Figure 3). Possibly the solids produced while drilling underbalanced would have been produced during the production phase in case the well had been drilled overbalanced. Secondly, the prolific nature of the well resulted in a minimum flowing bottom hole pressure (FBHP) of 215 bar. This is equivalent to a drawdown plus depletion of 90 bar, which is modest compared to 250 bar at abandonment conditions (Figure 1). More important, however, this high FHBP resulted in a downhole velocity across the 9-in. washed out borehole that was probably insufficient to transport sand effectively across most of the drainhole.

Sand control underbalanced
Given the expectation of more porous (weaker) sands in the second well and the inconclusive results of the sand test in the first well, the decision was made to complete the new well with a wire-wrapped screen (WWS). An aggressive-as-practical slot size of 300 micron was selected based on sand grain-size analyses on core material. The underbalanced installation of the WWS was made possible by combining field-proven technology in a novel manner by making use of both a drillable bridge plug and a drill bit plus mud motor ahead of the premium WWS. Well testing before and after running the sand screens showed identical inflow performance.

Acoustic/Intrusive sand detection
Since production startup, the sand production from both wells has been monitored using
Figure 3. Solids volume produced to surface while drilling K-101 drainholes (well features 2 drainholes, both contributing to flow).
both acoustic sand detection and intrusive erosion sand detection. The acoustic probe was installed downstream of the choke, resulting in a high level of background noise. There is a strong dependency of the background noise on choke opening, gas rate, pressure and liquid production. The intrusive erosion probe does not suffer from background noise and appears to respond only to solids production. It is, however, less sensitive and the response is slower. A sand calibration test on the second well showed that the intrusive probe was accurate within a factor 3 prior to any calibration.

Preventing downhole sand transport
For the last 3 months, the gas rate from well one had been limited to a maximum 0.9 times 10E6 m3/d because of the larger sand rates experienced at higher gas rates. Figure 4 shows the well performance at gas rates exceeding 1.0 times 10E6 m3/d: erosion accelerates at the same time as the inflow performance reduces, as is evident from a simultaneous reduction of FTHP and the gas rate. Well performance shows inflow instability at downhole gas velocities above about 6.6 ft/s (2 m/s).

The reduction of well performance is probably caused by temporary partial drainhole blockage associated with slugging of sand and liquid. Startup procedures are in place to make sure that the well is beaned up in a gradual manner to avoid excessive transient downhole velocities and instabilities.

Conclusions
• The prediction of sand production rates and volumes typically suffers from significant
Figure 4. Increased sand production when downhole gas production exceeds ~2 m/s observed simultaneously on erosion probe and well performance.
uncertainty. In critical cases it needs to be constrained by either relevant offset well experience and/or representative sand testing before it can be used to help decide on downhole sand control.
• The prediction of sand erosion helps explore a realistic sand limit of the surface facilities. The calculated acceptable sand production rates and volumes can be much higher than the default values sometimes used for engineering design.
• The underbalanced installation of a wire-wrapped screen was achieved by combining a drillable bridge plug with a drilling BHA ahead of the WWS.
• The intrusive erosion sand probe provides a simple quantitative measure of sand production and erosion. The acoustic sand probe is much more sensitive but can suffer heavily from background noise and requires calibration.
• In regard to horizontal drainholes, the transport of loose sand appears to be governed by the onset of slug flow. This means that sand production to surface can be managed by restricting the downhole gas velocity. A gradual startup of the well can help prevent sand slug flow across the horizontal drainhole.