The tool tested successfully for pump protection in a Canadian oil field, leading the manufacturer to create a line of downhole desanders specifically for oil and gas applications.
Steve Coffee, Enerscope Systems Inc.; and Michael Briffett, Husky Energy Inc.*
A field test in Canada demonstrated the effectiveness of a downhole desander to protect Electric Submersible Pumps (ESPs) from failure caused by the production of formation sand. An abrasive environment can cause radial wear of the pump bearings leading to vibration, which often results in equalizer and motor failure. Erosion of diffusers and impellers also reduces the efficiency of the pump, resulting in less than optimal production.
Renaissance Energy tested the desander as part of its efforts to reduce operating costs, 24% of which it attributed to well workovers and related servicing costs. High-volume lift systems and sand production were the dominant factors contributing to these high costs.
BACKGROUND
Renaissance, which was purchased by Husky Energy after this study, identified its Cantuar Unit in southwestern Saskatchewan as one of its critical areas for cost reduction. In this unit, the producing reservoir is fine- to medium-grained quartz arenite, semi-consolidated sandstone that is 39-46 ft (12-14 m) thick, laterally continuous and fairly homogenous. The sandstone has a 25-28% average porosity and a 1-darcy average permeability.
In June 1999, Renaissance had 30 operating ESPs in the Cantuar Unit, and plans called for doubling that number of pumps, but only if something could be done to control the pump damage and related costs. The average ESP run life in the unit was 406 days, with some pumps running as few as 103 days.
In search of a solution to formation sand production, Renaissance decided to test the Lakos PPD desander for downhole pumps, supplied by Enerscope Systems Inc. The selected desander had been widely successful in pumped groundwater wells.
The downhole desander works in a similar fashion to the company’s surface production-stream desander, in that both tools employ centrifugal action to separate the heavier-than-water particles from the water stream, Fig. 1.
 |
|
Fig. 1. The desander employs centrifugal action to separate the heavier-than-water particles from the water stream.
|
|
Sandy production fluid is drawn through tangentially aligned inlet slots at the top of the tool and into the separation chamber, where sand is centrifugally separated from the fluid and tossed to the perimeter of the chamber. The sand particles fall downward along the inner surface of the chamber to the bottom of the separator, where they discharge through a flapper valve into an unperforated section of the cased hole that acts as a sump. Meanwhile, suction of the downstream pump draws the sand-free fluid at the center of the chamber up through the vortex outlet.
Each unit operates within a prescribed flow range corresponding to a minimum and a maximum pressure loss regarding solids loading and other factors. The desander has no moving parts, no screens or filter elements to clean or replace, and requires no backwashing.
Production can proceed without interruption because removal of the separated particles does not require pump or system shutdown. The downhole desander uses a flapper valve that automatically discharges sand into the sump once the weight of the sand within the unit reaches a threshold. After the sand is dumped, the unit’s suction automatically closes the valve.
Use of these downhole desanders, patented in 1968, in the groundwater industry had produced at least a fourfold increase in pump life with millions of installations worldwide. To test the downhole desanders in an oilfield context, Renaissance installed 12 of the units in wells over a 2-yr period starting in August 1999. Unfortunately, only one unit was pulled and had performance data recorded before Husky bought the asset, after which data recording from the desanders was discontinued.
TEST OBJECTIVES
The chief test objectives established by Renaissance for the pump-protection desander were that it separate sand downhole and prevent the production equipment downstream of the desander from sanding up. Success of the first objective would be indicated by sand buildup in the well sump and a lack of erosion or wear to the ESPs, because Renaissance did not install sand-production monitoring equipment. Success of the second objective would be determined if the equipment could be removed from the well, trouble-free, at the time of inspection. Otherwise, the equipment would be expected to be stuck in the well.
The operator’s ultimate goal was for the desander to extend the ESP’s producing life. This would be evaluated based on past pump data, run-time measurement and pumping efficiency calculations.
At the same time, it was understood that the desander installation would reduce oil production rates, although sand buildup in the well would have also reduced oil inflow. This condition could be determined by monitoring and comparing total fluid and oil rates.
OPERATING PARAMETERS
Although oilfield water can be corrosive, the Cantuar Unit fluid did not appear to pose such a risk to standard carbon steel construction. An inspection of existing desanders in the Renaissance production streams also dispelled any concern for abrasive wear to the pump-protection desanders.
The supplier, however, made the test desanders with stainless steel and a Viton flapper valve to achieve a material hardness of 95 Rockwell B. This effectively neutralized any potential for wear to the desander so that the test objectives could be isolated from other possible failures, no matter how remote.
An External Upset End (EUE) pipe connection thread matched the field equipment specifications. No other operating conditions influenced any changes in material or design specifications.
Operating conditions were:
- Electric submersible pump
- 1,730-6,040 bpd (275-960 cu m/day) flow
- 2% maximum gas rate by volume
- 95% watercut
- Sand, silt and clay solids
- Solids specific gravity of 2.6
- 1 µm-9.5 mm particle size
- 1-10,000 ppm solids concentration (100-ppm average)
- Ambient temperature.
Because of the concern for excessive sand accumulation in the well, Renaissance selected wells with boreholes of 7-in. ID, which would provide ample annular space in the event that the pump sanded up and the equipment needed to be fished out. Adequate well depth below the pump setting was also a factor in well selection. Wells selected had at least a 66-ft (20-m) sump.
Oilfield ESPs offered no substantial operating differences from the groundwater pumps commonly installed with the pump-protection desanders. The most important selection factor was good pump history. This history established benchmarks for comparing pump run life and efficiency.
To maximize the testing objectives, Renaissance selected wells with pumps that had a short run life. It selected pumps with a moderate flowrate of 1,730-3,020 bpd (275-480 cu m/day) of fluid.
A steady flowrate, as compared with periodic slug flow, optimizes desander performance. The expected 95% watercut posed no risk to the desander’s performance.
The desander’s maximum gas handling capacity is 2% Gas-Oil Ratio (GOR). Higher GORs could be handled with additional sizing calculations.
The desander was installed using a packer, effectively isolating the ESP from the sandy fluids entering the wellbore through the production perforations and directing flow through the desander upstream of the pump intake. This method required two procedures to set the pump and the desander in the well, whereas the pump and the desander could have been set in one procedure along with a shroud for the ESP or with a perforated pup joint and seal cups, Fig. 2. However, setting the desander with a packer also would minimize retrieval complications if the well sanded up; only the desander would be sanded in, not the pump.
 |
|
Fig. 2. The desander and pump can be set in the well using only one procedure with an ESP shroud (a) or perforated pump joint and seal cups (b) to isolate the ESP from the sandy fluid stream, or the desander can be installed using a packer, and then the pump can be installed separately (c), as was done in the test well.
|
|
RESULTS
Renaissance shut down and pulled the first test desander for inspection in April 2000, after operating it for 203 days.
The test well had a history of severe sand damage. Its three previous pumps lasted 103, 132 and 143 days. All operating conditions at the time the first desander was retrieved appeared to be stable; i.e., the only reason Renaissance shut in the well was to gather test data. The inspection after 203 days clearly indicated that the pump would continue to operate at reasonable performance for some time to come. In fact, the pump was still operating at an efficiency of 60%, optimal for its design and application conditions, Table 1.
| TABLE 1. Test well operating and final conditions |
 |
|
The pump and desander were pulled without incident, and there was no evidence of sand accumulation above the packer, indicating little or no occurrence of the pump sanding up. The well sump was found to be 15.3 ft (4.67 m) shallower than its original 98.4-ft (30-m) depth due to an accumulation of 115.8 L of sand in the well sump, providing strong confirmation that the desander was removing sand from the production fluid, discharging it into the well sump, and keeping sand out of the pump.
On-site pump inspection indicated that the pump was in good condition, with no shaft play and clean oil throughout the equalizer and motor. After the ESP was shipped to the pump manufacturer for a full test and comparison to original bench tests, the pump manufacturer measured a head capacity reduction of only 4.6% at its best efficiency point.
A thorough breakdown inspection of the pump revealed that the impellers and diffusers had only minor sand erosion and the radial support bearings had minimal damage. This condition compares very well against the catastrophic failure of the three previous pumps after much shorter operating cycles in the same well.
The desander supplier inspected the desander and considered it to be in excellent condition. The company rated a minor polishing at the edges of the inlet slots as normal and insignificant wear, found no evident internal sidewall erosion, and noted minimal wear on the internal deflection pedestal. A slight tear of the flapper valve was observed, but its origin could not be determined. The tear did not appear to reduce operation or performance.
The test indicated no negative impact on oil production rates. The well continued to produce oil at an average rate of 157 bpd (25 cu m/day), which matched the predicted oil rate, based on inflow performance relationship data. The data indicated that the desander did not significantly reduce inflow.
CONCLUSIONS
The test desander matched or exceeded all of the established expectations. As a result of this testing, the supplier in 2001 began marketing a downhole series of pump-protection desanders engineered specifically for the oil and gas industry, featuring industry-standard materials, connections and specifications.
Although Husky did not study the results of the other 11 desanders in the initial trial, it has installed a total of 30 units in its wells since the first test. In some of these installations, the company is reporting an average run-time increase of about 110-120%. Other installations have been for Thums/Oxy, Schlumberger, Reda, Indian Oil and, recently, Plains Exploration and Production Company and Great Eastern Energy Corporation Ltd. 
|
THE AUTHORS
|
|
|
Steve Coffee is the new-business development manager for Enerscope Systems Inc., a Canadian supplier of equipment and engineering support for the oil and gas industry. Mr. Coffee has 15 year of filtration experience in the international market, with a focus on solids-from-liquid filtration for upstream and offshore production. He earned a BS degree in business administration with an emphasis on marketing at California State University and is an active member of the Produced Water Society, the Produced Water Club and the Society of Petroleum Engineers. Mr. Coffee can be contacted at steve@enersopesystems.com.
|
|
| |
Michael Briffett was a production engineer for Renaissance Energy at the time of this study, and was retained by Husky Energy Inc. in its acquisition of Renaissance. He now works as a completions engineer with Husky’s operations offshore Canada’s East Coast. Mr. Briffett holds a BS degree in mechanical engineering from Memorial University of Newfoundland. He can be contacted at mike.briffett@huskyenergy.com.
|
|
|
