Erosion monitoring manages sand production

Sept. 28, 1998
Sand monitoring based on erosion measurements has substantially reduced the cost of producing fields in terms of controlling damage to process equipment such as piping, chokes, valves, and fittings. In addition to being a safety device, the on-line sand monitoring equipment can lower capital expenditures by reducing the production system tolerances with respect to erosion.
Reidar Birkeland
Saga Petroleum a.s
Stavanger, Norway

Svein Erik Lilleland
CorrOcean Inc.
Houston

Roy Johnsen, Nils A. Braaten
CorrOcean ASA
Trondheim, Norway

Sand monitoring based on erosion measurements has substantially reduced the cost of producing fields in terms of controlling damage to process equipment such as piping, chokes, valves, and fittings.

In addition to being a safety device, the on-line sand monitoring equipment can lower capital expenditures by reducing the production system tolerances with respect to erosion.

An on-line system mounted on a subsea christmas tree will verify and quantify the sand being produced. Consequently, this allows the operator to take preventive measures before the sand reaches the surface and clogs flow lines and piping.

Most present sand control involves downhole techniques such as screens or gravel packs. A gravel pack will reduce the chance of producing sand and in some cases eliminate the problem altogether. However, production that does not include erosive particles is rare, so therefore many operators now incorporate sand monitoring into their production equipment.

During initial well testing, an on-line sand monitoring system can establish optimum flow rates with regards to tolerable sand production. However, downhole conditions will change with time, resulting in varying sand production. This leads to the requirement for continuous sand monitoring.

Monitoring system

The CorrOcean a.s. sand and erosion monitoring system (SMS) is based on an intrusive probe, consisting of two to four sensing elements made from an alloy that does not corrode in the particular environment. Fig. 1 [97,799 bytes] shows the probe mounted into a pipe through a hydraulic access fitting.

Each sensing element is fed with a constant electric current. By measuring the potential drop across the elements, the change in resistance of each element can be monitored. As illustrated in Fig. 2 [64,965 bytes], there is an inverse relationship between the change in electrical resistance, R, and element thickness, T.

When sand transported by the produced media hits the probe, the sensing elements will erode, and the increase in resistance can be continuously measured. The thickness reduction of the elements can then be quantified.

To compensate for temperature changes, which influence the element's resistance, a reference element is included at the back of the probe body. This element is exposed to the same temperature, but is not exposed to the erosive sand particles.

The corrosion resistant sensing elements are cast in a high-quality ceramic into the probe body, which is made from stainless steel to ensure that the probe has the necessary mechanical strength and corrosion resistance.

The sensing element thickness is designed based on expected production rates, maximum tolerable continuous sand production, and average reservoir grain size.

Total element erosion can then be calculated to give the necessary element life for a specified safety margin.

Calibration

The original sand monitoring development project included an extensive test program in a 6-in. multiphase flow rig at the Norsk Hydro a.s. research center at Porsgrunn, Norway. The program established a reliable model for quantifying sand concentrations, when flow conditions and average particle size are defined.

The model was derived from several hundred test runs. Fig. 3 [73,085 bytes] shows an example from the tests.

If sand injection is kept at a constant rate, the probe responds linearly, indicating a constant metal loss rate.

As the sand injection rate increases or decreases, the slope of the metal loss curve will change accordingly. This linear behavior is a prerequisite for establishing a reliable model for the probe response as a function of the relevant flow parameters.

Based on the test runs, the following mathematical relation was derived:

R= k 3 S 3 D 3 V(G(water, oil, gas))

where:

  • R = Thickness reduction of sensing elements
  • k = Constant
  • S = Sand rate
  • D = Average particle size
  • V = Mixture velocity
  • G = Function including fractions of oil, gas and water.
The equation is the basic relationship between erosion measured on the probe elements and sand content when the process parameters are known. The accuracy depends on the accuracy of the input parameters described previously.

The most important parameter is mixture velocity, because this gives the effective sand grain impact on the probe elements. It should, however, be noted that regardless of the process parameters, the elements would experience metal loss if erosive particles were produced. The process parameters can be included later to provide a sand rate calculation over the monitored period.

To quantify the sand content from the metal loss on the probe, the necessary input parameters are:

  • Produced oil, gas, and water volumes
  • Average particle size
  • Pressure
  • Temperature.
Based on these parameters, the metal loss reading from the probe can be converted to sand content in the fluid. Note, that even if one does not know the actual flow rate, the metal loss curve will always give an instantaneous indication of the status of sand production.

Field tests

Several tests by different oil companies, both under controlled environments in test loops as well as under real production conditions, have examined the reliability and accuracy of the system.

These studies have confirmed that the system does not require on site calibration to quantify the sand content within acceptable accuracy. An offshore Norway test showed that the sand quantity determined was within a ±10% accuracy.

Erosion modeling

Metal loss of the probe elements gives a direct value of the erosion caused by the sand particles in the fluid. From a safety and maintenance point of view, erosion on the piping system is of the greatest concern.

CorrOcean, together with Norsk Hydro, has developed a model for calculating erosion in critical components in piping and production systems. The model is based on a large number of erosion tests performed under controlled conditions. These tests generated an empirical erosion model for piping and production components such as 90° bends, T-bends, choke valves, and other flow obstacles.

Measured erosion data from the sand and erosion probe are input in the model for calculations, currently in a standard Excel spreadsheet.

The direct use of the metal loss on the probe will reduce the uncertainty and improve the accuracy of the erosion calculations compared to calculations made with the sand contents measured with a sand monitoring system.

System description

The sand monitoring systems are available for both platform topsides and subsea installation.

Topsides system

The topsides system can be used both on-line and off-line ( Fig. 4 [95,787 bytes]). The sand and erosion probe is installed in the pipe through a 2-in. access fitting (hydraulic or mechanical).

In an on-line system, the station interrogates the sand probe at typically 2-10 min intervals, converts the analog readings to digital format, and transmits the data back to the control computer. Software in the computer quantifies the metal loss and sand content.

The complete system (SMS) normally includes a rack-mounted topsides control unit, configured as an industry-type PC with a small screen and possibly a keyboard connection.

To interface with a topsides control system, a digital interface with a proper protocol between the SMS and the Scada system must be defined. The SMS can support different protocols such as Modbus, Excom, and Profibus. This can be either a one or two-way interface, depending on the configuration of the Scada system.

The two-way communication allows the SMS to read on-line production data or choke settings from the Scada system, and report back metal loss and sand rate data. If on-line production data are not available for the SMS, these must be entered manually.

Off-line is an alternative system, where data are transferred manually from stored data to a PC. In this configuration, a hand-held terminal replaces the field-bus master and cables. This terminal is used to program the data storage and download data to the PC.

Subsea system

The subsea version of the system has to be installed as an on-line system. The sand and erosion probe unit is normally placed downstream of the choke valve. The complete sand and erosion probe assembly (SPA) is mounted onto a standard 21/16-in. API flange.

The ohms read by the sand probe are transferred through a digital RS232 or RS485 interface (four-wire cable) to the subsea control module on the christmas tree. Here the signals are compressed and transmitted topsides through a cable with ±12 v or alternating 0.24-v current.

The sensor is powered through this four-wire cable with ±12 v supply. Current consumption is about 70 ma, on average.

All probes are interrogated regularly by the signal conditioning units (SCU) typically every 20 sec. These data are then transmitted to the subsea control module (SCM) on the respective christmas tree.

The data are subsequently divided into "packages" and sent topsides for storage in the internal data base on the main control unit.

Monitor history

CorrOcean developed the sand and erosion monitoring system together with Norsk Hydro. The project was initiated in 1989. Norsk Hydro had observed weaknesses with existing technologies (including acoustic sensors) and wanted a system based on electrical resistance principles.

After a 3-year development and test program, the first commercial system was introduced in 1992. The system was tested by different operators both in Norway and U.K.

The first system sold was expected to be for a topsides installation. However, in 1992 Saga Petroleum a.s. needed a sand monitoring system for its new subsea field, Tordis. It had identified that there was a potential for improving the operability and reducing the operating cost by including a sand monitoring system that would meet given tolerances, and operate continuously over the field life without requiring in situ calibration.

After a thorough evaluation, Saga selected the CorrOcean sand and erosion monitoring system in late 1992. Since then, the system has been selected for several subsea fields around the world (Table 1 [41,495 bytes]).

The latest installation is for the Exxon U.S.A. Production Co. Diana field in the Gulf of Mexico, scheduled to be onstream in early 1999.

The first topsides system was installed on the Gannett field for Shell U.K. Ltd. in 1993. After this installation, several hundred units have been installed worldwide for on-line and off-line sand and erosion monitoring.

On the Britannia field, Conoco U.K. Ltd. has installed sand and erosion sensors both subsea and on the topsides. The idea is to have a system that is secure to prevent sand produced downhole from clogging surface flow lines.

The system has also been extensively used during well testing for maximum sand free rate (MSFR) determination.

Tordis installation

Since 1994, the CorrOcean sand and erosion monitoring system has been in operation on Saga's Tordis subsea production system.

The subsea sand and erosion probe provides continuous control of sand production and erosion through optimized production rates. Sand production causes a rise in the metal-loss curve from the probe (Fig. 5 [47,814 bytes]). Because the erosion tolerance is fairly low on the subsea piping, full control of the local erosion rates is required.

The Tordis field consists of six subsea completed oil wells. Each well is equipped with two intrusive sand and erosion probes installed on the christmas tree, downstream of the choke valve.

Fig. 5 is taken from a particular incident on one well, where a slug of sand was produced over a short time period, and about 1,200 nanometers (10-9 m) metal loss was seen on the probe.

This was clearly a clean-up incident because the metal loss leveled off afterwards and has since been stable. An amount of sand corresponding to what the system measured was later found in the topsides separator tank.

Fig. 6 [50,466 bytes] shows data from a maximum sand-free rate test on another Tordis well. It had been choked back to about 12,000 bo/d. During the test, the well was gradually beaned up to a threshold level of 21,400 bo/d over a 2-day period, with continuous monitoring of the sand production.

At this threshold level, the well started to produce sand, as can be seen from the instantaneous steep rise in the metal-loss curve, about 10 min after the maximum production level was reached.

As a result of this, the well was choked back slightly, which resulted in a complete stop in sand production. Well stabilization took about 10 min, as seen on the metal-loss curve, which levels off at a new plateau.

A delay also occurred from the time when the choke was changed, until the sand was detected on the probe. This was mainly due to the transport time of the sand from the reservoir and up to the christmas tree on the sea floor.

After this, the probe responded immediately to the onset of sand production from the well.

This test allowed a substantial increase in the well's production, and hence, increased field profitability. An increase from 12,000 to 19,000 bo/d at an oil price of $15/bbl indicates an increased income of about $100,000/day.

Water breakthrough also can lead to increased sand production. This is demonstrated in Fig. 7 [69,295 bytes], which shows erosion data from one probe on Tordis over 6 months. The metal loss is correlated with the water cut, as derived from manual bs&w sampling. Clearly, the onset of sand production corresponds well with the increase in water cut over this period.

At the end of this period, Saga decided to choke back the well, which is marked in the figure. This had some effect for a few days, but it did not completely stop sand production, as can be seen from the continued metal loss increase at the end of the curve.

Other field examples

A different clean-up incident ( Fig. 8 [50,335 bytes]) is from another subsea installation, Vigdis, also operated by Saga. Vigdis produces oil with eight wells. First oil was recovered in January 1997.

Again, each christmas tree is equipped with a probe to determine sand and erosion. Fig. 8 shows accumulated metal loss over a 16-hr period in January 1997. The increase in metal loss correlates with a certain decrease in the well's temperature, which is believed to be related to remaining water being produced in slugs that contained debris from the well.

Hence, the conclusion was that the increase in metal loss is caused by a clean up of the well, and not by continuous sand production from the reservoir.

Another concern for the operating companies is erosion in the topsides process system. The most critical part is the choke, but also bends and reducers have often been subjected to erosion caused by sand particles.

One operator in the U.K. sector of the North Sea was faced with this problem and it decided to install monitoring equipment. Each well has been equipped with the CorrOcean sand and erosion monitoring system, and data from the system are used actively to operate the wells.

As soon as sand is detected in a well, the choke opening is reduced until sand production ceases. The company cannot accept any sand into the system from a safety and economical point of view.

In a recent North Sea case, an operator expected combined sand and back production of proppants from the well. As can be seen from Fig. 9, the erosive particles started to be produced after about 2 weeks. The particle erosive effect was high and after 1 month nearly 0.014 mm of the probe elements were eroded away.

By using the erosion model, the company estimated that the corresponding maximum metal loss in the choke to be close to 0.1 mm during the same period. The company decided to reduce the choke to try to stop producing solids.

As can be seen from the curve, the reduced choke stopped the metal loss, indicating no further solid production.

By combining an active use of the sand and erosion monitoring system with the erosion-modeling tool, considerably money can be saved on reduced maintenance and inspection on critical components such as chokes and bends. According to information from operators, they replace chokes on some wells monthly due to lack of monitoring equipment for sand and erosion. Others are spending a lot of money on ultrasonic inspection on both chokes and pipe components to fulfill their safety requirements. An active use of the sand and erosion monitoring system will reduce the operation cost.

Future developments

The original sand and erosion probe has been further developed to integrate a multipurpose subsea sand and erosion probe, which can monitor erosion, sand rate, pressure, and temperature. The pressure/temperature unit is a diaphragm type that can be installed either in the end of the probe or in the middle section. The middle section is an optimal position with respect to temperature measurements.

Typical measurement range is 0-1,000 bar for pressure, and 20-200° C. for temperature. The unit will typically communicate with the subsea control module through an RS485 interface with a Profibus protocol, for example.

CorrOcean also supplies a multiphase measuring system, dubbed Idun, that calculates flow rates of water, oil, and gas from pressure and temperature inputs from sensors downhole and upstream and downstream of the choke.

This combined unit will reduce the cost compared to using two or three separate units. The combined unit will also increase the accuracy and repeatability of pressure and temperature measurements.

CorrOcean has also further developed the sand and erosion probe and multiprobe so that they can be retrieved subsea with remote operated vehicles.

Bibliography

  1. Hutchnings, I.M., "The erosion of steels by the impact of sand particles," University of Cambridge, August 1984.
  2. S ntvedt, T., "Erosion of Ductile steel components, Effect of erosion on CO2 corrosion on 13% Cr and Duplex steel," Norsk Hydro report R064947.
  3. Heitz, E., and Kohley, "Corrosion Tests of 13% Chromium Steel in Sand Containing Formation Water," KP-Nr. 329, DECHEMA.
  4. Tong, K.N., "The erosion resistance of pipe bends in pneumatic conveying systems," PhD thesis, Thames Polytechnic, August 1981.
  5. Dugstad, A., "Erosion in a T-piece IFE," Contract No. KO2507-00/ UP-U010.
  6. Braaten, N., and S ntvedt, T., "A new concept for Sand Monitoring: Sand Probe based on the ER Technique," Paper No. OTC6985, Offshore Technology Conference, Houston, 1992.
  7. Skavang, K., Braaten, N.A., Sirnes, G., and Johnsen, R., "Erosion based sand sensor: Subsea system developed for the Tordis field," Paper No. OTC7504, Offshore Technology Conference, Houston, 1994.
  8. Braaten, N.A., Blakset, T., Johnsen, R., and Birkeland, R., "Field experience with a subsea sand monitoring system," Paper No. SPE35551, 1996.
  9. Salama, M., "Sand Production Management," Paper No. OTC 8900, Offshore Technology Conference, 1998.
  10. Conoco Inc., Test of CorrOcean erosion probe in DNV test rig, DNV Report No. 96-3620, 1996.
  11. AEA Sand Detector Testing, Final report, AEA-TSD-0437-C, 1995.

The Authors

Reidar Birkeland is a production engineer with Saga Petroleum a.s, Stavanger. He previously worked for BJ Services. Birkeland has an MS in petroleum technology from Stavanger College.
Svein Erik Lilleland is vice-president of sales and marketing for CorrOcean Inc. in Houston. He previously worked for Pall Norge, Pall Europe, and the Norwegian Trade Council. Lilleland has a BS in mechanical engineering from the University of Portsmouth, U.K.
Roy Johnsen is director of operations for CorrOcean ASA, Trondheim, Norway. His responsibilities include product development, sales, and production. He has 20 years of experience with materials and corrosion. Johnsen has an MS and PhD in materials and corrosion from the Norwegian Institute of Technology, Trondheim. He is a member of NACE International.
Nils A. Braaten is the principal engineer for CorrOcean ASA, Trondheim, Norway. He has a PhD in physics from the Norwegian Institute of Technology, Trondheim.

Copyright 1998 Oil & Gas Journal. All Rights Reserved.