Real-time automation optimizes production economics in Oman

Ron Cramer
Shell Services Co.
Houston

Cleon Dunham
Shell International Exploration & Production B.V.
Netherlands

Alley Al Hinai
Petroleum Development Oman
Muscat, Oman

Petroleum Development Oman's (PDO) on-line, real-time production automation and optimization system has increased oil production by an average of over 5%, and is saving PDO an estimated $7 million/year in capital and operating expenses, while significantly reducing vehicle mileage.

The on-line system monitors real-time well data and provides optimal control of well processes through a comprehensive suite of well optimization hardware and software applications. The benefits are derived from accelerated oil production with fewer associated costs and reduced manpower exposure in gas-lifted, beam-pumped, and electrical submersible pumped (ESP) oil fields.

PDO, so far, has installed electronic instrumentation and software on more than 1,200 individual wells and associated production facilities. Gas lift operations are expected to yield up to 5% more oil production with 10% less lift gas. Beam pumping production has improved 5% because of faster response time, with a 35% increase in mean time between pump failures. ESP production is expected to increase by 3%, with pump life improvement.

Equally important has been reduced manpower for visiting remote desert wellhead sites and, with that, much less driving and hazard exposure. Based on this success, PDO is in the process of applying this technology throughout its Oman operations (Fig. 1 [180,630 bytes]).

PDO's system is part of a program in which Shell E&P operating units are aggressively installing on-line, real-time production automation and optimization systems in key oil and gas producing areas around the world.

Software suite

The software used in the Oman production optimization program is an integrated suite of well testing and artificial lift applications developed by Shell Services Co. for use by Shell E&P operations worldwide. This software is called the Shell Oil Foundation System or SOFS. Lift methods covered include gas lift, ESP, and beam pumps. These applications are integrated in four primary ways.

1. All applications have the same look and feel, which expedites assimilation of the entire suite by operations and various support staff. Integration also alleviates confusion in interpreting results, which are presented in a similar manner for all applications.
2. The applications interface to a wide variety of electronic data sources (such as distributed control systems, remote terminal units, etc.) in a similar, logical, and easy to apply manner. This leads to fast and efficient applications.
3. The applications are integrated for the people who analyze, interpret, and use the associated real-time data. This is achieved by making the real-time field data available on the desktop of those who need to know. This includes operators and supervisors in control rooms throughout Oman, as well as engineers, analysts, and technicians in the head office (Fig. 2 [25,013 bytes]).
4. The applications have the capability to interact in real time, exchanging data and, where appropriate, initiating appropriate actions. For example, prior to a well test, the well test application checks with the artificial lift application to verify that the well is fit for testing. If the well is fit, the test continues. If not, the operator is flagged and a more suitable well is tested. Note, this functionality is not currently in use in Oman, but is being applied by Shell in the U.S.

The applications can be integrated or configured to plug and play, either alone or in any required combination. The package is easy to build, change, and amend. Once built, the system is flexible, allowing the applications to be simply, quickly, and interactively extended to reflect process changes such as new wells and test separators.

Extraction of real-time data and interfacing to other systems is easy because the suite is based on open standards. In Shell, in the U.S., for example, the suite of real-time software is being electronically linked to other systems, such as product accounting.

Beam pumps

During the late 1980s, PDO installed rod pump controllers (RPCs) on some beam-pumped wells (Fig. 3 [14,282 bytes]) in South Oman. The main objective was pump-off control to avoid fluid pound.

PDO soon realized, however, that just shutting down the pump was not the optimum solution. Transmitting the pump status information to production operators and analysts to quickly diagnose problems would lead to faster and more effective corrective action. This would reduce the need for staff to drive around looking for problem wells.

Hence, the system was updated and enhanced, using the SOFS real-time computer assisted operations (CAO) system. SOFS runs on HP 9000 workstations, with data gathered automatically and transferred in real time to manned operations coordination centers and to engineers in the head office for analysis and interpretation.

The supervisory pump-off control application enables real-time monitoring, control, and optimization by gathering load and position data from the pumps and converting the data to downhole conditions. Generated plots indicate the condition of the downhole pump, traveling valve problems, and tubing movements.

The module also reports and flags run time and pump cycle deviations and wells off production. The module also provides remote start/stop of individual wells or groups of wells and remote rod-pump controller (RPC) parameter adjustment.

Reports generated show rod loading, loads imposed on the pumping unit, and estimated pump efficiency based on comparing well test rates and calculated pump displacement.

Fig. 4a [165,754 bytes] shows a surface card (load-vs.-position) and a downhole card for a beam-pumped well. The surface card is measured by sampling the load and position data 20 times/sec during the pump stroke. The downhole card is calculated by the dynamic downhole pump model. In the case shown, the well is pumping normally with a full pump barrel.

The beam pump application is also used for on-line beam-pumped design and prediction. For example, the effects of improved rod designs, or new pump sizes, speeds, or stroke lengths can be evaluated on-line. Basic design data are entered and the software derives the surface pump card, downhole pump card, and torque. This allows the engineer to determine optimal beam- pumped designs that maximize production and minimize maintenance.

Additionally, the system includes a common well test application with real-time links to the beam- pumped module. To be sure that the pumps are suitable for testing, these links check well status before the well goes on test. Similarly, if a pump problem develops during the test, it can be detected and the test stopped.

At the end of the test, pumping time is accurately calculated and pump diagnostic cards automatically gathered and analyzed. Historic well test information is integrated, to help evaluate results.

In cases where beam- pumped wells are producing little or no gas, gross flow can be derived from on-line measurement of mass flow. Using measured density and known water and oil densities, water cut can then be established, allowing automatic and continuous calculation of net oil.

PDO initially applied this technology to a small number of wells. Production before and after installing the software was carefully benchmarked, and sufficient gains were confirmed to justify expanding the technology to all of PDO's 900 beam-pumped wells.

Experience to date indicates a 5% increase in sustained production as a result of early recognition of problems, remote start-up capability, and fewer mechanical failures. Additional benefits include a 35% increase in mean time between pump failure, reduced power consumption, and more effective use of manpower.

A key benefit has been a reduction in the need for operators to visit remote wellheads. This has improved safety in that there is less driving and less exposure of people to hazard.

Gas lift

Following successful application of CAO technology to beam pumps, PDO applied similar concepts to its gas-lift wells in the Yibal field.

About 55% of PDO's 840,000 bo/d total production is gas lifted (Fig. 5 [28,440 bytes]). A gain of up to 5% in production is expected from the suite of gas-lift optimization programs. The Yibal field was chosen as the pilot area for the gas-lift optimization program because it is the largest PDO field, with 320 vertical and horizontal wells, producing more than 230,000 bo/d, and using more than 200 MMscfd of lift gas.

Gas-lift performance models were constructed for each of the wells using WinGLUE, a PC-based Windows version of the Shell developed gas-lift user environment (GLUE) program. The WinGLUE program was selected as the standard for gas-lift surveillance because it provides a level of integration and functionality superior to other commercially available programs.

In WinGLUE, a complete data profile for each well can be accessed for well surveillance optimization or design changes. The program's primary benefit is its ability to match the production pressure model to reflect the actual flowing pressure gradient survey and gas-lift valve opening and closing pressures.

Having matched the model to the actual field measurements, lift-gas performance curves can be generated for all wells in a gas-lift system. From these, a lift-gas distribution run is made to provide an optimum field production rate for a given total lift-gas availability.

A relatively autonomous subset of the Yibal field was chosen as a pilot site. Electronic instruments were installed to measure gas-lift injection pressure and flow, and tubinghead and casinghead pressure. Also, bulk lift-gas flows and pressures were electronically measured for all supply and offtake points.

All remote real-time signals were routed back to the manned operations control center computer. During a 10-month period, 52 wells were reviewed and optimized through changes to lift rates, valve settings, or completion strings.

A critical success factor for the pilot was the ability of gas-lift algorithms generated by the software to optimize the wells both individually and collectively. The collective aspect provides a real-time open or closed-loop control system that can automatically distribute lift gas to individual wells while monitoring system gas supply and demand.

Once each minute the system samples field data to automatically determine the status of the gas-lift distribution system. If there is an imbalance caused by a significant change in either supply or demand, the system re-allocates the current supply to the wells. It does this by determining the appropriate amount of lift gas to inject into each producing well to optimize its production at the current level of gas availability.

The Yibal pilot included collective gas-lift-system monitoring, control, and optimization by keeping the gas-lift system in balance during compressor or flow station upsets, or during normal operational fluctuations, such as changes in compressor efficiency caused by ambient temperature changes.

Lift-gas wastage, which had been observed throughout the pilot area, was reduced by optimizing the gas supply distribution across the entire field and by keeping the lift-gas distribution under the control of the operator at all times. The operator has the option of either automatic allocation based on one of several available optimization methods, or direct manual allocation.

Under automatic control, the gas-lift system pressure is kept to within ±1 bar. The two primary gas-lift allocation methods are: one based on changes in rate, and one based on changes in pressure.

For the first one, several add-on extensions are included for fine-tuning to maintain stable system pressure.

Individual gas-lift wellhead monitoring and control was fostered by providing intelligent wellhead monitoring and alarm capabilities, and by maintaining pressures within operating limits to keep both the entire system and individual wells stable.

Exception reporting was used to monitor individual wells. This allows the operator to concentrate on problem wells and rapidly identify candidates for further well work. Increased production resulted from efficient identification, diagnosis, and correction of deviant lift behavior.

Definite gains resulted from having on-line information. Most of these were time-related, in that data were available immediately to enable the necessary corrective action.

The Yibal gas-lift pilot resulted in 5% acceleration in oil production, using 10% less lift gas. Based on this success, PDO decided to apply the same solution to the entire Yibal field. It now has been successfully in service for over 1 year.

PDO is now applying the optimization program to all its Oman gas-lifted wells.

Fig. 4b shows the gas-lift performance plot for a continuous, single string, gas-lift well in the Yibal field, Oman. The gas-lift distribution (line) pressure is very stable as is the gas-lift injection rate. However, the well is producing unstably as seen by the fluctuations in the injection pressure and the production pressure, which cycle about once/hr. As can be seen from the production pressure, the well actually stops producing in between heading cycles. This is very inefficient and requires corrective action. This well has a known leak which needs to be addressed.

Fig. 4c shows several variables for the total Yibal field lift gas supply including the system pressure and the net lift gas supplied to the wells. The field contains 320 continuous gas-lift wells, and the primary objective of the gas-lift delivery system is to maintain stable system pressure.

ESP optimization

Real-time artificial lift optimization also has benefited PDO's ESP wells by enabling them to produce at or near capacity, with minimum downtime and reduced capital and operating costs. For the ESP wells, real-time data are available to the operator in displays, trends, and reports designed to aid problem diagnosis.

For example, a combined pump and well lift performance index (LPI) is used to quickly identify the status of the total ESP system. LPI trends indicate impending deviations from normal operations.

To identify suboptimal production, the LPI also facilitates comparisons of ESPs in the same field and from field to field.

The software recognizes pump upsets or failures and automatically captures detailed information coincident with the incident. This can be replayed later to help with problem diagnosis. Much of these data are used to help design, or redesign, the installation so that when a new pump must be installed, it will better match downhole conditions. This will increase ESP productivity, efficiency, and lifetime.

Fig. 4d shows the performance of an ESP pumped well. The upper curve is the manufacturer's theoretical pump performance curve based on the pump specification, number of pump stages, and operating frequency. The lower curve is the actual pump performance curve based on the measured head (from measured downhole pressure) and production rate (from a well test).

The horizontal parallel lines show the theoretical head loss, and the vertical parallel lines show the theoretical deferred production. The dashed line, which trends to the upper right on the plot, is the well's inflow performance relationship (IPR) curve. The solid line, which curves to the right, is the break hosepower curve.

PDO has now applied this optimization technology to more than 200 ESP pumps and expects a 3% acceleration in oil production, as well as increased pump life.