Printable View

SPE 89870

Petrozuata - An Application of Multiphase Metering

Luigi Bortolin SPE, Petrozuata C.A; Parviz Mehdizadeh SPE, Production Technology Inc; Gordon Stobie, ConocoPhllllps

Abstract

The application of Multiphase Meters (MPM) over the past decade or so has, in the main, been in deployment of meters in small quantities (i.e. ones and twos) and there are few applications where MPM's have been deployed in bulk. Petrozuata in Venezuela is such an operation where 37 MPM's were deployed and have been in use for over 5 years.

This paper describes the facility and the operations where MPM's have been selected, tested and implemented. The paper also describes the diffculties experienced and the operational results from the extensive use of such measurement techniques.

Introduction

Petrozuata is a joint venture Strategic Association owned by eonocoPhillips (50.1 percent) and Petróleos de Venezuela, S.A. (PDVSA), the national oil company of Venezuela (49.9 percent). The project is a fully integrated crude oil processing and petroleum business, located in the state of Anzoátegui, Venezuela. It began commercial operations on April 12, 2001, however Extra Heavy erude Oil (EHeO) began flowing in mid 1998. Petrozuata's primary function is to produce EHeO from the Zuata region of the Orinoco Oil Belt; transport it to the Jose industrial complex on the north coast of Venezuela; upgrade it into 19 to 26.5 degree API synthetic crude; and market it along with 14 degree API gas oil and associated by products e.g. LPG, sulfur and petroleum coke.

The Petrozuata "project" is now an operational oil producing business with over 5 years production experience. The Strategic Association has a 35-year operating life and will require the driling of more than 750 wells with an estimated recovery of approximately 1.6 billion barrels of Extra Heavy erude Oil (EHeO) during this period. This facility uses the eonocoPhillips' proprietary coking technology to upgrade heavy crude oil into lighter synthetic crude and has a nameplate capacity of 120,000 barrels per day (BOPD). At present, Petrozuata produces more than 125,000 BOPD of EHeO. The synthetic crude oil produced by Petrozuata is used as a feedstock for eonocoPhillips' Lake eharles, Louisiana, refinery and the eardón refinery in Venezuela, operated by PDVSA.

Since 1997, Petrozuata has drilled more than 260 wells (at present there are 195 active producers) in an area of 56,000 acres of the Zuata region with the expectancy of drilling a further 490 wells over the next 30 years in order to drain the reservoir. Wells are clustered around 37 production pads as shown in Figure 1.

Early on in the design phase, it was perceived that a large number of wells and production pads over a substantial area would be required to successfully drain the reservoir. Traditional oil and gas extraction techniques would have required flow stations or pads containing a test and a production separator such that gas (and possibly water) could be separated from the hydrocarbon liquids. Knowing the crude oil viscosity and density, such separators would have been relatively large to accommodate the residence time and would have required heaters and chemicals to control foaming (and possibly assist in oil-water separation).

eonventional gravity based test separators and their associated equipment are high cost items and would have imposed a CAPEX cost of around $1-2MM per unit plus the associated operating costs. In order to reduce such costs, the project initiated a study in the conceptual engineering phase to determine whether multiphase production of fluids from the remote wells and pads to a central processing facility was a feasible option. The study concluded that multiphase pumping of the well stream products was feasible. Once the design had moved to a central processing focus for "pumped multiphase products", the requirement for the multiphase measurement at the well pads became an integral part of the design.

Qualification of Multiphase Metering Technologies

ConocoPhillips recognized in the 1980's that multi phase flow measurement would be a key to enhanced production operations and in reducing the cost of field production allocation and measurement. Early on, the operational limitations and cost of ownership of test separators and well testing was understood. eonsequently, the eompany has spent considerable efforts (in time and money) in the advancement and understanding of multi phase flow metering. The eompany concentrated on applications for extreme production areas - i.e. offshore, where small in-line MPM's would score heavily over relatively heavy, cumbersome and expensive test separators. As a result there was considerable support in-house for multi phase metering and measurement technology assessments. These were supported by pilot tests at the eonocoPhillips' multiphase test facility at North Maurice Field in Lafayette, LA. These activities were further supplemented by participation in a number of general and dedicated heavy oil based Multiphase Meter Joint Industry Programs (4,5).

The technology awareness heightened the potential benefits that multi phase measurement technology could bring to projects such as Petrozuata. Early in the Petrozuata project, based on a cost benefit analysis, comparing multi phase pumping and multi phase meters versus conventional well testing, the project concluded that there was a high probability of successful implementation of multi phase measurement technology if a number of multiphase measurement techniques were tested and a single vendor selected for further joint development and testing. In pursuit of this objective, a technology assessment was conducted in the early 1990's in order to establish what was available. From the current (2004) perspective, it is diffcult to remember that MPM technology and systems were in prototype form and some aspects of the technology we recognize today were not available. The technology review commenced with a desktop exercise and considered six MPM systems available.

The technology review considered such factors as:

• Current deployment
• Known/reported performance
• Type of service
• Whether it was an integral meter or a system
• Data communication
• Sizing, API gravity, viscosity and temperature history
• System availability, service and costs

The technology assessment identified four metering vendors, representing different technologies that were available and had already passed a number of initial performance qualification tests. These system vendors were selected for further evaluation.

These MPM systems used different measuring strategies and techniques. Their operating principles are described in references 2 and 3. After a number of initial qualification tests to check the performance of these meters under simulated field conditions (4), the Agar MPM was selected for this particular application and for further pilot testing in the field. The operating principle of the Agar metering system is described in reference 6. The primary reasons for selection of the Agar meter were:

• Meter performed well over a wide range of viscosities.
• The measurement strategy and techniques were less sensitive to fluid properties. This allowed the meter to operate in varying well fluids.
• The metering system provided capability to conduct volumetric verification tests of the measuring devices against the initial (factory) settings.

Production and Well Testing Pilot Tests

Two sets of pilot tests were conducted to assess the design and performance of the Agar meter under more realistic field conditions. These tests were conducted at the San Diego Norte and Bitor fields in Venezuela. The field and fluid conditions at these sites are summarized in Table 1. These conditions afforded the project an opportunity to check the performance of the meter for handling high viscosity production streams and varying field temperatures. It also afforded a chance to see how the meter reacted to the influence of low viscosity diluents, injected into EHeO, and the resulting variable mixtures of diluent and heavy crude. The flow rates available from the wells also provided an assessment of the meter's turndown performance.

Table 1 - Field/Fluid Conditions for the Pilot Tests

A further objective of these pilot tests was to address a number of 'issues of concern' with regards to the multiphase measurements in heavy oil production. The issues of concern are discussed in the following sections.

Diluent injection into an EHCO slugging flow stream would cause viscosity variations. These could range from the viscosity of heavy crude to that of pure diluent. Diluent injection in foamy oil tends to finger through the foamy mixture and is not homogeneous. The impact of varying viscosity on Venturi flow measurement devices used in MPM was not fully understood.

Diluent injection control is diffcult in a slugging flow and the diluent proportion is variable, giving rise to a range of densities in the fluid. This problem was known to have caused multiphase measurement problems in other heavy oil developments. Figure 5 is an example of the comparison between the MPM and tank test measurements for the three different wells and different diluent injection rates. The instantaneous tank test measurements for 3 wells are shown in Figure 5. In the actual qualification tests for MPM the tank tests were obtained after 30-40 hours residence time. These tests showed that the multiphase meter was able to address the issues discussed in the above sections and measured liquid rates under field conditions accurately.

Field Installation and Start Up Issues

The MPM's were installed during 1998 and 1999. Prior to dispatching from the vendor in Houston, the meters went through a Factory Acceptance Test (F AT). The FAT covered a wide range of flow rates, water cuts and gas volume fractions. During the deployment, design improvements were incorporated into each batch of the meters delivered as dictated by field experience and evolution II hardware and software.

The introduction of this 'new' technology highlighted the need for educating field personnel in the technology, equipment and its maintenance. Initially, the vendor provided training classes for key project engineers, and Petrozuata provided training for the field operators. On site, the vendor provided field support at the start up, and also assisted field technicians with training and hands on assistance with system maintenance and operations. The presence of the vendor at the site was a significant 'assist' to the technology acceptance.

Sand production was one of the issues that had to be addressed. Sand production rates are often specified as 0.5% by volume. The rate is however rarely a constant. It will rise and fall during production and may be 0.1 % for extended periods, and at that level is not a problem. However, events such as shut downs and start-ups can cause sand production to peak for a short period at 5% or more. The PD meter within the Agar multi phase metering system (6) is susceptible to foreign matter and may become worn, scored, or seized. Following the experience at Petrozuata, the site upgraded the external screens and the vendor was able to redesign the PD meter to increase its tolerance to foreign matter with sand tolerance up to the required leveL. The redesign consisted of more robust materials of construction and machined PD meter gear lobes to increase resistance to erosion products. It is important to note that the meter is still basically susceptible to gross ingress of large solids and foreign elements. Well start up is a time when sand may be present in the flow stream. During this time the multiphase meters are bypassed until the wells have cleared any erosional materials.

Another problem observed early on was plugging in the pressure and differential pressure transmitter impulse tubings. This instrument tubing was originally filled with silicone oil in order to prevent contamination and plugging. This required an intensive manual effort to ensure that the silicone oil was kept topped up and was not 100% successfuL. To overcome the manual effort, the vendor was requested to develop a more suitable system. A semi-automated system, requiring minimal effort to recharge the impulse tubing lines has been provided by the vendor and has performed welL. Subsequent instrument failures due to bitumen plugging have been significantly reduced.

Periodic Verification of the MPFM Performance

A periodic meter verification test procedure was developed and used in order to verifY that the meter is operating correctly. The volumetric verification test was accomplished by flowing a single-phase liquid at 10% and 50% of the fullscale flow rate of the PD meter through the MPM. This allows the metering components to be tested and compared with the initial factory calibration. The MPM is made up of three flow meters in series (6). When it leaves the factory, all three flow meters agree with each other, over the operating range, to better than :f0.5%. If any drift becomes apparent, it can be evaluated with respect to any action or replacement. The vendor recommends that this self-verification test (SVT) be carried out every 6 months. The current practice is to conduct SVT every 3 months and after every repair.

MPFM Performance Envelope

The MPM's performed well from start up, however some of the wells produced at rates, that were significantly lower than anticipated. These rates were below the low end of the meter's performance envelope, which was specified to max liquid rate of 5,000BPD. Rather than re-ordering smaller meters, the vendor was requested to study the problem and determine if the performance envelope could be extended. Initial tests with flows between 150 - 660 BPD showed that the MPM error could range from -40% to +25%. Following an extensive series of tests and a comprehensive analysis of the flow data using stock tank measurements the vendor was able to build a model to "curve fit" the meter at the low flow conditions. Upon implementation of the modified software, the tests were repeated. The "curve fit" removed the systematic bias and reduced the error band significantly. The improvement in MPM performance is shown in Figure 6.

It is believed that the 37 MPM's in service have logged about 75,000 hours of testing service. Tables 2 to 4 show the hierarchy of equipment, environmental and operational factors that contributed to the meter's initial malfunctions. Many of these factors can and have been addressed with proper maintenance and operator training. eonsequently, the frequency of malfunction associated with a number of these factors such as incorrect operation, calibration, grounding, etc have decreased. Shortly after start up, a major computerized management system called SAP was implemented and the SAP Maintenance Module has been used to monitor the MPM maintenance data.

Over a 21 months period, from January 2001 to September 2002, there were 97 well testing equipment failures (Table 2). This amounted to an average failure rate of 4.6 per month. A review of these reported failures has shown that some were due to mis-reporting and operational conditions and not related directly to MPM itself. Of the 97 failure reports, 25 were directly related to malfunctions in the MPM. The failure rate directly related to all 37 MPM's is therefore about 1.2 per month as an average over the period (0.032 per month per unit). A further review of the maintenance data from October 2002 to March 2004 (Table 3), when the MPM operations had gained some maturity, shows that the average reported failure rates had decreased from 1.2 failures per month to about 0.94 failures per month for all 37 MPM's (0.025 per month per unit ).

Table 2 - Maintenance effort logged for Jan 2001 to Sept 2002 (21 months) for Total 37 MPM's in service.

During the latter part of 2003, a maintenance check/audit was carried out on all of the meters. This audit assessed the role of a number of operational factors, listed in Table 4, which contributed directly to the meter malfunctions. Further studies of the MPM maintenance requirements reveals that on average each MPM requires about 120 man-hours per year of maintenance effort and about $2500/yr in parts. For 37 multiphase meters this might appear expensive but is considered attractive compared to the cost of a conventional well test system.

Table 3 - Total Maintenance effort logged Oct-02 to Mar-04 (18months) for Total 37 MPM's in service.

Table 4 - Hierarchy of Operational Factors Causing Multiphase Meter Malfunctions

Well Testing and Operational Lessons

A number of lessons were learned in the large-scale deployment of MPM technology and its application to well testing and production measurements. These are summarized in the following sections.

1-During the testing of high gas rate wells, the safety interlocks, used to protect the MPM from overspin, would react to small gas surges and put the meter into bypass. Working with the manufacturer and reviewing specific meter installations and details of the gas slugging and overs pin potential, the limits were broadened which allowed a higher number of wells to be tested without special equipment.

2- The type of MPM used in this project requires the manufacturer to build-in safeguards that are adapted to the flow range and which represents a conservative protection strategy. By working directly with the vendor and more accurately describing the operating parameters, the meter capacity can be increased without sacrificing safety or increasing costs significantly.

3- The MPM used in the project was designed for lower solids content than has been seen in the field. This primarily affects the PD meter in the MPM by increasing wear. The PD meter design was modified to add machined slots on the meter lobes and hardened surfaces within the meter body. This is now standard on all new meter purchases and the existing meters were retrofitted. The field has also implemented better upstream screens to catch sand and larger foreign bodies. To date this has virtually eliminated solids related failures (other than line plugging).

4- The initial instrument impulse tubing on the MPM had a buffering fluid, but repeated calibration failures due to bitumen plugging was experienced and required regular manual cleaning. Skids were retrofitted with a purge system using glycerine as the fill fluid. This system allows an operator to purge the instrument impulse tubing with a manual pump without opening the lines.

5- Initially the metering systems were not designed to interface with the field SCADA. These systems had to be retrofitted to allow full data exchange with the field SCADA system. This capability now allows on-line control room monitoring of well tests and a higher level of interaction for automatic well testing. eontrol room operators now initiate well tests without the need for field operator intervention. This is a great time saver as production pads are widely spaced. The ease of operation and reduced well test period means that wells are tested twice per month, increasing reservoir surveillance.

6 - Many of the wells have low water cuts (less than 2% WC). The water-in-oil metering section of the MPM was originally specified for 0-100% watercut application. At less than 2% of full scale, with 1-2% absolute watercut error, reported watercut does not always represent the water volumes produced. Manual shakeouts are used on wells to supplement water readings within the MPM. Operators perform these tests routinely to validate well tests and check the wells for solids content. Many EHeO fields operate with low water cuts. The need for full range (0- 100%) water cut measurements in the early stages of field development should probably be challenged. This section of the meter is expensive and having the ability to add it later when water breakthrough occurs may be an option to be considered.

MPM Field Performance - Production Measurements

The MPM's have performed well from the start of operations. Figure 7 shows the monthly allocation factors for the field during the period 1999 through 2003. The allocation factor is the ratio of the summation of well test divided by the fiscal oil volumes on a per month basis. The allocation factor reflects a number of parameters including the well testing practices used, the frequency and duration of well tests, and extrapolation of well test data to well rates. The allocation factor is also an indicator of the accuracy and repeatability of the MPM measurements. The drop to zero in field allocation factor represents the period when a national strike occurred and production was curtailed.

In the early production stages, the typical daily allocation factor was in the :t 20-25%. The monthly "fiscalized oil field allocation" factor varied from -0% to +45% as shown in Figure 7. This was perceived to be high and unacceptable and was probably due to a combination of MPM measurement uncertainty, as well as, external factors due to the plant start up and the management and operational systems being developed and fine-tuned. After the extension of the meter's dynamic range, the improved maintenance regime, the improved MPM calibration and system familiarization, the typical daily and monthly fiscalized allocation factor was enhanced to about -10% to 0 % which is considered to be satisfactory. However, it has been proven that daily allocation factors of up to :t3% could be achieved if MPM and well tests practices are followed properly. Gas is not a major production factor in the Petrozuata EHeo operations, and the actual total daily gas production is about 30 MMSeFD. The gas monthly allocation factor is around :f5%.

The high investment in prior research by PDVSA and eonocoPhillips was important in bringing this technology to fruition, especially as the proposal to move forward with multiphase metering technology was made in the early 1990's. In retrospect, the eonocoPhillips multiphase meter R & D was at the time focused towards severe service (offshore locations) so its implementation in an onshore area was surprising. The early testing programs developed knowledge of the crude characteristics and assisted the use of multi phase meters.

The Petrozuata partners have a number of heavy oil projects elsewhere in Venezuela and worldwide, where multi phase metering technology is being considered. In many cases the experience gained in the current operation wil be of great assistance elsewhere.

Acknowledgements

The authors wish to thank many current and former colleagues in PDVSA and eonocoPhillips in the assistance and encouragement in preparing this paper. They acknowledge the work done by lNTEVEP, Bitor and the many vendors, who participated in the early testing, and Agar eorporation who played a major role in making the MPM implementation a success.

References

1. Robles J. "Application of Advanced Heavy-Oil-Production Technologies in the Orinoco Heavy-Oil-Belt, Venezuela" SPE 69848, 2001 SPE lntemational Thennal Operations and Heavy Oil Symposium Porlamar, Margarita Island, Venezuela 12-14 March 2000.

2. Mehdizadeh P., "Multiphase Meters: Delivering Improved Production Measurements and Well Testing Today", Petroleum Engineering Intemational, May 1998.

3. Falcone, G. et aI, "Multi phase Flow Metering- Current Trends and Future Applications", SPE paper 71474 presented at the 2001 SPE-A TCE in New Orleans, Louisiana, Sept 30 - Oct. 3, 2001.

4. Colmenars, J. et.al."Multiphase Flow Metering Heavy Crude Oil Field Evaluation" 7th Intemational conference Multiphase 95, June 7-9, 1995: Cannes, France.

5. Stokes, Edward G., et aI, "Application of The First Multiphase Flowmeter in The Gulf of Mexico", SPE 49118, Annual Technology Conference and Exhibition, September 27-30, 1998, New Orleans, LA.

6. Mehdizadeh P., and Farchy D., "Multiphase Flow Metering Using Dissimilar Flow Sensors - Theory and Field trials Results", SPE 29847 presented at the SPE Middle East Oil Show, Bahrain, 1995.