Larry J. St. Germain
Louisiana Offshore Oil Port Inc.
Galliano, La.
Early in its operations, Louisiana Offshore Oil Port (LOOP) chose turbine meters for flow measurement in conjunction with mechanical displacement provers to provide the accuracy, resolution, and reliability to meet all regulatory and operational requirements.
LOOP's experience suggests basic design criteria for turbine meters and indicates factors that affect their performance in different crude types.
DEEPWATER FACILITY
LOOP is the only deepwater port in the U.S. for the importation of foreign crude oil. The facility is located in the Gulf of Mexico in Grand Isle Block 59, approximately 18 miles offshore.
It was designed specifically for unloading very large crude carriers (VLCCS) and ultra-large crude carriers (ULCCS) up to 700,000 dwt via single anchor-leg mooring (SALM) systems. The facility has a pumping platform complex submarine pipelines, a booster station, and a 48-in. pipeline interconnected to eight underground storage cavities each capable of a 4.5-5 million bbl capacity with a total combined capacity of over 40 million bbl (Fig. 1).
In LOOP's relatively short operational history of approximately 9 years, the port has handled 55 different crude types. These crudes range from Mexican Maya with a density of 0.9201 kg/I. and an API gravity of 22.2 to Algerian condensate with a density of 0.7269 kg/I. and an API gravity of 63.1.
From the onshore terminal, pipelines distribute these crude types to various parts of Louisiana, Texas, and the Midwest (Fig. 2).
High-accuracy flow measurement by turbine meter is a key feature of the operation of LOOP because of stringent line surveillance and custody-transfer requirements. The measurement system was designed to operate within the ranges of 318 cu m/hr (2,000 bbl/hr) to 15,900 cu m/hr (100,000 bbl/hr) with a very high degree of accuracy, resolution, and reliability.
The metering systems are installed at the inlet and outlet of the pipeline with on line provers to allow proving at any time during unloading operations. During the calendar year of 1989, LOOP received into the pipeline 49,1 72,050 cu m (309,422,058 bbl) and received on the other end of the pipeline 49,177,546 cu m (309,456,642 bbl). The difference was a 5,496 cu m (34,584 bbl) overage.
This translates into a difference of +0.0112% for the year. Such high accuracy is attained only by frequent meter provings.
RELIABILITY
In recent years, turbine meters have been competing successfully with positive displacement meters on many applications as a result of the economy of installation, low maintenance, and higher flow rates per comparable size.
Since the majority of turbine meters utilize tungsten carbide bearings and there is only one moving part in the flowing stream, little or no wear occurs after years of service (Fig. 3).
In LOOP's application of turbine meters, (eighteen 16 in., twelve 10 in., four 8 in., and four 2 in.), only three mechanical failures have occurred in approximately 9 years of operation and 540 million cu m (3.4 billion bbl) of measured volumes.
Stringent line surveillance and custody-transfer requirements mean that LOOP's metering has to be of the highest available accuracy and reliability.
LOOP decided early on turbine meters for flow measurement in conjunction with mechanical displacement provers. This combination would provide the needed accuracy, resolution, and reliability to meet all regulatory and operational requirements.
The turbine meter is not a new flow measurement device. Records indicate that in 1886 the first patent for a turbine meter was issued. In 1914 a patent was granted for a turbine meter which related flow rate to frequency.
But it was not until the 1950s that the turbine meter was developed into a precise flow-measurement device.
The development of jet engines and liquid propellant rockets fostered the need for an accurate, fast responding meter that could be used on exotic fuels and oxidizers at extreme temperatures.
The turbine meter met this need. It was soon applied to many other industrial flow measuring applications.
Turbine meters began to be applied extensively in the petroleum industry in the mid-1960s. Since publication of API Standard 2534 "Measurement of Liquid Hydrocarbon by Turbine Meter Systems" in March 1970, the turbine meter has gained broad acceptance for custody transfer of petroleum liquids such as LPG, light distillates, and light crude oils, primarily at large petroleum storage and transfer terminals.
TYPICAL LOOP OPERATION
In a standard LOOP application, the product enters a turbine meter (Fig. 3) complete with stainless-steel internals and two pickup coils with signal preamplifiers. The turbine meter housing is made of nonmagnetic stainless steel with raised face flanges.
Pulse generation occurs through the wall of the meter housing. The rotor contains the optimum number of blades properly contoured to provide the desired torque and speed.
High resolution is accomplished by a series of magnetic points on the rim of the rotor. The rotor is hydraulically balanced so there is no axial thrust.
The raw pulses from the turbine meter are sent to a factoring and pulse comparator totalizer. Pulses are compared from Probe A and Probe B to determine if there is any transmission failure or meter preamplifier failure.
The pulse from Probe A is conditioned and directed to an Anadex meter counter. This in turn is read by the computer for operating and ticketing purposes.
The pulse from Probe A is divided by the reciprocal of the pulse per cubic meter multiplied by 1,000. The result is displayed on the front of the factoring and pulse comparator and registered in pulses per cubic meter.
It is then sent to a combinator in the proportion of 1,000 pulses/cu m.
CALIBRATION CHECKS
The following parameters are vital for accurate measurement and should be routinely checked. Any time a device is not operating within the required limits, measurement becomes questionable.
Under most conditions any time a device is operating out of tolerance, the meter should be taken out of service if it cannot be corrected within a short period of time.
- Temperature. Proper temperature detection is extremely important. An error of 0.1- C. (0.18- F.) during measurement of 100,000 cu m (629,800 bbl) will introduce an error of 8 cu m (50 bbl),
All temperature transmitters in a flowing stream should be checked daily.
The device being used as the reference must have been checked for accuracy against an NIST (National Institute of Standards & Technology) standard. All temperature transmitters in measurement service should be maintained with an accuracy of 0 - 1 C. ( 0. 18 F.).
- Pressure. Pressure is critical to measurement but not as much as temperature. An error in pressure of 0.1 bar (1.5 psi) is equal to 0.75 cu m (4.7 bbl) in each 100,000 cu m (629,800 bbl) measured.
All pressure transmitters in a flowing stream should be checked for accuracy at least daily. The gauge used for checking the pressure transmitters must be calibrated with a dead-weight tester to verify the accuracy of the transmitter.
All pressure transmitters should be calibrated and maintained within an accuracy of 0.1 bar ( 1.5 psi).
- Density. A densitometer operating with a 0.1% error will produce a volume change of approximately 2 cu m (12.5 bbl) in each 100,000 cu m (629,000 bbl) measured. All densitometers in a flowing stream should be checked for accuracy daily.
The density is checked by taking manual API gravities and converting the observed gravity and temperature to density at the operating temperature for comparison to densitometer readings. The densitometer should be calibrated and maintained within an accuracy of 0.1% of the operating range.
- High integrity valves. A program of inspection of valves for proper operation and proper reporting of a seal to the programmable logic controller (PLC) and computer is done on every prove.
If a high integrity valve does not perform properly under the described conditions, the computer will abort the prove. At that time the valve is manually inspected to determine the problem.
If the valve is found not to have a seal, the meter run it represents is immediately taken out of service. A monthly visual inspection and test procedure is performed to verify that all components reporting a seal failure to the PLC are operating satisfactorily.
- Strainers/straightening vanes. Trash in a strainer, straightening vane, or meter will affect the meter operation and the factor.
Differential pressure is measured on all strainers on a constant basis with reports to the supervisory system. All strainers are inspected on a 6-month basis, and if a strainer is found to have an unusual amount of trash, the straightening vanes are rolled out for inspection of the inlet of the straightening vanes and meter.
PROVERS
Meter-proving systems have improved to the extent that pipe provers reduce the expense and difficulty of proving meters of large capacity and make more practical the application of these meters in high-volume pipeline, tanker, and barge metering operations.
These improvements also apply equally well to smaller meters and provers.
Pipe provers for measurement-reference vessels are widely used and may be straight or folded in the form of a loop. Both portable and stationary provers may be constructed on these principles.
Pipe provers have also been developed for pipeline use in which a calibrated portion of the pipeline itself, either straight, U-shaped, or folded, serves as the reference volume, Some provers are arranged so that liquid can be displaced in either direction, via the reciprocating or bi-directional types.
The chief advantage of a pipe prover over a tank prover is that during proving the flow of the liquid is not interrupted. This permits the meter to be proved at a uniform flow rate without having to start and stop. Rates can thus be very high if needed.
The reference volume required of a pipe prover, that is, the volume needed between detectors, depends in part upon the discrimination of the proving register, the reproducibility of the detectors, and the repeatability required of the proving system as a whole.
The relationship between meter rating and reference volume must also be taken into account. Provers of smaller volume than was once considered necessary can now be used because of the application of high-speed pulse generators on the meters, the use of precision displacer detectors, and the elimination of the inevitable errors in a start-stop method of proving.
PROVING FREQUENCY
The frequency of proving a particular turbine-meter system depends on so many aspects of its operating conditions that it is difficult to establish a fixed time or throughput interval for reproving.
LOOP has established certain reproving parameters which under normal conditions will provide good metering. But there are some conditions in which a meter cannot be proved often enough.
- Startup. After startup, all meters should be proved as soon as possible after conditions stabilize.
- Flow. All active meters should be reproved any time the flow rate changes - 300 cu m/hr (2,000 bbl/hr) or 10% of the previous flow rate.
On crude with a density of 0.8800 kg/I. or greater (29 API or lower), the active meters should be reproved when flow rate changes by 5% or more of the previous flow rate.
- Temperature. All active meters should be reproved whenever the temperature changes -- 2.0 C. (3.6 F.).
- Pressure. All active meters should be reproved whenever the pressure changes 3.5 bar (50 psi).
- Density. All active meters should be reproved whenever the density changes 0.0100 kg/I. (2 API).
- Viscosity. All active meters should be reproved whenever there is an observed change in viscosity of 20%.
- Duration. Time should be considered when conditions are unchanged since the last prove. This is recommended because foreign debris may become hung on a strainer, straightening vane, or meter and cause the characteristics of the meter to change, thus causing a factor change.
If this condition exists, a time interval of 8-10 hr for reproving should be considered.
All proving is automatic based on an operator-initiated command on these changes. All proving is preceded by a ticket which is cut on each meter.
The meter is proved at the new conditions and the new factor is rolled into the new volume.
PERFORMANCE EVALUATION
Measurement-system control charts can be used as a warning signal to show the extent to which conditions may have deviated from the acceptable norms.
The charts can be used to indicate trouble but not necessarily the nature of the trouble.
Historical meter-factor data are stored in LOOP's personal computer up to a maximum of 6,000 of the latest meter factors (up to 750 on each meter in the system.) These data are used for comparisons, on a daily basis, of any drifts or changes in the performance of the meters over a period of time without having to research files.
When measurement trouble is encountered, a systematic check of the measurement system is recommended. The following items should be checked but not necessarily in the order of listing:
- All valves affecting meter proving
- Strainers and flow conditioners
- Pulse counters, preamplifiers, signal-transmission system, power supply, pickup coils, and readout devices
- Moving parts, bearing surfaces of the turbine meter
- Other parts of the meter and meter run
- Detector switches in the prover
- Displacer in the prover
- Pressure, temperature, density-sensing devices
- Operation of the meterproving system at other than normal operating conditions.
METER CURVES, LINEARITY
Meter-characteristic curves can be developed from a set of proving results which will show, in graphic form, the relationship between a number of dependent and independent variables. These curves serve as measures of "operating conditions" and are sometimes called linearity curves.
Under the same conditions of proving, a meter should produce a factor within the range of linearity as established in the characteristic curves. If not in compliance, the cause may be attributed to reasons previously listed.
Fig. 4 is an example of two such characteristic curves for a 16-in. turbine meter on a high viscosity crude and a low viscosity crude.
When a meter is proved and the factor does not agree with the linearity of the characteristic curves, the reason for this deviation should be determined.
Of the many reasons that could cause a change, some of the most prevalent are the following:
- A change in temperature that will change viscosity, especially in higher density crudes
- Trash in a strainer, or on a straightening vane, or hung on a meter rotor
- Leaking valves in proving lineup, includes four-way valve
- Prover displacer
- Pressure, temperature, or density detection devices
- Prover detector switch malfunction
- Failure of pulse generator, or pulse-counting equipment.
It is important to note that even though many crudes follow similar linearity curves, some crudes do not follow the rules.
DECREASING METER FACTORS
In one experience on a 16-in. turbine meter at LOOP, foreign debris hanging on a rotor caused a shift in meter factor.
This shift was caused by a piece of neoprene hose, approximately 11 in. long, which wrapped itself on a blade of the rotor. This caused the turbine to pick up speed, which in turn produced an increase in pulse counts, causing a decrease in meter factor of approximately 0.0065.
It is important to emphasize that at those conditions, the factor is correct. A turbine meter for which the factor has changed can be used successfully in measurement, but it must be closely monitored because of the possibility of another shift in factor.
When a shift in a meter factor is encountered, the meter should be rechecked at least every 2 hr by running two round trip runs to verify that the pulse count has not changed.
If a decrease in pulse counts is witnessed, the meter should be reproved as soon as possible. The previous factor should be applied to the gross volume that it represented.
LOOP has used the following procedures with some success in an attempt to dislodge foreign debris in a flowing stream. Removing the meter from the flowing stream sometimes enables debris hung on the strainer basket to fall to the bottom of the rotor.
Flowing through the meter at a higher flow rate than normal will sometimes cause the debris to be dislodged from the rotor. In one such instance at LOOP, a 48-in. pipeline sphere disintegrated in a pig trap and a triangular shaped section tore through the strainer basket.
The section dislodged the straightening-vane section for a 16-in. turbine meter run. The straightening section along with the pig section came to rest against the upsteam stator of the turbine meter. As a result of the unavailability of heavy equipment needed to roll the spool section and meter for removal of the straightening vanes, the condition persisted for several hours.
Before the condition occurred, the meter factor was 1.0175 on Alaskan North Slope crude with a density of 0.8944 kg/I. and an API gravity of 26.7. The factor fell 0.0052 to a 1.0123 with the pig and straightening section still remaining against the upstream stator.
The pig section having lodged against the vanes caused an area in the straightening vanes to be void of flow which in turn caused a dead zone striking the rotor. This condition caused the rotor to pick up speed which caused an increase in pulse counts and a decrease in factor.
The meter run was drained and the pig section removed. The meter was placed back in service with only the vanes remaining against the upstream stator. The meter factor rose 0.0027 to 1.0150.
After the straightening vanes were replaced, the meter was placed back in service and proved. The meter returned to its original factor of 1.0175.
Fig. 5 demonstrates the effect on a meter characteristic curve as a result of a four-way valve seal leak. Note the dramatic effect on the meter factor by the seals' not sealing, causing crude to bypass to the outlet side of the four-way valve.
A properly seated valve would indicate a decrease in valve-cavity pressure. A seal leak would indicate an equalization to line pressure in the cavity of the valve.
The drop in factor is attributed to a portion of the crude by-passing to the outlet side of the four-way. This in turn causes the prover displacer to travel at a slower rate of speed between detector switches causing an increase in pulses which in turn caused a decrease in meter factor.
INCREASING METER FACTORS
Some of the instances LOOP has experienced with higher meter factors have been attributed to a prover sphere deflating and picking up speed between detector switches.
An instance of a defective tungsten washer caused drag on a rotor, decreasing the pulses, which caused a shift to a higher factor.
In both instances the defective part was replaced and the meter factor returned to its original characteristic curve.
Copyright 1990 Oil & Gas Journal. All Rights Reserved.