Custody Transfer: The Value Of Good Measurement And The Search For The Truth

July 2009 Vol. 236 No. 7

Daniel J. Rudroff, Chief Operating Officer, Welker Flow Measurement Systems, Sugar Land, TX

Custody transfer measurement in the oil and gas business has been described many ways. It has been called an accuracy in measurement that both the buyer and seller can agree upon and it has been called the best that can be achieved to meet the contract conditions.

I like to call it, “The Search for the Truth.” Ever since petroleum has been bought and sold, better ways to measure and better accuracies have been sought. A big advancement was the pipe prover. Today, the American Petroleum Institute (API) requires an accuracy of 0.02% when compared to a standard such as National Institute of Standards and Technology (NIST) traceable Seraphin cans. If we want to put 0.02% accuracy into perspective, that is 6.45 teaspoons, or a little more than two tablespoons, of oil in one barrel. That is very good measurement and that is worst case.

Converted to dollars on a small 10,000 barrels per day (bpd) Custody Transfer Liquid Metering Skid, a 0.02% error is about two bpd. At the average price of oil today, say $50.00/bbl, that is $100 dollars per day. In a year’s time, that is $36,500. That is why we all strive to exceed the 0.02% required by API. We know and understand the value that increased accuracy delivers to our companies.

Accuracy of measurement is important when oil is selling at $100/bbl and profits are good, but it is even more important when oil is at $50/bbl and the margins are tight. One lost barrel becomes a much larger percentage of the profit.

Welker Flow Measurement Systems (WFMS) has a patent pending on a new design bi-directional prover called the WFMS SCS Prover™. The SCS stands for straight calibrated section prover. With the SCS Prover, accuracies of 0.002% to 0.007% have been achieved regularly. The calibrated section is in one straight piece of pipe which eliminates the elbows and flanges in the calibrated section, the cause of most of the inaccuracies and problems in other bi-directional provers.

What Is Accuracy?

Accuracy is what we are after when we zero in on a target. Any good meter like any good rifle is repeatable. If every shot is within a small group, but not at the center of the target, we make an adjustment to the sights or to the scope to bring the pattern to the center of the target. Now our rifle is repeatable and accurate. It is the same with meters. The more repeatable they are, the more accurate we can make them.

A precise meter that is repeatable also needs to be zeroed in for accuracy. For this we use a “pipe prover.” This is a device whose volume between two switches has been checked and verified to a known volume. The length of pipe between the detector switches is determined by the volume needed to ensure good repeatability. A sphere is then inflated to a diameter larger than the inside of the pipe and placed in the prover. The two switches installed in the pipe detect when the ball reaches the beginning and end of the known volume in the pipe.

Normally, the meter prover volume between the switches is calibrated against a device called a Seraphin can whose precise volume is traceable to NIST. Before we can put the meter prover in service, it must be calibrated against a known volume such as these Seraphin cans. The prover must also conform with the API Manual of Petroleum Measurement Standards, Chapter 4 – Proving Systems, Section 2. The volume must match the calibrated Seraphin can’s volume three consecutive times at different flow rates to within 0.02% (0.0002).

Both the volume in the Seraphin cans and the prover are corrected for the effects of pressure and temperature on the fluid, usually water, and the temperature and pressure effect on the pipe and metal of the Seraphin tanks. Only when the prover passes this calibration can the prover – whose volume is now known – be used to prove the volume of a meter on site. Improving on the 0.02% required repeatability when compared to the Seraphin cans is always the goal of the pipe prover manufacturer. An error of about two tablespoons per barrel is then the maximum error API allows per barrel between the volumes of each of the three Seraphin tank calibration runs on the prover calibrated section.

Because there are many variables in a metering system, such as the way the meter is installed and the effect the fluid has on the meter, it is the meter system – including all the piping and components such as valves, flow conditioners, the location of the pressure and temperature transmitters and how the calculations are done – that affect the meter performance. The prover whose volume is known is used to verify the performance of the entire metering system.

There are several types of “provers.” They are bi-directional, uni-directional, bi-directional piston and small volume piston provers. All of these provers do an excellent job of proving meters. But in today’s market, companies are looking for a way to improve the performance of the prover and therefore the meter and meter system.

To indicate volume, all meters produce pulses in proportion to the amount of fluid that passes through the meter. Each volume, such as a barrel or a cubic meter, is indicated by a number of pulses. The number of pulses per unit volume is determined by the meter manufacturer. The amount of pulses produced by the meter, per barrel or cubic meter, as fluid passes through the meter is known as the “meter factor.”

Each type of meter – turbine, positive displacement, coriolis and ultrasonic – has unique characteristics that must be understood and addressed when proving. The turbine meter – because of the inertia of the rotor – does not immediately change its output with a change in flow. The positive displacement meter gear stack is not a direct couple and the distance between the pulses can vary as the flow rate changes even slightly. The coriolis and ultrasonic meters produce “manufactured” pulses. The method of measurement used by the coriolis meter or the ultrasonic meter is converted to pulses for proving. Because the number of pulses needs to be calculated, there is a slight time delay.

In the conventional “U” or “Serpentine” shape of the bi-directional or unidirectional prover there are elbows, welds and flanges within the calibrated section. Because the elbow has a smaller radius on the inside of the bend and a larger radius on the outside, and – because the elbow may not be perfectly round through its bend – the prover ball velocity will change as it passes through these elbows and some volume of fluid can become lost in the elbow’s imperfections. The changes in velocity change the flow rate and the imperfections in the elbows and flanges affect the volume and, therefore, the prover’s accuracy.

To counter the effects of imperfect elbows, the prover sphere is overinflated to assist its sealing against the pipe wall. However, overinflation of the ball increases the pressure drop across the ball and causes a greater change in velocity and, therefore, volume as the ball passes through elbows. Welds and flanges are also a problem. The inside diameter (ID) of the pipe and the ID of the flange are not always the same. If the flanges are undersize the ball will slow down and an oversize flange ID will cause the ball to speed up. The same holds true for welds in the calibrated section. The welds that are overly ground or not ground down sufficiently cause a velocity change in the calibrated section.

Crude oil meter proving systems in the U.S. are controlled by The API Manual of Petroleum Measurement Standards, Chapter 4 – Proving Systems, Section 2. It is recommended in this standard that a minimum of 10,000 pulses be accumulated between the detector switches. If 10,000 pulses cannot be accumulated between the switches, a method called pulse interpolation can be used. The reason 10,000 pulses were chosen is because of the error that occurs between when the prover’s detector switches activate and the meter produces a pulse. The prover ball detector switches – when they activate – fall somewhere between the pulses produced by the meter, and volume between the pulse and when the detector switch activates is the error.

Pulse interpolation is a computer-based method that reduces this error by averaging the time between the pulses produced by the meter or by adding a high-frequency pulse between the meter pulses. If the time between the pulses from the meter is known, the volume between the pulses can be determined, and the volume from when the prover detector switch activates to the first pulse can be determined. Likewise, the volume from the last pulse to when the second detector is activated can be determined. By using pulse interpolation. the meter-indicated volume between the ball detector switches can be determined very accurately

The SCS prover design developed by WFMS eliminates the problems created by elbows and flanges. There are no elbows or flanges in the calibrated section of the prover. It is a bi-directional spherical prover that uses a straight run of calibrated pipe between the two detection switches. Unlike the prior provers, there are no flanges, welds or elbows between the detection switches. The new prover has numerous advantages over conventional bi-directional spherical provers which use a U-shaped calibrated section of pipe or a serpentine-calibrated section of pipe which includes several elbows and flanges. These elbows, which need to be of the best quality, and the machined flanges that are required, are expensive.

Many experiments were conducted by WFMS on the prover. The velocity of the ball in the launching chamber was determined. Different size balls were used to demonstrate different ball-to-launcher relationships. The end of the 36-inch prover was removed and the ball was moved through the calibrated section at one foot per minute using water pressure. The contact of the ball with the wall was calculated at 10 inches at 3% overinflation of the prover sphere. There was zero leakage around the ball as it passed the entire distance between the detectors. Some advantages of the SCS prover are:

  1. There are no alignment flanges in the calibrated Section. Alignment flanges are expensive and machining on the flange or installing pins reduce the integrity of the alignment flange and therefore the piping system. Flanges in the pre-run are aligned using shoulder bolts that are approximately the same diameter as the flange bolt holes.
  2. There are no elbows in the calibrated Section. Elbows cause a pressure and flow change as the ball moves through the elbow and there can be loss of fluid if the elbow is not perfectly formed in the inside diameter. Because the ball does not have to go through elbows as it passes from switch to switch less inflation of the ball is required, making for better water draws and better proves with less pressure drop.
  3. The calibrated section can be rolled out and inspected without another water draw because no flanges are broken in the calibrated section. This is a cost savings both in time and water draw cost.
  4. Installing bells at the launchers which increases the pipe size into and out of the launchers from the four-way reduces the velocity at that point and lowers pressure drop and damage to the ball as it is pulled in front of the grating on conventional provers.
  5. It is ideal for coriolis and ultrasonic liquid meters with manufactured pulses because the flow before and in the calibrated section is not disrupted by the ball passing through elbows, welds or flange sets. Because the ball can be inflated less, it passes smoothly between the detectors, not disturbing the flow.
  6. Since the flow through the calibrated section is smooth, the pulses from conventional PD and turbines will be more evenly spaced, giving better proves especially when pulse interpolation is used.
  7. The cost is lower because special alignment flanges are not required.
  8. The ball does not have to be overinflated to compensate for irregularities in elbows and flanges
  9. API requires the repeatability of the prover volume when compared to the NIST-calibrated Seraphin tanks to be within 0.02%. Water draws on the SCS prover have come in from 0.002% to 0.007%, an order of magnitude better than that required by API.

There are several SCS provers in operation from 16 to 36 inches and all are performing better than anticipated. Some are on multiviscosity meters and some are on coriolis meters. Both meters produce manufactured pulses.

A new sphere removal tool arrangement has been built to allow a constant vacuum to be held on the ball as it is being removed from the launcher chambers.


Daniel J. Rudroff can be reached at 281-491-2445, E-Mail:,

Find articles with similar topics