Content of PetroWiki is intended for personal use only and to supplement, not replace, engineering judgment. SPE disclaims any and all liability for your use of such content. More information
Difference between revisions of "Continuous and fullbore spinner flowmeters"
|Line 115:||Line 115:|
Revision as of 16:41, 15 April 2015
There is no generic difference between a "continuous" spinner and a "fullbore" spinner. In the case of the fullbore, the spinner element folds into a diameter no greater than that of the tool when in the tubing, but expands into a larger diameter for surveying in the casing. The continuous spinner does not have this capability. The difference between the two is too small to justify a separate discussion of each.
The continuous meter derives its name from the need to move the tool fast enough to overcome frictional torque and start the spinner element rotating. It also derives its name from the in-situ calibration procedure that uses logging runs at several different cable speeds with the well shut-in at the surface. Neither continuous nor fullbore, however, can provide a log that is quantitative whenever the fluid velocity is sporadic, that is, changing in the logged interval.
The continuous and fullbore tools share three features:
- First, the spinner element on each is at the very bottom of the tool string. If the temperature tool is run in combination with the spinner, then the thermometer’s sensor will usually be located in the tool string above the spinner. Thus, the sensor will be 4 to 6 ft above the bottom of the string. Moreover, this 4 to 6 ft will include at least one centralizer. The mixing caused by the passage of the centralizer in front of the temperature sensor will decrease the vertical resolution of the temperature log by a few ft. Even so, the vertical resolution of the temperature tool to localized entries is still far superior to that of the spinner tool, and the temperature tool should be included in the tool string.
- The second common feature is the presence of at least two centralizers in any tool string containing the spinner flowmeter. The centralizers ensure that the spinner element samples the same location in the wellbore’s cross section at each depth. This consistency is necessary for the same tool calibration to apply at each depth and for the relative profile to be representative of flow rates.
- The third shared feature is that both tools use a four-blade propeller-type (or turbine-type) spinner element. Although the design of the spinner element may vary, the four-blade feature is retained.
The spinner element can rotate either clockwise or counterclockwise (as viewed down the tool barrel). The direction of rotation depends upon the movement of the fluid relative to the barrel of the tool, that is, upon the direction of fluid movement as seen by "rider" on the tool. Usually, the pitch of the spinner is such that relative movement of fluid up the barrel causes the spinner to turn clockwise, whereas relative movement of fluid down the barrel causes a counterclockwise rotation of the spinner. Consequently, movement of the tool downward in stagnant fluid causes relative movement of fluid up the barrel and rotates the spinner clockwise. Movement upward in a stagnant fluid causes relative fluid movement down the tool barrel and spins the element in a counterclockwise direction. Movement of the tool in a direction opposite to the direction of flow causes the spinner to turn in the sam direction, clockwise or counterclockwise, throughout the logged interval including those intervals with stagnant fluid. Passes made downward in a production well therefore cause a clockwise rotation over the interval from 100% to 0% flow. Passes made upward in an injection well cause a counterclockwise rotation. On the other hand, movement of the tool in the direction of flow at a line speed less than full-steam velocity causes a reversal in the direction of spinner rotation at some depth within the flow interval. The spinner first loses speed as flow velocity approaches tool speed. It then stops when fluid velocity reaches tool speed less the frictional threshold speed for the tool. The spinner remains stopped until fluid velocity changes by an amount that equals two threshold values, at which point the spinner begins to turn again, but in a direction opposite the previous one. As fluid velocity changes further, the spinner maintains its opposite rotational direction. Such a pass should not be used for percentage flow profiling, because two threshold values are "lost" in the record of spin rate.
Many spinners, however, do not record the direction of rotation. Even worse, some spinners have lower sensitivities when rotating counterclockwise. In event the direction of rotation changes and the spinner does not record it, the change can be recognized from its signature: a drop of the revolutions per second (RPS) to zero, followed by a resumption of the RPS to above-zero values.
Application and interpretation
Both continuous and fullbore tools are intended for quantitative use only in flow streams having a single component of velocity directed along the axis of the tool. Furthermore, the stream should be a single-phase or a high-rate multiphase (10,000 B/D or more). In the multiphase case, the well should have low deviation angle. The meters are designed to function quantitatively in environments such as that indicated in Fig. 1. This spinner pass downward is from a cased, vertical well producing single-phase gas from three isolated perforation sets, the deepest of which is a short interval at 11,700 ft. In the left track, the cable speed trace shows a logging speed of 56 ft/min. This speed produces the 1.2-RPS rate of spin below the deepest perforations. This part of the RPS trace is labeled "zero-flow reference" because the RPS response is caused by moving the tool downward through the stagnant fluid below the deepest set.
The completion in the figure has at least 50 ft of separation between successive perforation intervals. This is sufficient distance for any tangential velocity components associated with an entry at a given interval to die out before additional gas enters at the next interval above. Thus, the spinner attains a constant rotational rate above a given entry that reflects the axial velocity produced by the combined flow from that entry and all others below. In the figure, this constant rate is attained within approximately 10 ft above the entry. The amount of deflection to the right of the no-flow reference is proportional to the flow rate in the wellbore at the respective depth. Under these conditions, flow profiling is done simply from a determination of the fraction that a given stable deflection represents of the total stable deflection above all entries. Thus, the top set contributes approximately two-thirds of the total gas while the bottom set contributes only approximately 15% of the total.
The survey of the figure appears to show simple and direct a spinner record is to interpret. The validity of the interpretation, however, cannot be judged on the basis of the spinner log itself. Most wells that produce only oil or gas will have a stagnant column of mud or workover fluid standing to some depth in the wellbore unless the deepest entry is at a high rate. A wellbore-fluid density log for the present example could show a stagnant water column standing in the wellbore, at least to the bottom of the middle perforation set. If that is the case, then the spinner record above the bottom set is in response to lifting of water by the buoyant rise of gas through it and not a response to the single-phase velocity of gas flow upward. Then, if the response is assumed to be from single-phase gas flow, the contribution from the bottom set will be greatly exaggerated.
Furthermore, the type of completion in Fig. 1 is seldom encountered. The type of completion more likely to be associated with spinner logs is shown in Fig. 2. Here, there is a long bottom interval perforated over 65 ft and a short perforated interval only 5 ft above the bottom one. The figure shows a fullbore-spinner log on the right track and a wellbore-fluid density log on the left track. Both logs were run downward at a cable speed of 200 ft/min with the well producing at a combined rate of 10,000 B/D of oil and water at a 50% water cut. The objective of the logs was to profile the relative oil and water production within the long interval and to determine the relative contributions of each from the short interval.
Below the bottom of the long interval, the wellbore-fluid density is 1.0 g/cm3, identifying water. Above the bottom of the long interval, the fluid density decreases up to approximately Depth D because of the entry of oil. The increasing presence of oil in the water shows that the bottom interval is productive up to D. Above D, there is very little change of the fluid density, suggesting that the upper part of the long interval and the short interval are not significantly productive.
Below D, the RPS trace in the 7-in. casing is ever-changing. The trace never attains a stable RPS value, which is necessary for quantitative analysis of the flow velocity. This fact alone means that the production from the long part of the bottom interval cannot be profiled in the straightforward manner of Fig. 1. Possibly, the absence of a stable RPS response is because of the high logging speed, which raises the rotational inertia of the spinner element. Quantitative analysis also requires that the flow velocity be entirely axial. But such is not the case in this example. At Depth A, there is a jetting entry, the tangential velocity of which causes a spiked increase in the RPS response. The fluid-density log at this depth decreases for a short interval, identifying oil. At Depth C, the fluid-density trace spikes in response to a fluid jet that appears to put water between the sensor ports spaced 2 ft apart. If this is the case, then the rate is insignificant because the density is the same on either side of the spike. At D, there is a spiked increase in the RPS response resulting from the tangential velocity of a jet entry. This is not to say that all perforation jets cause an increase of the RPS response. In other cases, a jet entry may have a tangential velocity that decreases the RPS trace. This happens at 7,211 ft, for example.
The density trace shows that essentially all the oil has entered the wellbore by a depth of 7,180 ft. The spinner shows that essentially all the flow is in the wellbore by this depth. Consequently, both oil and water enter below this depth. The reader can apply similar reasoning to conclude that approximately 70% of the flow enters below 7,200 ft, bringing with it a major part of the oil and water production.
Above D, the RPS trace slowly diminishes. This means that the flow at D has a swirling (tangential) component that dies away as the flow moves up. Swirls can require several hundred feet of travel to decay completely. At D, the swirl is contributing to the RPS response; as the swirl dies away, the contribution diminishes and the RPS decreases. In other cases, the swirl may detract from the RPS response; as the swirl decays, the RPS increases. Even if the RPS trace were to show some evidence of production from the short interval, the presence of the swirl would preclude a quantitative analysis of the flow velocity.
The logging operator used the high logging speed in this example to minimize the distortion of the RPS trace by jet entries, as at Depth A. Actually, this procedure maximizes the distortion by biasing all fluctuations to the high side of their excursions rather than to their average.
The previous two examples should remind the reader that an apparently simple record from a direct measurement can have subtle meanings and may fail to present a complete accounting of the situation.
Continuous tools are available in a wide range of configurations, more so than the fullbore tools. The diameters range from 1 3/8-in. to 2 1/8-in. The 1 3/8-in. versions, with centralizers, should pass through 2-in. tubing. If the tubing includes landing nipples, such as the 1.82-in. size, it is difficult to force the centralizers through the nipples. Safety considerations preclude even an attempt.
The centralizers also come in a variety of configurations. Powered centralizers offer the least problem to entry through tubing. These are closed by strong springs when in the tubing. In the casing, a downhole motor opens the centralizer against the spring force. In event of a failure of the motor downhole, the powered centralizer has a shearing mechanism so that the constriction at the tubing’s end can be used to close the centralizer when re-entering the tubing. The chance of sticking a centralized string is greater than that for an uncentralized one.
Some continuous tools have the spinner element inside a bow-spring cage with no additional protection from damage. In others, the element is inside a rigid cage having the same diameter as the tool. Still others place the element inside a short section of tubing having the same diameter as the barrel. The latter are immune to the tangential velocity of a jet-entry; they are not immune to swirl, which has both axial and tangential components. If the flow is multiphase, the trace from such a tool is noisier because of the light phase’s tendency to pass through the "chimney."
To measure the RPS of the spinner element, the most common means is a magnet and pickup coil. A narrow magnet is attached lengthwise to a section of the spinner shaft. The magnet rotates under the pickup coil, which is divided into independent segments so that the coil generates a number of inductive current spikes per revolution of the shaft. The resolution, however, does not approach that from tools utilizing an optical sensing assembly, which also detects the direction of rotation. Another variation uses a single sector of pickup coil with three bar magnets embedded in the rotor at unequal azimuth angles. While reducing the tool’s resolution, this approach does detect the direction of rotation.
Magnets should be located inside the tool barrel so that iron particles in the wellbore fluid attach themselves to the outside of the tool and do not interfere with rotation of the spinner. Magnets located on the spinner shaft before it enters the barrel can attract iron particles and eventually make rotation impossible.
Most continuous spinners are rated for pressures in the range of 15,000 to 20,000 psi and temperatures of 350 to 400°F. Some tools can accommodate 500°F, but they employ vacuum flasks and thus have a diameter of at least 2.5 in.
Practically all fullbore spinners are copies of the original Schlumberger version with slight modifications here and there. Fullbore tool diameters range from 1½ to 1∕1611 in.
These tools use specific diameter spinner elements to accommodate specific casing sizes. Typically, three different diameters are used to cover casings in the range of 4 to 9 5 / 8 in. Pressure and temperature ratings are the same as for the continuous spinners, which were stated previously.
The difference in resolution between continuous and fullbore spinners is small; the following comments apply to both types. Neither is very effective for quantitative resolution of low-rate, multiphase production. In the U.S., the spinners’ greatest quantitative application is injection profiling. Table 1 summarizes typical rate resolution under different flow conditions in vertical wellbores, although actual resolutions are quite dependent on flow conditions.
These numbers should be viewed in the sense that the given amount of flow is lost, even qualitatively, for entries above the deepest entry. The qualitative resolution of the deepest entry is better than the numbers in the table.
Refer to Fig. 3, which pertains to the record of a high-quality, high-resolution continuous tool that detects the spinner element’s direction of rotation. The well, with three perforation sets, produces 1,800 B/D at 28% oil, 72% water, and no gas. The survey shows two logging runs, one up and one down, with each at 22 ft/min cable speed (left track). There are two respective RPS records (right track). Zero RPS is at the fifth chart division from the left, and 1 RPS is spread over four chart divisions (a sensitivity of 0.25 RPS per chart division), showing the tool’s high resolution.
On the up run (dashed trace) in the stagnant fluid below the bottom perforation set, an imaginary observer riding on the tool would perceive the fluid velocity as down the barrel so that this trace records counterclockwise rotation (CCW) below the bottom set. On the down run through the stagnant fluid below the bottom set, the imaginary observer would perceive the fluid velocity as up the barrel; therefore, the rotation is clockwise (CW). Below the bottom set, the RPS of the up run is more irregular than on the down run, because the tool, when jerked off the bottom, requires some distance to reach a steady speed. Thus, the comments below pertain to the down run.
On the down run in the stagnant fluid (which is single-phase formation brine or workover fluid) below the bottom set of perforations, the RPS shows a very steady value of three chart divisions; that is, 0.75 RPS. The tool used for this survey has a threshold velocity of 5 ft/min. This amount of line speed in the stagnant fluid is required to overcome frictional torque and start the spinner rotating. Therefore, the velocity driving the spinner on the down run through the stagnant fluid is 22 – 5 = 17 ft/min. The sensitivity of the tool is 17/0.75 = 22.67 ft/min/RPS. In the stagnant fluid, a sustained defelection of 0.2 chart division would be recognizable on the RPS trace. The single-phase sensitivity of this tool is
0.2(chart div) × 0.25(RPS / chart div) × 22.67 ft / min / RPS = 1.13ft / min.
The 7-in., 23-lbm/ft casing has a capacity of .0393 bbl/ft; thus, this velocity corresponds to a flow rate of
1.13(ft / min) × 0.0393(bbl / ft) × 1,440(min / D) = 64 B / D.
This value is at the low end of the range listed in Table 1 because the flowmeter in this case has very high resolution. On the down run, the flow from the bottom two sets causes an average deflection of 0.8 chart divisions to the right of the steady response in the stagnant fluid below the bottom set (the no-flow reference). In full flow, the average deflection is 6.5 chart divisions to the right of reference. Thus, the relative contribution of the bottom two sets is (0.8/6.5) × 1,800 = 222 B/D. A flowing temperature survey in the same well (not shown), analyzed separately, establishes a more reliable estimate of the contribution of the bottom two sets: 430 B/D. The flowmeter lost 200 B/D, which is consistent with the resolution stated for oil and water flows (see Table 1). As oil, because of its buoyancy, rises through water in the wellbore interval defined by the bottom two sets, the oil churns and circulates the water even if the water is flowing; in turn, this action in the heavy phase can either slow down or speed up the spinner element, depending upon what component of circulation in the heavy phase most affects the tool. On the down run, the spinner is slowed, diminishing the spinner-estimated relative contribution of these sets. On the other hand, the spinner element speeds up slightly on the up run in response to a circulation downward.
Some additional comments are in order relative to the spinner traces in Fig. 3. Note first the pass upward at a line speed of 22 ft/min (the dashed-line trace). This line speed is close to the value of the upward superficial velocity for the full-flow stream (18 ft/min). Consequently, the spinner deflection is small approximately 0.25 rps counterclockwise above 6,050 ft. Below this depth, this low speed accentuates the fluid turbulence associated with the lifting and fallback of water as oil moves through it. This churning and circulatory action causes the spinner speed to flip-flop across the zero-rps axis with a reversal in the direction of spin on each crossing. On the downward pass at 22 ft/min (the solid-line trace), the higher-frequency oscillations are much smaller than on the upward pass. With the relative sped of the fluid to the tool being approximately 10 times larger than on the up pass, the sample time is too small for the spinner response to mirror the full extent of the high-frequency fluctuations seen on the up pass. As a result, one sees on the downward pass primarily the occasional lower-frequency events associated with gross slugging or "heading" in the flow.
The reduced effect of turbulence on the pass downward illustrates an adage that one still finds repeated in the literature to the effect that a high logging speed should be used to "minimize" the effect of turbulence. Although the claim is true, the implication that the response is more accurate is not. The reduced sample time caused by high relative speed between fluid and tool introduces a "hidden" bias toward higher spinner speed in flows affected by multiphase turbulence. The positive bias results from the greater influence of the life surges relative to the fallback flow along the casing wall. Such a bias is evident on the record from the upward pass, the dashed line in Fig. 3. An average value for the spinner speeds on the section of record in the interval of 6,090 to 6,130 ft is clearly to the clockwise side of the zero axis by approximately 0.25 rps, whereas the value above 6,050 ft is approximately 0.25 rps counterclockwise. The gross bias of approximately 0.5 rps represents a flow of 160 B/D in this example. The bias increases as the relative speed increases. Some spinner tools rectify the output pulses before conversions to rps values. On records from such tools, turbulence of any type produces a bias to higher speeds, owing to the interaction between sample time and spinner inertia.
The calculation immediately preceding is an example of relative profiling; that is, using the flowmeter log to establish relative contributions, with the total flow rate known independently. When an absolute flow rate is needed, the method of downhole calibration can be used. Downhole calibration is appropriate for injection flows or single-phase production except for, perhaps, the deepest entry, which may be submerged in captive completion fluid.
To perform a downhole calibration, the well is shut in at the surface. Both up and down runs are made through the static fluid at various cable speeds. On the calibration plot, cable velocities appear on the horizontal axis, with downward logging speeds as positive velocities and upward speeds as negative velocities. Values of RPS appear on the vertical axis, with values positive for down runs and negative for up runs. For each run, a point is plotted that corresponds to the cable velocity and its respective RPS value. A "best-fit" straight line is constructed for the plotted points of the down runs; this line corresponds to clockwise rotation and is located in the first quadrant of the plot. A second best-fit straight line is constructed for the plotted points of the up runs; this line corresponds to counterclockwise rotation and appears in the third quadrant.
The line for the down runs intersects the cable velocity axis to the right of the origin, and the velocity value at this intersection is the ideal or extrapolated threshold velocity for down runs. This value is slightly less than the actual speed needed to overcome frictional torque and start the spinner element rotating. The slope of this line is the sensitivity of the tool during down runs. The line for the up runs intersects the cable velocity axis to the left of the origin, and the absolute velocity value at the intersection is the threshold velocity for up runs. The value of the slope of the line is the sensitivity of the tool during up runs. For a quality tool, the two threshold velocities are nearly the same, as are the two sensitivities.
As an example of the use of the downhole calibration, consider a down run with the well producing. The RPS value at a depth of interest is taken from the log and on the plot is projected horizontally onto the straight line for the down runs. The corresponding value on the velocity axis is the velocity that drives the spinner element at this location. From this velocity, subtract the cable speed to obtain the net velocity. This velocity value is the apparent flow velocity. Multiplication of the net velocity (ft/min) by the pipe capacity (bbl/ft), by 1,440 (min/D) and by an independently determined calibration factor yields the desired flow rate. The calibration factor is needed because the apparent velocity as measured by a spinner is larger than the average or superficial velocity of the stream.
As another example, consider an up run with the well producing and with the spinner turning counterclockwise. The RPS value at a depth of interest is taken from the log and projected onto the line for the up runs. The corresponding value on the velocity axis is the velocity that drives the spinner element. From the absolute value of this velocity, subtract the positive value of the cable speed to obtain the net velocity. The rate is calculated from this net velocity in the same way. Had the spinner reversed to spin clockwise, the speed would be projected onto the calibration line for down runs and the cable speed added to the result to obtain the net velocity. An algebraic formulation of spinner speed as a function of cable and fluid velocities can be found on Production logging application tables under the category of injection profiling.
Noteworthy papers in OnePetro
Use this section to list papers in OnePetro that a reader who wants to learn more should definitely read
Use this section to provide links to relevant material on websites other than PetroWiki and OnePetro