The invention is generally directed to hydraulic lifting system and particularly to controlling the speed of a sucker rod pumping system.
A pumping system is typically used to lift oil and other wellbore fluids from a subterranean reservoir to the surface. One commonly used pumping system is known as a “sucker rod” pump. A sucker rod pumping system incorporates a downhole reciprocating pump comprised of a reciprocating piston inside a pump barrel that is attached to a production tube. The barrel is located in a subterranean reservoir which is at least partially filled with the well bore fluids. The piston is linked to a prime mover at the surface by a mechanical system that translates the rotational movement provided by the prime mover to the reciprocal movement required for the pump piston. The mechanical mechanism includes a rod string, a polished rod, a bridle, a horsehead, a pivotally supported walking beam and a rotating arm. The rod string is connected to the piston and runs inside the production tube through which the wellbore fluids in the subterranean reservoir are lifted to the surface. The rod string is connected to the polished rod at the surface end of the production tube and the polished rod is attached to the bridle which is coupled to the horse head. The horse head is attached to one end of the walking beam and translates its pivotal movement to the reciprocal movement required for the piston. The rotating arm is connected between the other end of the walking beam and the prime mover. The downward stroke starts at the highest point of the horsehead and continues until the horsehead has reached its lowest point. During the down stroke the rod string and piston in the downhole reciprocating pump descend as gravity pulls them downward. The upstroke is powered by the prime mover, which lifts the rod string and piston upward until the horsehead has reached its highest point again.
As the piston descends on the down stroke a check valve (sometimes called the delivery valve or traveling valve) in the piston opens to let wellbore fluids in the barrel pass though. At the same time a check valve (sometimes called the inlet valve or standing valve) in the barrel closes to prevent wellbore fluids in the barrel from escaping into the subterranean reservoir surrounding the barrel. As the piston is raised on the up stroke the delivery valve is closed such that wellbore fluids that are above the piston are lifted upward into the production tube and towards the surface. At the same time the piston is being raised on the up stroke the inlet valve in the barrel opens permitting wellbore fluids in the subterranean reservoir surrounding the barrel to be sucked into the barrel. The cycle described here repeats during each complete stroke of the sucker-rod pumping system.
To operate a sucker-rod pump in a cost effective manner, the pump fillage level and speed of the stroke should be set such that a profitable amount of wellbore fluid can be extracted by the pumping system while avoiding conditions where the well is pumped off. A pump off condition occurs when the rate at which the subterranean reservoir is supplying wellbore fluids to the barrel is exceeded by the rate at which wellbore fluids are being pumped to the surface. When a well is operating in a pumped off condition it is not operating in an effective and efficient manner. If the well is allowed to continue operating in a pump-off condition damage to the rod string and the downhole reciprocating pump will most likely occur. Any damage to the rod string or downhole reciprocating pump will result in down time for the well and expensive repairs to the damaged components. Therefore, an accurate means for determining the wellbore fluid level, pump fillage and adjusting the speed of the pumping system to maintain a cost effective operating level is desirable.
The present invention determines an optimal speed for a sucker rod pump by monitoring the torque of the prime mover providing motive force to the pump system. Since gearbox input torque, and crankarm torque are proportional to the prime mover torque, these torque values could also be used to provide similar results. The torque values are processed by a microprocessor according to an algorithm stored in a memory associated with the microprocessor. The results of the processing provide an accurate indication of pump fill which is then used by the microprocessor to adjust the pump an optimal speed for maintaining a cost effective operation of the pumping system.
The microprocessor performs the following operation according to the algorithm stored in the associate memory:
The present invention provides a method for accurately determining pump fill and adjusting pump speed to an optimum level for conventional or air balanced sucker rod pump using the API Spec. 11E geometry (also known as Rear-mounted geometry and Class I lever systems with crank counterbalance) and Mark II pumps that use the API Spec. 11E standard geometry (also known as a Front-mounted geometry and Class III lever systems with crank counterbalance). Referring to
During one complete stroke of the pumping system 10 the horsehead 38 falls from its highest position to its lowest position and returns to its highest position. As the horsehead 38 falls to its lowest position (
To operate a sucker rod pumping system 10 described above in an efficient manner the speed at which the pumping system 10 operates must be controlled such that the maximum amount of well bore fluids 86 are delivered to the exit tube 106 at the end of each upward stroke without lowering the level of well bore fluids 86 in the subterranean reservoir 82 to a point at which a pump-off condition results.
Referring now to
At step 200 the microprocessor 110 initiates the pump control algorithm 118 as the pumping system 10 begins a pump stroke. At step 205 the pumping system 10 begins to monitor, at predetermined regular intervals, raw torque of the prime mover 14 with respect to a particular horsehead 38 position. Raw torque can also be monitored at several points in the mechanical linkage 34, however, the prime mover 14 provides the easiest point for monitoring and will be indicated as the torque monitoring point in the example discussed herein. The number of intervals monitored should be sufficient to produce a graphical representation of the pump stroke that appears smooth to the naked eye and is limited only by the technology used. It is also understood that at any time during the disclosed process the number of intervals can be downsampled or filtered by any known means such as averaging, moving average, interpolating, removing outlying torque samples, decimation, low-pass, exponentially weighted moving average (EWMA), finite or infinite impulse response, or frequency domain filtering, etc. to make the calculations more manageable and to make the graphic representation of the array smoother. The torque of prime mover 14 can be measured or determined by using a torque sensor, calculated by the system controller 18 or estimated from ammeter or power meter measurements.
At step 210 microprocessor 110 stores the monitored prime mover 14 raw torque and associated horsehead 38 positions of a complete pump stroke in memory 114 as a raw torque array Traw, as shown below where N is the number of intervals monitored.
T(raw)=[T(raw0),T(raw1),T(raw2), . . . T(rawN)]
At step 215 microprocessor 110 creates a filtered torque array Tf from the raw torque array (Traw) and stores the filtered torque array Tf in memory 114. As indicated above, downsampling or filtering can be done by any know means, for example a moving average as indicated below.
Tf=(T(n)=T(n−1)+T(n−2))/3
At step 220 microprocessor 110 creates a rotatum array R of the down stroke from the filtered torque array Tf, shown in
R(n)=[(Tf(n)−Tf(n+B))]
The value of B can be selected by examining torque data from any well, or collection of wells. The selected value of B should accentuate the effects of pump fill in the generated rotatum array R. Torque curves, and downhole cards from one or more wells, can be compared with rotatum arrays from the same wells to see if there was a strong correlation between pumpfill as shown by the rotatum minimum and pump-fill as shown by the torque curve or downhole card.
When the piston 74 of the down hole pump 66 encounters the well bore fluids 86 there will be a change in prime mover 14 torque. The magnitude of torque change and span of horsehead position over which these changes occur determines the range for value of B such that:
1. The minimum value of B is limited because Tf(n+B) must be spaced far enough apart in time from Tf(n) so that when viewing the resulting rotatum curve or scanning of the rotatum array R by the microprocessor 110, there will be a detectable difference in torque value between them at the point when the piston 74 encounters the well bore fluids 86. B must be greater than 1 because the closest sample to compare is the adjacent sample.
2. The maximum value of B is limited because Tf(n+B) must be spaced close enough in time to Tf(n) so that there will not be a greater difference in torque between them than could be caused by things (such as differences in mechanical advantage of the crankarm to the linear motion of the bridle at different points in the stroke, or changes in counterweight balance position) other than the piston 74 encountering the well bore fluids 86. To reduce the effects of the above phenomena, the torque samples being compared should generally be less than 25% of the downstroke apart from each other.
3. The value of B that best accentuates the effects of pump fill in the rotatum curve is selected from values between the maximum and the minimum of Tf(n+B).
In some instances a non-integer value of B is selected to best accentuate the effects of pump fill in the generated rotatum array R, the value of torque at (n+B) can be estimated by using linear interpolation between points (n+A) and (n+C). The following formula is used to determine the portions of point (n+A) and (n+C) required to produce the non-integer (n+B).
R(n)=[a*(Tf(n)−Tf(n+A))+c*(Tf(n)−Tf(n+C))]
As an example, in a pumpjack where 128 samples per stroke were stored, comparison between points that are 1.2 samples apart was selected for (n+B) based on the description provided above for comparing pumpfill as shown by the rotatum minimum and pump-fill as shown by the torque curve or downhole card and determining the minimum and maximum values for (n+B).
The following chart shows values that can be used in the formula for the example above.
At step 225, microprocessor 110 determines whether the pump is a conventional pump or a Mark II pump. Information relating to whether the pump is conventional or not conventional (Mark II) is usually provided by well management personnel during commissioning of the pumping system 10 and stored in memory 114. If it is determined at step 225 that the pump is not conventional the microprocessor will proceed to step 230, which will be discussed in detail later. If it is determined at step 225 that the pump is conventional the microprocessor will proceed to step 245.
At step 245 the microprocessor 110 will determine if the well is suspected of having low pump fill and therefore a low producing well. Information indicating that a well is known to have the possibility of low pump fill is stored in a flag. This flag can be set at well commissioning or any time it is learned or suspected that the well has a possibility of having low pump fill. This flag is stored in memory 114 for use at step 245. The flag can be set by the well manager, operator or microprocessor 110 after determining that the pumpfill trend from one stroke to the next is decreasing consistently and trending in a way that suggest true pumpfill will drop below 50%. Other indicators such as the peak raw torque being in the upper half of the down stroke, as shown in
At step 230 the microprocessor 110 determines pump fillage. In a conventional well this is accomplished by scanning the down stroke portion of the rotatum array R for a rotatum minimum Rmin and a maximum horsehead 38 position, as shown in
In a conventional well the pump fill is determined by dividing the horsehead 38 position associated with the rotatum minimum Rmin by the maximum horsehead 38 position. In
Prime mover 14 torque is applied slightly different in a non-conventional Mark II pump and therefore the graphical representation of the array TrawMII for a full pump stroke is different, as shown in
At step 235 the microprocessor 110 determines the optimal pump system 10 speed from the determined pump fill by comparing the determined pump fillage with a previously determined target pump fillage. The difference between the target pump fillage and the determined pump fillage is the fill error. The pump speed is adjusted to eliminate or reduce the fill error. To prevent extreme speed changes, the speed will be increased or decreased by no more than a predetermined percentage at each pump speed change.
Steps 250 through 260 are for conventional pumps that are operating on wells that have been suspected of being low producing wells in step 245. Steps 250 and 255 provide a more accurate determination that the well is truly a low producing well and step 260 provides a more accurate determination of the pump fillage position in a low producing well.
At step 250 the microprocessor 110 determines whether the peak torque Pt as indicated in
At step 255 the microprocessor 110, using the rotatum minimum Rmin of
At step 260, microprocessor 110 will modify the rotatum vs horsehead 38 position array R of
Number | Name | Date | Kind |
---|---|---|---|
2691300 | Morris | Oct 1954 | A |
4490094 | Gibbs | Dec 1984 | A |
7558699 | Beck et al. | Jul 2009 | B2 |
7891237 | Mills | Feb 2011 | B2 |
8322995 | Ehimeakhe | Dec 2012 | B2 |
8328527 | Ehimeakhe | Dec 2012 | B2 |
8793080 | Erlandsson | Jul 2014 | B1 |
8892372 | Peterson | Nov 2014 | B2 |
8910710 | Krauss | Dec 2014 | B2 |
20110103974 | Lamascus | May 2011 | A1 |
20120251335 | Hurst | Oct 2012 | A1 |
20140088875 | Krauss | Mar 2014 | A1 |
20140129037 | Peterson | May 2014 | A1 |
20150275651 | Chen | Oct 2015 | A1 |
20150345280 | Krauss | Dec 2015 | A1 |
20160265321 | Elmer | Sep 2016 | A1 |
20170016313 | Pons | Jan 2017 | A1 |
20170037845 | Al Assad | Feb 2017 | A1 |
20170167482 | Assad | Jun 2017 | A1 |
Number | Date | Country | |
---|---|---|---|
20180038366 A1 | Feb 2018 | US |