Professional Documents
Culture Documents
SPE 10255
Member SPE-AIME
tendency of the air cylinder to require frequent al calculations would be necessary to include the
maintenance. operating limits of maximum rod stress and maximum
beam load.
The Type III unit shown in Figure 1 has a similar
geometry as a Type II unit. Consequently it enjoys In order to solve this problem a computer pro-
the same reduced acceleration loads when compared gram has been written that will calculate gear box
with a Type I (conventional) unit. However, the Type torque, maximum beam load, and maximum rod stress
III unit has the equalizer arm moved forward on the for a given pumping system over a given range of
beam and is no longer centered over the crank shaft pumping speed and pumping depth. The calculated
of the gear box. The counter weights are also out of values are then plotted as a three dimensional
phase with the cranks. As shown in Appendix B this surface of gear box torque as a function of pumping
has the additional advantage of further reducing the speed and pumping depth. The program can generate
acceleration loads on the upstroke and increasing a surface for any pumping system, using any of the
them on the downstroke. Thus, the maximum and the four types of pumping units, over any pumping speed
minimum beam loads are brought closer together. This and pumping depth range. Either the API or the
allows for more efficient counterbalance and conse- classical method can be used in the calculations.
quently for reduced gear box torque loads when com- Figure 4 is an example of the surface generated
pared to an equally loaded Type I (conventional) unit using the data in Table 1 and API method. The
or a Type II (air balanced) unit. torque limit for the data in Table 1 has been drawn
across the surface as well as the maximum stress
The Type IV unit shown in Figure 1 has a geometry limit for the rods and the maximum beam load. From
similar to a Type I unit except that it has the samp- Figure 1 it can be seen that the operating limit
son post centered on the beam and the equalizer arm, for the pumping system in Table 1 occurs at the
at the rear of the beam. is no longer centered over point where the beam load limit and the gear box
the crank shaft. The counter weights have been torque limit intersect. This occurs at a pumping
placed out of phase with the crank. These changes, speed of 10.5 strokes/min. and a pumping depth of
as shown in Appendix B. have given this unit 8600 feet. At deeper pumping depths or faster
acceleration loads that are similar to a Type III pumping speeds the pumping system would develop gear
unit except that the maximum acceleration load still box torques or beam loads that would exceed the
occurs at the bottom of the downstroke when the fluid operating limits of the system. For this particular
load is transmitted to the rod string, just as a Type set of data the maximum rod stresses that are
I (conventional) unit. Therefore, the Type IV unit developed do not limit the operating range of the
will experience loads and torques that are less than system.
an equivalently loaded Type I unit, but they will not
be as low as the loads and torques carried by the Figure 4 illustrates several interesting points
Type III unit. For most Type IV installations the concerning the API calculation method. The surface
loads and torques carried by the unit should be is irregular. There is a pronounced valley that
roughly the same as a Type II (air balanced) unit. runs diagonally across it from 6500 feet and 15 spm
to 9500 feet and 5 spm. This valley is formed by
The API design approach has been generally a leveling and in some areas a decrease in calculated
recognized as a superior design method for Type I gear box torque with increases in depth for the lower
(conventional) pumping units. Design methods for to medium pumping speeds. This calculational
Type II and Type III units have been presented by the phenomenon is illustrated for low pumping speeds in
manufacturers of these units ll • l2 and are generally Figure 2. The decrease in peak torque, using the
empirical variations and combinations of the API and API method, at low pumping speeds and deep pumping
the classical design approaches. No design approach depths is due to large damping effects. Damping
that satisfies the geometry of the Type IV unit has effects increase with increases in depth, if the
been presented. Appendix B presents a set of design pumping system is held constant, because the rod
equations that generate results that fit the geometry string becomes more flexible as pumping depth is
of this unit. These equations have been used in increased. Also, at low pumping speeds the dynamic
this paper to illustrate the results predicted by loads are very small and are quickly nullified by
the analysis of the Type IV unit geometry. The the damping effects at deep pumping depths. However,
equations have not been tested by field measurements. as the pumping speed is increased the dynamic loads
Until such data becomes available the equations become more important than the damping effects and
presented in Appendix B for the Type IV unit should the calculated gear box torque increases steadily
be used with caution. as pumping depth is increased. This feature of
the API method is illustrated in Figure 3.
THREE DIMENSIONAL DESIGN
The classical methods do not consider damping
Figures 2 and 3 show gear box torque versus effects. Thus, as Figures 2 and 3 indicate, the
pumping depth at a fixed pumping speed for the classical methods always predict increases in gear
pumping system shown in Table 1. The figures box torque with increases in pumping depth. Figure
present the gear box torque predicted for Type I 5 is a surface generated using the data in Table I
(conventional) units. using the API method and the and the classical method of design. This surface is
classical method. From these figures it is obvious smooth and always predicts an increasing torque with
that it is possible to predict the operating limit increases in pumping speed and pumping depth.
of a given pumping system operating at a given speed. Figure 5 predicts an operating limit of 10 spm and
However, many such graphs would be required in order 9170 feet using the classical method. This illu-
to determine the optimum combination of pumping speed strates the tendency of the classical methods to
and pumping depth for a given system. Then addition- predict smaller loads on the pumping system at
SPE 10255 PAUL M. BOMMER 3
medium to high pumping speeds than the API method. load that this unit can achieve when compared to any
At low speeds the API method predicts generally other unit type (see Appendix B). Had a heavier
smaller loads than the classical methods due to the beam been used with this unit it would be capable
effects of damping. of pumping from 9500 feet at a maximum pumping speed
of 14 strokes per minute. The equations used in the
Pumping unit designs tend to vary somewhat be- design of this unit are a blend of the API and the
tween manufacturers, even for the same type of unit. classical methods. The equation for gear box torque
Variations in the safety factors built into the de- comes from the classical approach. As was pointed
sign of unit components and variations in the way out earlier, the result is a smooth torque surface.
the high stress points of the units are held together
can cause large differences in the safe loads that Figure 11 shows a surface for a Type IV unit.
the units can carry. Thus, when units are being Again, the loads predicted for this unit are less
used that may have lower safety factors or somewhat than the Type I unit. However, the loads are still
weaker couplings than other brands of the same unit considerably larger than those predicted for a Type
type it is advantageous to use the design method III unit. Figure 11 was generated using,the equa-
that predicts the largest possible load on the unit. tions presented in Appendix B. These equations are
This practice causes a larger unit to be used than a blend of the API and the Classical methods and
would be required for units with larger safety fac- they predict loads that follow the trends suggested
tors or more heavily reinforced stress points. The by the theory of Appendix B. It must be pointed out
three dimensional presentation can be used to form that, unlike the equations for the other three types
a composite surface using the largest predicted loads of unit, the equations for the Type IV unit have not
from the API and the classical methods. Figure 6 been tested against measured data.
shows a composite surface using the largest predicted
values from both methods. This figure demonstrates Many operational effects can be examined using
that the API method calculates larger loads except the three dimensional surface. Perhaps the most
at medium to low speeds and deep pumping depths important operational consideration that is commonly
where the classical method predicts larger loads. If neglected is that of pump submergence. When a well
a machinery factor is used in the classical method has a fluid level in the annulus that is above the
then the classical method predicts larger loads than pump setting depth a U-tube effect oCCurs. Thus,
the API method except at lower speeds and shallower the net lift of the pumping system is from the
pumping depths. This is shown in Figure 7. The fluid level to the surface rather than from the pump
machinery factor, as described in Appendix A, is a depth to the surface. This effect can greatly re-
of crank and pitman arm length. This duce the loads on the pumping system because the
has been described as a correction for the fluid is actually being lifted from a depth shallower
crank angle of the pitman arm and for variations in than the pump depth. Figure 12 presents a surface
angular velocity of the crank that lead to increases for a Type I unit, using the data of Table 1. and
in the acceleration loads on the system. As shown 3000 feet of pump submergence at every depth. The
in Figure 7 using a machinery factor of 1.318, with loads carried by this system are much less than the
the data of Table 1, has decreased the area of safe loads on the same system without any pump submer-
operations across the surface when compared with gence (Figure 4). Unlike the system with no pump
the API method or the classical method with no submergence the maximum beam load and rod stress are
machinery factor. never reached and the gear box torque limit is
reached at a much deeper pumping depth and faster
The surface can be used to examine several pumping speed. Comparing the system with 3000 feet
torque limits over an operating range. Figure 8 of pump submergence (Figure 12) to the system with
shows a surface that was generated using the API no pump submergence (Figure 4) it can be shown that
method and the data in Table 1. The beam load and the same pumping unit operates at deeper depths and
rod stress limits have been drawn on the surface and at higher speeds if pump submergence exists and
a torque limit line for 228000 in-Lbs and 160000 where pump submergence exists a smaller pumping unit
in-Lbs has been constructed. This generates a series can be used to do the same amount of work.
of isotorque lines across the surface. These iso-
torque lines can be used to check at a glance the CONCLIJSIONS
operating limits of several sizes of pumping units.
A three dimensional surface of gear box torque,
A surface for a Type II (air balanced) unit is beam load, and rod stress as a function of pumping
shown in Figure 9. This surface is exactly the depth and speed has been presented for four types
same shape as the surface for the Type I (conven- of pumping units. The surface can be generated
tional) unit shown in Figure 4. However, the loads for any size pump, any sucker rod combination, any
carried by the Type II unit are not as large as those pumping unit type and size, and over any range of
calculated for the Type I unit. The surface is the pumping depth and pumping speed.
same shape for both unit types because the API method
has been used for both units. The Type II unit loads The surface can be used to quickly define the
are less than the Type I units because the API operating limits of a pumping system. It can also
equations for the Type II unit have been adjusted to be used to compare the importance of the many
account for the changes in the dynamic loads that are variables that make up a pumping system. Further-
brought about by the geometry of the Type II unit more, the surface can be used to examine the effects
(see Appendix A). of operational conditions such as pump submergence
and unanchored tubing.
Figure 10 presents a surface for a Type III unit.
This surface illustrates the dramatic reduction in The various pumping system design methods have
4 SUCKER ROD PUMPING SYSTEMS DESIGN - ANOTHER LOOK SPE 10255
been examined and compared using the surface. The 3. J. C., "Vibration Problems in Oil
equations of each method for each type of pumping API. New York (1937),
unit have been studied and a set of equations has p.
been for use with the new Type IV pumping
unit. 4. Coberly, C. J., "Problems in Modern Deep-Well
Pumping", Oil Gas J. (May 12, May 19, 1938), p. 126-
A composite surface has been constructed using 137, p. 88-95.
the various design methods. The composite surface
depj~ts the largest loads predicted by each method. 5. Craft, B. C., Holden, W. R., and Graves, E. D.,
This surface could be used in the design of pumping
units that had smaller manufacturers safety factors
or less heavily reinforced stress points. p. 280-367.
Weight of rods in well fluid N (Lb) Here w is the constant angular velocity of the
crank. Using the quadratic formula the displacement
x Displacement of pitman arm from highest of the equalizer arm (x) can be obtained
point of travel m (ft)
z Pumping unit geometry factor (dimen- x = h+c(l-cos(wt» +
sionless)
2 k
(wt) + (h ] 2 . (A2
2 2 2 2
[c cos (wt)+(h -c ) . . . (A3) Equations (An and (AS) when multiplied by the
weight that is supported by the polished rod at
An expression for the acceleration of the equalizer the appropriate crank angle will predict the
arm (point B) can be obtained by taking the second maximum and the minimum force due to acceleration.
derivative of displacement with respect to time in
equation (A3). Type II (air balanced) units have the crank
and pitman assembly placed in front of the sampson
2 2 2 post. as shown in Figure 1. For Type II units
d x/dt cw cos (wt) +
angle wt is zero or an even multiple of Pi, again
when the equalizer arm is at the highest point of
2 2 2 2 -3/2
~[c cos (wt)+(h -c )] travel. However, since the crank assembly is in
front of the sampson post the highest point of
travel occurs at the end of the upstrok~ rather
than the end of the downstroke as with a Type I
unit. Thus, the maximum acceleration load occurs
2 2 2 2 -~
~[c cos (wt)+(h -c )] at the end of the upstroke with a Type II unit
and the minimum acceleration load occurs at the
2 2 end of the downstroke. A Type II unit experiences
c 2w cos(2wt) • (A4)
the minimum acceleration load at the time when the
fluid load is transferred to the rod string. When
The maximum acceleration of the equalizer arm occurs
compared with a Type I unit the maximum polished
for wt equal to zero or any even multiple of Pi.
rod load for a Type II unit is reduced and the
The cos(wt) for wt equal to zero or any even
minimum polished rod load is increased. Therefore,
multiple of Pi is a positive one. Thus. the ex-
the counterbalance for a Type II unit is more
pression for the maximum acceleration of a Type I
effective in balancing the loads and the gear box
pumping unit is torque is reduced as a result.
The minimum acceleration load occurs for wt is the upstroke acceleration term (Equa. CA7)
equal to Pi or any odd multiple of Pi. At these Type I units and Equa. (AB) for Type II units).
points the cos(wt) is equal to a negative one.
F is the downstroke acceleration term (Equa. (AB)
Z
for Type I units and Equa. CA7) for Type II units).
2 2 2
= w c[l-c/h] (m/s or ft/s ) . . . (A6)
reversals are minimal. The counterweights are unit. This is necessary due to the increased
placed 7.5 degrees out of phase with the crank so acceleration loads during the downstroke of a
that the counterweight effect will be in phase Type IV unit.
195 degree upstroke and 165 degree down-
CBL 0.5 (PPRL + MPRL) (B3)
PPRL + +
. . . (Bl)
Here:
r--
UJ
Ln I
+ m
Ul
95
0
~··"i
95 ~
0
Ul
MF 1. 318
95 C)
CD
95 C) Classical.
Mi' 1.0
I
Z 95 C)
(\J /
(') I
I
('i C)
I
I
W ::r
/
("\J Cl
Q'!
o
f---
6 50
DE TH (F T x 1 0 3)
Fig. 2 - Gear box torque vs_ pumping depth at 5 spm
('J
::r
/C'J
(')
+ (\J
UJ (!j
.,----\
(')
"
I>
(') Clas:sic.31
HF 1.318
(1i
/ C)
/
aJ (') ;('i C)
/
(')
(J'j C)
Z /
/
MF 1.0
CD C)
CD
(\J
W (') C)
UJ
(\J
0 C)
er:: CD
(')
f- (\J ----L__
6.50 7. 10 7.70 .so
DEPT (F T X 1 0+3 )
V;
~;z:
276282
W IBq168
::0
C!!
~
.....
~
0::
w
"-
226 C~NVENtlONRL
21300 L6 aERM
9S IN STR~KE
1, 2S I~ PLUNCER
TYPE 76 RODS
_ l~RgU< LIMIT
. qOOOO PSI ROD STRESS LIMIT
BEAM LIMIT
3~7976
If)
':lI 2S09B2
:z
::
w 173908
::0
C!!
0::
C
I-
e699~
"-
a:
w
CL
228 CONVENTlONRL
21300 La aERM
B6 IN STROKE
1.25 IN PLUNCER
TYPE 76 R"DS
lOR(,IUE LIMI1
PSI ROD STRESS LIMIT
LIMIT
368376
V;
co 276282
....J
I
:z
,~
w le~188
::0
0
'"
C
I-
~ 9209~
cr:
w
CL
Fig, 6 - API and classical method composite surface - machinery factor =1.0
~1~I05
U;
a:l 309079
-'
I
::z
w 206053
5
""
~
103026
:''"5
<L
228 CeNVENTiONAL
21300 LB BEAM
8S IN STROKE
1.25 IN PLUNCER
TYPE 76 RODS
TeRQUE LIMIT
•. ~oooo PSI ~oo STRESS LIMIT
_ _ BEAM LlMlT
Fig. 7 - API and classical method composite surface - machinery factor =1.318
36837S
!.f1
cc
-'
I
::z
::;
UJ la~la8
:::l
c:l
'".....
<:J
'"tt
UJ
a..
3536~1
;;:;
a:l 265231
-7
z
::;
UJ 176620
:::l
c:l
:;s,...
'"tt
UJ
a..
U;
~
I
Z
::
lI.J U3Y66
::J
C>
::'"
'"ct:
lI.J
"-
J._ Dovml:iLrokc
Hid u p;~; t roke S ta rt: downs t roke Mid downs t roke Start upst roke
Type I
LS"
8.0'
9.2.') ,
Type II
a = 4.67' 20075 Ib
6.25' 35075 Ib
crank supports F3 15000 Ib
0
11. 46 1b
a ~ 7.29' F ~ 22 1.79 Ib
Z
)13 15000 Ib
ail;: cylinder suppot'ts
III
F 2903 Ib
,~ 1
'" F 2 = 18624 Ib
J !C: 1
1 F
2
F
3
F) ~ 15000 1b
Type I V
b 8. I ' F 15741 Ib
I I I 1
8.5' ~ :JQ741 Ib
!F1
f
F2
1
F
)
F
3
15000 Ib