压缩机-泵-制冷工程.doc

外文翻译
Chemical and Petroleum Engineering, Vol. 40, Nos. 11–12, 2004
COMPRESSORS, PUMPS,
REFRIGERATION ENGINEERING
UPDATING PISTON PUMPS FOR OIL
PRODUCTION

B. S. Zakharov,1 G. N. Sharikov,2
and E. G. Kormishin2

The three-plunger acid treatment pump SIN32 and the two-cylinder double-acting pump NPTs-32 with four working chambers (for cementing units) have been updated to control pump delivery. The fluid delivery diagrams for pumps of various designs are examined and the test results are reported.

In drilling and oil production, single-acting three-plunger (triplex) pumps or double-acting two-cylinder (duplex) pumps are used.
In injecting reagents (clay drilling mud, water, cement, acid, etc.) into wells, depending on the technology applied，it is required to inject the fluid in amounts ranging from the maximum to the minimum in a single operation. If the bed accepts the injected fluid well, it becomes necessary to maximize pump delivery for quick completion of the operation. If on the other hand, the bed does not accept the fluid well, it becomes necessary to reduce pump delivery so as to restrict the injection pressure to the safe limit. At present, because of wear of well (down-hole) equipment, the permissible injection
pressure is not higher than 10–15 MPa..
The delivery of a piston (reciprocating) or a plunger (displacement) type of pump can be controlled in the following ways:
• by installing several pumps with identical or different pumping capacities;
• by changing the drive rotation speed;
• by using cylinders (plungers) of the required size;
• by channeling a part of the fluid into a bypass; and
• by dismounting one or several valves.
The first version is used essentially in drilling. In oil production, generally all versions are used either individually or in some combination.
All pumping units designed for injection of various fluids (fluidal materials) for cementing, hydraulic formation fracturing, hydraulic sand-jet flushing of sand bridges, and other flushing operations in oil and gas wells are mounted on the chassis of motor vehicles (trucks), tractors, caterpillar (tracked) carriers, and specially made carriages.
The operating parameters of the pumps (delivery and injection pressure) depend on the power of the drive and maximum and minimum speed of the engine and the pump. The pump delivery can be changed by changing the number of pump strokes without stopping the engine with the help of a gearbox (by gear shifting) and with stopping of the engine by installing cylinders of the required size. Replacement of the cylinders takes a lot of time and is not always possible in a continuous echnological process. In the existing pumping plants, the delivery variation range is inadequate. At the minimum rotation speed and cylinder diameter, the delivery remains extremely high, and for injecting the fluid into the bed the pressure has to be raised above what is permissible.

Assigned by NGDU Zainskneft’, Ékogermet carried out updating of two types of pumps, namely, SIN32 and NPTs-32.
In the three-plunger (triplex) acid treatment pump SIN32, for reducing the minimum delivery down to 1.0 m3/h,plungers having a diameter of 125 mm were replaced with plungers having a diameter of 55 mm. As a result, the theoretical pump delivery was reduced from 16 down to 3.3 m3/h. Further reduction of the pump delivery was achieved by reducing the rotation speed of the vehicle engine to the possible minimum (500–600 rpm).
Simultaneously with this, a new design of packing glands (sealing devices) of plungers of the UPN55 type was developed.It was based on Zakharov mechanical seal [1], which demonstrated high reliability and durability in sucker-rod (oil) pumps. The sealing units and the pistons with a diameter of 55 mm were made for the SIN32 pump by ÉLKAMneftemash in Perm. Its finishing and testing were done by Ékogermet jointly with NGDU Zainskneft’.
The design of the UPN55-type plunger seal is shown in Fig. 1. The combined seal consists of the main threestage mechanical seal 4 and an elastic sealing collar 2. Each stage of the mechanical seal consists of ten rings that are elastically pressed against each other and simultaneously against the plunger surface. The rings are pressed against the
plunger in pairs from the opposite sides. The next pair is turned relative to the preceding one by 90º. The rings are pressed in the axial direction by rubber rings of round cross section and in the radial direction, by rubber girdles with eccentric collars. The plunger 5 is made of steel 45 and is chromium-plated and the sealing rings are of bronze. Three
cartridges with mechanical seals were installed in the housing bore 3 with a clearance that helps self-centering of the seals relative to the plunger. The cartridges are pressed together by a round nut 1 through a bushing with the sealing collar 2. There are holes in the housing for injecting oil and draining out the overflow into the receiving (suction) line of the pump.
In contrast to the well-known elastic glands, the mechanical seal does not require periodic adjustments and ensures reliable operation of the assembly over a long period [2]. Use of the updated SIN32 pump having a UPN55 type of mechanical plunger seals confirmed that the proposed design operationally fit.
From August through December 2003, NGDU Zainskneft’ carried out seven bottom-hole treatments (BHT) of six wells using the updated SIN32 pump. Different types of technological operations were carried out in the wells: mud acid BHT, muriatic (hydrochloric) acid BHT, injection of the reagents SNPKh-9021, MIAPROM, and RMD, for which SIN32 and ATs-32 pumping units were generally used. If acid or any other reagent could not be forced through (injected) at 12–15 MPa pressure, a low-capacity unit was connected with the SIN32 pump. In that case, the injection pressure dropped by 2–4 MPa。Injection was completed at the third-gear speed of the engine.

The NGDU technologists believe that connecting a low-delivery unit with an SIN32 pump offers the following advantages:
• possibility for continuous injection of acids and reagents in case of low intake capacity of the bed and for prevention of opening up of the fractures (hydraulic fracturing) of the collector and excessive rise in flow string testing pressure;
• extended operating life of the flow string by virtue of pressure stabilization during injection; and
• action of the acid throughout the perforation period and more complete reaction with the rock when the acid infiltrates the bed.
Since the maximally possible delivery of the SIN32 pump is reduced at least fivefold, NGDU Zainskneft’ proposed to perform all BHTs by injecting acids into the bed with the aid of a low-capacity unit and all other operations, with a standard unit. In that case, however, it would be necessary to place in the well, instead of one, two units, which have to be handled，by two teams, i.e., it will entail additional manpower and costs. Moreover, a low-capacity unit is not always fully utilized(does not operate to full capacity) and often stalls.
Thus, for a specific size of the cylinder it is necessary to reduce the pump delivery down to the minimum and, consequently,to broaden the range of control of the pump capacity toward its reduction while maintaining maximally possible delivery.
In multichamber pumps, this issue is resolved by shutting down (disengaging) one or several working chambers.
In duplex plunger pumps, disengaging one or two chambers will cause significantly uneven delivery, hydraulic shocks, disruption of the balance of loads on the drive, and failure of the pump.
In double-acting two-cylinder (duplex) pumps having four working chambers of the NPTs-32(9T) type, which are installed, for example, in ATs-32 cementing units, the delivery can be reduced by disengaging two rod chambers, which is achieved by removing two delivery (pressure) valves (Fig. 2).
The delivery of the NPTs-32 type of pump (duplex) having four chambers is
Q = 2(2F – ƒ)Sn,
where F is the cross-sectional area of the cylinder with a diameter Dc, dm2; ƒ is the cross-sectional area of the rod with a diameter dr, dm2; S is the stroke length, dm; and n is the number of double strokes per minute.
If the delivery (pressure) valves are removed from the rod chambers, the four-chamber pump turns into a two-chamber one with differentially acting cylinders. The delivery of such a pump Q1 = 2FSn.

If the valves from the front chambers are removed, the pump delivery can be determined by the equation Q2 = 2(F – ƒ)Sn.
Reduction of delivery by disengaging the rear (rod) chambers depends on the factor k1 = (2 – ƒ/F) and by disengaging the front chambers, on the factor k2 = [2 + ƒ/ (F – ƒ)]. It can be readily seen that for reducing delivery the front chambers have to be disengaged. However, theory and practice show that disengagement of the rod chambers is more advisable.
Thus, in NPTs-32 type of pump having cylinders of 90, 100, 115, and 127 mm diameter and rods of 45 mm diameter the delivery can be reduced 1.75–1.87 times by removing the valves from the rod chambers. At low loads (pressure drop
not more than 15 MPa and minimal delivery), the engine of the motor vehicle KrAZ-250 can run steadily at a rotation speed of 550 rpm. In the second gear with minimum engine rotation speed, the delivery of a pump with a cylinder of 90 mm diameter can be reduced down to 1.0 m3/h.
Unlike the SIN32 pump, the delivery of the NPTs-32 pump can be controlled during the technological operation and reducing or raising the delivery can change the pump output. Removal and installation of two valves do not take too long.
Let us see how the uniformity of pump delivery will change upon removal of the valves.
It is well known that the instantaneous output of a single-cylinder single-acting pump is
q = Frsiná = 0.5FSsiná,
where r is the radius of the crank and á is the crankshaft-turning angle.
The ratio of the maximum instantaneous delivery to the average delivery of the pump is called coefficient of delivery nonuniformity: ä = Qmax/Qav.
The average delivery of a four-chamber pump in one turn of the crank Qav = 2(2F – ƒ)S/2·3.14.
The maximum instantaneous delivery of a pump having four chambers and cranks turning at a 90° angle (Fig. 3a)Qmax = FSsin45° = 0.7FS. For the NPTs-32 type of duplex pump (Dc = 90–127 mm and dr = 45 mm), ä = 1.25–1.17.
After this, as the delivery (pressure) valves are removed from the rod chambers, the average delivery of a two-chamber differential (differentially-acting) pump (Fig. 3b) Qav = 2FS/2·3.14 = FS/ 3.14.
For such pumps, the maximum instantaneous delivery
Qmax = (F – ƒ)Ssin45° = 0.7(F – ƒ)S; ä = 1.65–1.91.
For all other types of delivery variation on account of removal of valves (in succession, all front pressure valves or crosswise, one of the front chambers and another of the rod chambers in another cylinder), the coefficient ä will be much higher. In general, in differential pumps, to reduce the nonuniformity in the pump delivery, the rod diameter is so chosen that its cross-sectional area is half that of the cylinder, i.e., ƒ = 0.5F. In that case, the delivery nonuniformity coefficient will be the lowest for two-cylinder differential pumps: ä = 0.7·0.5FS·3.14 /FS = 1.099.
For each cylinder, in order to get the coefficient ä = 1.099, it will be necessary to make a rod of a fixed diameter (63,70, 80, and 90 mm, respectively). But then, if the pump operates with all the valves, there will be a substantial increase in delivery nonuniformity and decrease in pump delivery. If the NPTs-32 pump is required to operate in two modes, it is perhaps advisable to make a rod of 55 mm diameter (for cylinders of 90 and 100 mm diameter) and of 70 mm diameter (for cylinders of 115 and 127 mm diameter). In that case, the delivery nonuniformity coefficient will be identical for both modes of pump operation: ä = 1.35–1.38. The theoretical pump delivery on account of increase in the diameter of the rod in a duplex-type pump will decrease roughly by 10%.
Let us see how the delivery nonuniformity will change if the pressure (delivery) valves are removed from the front chambers (Fig. 3c). The delivery, as was noticed earlier, will decrease more than twofold.
The average delivery of a series-produced NPTs-32 pump operating with two rod-chambers is
Qav = 2(F – ƒ)S/2·3.14 = (0.75–0.87)FS/ 3.14.
The maximum instantaneous pump delivery (forward stroke) is
Qmax = FSsin45° = 0.7FS.
The minimum instantaneous delivery (back stroke) is
Qmin = FSsin45° = 0.7FS = 0.7(0.25–0.125)FS.
The delivery nonuniformity coefficient is
ä = (Qmax + Qmin)/Qav = 3.87–3.45.
If the rod diameter of such a differential pump is increased, the delivery nonuniformity will increase further and,therefore, it will be necessary to remove valves only from the rod chambers.
In series-produced NPTs-32 type of pumps having four working chambers, the nonuniformity of the fluid flow in the delivery (pressure) and suction (intake) lines will be identical and will depend on the rod diameter. The smaller the crosssectional area of the rod, the greater will the fluid flow uniformity be.
In the suction (intake) line of a differential pump, the flow nonuniformity increases considerably because only two chambers operate and the cranks of the crankshaft are turned by not 180°, as is usual for single-acting two-cylinder (duplex) pumps, but by 90°.
The delivery nonuniformity coefficient in this case will be ä = Qmax/Qav = 2.199.
Suction conditions of differential pumps, just as of all other types of piston pumps, can be improved by installing air suction surge chambers in the suction line and placing the fluid tanks above the pumping unit.
An NPTs-32 type of differential pump was tested in field conditions in two modes: with two front working chambers (the delivery valves were removed from the rod chambers) and with two rear rod chambers (the delivery valves were removed from the front chambers). During the tests, the delivery Q and the pressure p were measured at various rates. Cylinders with a diameter of 115 mm and rods with a diameter of 45 mm were installed in the pump. The test results are reported in Tables 1 and 2, respectively.

The tests of the differential pump were performed in two wells. In one well (Table 1), the residual pressure was 10 MPa and in the other (Table 2), 5 MPa.
It is evident from Table 1 that the measured deliveries are in accord with the calculated with due regard for the volumetric efficiency. The average volumetric efficiency of the pump operating at the second-gear speed is 0.77 and at the thirdgear
speed, 0.65. With increase of the rotation speed, the volumetric efficiency decreases and pressure fluctuations rise from 5 to 30%. In spite of high delivery nonuniformity coefficient (ä = 1.86), the pump functioned satisfactorily.
It follows from Table 2 that the mea