Subsea mud pump and control system

Abstract

The invention is a subsea pump that includes pumping elements, each pumping element including a pressure vessel with a first and a second chamber therein and a separating member disposed between the first and second chambers. The first and second chambers are hydraulically connected to receive and discharge a hydraulic fluid and a drilling fluid, respectively. The separating member moves within the pressure vessel in response to a pressure differential between the first and second chambers. A hydraulic power supply is arranged to pump the hydraulic fluid to the first chamber of each of the pumping elements. A valve assembly is hydraulically coupled to the first chambers of the plurality of pumping elements and to the hydraulic power supply. Volume measurement devices are arranged to measure volumes of each of the first chambers and the second chambers. A valve controller is connected to the valve assembly and to the volume measurement devices, and the valve controller is arranged to control a rate and timing of a flow of the hydraulic fluid into the first chambers and a rate and timing of a flow of the hydraulic fluid out of the first chambers in response to the volume measurements. The valve controller is configured to maintain at least one of a substantially constant pump inlet pressure, a substantially constant pump discharge pressure, and a substantially constant total volume of the first chambers.

Description

BACKGROUND OF INVENTION

1. Field of the Invention

The invention relates generally to offshore drilling systems which are used to drilling subsea wells. More particularly, the invention relates to a subsea pump and an associated control system for use in offshore drilling systems.

2. Background Art

In conventional offshore drilling operations from, for example, a floating drilling vessel, a large diameter marine riser (e.g., a 21 inch marine riser) generally connects surface drilling equipment on the floating drilling vessel to a blowout preventer stack connected to a subsea wellhead located on the seabed. The marine riser is generally filled with drilling fluid (or “drilling mud”) so that a total hydrostatic pressure on a formation being drilled in a wellbore is determined by the hydrostatic pressure of the mud in the drilled wellbore (below the seabed) plus the hydrostatic pressure of the mud in the marine riser (above the seabed). In many cases, the total hydrostatic pressure of the “mud column” may exceed a fracture pressure of the formation being drilled. Accordingly, a large number of casing strings may need to be placed in the wellbore to protect the formation and maintain well control. In deep water drilling operations, the total cost of installing a large number of casing strings, combined with smaller oil and gas production rates possible through reduced diameter casing, can often result in wells which are uneconomical to drill and produce.

It has been determined that an important aspect of improving the economics and well control of deep water wells lies in reducing the hydrostatic pressure of the mud in the marine riser to that of a column of seawater, while at the same time filling the wellbore with drilling mud of sufficient weight to maintain well control. Various concepts have been presented in the past for achieving this goal, and the concepts can be grouped into two categories: mud lift drilling with a marine riser and riserless drilling.

Mud lift drilling with a marine riser typically includes a dual density mud gradient system, and the density of the mud return in the riser is generally reduced so that the hydrostatic pressure of the mud column in the riser, measured at the seabed, more closely matches that of seawater. The mud in the well bore remains weighted at a higher density to maintain proper well control. For example, U.S. Pat. No. 3,603,409 issued to Watkins et al. and U.S. Pat. No. 4,099,583 issued to Maus both disclose methods of using injected gas to reduce the density of the mud column in the marine riser, thereby reducing the hydrostatic pressure of the mud in the marine riser as measured at the seabed.

Riserless drilling generally includes eliminating the riser as a mud return path and replacing it with one or more small diameter mud return lines. For example, U.S. Pat. No. 4,813,495 issued to Leach discloses a system that eliminates the need for the marine riser and, as an alternative, uses a centrifugal pump to lift mud returns from the seafloor to the surface through a mud return line. A rotating apparatus isolates the mud in the wellbore annulus from seawater as the drillstring is run into and out of the wellbore.

U.S. Pat. No. 6,102,673, issued to Mott et al. and assigned to the assignee of the present invention, discloses a dual gradient riserless drilling system that uses a pressure actuated drillstring valve to control mud free fall, rotating and non rotating subsea diverters to isolate the mud in the wellbore from fluids, such as seawater, above the wellbore, a solids control system to control the size of solids in mud return lines, and a subsea positive displacement pump actively controlled in a coordinated manner with surface equipment on a drilling vessel to maintain the volume of mud in the wellbore.

Generally, the riserless drilling is preferred over the mud lift system because riserless drilling employs a pressure barrier between the wellbore and the surrounding environment. The pressure barrier allows the wellbore to be drilled in an “underbalanced” condition where formation pressures typically exceed the pressure of the drilling mud in the wellbore. Underbalanced drilling may significantly improve the rate of penetration of a drill bit and also helps reduce the risk of formation damage.

U.S. Pat. No. 6,102,673 issued to Mott et al discloses a subsea positive displacement pump with multiple pump elements, each pump element comprising a pressure vessel divided into two chambers by a separating member and powered by a closed hydraulic system using a subsea variable displacement hydraulic pump. The subsea positive displacement pump includes hydraulically actuated valves to ensure proper valve seating in the presence of, for example, cuttings from the drill bit that are present in mud returns from the wellbore. The hydraulically actuated valves also provide flexibility in valve timing (which is typically not available with conventional spring biased check valves) and provide quick valve response in high flow coefficient (Cv) arrangements necessary for high volume pumping (e.g., substantially high flow rates).

The subsea positive displacement pump disclosed in U.S. Pat. No. 6,102,673 issued to Mott et al is controlled by a unitary control module which receives the following signals: (1) position signals from a position indicator on the separating member in each pump element, wherein the position signals are converted into volume measurements; (2) flow and pressure signals from devices on a return side of the closed hydraulic system; (3) flow signals from a supply side of the hydraulic system (usually positioned proximate the variable displacement hydraulic pump); and (4) pressure signals from a mud suction pressure transducer.

Control signals from the control module: (1) control the operation of the flow control valve on the hydraulic fluid return to ensure that the flow rate from the variable displacement hydraulic pump is equal to the flow rate returning to the hydraulic reservoir; (2) operate the two hydraulic control valves and two hydraulically actuated mud valves on each pumping element to control the pumping rate of the subsea mud pump; and (3) control the flow rate of the variable displacement hydraulic pump. The control module algorithm is designed to provide “pulsationless” flow by precisely controlling the “phasing” of the multiple pumping elements to overlap both the fill and discharge cycles of the pumping elements.

The control system is difficult to precisely adjust because it has proven difficult to accurately model both the non-linear responses of many of the hydraulic components of the system and the wellbore hydraulic characteristics over time. In practice, significant load changes from a stable pump operating condition, such as step load changes of plus or minus fifty percent, have been found to cause instability in the system. Further, the response of the variable displacement hydraulic pump to the control signals, which is adequate at low and steady pumping rates, has proven to be inadequate at higher mud pump rates (e.g., pump rates above about 4-5 strokes per minute).

The subsea pump disclosed in U.S. Pat. No. 6,102,673 issued to Mott et al generally requires that the hydraulic power source be located proximate the subsea mud pumping elements with high flow capacity (e.g., high Cv piping between the hydraulic pump and the mud pumping elements) to minimize lag in the hydraulic response. This precludes, for example, using high pressure pumps located on the floating rig as a source of hydraulic power. Moreover, because the hydraulic valves controlling the mud pumping elements in the disclosed arrangement must have a high Cv to allow the mud pumps to operate at high flow rates, the disclosed control valve arrangement may be prone to hydraulic “water hammer” effects whenever the large bore valves open or close under differential pressure during the pumping cycle, especially at high pump rates.

It would be advantageous, therefore, to design a subsea mud pump and a coordinated control system that would enable stable, efficient operation of deep water drilling systems, including riserless drilling systems. It would also be advantageous to design a control scheme that insures that bottom hole pressure (BHP) is maintained whenever drilling mud pumps are stopped, for example, to add lengths of drillpipe to the drillstring (e.g., when “making a connection”).

Finally, it would be advantageous to design a control system that can compensate for drilling mud that has some degree of compressibility, whether because of the high hydrostatic pressures encountered in deepwater subsea operations (e.g., at depths of 10,000 feet, fresh water exhibits compressibility on the order of 2.5-3%) or because of entrained gas or volatile liquids/hydrocarbons that may be present in the drilling mud leaving the wellbore.

SUMMARY OF INVENTION

In one aspect, the invention is a subsea pump comprising a plurality of pumping elements. Each pumping element comprises a pressure vessel with a first and a second chamber therein and a separating member disposed between the first and second chambers. The first and second chambers are hydraulically coupled to receive and discharge a hydraulic fluid and a drilling fluid, respectively, wherein the separating member moves within the pressure vessel in response to a pressure differential between the first and second chambers. A hydraulic power supply is adapted to pump the hydraulic fluid to the first chamber of each of the pumping elements, and a valve assembly is hydraulically coupled to the plurality of pumping elements and to the hydraulic power supply. Volume measurement devices are adapted to measure volumes of each of at least one of the first and second chambers. A valve controller is operatively coupled to the valve assembly and to the volume measurement devices, and the valve controller is adapted to control a rate and timing of application of the hydraulic fluid to each of the first chambers and a rate and timing of discharge of hydraulic fluid therefrom in response to the measurements of volume so as to maintain at least one of a substantially constant pump inlet pressure, a substantially constant pump discharge pressure, and a substantially constant total volume of the first chambers.

In another aspect, the invention is a method for operating a subsea pump comprising a plurality of pumping elements. Each pumping element comprises a pressure vessel with a first and a second chamber therein and a separating member disposed between the first and second chambers. The first and second chambers are hydraulically coupled to receive and discharge a hydraulic fluid and a drilling fluid, respectively, wherein the separating member moves within the pressure vessel in response to a pressure differential between the first and second chambers. The method comprises measuring a volume of at least one of the first and second chambers, and applying hydraulic fluid to each of the first chambers at a selected rate and a selected time and enabling discharge of the hydraulic fluid at a selected time therefrom in response to the measurements of volume so as to maintain at least one of a substantially constant pump inlet pressure, a substantially constant pump discharge pressure, and a substantially constant total volume of the first chambers.

Other aspects and advantages of the invention will be apparent from the following description and the appended claims.

BRIEF DESCRIPTION OF DRAWINGS

FIG. 1A shows a simplified schematic view of an embodiment of the invention.

FIG. 1B shows a schematic diagram of an embodiment of a diaphragm pump module of the current invention.

FIG. 2 shows a graph of pump discharge pressure versus time for an embodiment of a pump system of the current invention where the pump system operated without compression control valves.

FIG. 3 shows a flow chart of an operation sequence of a pump in an embodiment of the invention.

FIG. 4A shows an embodiment of a hydraulic control system.

FIG. 4B shows another embodiment of a hydraulic control system.

FIG. 5 shows a graph of mud chamber volume versus linear displacement of the pump diaphragm in an embodiment of the invention.

FIG. 6 shows a graph of flow rate versus pressure in a mud suction line in an embodiment of the invention when pumps are stopped to make a connection.

FIG. 7 shows a graph of mud chamber volumes in a triplex pump embodiment of the invention during stable operation.

DETAILED DESCRIPTION

FIG. 1A shows a simplified schematic view of an embodiment of the invention. A subsea pump 200 comprises a hydraulic power supply 210 and pumping elements 220 . The hydraulic power supply 210 is hydraulically coupled to the pumping elements 220 by a hydraulic fluid supply line 230 . The hydraulic fluid supply line 230 is also coupled to a valve 240 . The valve 240 is operatively coupled to a valve controller 300 that is adapted to control a rate and time of application of hydraulic fluid to the pumping elements 220 . A flow of drilling fluid is supplied to the pumping elements through an inlet line 260 .

The flow of hydraulic fluid energizes the pumping elements 220 , and the flow of hydraulic fluid into the pumping elements 220 generates a flow of drilling fluid out of the pumping elements 220 through a discharge line 270 . Similarly, a flow of drilling fluid into the pumping elements 220 generates a flow of hydraulic fluid out of the pumping elements 220 . Hydraulic fluid flows out of the pumping elements 220 through a hydraulic fluid discharge line 280 . In some embodiments, a valve 290 is hydraulically coupled to the hydraulic fluid discharge line 280 and is operatively coupled to the valve controller 300 . The valve controller 300 is adapted to operate the valve 290 so as to control a rate of discharge of the hydraulic fluid from the pumping elements 220 .

By controlling, for example, the timing and rate of the application and discharge of hydraulic fluid to and from the pumping elements 220 , respectively, operating characteristics of the subsea pump 200 such as a pump inlet pressure, a pump discharge pressure, and a total volume of drilling fluid in the pumping elements 220 may be selectively controlled. These and other aspects of the invention are described in detail below.

FIG. 1B shows a detailed schematic diagram of an embodiment of a subsea diaphragm pump AA used in the invention. The subsea pump AA comprises three diaphragm pumping elements A, B, C connected by, for example, manifolds (not shown). The subsea pump AA shown in FIG. 1B essentially emulates a triplex positive displacement reciprocating pump. In some embodiments, a hydraulic fluid used to power the subsea pump AA and the pumping elements A, B, C comprises filtered seawater. However, other types of hydraulic fluid may be used to drive the subsea pump AA, and the use of filtered seawater is not intended to be limiting. Filtering of the seawater may be performed with equipment (not shown) located at the surface (e.g., on a drilling vessel (not shown)) or located proximate the seafloor.

Further, the embodiment shown in FIG. 1B includes three diaphragm pumping elements A, B, C. However, the number of pumping elements used with other embodiments of the invention may vary depending, for example, on factors such as a maximum flow rate required during operation, a desired redundancy of pumping elements, and packaging issues. Accordingly, embodiments of the invention may include, for example, from two diaphragm pumping elements or six diaphragm pumping elements. Moreover, linear piston-type pumps may also be used in some embodiments of the invention, and the examples below describing the operation of diaphragm pumps are not intended to be limiting.

Each of the pumping elements A, B, C comprises a vessel 1 a , 1 b , 1 c with two chambers. The chambers comprise mud chambers 2 a , 2 b , 2 c and hydraulic power chambers 3 a , 3 b , 3 c , where the chambers are typically separated by separation elements, one example of which is substantially impermeable pump diaphragms 4 a , 4 b , 4 c . In some embodiments, the diaphragms 4 a , 4 b , 4 c comprise an elastomeric material. However, the diaphragms 4 a , 4 b , 4 c may be formed from other materials, such as non-elastomeric materials or reinforced elastomeric materials, and the type of diaphragm material is not intended to be limiting.

In some embodiments of the invention, it is desirable to maintain a substantially constant inlet pressure (e.g., in the mud suction line 27 ). In other embodiments, it is desirable to maintain a substantially constant discharge pressure at a pump outlet (e.g., in the mud discharge line 28 ). If, for example, the inlet pressure is maintained at a substantially constant level, it is typical to let the discharge pressure “float” or vary during drilling operations. The opposite is also true when, for example, the discharge pressure is maintained at a substantially constant level. Various aspects of these embodiments of the invention are described in detail below. Note that operator preference, drilling conditions, etc. help determine which of the inlet pressure or the discharge pressure is maintained at a substantially constant level during drilling operations. Accordingly, the invention contemplates operating at all of the aforementioned conditions and incorporates the flexibility necessary to, for example, change from maintaining a substantially constant inlet pressure to maintain a substantially constant discharge pressure (and, if required, back again) during the process of drilling a well.

At the time interval shown in FIG. 1B , drilling mud has completely filled the mud chamber 2 a of the first pumping element A, mud is filling the mud chamber 2 b of the second pumping element B, and mud has been completely expelled from the mud chamber 2 c of the third pumping element C. During operation of this embodiment of the invention, mud flows from a mud suction line 27 (which is operatively connected to all three pumping elements A, B, C) into the diaphragm pumping module AA. Mud in the mud suction line 27 is generally a mud return from a wellbore (not shown) being drilled. For example, mud may be stored and processed (e.g., degassed, desilted, weighted, etc.) at the surface before being pumped (e.g., via surface pumps) downhole (e.g., through a drillstring comprising drillpipe and a bottom hole assembly (BHA)) into the wellbore. Mud then flows uphole through an annulus between the drillstring and walls of the wellbore and into the mud suction line 27 .

Mud from the mud suction line 27 then flows through actuated mud suction valves 9 a , 9 b , 9 c and into the mud chambers 2 a , 2 b , 2 c of the pumping module AA. After the mud chambers 2 a , 2 b , 2 c have been filled, mud may then be pumped from the mud chambers 2 a , 2 b , 2 c through actuated mud discharge valves 8 a , 8 b , 8 c and into a mud discharge line 28 . The mud discharge line 28 is typically connected to a mud return line (not shown) that is connected to mud handling and processing equipment (not shown) located at the water surface.

In some embodiments of the invention, the mud suction valves 9 a , 9 b , 9 c and mud discharge valves 8 a , 8 b , 8 c are power actuated valves of the type described in U.S. Pat. No. 6,102,673 issued to Mott et al. Power actuated valves are preferable, for example, when pumping mud returns from a drilled wellbore because the suction valves 9 a , 9 b , 9 c and the discharge valves 8 a , 8 b , 8 c may have to close and seal against large and irregularly shaped obstructions such as formation cuttings. Accordingly, power actuated valves are desirable because conventional spring biased check valves may be unable to close against such obstructions and thereby form an effective seal. However, conventional spring biased check valves may be used with embodiments of the invention. For example, spring biased check valves may be used with embodiments of the invention that use a diaphragm type mud pump of the type disclosed in U.S. Pat. No. 2,703,055 issued to Veth et al.

Hydraulic fluid is pumped into the hydraulic power chambers 3 a , 3 b , 3 c from a flow regulated hydraulic fluid source 23 through respective hydraulic inlet control valves 6 a , 6 b , 6 c . Hydraulic pressure in the hydraulic power chambers 3 a , 3 b , 3 c is monitored by respective hydraulic chamber pressure transducers 11 a , 11 b , 11 c . The inflow of hydraulic fluid moves the pump diaphragms 4 a , 4 b , 4 c and displaces the diaphragms 4 a , 4 b , 4 c so as to pump the mud out of the respective mud chambers 2 a , 2 b , 2 c . For example (referring to FIG. 1 B), when hydraulic fluid flows into the “upper” hydraulic power chambers 3 a , 3 b , 3 c , mud is forced out of the “lower” mud chambers 2 a , 2 b , 2 c and into the mud discharge line 28 .

In contrast, when the mud chambers 2 a , 2 b , 2 c are filling with mud, respective hydraulic outlet control valves 7 a , 7 b , 7 c are opened and hydraulic fluid in the hydraulic power chambers 3 a , 3 b , 3 c flows out through a discharge line 25 . Note that in some embodiments that use seawater as the hydraulic fluid, the discharge line 25 may dump the seawater hydraulic fluid into the ocean proximate the subsea pump AA. The seawater embodiments are advantageous in that additional equipment (such as a hydraulic fluid recirculation system (not shown)) is not required to further transport the seawater hydraulic fluid. However, other embodiments may include a hydraulic fluid recirculation system (not shown) attached to the discharge line 25 so that the hydraulic fluid is reusable by returning the hydraulic fluid to the surface. For example, some embodiments of the invention may use oil as the hydraulic fluid. The oil-based hydraulic fluid may be recirculated rather than dumped into the sea. The oil-based hydraulic fluid is also advantageous because a pump pressure required to pump the oil-based hydraulic fluid at depth is typically less than a pump pressure required to pump the seawater hydraulic fluid at a similar depth.

Substantially instantaneous positions of the pump diaphragms 4 a , 4 b , 4 c may be determined by position transducers 5 a , 5 b , 5 c attached to the pump diaphragms 4 a , 4 b , 4 c of each of the pumping elements A, B, C. In the embodiment shown in FIG. 1B , the position transducers 5 a , 5 b , 5 c are magnetorestrictive linear displacement transducers (LDT) of the type disclosed in U.S. Pat. No. 6,102,673 issued to Mott et al. However, other types of position transducers may be used to measure the absolute position of the diaphragm, including but not limited to linear variable differential transformers (LVDT) and ultrasonic measurement devices. Accordingly, the type of position transducer is not intended to limit the scope of the invention.

The position of pump diaphragms 4 a , 4 b , 4 c determined by the diaphragm position transducers 5 a , 5 b , 5 c is used by sequencing devices 21 a , 21 b , 21 c to determine when the pump diaphragms 4 a , 4 b , 4 c have reached the “end” or limit (e.g., the top or bottom) of their stroke. In addition, the diaphragm position information may be conveyed to personnel and equipment aboard the floating drilling vessel (not shown) and is used in the operation of a constant volume flow control system (D in FIG. 4 ) to control the flow regulated hydraulic power source 23 . Note that similar position transducers 5 a , 5 b , 5 c may be used with embodiments of the invention that use linear piston-type pumps.

In order to ensure substantially constant discharge pressures from the subsea pump AA, it is important to compress drilling mud to a desired discharge pressure before it is discharged from the pump. Drilling mud returns at the seabed are likely to be more compressible than drilling mud pumped by mud pumps on the surface. For example, mud pumped through the mud pumps at the surface is typically cleaned of large cuttings (e.g., shale), sand, silt, and fluid returns from the well such as oil and/or brine, and is degassed before it is returned to the mud pumps for recirculation into the wellbore. On the seabed, it is possible that drilling mud returned directly from the wellbore to the subsea pump AA contains quantities of entrained gas or volatile liquid petroleum fractions, and even small quantities of gas and/or volatile liquids may substantially increase the compressibility of the drilling mud. Furthermore, at the high hydrostatic pressures encountered in deepwater drilling (e.g., at 10,000 feet below the surface, hydrostatic pressure is about 4500 psi), even completely gas free water based drilling mud may compressible by 2-3% (with respect to a unit volume). Oil based drilling mud and certain drilling mud additives will typically be even more compressible than water based drilling mud.

Prior art subsea pumping relied on the belief that complete compression of the drilling mud could be achieved by properly controlling the hydraulic inlet valves. However, it has been determined that if a flow regulated hydraulic power source is used to finish the compression of the drilling mud prior to pumping the mud, it can result in negative pressure spikes in the mud chamber (e.g., mud chamber 2 a ) of the affected pump element (e.g., pump element A) and, as a result, can transmit the negative pressure spikes through the mud return line 28 (and, as a result, possibly damage other equipment).

FIG. 2 shows a graph of the discharge pressure versus time of a prior art subsea pump such as that shown in the Mott patent. The pump in FIG. 2 is in a duplex configuration (e.g., it comprises two diaphragm pumping elements), but the graph is typical of a pump with any number of diaphragm pumping elements. The graph shows that the pump typically operates at a relatively stable average pressure P avg . However, note the extremely large negative pressure spikes P spike that are generated at times T1, T2, T3, T4 when the mud discharge valves open.

At times T1-T4, if the mud in the mud chamber is compressible, the mud in the mud discharge line will flow into the mud chamber and cause the discharge pressure to drop suddenly (generating the pressure spikes P spike ) until the pressure in the mud chamber is equalized to the mud pressure in the mud discharge line.

In practice, it has been determined that negative pressure spikes are generally more severe at higher pump rates (e.g., at pump rates of approximately 4-5 strokes per minute or greater) because there is less time during the pump cycle for compressible mud to be compressed to the desired discharge pressure. In addition, high flow coefficient (Cv) piping is generally required for higher pump rates, and the high flow coefficient piping makes it difficult to precisely control the hydraulic inlet control valves at lower flow rates.

Referring again to FIG. 1B , the subsea pump AA of this embodiment of the invention comprises a hydraulic power source 24 (that is not flow regulated) and compression control valves 10 a , 10 b , 10 c coupled to respective pump chambers 3 a , 3 b , 3 c . The compression control valves 10 a , 10 b , 10 c have flow coefficients (Cv) on the order of 0.1 to 0.01 times the Cv of the hydraulic inlet control valves 6 a , 6 b , 6 c to help ensure smooth compression of the drilling mud in the mud chambers 2 a , 2 b , 2 c . For example, hydraulic fluid from the hydraulic inlet control valve 6 a , 6 b , 6 c flows into the respective hydraulic power chambers 3 a , 3 b , 3 c and displaces the respective pump diaphragms 4 a , 4 b , 4 c , thereby pressurizing the mud in the respective mud chambers 2 a , 2 b , 2 c . Then, after the mud has been pressurized, but before the mud has been released from the respective mud chambers 3 a , 3 b , 3 c through the mud discharge valve 8 a , 8 b , 8 c , the compression control valve 10 a , 10 b , 10 c opens briefly and allows pressurized hydraulic fluid from a non flow regulated hydraulic power source 24 to flow into the hydraulic chamber 3 a , 3 b , 3 c and thereby pressurize the mud in the mud chamber 2 a , 2 b , 2 c to substantially the same pressure as the desired mud discharge pressure.

Further, it has been determined that rapid opening of the hydraulic outlet control valves 7 a , 7 b , 7 c (as required by high pump stroke rates), combined with the relatively high flow coefficients (Cv) of the hydraulic outlet control valves 7 a , 7 b , 7 c , can cause severe hydraulic hammering of the system. Hydraulic hammering is produced by the “water hammer effect,” where a sudden release of high pressure fluid into, for example, a flow conduit that is at a lower pressure generates a hydraulic “shock wave” in the system. The hydraulic hammering may damage the system by, for example, fatiguing tubular joints, valves, etc. after repeated occurrences.

Accordingly, embodiments of the invention include decompression control valves 12 a , 12 b , 12 c that have flow coefficients (Cv) on the order of 0.01 to 0.1 times the Cv of the hydraulic outlet control valves 7 a , 7 b , 7 c . Activation of the decompression control valves 12 a , 12 b , 12 c produces a gradual reduction in pressure and helps ensure smooth discharge of the hydraulic fluid from the hydraulic power chambers 3 a , 3 b , 3 c without hydraulic hammering. For example, after the mud has been completely pumped from the mud chamber 2 a , 2 b , 2 c through the mud discharge valve 9 a , 9 b , 9 c and hydraulic inlet control valve 6 a , 6 b , 6 c is completely closed, the decompression control valve 12 a , 12 b , 12 c is opened to gradually relieve pressure from the hydraulic power chamber 3 a , 3 b , 3 c.

In some embodiments that use filtered seawater as the hydraulic fluid, the decompression control valves 12 a , 12 b , 12 c are vented to the sea. However, as previously explained, other arrangements are possible when, for example, the hydraulic fluid comprises a fluid other than seawater.

Both the compression 10 a , 10 b , 10 c and decompression 12 a , 12 b , 12 c control valves can be actuated for a selected period of time (for example, a fixed number of seconds or a fraction of the time required to complete a pump cycle), selectively actuated with reference to the pressure in the hydraulic power chamber 3 a , 3 b , 3 c measured by pressure transducers 11 a , 11 b , 11 c , or controlled by an algorithm that evaluates both time and pressure at any selected instant and actuates the valves accordingly.

In the embodiment shown in FIG. 1B , each diaphragm pumping element A, B, C is controlled by a sequencing device 21 a , 21 b , 21 c which receives data signals from different parts of the diaphragm pumping elements A, B, C, and provides control signals to the various control valves as shown in the Figure. Data are transmitted to the sequencing devices 21 a , 21 b , 21 c by, for example, a diaphragm position data link 20 a , 20 b , 20 c and by a hydraulic chamber pressure link 16 a , 16 b , 16 c.

The pump system operator can set various operational parameters via sequencing device data links 29 a , 29 b , 29 c . The operational parameters are described in detail in the description of FIG. 3 below. Control signals are transmitted from the sequencing devices 21 a , 21 b , 21 c to the various control valves by, for example, a decompression control valve data link 13 a , 13 b , 13 c , a hydraulic inlet control valve data link 14 a , 14 b , 14 c , a mud suction valve data link 15 a , 15 b , 15 c , a mud discharge valve data link 17 a , 17 b , 17 c , a decompression control valve data link 18 a , 18 b , 18 c , and a hydraulic outlet control valve data link 19 a , 19 b , 19 c . Moreover, the data and control signal links are understood to incorporate the necessary Input/Output (I/O) devices to accommodate the required signals to and from the sequencing devices 21 a , 21 b , 21 c . Accordingly, the type of I/O devices used with the sequencing device system is not intended to be limiting.

Each sequencing device 21 a , 21 b , 21 c may in turn be bussed together with the other sequencing devices through a sequencing device controller bus 22 so that the sequencing devices 21 a , 21 b , 21 c may exchange data with each other. For economy, ease of programming, maintenance, and ease of trouble-shooting, it is preferable that the sequencing devices 21 a , 21 b , 21 c be separate entities. In this manner, each sequencing device 21 a , 21 b , 21 c controls the operation of one diaphragm pumping element A, B, C. However, it will be understood by those skilled in the art that one sequencing device could be used to control all three diaphragm pumping elements A, B, C, which would allow the elimination of the sequencing device controller bus 22 because the function of the bus would be handled internally by the independent sequencing devices 21 a , 21 b , 21 c . Alternately, the sequencing devices 21 a , 21 b , 21 c could be separate “virtual machines” that are physically operated and controlled by, for example, a single computer.

The absolute diaphragm position data from the diaphragm position LDTs 5 a , 5 b , 5 c are transmitted by a diaphragm position data link 20 a , 20 b , 20 c to the sequencing devices 21 a , 21 b , 21 c and are compared to “full” and “empty” set points to determine if the mud chambers 2 a , 2 b , 2 c have reached the point where they are full or empty of drilling mud. The full or empty status for each pumping element is used to trigger steps in the logic sequences performed by the sequencing devices 21 a , 21 b , 21 c . The full and empty set points may be selected by the pump system operator or may be stored in a memory (not shown) of the sequencing devices 21 a , 21 b , 21 c . Further, the set points may be modified by the pump system operator at any time during the operation of the pump AA.

For example, the pump arrangement AA shown in FIG. 1B comprises three diaphragm pumping elements A, B, C. In the embodiment, an “A Full” status (that indicates that mud chamber 2 a is full of drilling mud) is an instruction for the sequencing device 21 a to begin the process of compressing drilling mud in the mud chamber 2 a (in diaphragm pumping element A), and for sequencing device 21 b to begin the process of filling mud chamber 2 b (in diaphragm pumping element B) with drilling mud. Similarly, a “C Empty” status (that indicates that mud chamber 2 c is empty) is an instruction for the sequencing device 21 a to begin pumping drilling mud from mud chamber 2 a (in diaphragm pumping element A), and the “B Empty” status is an instruction for the sequencing device 21 c to begin pumping drilling mud from mud chamber 2 c . The sequencing of the embodiment shown in FIG. 1B is covered in more depth in the detailed description of FIG. 3 below. Diaphragm position data from each of the diaphragm position transducers 5 a , 5 b , 5 c are sent from the pump elements A, B, C, through diaphragm position totalizer data links 26 a , 26 b , 26 c , to a constant volume flow control system E, as shown in FIG. 4 and as described in detail below.

FIG. 3 shows a simplified flow chart of a logic sequence BB that may be used by the sequencing device 21 a for pump element A (of the three pumping element embodiment shown in FIG. 1B ) in an embodiment of the invention. However, similar sequences could be used for systems that include more or fewer pump elements, and the description of the logic sequence BB shown in FIG. 3 is not intended to be limiting with respect to, for example, a number of pumps in an embodiment and/or a type of logic used to form the sequence. Further, it will be understood by those skilled in the art that, for example, an event driven logic sequence could be substituted for the boolean logic sequence BB shown in FIG. 3 .

Logic Sequence

The embodiment of the logic sequence BB shown in FIG. 3 is divided into two parts separated by the dashed line: a “Pump Filling Sequence” B 1 (shown above the dashed line), and a “Pump Emptying Sequence” B 2 (shown below the dashed line). In this embodiment, the logic sequence BB starts with pump element A (e.g., mud chamber 2 a ) empty of drilling mud and with diaphragm pumping element C (e.g., mud chamber 2 c ) full of drilling mud. However, the “start-up” condition is not intended to be limited by any single set of empty/full conditions for a single pumping chamber. For example, a similar “start-up” condition could include a check of an empty/full status of pump elements B and C.

After a START signal 30 (which may be initiated, for example, by a signal from the sequencing device 21 a or by the pump system operator), the logic sequence BB queries a pump A standby status register 32 at a pump A standby decision step 31 . Note that “standby status” is typically designated by the pump system operator. However, standby status could be designated by, for example, pump monitoring software or by downhole sensors. Accordingly, the method of designating standby status is not intended to be limiting.

For example, in some embodiments of the invention, if any of the pump elements A, B, C require service while the subsea pump AA is running, the standby status of the pump element A, B, C requiring service can be set to a “YES” value by a signal from the pump system operator. When a pump element A, B, C of, for example, a triplex pump arrangement, is set to “STANDBY,” the standby status will have the effect of temporarily converting the operation of the triplex subsea pump into a duplex pump (e.g., the standby setting will effectively remove the standby pump element from the pumping sequence).

If the pump element A has a “NO” value as its standby status, the logic sequence BB then queries a pump C fill status register 36 at a pump C fill status decision step 35 to determine whether pump element C (e.g., mud chamber 2 c ) is full of drilling mud. Note that a “FULL” set point 37 of the pump C fill status register 36 may be defined by personnel on the floating drilling vessel (not shown) or may be preprogrammed into the logic sequence BB.

If pump element C is not full (e.g., if the pump C fill status register 36 has a “NO” value), the logic sequence BB loops until it receives indication that pump element C (e.g., mud chamber 2 c ) is full of drilling mud (e.g., until the Pump C fill status register 36 is set to “YES”). When pump element C is full of drilling mud, the sequencing device 21 a sends signals 41 a , 41 b to open the mud suction valve 8 a and the hydraulic outlet control valve 7 a , respectively. Thereafter, mud begins flowing from the mud suction line 27 , through the mud suction valve 9 a , and into the mud chamber 2 a . As the mud chamber 2 a is filling, hydraulic fluid is displaced from the hydraulic power chamber 3 a and flows out of the hydraulic power chamber 3 a through the hydraulic outlet control valve 7 a into the discharge line 25 . The aforementioned process of filling mud chamber 2 a and simultaneously emptying hydraulic power chamber 3 a continues until a signal 39 from the diaphragm position transducer 5 a matches a pump A “FULL” status set point 38 . At this point, a pump A fill status 40 is set to a “FULL” value.

When the pump element A mud chamber 2 a is full, a signal from a pump A full decision step 42 starts a mud suction close timer 43 , which delays the logic sequence BB for a delay time 44 . After the delay time 44 has expired, a “close” signal 45 is transmitted to the mud suction control valve 9 a . Similarly, there is then a delay of delay time 47 initiated by a hydraulic outlet close timer 46 before a “close” signal 48 is sent to the hydraulic outlet control valve 7 a . Further, a delay of delay time 50 is initiated by a compression valve open timer 49 before an “open” signal 51 is sent to the compression valve 10 a . The delays are used to ensure that the drilling mud and hydraulic flow paths to the next chamber have been established prior to closing the currently filling or emptying chamber. Accordingly, the delays help prevent system damage that may occur if there is no flow path open on either the mud or hydraulic side of the system at a selected time.

Note that operation of the compression valve 10 a is shown within the pump filling sequence B 1 because the mud discharge valve 8 a is still closed. Compression of the mud in the mud chamber 2 a should be understood as a step to “condition” the mud to be pumped, rather than as a part of the pumping process (e.g., a part of the pump emptying sequence B 2 ).

The compression valve 10 a generally remains open until a pressure 55 in the hydraulic power chamber 3 a , as measured by a pressure transducer 11 a , reaches a predetermined set point 56 (as determined by a comparator 57 ), or until a Pump C status 60 a is “YES.” A condition satisfying an “OR” element 54 initiates transmission of a signal 58 to close the compression valve 10 a when the pressure 55 is achieved or the Pump C “YES” status 60 a has been achieved.

After the compression valve 10 a is closed, the logic sequence BB again polls the pump A standby status register 30 for pump element A at a pump A standby decision step 59 . Note that this means that both the pump emptying B 2 and pump filling B 1 sequences start with a determination of whether the particular diaphragm pump element A, B, or C is in active or standby status. Consequently, if a pump element is placed on standby status during operation, the pump element (that is placed on standby during operation) will finish the current half cycle (e.g., filling B 1 or emptying B 2 ), and thereafter that particular pump element will be bypassed in the pumping order of the subsea pump AA.

If pump element A is not on standby status, the sequencing device 21 a then polls a pump C fill status register 61 to determine if the mud chamber 2 c of pump element C is empty of drilling mud. The “empty” condition is defined by an empty set point 62 .

Note that the only external references in the logic sequence BB available to the sequencing devices 21 a , 21 b , 21 c for each pump element A, B, C is the “full” and “empty” status of its “partner” pump element in the sequence, which is polled twice during each pump stroke (e.g., once before the pump filling sequence B 1 and once before the pump emptying sequence B 2 ). For example (and to further describe the pumping element “partners”), the only external reference for the sequencing device 21 a for pump element A is the pump status register for pump element C. Similarly, the sequencing device 21 b for pump element B refers to the fill status register for pump element A, and the sequencing device 21 c for pump element C refers to the fill status register for pump element B. Note that while prior art diaphragm pump controls attempt to keep multiple diaphragm pumping elements strictly in a selected phase relationship, the sequencing devices 21 a , 21 b , 21 c of the current embodiments only keep the pump elements A, B, C in a selected operating sequence.

Referring again to FIG. 3 , if pump element C is empty (of drilling mud), the sequencing device 21 a sends a signal 63 a to ensure that the compression valve 10 a is closed, a signal 63 b to open the hydraulic inlet control valve 6 a , and a signal 63 c to open the mud discharge valve 8 a . At this point, hydraulic fluid flows from the flow regulated hydraulic power source 23 , through the hydraulic inlet valve 6 a , and into the hydraulic power chamber 3 a , thereby displacing the pump diaphragm 4 a and forcing drilling mud from the mud chamber 2 a out through the mud discharge valve 8 a , into the mud discharge line 28 and, subsequently, back to the floating drilling vessel (not shown) on the surface. The process of filling the hydraulic power chamber 3 a and emptying the mud chamber 2 a continues until the signal 39 from the diaphragm position transducer 5 a matches a pump A “EMPTY” fill status set point 64 , and the pump A fill status register 40 is then set to “EMPTY.”

When the pump A fill status register 40 is set to “EMPTY,” a pump A empty decision step 65 then sends a signal 66 to close the mud discharge valve 8 a . There is a delay of delay time 68 (controlled by the hydraulic inlet close timer 67 ) before a signal 69 is sent to close the hydraulic inlet control valve 6 a . There is then a further delay of delay time 71 (controlled by a decompression valve open timer 70 ) before a signal 72 is sent to open the decompression valve 12 a . When the decompression valve 12 a opens, hydraulic fluid is expelled (e.g., into the sea or into a hydraulic fluid recirculation chamber (not shown)) as pressure is gradually released from the hydraulic power chamber 3 a.

The decompression valve 12 a remains open until either a selected compression time 74 has passed, as determined by decompression open timer 73 , or a pressure 76 in the hydraulic power chamber 3 a , as measured by a pressure transducer 11 a , reaches a predetermined set point 77 as determined by a comparator 78 . A signal 79 to close the decompression valve 12 a is initiated by an “OR” function 75 that is connected to the decompression open timer 73 and the comparator 78 .

At this point, the second half (e.g., the pump emptying sequence B 2 ) of the logic sequence BB has been completed. Pump element A is now ready to begin the logic sequence BB again after being activated by the sequencing device 21 a.

Hydraulic Control System

FIG. 4A shows an embodiment of a hydraulic control system CC that can be used to regulate the flow of hydraulic fluid in and out of the pump elements (A, B, C in FIG. 1B ) in embodiments of the invention. Note that the flow rate of drilling mud is not directly measured by the hydraulic control system CC because drilling mud returns from a wellbore may be extremely erosive, and flow measurement of the erosive drilling mud can be unreliable. Alternatively, the flow rate of the drilling mud can be accurately derived from either diaphragm displacement data or from flow rate measurements of the relatively “cleaner” hydraulic fluid. One advantageous characteristic of the subsea pump AA is that the pump elements have a substantially 1:1 pumping ratio (e.g., where there is no hydraulic “slip”) so that the flow rate of hydraulic power fluid into the subsea pump AA is proportional to the flow rate of drilling mud out of the subsea pump AA.

The subsea pump AA shown in FIG. 4A is a simplified representation of the pump shown in FIG. 1 B. The subsea pump AA has inputs comprising the flow regulated hydraulic power source 23 , the non flow regulated hydraulic power source 24 , and the mud suction line 27 . The subsea pump AA also has outlets that comprise the discharge line 25 , the diaphragm position totalizer data links 26 a , 26 b , 26 c , and the mud discharge line 28 .

The subsea pump AA comprises a self contained, self controlled pumping unit which pumps drilling mud at a selected flow rate and pressure increase from the mud suction line 27 to the mud discharge line 28 , depending only on the hydraulic power supplied by the flow regulated hydraulic power source 23 , the non flow regulated power source 24 , and flow restriction, or throttling, applied to the discharge line 25 .

In the embodiment shown in FIG. 4A , hydraulic power is supplied by hydraulic fluid from the hydraulic power source 81 , which, in some embodiments, comprises a pump preferentially located on a floating drilling vessel (not shown). For example, positioning the hydraulic power source 81 on the floating drilling vessel (not shown) would allow using conventional drilling mud pumps as the hydraulic power source 81 , wherein the hydraulic fluid is conveyed from the surface to the subsea pump AA via a high pressure fluid conduit (not shown). Alternatively, the hydraulic power source 81 may comprise a submersible hydraulic pump (not shown) located proximate the subsea pump AA on the seabed. For example, in some embodiments the hydraulic power source 81 may comprise a submersible electric pump (not shown) that receives electric power from the floating drilling vessel (not shown) on the surface.

The pressure of the inflow of hydraulic fluid at a hydraulic manifold 93 is controlled by a hydraulic pressure control system D. The hydraulic pressure control system D is designed to maintain the hydraulic fluid at a higher pressure than the mud being discharged from the subsea pump AA to ensure that there are no negative pressure spikes in the mud discharge line 28 .

For example, the pump system operator can select a desired pressure differential between the mud discharge line 28 and hydraulic manifold 93 by controlling a pressure differential set point 94 . Typically, the selected pressure differential will be between 50 and 150 psi, and a pressure differential in this range is generally high enough to prevent negative pressure spikes in the system when the mud discharge valves ( 8 a , 8 b , 8 c in FIG. 1B ) are opened but low enough to avoid hydraulic hammering of the system when the hydraulic outlet control valves ( 7 a , 7 b , 7 c in FIG. 1B ) are opened.

Pressure in the hydraulic manifold 93 is regulated by a dump valve 85 , and the dump valve 85 is modulated by a dump valve controller 82 via a dump valve controller data link 82 a . The dump valve controller 82 operates in response to a differential pressure calculated by subtracting a value equal to a pressure in the mud discharge line 28 (typically measured by a mud discharge pressure transducer 84 , preferentially located on or proximate the subsea pump AA) from a value equal to a pressure in the hydraulic manifold 93 (typically measured by a pressure transducer 83 located on the hydraulic manifold 93 ), and then modulates the dump valve 85 to achieve the preselected differential pressure. However, the differential pressure described above may also be measured by subtracting pressures measured at alternative locations in the pumping system, and the location at which the differential pressure is calculated is not intended to be limiting.

Pressure modulation via the dump valve 85 helps ensure that the pressure in the hydraulic control system CC is greater than the pressure of the discharged mud so that when the mud discharge valves 8 a , 8 b , 8 c open during the pumping cycle, the mud inside the mud chambers 2 a , 2 b , 2 c is generally at a higher pressure than the mud in the mud discharge line 28 . Moreover, in some embodiments of the invention, the dump valve 85 may be modulated to maintain a substantially constant mud discharge pressure. In these embodiments, the dump valve controller 82 monitors the discharge pressure measured by the pressure transducer 84 and adjusts the dump valve 85 to maintain the substantially constant discharge pressure.

Hydrostatic pressure (e.g., ambient pressure at depth) is measured by a hydrostatic pressure transducer 95 and is communicated to the dump valve controller 82 via a hydrostatic pressure data link 95 a . If desired, the measured hydrostatic pressure can be used by the dump valve controller 82 as a reference pressure. For example, pressure in the hydraulic manifold 93 could be regulated at 150 psi above pressure in the mud discharge line 28 , but in no case less than the reference hydrostatic pressure. Also note that hydraulic fluid in the hydraulic manifold 93 flows directly into the subsea pump AA as the non flow regulated hydraulic power source 24 and through a total volume control valve 86 as the flow regulated hydraulic power source 23 .

The hydraulic pressure control system D is advantageous because, in prior art designs, a pressure of the hydraulic fluid is not controlled relative to a mud discharge pressure measured proximate a subsea positive displacement pump, which can result in mud discharge pressure “spikes” if the hydraulic pressure drops so that the mud in the discharge piping is at a higher pressure than the mud in the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) when a mud discharge valve ( 8 a , 8 b , 8 c in FIG. 1B ) opens during the pumping cycle. As described above with reference to FIG. 1B , if the pressure of the drilling mud in the discharge pipe 28 is greater than the pressure of the drilling mud in the mud chambers ( 2 a , 2 b , 2 c in FIG. 1 B), a back flow characterized by a negative pressure spike may result when drilling mud flows from the discharge pipe 28 into the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) when the mud discharge valves ( 8 a , 8 b , 8 c in FIG. 1B ) are opened.

Constant Volume Flow Control System

One of the fundamental control strategies used to control fluid flow both into and out of the subsea pump AA is to maintain a constant volume of drilling mud in the hydraulic power chambers 3 a , 3 b , 3 c at any selected time by regulating the flow of hydraulic fluid into the subsea pump AA. A net result is maintenance of a selected total volume of drilling mud in the mud chambers 2 a , 2 b , 2 c at any selected time.

The flow rate of hydraulic fluid in the flow regulated hydraulic power source 23 is regulated by the constant volume flow control system E, the goal of which is to maintain the total volume of drilling mud in the subsea pump AA (e.g., in the mud chambers ( 2 a , 2 b , 2 c in FIG. 1 B)). The embodiment shown in FIG. 4A uses a measurement of a total instantaneous volume of the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) to keep the pump elements (A, B, C in FIG. 1B ) in phase. A mathematical proof of the relationship between total mud volume and pump phase is discussed below in the section entitled Phase and Total Volume.

A pump volume totalizer 88 determines an instantaneous total volume of the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) by summing the instantaneous total mud volume of the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) based on positions of the individual pump diaphragms ( 4 a , 4 b , 4 c in FIG. 1B ) and an algorithm which relates diaphragm position to mud volume of the related mud chamber ( 2 a , 2 b , 2 c in FIG. 1 B). For an example of how to determine the instantaneous total mud volume of the mud chambers ( 2 a , 2 b , 2 c in FIG. 1 B), refer to the section below entitled Measuring Mud Chamber Volume.

Total volume control valve 86 receives control signals from total volume valve controller 87 via a valve control signal 87 a . The total volume valve controller 87 compares a total volume set point 97 a with an instantaneous volume of the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) supplied to the total volume valve controller 87 by the pump volume totalizer 88 via the total volume data link 88 a . If, for example, the instantaneous volume of mud in the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) is greater than the total volume set point 97 a (which generally indicates that the pump rate is too low), the total volume control valve 86 will be opened slightly, thereby increasing the pump rate of the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) and tending to bring the pump elements (A, B, C in FIG. 1B ) back into a desired phase relationship.

Alternative Embodiment

FIG. 4B shows another embodiment of the hydraulic control system CC.

The embodiment shown in FIG. 4B is similar to the embodiment shown in FIG. 4A except, for example, for the absence of a designated hydraulic pressure control system (e.g., hydraulic pressure control system D in FIG. 4 A). The embodiment in FIG. 4B essentially combines the hydraulic pressure control system (D in FIG. 4A ) and the constant volume control system (E in FIG. 4A ) into a unitary constant volume flow control system G.

As in the previous embodiment, the subsea pump AA is a simplified representation of the schematic diagram shown in FIG. 1 B. The subsea pump AA has the hydraulic power source 81 as an input. The subsea pump AA also has outlets that comprise the discharge line 25 , the diaphragm position totalizer data links 26 a , 26 b , 26 c , and the mud discharge line 28 . Note that the total volume control valve ( 86 in FIG. 4A ) has been eliminated and that the dump valve controller 82 has replaced the flow rate valve controller ( 87 in FIG. 4 A). The dump valve 85 now performs the function of regulating total volume in the system CC subject to control inputs from the dump valve controller 82 . An accumulator 98 may be used with the system to condition a flow of hydraulic fluid. For example, the accumulator 98 may be adapted to maintain a sufficiently high pressure in hydraulic flow lines 99 , 100 (e.g., to prevent negative pressure spikes in the system).

The flow rate of hydraulic fluid in the flow regulated hydraulic power source 23 is regulated by the constant volume flow control system E, the goal of which is to maintain a substantially constant total volume of drilling mud in the subsea pump AA (e.g., in the mud chambers ( 2 a , 2 b , 2 c in FIG. 1 B)) at any selected time. The embodiment shown in FIG. 4B uses a measurement of a total instantaneous volume of the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) to keep the pump elements (A, B, C in FIG. 1B ) in sequence. As in the previous embodiment, refer to the section entitled Phase and Total Volume below for a mathematical proof of the relationship between total mud volume and phase.

The pump volume totalizer 88 determines an instantaneous total volume of the mud chambers 2 a , 2 b , 2 c by summing the instantaneous total mud volume of the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) based on positions of the individual pump diaphragms ( 4 a , 4 b , 4 c in FIG. 1B ) and an algorithm which relates diaphragm position to mud volume of the related mud chamber ( 2 a , 2 b , 2 c in FIG. 1 B). For an example of how to determine the instantaneous total mud volume of the mud chambers ( 2 a , 2 b , 2 c in FIG. 1 B), refer to the section below entitled Measuring Mud Chamber Volume.

The dump valve 85 receives control signals from the dump valve controller 82 via the valve control signal 87 a . The dump valve controller 82 compares the total volume set point 97 a with an instantaneous volume of the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) supplied to the dump valve controller 87 by the pump volume totalizer 88 via the total volume data link 88 a . If, for example, the instantaneous volume of mud in the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) is greater than the total volume set point 97 a (which generally indicates that the pump rate is too low), the dump valve 85 will be opened slightly, thereby increasing the pump rate of the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) and tending to bring the pump elements (A, B, C in FIG. 1B ) back into a desired phase relationship.

Hydrostatic pressure is measured by the hydrostatic pressure transducer 95 and is communicated to the dump valve controller 82 via a hydrostatic pressure data link 95 a . If desired, the measured hydrostatic pressure can be used by the dump valve controller 82 as a reference pressure. For example, pressure in the hydraulic manifold 93 could be regulated at 150 psi above pressure in the mud discharge line 28 , but in no case less than the reference hydrostatic pressure.

It has been determined that, in some embodiments of the invention, the accumulator 98 and the inherent compressibility of the hydraulic fluid enables adequate conditioning (e.g., compression, filtering, etc.) to pressurize the hydraulic fluid to a sufficient level relative to the mud being discharged from the subsea pump AA to ensure that there are no negative pressure spikes in the mud discharge line 28 . Note that although the accumulator 98 is shown to be a separate item in FIG. 4B , the accumulator 98 may be included in, for example, the hydraulic flow lines 99 , 100 . Accordingly, the embodiment shown in FIG. 4B is not intended to limit the location or type of accumulator 98 that may be used with the hydraulic control system CC.

Fill-rate Flow Control System

In some embodiments of a pump according to the invention, the rate at which the hydraulic fluid is discharged from the pump elements (A, B, C in FIG. 1B ) controls a pump rate of the subsea pump AA. Subsea pumps AA used in the various embodiments of the invention have a 1:1 ratio between a volume of hydraulic fluid and a volume of drilling mud that flow through the subsea pump AA in a selected cycle. Further, the individual pump elements (A, B, C in FIG. 1B ) are maintained in proper sequence by the sequencing devices ( 21 a , 21 b , 21 c in FIG. 1 B). Accordingly, because the total volume of the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) remains substantially constant, the flow rate of the drilling mud can be precisely controlled (e.g., with very little control error) by controlling the discharge rate of the hydraulic fluid (e.g., the flow rate of the hydraulic fluid discharge can be controlled to equal the desired flow rate of the drilling mud).

The hydraulic fluid discharged from the subsea pump AA passes through the discharge line 25 . The discharge flow rate is measured by a discharge flow meter 92 . The flow rate in the discharge line 25 is regulated by a discharge control valve 89 . The discharge control valve 89 is, in turn, controlled by a discharge controller 90 , and the discharge controller 90 uses data received through an inlet pressure data link 91 a (from an inlet pressure transducer 91 that measures the pressure of the drilling mud in the mud suction line 27 ) or through a flow data link 92 a (from the discharge flow meter 92 ).

There are three particular drilling mud flow rate modes that are typically required during subsea mudlift drilling operations: a constant annulus pressure mode, a constant flow rate mode, and a “make connection” mode. The constant annular pressure mode is designed to maintain a substantially constant pressure at the subsea pump inlet regardless of flow rate. Assuming that the hydrostatic and friction pressures in the wellbore annulus are generally constant, a substantially constant inlet pressure results in a substantially constant bottom hole pressure (BHP), which is required to maintain well control. In the constant pressure mode, the control system CC adjusts the pump rate of the subsea pump AA to maintain pump inlet pressure at a substantially constant level. For example, if the wellbore annulus pressure starts to rise above a preselected pressure set point, the subsea pump AA must operate at a higher stroke rate (e.g., pump at a higher flow rate) to maintain the inlet pressure at a preselected level. Moreover, if the inlet pressure drops below another preselected pressure set point, the stroke rate of the subsea pump AA must be decreased to maintain the inlet pressure at the preselected level.

In contrast, the constant flow rate mode seeks to maintain a constant volumetric flow rate from the wellbore annulus regardless of wellbore pressure. The constant flow rate mode is analogous to the “pulsation free” pumping method disclosed in U.S. Pat. No. 6,102,673 issued to Mott et al. The ability to pump at a substantially constant flow rate is required for selected well control activities used in dual gradient drilling systems.

The “make connection” mode is used when, for example, surface pumps must be stopped to add more drillpipe to a drillstring (e.g., when drilling personnel on the floating drilling vessel “make a connection”). The make connection mode is described in detail below.

When the subsea pump AA is operating (in a substantially steady state mode during, for example, normal drilling operations), the bottom hole pressure (BHP, or annular pressure at a drill bit) may be defined as:

BHP=P HYD +P AFP +P INLET   (1)

where P HYD is a hydrostatic pressure of a mud column from the drill bit to the pump inlet, P AFP is an annular friction pressure generated by resistance to drilling mud flow in the annulus between the drillpipe and the walls of the wellbore (not shown), and P INLET is a pressure in the wellbore annulus measured at the suction line 27 to the subsea pump AA. Note that the relationship between P AFP and flow rate will usually be linear at the flow rates expected in normal drilling operations (e.g., return flow of drilling mud in the wellbore annulus will generally be laminar). Further, during normal drilling operations, the largest contribution to BHP will be the hydrostatic pressure of the mud column in the wellbore annulus above the drill bit.

In order to prevent the well from flowing (e.g., to prevent formation fluids from entering the wellbore and generating a “kick”), the BHP must generally be maintained at a known, constant level during the entire drilling process, including during the “make connection” process. If the surface pumps on the floating drilling vessel are stopped, the contribution of P AFP is lost. The pressure drop attributable to the loss of P AFP must be compensated for in order to maintain a substantially constant BHP so as to maintain proper well control. Because the hydrostatic pressure of mud in the drillpipe (P HYD ) is substantially constant, the inlet pressure (P INLET ) must be increased to compensate for the loss P AFP .

FIG. 6 shows a graph of flow rate versus inlet (suction) pressure in the mud suction line ( 27 in FIG. 4 ). During normal drilling operations, suction pressure (P INLET ) in the mud suction line ( 27 in FIG. 4 ) is maintained at a suction pressure set point 100 . P INLET 101 required to maintain the BHP at zero flow rate is greater than the suction pressure set point 100 by an amount equal to the annular friction pressure (P AFP ) 102 . P AFP is linear with respect to flow rate, so P INLET required to maintain a substantially constant BHP at any flow rate may be graphically represented by a line 103 drawn between the suction pressure set point 100 and P INLET required to maintain BHP at zero flow rate 101 . Accordingly, whenever surface pumps (not shown) are turned off to “make a connection,” P INLET must be maintained at a pressure that graphically falls to the right of the line 103 in order to maintain a substantially constant BHP.

In practice, because there are several “throttles” or controls in a fluid path between the surface pumps (not shown) and the wellbore annulus (not shown) (including, for example, nozzles (not shown) in the drill bit (not shown)), if the surface pumps (not shown) are stopped abruptly, P INLET and flow rate will drop to the left of the line 103 . If this occurs, then the BHP at zero flow rate will be at a level below P INLET required to maintain a substantially constant BHP (e.g., lower than the minimum P INLET 101 required to maintain BHP at zero flow rate). This situation may be avoided by implementing control schemes.

In order to maintain at least the minimum P INLET 101 , the BHP can be automatically controlled during a surface pump shut down process. For example, when the “make connection” mode is activated (e.g., either manually or automatically upon initiation of surface pump shutdown), the surface pumps (not shown) are stopped. The flow rate valve controller ( 90 in FIG. 4 ) may be instructed to maintain a selected P INLET to the right of the line 103 . The suction pressure set point is continuously and automatically updated to reflect the current mud flow rate as measured by the flow meter 92 . For example, if the subsea pump (AA in FIG. 4 ) slows at a faster rate than the surface pumps (not shown), the subsea pump (AA in FIG. 4 ) may be used to maintain the BHP by compensating for the loss of P AFP . This embodiment of an automatically controlled shut down procedure is shown by as curve 104 in FIG. 6 .

Alternatively, the loss of P AFP may be compensated for just prior to shutting down the surface pumps (not shown) by controlling the flow rate valve controller ( 90 in FIG. 4 ) to raise the inlet pressure (P INLET ) by a required suction pressure offset 105 to a selected offset pressure 106 before the surface pumps (not shown) are shut down to “make a connection.” The offset pressure 106 value may be determined by a linear function relating a change in flow rate to a change in P INLET with respect to time. For example, one embodiment of a linear function relating the change in flow rate to the change in suction pressure over time is represented as a curve 107 . In this embodiment, when the surface pumps (not shown) are shut down, P INLET follows the curve 107 so that, as the flow rate of the surface pumps (not shown) decreases in a controlled manner, the curve 107 to P INLET 101 required to maintain a substantially constant BHP at zero flow rate. Note that this control scheme must be carefully monitored because when P INLET is raised momentarily to the selected offset pressure 106 , the BHP also increases accordingly, which may induce a risk of fracturing the formation in open hole intervals of the wellbore.

Another advantageous method would be to combine the two previously described methods by, for example, automatically raising P INLET by some small amount during drilling operations in anticipation of shutting down the surface pumps (not shown) to make a connection (e.g., this essentially involves creating a BHP “safety margin” a selected level above the formation pressure). Next, a control algorithm could be implemented (as described above) to automatically control BHP during surface pump shutdown. This method avoids a sudden increase in BHP that may be experienced when achieving the desired offset pressure 106 .

Phase and Total Volume

FIG. 7 shows a diagram depicting time varying mud volumes in the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) of the subsea pump (AA in FIG. 1B ) (during, for example, steady state operation). Curves 108 , 109 , and 110 show depict mud volume in the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) over time, respectively. Vmin represents a minimum volume of each mud chamber ( 2 a , 2 b , 2 c in FIG. 1 B), and Vmax represents a maximum volume of each mud chamber ( 2 a , 2 b , 2 c in FIG. 1 B). A last measured fill time for a mud chamber ( 2 b in FIG. 1B ) (where a fill cycle of the mud chamber ( 2 b in FIG. 1B ) is represented by curve 109 ) is represented by Tf, a compression time for the mud chamber ( 2 b in FIG. 1B ) is represented by Tc, and a decompression time for the mud chamber ( 2 b in FIG. 1B ) is represented by Td. Note that the diagram in FIG. 7 is idealized to the extent that the effect of volume change during compression and decompression are not considered. These effects have no effect on the idealized calculations because both compression and decompression times shown in FIG. 7 include “wait” times and thereby balance any small volumetric changes within a single cycle.

At time T, the mud volumes of the mud chamber ( 2 a , 2 b , 2 c in FIG. 1B ) are:

Volume 2 a =V min  (2)

Volume 2 b =V max  (3) $$ Volume ⁢   ⁢ 2 ⁢ c = V ⁢   ⁢ min + Tc ⁡ ( V ⁢   ⁢ max - V ⁢   ⁢ min Tf ) ( 4 )

where the volume of mud chamber 2 c equals the minimum volume plus the elapsed time Tc multiplied by the slope of the curve 110 . Equation (4) may be rewritten as: $$ Volume ⁢   ⁢ 2 ⁢ c = V ⁢   ⁢ min + Tc Tf ⁢ ( V ⁢   ⁢ max - V ⁢   ⁢ min ) ( 5 )

and a total volume at time T can be expressed as: $$ Total ⁢   ⁢ Volume ⁢   ⁢ at ⁢   ⁢ T = Vt = V ⁢   ⁢ min + ( 1 - Tc Tf ) ⁢ V ⁢   ⁢ min + ( 1 + Tc Tf ) ⁢ V ⁢   ⁢ max . ( 6 )

Moreover, phase (Φ) may be defined as a difference between the compression and the decompression times, normalized by the fill time: $$ Φ = Tc - Td Tf . ( 7 )

Further, because:

Tc+Td=Tf   (8)

then: $$ Φ = 1 - Tc Tf . ( 9 )

By substituting Equation (9) into Equation (6): $$ Vt = 3 2 ⁢ ( V ⁢   ⁢ max + V ⁢   ⁢ min ) + Φ 2 ⁢ ( V ⁢   ⁢ max ⁢   -   ⁢ V ⁢   ⁢ min ) ⁢   ⁢ and ( 10 ) Φ = 2 ⁢ Vt - 3 ⁢ ( V ⁢   ⁢ max - V ⁢   ⁢ min ) V ⁢   ⁢ max - V ⁢   ⁢ min . ( 11 )

Accordingly, once the values of Vmin and Vmax are selected, there is a direct linear correlation between the steady state values of total volume (Vt) and phase (Φ). As long as the total volume in the system is being controlled during drilling operations, the pump cycle should not shift out of phase.

Measuring Mud Chamber Volume

In some embodiments, the pump diaphragm ( 4 a , 4 b , 4 c in FIG. 1B ) may be designed to “roll” along sides of the vessel ( 1 a , 1 b , 1 c in FIG. 1B ) when the diaphragms ( 4 a , 4 b , 4 c in FIG. 1B ) are displaced, contrasting with designs where diaphragms comprise tightly stretched membranes in which the volume displacement of the diaphragm is limited by the maximum allowable strain of the diaphragm material. The rolling of the diaphragms ( 4 a , 4 b , 4 c in FIG. 1B ) enables the pump elements (A, B, C in FIG. 1B ) to have larger effective displacements (e.g., a higher percentage of the volume of the vessels ( 1 a , 1 b , 1 c in FIG. 1B ) can be displaced with each “stroke” of the diaphragm ( 4 a , 4 b , 4 c in FIG. 1 B)), and provides improved fatigue life for the diaphragms ( 4 a , 4 b , 4 c in FIG. 1 B), both of which are important aspects of some embodiments of the invention. For example, it has been determined that if the drilling mud being used in drilling operations is very compressible (e.g., the drilling mud includes entrained gas or volatile liquids), the diaphragm should be able to be fully stroked in the mud discharge direction to achieve a sufficient compression ratio to move the gas and/or volatile liquids through the pump without the pump becoming gas locked.

Some embodiments of the invention use diaphragm pump elements similar to diaphragm type pulsation dampers such as those disclosed in U.S. Pat. Nos. 2,757,689, 2,804,884, 3,169,551, 3,674,053, and 3,880,193, all assigned to the assignee of the present invention. The diaphragms disclosed in these references are generally in a fully “unfolded” position when a pump element is empty of drilling mud. Diaphragms according to these designs help avoid gas lock that may be caused by compressible fractions of drilling mud. Other embodiments comprise diaphragms such as those disclosed in U.S. Pat. No. 4,755,111, where a thickness of the diaphragm tapers from a thickest portion near edges of the diaphragm to a thinnest portion near a middle of the diaphragm. These diaphragms are stiffest in bending near the edges and less stiff near the middle, and this design encourages the diaphragm to roll back on itself (rather than simply bending back and forth) during displacement. Further, other types of diaphragms may be used with the invention, and the type of diaphragm is not intended to be limiting.

When rolling diaphragms are used in embodiments of the invention, volume measurement is complicated by the fact that the volume displaced by the diaphragm is a nonlinear function of the linear displacement of the diaphragm as measured by, for example, the diaphragm LDT. Further, because the diaphragm rolls differently depending upon the direction of displacement (e.g., when the pump element is either filling with mud or discharging mud), the function relating volume displacement to lineal displacement is “path dependent.”

FIG. 5 shows a curve defining the relationship between volume displaced and linear displacement of the diaphragm (as measured by an LDT) for a nominal 20 gallon displacement diaphragm pump element (such as pump elements A, B, C in FIG. 1 B). Note that a fill curve and an emptying curve diverge substantially proximate a middle of a stroke into path dependent functions. It has been determined that the path dependent curves are repeatable and that they can be reliably modeled mathematically with, for example, as few as four equations. The following discussion describes equations derived for four curves 51 , 52 , 53 , 54 that model volume displaced versus linear displacement of the pump element diaphragm referenced above.

A lower substantially linear segment 51 may be modeled with the following equation:

y= 1.75 x−a   (12)

where “y” is a volume displaced, “x” is a linear displacement, and “a” is a coefficient related to an output of the LDT.

A nonlinear segment of a filling curve 52 may be expressed as:

y= 0.085 x 2 +3 x−b.   (13)

An upper substantially linear segment 53 may be expressed as:

y= 1.4 x−c.   (14)

Finally, a nonlinear segment of an emptying curve 54 may be modeled as:

y= 0.125 x 2 −2.5 x+d.   (15)

Note that “b,” “c,” and “d” are also coefficients related to the output of the LDT.

Modeling functions for other sizes of torispherical-type diaphragm pumping elements would be similar to equations (12)-(15) above, but the functions relating linear displacement to volume displaced for any size and type of rolling diaphragm pump element must generally be determined separately by empirically measuring the displaced volume per length of diaphragm stroke and fitting a function to the measured curve by, for example, regression or other curve fitting techniques known in the art. Moreover, other methods, such as look-up tables, may be used in determining instantaneous volume measurements. If, for example, linear piston-type pumps are used in embodiments of the invention, volume calculations are much simpler and are known in the art.

In practice, equations such as equations 51 - 54 may be used to calculate instantaneous mud chamber volumes. First, a determination must be made relating to whether the mud chamber is in a filling mode or a discharge mode. This determination may be made, in some embodiments, by evaluating a status of the mud discharge valves. For example, if the mud discharge valves ( 8 a , 8 b , 8 c in FIG. 1B ) are open, then the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) are emptying, and the functions governing sections 53 , 54 , and 51 of the curve shown in FIG. 5 are applicable. If, in contrast, the mud discharge valves ( 8 a , 8 b , 8 c in FIG. 1B ) are closed, the mud chambers ( 2 a , 2 b , 2 c in FIG. 1B ) are filling, and the functions governing sections 51 , 52 , and 53 are applicable. The state of the mud discharge valves can be communicated from the sequencing devices ( 21 a , 21 b , 21 c in FIG. 1B ) to the pump volume totalizer ( 88 in FIGS. 4A and 4B ) by the sequencing device data links ( 29 a , 29 b , 29 c in FIG. 1 B).

While the invention has been described with respect to a limited number of embodiments, those skilled in the art, having benefit of this disclosure, will appreciate that other embodiments can be devised which do not depart from the scope of the invention as disclosed herein. Accordingly, the scope of the invention should be limited only by the attached claims.

Claims

1. A subsea pump comprising:a plurality of pumping elements, each pumping element comprising a pressure vessel with a first and a second chamber therein and a separating member disposed between the first and second chambers, the first and second chambers being hydraulically coupled to receive and discharge a hydraulic fluid and a drilling fluid, respectively, wherein the separating member moves within the pressure vessel in response to a pressure differential between the first and second chambers; a hydraulic power supply adapted to pump the hydraulic fluid to the first chamber of each of the pumping elements; a valve assembly hydraulically coupled to the first chambers of the plurality of pumping elements and to the hydraulic power supply; volume measurement devices adapted to measure volumes of each of at least one of the first and second chambers; and a valve controller operatively coupled to the valve assembly and to the volume measurement devices, the valve controller adapted to control a rate and timing of application of the hydraulic fluid to each of the first chambers and a timing of discharge of hydraulic fluid therefrom in response to the measurements of volume so as to maintain at least one of a substantially constant pump inlet pressure, a substantially constant pump discharge pressure, and a substantially constant total volume of the first chambers.

2. The subsea pump of claim 1, wherein the valve assembly comprises a dump valve adapted to control a pressure of the hydraulic fluid, the pressure of the hydraulic fluid adjustable to select the rate of application.

3. The subsea pump of claim 1, further comprising at least one sequencing device operatively coupled to the volume measurement devices and the valve assembly, the at least one sequencing device adapted to determine a fill status of the first and second chambers from the volume measurements and to operate inlet and outlet valves connected to the first chambers and second chambers at selected times so as to enable filling and emptying of the first and second chambers with hydraulic fluid and drilling fluid, respectively.

4. The subsea pump of claim 1, wherein the plurality of pumping elements comprise diaphragm pumps.

5. The subsea pump of claim 4, wherein each separating member comprises a rolling diaphragm.

6. The subsea pump of claim 1, wherein each separating member comprises a piston.

7. The subsea pump of claim 1, wherein the volume measurement devices comprise position measuring sensors operatively coupled to each of the separating members.

8. The subsea pump of claim 7, wherein the position measuring sensors comprise magnetostrictive transducers.

9. The subsea pump of claim 7, wherein the volume measurement devices comprise an integrator.

10. The subsea pump of claim 1, wherein the hydraulic power supply comprises a submersible pump disposed proximate the subsea pump on the seafloor.

11. The subsea pump of claim 1, further comprising compression valves hydraulically coupled to the hydraulic power supply and to each of the first chambers, the valve controller being operatively coupled to the compression valves and adapted to open the compression valves and apply hydraulic fluid at a selected rate and time after the second chambers are substantially full of drilling fluid so as to pressurize the drilling fluid in the second chambers to a selected pressure substantially equal to a pressure in a drilling fluid discharge line hydraulically coupled to the second chambers, the valve controller adapted to close the compression valves after pressurization and before opening hydraulic fluid inlet valves coupled to the first chambers and to a hydraulic fluid inlet line and drilling fluid outlet valves coupled to the second chambers and to the drilling fluid discharge line.

12. The subsea pump of claim 11, wherein the hydraulic power supply comprises a flow regulated hydraulic power supply hydraulically coupled to the hydraulic fluid inlet valves and a non-flow regulated hydraulic power supply hydraulically coupled to the compression valves, the flow regulated hydraulic power supply comprising a dump valve adapted to control a pressure of hydraulic fluid supplied therefrom.

13. The subsea pump of claim 1, further comprising decompression valves hydraulically coupled to the hydraulic power supply and to each of the first chambers, the valve controller being operatively coupled to the decompression valves and adapted to open the decompression valves and release hydraulic fluid at a selected rate and time after the first chambers are substantially full of hydraulic fluid so as to depressurize the hydraulic fluid in the first chambers to a selected pressure substantially equal to a pressure in a hydraulic fluid discharge line hydraulically coupled to the first chambers, the valve controller adapted to close the decompression valves after depressurization and before opening hydraulic fluid outlet valves coupled to the first chambers and to the hydraulic fluid discharge line and drilling fluid inlet valves coupled to the second chambers and to a drilling fluid inlet line.

14. The subsea pump of claim 1, wherein the valve controller is adapted to maintain a selected differential pressure between hydraulic fluid supplied by the hydraulic power source and drilling fluid in a drilling fluid discharge line so as to reduce pressure surges into the second chambers when drilling fluid outlet valves are opened.

15. The subsea pump of claim 14, wherein the differential pressure is selected from a range comprising approximately 50 psi to approximately 150 psi.

16. The subsea pump of claim 1, wherein the valve controller is adapted to maintain a selected differential pressure between hydraulic fluid supplied by the hydraulic power source and drilling fluid in a drilling fluid discharge line so as to reduce hydraulic hammering when hydraulic fluid outlet valves are opened.

17. The subsea pump of claim 16, wherein the differential pressure is selected from a range comprising approximately 50 psi to approximately 150 psi.

18. The subsea pump of claim 1, further comprising an accumulator hydraulically coupled to a hydraulic fluid supply line, the hydraulic fluid supply line hydraulically coupling the hydraulic power supply to each of the first chambers, the accumulator adapted to minimize pressure fluctuations in the hydraulic fluid supply line.

19. The subsea pump of claim 1, further comprising a pressure transducer operatively coupled to the valve controller and adapted to measure a hydrostatic pressure proximate the subsea pump, the valve controller adapted to maintain the hydraulic fluid in a hydraulic fluid supply line at a pressure at least equal to the measured hydrostatic pressure.

20. A method for operating a subsea pump comprising a plurality of pumping elements, each pumping element comprising a pressure vessel with a first and a second chamber therein and a separating member disposed between the first and second chambers, the first and second chambers being hydraulically coupled to receive and discharge a hydraulic fluid and a drilling fluid, respectively, wherein the separating member moves within the pressure vessel in response to a pressure differential between the first and second chambers, the method comprising:measuring a volume of at least one of the first and second chambers; and applying hydraulic fluid to each of the first chambers at a selected rate and time, and enabling discharge of the hydraulic fluid at a selected time therefrom in response to the measurements of volume so as to maintain at least one of a substantially constant pump inlet pressure, a substantially constant pump discharge pressure, and a substantially constant total volume of the first chambers.

21. The method of claim 20, wherein the rate and timing of the application of the hydraulic fluid is controlled to maintain a substantially constant total volume of the drilling fluid in the second chambers.

22. The method of claim 20, wherein the rate and timing of the application of the hydraulic fluid is controlled to maintain a selected pressure differential between the hydraulic fluid applied to the first chambers and drilling fluid in a drilling fluid discharge line.

23. The method of claim 20, wherein applying hydraulic fluid comprises determining a fill status of the first and second chambers from the volume measurements and operating inlet and outlet valves connected to the first chambers and second chambers at selected times in response to the fill status so as to enable filling of the first chambers with hydraulic fluid and emptying of drilling fluid from the second chambers.

24. The method of claim 20 wherein enabling discharge of hydraulic fluid comprises determining a fill status of the first and second chambers from the volume measurements and operating inlet and outlet valves connected to the first chambers and second chambers at selected times in response to the fill status so as to enable emptying of hydraulic fluid from the first chambers and filling of the second chambers with drilling fluid.

25. The method of claim 20, further comprising applying the hydraulic fluid to the first chambers at a selected rate and time after the second chambers are substantially full of drilling fluid so as to pressurize the hydraulic fluid in the first chambers and the drilling fluid in the second chambers to a selected level substantially equal to a pressure of drilling fluid in a drilling fluid discharge line hydraulically coupled to the second chambers before enabling an inflow of hydraulic fluid into the first chambers and an outflow of drilling fluid from the second chambers.

26. The method of claim 25, further comprising delaying the application of the hydraulic fluid by a selected time so as to ensure that a flow path has been established between a hydraulic fluid supply line and the first chambers.

27. The method of claim 25, further comprising delaying the inflow of hydraulic fluid by a selected time so as to ensure that a flow path has been established between a hydraulic fluid supply line and the first chambers.

28. The method of claim 25, further comprising delaying the outflow of the drilling fluid by a selected time so as to ensure that a flow path has been established between the second chambers and the drilling fluid discharge line.

29. The method of claim 20, further comprising releasing the hydraulic fluid from the first chambers at a selected rate and time after the second chambers are substantially empty of drilling fluid so as to depressurize the hydraulic fluid in the first chambers to a selected level substantially equal to a pressure of hydraulic fluid in a hydraulic fluid discharge line hydraulically coupled to the first chambers before enabling an outflow of hydraulic fluid from the first chambers and an inflow of drilling fluid into the second chambers.

30. The method of claim 29, further comprising delaying the releasing of the hydraulic fluid by a selected time so as to ensure that a flow path has been established between the first chambers and the hydraulic fluid discharge line.

31. The method of claim 29, further comprising delaying the outflow of hydraulic fluid by a selected time so as to ensure that a flow path has been established between the first chambers and the hydraulic fluid discharge line.

32. The method of claim 29, further comprising delaying the inflow of the drilling fluid by a selected time so as to ensure that a flow path has been established between the second chambers and a drilling fluid supply line.

33. The method of claim 20, wherein the measuring a volume comprises measuring a position of the separating member and integrating the result.

34. The method of claim 20, wherein the measuring a volume comprises measuring a position of the separating member and converting the position measurement into a volume measurement using an empirically determined algorithm.

35. The method of claim 20, wherein the measuring a volume comprises measuring a position of the separating member and converting the position measurement into a volume measurement using an empirically determined look-up table.