DEVICE AND METHOD FOR DETERMINING THE PHASE INTERFACE LEVEL IN A TANK

Abstract

A method for determining the phase interface level of a multiphase system, includes a tank intended to receive a multiphase system including at least two fluids having distinct phases, and a tube vertically immersed inside the tank. The tube is intended to be filled with a fluid at equal pressure with the fluid contained in the tank at the level of a headspace of the tank. The tube has a plurality of differential pressure sensors per membrane which are spaced apart vertically from each other along the tube to measure the pressure difference between the fluids contained and stratified in height in the tank and the fluid contained in the tube.

Claims

1.-10. (canceled)

11. A device for determining the phase interface level of a multiphase system, comprising: a tank intended to receive a multiphase system comprising at least two fluids having distinct phases; a tube vertically immersed inside the tank, said tube being intended to be filled with a fluid at equal pressure with the fluid contained in the tank at the level of a headspace of said tank, the tube comprising a plurality of differential pressure sensors per membrane which are spaced apart vertically from each other along the tube to measure the pressure difference between the fluids contained and stratified in height in the tank and the fluid contained in the tube each differential pressure sensor comprising a corrosion-resistant sealed membrane which is mounted in an orifice of the tube and which is fixed to a Bragg grating optical fiber strain sensor, the strain of said membrane being intended to measure a differential of pressure between, on the one hand, the multiphase mixture contained in the tank and, on the other hand, the fluid contained in the tube; and the orifices of the tube in which the membranes of the differential pressure sensors are mounted being positioned on several lateral faces of the tube, each lateral face of the tube being provided with the same number of differential pressure sensors spaced apart at the same pitch in height from a different initial height from a lower end of the tube in order to increase the spatial resolution of the measurement.

12. The device according to claim 11, wherein the membranes of the differential pressure sensors extend over the entire height of the tank.

13. The device according to claim 11, wherein the Bragg grating optical fiber strain sensors are fixed to the center of the membrane of each differential pressure sensor.

14. The device according to claim 11, wherein the Bragg grating optical fiber strain sensors are distributed in several independent strain sensor gratings in order to obtain redundancy of the measurement chain.

15. The device according to claim 11, wherein the tube comprises a central cavity intended to be filled with a fluid at equal pressure with the fluid contained in the tank at the level of a headspace of said tank, said central cavity being closed at a lower end and obstructed at an upper end by an expansion bellows.

16. The device according to claim 11, wherein the Bragg grating optical fiber sensors progress longitudinally inside the tube.

17. A method for determining the phase interface level in a tank receiving a multiphase system by means of a device according to claim 11, comprising: for each differential pressure sensor Cj, with 1<j≤n, the calculation of the slope of the differential pressure measurement line associated with the two differential pressure sensors Cj−1, Cj; for each differential pressure sensor Cj, with 2<j<n, the calculation of the variance associated with each set of sensors C2 to Cj of the slope of the differential pressure measurement line associated with the two differential pressure sensors Cj−1, Cj; for each differential pressure sensor Cj, with 2<j<n, the calculation of the variance associated with each set of sensors Cj+1 to Cn of the slope of the differential pressure measurement line associated with the two differential pressure sensors Cj−1, Cj; for each differential pressure sensor Cj, with 2<j<n, the calculation of the sum of the two previously calculated variances; the localization of at least a local minimum of the sum of the two previously calculated variances in order to identify the sensor Cj corresponding to a phase interface level; and the intersection of the linear regression line of the measurements from the sensors Cj to C1 with the linear regression line of the measurements from the sensors Cj+1 to Cn in order to identify the height of the phase interface level.

18. The method according to claim 17, wherein, in case of localization of two local minima which identify two sensors Cj and Ck, the intersection of the linear regression line of the measurements from the differential pressure sensors Cj to C1 with the linear regression line of the measurements from the differential pressure sensors Cj+1 to Ck accurately identifies a first phase interface level height, and the intersection of the linear regression line of the measurements from the differential pressure sensors Ck+1 to Cn with the linear regression line of the measurements from the differential pressure sensors Cj+1 to Ck accurately identifies a second phase interface level height different from the first height.

19. The method according to claim 17, wherein the calculation of the slopes of linear regression lines is carried out by the method of least squares.

20. The method according to claim 17, wherein the tube of the device is filled with a liquid having a low freezing point and a high boiling point.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Other characteristics and advantages of the present invention will emerge from the description given below, with reference to the appended drawings which illustrate an exemplary embodiment thereof without any limitation. In the figures:

[0025] FIG. 1 is a sectional view of a cylindrical multiphase gravity separation tank equipped with a device for determining the phase interface level according to the invention;

[0026] FIG. 2 is a perspective view of the tube of the device of FIG. 1;

[0027] FIG. 3 is a longitudinal sectional view of FIG. 2;

[0028] FIG. 4 is a sectional view along IV-IV of the tube of FIG. 3; and

[0029] FIGS. 5A and 5B illustrate two examples of curves of pressure differential as a function of the height in a gravity separation tank obtained by the implementation of the method for determining the phase interface level according to the invention.

DETAILED DESCRIPTION OF THE INVENTION

[0030] The invention relates to a device for determining the phase interface level in a multiphase system tank, in particular in a multiphase gravity separation tank, for example a subsea multiphase gravity separation tank, that is to say a device making it possible to measure inside a tank the levels between the different phases separated into several superimposed strata of an initially multiphase mixture (in particular water, gas and oil).

[0031] By “multiphase system” is meant here a mixture of at least two fluids having different phases, namely in particular a mixture of fluids in different phases (i.e. liquid phase, aqueous phase and gas phase), or a mixture of at least two liquids that are immiscible with each other.

[0032] Typically, a subsea multiphase gravity separator comprises in particular several cylindrical-shaped ducts which are aligned horizontally and parallel to each other. These ducts form tanks in which the multiphase mixture (for example a mixture extracted from a hydrocarbon production well) is injected at one of the ends. The ducts are long enough to allow the different phases of the mixture to be separated by gravity inside the ducts. At the end opposite to their injection, the ducts are connected to one or several pump(s) which will suck the oil and gas phases of the mixture to send them on the surface (to an FPSO), while the aqueous phase is sucked by another pump to be returned to the production well.

[0033] FIG. 1 represents, in cross section, a duct forming a tank 2 of such a subsea multiphase gravity separator. Of course, the present invention is not limited to the subsea gravity separators but also applies to the terrestrial gravity separators and in general to any type of tank, whether in the open air or not.

[0034] The tank 2 has a circular-shaped cross section. It is filled with a multiphase mixture 4 whose phases, under the effect of gravity, are separated into a fluid in a liquid state (water layer 4a at the bottom of the tank), a fluid in an oil state (oil layer 4b above the water layer), and a fluid in a gas state (gas layer 4c above the oil layer).

[0035] A tube 6 of longitudinal axis X-X is mounted vertically inside the tank 2 by being immersed in the multiphase mixture 4. More specifically, the tube 6 passes right through the tank and is fixed in its upper part on a vertical tap 8 of the tank.

[0036] As represented in FIGS. 2 and 3, the tube 6 may have, in its part immersed in the tank, a substantially square-shaped cross section with four lateral faces 10. More generally, this cross section of the tube may have a polygonal (for example hexagonal) shape. Moreover, in the case of a very high tube, it can have only one face equipped with differential pressure sensors.

[0037] Each of these lateral faces 10 preferably comprises the same number n of differential pressure sensors 12 (in the example illustrated in FIGS. 1 to 3, the number n of differential pressure sensors is equal to 11 on one face and to 12 on a second face).

[0038] The differential pressure sensors 12 are distributed over the entire height of the tank. For each lateral face 10 of the tube, the differential pressure sensors are spaced apart at the same pitch p in height from an initial height, respectively H1, H2, H3, H4, different from a lower end of the tube.

[0039] FIG. 3 represents the respective initial heights H1 to H4 for all of the lateral faces 10 of the tube. It will be understood that these initial heights are offset longitudinally relative to each other by the same distance h.

[0040] Preferably, the distance h between the respective initial heights H1 to Hf (in the case of a tube 6 with a cross section in the form of a polygon with f lateral faces) is equal to p/(f−1), p being the pitch spacing between two adjacent differential pressure sensors. Thus, in the example of a tube 6 having four lateral faces, the distance h between the respective initial heights H1 to H4 is preferably equal to p/3.

[0041] Furthermore, as more specifically represented in FIG. 4, each differential pressure sensor 12 is composed of a sealed membrane 14 (or tympanum) made of corrosion-resistant material (for example alloy or glass) which is sealingly mounted in an orifice 16 made in the lateral faces 10 of the tube and a Bragg grating optical fiber strain sensor 24 which is fixed to the membrane 14.

[0042] The membrane 14 of these differential pressure sensors thus separates the interior of the tube from the interior of the tank 2. The membrane is thus subjected to a pressure differential between, on the one hand, the multiphase mixture 4 contained in the tank and, on the other hand, the fluid 22 contained in the tube.

[0043] It will be noted that the membranes 14 form, for each lateral face 10 of the tube, a string of point sensors which cover the entire height of the tank (the membranes are distributed over the entire height of the tank).

[0044] It will also be noted that the orifices 16 (for example of circular shape) made in the lateral faces 10 of the tube and in which the membranes 14 of the differential pressure sensors are mounted can be aligned, for each lateral face of the tube, on an axis parallel to the longitudinal axis X-X of the tube.

[0045] In addition, the interior of the tube 6 comprises a central cavity 18 which is closed at the lower end of the tube, obstructed in its upper part by a metal expansion bellows 20 and which receives a fluid 22 having a low freezing point and a high boiling point (typically glycol or hydraulic oils).

[0046] When the temperature varies, the volume of the fluid 22 contained in the central cavity 18 of the tube will vary and this variation is taken up by the expansion bellows 20. Thanks to the flexibility of the expansion bellows, any internal volume variation will not result in pressure variation inside the central cavity 18 of the tube.

[0047] Furthermore, the great flexibility of the expansion bellows makes it possible to affirm that the pressure inside the tube is equal to the external pressure, so that the expansion bellows transmits the external pressure to the fluid 22 contained inside the tube. We are talking about equal pressurization of the interior of the tube with the headspace of the fluid 4c.

[0048] In practice, the tube 6 is generally prepared in the following way: application of a vacuum, then filling with the fluid 22, in this case glycol or hydraulic oils for application to a hydrocarbon separator and closing with a sealed plug.

[0049] The expansion bellows 20 is more specifically positioned at the upper part of the tube which passes through the vertical shoulder 8 of the tank.

[0050] Still according to the invention, the differential pressure sensors 12 comprise, for the acquisition of the measurement of the pressure differential, Bragg grating optical fiber strain sensors 24 fixed to the membranes 14 (the strain of the membranes under the effect of the pressure differential results in elongation or contraction of the Bragg grating optical fiber).

[0051] For this purpose, as represented in FIG. 4, the tube 6 comprises four lateral holes 26 which extend longitudinally over the entire height of the tube. The optical fibers of the Bragg grating optical fiber strain sensors 24 have a measurement point 24a fixed by bonding to the center of the membranes 14 and a part 24b which progresses between two adjacent measurement points via the lateral holes 26.

[0052] Different configurations are possible for the Bragg grating optical fiber strain sensors 24. It is possible to envisage a chain of 11 or 12 optical fiber strain sensors for each lateral face 10 of the tube (each chain showing a forward and possibly a return direction in one of the lateral holes 26 of the tube), or two chains of two times 11 or 12 optical fiber strain sensors (with the forward direction through two lateral holes and return direction through the two other lateral holes), or a chain of four times 11 or 12 optical fiber strain sensors for all four lateral faces.

[0053] Preferably, in order to obtain measurement redundancy by the acquisition of two independent measurement chains, it is still possible to position two chains of Bragg grating optical fiber strain sensors side by side on the same membrane.

[0054] An example of a non-limiting algorithm for determining the phase interface level by the device according to the invention as described above is described below in relation to FIGS. 5A and 5B.

[0055] For each set of two consecutive differential pressure sensors C.sub.j, C.sub.j+1, software means calculate:

for each differential pressure sensor C.sub.j, with 1<j≤n, the slope of the differential pressure measurement line associated with the two differential pressure sensors C.sub.j−1, C.sub.j;
for each differential pressure sensor C.sub.j, with 2<j<n, the variance associated with each set of sensors C.sub.2 to C.sub.j of the slope of the differential pressure measurement line associated with the two differential pressure sensors C.sub.j−1, C.sub.j;
for each differential pressure sensor C.sub.j, with 2<j<n, the variance associated with each set of sensors C.sub.j+1 to C.sub.n of the slope of the differential pressure measurement line associated with the two differential pressure sensors C.sub.j−1, C.sub.j; and
for each differential pressure sensor C.sub.j, with 2<j<n, the sum of the two previously calculated variances.

[0056] From these calculations, the software means localize at least a local minimum of the sum of the two previously calculated variances, this local minimum being identified at a sensor C.sub.k corresponding to a phase interface level.

[0057] The software means then construct the linear regression line of the measurements from the sensors C.sub.k to C.sub.1 on the one hand, and the linear regression line of the measurements from the sensors C.sub.k+1 to C.sub.n on the other hand. The intersection of these two linear regression lines accurately identifies the height of the phase interface level.

[0058] When there are two local minima which identify two sensors C.sub.l and C.sub.k, the intersection of the linear regression line of the measurements from the sensors C.sub.l to C.sub.1 with the linear regression line of the measurements from the sensors C.sub.l+1 to C.sub.k accurately identifies a first phase interface level height, and the intersection of the linear regression line of the measurements from the sensors C.sub.k+1 to C.sub.n with the linear regression line of the measurements from the sensors C.sub.l+1 to Ck accurately identifies a second phase interface level height.

[0059] The number of extrema sought corresponds to the number of phase interfaces sought (a single extremum for a fluid stratified in two phases, and two extrema for a fluid stratified in three phases).

[0060] A first example of data thus calculated is proposed in Table 1 below and illustrated by the curve of FIG. 5A. This example applies to the case of a fluid present in a gravity separation tank which is in a stratification of three different phases, namely a gas phase, an aqueous phase and a liquid phase. The measurements were carried out using 31 differential pressure sensors spaced apart from each other between a tank depth equal to 0 mm (for the uppermost sensor) up to a tank depth equal to −600 mm (for the lowest sensor).

[0061] FIG. 5A represents for this first example the pressure differential curve ΔP (in mbar) measured by the pressure differential sensors as a function of the depth (in mm) in the gravity separation tank.

TABLE-US-00001 TABLE 1 slopes of the lines connecting 2 variance variance ΔP depth consecutive 1 of the 2 of the sum of the C.sub.j (mbar) (mm) sensors slopes slopes variances 1 0.78 0 2 2.33 −20 −0.08 3 4.16 −40 −0.09 0.00005 0.00289 0.00294 4 6.23 −60 −0.10 0.00011 0.00272 0.00284 5 7.50 −80 −0.06 0.00023 0.00274 0.00296 6 9.33 −100 −0.09 0.00019 0.00259 0.00278 7 11.67 −120 −0.12 0.00030 0.00220 0.00250 8 12.89 −140 −0.06 0.00037 0.00216 0.00253 9 14.73 −160 −0.09 0.00032 0.00189 0.00221 10 16.73 −180 −0.10 0.00030 0.00148 0.00178 11 18.76 −200 −0.10 0.00029 0.00095 0.00124 12 18.80 −220 0.00 0.00091 0.00099 0.00189 13 19.54 −240 −0.04 0.00098 0.00091 0.00189 14 19.82 −260 −0.01 0.00120 0.00091 0.00211 15 19.80 −280 0.00 0.00148 0.00095 0.00243 16 20.49 −300 −0.03 0.00145 0.00083 0.00228 17 21.36 −320 −0.04 0.00139 0.00058 0.00197 18 21.00 −340 0.02 0.00168 0.00062 0.00230 19 21.76 −360 −0.04 0.00161 0.00032 0.00193 20 21.60 −380 0.01 0.00175 0.00030 0.00204 21 21.00 −400 0.03 0.00200 0.00032 0.00233 22 19.71 −420 0.06 0.00250 0.00024 0.00274 23 19.55 −440 0.01 0.00251 0.00020 0.00271 24 18.80 −460 0.04 0.00267 0.00023 0.00290 25 18.18 −480 0.03 0.00276 0.00026 0.00302 26 17.38 −500 0.04 0.00287 0.00029 0.00316 27 17.30 −520 0.00 0.00281 0.00017 0.00298 28 16.46 −540 0.04 0.00290 0.00021 0.00311 29 16.17 −560 0.01 0.00286 0.00005 0.00291 30 15.19 −580 0.05 0.00296 0.00000 0.00296 31 14.49 −600 0.04

[0062] In this first example, two extrema are identified for the sum of the variances, namely for the “sensor 11” located at a depth of −200 mm and for the “sensor 19” located at a depth of −360 mm.

[0063] It is therefore deduced that the interface between the gas and aqueous phases of the fluid contained in the gravity separation tank is located at the sensor 11 at a depth more accurately calculated by linear regression of −208.56 mm in the tank and that the interface between the aqueous and liquid phases of the fluid is located at the sensor 9 at a depth more accurately calculated by linear regression of −365.40 mm in the tank.

[0064] A second example of data is proposed in Table 2 below. This example applies to the case of a fluid present in a gravity separation tank which is in a stratification of two different phases, namely a gas phase and a liquid phase.

[0065] As in the previous example, the measurements were carried out using 31 differential pressure sensors spaced apart from each other between a tank depth equal to 0 m (for the uppermost sensor) up to a tank depth equal to at −15 m (for the lowest sensor).

[0066] FIG. 5B represents for this first example the pressure differential curve ΔP (in mbar) measured by the pressure differential sensors as a function of the depth (in m) in the tank.

TABLE-US-00002 TABLE 2 slopes of the lines connecting 2 variance variance ΔP depth consecutive 1 of the 2 of the sum of the C.sub.j (mbar) (m) sensors slopes slopes variances 1 38.32 0 2 54.35 −0.5 −32.04 3 55.45 −1 −2.21 14.92 15.82 30.73 4 75.56 −1.5 −40.22 16.34 15.37 31.71 5 93.17 −2 −35.22 14.84 15.13 29.97 6 105.41 −2.5 −24.47 13.33 15.28 28.61 7 131.62 −3 −52.41 15.46 13.33 28.79 8 148.53 −3.5 −33.83 14.34 12.80 27.15 9 161.15 −4 −25.24 13.58 12.72 26.30 10 186.96 −4.5 −51.62 14.39 9.21 23.59 11 202.45 −5 −30.98 13.66 7.89 21.56 12 208.69 −5.5 −12.47 14.28 8.01 22.29 13 204.52 −6 8.33 17.46 7.34 24.80 14 210.49 −6.5 −11.93 17.30 7.48 24.78 15 211.18 −7 −1.39 17.88 7.54 25.42 16 221.46 −7.5 −20.55 17.30 7.04 24.34 17 221.22 −8 0.47 17.80 6.97 24.77 18 225.10 −8.5 −7.76 17.63 7.23 24.86 19 231.11 −9 −12.02 17.29 7.42 24.70 20 236.76 −9.5 −11.30 16.98 7.65 24.63 21 242.53 −10 −11.54 16.67 7.88 24.56 22 246.08 −10.5 −7.10 16.52 8.31 24.82 23 244.98 −11 2.21 16.77 8.18 24.95 24 255.35 −11.5 −20.74 16.41 6.96 23.37 25 256.07 −12 −1.45 16.44 7.27 23.71 26 261.22 −12.5 −10.28 16.18 7.74 23.92 27 266.38 −13 −10.33 15.93 8.27 24.20 28 271.03 −13.5 −9.29 15.71 9.02 24.73 29 269.81 −14 2.44 15.85 9.98 25.84 30 277.69 −14.5 −5.76 15.58 0.00 15.58 31 275.59 −15 4.20

[0067] In this second example, a minimum is identified for the sum of the variances, namely for the “sensor 11” located at a depth of −5 m.

[0068] It is therefore deduced that the interface between the gas and liquid phases of the fluid contained in the gravity separation tank is located at the sensor 11 at a depth more accurately calculated by linear regression of −5.04 m in the tank.

[0069] Of course, it is possible to consider other algorithms than the one described previously to determine the phase interface level from the strain measurements collected from the Bragg grating optical fiber strain sensors.