CA2150325A1

CA2150325A1 - Multiphase flow meter

Info

Publication number: CA2150325A1
Application number: CA002150325A
Authority: CA
Inventors: Ian Atkinson; Peter Long; Gerard Segeral
Original assignee: Ian Atkinson; Peter Long; Gerard Segeral; Schlumberger Canada Limited
Current assignee: Schlumberger Canada Ltd
Priority date: 1994-05-27
Filing date: 1995-05-26
Publication date: 1995-11-28
Also published as: GB2289766B; MY113066A; US5591922A; EP0684458A2; TNSN95059A1; NO952088L; DE69529189D1; GB2289766A; FR2720498A1; NO952088D0; EG20705A; BR9502559A; EP0684458B1; AU699448B2; FR2720498B1; AU2032095A; EP0684458A3; GB9510315D0

Abstract

The invention relates to a device for measuring flow rate of multiphase fluids such as oilwell effluents, containing liquid hydrocarbons, gas, and water.
The device comprises:
first and second sections situated at a distance one from the other in the flow direction, each including a passage provided with means for inducing a change of speed therein; and respective means for measuring the resulting pressure differences, the pressure different signals obtained in said respective sections being suitable for cross-correlation to produce a signal representative of the total volume flow rate.

Description

MULTIPHASE FLOW METER

The present invention relates in general to studying the flow ch~cteristics of m~ltiph~ce fluids, especially fluids having three phases. In particular, the invention relates to measuring the flows of fluids such as oil effluents composed of crude oil, water, and gas.
The problems associated with measuring flow rate in a multiphase fluid are well known. The techniques that can be applied to flows of a single phase such as systems in~ lding a turbine or the like turn out to be unsuit~ble in multiphase flows, and conse-quently it is often ~ecess-.y to L,e.fwll~ a plurality of llR~u~ cll~s in order to be able to obtain a result that is accurate with respect to the volume flow rates of the various phases. In the oil and gac industry, and in co~lp~dble industries, it is common practice to measure mul-tiphase flows, and in a production well, it is usual for the fluid products to comrrice a mix-ture of oil, of water, and often of gas. Given that it is i,l,polldnt to determine the volume flow rates of the individual phases at certain time intervals throughout the lifetime of the well, in order to de~.,lline whether any col~ e action is required to improve or restore the productivity of the well, and also to determine exactly how much oil is being produced, various m~tho ls have been proposed in the prior art for measuring flow rates in multiphase flows. In convention~l manner, when testing exploration wells, the fluids are sent to separa-tors and the volumes of each phase are determined separately. However, that requires an in-st~ tion that is voluminous and expensive and it does not make it possible to obtain infor-mation inst~nt~neously about the production of the well.
Proposals have also been made to combine a venturi type device, a gamma ray densilo,lletel, and a turbine meter (R. Franck et al., European Two Phase Flow Group Meet-ing, University of Str~theli-le, Glasgow, June 3-6, 1980). That system was developed for two-phase flows (wa~el/s~ea~ll) and is not capable of measuring three-phase flows.
US patents 4 856 344 and 4 974 542 describe multiphase flow rate measure-ment devices of the gradio-venturi type which, by means of differential pressure measure-ments distributed along the length of the device, provide information relating to the flow rates of the phases.
A gradio-venturi cc-llbines a section that measures a static pl~,s~u~ gradient and a section that measures variation in lllo~.fn(l..,~ That device provides accurate mea~ul~ nt, in par-ticular for the oil-water and oil-gas flows encountered in the oil industry. Nevertheless, it has been observed that ,l,easulcmeut accuracy is limited once the proportion of gas by vol-ume eYcee~lc 60% to 70%.
Various systems for measuring multiphase flows are proposed in British pat-ents 1 272 152 and 1 461 537, and in the document 6.2 North Sea Flow Measurement Work-shop 1990, National Engineering Laboratory, Glasgow, entitled "Simple full-bore water-cut 2 21~0325 meaaulclllcnt technique" by D. Brown and JJ. der Boer. In each case, the apparatus de-scribed is in the form of an upside-down U-shape and pl~sau~e meaaurc~ ts are pclr~,lllled at various heights along each of the branches of the U-shape and then colllp~,d to determine the pala~llcte~a of the flow. Given that none of the measul~mell~à relates to ch~nges in dy-namic l~lcssulc, it is not possible to calculate the individual flow rate of each phase on the basis of those l)lesaulc meaaulclll~ à alone.
Patent WO ~3/17305 describes a combin~tinn of two gradio-venturis ~icposeA
in sections where the effect of gravity makes it possible to measure the density of the mix-ture. In addition, means are provided for measuring the water fraction and to extract a full sample of the mixture. That system is fully capable of ~.f~....;ng multiphase flows, but it likewise suffers from a limit in terms of measu~ ent accuracy once the volume fraction of the gas exceeAc 60% to 70%.
Patent WO 93/19347 describes a mllltirh~ce flow meter accoci~ting in series a volume-measuring flow meter and two flow-measuring sensors responsive to the molllenlulll of the measured fluids, and in which the ratio of gas speed to liquid speed is m~inr~ineA at a constant value, e.g. equal to 1. The volume-measuring flow meter consists, for example, in a turbine or in a flow meter having oval gears. Such an intrusive type of device suffers from well-known drawbacks. In addition, keeping the ratio of gas speed to liquid speed at a con-stant value con~ es an awkward constraint.
Mention may also be made of the following ~oc~ments concerning flow rate meaaul~illlcnla in multiphase flows of the oil effluent type:
publication SPE 23065 by Gold et al., entitled '~Meaaulelllent of multiphase well fluids by positive displ~ement meter";
publi~tion SPE 25620 by Cary et al., entitled "New well testing technology: a portable three-phase test unit";
patent application WO 93/24811; and US patent 5 287 752.
Finally, mention is made of US patent 4 397 190 which describes a device adapted to making meaaul~lllellts in production wells.
The present invention seeks to pclr~ llll flow rate meaaulements on multiphase fluids, and more particularly on oilwell çMuentc, which may have a gas content eXcee~ing 80% or even 90%.
In a first aspect, the present invention provides a method of llleaauling flow rate for a multiphase fluid, such as the effluent of an oil well, which may contain a mixture of liquid hydrocarbons, gas, and water, comprising the steps of:
changing the flow velocity respectively in a first section and in a second sec-tion spaced from each other in the flow direction;

3 2l~o325 cll. ;r g a first and a second p-~s~w~ differences along le;,~e.,~B/ely said first and said second sections to obtain first and second pl. s~ule difference signals; and co...p~.;..g said first and second ~ ,s~ule dirrer~ ce signals to derive a thirdsignal indicative of the total volume flow rate q.
Preferably said third signal is formed by cross-correlation belween the first and second plessul~ (lirr~ ~nt signals.
In accordance with a further aspect, in order to deterrnine the flow rates of the phases, a third pressure dirr~nce is measured in a flow section to obtain a fourth signal that is a function of the total mass of flow rate Q and of the density p of the Illib~lUl~, and a fifth signal is formed l~ se~ re of said density p .
On the basis of the above inrol..~tion in another aspect of the invention, two possible values for the total mass flow rate Q are formed: a first value from the fourth signal which is a function both of the total mass flow rate Q and of the density p of the fluid, and from the fifth signal which is represçnt~tive of the density p; and a second value from said fourth signal and from the third signal which is represent~tive of the total volume flow rate q the first value being appropriate when the gas fraction of the fluid is moderate and the sec-ond value being appropriate when the gas fraction is high.
The invention will be well understood from reading the following description made with reference to the acco,l,panying drawings, listed as follows:
Figure 1 is a diagram of a device providing an indication of total flow rate;
Figure 2 is a diagram of a device providing a set of mea~ur~.llel1~s for determin-ing the flow rates of each phase in a three-phase fluid such as the effluent of an oil well;
Figure 3 is a diagram showing a variant embodiment of the Figure 2 device;
Figure 4 is a block diagrarn showing the processing applied to the information provided by the device of Figure 2 or of Figure 3;
Figure S shows a possible disposition for the device of Figure 3; and Figure 6 shows another possible disposition.
Figure 1 shows a duct 10 along which a multiphase fluid flows in the direction of the arrow, which fluid may be an effluent from an oil well and composed of a ~ ule of crude oil (referred to below merely as "oil"), of gas, and of water.
Two measul~."cnt sections A and B are provided along said duct and they are sep~tçA by a ~ict~nce L defined in the flow direction, said (iict~nce L coll,~sponding to a volume of fluid V (acsuming that the duct is of constant section S, then V = L.S). Each ,l~e&sulelllent section includes means for inducing a change in the fluid velocity, IG~lesented in the form of venturis 1 lA and 1 lB each forming a constriction in the flow cross-section. Each venturi is associated with a pair of pressure takeoffs spaced from each other inthe flow direction. The sections A and B thus include the following pressure takeoffs respec-tively: 12A, 13A, and 12B, 13B. Each pair of ~I~SSulc; takeoffs is connecteA to a respective dir~rcnlial ~ S~UlC sensor 14A, 14B r~,al~onsi~e to the pl.,s~ul~, difference genc.aled in each section. The sensors 14A, 14B provide l.,sl,e~ e signals ~pX(A) and ~pX(B).
These signals are applied to a cross-correlation device 15 which, using a tech-nique that is known per se, deterrnines the time lag ~ collc~onding to maximum correlation b~ " the signals coming from the sections A and B. Given the volume V between the seC*onc A and B, the time lag deterrnined in this way makes it possible to determine the total volume flow rate q whic}~is inversely pl~ollional to time ~ .
It will be noted that in the embodiment shown in Figure 1, the ~lcs~ule takeoffsare located l.,sl~e-;~i./ely Up:j~lcalll from the constriction and in the constriction. However, other arr~n~e---çnt~ are possible. For inct~nce7 the IJrcssul~ takeoffs may be located respec-tively in the constriction and dow"s~ l of the constriction. It is also possible to provide plessule takeoffs located respectively u~ ~ll and downstream of the constriction, in which case the pl~ Ul~c difference will be based on the pl~ ul~; drop ent~ile~ by the venturi. These alternative arrang~ c of pressure takeoffs may be used also with the other embo iim~ntc described hereinbelow.
Means other than venturis can be used, e.g. devices having orifices. Such means could be considered. It is possible to consider using means of different types in the sections A and B; however for correlation purposes it is preferable to use means that are of the same type, and better still that are identic~l The geolllellical disposition of the sections A and B may be implçm~nte-l in different manners: the sections may be disposed in a duct that is horizontal, or vertical, or one section may be placed in a vertical portion of duct while the other is in a horizontal por-tion, or indeed, as shown in Figure 2, section A may be placed in a vertical portion where the flow is up while section B is in a vertical portion where the flow is down.
The ~iict~nce L which determines the volume V is selecte~ in appropriate man-ner as a function of the expected range of flow rates: it should be short for low flow rates and longer if higher flows are expected.
In order to enlarge the dynamic range of the device, a third mea~ull,llæ~lt sec-tion (not shown) may be provided similar to the sections A and B and c~it~t~l at a ~lict~nce L' from the section A coll~ ollding to a volume V' of fluid that is different from V. For example, if the volume V is apl)rupliate for relatively small flow rates, then a volume V' greater than V will be applu~liate for larger flow rates. The signal from the sensor associ-ated with the third section is sllb~ ule~ for that coming from the section B, as a function of the flow rate values as e~l.ecle-l or measured.
Concerning correlation techniques as applied to flow rate measurement, men-tion may be made of the work by M.S.Beck and A. Plaskowski "Cross-correlation flow me-ters - their design and applications", Adam Hilger, 1987. It may be observed that for deter-mining the time ~, there exist methods other than cross-correlation for co~ )~ing the signals 21503~5 such as the signals delivered by the sections A and B. Reference may be made on this topic to the li~ tulc on signal proces~ing In the study of multiphase fluids, the device described has the advantage of a large dyllalllic range concerning the plup<:~lLion of gas and liquid in the fluid, and it is of moderate cost.
The device shown dia~ tir~lly in Figure 2 provides inÇollllalion that en-ables the flow rate of ea~h phase in a three-phase fluid such as oil well effluent to be deter-mined. The device as shown comprises a portion of duct 20 where the flow is up~v~uds, a portion of duct 20' where the flow is horizontal, and a portion of duct 20" where the flow is dowllw~u~ls.
A device of the type shown in Figure 1 is provided that cornrrices mea~ul~ ent sections A and B respectively comprising venturis 21A and 21B, pairs of plCS~ulc t~keoffc 22A, 23A and 22B, 23B, and dirrere.llial ~l~ssulc sensors 24A, 24B. The sections A and B
are ~it..~ted ,~ ec~ ely in the up portion 20 and in the down portion 20", however as ex-plained above they could be placed differently, providing the volume of fluid V that corre-sponds to the rlict~nse L between them in the flow direction is a~,u~Jliate.
A gradio-venturi type device is also provided that comprices a measurement section C and a mcasul. .lRnt section D citn~ted in the portion of duct 20 where the fluid flows upwards.
The section C is provided with means for creating dynamic ples~ur~ such as the venturi 21C (however it would also be possible to use a device having orifices or the like) and with plCSSulc takeoffs 22C and 23C placed u~sLI~ alll from the venturi and at the venturi, and a differential ~)lC;~UlC sensor 26 is connected to the ~ll ssUle takeoffs 22C and 23C. The dirr~ ial prcssule signal ~pV provided by the sensor 26 is a function of the total mass flow rate Q and of the density p of the mixture, and more precisely, to a first appr~illlation, it is l,lo~llional to the expression Q2/p .
Section D of the "gradiomanolllet~ l" type includes a portion of duct 20D that is of coll~t~nt section, together with two ples~ulc takeoffs 27D and 28D that are spaced apart by a ~lict~nce h in the (vertical) direction of the flow. These prcssulc takeoffs are connected to a dirr~.~..,ial ~ S~ul~ sensor 29 that produces a signal ~pG. In conventional manner, the ~iirr. .~ nce in level b~ . xn the pl~s~ule takeoffs, equal to the ~lict~nce h for a vertical duct, creates a static pl~,S~UU~ difference that is pl~Jpollional, to a first ap~lu~illlalion, to the den-sity p of the llli~lule.
The device shown also insluAes a device 30 that provides one or more indica-tions about the composition of the multiphase fluid, in other words relating to the ~lOpOl lions by volume or by mass of the phases conctituting the fluid. For convenien~e this device is shown in Figure 2 as being sit~l~ted in the horizontal portion of duct 20', however such a dis-position is not essenti~l ~- In the above-men~ioned case of oil well efflllent (a mixture of water, oil and gas) there exist various app~tuses for determining composition, and more particularly for d.,t~ fi~ g water fraction, which al)p~u~tuses implem.ont various physical principles: nu-clear m~tho lc (interaction with gamma rays), electrom~gnetic (microwaves), etc. The meas-ul~ cnls may also be pc.Ç~-Illed in various dirf~cnl ways: directly on the flow, or on sarn-ples taken from a shunt flow, as desçrih~ed in patent application WO 93/17305. Depending on the technique used, it is possible to obtain an in~ tion of the water fraction in terms of the ratio of water to all of the liquid phases comhined (WLR), which comes to the same as obtaining the wate~/oil ratio (WOR); or else the water hydl~lon cut (WHC) is obtained, i.e. the fraction of water relative to all of the hydrocarbons (oil + gas).
Figure 3 shows an advantageous variant embodillle,ll of the Figure 2 device, in which a single venturi 21AC replaces the venturis 21A and 21C of Figure 2. The measure-ment sec*ons A and C of Figure 2 now coin~ e in a single sec*on AC.
It may be observed that the venturi 21AC is associated, as in Figure 2, with twodirfe,r~-llial ~ ,S~UIC sensors 24A and 26. The reason for ret~ining two dis*nct sensors is that the set of characteristics required for a sensor depends on the use to which the mea~urellRn are put. The characteristics that are desirable with respect to accuracy, resolu*on, and band-width are very dirfel.,nt depending on whether the measurements are used for obtaining an absolute value, as is the case for the signals ~pV, or merely a relative value, as is the case for the signals ~pX which are subjected to correla*on ~,locessing. Although in theory it is not impossible that a sensor could exist that is suitable for both purposes, in the present state of affairs, such a sensor is not available. Figure 3 also shows two pairs of ~lcs~ule takeoffs 22A, 23A and 22C, 23C conneclcd l~sl)e~ ely to the sensors 24A and 26. However it would be possible to use a single pair of ~ ssure takeoffs for the section AC and conne~lcd to both dirf, .ential pres~ulc sensors. An alternative emb~lim~nt, not shown, includes a plu-rality of pairs of p~S~UlC takeoffs located at angularly distributed positions on the flow duct so as to remove the effect of local variations and provide pl~SSUlc difference mea~ule,l~enls averaged for a given flow section. Such an embodiment may comprise for inct~nce four pairs of ~ S~Ul~ takeoffs ~ngnl~rly distributed i.e. at 90 sp~cingc~ and the takeoffs are connected so as to provide for each IJIeS:.UI~, difrel~,nce two lleasule-ll.,llls at ~ m~ohic~lly opposite po-sitions, which are averaged in any ap~l~,pliate ,llannel.
Figure 4 is a block diagram showing one way of procescing the info~n~tion provided by a device as described above with Icfe-ence to Figures 2 and 3. The following descl;~lion refers to the case where the multiphase fluid is an oilwell effluent. The magni-tudes Q (mass flow rate), q (volume flow rate), and p (density) are associated with the fol-lowing indices: ~ (gas), I (liquid), o (oil), and _ (water); while the absence of an index means, as above, the total flow rate or the density of the mi~clul~.

7 215032~S

In the explanation helow~ it is ~c~....... ~ that the denciti--s PO, PW. and pg of the colllponents of the mllltiph~se fluid are known.
~ ei~ing is based on the following principle. As explained above, providing the gas fraction is not too great then the gradio-venturi, in combination with a sensor for water fraction, suffices to provide infullllation en~hling the flow rates of the fluids to be de-termineA namely a signal l~ sen~ ;ve of the density p of the IlliAlule and a signal /~pV
e~l~sen~ e of the eA~ ion Q2/p. When there is a large gas fraction, e.g. greater than 65%, the density measure.llent provided by the gr~lio~ no~ te becQmes unusable. How-ever, even when there is a large gas fraction, the info....~l;Qn obtained by correlating the dif-ferential pl~,S:~Ul~, signals ~pX and lepl~sent~l;./e of the total volume flow rate q makes it possiblc in co...l-in~;on with the inro.lllalion ~pV from the gradio-venturi to c~lc~ te the total mass of flow rate Q. If the density p of the ~llixlulc; is not available, the expression Q /p can be eAp~ssed in the form of the product Q.q, and given the volume flow rate q, it is possible to determine the mass flow rate Q.
The mass flow rate Q is calculated continuously via two parallel paths, the first on the basis of the signals ~pV and ~pG provided by the gradio-venturi, and the second on the basis of the signal ~pV and the signal obtained by correlation of the signals ~pX. Two possible values are thus obtained for the mass flow rate Q, one of which is applupliate when the gas content is moderate, while the other is applu~liate when the gas content is high.
Given the ~en~itiÇs of the individual phases, and also the density P 1 of the liquid fraction, which is obtained from the measul~,.l.ent of the water fraction, values of the gas flow rate Qg and qg and of the liquid flow rate Ql and ql are c~lc~ te~ that cc.ll~,s~olld l.,i,l,e~Li./ely to the two values of the flow rate Q. On the basis of each of the resulting pairs of values qg~ ql, a gas content is calculated. In addition, by tracing curves showing the accuracy of measure-ments as a function of gas content, the limiting value of gas content beyond which the accu-racy of the information ~pG from the gradiomanometer becomes unacceptable is established ol~hand. The limit value depends on nulllc.uus ~llete.~ (characteristics of the sensors, ~ e~ e of the device, plup~Lies of the fluid, ...). The gas content values obtained as mentiQne~ above are colllpal~ with said limit value and a selection is made accordingly.
This kind of processing is illustrated by the block diagram of Figure 4. The signals from the gradio-venturi sensors are l,rùcessed by m~-les 40 and 41 that provide data ,senl~l;./e of the density p and of the eA~lession Q2/p or Q.q. The volume flow rate q is provided by the correlator 15 as described with reference to Figure 1. A module 42G corre-sponding to the first above-mentioned path calculates a value for the mass flow rate Q(G) on the basis of data coming from the modules 40 and 41. The module 42G receives the infor-mation Pl (density of the liquid phases) from a module 43 connected to a module 44 for I,lucessing the signals from the sensor 30 in Figure 2. Given the gas density p g and the liq-uid density p l, the module 42G provides coll.,;,ponding values for the gas flow rate Qg(G) 21~0325 and qg(G) and for the liquid flow rate Ql(G) and ql(G). The module 42X c~jl,csponding to the second path calculates the collG;,~onding values Qg(X), qg(x) and Ql(X), ql(X) on the basis of the data coming from the module 41 and from the correlator 15. The two sets of values are applied to a selector module 45 which c~lc~ tes a gas content value on the basis of the gas and liquid flow rate values, and it co~ ucs said value with said limit value men-tioned above in order to select be~ ,n the two sets of values, with selection being symbol-ized by a switch having two positions, LG (moderate gas content) and HG (high gas con-tent). The values selected are applied to a module 46 which also receives the value for the water fraction WOR. This module provides the flow rates for each of the phases.
Figure 5 shows one possible di~osi~ion of the device of Figure 3. This dispo-sition cullc~,~ollds to the archit~ct-lre desç~ihe~ in above-m~ntionecl patent application WO
93/17305, to which lcfe,~l,ce may be made for further details. A measul~.lRnt circuit 50 in the general form of an upside-down U-shape is mollnteA in shunt on the horizontal duct 51 via which the fluid is conveyed. A bypass circuit 52 is provided for lelllpol~ily confinin~ a certain volume of fluid for analysis ~ul~oses. Quick closing valves QCV control fluid ac-cess to the various ducts. A gradio-venturi 53 is provided in the up branch of the circuit 50.
A venturi 54 is provided in the down branch of the circuit 50. According to the tçaching of the above-mentioned patent application, this disposition of the mea~ulclll~,nl circuit makes it possible to co.~ c~te for slip between phases.
The eleme~tc 55 and 56 mounted on the measul~ ent circuit represent respec-tively a water fraction sensor and a multiphase analysis device operating on samples such as an icol-inetic sarnpling device.
Other elc.l,en~ (not shown) are also provided such as absolute pressule sensors and lcll~l~el~ture sensors. The information obtained is used for calculating clencities P O~ P w~
p g under the con~itions of the flow. In addition, a device may be provided to homogenize the fluid as it enters the mea~uiclllent circuit.
Figure 6 shows another possible disposition that is more compact, and in which all of the mea~u~clll.,nt means are disposed on a vertical section of duct 60. In this disposi-tion, the gradio-venturi 61 has its venturi section 62 placed upsll~,~ll from the constant sec-tion gr~rliom~nomçter section 63. This disposition has a favorable influence on the quality of the mea~ulclllellt~ provided by the gr~iom~nomçter, because of the homogenizing effect of the upstream venturi 62 on the fluid. The second venturi 64 is placed immç~ tçly after the gradiomanometer section 63 at an appropl;ate rlict~nce L from the venturi 62 for correla-tion ~ oses. In addition, provision is made to measure the water fraction at thegradiomanometer section 63 by means of a sensor l~l,lcsented by block 65. A device for multiphase analysis by sampling (not shown) may also be mounted on the duct 60, e.g. up-stream from the gradio-venturi.

Claims

1. A method of measuring flow rate for a multiphase fluid, such as the effluent of an oil well, which may contain a mixture of liquid hydrocarbons, gas, and water, comprising the steps of:
changing the flow velocity respectively in a first section and in a second sec-tion spaced from each other in the flow direction;
measuring pressure differences along respectively said first and said second section to obtain a first and a second pressure difference signals; and comparing said first and second pressure difference signals to derive a third signal indicative of the total volume flow rate ?.

2. A method according to claim 1, in which said third signal is formed by cross-correlation between the first and second pressure different signals.

3. A method according to claim 1 or claim 2, in which said measuring step comprises the steps of measuring pressure differences at a plurality of angular positions distributed around the respective sections, and averaging the plurality of measurements thus produced to obtain said first and second signals.

4. A method according to any one of claims 1 to 3, in which a further pressure difference resulting from a change in flow velocity is measured in a flow section to obtain a fourth sig-nal that is a function of the total mass flow rate Q and of the density p of the mixture, and a fifth signal is formed representative of said density p .

5. A method according to claim 4, in which said signal representative of the density p is formed by measuring a static pressure difference between two points of the flow spaced from each other in the vertical direction.

6. A method according to claim 4 or 5, in which said flow section is selected from said first and second sections.

7. A method according to claim 6, in which said further pressure difference is measured at positions angularly distinct from the positions at which said first or second pressure differ-ences are measured.

8. A method according to claim 4 or 5, in which said flow section is distinct from said first and second sections.

9. A method according to any one of claims 4 to 8, in which two possible values for the total mass flow rate Q are formed: a first value from the fourth signal which is a function both of the total mass flow rate Q and of the density p of the fluid, and also from the fifth signal which is representative of the density p; and a second value from said fourth signal and from the third signal which is representative of the total volume flow rate q, the first value being appropriate when the gas fraction of the fluid is moderate and the second value being appro-priate when the gas fraction is high.

10. A method according to claim 9, in which the density p1 of the liquid fraction of the fluid is determined, with the first and second values of the total flow rate and of the density p1 of the liquid fraction being used to determine respectively first and second values of the gas flow rate and of the liquid flow rate, the corresponding gas content is determined and one of the first and second values is selected by comparing the resulting gas content with a prees-tablished limit value.

11. A method according to claim 10, in which the density p1 is determined on the basis of a measurement of the fraction of one of the liquid phases.

12. A device for measuring the flow rate of a multiphase fluid such as the effluent of an oil well, which may contain liquid hydrocarbons, gas, and water, comprising:
first and second sections spaced from each other in the flow direction, each in-cluding a passage provided with means for inducing a change of velocity therein; and first and second means for measuring a pressure difference along respectively said first and second sections and generating first and second pressure difference signals suit-able for cross-correlation with each other, said cross-correlation producing a third signal in-dicative of the total volume flow rate.

13. A device according to claim 12, in which each of said passages includes a venturi.

14. A device according to claim 13, in which each of said first and second measuring means comprises a pressure takeoff located upstream of said venturi and another pressure takeoff located in said venturi.

15. A device according to claim 13, in which each of said first and second measuring means comprises a pressure takeoff located in said venturi and another pressure takeoff located downstream of said venturi.

16. A device according to claim 13, in which each of said first and second measuring means comprises a pressure takeoff located upstream of the venturi and another pressure takeoff lo-cated downstream of the venturi

17. A device according to any one of claims 13 to 16, in which each of said sections com-prises a plurality of angularly distributed pairs of pressure takeoffs spaced from each other in the flow direction.

18. A device according to claim 12 or 13, comprising a third section including a passage provided with means for inducing a change of velocity therein, and third means for measur-ing a pressure difference along said third section, the signal obtained being a function of the total mass flow rate and of the density p of the mixture, and a fourth section where the flow is upward. provided with means for measuring the static pressure difference between two points spaced from each other in the vertical direction, said pressure difference being indica-tive of the density p of the mixture.

19. A device according to claim 18, in which said third means is distinct from said first means.

20. A device according to claim 18, in which the third section coincides with said first sec-tion.

21. A device according to claim 20, in which said third means includes a plurality of pairs of pressure takeoffs arranged at angularly distributed locations.

22. A device according to any one of claims 12 to 21, in which the fluid flows upwards in the first section and downards in the second section.

23. A device according to claim 12, for measuring the flow rate of the effluent of an oil well containing a pressure of liquid hydrocarbons, gaseous hydrocarbons, and water, the device comprising a gradio-venturi that includes a venturi mounted upstream from a gradioma-nometer section of a constant section, and a second venturi placed downstream from the gradio-venturi.

24. A device according to claim 23, in which the second venturi is placed immediately after the gradio-venturi.

25. A device according to claim 23 or 24, including a sensor responsible to water fraction, and placed at the level of the gradiomanometer section.