FR2853416A1 - Determining composition of a homogeneous or heterogeneous fluid comprises comparing differences in attenuation of microwaves passed through different layers - Google Patents

Abstract

The process is carried out in a number of stages: in Stage A the fluid contained in a given volume is divided into layers; in Stage B values representing modifications in microwave beams transmitted through the first of the layers are measured. In Stage C the composition of the fluid in the first layer is determined by comparing the values measured in Stage B with predetermined values for a fluid of known composition. Stage D consists of repeating Stages B and C for the remaining layers of the fluid, and the overall composition of the fluid is determined from the results obtained in the previous stages.

Description

The present invention relates to a method for determining the

  composition of a fluid. More specifically, the invention proposes a method for processing values representing the characteristics of wave beams having passed through the fluid.

  The method can be applied to a homogeneous or heterogeneous fluid. For example, the method according to the invention is suitable for determining the position and the composition of several superposed layers of fluids. For example, the method can be applied to a petroleum effluent contained in a separation tank or circulating in a pipe. According to the operating conditions of this effluent, for example under the effect of gravity, the different phases of the effluent (water, oil and gas) are distributed in laminated form, that is to say in several layers superimposed on distinct and homogeneous compositions. The method makes it possible to determine the position of the layers of water, oil, gas, emulsion composed of water, oil and gas, and of foam composed of oil and gas, etc. In addition, the method makes it possible to determine the composition of the monophasic, emulsion and foam layers.

  Document FR 01 / 16.096 proposes a separator balloon provided with a first rod provided with microwave beam transmitters, with a second rod provided with microwave beam receivers. A microwave source is connected to the transmitters. Processing means record the characteristics of the beams picked up by the receivers, as well as the microwave beam emitted by the source.

  The present invention provides a method of processing the characteristics of the wave beams picked up by the receivers in order to determine the composition of the fluid traversed by the wave beams.

  In general, the present invention relates to a method for determining the composition of a fluid contained in a volume, in which the following steps are carried out: a) several slices which divide said volume are defined, each slice comprising a part of said fluid, b) a family of values representing the modifications of a wave beam having traversed a path in a first of said sections is measured, c) the composition of the part of said fluid located in said first section is determined by comparing the measured values in step b) with a set of families of previously determined values, each family of said set characterizes a fluid of known composition, d) steps b) and c) are repeated for each slice defined in step a), and e) the overall composition of said fluid is determined from the compositions determined in step c) and from the definition of the slices from step a) .

  According to the invention, in step b), it is possible to measure the attenuation and the phase shift of a microwave beam and, in step c), one can determine the composition of the fluid located in said first wafer by comparing the attenuation and the phase shift measured in step b) with a set of pairs of attenuation and phase shift values determined beforehand, each pair of said set characterizes a fluid of known composition.

  If the fluid comprises at least two phases in layered layers, the position of said layers can be determined from the compositions determined in step c) and from the definition of the wafers from step a).

  The slices defined in step a) can be superimposed and separated by parallel planes.

  According to the invention, in step c), at least one of the following conditions can be taken into account: the pressure, the temperature and the salinity of said fluid.

  According to the invention, in step c), a model can be used to assign a composition to a family of values measured in step b), the model being constructed from said set of families of values. One can use a statistical or behavioral model of the linear, quadratic and neural network type and one can use a regression of the linear, PLS or neural network type.

  If the fluid is a petroleum effluent comprising water, oil and gas, the model performs operation 1: the family of values is classified by discriminant analysis into one of the following categories: monophasic water, monophasic oil , gas single-phase or multi-phase.

  If the family of values is classified in a monophasic category, then the composition of the fluid is determined.

  If the family of values is classified in the multiphase category, operations 2 and 3 are carried out: Operation 2: the family of values is classified in a subcategory corresponding to a fluid comprising a phase in a proportion comprised in a fixed interval at l 'inside [0%; 100%] Operation 3: the proportion of the phase corresponding to the sub-category is determined by regression on the fixed interval.

  Other characteristics and advantages of the invention will be better understood and will appear clearly on reading the description given below by way of example with reference to the drawings among which: - Figure 1 shows schematically a measuring device, - Figures 2 and 3 show two examples of fluids to which the method according to the invention can be applied, - Figure 4 shows a diagram of the method according to the invention.

  The method according to the invention can use the measurements obtained by means of the measurement device shown diagrammatically in FIG. 1. The measurement device comprises a source 1 of beams of waves connected to the transmitters E arranged on the first line 2. Receivers of R wave beams arranged on a second line 4 are connected to the measuring means 6. A fluid 7 is placed in the space between the first line 2 and the second line 4. The multiplexer 3 makes it possible to connect the source 1 to any one of the emitters E. Thus, the emitter which is connected to the source 1 emits a wave beam which propagates in the fluid 7. It is considered that the beams of waves propagate along straight lines in the fluid 7. The multiplexer 5 makes it possible to connect any of the receivers R to the measuring means 6. The measuring means 6 are also connected to the source 1. Thus, the measuring means 6 can record the characteristics wave beam patterns picked up by the receiver which is connected to the measuring means 6 and the characteristics of the wave beam 25 directly emitted by the source 1.

  Lines 2 and 4 can be in the form of curves, in the form of parallel or non-parallel lines. The transmitters can be arranged at regular intervals on the first line 2 (that is to say that the value of the distance which separates two adjacent transmitters is identical for all the transmitters) and / or the receivers can be arranged at regular intervals on the second line 4 (that is to say that the value of the distance which separates two adjacent receivers is identical for all the receivers). When the transmitters and receivers are arranged at regular intervals on lines 2 and 4, the value of the interval 5 separating two adjacent transmitters can be equal to, or be, an integer multiple of the value of the interval separating two receivers adjacent.

  Lines 2 and 4 can be located on the walls of an enclosure which contains the fluid 7, for example on the wall of a pipe in which the fluid 7 circulates. Lines 2 and 4 can also be located on rods which immerse in the fluid 7 contained in an enclosure, for example in a separator flask used by the petroleum industry.

  The transmitters E and the receivers R are antennas known to those skilled in the art, for example described in relation to FIG. 2 of the French document E.N. 01/16096. The transmitters and receivers can be of identical construction.

  According to the invention, the type of wave produced by the source 1 can be chosen from electromagnetic radiation waves such as microwaves, infrared, nuclear radiation, optical light, or even from waves acoustic such as ultrasound.

  In general, the fluid 7 can be a homogeneous fluid, that is to say a fluid whose composition is substantially identical for any part of the fluid or, expressed in another way, that the constituents of the fluid are distributed substantially identical in the whole fluid. The fluid 7 can also be a heterogeneous fluid, that is to say a fluid whose composition 25 is not identical for any part of the fluid or, expressed in another way, that the constituents of the fluid are distributed in an inhomogeneous manner throughout the fluid.

  In general, the fluid 7 can be a monophasic fluid, that is to say composed of one or more constituents which mix. The fluid 7 can also be a multiphase fluid, that is to say composed of several constituents which do not mix.

  For example, the fluid 7 can be a petroleum effluent comprising water, oil and / or gas.

  With reference to FIG. 2, the petroleum effluent 10 can be in turbulent flow in a pipe 11. In this case the composition of the effluent can be homogeneous, that is to say that the water, the oil and the gases are distributed uniformly throughout the pipe 11. In this case, the method makes it possible to determine the water, oil and gas content of the effluent.

  According to another alternative shown diagrammatically in FIG. 3, the water 20, the oil 22 and the gas 24 can be distributed in the form of laminated layers, for example when the effluent is in laminar flow in a pipe 25 or when the effluent decanted in a separation flask. More specifically, the effluent can be in the form of superimposed layers: a layer of water 20, a layer of water / oil / gas emulsion 21, a layer of oil 22, a layer of oil / gas foam 23 and a gas layer 24. In this case, the method notably makes it possible to determine the position of the different layers and to determine the composition of each layer. The method according to the invention is described in relation to the diagram in FIG. 4.

  Step A: Measurements The measuring device described in relation to FIG. 1 operates by double multiplexing in transmission and in reception. Source 1 produces a wave beam of selected frequency f1. The multiplexer 3 connects the wave source 1 successively with each of the transmitters E. Thus the transmitters emit one after the other a beam of waves. During the transmission time of one of the transmitters E, the multiplexer 5 successively connects each of the receivers R with the measuring means 6. The measuring means 6 also measure the wave beam emitted by the source 1. Thus, the measuring means 6 measure the characteristics of the wave beam emitted 5 by a transmitter and picked up by each of the receivers. In this way, it is possible to measure the wave beam emitted by a transmitter and received by a receiver, for all the pairs of transmitter-receiver available.

  The characteristics of the waves produced by the source 1 are chosen so that the wave beam is modified by at least one of the components of the fluid 7 when it propagates in the fluid 7.

  For example, the frequency f1 of the microwave beam produced by the source 1 can be chosen so that the microwave beam is substantially modified by one of the constituents of the fluid 7 when the beam 15 propagates in the fluid 7, for example the attenuation and the phase shift are sensitive when the beam propagates in the fluid 7. For example, the frequency f 1 is chosen so that the microwave beam is appreciably modified when it crosses natural gas (CH4 ). Thus, by measuring the attenuation and the phase shift of a beam of frequency fl having passed through a petroleum effluent, information is obtained on the content of natural gas contained in this effluent.

  The measurements can be repeated for different wave characteristics produced by source 1.

  For example, the measurements can be repeated for several frequencies of 25 microwave beams. The frequencies can be chosen so that for each frequency the microwave beam is substantially modified by a constituent of the fluid and that at each different frequency, the beam is substantially modified by a different component. For a petroleum fluid containing water, oil and / or gas, the frequencies can be chosen between 1 GHz and 100 GHz, preferably between 2 GHz and 10 GHz, for example in the following four frequency ranges : fa [2 GHz; 2.7 GHz] fb [2.7 GHz; 3.7 GHz] fc [3.7 GHz; 6 GHz] fd [6 GHz; 10 GHz] Step B: Data structure The modifications of the wave beam which has passed through the fluid are determined according to the distance separating a transmitter E and a receiver R, by knowing the characteristics of the wave beam emitted by this transmitter ( ie the beam emitted by the source 1) and the characteristics of the wave beam picked up by this receiver. These modifications are quantified by a family of values, for example by the attenuation, the phase shift, the direction or the polarization of the wave which has passed through the fluid. For example, the attenuation and phase shift of a microwave beam which has passed through the fluid over the distance separating a transmitter E and a receiver R are determined, by knowing the characteristics of the microwave beam emitted by this transmitter. (ie the beam emitted by the source 1) and the characteristics 20 of the microwave beam picked up by this receiver.

  For example, the attenuation and phase shift values can be classified in a matrix of complex numbers whose component Mij corresponds to the attenuation and phase shift values of the frequency beam quivered by the transmitter i and picked up by the receiver j. When the measuring device comprises n transmitters and m receivers, the matrix is of dimension n × m. By carrying out the measurements for the four frequencies fa, fb, fc and fd, one obtains four matrices, each corresponding to measurements carried out at each of the four frequencies.

  When the receivers and transmitters are arranged on two straight and parallel lines at regular and equal intervals, the values located on the same diagonal of the matrix correspond to the attenuation and phase shift values of the beams having traveled a path between a transmitter and a 5 receivers separated by the same distance. Thus, the corresponding measurements of the wave beams having traveled the same distance in the fluid are located on the same diagonal of the matrix.

  Step C: Determining the composition of the fluid To construct a mapping of the fluid 7, the volume in which the fluid is located, that is to say the space between lines 2 and 4, is cut into slices T. Each slice T can be defined by the volume located between two surfaces, for example between two planes, two parallel planes or two horizontal planes. In FIG. 1, the slices T consist of volumes 15 located between horizontal planes, the slices being superimposed. In this case, the slices can be of constant or variable thickness. The slices T can be chosen so that a slice T comprises at least one transmitter E and at least one receiver R arranged on lines 2 and 4. In the slice T, the transmitter and the receiver can be positioned at any location 20 respectively on lines 2 and 4. For example, the transmitter E, respectively the receiver R, can be arranged on line 2, respectively on line 4, on one of the border surfaces of the section T The distance between the transmitter and the receiver is known.

  In addition, an experimental database of the 25 possible configurations of the fluid 7 (for example monophasic, multiphase, stratified, dispersed, emulsified, etc.) and experimental configurations (for example pressure, temperature, salinity, etc.) have been created. .), the composition of the fluid 7 being known. Each experiment gives rise to measurements associated with the configuration of the fluid 7, with the experimental conditions and with the modifications of beams of waves having crossed the fluid. The modifications of wave beams as well as the experimental conditions are characterized by a family of values which includes, for example, the values of attenuation, phase shift, temperature, pressure, salinity. Thus, the data bank groups together, for each experiment carried out on a fluid of known composition, the measurements of the experimental configurations and the values representing the modifications of wave beams having passed through the fluid. The values used can be direct values and / or combinations of direct values (by multiplication, by trigonometric, logarithmic, exponential transformations, etc.) This family of values will be used subsequently as input to the model, described below , to determine the type of fluid (monophasic, emulsified,) and the proportions of each phase composing the fluid.

  From the database, a model is constructed which makes it possible to assign a known composition to a family of values, that is to say that by entering into the model a family of values, it responds by giving the type and the fluid composition. The model is constructed using methods known to those skilled in the art such as a statistical or behavioral model using a polynomial function, a neural network or a discriminant analysis.

  Thus, to determine the composition of the fluid located in a slice T, we use the model previously constructed from the database: the model determines the composition of the fluid from the family of values representing the modifications of the wave beam having passed through the fluid of unknown composition in this slice T. In addition, to improve the accuracy of the estimation of the composition, it is possible to take into account in particular the salinity, the pressure and / or the temperature of the fluid of unknown composition. Salinity, pressure and temperature can be measured by sensors located in the fluid.

  In this case, the database also includes the values of salinity, pressure and / or temperature: each family of values of the database characterizes a known fluid at known salinity, pressure and / or temperature.

  The model built from the database also takes into account the values of salinity, pressure and / or temperature. Thus, the model makes it possible to assign a composition of fluid to a family of values 10 representing the modifications of the beam of waves having crossed the fluid of unknown composition, supplemented by values of salinity, pressure and / or temperature of this fluid of unknown composition. In other words, by giving the model a family of values as well as the salinity, pressure and / or temperature values, it responds by giving a fluid composition. According to the invention, the model constructed from the database can operate according to a succession of operations described below.

  In general, the fluid of unknown composition is likely to comprise several phases. For example, the fluid is a petroleum effluent comprising water, oil and gas.

  To determine the composition of the fluid located in a slice T, the family of values representing the modifications of the wave beam having passed through the fluid of unknown composition in a slice T is used, and optionally values of salinity, pressure and / or temperature.

  Operation 1: First, the families of values in the database are classified in one of the "monophasic" categories or in the "mixture" category. Each "monophasic" category groups together the families of values characterizing a fluid composed of a given single phase. There are as many "monophasic" categories as there are different phases capable of composing the fluid of unknown composition. The category "mixture" groups together 5 families of values characterizing a fluid composed of several phases, for example an emulsion of two phases, of three phases.

  The classification can be carried out using a discriminant analysis which may be of the linear, quadratic or neural network type.

  If the model affects the family of values characterizing the fluid of unknown composition in one of the "monophasic" categories, then the composition of the fluid in slice T is clearly identified: the fluid is composed of 100% of the phase corresponding to the category "monophasic" affected.

  If the model classifies the family of values in the category "mixture", then we carry out Operations 2 and 3.

  In the case where the fluid 7 is a petroleum effluent, the families of values of the database are classified in the following four categories: "water", "oil", "gas" and "mixture".

  From this classification, the model affects, by discriminant analysis, the family of values representing the modifications of the wave beam having crossed the fluid of unknown composition in a slice T in one of the categories.

  If the family is classified in the “water” category, then the following composition is assigned to the fluid in section T: 100% water, 0% oil and 0% gas.

  If the family is classified in the "oil" category, then the fluid of tranche T is assigned the following composition: 0% water, 100% oil and 0% gas.

  If the family is classified in the "water" category, then the fluid of tranche T is assigned the following composition: 0% water, 0% oil and 100% gas.

  If the family is classified in the "mixture" category, the operations 2 and 3 described below are carried out to determine the composition of the fluid.

  In operations 2 and 3, a family of values is treated which has been classified in the mixture category, that is to say that the family of values characterizes a fluid which comprises at least two different phases.

  The proportion of the different phases can be estimated by regressing the interval 0% to 100% for each phase. To improve the reliability of the model, the present invention proposes to cut the interval [0%; 100%] in at least two intervals (Operation 2) and step by step regression is performed (Operation 3).

  Operation 2: The families of values which have been classified in the "mixture" category correspond to multiphase mixtures comprising phases in proportions of between 0% and 100%. These families of values are classified into subcategories. Each subcategory groups the families which correspond to a fluid comprising a phase in a proportion included in a fixed interval within [0%; 100%]. Preferably, for each phase, there are at least two sub-categories whose intervals do not overlap and whose intervals total overlap the interval [0%; 25 100% o]. Thus, for each phase, the families of values are classified in one of the subcategories corresponding to a fixed proportion interval.

  The classification can be carried out using a discriminant analysis which may be of the linear, quadratic or neural network type.

  From the classification into subcategories, we construct the discriminant analysis model allowing to assign one of the subcategories to a family of values characterizing a fluid of unknown composition. In other words, the model indicates, for each phase, to which sub-category 5 belongs to the family of values characterizing a fluid of unknown composition. The model can be of the linear, quadratic or neural network type.

  In the case where the fluid 7 is a petroleum effluent, the families 10 of values of the category "mixture" are classified, for example, in one of the following segments: - the subcategory corresponding to a proportion of water included between 0% and 30%, - the sub-category corresponding to a proportion of water between 15 30% and 70%, - the sub-category corresponding to a proportion of water between 70% and 100%, - the sub-category corresponding to an oil proportion between 0% and 30%, - the sub-category corresponding to an oil proportion between 30% and 70%, - the sub-category corresponding to an oil proportion between 70% and 100%.

  From this classification, a discriminating analysis sub-model 25 assigns the family of values representing the modifications of the wave beam having crossed the petroleum effluent in a section T in one of the sub-categories giving the proportion of water. , and in one of the subcategories giving the proportion of oil.

  Operation 3: When the family of values representing the modifications of the wave beam having crossed the fluid of unknown composition in a section T is classified in one of the subcategories, the 5 proportion of the phase corresponding to this subcategory is determined. -category by performing a regression on the interval corresponding to this sub-category. The regression can be of the linear, PLS or neural network type.

  For example, in the case where the family of values representing the 10 modifications of the wave beam having crossed the petroleum effluent in a T tranche has been classified in the subcategory corresponding to a water proportion of between 0% and 30 %, and in the subcategory corresponding to an oil proportion between 70% and 100%, determines the water proportion by performing a regression on the interval 0% to 30%, and the oil proportion is determined regressing from 70% to 100%.

  Step D: Calculation of the overall proportions By knowing the proportions of the fluids located in each slice T determined in step C and by knowing the geometry of the slices T defined in step C, the total proportions of each component phase are calculated by summation the fluid of unknown composition.

  With reference to FIG. 1, the total volume of the fluid studied is divided into 25 slices T. The slices T are defined by volumes located between horizontal planes, the slices being superimposed. The volume of the fluid in each section T is known. In addition, in step C, the composition of the fluid included in the volume defined by this section was determined for each slice. For example in the case of a petroleum effluent, in step C, the proportion of water, oil and gas contained in each slice T was determined. Considering the volumes of each slice and knowing the proportions of water, oil and gas included in each slice, the overall proportions of water, oil and gas of the total volume of fluid are determined Step E: Determination of the position of the layers of fluids In the application to a stratified fluid as illustrated in FIG. 3, knowing the geometry of the slices T defined in step C and knowing the composition, determined in step C, of the fluid located in each slice, the position of the different layers of fluids. Referring to Figure 1, the total volume of the fluid studied is divided into

  slices T. Slices T are defined by volumes located between horizontal planes, the slices being superimposed. Thus, the position of the different layers of fluids can be determined as a function of the height. The greater the number of wafers, the better the accuracy of the position of the layers of fluids.

  By repeating steps A, B and C, one can follow the evolution over time of the composition of the fluid in the slices T and the position of the interfaces if the fluid is a stratified fluid. Monitoring can be done in real time and represented visually.

Claims (8)

  1) Method for determining the composition of a fluid contained in a volume, in which the following steps are carried out: a) several slices which divide said volume are defined, each slice comprising a part of said fluid, b) a family of values representing the modifications of a wave beam having traversed a path in a first of said sections, c) the composition of the part of said fluid located in said first section is determined by comparing the values measured in step b) with a set of families of previously determined values, each family of said set characterizes a fluid of known composition, d) repeating steps b) and c) for each slice defined in step a), and e) determining the composition overall of said fluid from the compositions determined in step c) and from the definition of the wafers in step a).   2) Method according to claim 1, in which in step b), the attenuation and phase shift of a microwave beam are measured and, in which in step c), the composition of the fluid is determined located in said first section by comparing the attenuation and the phase shift measured in step b) with a set of pairs of previously determined attenuation and phase shift values, each pair of said set characterizes a fluid of known composition.   3) Method according to claim 1, in which said fluid comprises at least two phases in layered layers, and in which the step is carried out: f) the position of said layers is determined from the compositions determined in step c) and from the definition of the slices in step a).   4) Method according to one of the preceding claims, wherein the slices defined in steps a) are superimposed and separated by parallel planes.   5) Method according to one of the preceding claims, wherein in step c) at least one of the following conditions is taken into account: the pressure, the temperature and the salinity of said fluid.   6) Method according to one of claims 1 to 5, in which in step c) a model is used making it possible to assign a composition to a family of values measured in step b), the model being constructed from of said set of value families.   7) Method according to claim 6, in which a statistical model or a behavioral model is used, of the linear, quadratic and neural network type and a regression of the linear, PLS or neural network type is used.   8) Method according to claim 7, in which the fluid is a petroleum effluent comprising water, oil and gas, and in which the model performs operation 1: the family of values is classified by discriminant analysis in one of the following categories: monophasic water, monophasic oil, monophasic gas or multiphase, If the family of values is classified in a monophasic category, then the composition of the fluid is determined If the family of values is classified in the category multiphase one carries out the operations 2 and 3: Operation 2: the family of values is classified in a subcategory corresponding to a fluid comprising a phase in a proportion included in a fixed interval within [0%; 100%] Operation 3: the proportion of the phase corresponding to the subcategory is determined by regression on the fixed interval.