WO2011119045A1

WO2011119045A1 - A method of conducting x-ray tomography measurements of petroleum- containing multiphase flows through a pipe

Info

Publication number: WO2011119045A1
Application number: PCT/NO2011/000102
Authority: WO
Inventors: Bin Hu; Sven Nuland; Chris Lawrence
Original assignee: Institutt For Energiteknikk
Priority date: 2010-03-24
Filing date: 2011-03-24
Publication date: 2011-09-29

Abstract

An X-ray tomography multiphase hydrocarbon flow measurement system using broad-spectrum X-ray sources and two or more sets of sources and transversely arranged detector arrays for tomography calculation of cross- sections of the flow. A velocity profile is calculated from the cross correlation of the sets of sources and detectors. These sources and detectors are housed in a pressure vessel around a pipe with a carbon fibre pipe separating the hydrocarbon flow from the neutral, pressurized atmosphere.

Description

A METHOD OF CONDUCTING X-RAY TOMOGRAPHY MEASUREMENTS OF PETROLEUM- CONTAINING MULTIPHASE FLOWS THROUGH A PIPE.

Field of the invention

The present invention relates to X-ray tomography used in measurement systems for multi-phase petroleum fluid flows in the oil and gas industry. More specifically, the invention relates to a method and an apparatus for determining production rates of oil, water and gas in a multi-phase petroleum- containing flow in a pipe and further for characterising the flow based on X-ray tomography.

Background of the invention

Flow measurement systems, including multiphase flow measurement systems (MPFMS) , available to the oil and gas industry today are limited in accuracy and the ability to provide sufficient amount of detailed information.

A producing oil field reservoir may cover several concessions with different ownership. Thus different wells may have different ownership. For practical reasons they may be produced to the same platform, through the same process plant. Then the question of allocating the production to different wells arises, stimulating the need for good multiphase flow meters that can determine the production of water (which might be a liability whose processing has to be paid for) and the oil and gas, both with different values and costs.

Generally reservoirs produced through wells to a stationary platform achieve a better recovery of reserves than reservoirs produced through subsea wells. This is due to better monitoring of production. It can be discrete events that waste the producing energy, such as gas break-through, or water breakthrough which means that volumes of oil are left stranded with no communication to the well . Generally there is an optimal production profile, based on extensive reservoir studies that should be followed to achieve the best recovery of reserves. The first kind of destructive event could be detected by a phase fraction meter, where the ability to measure even small fractions of one phase, principally water, is important. For the more general control of production a full multiphase flow meter is needed.

In general, the liquid phases will often have a structure due to entrained bubbles and droplets that can be used to correlate the flow rates and thereby production.

Multiphase flow meters generally consist of one unit for phase fraction determination and another for flow rate determination usually based on pressure drop measurement, but in some cases also on cross correlation of flow structures .

There are mainly two classes of methods for phase fraction determination, dual energy gamma densitometers, and electrical methods such as microwave and impedance measurements .

Dual energy measurement is an established method, although in the field there is the problem of health and safety associated with the use of radioactive sources, especially so because stronger sources are needed than are used in the laboratory. There is also a problem with changes in salinity which change the radiation absorption.

The electrical impedance and microwave methods are used especially to improve the determination of water fractions. The basic problem is that the response depends very much on the spatial distribution of the phases . It is especially crucial whether the water forms a continuous phase or not. Therefore it may perform poorly in the transition region between continuous and dispersed water. Falcone et al . state that despite success claims, the application of such devices remains questionable. For known fluid properties pressure drop measurement is a good method for determining flow rates. However, for multiphase flow the rheological properties such as viscosity, may be influenced by the degree of mixing and the phase fraction, especially around the inversion point for oil-water emulsions. Gas in oil foams may also influence rheology, but to a degree that is less well known. These changes in rheology can drastically change the relationship between flow rate and pressure drop.

Cross correlation methods determine the speed of some large scale features of the flow, usually slugs due to the poor spatial resolution. However, the relation between slug

velocities and flow rates is complex and not unique. Therefore this method is questionable and needs extensive calibration.

Tomography is best known from is medical implementation, in which a detailed picture is constructed after illuminating the specimen from many angles as the instrument or the object rotates. This requires a long exposure time, and is not of practical use in multiphase flow, where the features of interest move at high speed.

Detailed determination of phase distribution over the pipe cross section in the flow of oil, gas and water, with possible further complications due to sand etc, is of fundamental importance in understanding the flow and building improved models for design and analysis of multiphase transport

facilities. It also enables field application such as in multiphase meters or in monitoring equipment such as water cut meters or detection of changes in salt content of water which might indicate conditions requiring attention.

The mainstay of phase fraction measurement has been the gamma densitometer, which relies on the same physical phenomenon of radiation absorption as the X-ray meters discussed here. They differ in the detector technology and the radiation source. They come either as a narrow beam or a broad beam version. The narrow beam has a small detector and gives the phase fractions along one line, while the broad beam has a detector that covers the whole pipe section. The spatial and time resolution are low due to the relatively low intensity of the gamma sources. The sources used are radioactive isotopes. This has two drawbacks. First as high radiation intensity is needed to get good counting statistics at high sampling rates, the strong sources needed become dangerous and cumbersome to handle; while the X-ray source can be turned off, the isotopes cannot. The other drawback is that the photon energies of the isotopes may not have the values that are needed, nor may the strength of the useful spectral lines relative to other lines be what is desired to distinguish between oil and water. The X-ray source has a greater freedom in deciding the energy of the maximum intensity, but gives a continuous spectrum of energies, and with absorption coefficient that changes with energy, this introduces new problems . The narrow beam densitometer has also been used as a scanning or traversing device in order to determine average phase fractions as a function of the height above the pipe bottom. This type of measurement is restricted to stationary flows without important transient phenomena such as waves or slugs, due to the time needed to traverse the cross section of the pipe.

Imperial College has pioneered the use of X-ray tomography for multiphase flow phase fraction measurement, using photon energies that enable three-phase discrimination. However, lacking detectors with the necessary energy discrimination, they have resorted to moving filters to give alternating high and low energy exposure. The result is a low time resolution, with 5Hz as maximum sampling rate.

New development of X-ray detector technology has made detailed examination of two-phase gas liquid flow possible through a tomography approach. Hu et al. (2009) describes a similar device using more modern detectors. In this device three-phase capability is obtained using the fact that the detectors have a width of about 6 mm and 64 pixels in the flow direction. Half of these are covered by a low energy filter in the form of a thin metal film. Half the pixels are exposed only to high energy radiation, half to a mixture of high and low energy. Due to the difference in absorption of these two energy spectra, it is possible to distinguish between oil and water. But this comes at some costs. First there is a loss in count rates, resulting in higher statistical error, which is compensated for by a reduction in sampling frequency. Second, there is a blurring of the spatial resolution. Use of high-pass filters (metal foil) allow to exploit the different X-ray absorption in oil and water at low energies to discriminate between three- phases: oil, gas and water. However, this leads to loss in accuracy due to lost count rates due to absorption in the foil. Furthermore, the discrimination achieved is not very good since the method compares absorption of strongly overlapping spectra. The key to better three-phase discrimination lies with the detector. New detectors are being developed combining small pixel size with multi-energy discrimination.

Flow measurements using tomography are well known and state of the art is discussed in:

Bin Hu, Morten Langsholt, Sven Nuland, Chris Lawrence, Void distribution in liquid layer in stratified wavy flows measured by an X-ray CT instrument , 14TH International conference

Multiphase Production Technology, Cannes, France, 2009.

Hu, B. , Stewart, C. , Hale, CP., Lawrence, C.J., Hall, A.R.W. , Zwiens, H. , Hewitt, G.F. Development of an X-ray computed tomography (CT) system with sparse sources: application to three-phase pipe flow visualisation. Exp. Fluids, 39 (4) , 667-678 (2005)

F. Ricard, C. Brechtelsbauer, X. Xu, C. Lawrence and D.

Thompson. Development of an electrical resistance tomography reactor for pharmaceutical processes . Canadian Journal of Chemical Engineering 83(1), 11-18 (2005).

F. X. Ricard, C. Brechtelsbauer, X. Y. Xu, C. J. Lawrence, Monitoring of multi -phase pharmaceutical processes using electrical resistance tomography. Chemical Engineering Research and Design (Transactions of the IChemE, Part A) 83 (A7) , 794-805 (2005) .

Falcone, G. , Hewitt, G.F., Alminti, and Harrison. B. Multiphase flow metering: Current trends and future developments. SPE paper 74689 10 An early disclosure is in:

M.C. Clarijs & al . , Optimized X-ray spectra for multiphase-flow measurements, DGZfP Proceedings BB 67-CD. 153. Poster 1,

(1999) , (www. dgzfp . de/Portals/24/PDFs/BBonline/bb_67- CD/bb67_p01.pdf)

Linear inverse solutions and ill-posed problems are discussed in Grave de Peralta Menendez, R. , Murray, M.M. and Gonzalez Andino, S.L. Improving the performance of linear inverse solutions by inverting the resolution matrix. IEEE-TBME 2004 and in

A. Tikhonov and V. Arsenin , Solution of -Ill-Posed Problems. , Wiley & Sons (1977)

Spectral X-ray computed tomography (CT) imaging technology is disclosed in patent US7170049, Pixelated cadmium zinc telluride based photon counting mode detector, and in patent application US20080099689, Photon counting imaging detector system. Such a detector can determine the energy of a photon within an energy window where the boundaries can be set.

Summary of the invention

The present invention is advantageous in that it provides minimal flow intervention, significantly improves volumetric fraction measurements, spatial and temporal resolution, phase velocity measurements and flow pattern visualization. In an embodiment of the invention the system comprises sources and detectors that are thermally, barically and chemically insulated from the flow. A possible configuration is for them to be housed in a pressure vessel around the pipe with e.g. a carbon fiber pipe separating the hydrocarbon flow from the neutral but pressurized atmosphere around it.

The sources are broad- spectrum X-ray. The system has detectors that determine the energy of a photon within energy windows, the boundaries between these can be set through software .

The system uses two or more sets of X-ray sources and detectors in parallel. The computed cross correlation of these systems gives the velocity profile. Together with the spatial

distribution of the phases, this gives the production of each phase . The system comprises software that is run on one or more computers connected to and controlling the sources and

detectors. One or more computers are also used for computing the results and presenting them graphically. According to one aspect of the invention it is a method of conducting X-ray tomography measurements of petroleum- containing multiphase flows through a pipe,

- arranging an X-ray transparent pipe section (2) as an in-line part of said pipe of generally the same diameter;

- emitting, at a first position along said pipe section (2) , wide-angle beams of X-ray photons with a broad energy spectrum from two or more first pairs of first X-ray generators (6a, 6b,..), transversely across said pipe section (2) towards one or more first X-ray sensor linear arrays (7a, 7b, ..) , and - emitting, at a second position along said pipe section (2) , wide-angle beams of X-ray photons with a broad energy spectrum from two or more second pairs of second X-ray generators (6a, 6b, ..) transversely across said pipe section (2) towards one or more second X-ray sensor linear arrays (7a, 7b, ... ) ,

- said X-ray generators (6a, 6b, ..) radiating X-ray photons under first and second different angles (al, a2,..) about an axis through said pipe section (2) ,

- said first and second positions have a separation (L) sufficiently short to allow persistence of flow structures, characterised in that

- using X-ray sensor linear arrays (7) subdivided into pixel sensors (71) and arranged transversely with respect to said pipe section (2) ,

- conducting single X-ray photon mode counts sorted into two or more energy levels of higher-energy photons and lower-energy photons from an integrated circuit connected to each said pixel sensor (71) ;

- sampling said sorted counts at said first and second

positions generally simultaneously and at a sampling rate generally higher than a flow time over said separation (L) ;

- calculating two-dimensional tomographic image sections of said flow at said first and second positions based on said energy sorted single X-ray photon mode counts; and

- calculating correlations of said flow structures of at least one phase of oil, gas or water phases of said multi-phase flow between said first and second positions based on said

tomographic image sections .

In another aspect the invention is the apparatus defined in the attached claims. Further inventive embodiments are indicated in the dependent claims of the method and the apparatus,

respectively .

Some differences and advantages of the invention over the prior art

Schlumberger ' s WO1999060387 presents a method of characterising a multiphse mixture flowing along a flow path, generally comprising the following steps:

(i) directing X-ray photons toward a first location along the flow path of the mixture; (ii) collecting photons toward a first location along the flow path of the mixture and producing a signal in response to the collected aggregate energy,

(iii) directing a plurality of X-ray photons toward a second location along the flow path of said mixture;

(iv) collecting multiple photons of X-rays transmitted through said second location and producing a second signal in response to the collected aggregate energy; and

(v) analysing said first and second signals.

In page 3, third paragraph, it states that for a flow velocity of 10 m/s, 10 microsecond "flashes" are used. The fluid moves 0.1 mm during the flash. If an interrogation region of 1 cm is used, 1% of the fluid moves out of the interrogation region during the flash. WO1999060387 further states that "In general, for a three component fluid, two independent measurements must be made during each flash in order to completely characterize the fluid concentrations. However, the rather crude image of fluid concentrations provided by WO1999060387 may not be sufficient for forming tomographic images for structures less than so-called "slugs".

In general, the present invention has several advantages over Schlumberger ¹ s WO1999060387. Considering velocity measurements, WO1999060387 describes the use of a Venturi restriction so as for flow rates to be determined via pressure differential measurements. However, there may be different velocities for gas, oil and water. The velocity may also be calculated by correlations of "flash" X-ray measurements.". The present invention works without a Venturi and thus does not introduce any throat which would otherwise disturb the flow to be measured, whereas the present invention does not require a Venturi restriction but correlates detected detailed structures that are sufficiently persistent to appear at the sensors at both the first and the second locations. Further, a Venturi meter for multi-phase is usually arranged with its axis vertically, a feature which deviates from the pipe if the pipe has an orientation other than vertical, such as for horisontal pipeline arrangements.

Considering resolution properties, the resolution of

WO1999060387 is significantly less than for the present invention. The sensors of WO1999060387 are not subdivided into pixels. Each sensor is rather large, using an interrogation region of about 1 cm as stated above, it thus provides an average measurement over a rather large area, incurring insufficient resolution of details of the flow. The present invention uses sensors of smaller extent which provides finer details .

With regard to energy discrimination, in paragraph four on page 3, WO1999060387 states that only the flux of X-ray radiation is measured, which is in line with the term "the collected aggregate energy" above. Thus the energy of individual photons are not possible to discriminate; two photons of 30 keV cannot be distinguished from one photon of 60 keV other than through use of filters . The energy distribution of the photon flux received is not measured directly but is assessed through the use of a high-pass filter. WO1999060387 uses an X-ray source with a continuous spectrum up to 60 keV and a low energy detector generally sensing between 10 to 25 keV, a filter and a so-called "black" totally absorbing high energy detector generally sensing between 45 and 60 keV, please see Fig. 2 a, b, c of WO1999060387. The energy distribution of the photons captured is thus subdivided through the use of one or more filters and/or the use of two sensors.

In an alternative embodiment of WO1999060387 the energy distribution of the photons captured may be varied using an X- ray source sending out different energies in different pulses such as alternating between a 30kV and a 60kV power supply for the same target for X-ray emission, please see Figs. 3a and 3b of WO1999060387. Such an arrangement of X-ray generators requires much space and may more space than is actually available for a tomographic arrangement of two, three or more X-ray sources arranged around a pipeline at the same axial position for a tomographic section arrangement .

The present invention uses sorting incoming photons into distinct energy bins set in the sensor, please see Fig. 4b. This allows selecting sensor data from the lower energy for better to discriminate between water and oil, a contrast which is reduced for higher-energy photons. It is thus advantageous to have an X-ray transparent pipe section such as a carbon fibre pipe section which allows passage also of the lower- energy photons such as in "band 1" of Fig. 4 with a good energy range for discriminating water and oil .

With regard to the tomographic arrangement, tomographic measurements are suggested WO1999060387 using several detectors about the circumference of the venturi restriction. Such an arrangement of several separate, single sensors in WO1999060387 is not specified but it would provide a rather crude

tomographic image, while the present invention uses a setup of several linear arrays of sensors with the linear array arranged transversely to the path of the fluids. Using the detectors described in WO1999060387 each of the detectors will require too much space to enable a detailed array required for

conducting tomographic analysis required for finding flow structures other than entire so-called slugs of water or gas. Such coarse structures ¹ velocities do not represent the velocities of the components of the flow.

With regard to X-ray attenuation, WO1999060387 describes measurements conducted on a pipe. In page 15, line 13, it is specified that the pipe in question must have a window in the Venturi restriction wherein the window is transparent to X- rays . In an embodiment of the present invention there are less limitations with regard to the pressure of the three-phase flow of fluids in the pipe. The X-ray sources and the sensors are arranged in a pressurised neutral gas chamber which envelopes an X-ray transparent carbon fibre pipe, with a relatively small pressure gradient across the carbon fibre pipe wall. This allows for a simpler arrangement of transversely arranged linear arrays of X-ray sensors without limitations imposed by narrow window slits. Further, ring-shaped window slits which would pose significant pressure restrictions are avoided. US7170049 to Iwanczyk describes a rectangular array of

pixelated cadmium zinc telluride based photon counting mode sensors. One advantage of Ivanczyk's patent is the very small size which may be achieved for each pixel of the sensor: about 1 mm. Each pixel of the sensor may be coupled to an amplifier wherein programmable threshold signals may be set to

discriminators which may discriminate between high-energy photons and low-energy photons. Thus the energy of discrete incoming photons may be measured. The pixel sensors of Ivanczyk are arranged in a rectangular grid subdivided into e.g. 16 x 16 pixels. However, such rectangular arrays provide a large number of pixel measurements from a rectangular sensor area which requires much processing when strictly only one long single strip of pixels across the direction of the pipe is required. An advantage of the sensors used in the present invention is their linear arrangement in a 1 x 160 arrangement which may be arranged transversely to the flow direction of the pipe.

Sensors integrating two or more pixels in the flow direction may also be used. Other arrangements with more pixels along the sensor may be imagined. Instead of capturing a rectangular image of a "volume" passing by, rapid acquisition of several linear transverse images is conducted while the multi-phase flow with its structures pass, and the linear images of different orientation are processed and combined to a

tomographic image of the successively passing so formed sections of the flow.

Brief description of the drawings The following detailed description of the preferred embodiment is accompanied by drawings in order to make it more readily understandable . In the drawings : Figure 1 is an X-ray tomograph with two detector (X-ray camera) source pairs .

Fig. lb is an illustration of an X-ray tomograph arrangement according to the invention using three X-ray source - X-ray detector pairs. The apparatus is illustrated in a section across the pipe section with the sources and sensors .

Fig. lc is a longitudinal section view through a first and a second position along an x-ray transparent section of a pipe in line with the pipe conducting the three-phase petroleum fluid flow, corresponding to the setup of Fig. lb. An X-ray source and detector pair is illustrated at each positions. There are at least two, preferably three such pairs of source and detector.

Fig. Id comprises illustrations of linear arrays of X-ray sensors (7, 71) used in the method and the apparatus according to the invention.

Figure 2a shows time-history projection views of X-ray

measurements a multi -phase flow as registered by two transverse detectors .

Figure 2b shows a sequence of images from two detectors .

Figure 2c shows a tomograph and void fraction profiles

reconstructed from data from two detectors

Figure 3 shows the results of two-dimensional reconstruction from 1 -dimensional projections.

Figure 4 shows the difference between X-ray absorption between water and a carbon-rich material.

Fig. 4b shows in the lower part of the diagram an X-ray spectrum between 10 keV and 70 keV as may be produced by the X- ray sources of an embodiment of the invention, and in the upper part of the drawing is shown diagrams of the different

attenuations of the three phases water, oil and gas. The attenuation coefficient scale is in units of "/cm". Figure 5 shows X-ray tomography for a high pressure pipeline compiled from a series of line sensor captions as a function of time . The upper image is a top view and the lower image is side view . Fig. 5b presents reconstructed tomograms of a flow structure, here a slug, passing through the sensor.

Description of embodiments of the invention In the following, the present invention will be discussed and example embodiments described.

The present invention is a method of X-ray high-temporal and high-resolution tomography measurement of a petroleum- containing multiphase flow, allowing distinguishing between three phases, oil/gas and water, and allowing computation of flow velocities of at least one of the phases. The invention is also an X-ray tomography system for such a multiphase flow with high temporal and spatial resolution.

The system of the present invention has a detector arranged for determining the energy of a photon within energy windows, the boundaries between these can be set e.g. through software. The individual pixels of such detectors are preferably according to Iwanczyk's patent application US2010/0316184 published 16. Dec 2010 describing rectangular arrays of such pixel detectors, after the priority date of the present invention, and according to the pixels of otherwise rectangular arrays of such pixel detectors described in Iwanczyk's patent US7170049.

The present invention uses sorting incoming photons into distinct energy bins set in the sensor, please see Fig. 4b. The energy level of individual photons is measured and a distribution into e.g. four energy bands may be used, such as Band 1: 10-20 keV

Band 2: 20-30 keV

Band 3: 30-40 keV

Band 4: 40-70 keV.

The separation between energy bins are set electronically in the sensor.

Fig. Id comprises illustrations of linear arrays of X-ray sensors (7, 71) used in the method and the apparatus according to the invention. In the upper left part a detector module card with a linear sensor array module of width 1 pixel and a length of 32 pixels. The detector module card may be arranged side by side so as for forming a number of e.g. five consecutive linear arrays, here forming 1 x 160 pixels sensors. A cross section of the detector module is seen in the middle upper part of the illustration. Each module may comprise an application- specific integrated circuit "ASIC" with parallel circuits with

discriminators and counters for lower and higher energy X-ray photons similar to what is illustrated in Fig. 4A of

US20100316184. The detector module is provided with a connector to a motherboard which is illustrated in the upper right part of the drawing. One linear X-ray sensor array according to the invention may be assembled on the motherboard and enclosed in an aluminium profile with an X-ray window along the array of sensor pixels, as illustrated in cross-section in the lower right part of the drawing, and is an enlarged cross- section of the detector cross-section illustrated in Fig. lc.

In a method aspect the invention it is a method of conducting X-ray tomography measurements of petroleum-containing

multiphase flows through a pipe,

- emitting, at a first position along said pipe section (2) , wide-angle beams of X-ray photons with a broad energy spectrum from two or more first pairs of first X-ray generators (6a, 6b,..), transversely across said pipe section (2) towards one or more first X-ray sensor linear arrays (7a, 7b,..), and

- emitting, at a second position along said pipe section (2) , wide-angle beams of X-ray photons with a broad energy spectrum from two or more second pairs of second X-ray generators (6a, 6b, ..) transversely across said pipe section (2) towards one or more second X-ray sensor linear arrays (7a, 7b, ... ) ,

- said X-ray generators (6a, 6b, ..) radiating X-ray photons under first and second different angles (al, a2 ,..) about an axis through said pipe section (2) ,

- said first and second positions having a separation (L) sufficiently short to allow persistence of flow structures, characterised by

- said X-ray sensor linear arrays (7) subdivided into pixel sensors (71) and arranged transversely with respect to said pipe section (2) ,

- using an integrated circuit connected to each said pixel sensor (71) conducting single X-ray photon mode counts and sorting said counts into two or more energy levels of higher- energy photons and lower-energy photons

- sampling said sorted counts from said one or more linear arrays (7) of pixel sensors (71) at said first and second positions at a sampling rate generally higher than a flow time over said separation (L)

- calculating, based on one or more of said sorted single photon counts from each pixel sensors (71) from said linear arrays of X-ray sensors (7) , two-dimensional tomographic representations of X-ray attenuation representing one or more phase fraction distributions of oil, water or gas in sections of said flow at said first and second positions; and

- using said phase fraction distributions in said sections at said first and second positions, calculating correlations of patterns representing persistent recognisable flow structures .

The tomographic X-ray attenuation is calculated based on that the X-rays originate from at least two X-rays sources and are detected at least one linear X-ray sensor array, for the first and second positions . Each single pixel measurement represents only an average of the X-rays traversing the fluid between the source and the sensor. The instrument may be calibrated using fluids of known compositions . One may calibrate the instrument using samples of only water, only oil, and only gas. One may also conduct the calibration for all the energies.

With the term "a sampling rate generally higher than a flow time over said separation (L) " is meant that the time interval between two consecutive samples and thus tomographic sections of the flow is far shorter than the time for the multi-phase fluids to flow from the first to the second positions.

The phase fraction distributions at each said first and second position may be solved from a set of equations with respect to phase fraction in a two-dimensional pattern based on the X-rays from the at least two X-ray sources detected at the arrays. The so mapped phase fraction distribution at the first and second two-dimensional phase fraction distributions, representing either oil, water or gas in each position, may then be

correlated for persistent flow structures for different time intervals between the first and second position. Different phases will usually have different velocities. The correlation of different structures in the flow will allow calculation of the different velocities of different parts of the three-phase flow. Based on the correlations, velocity profiles through the flow may be calculated. Using the different velocities and phase distributions the flow velocity of each phase through the apparatus may be calculated. The results may be used as a basis for fiscal calculations or technical considerations about the three-phase flow.

According to an embodiment of the method, one may discriminate between oil, gas and water phases respectively, in said tomographic image sections based on calculated X-ray

attenuation of pixel volumes of said flow based on

corresponding calculated tomographic pixels.

Further, selected higher- or lower-energy bin pixel

measurements may used for calculating X-ray attenuation differences between particularly oil and water.

There are several advantages of the method and the apparatus according to the invention: Using an integrated circuit in a high-density readout X-ray single photon counting mode for each said pixel sensor (71) allows the tomographic imaging of a correspondingly high number of cross-sections of the multiphase flow as the flow is "sectioned" into a continuous stack of circular tomographically mapped transverse sections of the three-phase flow. Rectangular sensors initially tried proved to provide two-dimensional area-wide measurements which were slow to sample and which provided unnecessary sensor data in the length direction of the pipe section. Sampling a fluid volume at each position as is done with a rectangular array of sensor pixels is strictly not desirable because the flow structures may change rapidly. Thus the 2D sensors of US7170049 are not practically adapted to the present invention. The present sensor setup with one pixel in the flow direction and n pixels across the pipe significantly reduces the sampling time at each of the two positions and allows a far higher number of relevant tomographic cross sections of the passing flow structures per time unit, which allows detailed measurement of fine structures of the flow, such as eddies, bubbles, droplets, which all may improves the calculation of mass transport through the sensor.

The attenuation coefficients of oil and water in large parts of the energy spectrum, particularly for higher energies, e.g. near 70keV, are not very different, please see Fig. 4b. Due to the fact that low-pass filtering is not available for X-rays, filtering may not feasibly be used to discriminate the lower- energy photons, and subtraction between one filtered and one unfiltered pixel would require more sensors. The attenuation differences for lower energy spectra in the bands between lOkeV, 20keV, 30keV and 40 keV are better. For the attenuation differences between gas and oil or water the difference is significantly higher and good even near 70keV. Utilising the features of each said pixel sensor (71) to discriminate at least between two energy levels of higher-energy X-ray photons and lower-energy X-ray photons, and using an integrated circuit in a high-density readout X-ray single photon counting mode for each said pixel sensor (71) will provide better contrast in the resulting "grayscale" binned data of each sensor and thus provide better data as input to the tomography mapping of each cross-section, and thus improve discrimination between oil and water. Within wide radiation ranges all photons may be counted, and little smearing between high- and low-energy photons will occur, thus less radiation is required while higher tomographic resolution is allowed.

A significant advantage of omitting a Venturi pipe by having a full-bore pipe section (2), i.e. having the same diameter as the pipe, avoids disturbing the multi-phase flow. The present invention has the advantage over the Venturi pipe section of WO1999/60387 that it may have any orientation as opposed to a Venturi-meter which usually has to run vertically, interrupting an otherwise horizontal pipeline and introducing obstacles to other parts of the plant or making the pipeline more vulnerable to mechanical damage from passing ROV's, wires, or mechanical tools .

A further, rather practical advantage of omitting a Venturi restriction of any kind by having a full-bore pipe section (2) , i.e. having the same diameter as the pipe, allows the passage of a so-called "pig" through the apparatus of the invention, thus not restricting inspection and maintenance of the pipe unnecessarily .

Sorting energies in bins through programming energy bands into the receivers further avoids using several X-ray sources of different energies such as 30 keV and 60 keV sources in the same source position, as exemplified by Figs 3a and 3b of WO1999/060387. Such double power supplies to the same radiating target may be rather voluminous and the space available for such multiple generators for each radiation angle is small. In the preferred embodiment the present invention makes use of only one X-ray source per radiation angle for the first and second position along the pipe section.

With a detector with n pixels the system determines the distribution of the phases over the cross section in n² spatial elements. One exposure yields just 2n equations to determine the 2n² unknowns. In medical tomography the necessary

information is gathered by rotating the source and detector and exposing the object at many angles, each exposure giving a new set of n equations. This is not possible for fluid flow because of the high time resolution necessary, a sampling rate of 100 to 300 Hz or more. In the present invention use several X-ray source and X-ray linear sensor pairs are used. Additionally it is possible to illuminate the same detectors from several angles using pulsed X-ray sources in sequence.

Still it is not possible to gather enough information to obtain a fully determined system of equations . Software is thus used using mathematical methods to obtain as much information as possible from such underdetermined systems .

Having obtained sufficient spatial resolution it is possible to determine the velocity profile from the cross-correlation of the pattern of such flow details (droplets, bubbles, particles) at two axial positions . This makes it possible to study in detail several fundamental multiphase flow problems, e.g. gas leakage through slugs.

Tomography is best known from is medical implementation, in which a detailed picture is constructed after illuminating the specimen from many angles as the instrument or the object rotates. This requires a long exposure time, and is not of practical use in multiphase flow, where the features of interest move at high speed. To obtain spatial resolution at high time resolution, the pipe cross section is illuminated from fixed angles. From practical and cost concerns only a limited number of angles, until now two or three, can be used.

Figure 1 and lb shows illustrations of X-ray tomographs with two, respectively three X-ray detector - X-ray source pairs. Each detector source pair of Fig. 1 produces a set of

instantaneous "shadow" images as seen in figure 2.

If the number of pixels in the camera normal to the flow direction is n, and we assume that the pipe cross section is divided into (it/4)n² squares, we have 2(π/4)η² unknown phase fractions and each ray from the source to a detector pixel defines two equations. If we have X-ray sensors with m energy bins, here at least two, but preferably three or four, each ray defines m equations. Thus unless we have illumination at almost as many angles as we have pixels in each direction we have an underdetermined set of equations. This is clearly not directly solvable and one must resort to pseudo inverses to approximate the phase distribution. This is a field of wide interest as similar underdetermined systems are important in many areas from geophysical measurements to PET scanning, thus many possible ways are well known to those skilled in the art . The result of such reconstructing cross sectional phase fraction distribution is shown in figure 3. Due to the limited number of illumination angles it can be seen in figure 3 that "ghost" structure arises in the reconstruction process. These might interfere in a correlation process for example, where one wants to infer flow velocities from two closely placed tomographic planes .

A significant improvement represented by the method and system of the present invention is the ability of the system's detectors to utilise the discrimination between different energy levels. A further advantage is the ability to sort individual photon counts into the different energy levels, instead of integrating and smearing the photon energies as with WO1999/060387 which has no single photon counting and no energy binning. The present invention opens the possibility to use more precise values for the absorption coefficients that change rapidly with photon energy, please see figures 4 and 4b. This is an improvement over averaging the absorption coefficient over a wide energy range. The ability to discriminate between oil, water and gas is important both for laboratory and field applications. To distinguish between phases using X-rays, the absorption coefficients have to be different for the different phases. Unfortunately for photon energies higher than about 70 keV, the X-ray absorption depends entirely on the density of the material, and the density of water and oil is sufficiently close to make discrimination impossible. However, for energies below 70 keV there is a difference in absorption coefficient between water and carbon rich oil . Figure 4 uses polyethylene as an example of a carbon rich material, similar to oil, and compares this to water. It can be seen that the difference between the X-ray attenuations of oil and water increases with decreasing photon energy.

In an embodiment of the present invention a band of pixels is covered with a high pass filter, in the form of a thin copper film. Thus half the pixels receive only the high energy part of the spectrum, the other part the whole spectrum including the low energy part. This allows us to define two equations for the absorption of the radiation, where the absorption coefficients are sufficiently different to give linear independent

equations. However, due to the broad energy spectrum the equations are poorly conditioned, which increases the

uncertainty in the discrimination.

Thus to improve the discrimination between gas, oil and water it is crucial to discriminate between different photon

energies, and not mix too wide ranges of energy in the

detector. Photons are sorted into energy bins in the detectors such as described in Iwanczyk US7170049, please see also the attached Fig. 4b, with a dividing energy that is set by software . A rectangular pixelised detector with multiple bins is described in Iwanczyk' s application US20100316184. A number of bins for each individual sensor pixel can be introduced at the expense of more complex electronics. With two energy boundaries (three bins) that can be set in software the present invention has an unprecedented flexibility in finding the optimal configuration. The use of three or more bins also introduces redundancy, hence robustness, into the

discrimination .

The accuracy of the two-dimensional reconstruction depends very much on the number and arrangement of detector source pairs . There is an increase in quality with the number of such pairs until we have a completely determined system of equations.

However, the cost and complexity grow very fast, and there is a rapidly diminishing return of such investment in terms of accuracy.

In one embodiment the detectors are grouped in a semicircle and illuminate the detectors from several angles. With more than two angles, some detectors will be illuminated from two or more sources, and it is necessary to use pulsed sources and a synchronization algorithm to keep track of which data belongs to which angle/source.

It is very expensive to test all such arrangements in practice . Therefore a simulation software is used that can predict the X- ray raw data from a given phase distribution, and these data are used to study the reconstruction algorithm functions for a given source configuration.

For the reconstruction algorithm we assume that the pipe cross section is divided into square elements with the same length as the pixel size of the detector. For each element there is an unknown phase fraction. For the sake of simplicity we discuss the two-phase system and use the fact that phase fractions sum to 1. Thus the number of unknowns increases with the square of the pixel number n. For each pixel we can describe the

absorption along a ray from the source through a linear equation. So for each detector-source pair we get n equations. The basic problem is that the information gives an

underdetermined system for all practical numbers of

illumination angles. Formally we have a set of linear

equations : Ax = b Equation 1

Where A is an (n²)x(mn) matrix of rank mn where m is the number of detector source pairs.

We obtain an approximate solution: x = A^*b

Equation 2 Using equation 1 we have x = A Ax = Rx „ . . _

Equation 3

We want a pseudo inverse which makes R as close to the identity matrix as possible, Grave de Peralta Menendez et al, 2004. Thus the reconstruction problem can be formally stated as a

minimization problem, but due to the high dimensionality this is practically impossible to solve directly. Therefore

simplifications are needed and there are many examples in literature on this question.

The Data processing and tomographic reconstruction process includes finding efficient way to extract and convert the raw count values to values of phase holdups and velocities for each illumination angle. The holdup values calculated from all different illumination angles are then used to reconstruct tomographs which allow us to visualize internal flow

structures .

Most importantly in the present invention, we mix energies with very different absorption coefficients. The reconstruction algorithm is based on averaging the properties, and the good results rely very much on calibration.

With the great spatial and temporal resolution possible with the present invention it is possible to follow structures in the flow by using two sets of detectors separated by a suitable distance, possibly close enough to be illuminated by the same sources. With the much better spatial resolution it is possible to follow much smaller structures than the waves and slugs that are possible with present equipment, and thus correlate structures that are directly related to local fluid speed.

Coupling holdup distributions with the velocity measurements, one can also obtain the flow rate of each phase.

The key factor is the spatial resolution obtained, the time resolution will be sufficient. The present invention may not be able to resolve individual sub-millimeter bubbles or droplets, but these will not be uniformly distributed but form structures that will survive a short distance. Thus there is a possibility to determine their velocity by cross correlation.

With three phase fraction determination, especially a reliable water fraction determination, and a velocity profile we can determine production rates. However, not all of the phases have structures that can be followed by cross correlation, most importantly the gas phase. The liquid phases will for most practical production rates have droplets of the other phase entrained. For higher production rates showers of droplets entrained in the gas phase create structures that can be followed. But generally the gas phase is without sufficient structures for the velocity to be determined by cross

correlation.

Thus, to overcome this problem, in one embodiment the phases are mixed using static mixers.

A number of modelling problems have resisted attempts at solution due to lack of methods to examine the detailed flow structure. The present invention, a tomography system with good spatial and temporal resolution as well as good three-phase capability enables cross correlation to determine local velocities, at least in part of the flow where there are structures that can be followed. This make possible completely new ways of understanding what is going on: - Gas leakage through slugs may be crucial to the modeling of slug flow, and in many situations it is the most sensitive parameter and the least known in the model.

With most prior art instrumentation we can only see part of the problem, namely gas fraction in slugs, not the real parameter: the leakage of gas through the slugs - relative to the slug body. With the instrument of the present invention we can measure both the total bubble fraction, its spatial distribution, and through cross correlation the mean velocity profile, the problem is attacked much more directly.

- Droplet generation and distribution are important for modelling the accumulated liquid fraction in a stratified flow. This is especially important for long gas condensate transfer lines, in both design and operational analysis. Until now this problem has been examined experimentally using isokinetic probes to extract samples from the flow. This is cumbersome, which limits the parameter space that has been examined. Fast transients associated with waves for example are averaged out . Waves are important for droplet generation, and the present invention makes it much easier to examine this correlation.

- Three-phase mixing is important for the relative accumulation of the three phases, but also for determining dispersion properties such as viscosity or yield stress. However, the dispersions, for example a gas-in-oil foam, are dynamically stabilized by the flow itself, and cannot be removed from the test pipe for closer examination. The present invention makes it possible to determine both the velocity profile and the spatial distribution of the phases, giving valuable information about how the rheology has changed as a result of the mixing.

Such capabilities also make a great difference to field applications such as multiphase flow meters, using thick pressure resistant piping. This embodiment is shown in figure 5, where the equipment is contained in a

pressurized compartment with an inert gas. The compartment is separated from the process flow by a radiation

transparent wall, with a pressure difference over this wall less than 10 bar. The detectors are all solid state and can easily be made to withstand the pressure. The most fragile parts, the X-ray sources are contained in low

, pressure compartments . Only a narrow slit in these f

I compartments facing the flow, need to be both pressure

J??-^'

½ resistant and transparent to radiation. In the rest of the compartment steel can be used without any consideration of radiation transparency.

In the preferred embodiment of the present invention:

- The number of projection angles the system incorporates is three. More projections improve image quality, but add cost and complexity.

- Two sets of positions for the detectors and X-ray sources are used.

■* - The X-ray sources have a broad spectrum.

- Detector including sensor pixel geometry and size of crystal is based on US7170049 or US20080099689. A CdTe direct conversion sensor crystal is used.

- Three discriminating energy bins are used.

^'Other parameters include

- Number of bits to use for trim-DACs for setting of energy level and counters .

- Optimal distance between the two positions with the parallel displaced linear fine-pixel (or so-called strip) sensors, which will give an additional image plane and hence obtain measurement of phase velocity.

- Noise, gain, linearity, stability and dynamic range.

- Imaging frame-rates, data transfer speed and protocol

- Interface for communicating with detector unit .

The system uses high-density readout X-ray single photon counting mode integrated circuit with spectroscopic

capabilities such as described in US7170049.

In another embodiment the number of discriminating energy bins are higher than three, please see Fig. 4b, which gives significant improvement in image quality and mass flow detection through exploiting the inherent spectroscopic or 'colour' information. The detector is built using an ASIC including:

- low-noise preamplifier,

- filtering shaper,

- energy discriminators,

- adjustment DACs for calibration of each and every channel, and

- single-event counters.

The design balances these features :

- low noise and high gain to maximize SNR,

- fast shaping to avoid pile-up,

- low power consumption,

- high dynamic range and linearity,

- precise adjustment DACs for accurate calibration, low drift and high stability.

The present invention requires frequent readout of data. The high frame-rate requires as short as possible dead-time by using high-speed transfer protocols, which ensure sufficient photon statistics in each frame. Many of the mentioned design goals can not be maximized simultaneously and making the right compromises is essential.

In the choice of CdTe direct conversion sensor crystal for preferred embodiment the following properties are important

- pixel size,

- crystal thickness,

- biasing conditions,

- size of crystal, - type of anode contacts and passivation.

These sensor parameters have great impact for detection quantum efficiency (DQE) , photon statistics, spatial resolution, speed of charge collection, and diffusion and spectroscopic

resolution.

Other sensor types than CdTe are well known and may also be used in other embodiments of the present invention, such as CZT, HgI2, GaAs or Si.

In the preferred embodiment the sources and detectors are housed in a pressure vessel around the pipe with a carbon fiber pipe separating the hydrocarbon flow from the neutral but pressurized atmosphere. The entire apparatus may be arranged subsea, even at deep water while maintaining high X-ray transparency for the pipe-section (2) , avoiding loss of low- energy X-ray photons.

Figure 5 shows X-ray tomography for a high pressure pipeline compiled from a series of line sensor captions as a function of time . The upper image is a top view and the lower image is side view. Black represents 100% oil. The sampling rate is 120 Hz. Fig. 5b presents reconstructed tomograms of a flow structure, here a slug, passing through the sensor. The tomograms of Fig. 5b may be calculated from the transverse line data from Fig. 5. The images are indexed by the times corresponding to Fig. 5. Fig. 5b, image (a) is interpreted as a gas pocket, (b) is interpreted as a cross section of the front of a slug, (c) , (d) and (e) inside the slug, and (f) at 0.96 s the tail of the slug.

Claims

claims

1. A method of conducting X-ray tomography measurements of petroleum-containing multiphase flows through a pipe,

- arranging an X-ray transparent pipe section (2) as an in-line 5 part of said pipe of generally the same diameter;

- emitting, at a first position along said pipe section (2) , wide-angle beams of X-ray photons with a broad energy spectrum from two or more first pairs of first X-ray generators (6a, 6b,..), transversely across said pipe section (2) towards oneo or more first X-ray sensor linear arrays (7a, 7b,..), and

- emitting, at a second position along said pipe section (2) , wide-angle beams of X-ray photons with a broad energy spectrum from two or more second pairs of second X-ray generators (6a, 6b, ..) transversely across said pipe section (2) towards ones or more second X-ray sensor linear arrays (7a, 7b, ...),

- said first and second positions having a separation (L)o sufficiently short to allow persistence of flow structures, characterised by

5 - using an integrated circuit connected to each said pixel

sensor (71) conducting single X-ray photon mode counts and sorting said counts into two or more energy levels of higher- energy photons and lower-energy photons

- sampling said sorted counts from said one or more linearo arrays (7) of pixel sensors (71) at said first and second

positions at a sampling rate generally higher than a flow time over said separation (L) ;

- calculating, based on one or more sorted single photon counts from each pixel sensors (71) from said linear arrays of X-ray5 sensors (7) , two-dimensional tomographic representations of X- ray attenuation representing one or more phase fraction distributions of oil, water or gas in sections of said flow at said first and second positions; and - using said phase fraction distributions in said sections at said first and second positions, calculating correlations of patterns representing persistent recognisable flow structures. 2. The method of claim 1, wherein lower-energy bin pixel measurements are used for calculating X-ray attenuation differences for distinguishing particularly between oil and water. 3. The method of claim 1, using said correlations of flow structures of at least one of said oil, gas or water phases and said separation (L) for calculating a flow velocity of said at least one phase represented by said correlated flow structure. 4. The method of claim 1, using a finely subdivided X-ray sensor array (7) , preferably subdivided into a number of 160 pixel sensors (71) of length 1 mm and width 1 mm across the pipe direction and 1 pixel width in the pipe direction. 5. The method of claim 1, keeping said X-ray sources (6) contained in low internal pressure chambers in radiation opaque steel compartments (61) with a high-pressure resistant

radiation transparent window (62) facing the X-ray transparent pipe section.

6. The method of claim 1, using said pipe section (2) with any desired inclination between and including horisontal and vertical . 7. The method of claim 1, setting boundaries between said energy levels of said X-ray sensors digitally on said

integrated circuits by means of an algorithm in a computer.

8. The method of claim 7, setting one or more of said energy levels of X-ray photons to below 70 keV so as for allowing better discrimination between oil and water attenuation.

9. The method of claim 8, controlling each said pixel sensor (71) through individually programmable threshold signals for setting discrimination between said at least two energy levels of higher-energy X-ray photons and lower-energy X-ray photons. 10. The method of claim 1, using fibre composite such as carbon fibre reinforced composite for the wall of said pipe section (2) .

11. The method of claim 1, using a relatively low pressure tolerant pipe section (2) , less than about 10 Bar, while said petroleum- containing multi-phase flow being under high

pressure, said pipe section (2) and said first and second X-ray generators (6a, 6b,..) and said first and second X-ray sensor linear arrays (7a, 7b,..) enveloped in one or more inert-gas filled high-pressure compartments (4) balancing said high pressure of said multi-phase hydrocarbon flow.

12. The method of claim 1, using pulsed emissions from said first and second X-ray generators (6) at points of time controlled from a computer.

13. An X-ray multiphase flow tomography apparatus for

petroleum-containing multiphase fluids in a pipe,

characterised by

- an X-ray transparent pipe section (2) for being connected as an in-line part of said pipe of generally the same diameter;

- two or more first pairs of first X-ray generators (6a, 6b, .. ) and first X-ray sensor linear arrays (7a, 7b,..) arranged at a first position along said pipe section (2) ,

- two or more second pairs of second X-ray generators (6a, 6b, ..) and second X-ray sensor linear arrays (7a, 7b, ...) arranged at a second position along said pipe section (2),

- said X-ray generators (6a, 6b, ..) directed at first and second different angles (al, a2, ..) about an axis through said pipe section (2) so as for allowing two-dimensional tomographic image sections of said flow at said first and second positions,

- said first and second positions having a separation (L) sufficiently short to allow persistence of flow structures to allow correlation of said flow structures of said first and second tomographic image sections;

- each said X-ray generator (6a, 6b, ...) arranged for emitting a wide-angle beam with a broad energy spectrum of X-rays, - each said X-ray sensor linear array (7) subdivided into pixel sensors (71) and arranged transversely with respect to said pipe section (2) ,

- each said pixel sensor (71) arranged for discriminating at least between two energy levels of higher-energy X-ray photons and lower-energy X-ray photons and connected to a high-density readout X-ray single photon counting mode integrated circuit separately counting said higher-energy X-ray photons and said lower-energy photons separately. 14. The apparatus of claim 13, comprising a computer arranged for calculating, based on one or more sorted single photon count measurement from each pixel sensors (71) from said linear arrays of X-ray sensors (7) , two-dimensional tomographic representations of X-ray attenuation representing one or more phase fraction distributions of oil, water or gas in sections of said flow at said first and second positions; and

- arranged for using said phase fraction distributions in said sections at said first and second positions, for calculating correlations of patterns representing persistent recognisable flow structures.

15. The apparatus of claim 13, each said linear X-ray sensor array (7) being finely subdivided, preferably in a number of 160, of pixel sensors (71) of length 1 mm and width 1 mm.

16. The apparatus of claim 15, each said sensor array (7) subdivided into groups of linear sensor arrays of preferably 32 pixel sensors (71) , each group of extending across a detector module card of 32 mm width, each detector module comprising an ASIC (application-specific IC) and a connector to a

motherboard, said motherboard arranged in an extruded aluminium rail with a longitudinal slit for exposing said linear X-ray sensor array (7) .

17. The apparatus of claim 13, each said X-ray generator (6) provided with a collimator (61) shaped for restricting said wide-angle beam of X-rays towards said corresponding

transversely arranged X-ray sensor array (7) .

18. The apparatus of claim 13, said diameter of said pipe and X-ray transparent section (2) being about 10 cm, and said separation (L) between said first and second position being 10 to 20 cm.

19. The apparatus of claim 13, said X-ray sources (6)

contained in low pressure in radiation opaque steel

compartments (61) with a high-pressure resistant radiation transparent window (62) facing the flow.

20. The apparatus of claim 13, said pipe section (2) arranged for being directed with any inclination between and including horisontal and vertical.

21. The apparatus of claim 13, wherein boundaries between said energy levels of said X-ray sensors are set digitally by means of an algorithm in a computer.

22. The apparatus of claim 13, one or more of said energy levels of X-ray photons defined below 70 keV so as for better discriminating between oil and water attenuation.

23. The apparatus of claim 13, each said pixel sensor (71) arranged for being controlled through individually programmabl< threshold signals for discriminating between said at least two energy levels of higher-energy X-ray photons and lower-energy X-ray photons.

24. The apparatus of claim 13, said pipe section (2) being straight .

25. The apparatus of claim 13, said pipe section (2) made in fibre composite such as carbon fibre reinforced composite.

26. The apparatus of claim 13, said pipe section (2) being relatively only low pressure tolerant, less than about 10 Bar, said petroleum-containing flow being under high pressure.

27. The apparatus of claim 13, said pipe section (2) enveloped in one or more inert-gas filled high-pressure compartments (4) balancing said high pressure of said multi -phase hydrocarbon flow (F) , and provided with said first and second X-ray generators (6a, 6b,..) and said first and second X-ray sensor linear arrays (7a, 7b, .. ) .

28. The apparatus of claim 27, said X-ray sources (6) contained in low pressure in radiation opaque steel

compartments (61) with a high-pressure resistant radiation transparent window (62) directed facing towards the pipe section (2) .

29. The apparatus of claim 13, said first and second X-ray generators (6) arranged for pulsed emission at points of time controlled from a computer.

30. The apparatus of claim 13, said pixel sensors (71) comprising CdTe direct conversion sensor crystals.

31. The apparatus of claim 13, said X-ray sources (6) controlled by a computer (8) using a first control algorithm arranged for triggering said X-ray sources, said computer (8) comprising a second algorithm for commanding receiving measurement data from said sensors (7) .