WO2015015215A1

WO2015015215A1 - Characterization of an oil and gas industry sample

Info

Publication number: WO2015015215A1
Application number: PCT/GB2014/052354
Authority: WO
Inventors: John Francis Gregg; Martin Roy Harrison; Peter WHERRIT
Original assignee: Salunda Limited
Priority date: 2013-07-31
Filing date: 2014-07-31
Publication date: 2015-02-05
Also published as: EP3028037A1

Abstract

A sample acquired during hydrocarbon exploration or production, and comprising at least one of water, hydrocarbons or solids, is sensed by performing continuous wave nuclear magnetic resonance measurements of the amount of a relevant type of nucleus from which a measure of the composition of the sample is derived. During a scan through resonance of the one or more types of nuclei, there is detected an absorption measurement of the absorbed energy of the electromagnetic field, that may be an integrated measure of instantaneous absorption. Hydrogen nuclei may be measured to derive a measure of the amount of water in the sample. Chlorine nuclei and/or sodium nuclei may be measured to derive a measure of the salinity of the sample.

Description

Characterization Of An Oil And Gas Industry Sample

The present invention relates to characterisation of samples acquired by the oil and gas industry, typically during exploration and production.

The oil and gas industry has a considerable need to analyse and monitor a wide range of samples acquired during hydrocarbon exploration and production.. Non-limiting examples include evaluation of formation fluids downhole and at the surface , inline monitoring of production fluids and analysis of drilling fluids, drilling cuttings and waste. Depending on the situation, the sample may include water, hydrocarbons and/or solids, or various mixtures thereof, and may also include gas.

Merely by way of example, salinity is an important property of fluids encountered in hydrocarbon exploration and production. For example, it is desirable during exploration and production of hydrocarbons to measure the salinity of water in an oil water mixture, for example during drilling. Oil and water flowing in a pipe can co-exist in oil continuous (i.e. water suspended in oil) or water continuous (i.e. oil suspended in water) phases, or switch between these two states in some random and unpredictable fashion. It is particularly difficult to measure the salinity of droplets of water suspended in a

hydrocarbon continuous phase, for example when the water liquid fraction (WLF) or water cut is under fifty per cent.

One situation where salinity needs to be sensed in enhanced oil recovery (EOR) typically involves the injection of water, or waterflood. During a sweep, water is injected to maintain reservoir pressure and to maximise recovery of produced hydrocarbons. In this fashion recovery can increased from 30% up to a typical recovery target of 65%. Some companies inject produced water that has been purified and cleaned. Others simply inject 'raw' produced water, or inject water from another well. Sea water is usually injected offshore, whereas river water or produced water is used onshore. The salinity of injected water may be adjusted to boost hydrocarbon recovery. In each case, it is important that the salinity of the water can be measured before injection.

By measuring the salinity of well fluids, or water produced from a well at the surface, it is possible to distinguish between injected water and formation water. Injected water, that may be re-injected produced water, can generally be distinguished from formation water by its salinity. Therefore, in order to determine the efficiency of a 'sweep' or waterflood, it is important to discriminate between fluids by measuring their salinity both in the borehole, and in the formation.

Salinity can vary greatly within the borehole due the presence of brine, completion water, condensation water and fresh water from an aquifer. Salinity also can vary seasonally. For example, in Alaska salinity can change during the summer due to the melting of glaciers. Finally, salinity can vary on shorter timescales; for example during a sweep, a significant change in the salinity of produced water could indicate that injected water has reached the producing well, which is an event known as 'breakthrough' . Water is a big problem in shut in gas wells since liquid loading, or breakthrough, can kill a well.

Samples are typically taken at the surface at the well head and then expedited for analysis, frequently resulting in delays before results are available. Delays are particularly critical during breakthrough and can potentially lead to saturation of a well and curtailment of production. Similarly, collecting samples can be relatively difficult where flows are co- mingled, for example offshore, along a pipeline network or where samples are currently collected onshore and many miles from producing wells. In these circumstances results can either be too late, or erroneous, and a local salinity sensor, at the well head or inline with a separator is advantageous.

Commercially available capacitance-based sensors are often not capable of measuring above approximately 40% water hold-up, and cannot measure salinity.

Traditional conductivity based sensors cannot measure salinity in the presence of hydrocarbons mixed with water. Therefore, it would be desirable to expand capability to a wider range of water liquid fractions, and a wider range of salinity from fresh water up to saturated brine.

It has been proposed to sense the composition of a sample acquired during hydrocarbon exploration or production using nuclear magnetic resonance ( MR). NMR measures energy absorption or release by atomic nuclei, which may be affected

intramolecular magnetic field around an atom in a molecule. NMR is used widely outside the oil and gas industry, generally in a pulsed manner in which a radio frequency (RF) pulse of electromagnetic (EM) field containing multiple frequencies is used to excite different types of nuclei in a sample and the energy released by the nuclei is detected a finite time after energisation. However, whilst NMR is a powerful technique for studying many different properties, conventional pulsed NMR equipment is complex, and relies on expensive, heavy and bulky magnet technology, making widespread application within the oil and gas industry during hydrocarbon production and exploration impractical.

According to a first aspect of the present invention, there is provided a method of sensing the composition of a sample acquired during hydrocarbon exploration or production and comprising at least one of water, hydrocarbons or solids, the method comprising:

performing on the sample continuous wave nuclear magnetic resonance

measurements of the amount of one or more types of nuclei; and

analysing the measurements to derive a measure of the composition of the sample. The present invention therefore makes use of continuous wave MR. This provides advantage when applied to a sample acquired during hydrocarbon exploration or production in that it allows performance of measurements of the amount of one or more relevant types of nuclei from which useful measures of the composition may be derived, but using a simple NMR apparatus and technique.

Particular advantage may be achieved by detecting an absorption measurement of the energy of the electromagnetic field absorbed by the sample during a scan through resonance of the one or more types of nuclei that is an integrated measure of the instantaneous absorption. In contrast to a spectroscopic technique that detects the instantaneous absorption at different frequencies, the scan may be performed at relatively high rates, providing for rapid acquisition of measurements.

Various types of nuclei may be detected, including hydrogen nuclei that may be used to derive a measure of the amount of water in the sample and including chlorine nuclei and/or sodium nuclei that may be used to derive a measure of the salinity of the sample.

Advantage is also achieved when the method is applied to plural types of nuclei. In this case as each type of nucleus is in the same sample, the measurements are effectively calibrated in scale with respect to each other, allowing the measurements to be accurately combined in deriving a measure of the properties of the composition of the sample. This improves accuracy generally and in the case of a flowing fluid sample mitigates the effects of varying flow rates to some extent.

According to a further aspect of the present invention, there is provided a sensor system arranged to perform a similar sensing method.

Embodiments of the present invention will now be described by way of non- limitative example with reference to the accompanying drawings, in which:

Fig. 1 is a diagram of a sensor system including an NMR apparatus and an analysis unit;

Fig. 2 is a diagram of the NMR circuit of the NMR apparatus;

Fig. 3 is a graph of the variation with time of the amplitude of the magnetic field produced by the Helmholtz coils of the sensor system;

Fig. 4 is a diagram of a sensor system that simultaneously performs NMR measurements of the amount of two types of nuclei;

Fig. 5 is a plot of the signal output by the detector circuit of the NMR circuit;

Fig. 6 is a plot of the signal of Fig. 5 after rectification;

Fig. 7 is a graph of the measurement of hydrogen nuclei against the sweep frequency of the magnetic field; and

Fig. 8 is a graph of the variation with time of the amplitude of the magnetic field produced by the Helmholtz coils of the sensor system for measuring hydrogen nuclei in water and hydrocarbons.

A method and sensor system for sensing a sample acquired during hydrocarbon exploration or production will be described.

The sensing may be applied to a wide range of samples, depending on the situation in which it is acquired and the composition that it is desired to sense. Generally the sample may comprise water, hydrocarbons or solids, or it may comprise a mixture thereof. The hydrocarbons may be any of the wide range of types of hydrocarbon that are present during hydrocarbon exploration or production.

In many situations, the sample may be a fluid sample, typically a liquid sample. Such a fluid sample may comprise a mixture of hydrocarbon and water, for example an oil-continuous phase (i.e. water suspended in oil) or a water-continuous phase (i.e. oil suspended in water). Mixtures of water and oil can take either of these two forms, depending on the water content, the water salinity, the type of hydrocarbon, the

temperature of the mixture and the history of the mixture, i.e. how the mixture was created. The transition between the states shows hysteresis, and occurs very rapidly.

The water-continuous phase may comprise isolated oil droplets and may further comprise gas bubbles and/or solid particles floating in a water matrix. If the water is saline, the mixture is electrically conducting.

The oil-continuous phase may consist of isolated water droplets floating in a oil matrix that may also contain gas bubbles and/or solid particles. The mixture is not electrically conducting.

In the case of fluid sample, the sample may be flowing as it is sensed. In this case the sensing has particular advantages as described below, but the sample could equally be a static sample.

A fluid sample may further comprise solids, for example solids suspended in the fluid.

A fluid sample may further comprise gas, for example bubbles of gas or dissolved gas.

An example of a fluid sample is a sample of a drilling fluid often referred to as a drilling mud. Drilling mud typically contains a mixture of water and oil, together with variable amounts of gas and solid material such as rock particles. The drilling fluid may be a water-based mud (which can be dispersed and non-dispersed), or a non-aqueous mud, sometimes called an oil-based mud.

Other examples of a fluid sample are a sample of a completion fluid, drilling waste, drilling cuttings, produced fluids, or crude oil.

In other situations, the sample may be a solid sample, for example rock. Such a solid sample may further comprise water and/or hydrocarbons, for example absorbed in the solid sample, or in pores.

A solid sample may further comprise gas.

Fig. 1 shows a sensor system 1 including an NMR apparatus 2 and an analysis unit

3. The NMR apparatus 2 is arranged as follows to perform continuous wave NMR measurements.

The NMR apparatus 2 comprises a sample holder 10 that forms part of a pipe 11 through which the sample flows. This arrangement is suitable for a flowing fluid sample. For a static sample, the sample holder 10 could be replaced simply by a container holding the sample.

The sample holder 10 is made from a non-conductive material so that it does not interfere with the magnetic field and RF EM field. The sample holder 10 may be made from an electrically insulating material such as glass. The sample holder 10 may have any suitable dimensions commensurate with the formation of magnetic field and RF EM field. In one example, the sample holder 10 has a diameter of 8mm.

The NMR apparatus 2 further comprises a permanent magnet 12 that generates a uniform, static magnetic field. The sample holder 10 is rigidly held within the poles of the permanent magnet 12 so that its magnetic field is applied to the sample holder 10 perpendicular to the pipe axis of the pipe 11. The permanent magnet 12 may have any suitable strength and dimensions. In one example, the permanent magnet 12 provides a magnetic field of approximately 0.5T over a circular region approximately 50mm in diameter, with poles separated by 20 mm.

The NMR apparatus 2 further comprises a pair of Helmholtz coils 13 that on driving by an alternating drive current generate a magnetic field aligned with, and in the same direction as, the magnetic field from the permanent magnet 12. Thus, an overall magnetic field that is the sum of that from the permanent magnet 12 and the Helmholtz coils 13 is applied to the sample holder 10. In use, the magnetic field from the Helmholtz coils 13 may be varied by control of the drive current to modulate the overall magnetic field.

The NMR apparatus 2 further comprises a RF coil 14 that is used to apply an RF EM field to the sample holder 10. The RF coil 14 is coiled around the sample holder 10 with its coil axis aligned with the perpendicular to the pipe axis of the pipe 11. The RF coil 14 may take any suitable form for example being made from enamelled copper wire. Thus, in use it generates an EM field having a magnetic field component perpendicular to the magnetic field created by the permanent magnet 12 and the Helmholtz coils 13.

The sample holder 10 and the RF coil 14 is surrounded by a screening box 15 that may be metallic. The screening box 17 allows the magnetic field to pass, but screens the RF coil 14 from electrical noise. The screening box 17 has apertures for the pipe 11 and the electrical connections to the RF coil 14.

The RF coil 14 is connected to an NMR circuit 20 in a screening enclosure 16 of the NMR apparatus 2, for example using two lengths of rigid copper conductor 17 that are supported by plastic spacers 18 inside a screening tube 19 that may be metallic. The screening tube 19 is connected to sample holder 10 and the RF coil 14 and to the screening enclosure 16 containing the NMR circuit 20. This arrangement provides low-resistance, low-capacitance connections between the RF coil 14 and the NMR circuit 20.

The NMR circuit 20 of the NMR apparatus 2 is shown in Fig. 2 and will now be described. The circuit 15 includes an oscillator 21. The oscillator 21 generates an oscillating signal in a tank circuit formed by the coil 15, together with a capacitor 22. The oscillating signal generates an RF EM field in the RF coil 14 that is therefore applied to the sample in the sample holder 10.

The oscillator 21 may be a marginal oscillator or a Robinson oscillator. In this case, the oscillating signal is dependent on the impedance of the RF coil 14 and hence on the absorption of the energy of the RF EM field in the sample. The amplitude of the oscillating signal represents the instantaneous absorption and a signal representing that may be output by a detection circuit 25 as a measure of the instantaneous absorption, as described below. In principle, the oscillator 21 could be of any other type that can be used to detect the absorption of the energy of the RF EM field in the sample.

In this example, therefore, the RF coil 14 is used both to apply the RF EM field and as a sensor element to detect absorption of the energy of that RF EM field by the sample. That is convenient but is not essential in that a separate sensor element, such as a further RF coil, may be used to detect absorption of the energy of that RF EM field.

The MR circuit 20 also includes a drive circuit 23 that drives the Helmholtz coils 13.

The NMR circuit 20 also includes a control circuit 24 that controls the oscillator 20 and the drive circuit 22. In particular, the control circuit 15 performs a control causing the NMR apparatus to perform continuous wave NMR measurements on the sample in the sample holder 10 as follows.

The control circuit 24 controls either the drive circuit 23 or the oscillator 21 to perform a scan of the sample through the resonance of a type of nucleus under test.

In the former case, the drive circuit 23 is controlled to vary the magnitude of the magnetic field applied to the sample. The magnitude of the magnetic field is scanned through the resonance. In this case, the drive circuit 23 may include a programmable waveform generator that enables the sweep frequency and the field slew rate to be independently controlled by the control circuit 24.

In the latter case, the oscillator 21 is controlled to vary the frequency of the RF EM field through the resonance. In this case, the oscillator 21 is designed to provide a controllable frequency of the oscillating signal, for example by the capacitor 22 being a tuning capacitor that is tuned to adjust the frequency of the oscillating signal.

In either case, the scan is performed so that the sample passes through the resonance of type of type of nucleus under test. In a magnetic field, the energy of nuclear spin states of a nucleus split into different levels with the splitting proportional to the strength of the magnetic field. In thermal equilibrium, there is a slight imbalance in the populations of the nuclear spin states, with slightly more nuclei occupying the state of lower energy. At resonance, the nuclei absorb photons having an energy equal to the energy difference between the nuclear spin states. Thus, resonance occurs when the frequency of the EM RF field is such that the energy of the photons of the EM RF field is equal to the energy difference between the nuclear spin states of the type of nucleus under test. As this energy difference is proportional to the applied magnetic field, the resonance can be achieved by scanning the magnetic field or the frequency of the EM RF field.

The type of nucleus under test may be selected by selection of the magnitude of the applied magnetic field and the frequency of the EM RF field in combination. Thus, the resonance frequency is determined by the type of nucleus (i.e. which isotope of the element under study) and by the strength of the magnetic field. By way of example, in a field of 0.5T (which is a typical value that can be easily generated using currently available permanent magnets), the hydrogen- 1 isotope ("proton") resonance occurs at 19.97MHz, the sodium-23 isotope resonance occurs at 5.28MHz, and the chlorine-35 isotope resonance occurs at 1.96MHz. If the strength of the magnetic field changes, the resonance frequencies change in direct proportion.

In general, the magnitude of the applied magnetic field may be freely chosen, provided that the frequency of the RF EM field is chosen accordingly. Higher magnetic fields are desirable, mainly because the splitting of the energy levels of the nuclear spin states is greater, which gives a greater imbalance in the populations of the two states. This allows more energy to be absorbed during the scan through the resonance condition, giving stronger MR signals. However, in practice the size of the sample holder 10 may restrict the size of the permanent magnet 12 and hence the magnetic field.

There is no fundamental lower limit on the value of the magnetic field that is used, but a practical one of difficulties created by a low frequency of the RF EM field and a weak output signal. NMR signals of more types of nuclei having a high gyromagnetic ratio, such as hydrogen, can be observed even in the earth's magnetic field of -0.0005T, although the resonance frequency is very low and the signals are very weak.

Fig. 3 shows a suitable variation of the magnitude of magnetic field B in the case that the drive circuit 23 is controlled. The magnetic field B is scanned alternately up and down in scan periods 30 during which resonance occurs at a time 31 within the scan period. During the scans in scan periods 30, the magnetic field B is changed at a controlled slew rate. Between scan periods 30, the magnetic field B is maintained constant. The scan periods 30 repeat at a field sweep frequency fs (being twice the frequency of the overall waveform of the magnetic field) providing an interval 32 between successive resonances. The field sweep frequency fs may be chosen to allow spin relaxation of the type of nucleus under test during the interval 32. The field sweep frequency fs varies but is typically in the range from 0.01Hz to 100Hz.

Any type of nuclei may be studied to provide different information on the composition of the sample. For example the type of nucleus may be hydrogen nuclei, that provides information on water and hydrocarbons in which it appears, or may be sodium nuclei or chlorine nuclei which provides information on salinity.

Advantage is also achieved when plural types of nuclei in the sample are studied. In this case as each type of nucleus is in the same sample, the measurements are effectively calibrated in scale with respect to each other, allowing the measurements to be accurately combined in deriving a measure of the properties of the composition of the sample. This improves accuracy generally and in the case of a flowing fluid sample mitigates the effects of varying flow rates to some extent.

The NMR apparatus 2 can perform continuous wave NMR in respect of a single type of nucleus at any given time. The NMR apparatus 2 may in principle be controlled to perform continuous wave NMR in respect of different types of nuclei at different times, by varying the magnetic field and/or frequency of the RF EM radiation. However, in practice, the components of the NMR apparatus 2 such as the Helmholtz coils 13, the RF coil 14 and the oscillator 21 may be optimised to work in a particular band that effectively limits the type of nuclei that can be studied.

Fig. 4 illustrates the NMR apparatus 2 modified to simultaneously perform NMR measurements of the amount of two types of nuclei, simply by replicating the RF coil 14 within the screening box 15 and the NMR circuit 20 within the screening enclosure 16. Thus, each of the coils 14 and the associated NMR circuit 20 is arranged to perform NMR measurements of one of the types of nuclei. Similarly further replicas could perform NMR measurements of other types of nuclei.

At resonance during the scan, energy of the electromagnetic field is absorbed by the sample. The NMR circuit 20 detects an absorption measurement of this absorbed energy as follows.

The NMR circuit 20 includes a detection circuit 25 that detects the absorption measurement from the oscillator 21. In the case that the oscillator 21 is a marginal oscillator or a Robinson oscillator, the detector circuit 25 detects and outputs a signal representing the amplitude of the RF EM field in the RF coil 14 as a measure that represents the instantaneous absorption.

The detector circuit 25 may include filters to remove RF signals and extract the desired signal component. The filters may also isolate the sharp NMR signals from liquids from the broad signals from solid matter, in the case that it is desired to detect the type of nucleus under test in a liquid phase only. The detector circuit 25 may also include a lock-in amplifier to improve the signal to noise ratio of the output signal.

Fig. 5 shows a typical output signal from the detector circuit 25 recorded, by way of example, from hydrogen- 1 as the magnetic field is scanned through the resonance. Signals from other types of nuclei are similar in form, but typically much weaker in intensity. As can be seen, besides the main peak which derives from the nuclei being excited from a spin state of low energy to a spin state of high energy, there is relaxation ringing on the trailing edge as the nuclei relax back into the spin state of low energy. The main peak is dependent on the amount of the type of nucleus under test. On the other hand, the relaxation ringing is less useful, and so is ignored by removing this part of the signal as follows.

The output signal passes through a rectifier 26 that extracts the positive part of the signal, thereby removing the relaxation ringing on the trailing edge. Fig. 6 shows a typical output signal from the rectifier 26. As an alternative to using a rectifier 26, the averaging circuit 27 described below can simply average the positive part of the signal only.

The rectified output signal from the rectifier 26 is supplied to an averaging circuit 27 that averages the output signal over plural scans, so that the final output signal of the NMR circuit 2 is an absorption measurement that is averaged over plural scans. This improves the accuracy of the absorption measurement.

The output signal from the averaging circuit 27 is supplied to an integrator 28 that integrates the output signal. Therefore, the output signal from the integrator 28 is an integral of the measure of instantaneous absorption represented by the output signal from the oscillator 21. This is therefore an absorption measurement of the type of nucleus under test.

The output signal from the integrator 28 is supplied to an analog-to-digital converter 29 to convert the absorption measurement into a digital signal that is output from the NMR circuit 20 to the analysis unit 3.

The analysis unit 3 analyses the absorption measurements from the NMR apparatus and derives a measure of the composition of the sample. The analysis unit 3 may be implemented by any suitable data processing system, for example a computer apparatus including processor executing an appropriate program or dedicated hardware.

Various measures of the composition of the sample may be derived from the adsorption measurements arising from various different types of nuclei.

In the case that the NMR apparatus 2 provides adsorption measurements of the amount of hydrogen nuclei, the analysis unit 3 may derive a measure of the amount of water in the sample from those measurements as follows.

Hydrogen nuclei are present in any water and in any hydrocarbons in the sample. In general a signal may be derived from either or both of these components. Hydrogen nuclei are also present in any gas in the sample, but in negligible amounts compared to the other components, so effectively no signal is derived from gas constituents. Hydrogen nuclei are also present in any solid material in the sample, but the signal from this component has a very broad bandwidth compared to the signal from water or hydrocarbons, and thus is not detected by the detection circuit 25, particularly if this is designed with appropriate filters, so effectively no signal is derived from solid constituents.

However, it remains necessary to separate the signals from water and

hydrocarbons. This is done on the basis of their different spin relaxation behaviour. At resonance, the hydrogen nuclei in both water and hydrocarbons absorb energy of the RF EM field. If the RF EM field persists, this may occur until the spin population in the two spin states becomes equal, at which point no further energy absorption occurs and the NMR signal disappears, but at a high rate of scan through resonance such an equilibrium is not reached. Irrespective of whether equilibrium is reached, as the scan continues and moves away from resonance, the hydrogen nuclei that have been excited return to thermal equilibrium at the spin relaxation rate. If thermal equilibrium is restored before returning to resonance, the NMR can be performed again and the NMR signal remains at a constant value for each field sweep. However, if a scan through resonance is repeated before hydrogen nuclei have relaxed, then those excited hydrogen nuclei do not contribute to the output signal. Spin relaxation of hydrogen nuclei in water occurs significantly slower than spin relaxation of hydrogen nuclei in hydrocarbons, and the NMR apparatus 2 uses this phenomena to select the absorption measurement accordingly.

Fig. 7 illustrates the effect of the spin relaxation behaviour of hydrogen in water and in hydrocarbon. In particular, Fig. 7 is a graph of the output signal against the field sweep frequency fs, which may be obtained simply by plotting results observed whilst scanning the field sweep frequency fs in the NMR apparatus 1. Fig. 7 shows separation of the signals as a result of spin relaxation of hydrogen nuclei in water occurring relatively slowly compared to spin relaxation of hydrogen nuclei in hydrocarbons. Thus, at a low field sweep frequency fs, corresponding to a long interval 32 between successive resonances greater than the period needed for spin relaxation of hydrogen nuclei in water (and therefore also hydrocarbons), the output signal V_T is derived from the hydrogen nuclei in both water and hydrocarbons. Conversely, at a high field sweep frequency fs, corresponding to a short interval 32 between successive resonances less than the period needed for spin relaxation of hydrogen nuclei in water, albeit greater than the period needed for spin relaxation of hydrogen nuclei in hydrocarbons, the output signal V_H is derived from the hydrogen nuclei in hydrocarbons alone.

The low field sweep frequency fs may be around 0. lHz. The high field sweep frequency fs may depend on the types of hydrocarbons expected to be present in the sample, which may in turn depend on the type of sample. The high field sweep frequency fs may typically be 10Hz or 100Hz. These frequency values may be derived

experimentally for any given type of sample by taking measurements at different frequencies to observe how the response varies with frequency, for example by plotting results in a similar manner to Fig. 7.

A corresponding signal representing the amount of hydrogen nuclei in water may be derived from the difference (V_T-V_H) between the output signals at low and high sweep field frequencies.

The spin relaxation properties of hydrogen nuclei in water and hydrocarbon are of themselves known. For example, US-4,785,245 discloses a technique in which the periods required for spin relaxation properties of hydrogen nuclei in water and hydrocarbon are used to distinguish the signal from hydrogen nuclei in hydrocarbon in a pulsed wave NMR technique. However, in the sensor system 1 , separation of the signal from hydrogen nuclei in water is performed in the NMR apparatus 1 using continuous wave NMR by changing the field sweep frequency fs used when scanning through resonance of hydrogen nuclei.

In particular, the control circuit 24 controls the drive circuit 23 to perform scans with different field sweep frequencies fs.

First scans are performed through resonance of hydrogen nuclei at a field sweep frequency which provides an interval between scans that is sufficiently long to allow spin relaxation of hydrogen nuclei in water (and hence also in hydrocarbons). Thus, each of the first scans is performed after such an interval since the previous scan.

Second scans are performed through resonance of hydrogen nuclei at a field sweep frequency which provides an interval between scans that is sufficiently short to prevent spin relaxation of hydrogen nuclei in water but sufficiently long to allow spin relaxation of hydrogen nuclei in hydrocarbons. Thus, each of the second scans is performed after such an interval since the previous scan.

A first absorption measurement V_T is detected by the NMR apparatus during the first scans and a second absorption measurement V_H is detected during the second scans.

In one possible type of control, the first scans are all performed together using the waveform of the magnetic field B shown in Fig. 3 with a constant interval 32 between scans and at a different time the first scans are all performed together using the waveform of the magnetic field B shown in Fig. 3 with a different, constant interval 32 between scans. In this case, during each of the set of first scans and the set of second scans, the NMR apparatus 2 operates as described above and the averaging circuit 29 outputs an averaged absorption measurement at the end of each set of scans.

In another possible type of control, the first and second scans are performed alternately. This may be performed using the waveform of the magnetic field B shown in Fig. 8. The first scans 33 are performed when the magnetic field B rises and the second scans 34 are performed when the magnetic field B falls, so that the first scans 33 and second scans 34 alternate. Each first scan 33 is performed after an interval 35 since the previous second scan 34 that is sufficiently long to allow spin relaxation of hydrogen nuclei in water. Each second scan 34 is performed after an interval 36 since the previous first scan 33 that is sufficiently long to allow spin relaxation of hydrogen nuclei in water. . In this case, the NMR apparatus 2 operates generally as described above, except that the averaging circuit 29 is controlled to separately average the absorption measurements from the first scans 33 and from the second scans 34. This type of control slightly increases the complexity of the averaging circuit 29 since it must average two alternate signals that are serially received. However, this type of control has the advantage that the first and second absorption measurements are derived under identical conditions, which reduces error.

The analysis unit 3 uses the first and second absorption measurements Vx and V_H to derive a measure Vw of the amount of water in the sample therefrom, i.e. as the difference The (i) first absorption measurement V_T, (ii) second absorption measurement V_H, and (iii) measure Vw are proportional to the amount of hydrogen nuclei in (i) water and hydrocarbons combined, (ii) hydrocarbons and (iii) water, respectively, with the same proportionality constant that may be determined for the NMR apparatus 1 at a given applied magnetic field by calibration with known samples.

The measure Vw of the amount of water in the sample may be expressed as an absolute amount or as the water cut, i.e. the proportion of water in the sample, sometimes referred to as the On-Line Water Determination.

In the case that the NMR apparatus 2 provides adsorption measurements of the amount of sodium nuclei or chlorine nuclei, the analysis unit 3 may derive a measure of the salinity of the sample from those measurements as follows.

The presence of salt ions (which may include sodium, potassium, magnesium, and calcium ions together with chloride ions) in water in the sample does not significantly modify the spin relaxation rate of the hydrogen nuclei in the water. This means that it is difficult to derive a measurement of the water salinity by studying the hydrogen nuclei in isolation.

Therefore a measure of salinity is derived using adsorption measurements of the amount of sodium nuclei or chlorine nuclei.

The NMR measurements from sodium nuclei or chlorine nuclei originate from saline water in the sample, irrespective of whether the sample is in the oil-continuous or water-continuous state. There are negligible quantities of sodium or chloride ions in oil or gas, and the signals from any solid material will be too broad to detect as discussed above.

The output signals from sodium nuclei and chlorine nuclear are similar in shape to that from hydrogen nuclei, but much weaker in intensity so the averaging performed in the averaging circuit 28 is more important. Since the spin relaxation rate for the spin 3/2 sodium and chlorine nuclei is very fast, a fast field sweep frequency fs can be used without saturating the nuclei (i.e. the spins states always have time to return to thermal equilibrium before the next sweep through the resonance). This means that many sweeps can be generated and combined to improve the signal-to-noise ratio in a reasonable period of time.

Chlorine nuclei provide a better measure of salinity than the sodium nuclei, since other salts such as potassium, magnesium and calcium chloride may be present in the water. The current definition of salinity is based on the electrical conductivity of the sample. Since potassium, magnesium and calcium ions also contribute to the electrical conductivity, a salinity value derived from the concentration of sodium ions will be lower that the correct salinity value.

Against that sodium nuclei provide a stronger output signal than chlorine nuclei, so sodium nuclei may alternatively be used and still provide a useful measure of salinity, subject to the issues mentioned in the previous paragraph. Thus, although the following discussion refers to chlorine nuclei, sodium nuclei could be substituted.

The adsorption measurement of the chlorine nuclei is a measure of the number of chloride ions in the water within the sample. By combining this information with the water cut, i.e. the proportion of water in the sample, which might be derived by the sensor system 1 from the hydrogen nuclei as described above or determined by a separate sensor, the salinity of the water in the sample can be determined.

The analysis unit 3 may derive a measure of the salinity of the sample from the adsorption measurement and from the measure Vw of the amount of water in the sample derived as above, as follows.

The adsorption measurement from the chlorine nuclei is proportional to the water cut, i.e. the proportion of water in the sample, and to the salinity S of the water, because it is proportional to the total number of chlorine nuclei inside the sensing region of the RF coil 14. Thus, analysis unit 3 derives the salinity S from the adsorption measurement Vci and the measure Vw of the amount of water in the sample in accordance with the equation:

Vci = k.V_w.S

In this equation, k is a proportionality constant that may be determined for the MR apparatus 1 at a given applied magnetic field by calibration with known samples. As the adsorption measurement Vci and the measure Vw of the amount of water in the sample are both determined using the same NMR apparatus 2 with the same sample, in the case of a flowing fluid sample, that is with the same instantaneous flow rate, they scale with respect to each other with any experimental perturbations and at different flow rates without affecting the determined salinity S.

Thus, continuous wave NMR performed by the NMR apparatus 2 enables the concentration of the types of nuclei, for example hydrogen nuclei, sodium nuclei and chlorine nuclei in the water component of the sample, for example a mud, to be

determined, regardless of whether the sample is in an oil-continuous or a water-continuous phase.

Alternative solutions might rely on deploying multiple sensors to determine other properties of the flow that vary with water cut and water salinity.

A variety of alternative techniques can be used, but each technique has limitations.

Electrical permittivity measurements work well in the oil-continuous phase, but become insensitive to changes in the fluid composition in the water-continuous phase. They are also affected by the presence of gas and solids in the mixture.

Electrical conductivity measurements work well in the water-continuous state, but become insensitive to changes in the fluid composition in the oil-continuous state. They are much less affected by the presence of gas and solids in the mixture.

Density measurements work in both fluid states, but are greatly affected by the presence of oil and solids in the mixture.

By combining the information from all of these techniques, and making certain assumptions about the properties of the oil and amount of solid material, it is possible to calculate the properties of the water in the sample. But the results are not very accurate, particularly in the oil-continuous state, and the presence of gas and rock in the flow would affect the density measurement and skew the water cut and salinity calculation.

The sensor system 1 can be applied in a range of applications in hydrocarbon exploration or production to study a range of samples. Some non-limitative examples are as follows.

During drilling, fluids are routinely analysed by means of a 'mud check' to improve drilling efficiency. However today this process is largely manual and based on a slow measurement using a retort, and a time-consuming chemical titration. Drilling fluid characteristics that may be automatically measured (either offline using a stand-alone unit or by an online sensor) by the invention (alone or in combination with other sensors) include the oil/water ratio, solid content, density, viscosity, chloride content, sodium content, bromide content and overall salinity. Likewise drilling waste is analysed prior to disposal, transportation or storage. Characteristics that may be measured by the invention include oil, solid and chloride content. Expensive, specialist fluids used during drilling and well-completion such as completion water and brine may be analysed for Chloride, Sodium, Bromide or Caesium content using an NMR apparatus based on the invention. Other species that may be detected utilising an apparatus based on the invention and that are found in drilling fluids, waste and cuttings as well as in 'flowback' water (as used during hydraulic fracturing of rock such as shale to extract so-called 'tight' oil,

'unconventional' oil and gas reserves) could include normally occurring radioactive material (NORM) such Uranium, Thorium, Potassium, Radium and Radon. Finally, other species of interest present in formation water, connate water, injection water or production water during increased oil production (IOR) that may be detectable by a NMR apparatus based the invention include Sodium, Potassium, Magnesium, Calcium, Barium, Strontium, Iron (dissolved or total iron content), Sulphate, Chloride, Phosphorous, Silicon, Boron, Lithium, Aluminium, Bicarbonate and Carbonate.

There are several applications in IOR. For IOR, the sample may be a sample of injected water or waterflood that is injected. In this case, the salinity of the water is measured before injection. Also for IOR, the sample may be a sample from the borehole or from the formation. In this case, the salinity of the sample is measured in order to distinguish between injected water and formation water, in order to determine the efficiency of a 'sweep' or waterflood.

The sensor system 1 therefore offers the ability to optimize oil recovery from oilfields employing water injection (i.e. 'waterflood'), and the injection of desalinated water in particular. In waterflood, unless water 'breakthrough' is quickly brought under control there can be a significant reduction in hydrocarbon production, and the lifetime of the well can be severely curtailed. Therefore obtaining an instant measurement is of critical value. Obtaining an instant salinity measurement will be key to optimizing the process of injecting water to boost oil extraction and it is thought that as much as 10% to 30% more oil may be recoverable, and rate of extraction increased accordingly. The sample may be a sample from a borehole. In this case, the salinity of the sample is measured to monitor the performance and operation of the well. The sample may be taken at the surface at the well head and measured by the sensor system 1. This may occur on site, thereby reducing delays compared to remote analysis.

The sensor system 1 may be employed at a well head or downhole.

In a well, multiple layers of water can be present and breakthrough can take place at more than one of these layers. In these circumstances, downhole application of the sensor system is useful to identify where breakthrough has occurred, by directly logging salinity along the well. This provides advantage over retrieval of samples to the surface using special vessels. The capability to directly measure salinity downhole helps operators understand how water moves, in time and by distance. Likewise, if a significant change in salinity is detected in certain a region, this may indicate where water has broken through. Operators can then use this information to actuate valves, or by some other means isolate a water producing region, thereby optimising hydrocarbon recovery under flowing conditions.

The sensor system 1 may be operated in flowing conditions from 100% oil or gas to 100%) water. Under flowing conditions (e.g. at the well head), sensor system 1 may be coupled in-line with the Christmas tree, or a separator, and may operate to identify water in some kind of emulsion. This emulsion would be typically in the ratio of approximately 20%) oil and 80%> water. Temperature capability includes operation at up to 150 C for up to 20 hours.

For a downhole salinity sensor the RF coil 14, permanent magnet 12 and Helmholtz coils 13 are configured as a probe mounted either as a centralised, tool or as part of an array tool. Typically, outside diameters for logging tools are between 1 11/16" and 2 1/8". Narrow tools are preferred since they can run through constrictions, casing and tubing. Similarly, short tools are more flexible and since they are less like to snag during well logging. Casing internal diameters can be between 41/5" and 95/8".

In the case of an array tool, more probes are preferred since they can provide data when fluids are present in multiple layers, for example in high angle or horizontal wells. Array tools provide valuable data when fluids are present in stratified layers in horizontal well, whereas a centralised tool is sufficient in vertical wells. Array data is particularly useful since fluids move at different velocities. For example in vertical wells water tends to flow in an annulus between oil and the case. In horizontal wells lighter fluids are present in upper layers, with denser fluids such as completion water, formation water present in lower layers. The array tools may be implemented with six, eight or even twelve probes support on individual arms, or on bowsprings. Clearly miniaturised sensor probes are preferable since they may be deployed in larger numbers. The sensor system 1 may be implemented as a small probe with a miniature driver board that could be packaged within the arms of an array tool.

Claims

1. A method of sensing the composition of a sample acquired during hydrocarbon exploration or production and comprising at least one of water, hydrocarbons or solids, the method comprising:

performing continuous wave nuclear magnetic resonance on the sample to obtain measurements of the amount of one or more types of nuclei; and

analysing the measurements to derive a measure of the composition of the sample.

2. A method according to claim 1, wherein the step of performing continuous wave nuclear magnetic resonance measurements comprises:

applying a magnetic field and radio frequency electromagnetic field to the sample, performing at least one scan of one of the magnitude of the magnetic field or the frequency of the electromagnetic field through resonance of the one or more types of nuclei; and

detecting an absorption measurement of the energy of the electromagnetic field absorbed by the sample during said at least one scan through resonance of the or each type of nucleus.

3. A method according to claim 2, wherein the step of detecting an absorption measurement comprises detecting a measure of instantaneous absorption during said at least one scan, and integrating said measure of instantaneous absorption to derive said absorption measurement.

4. A method according to claim 2 or 3, wherein said step of performing at least one scan comprises performing plural scans, and said step of detecting an absorption measurement comprises detecting an absorption measurement averaged over the plural scans.

5. A method according to any one of the preceding claims, wherein the one or more types of nuclei include hydrogen nuclei.

6. A method according to claim 5, wherein the step of analysing the measurements derives a measure of the amount of water in the sample from said measurements of the amount of hydrogen nuclei.

7. A method according to any one of claims 2 to 4, wherein:

the one or more types of nuclei include hydrogen nuclei;

said step of performing at least one scan comprises:

performing at least one first scan through resonance of hydrogen nuclei after an interval since any previous scan through resonance of hydrogen nuclei a previous scan that is sufficiently long to allow spin relaxation of hydrogen nuclei in water, and

performing at least one second scan through resonance of hydrogen nuclei after an interval since any previous scan through resonance of hydrogen nuclei that is sufficiently short to prevent spin relaxation of hydrogen nuclei in water but sufficiently long to allow spin relaxation of hydrogen nuclei in hydrocarbons; and

said step of detecting an absorption measurement comprises:

detecting a first absorption measurement of the energy of the electromagnetic field absorbed by the sample during said at least one first scan, and

detecting a second absorption measurement of the energy of the electromagnetic field absorbed by the sample during said at least one second scan; and

the step of analysing the measurements derives a measure of the amount of water in the sample from the first and second absorption measurements.

8. A method according to any one of the preceding claims, wherein the one or more types of nuclei include chlorine nuclei and/or sodium nuclei.

9. A method according to claim 8, wherein the step of analysing the measurements derives a measure of the salinity of the sample from said measurements of the amount of chlorine nuclei or sodium nuclei.

10. A method according to claim 9, wherein the step of analysing the measurements derives a measure of the salinity of the sample from said measurements of the amount of chlorine nuclei or sodium nuclei and from a measure of the amount of water in the sample.

11. A method according to any one of the preceding claims, wherein the one or more types of nuclei comprise plural types of nuclei.

12. A method according to any one of the preceding claims, wherein the sample is a fluid sample.

13. A method according to claim 12, wherein the sample comprises water and hydrocarbons.

14. A method according to claim 12 or 13, wherein the sample is a flowing fluid sample.

15. A method according to any one of the preceding claims, wherein the sample further comprises gas.

16. A sensor system for sensing the composition of a sample acquired during hydrocarbon exploration or production and comprising at least one of water, hydrocarbons or solids, the sensor system comprising:

an NMR apparatus arranged to perform continuous wave nuclear magnetic resonance on the sample to obtain measurements of the amount of one or more types of nuclei; and

an analysis unit arranged to analyse the measurements to derive a measure of the composition of the sample.

17. A sensor system according to claim 16, wherein the NMR apparatus is arranged to perform continuous wave nuclear magnetic resonance on the sample by:

18. A sensor system according to claim 17, wherein the NMR apparatus is arranged to detect an absorption measurement by detecting a measure of instantaneous absorption during said at least one scan, and integrating said measure of instantaneous absorption to derive said absorption measurement.

19. A sensor system according to claim 17 or 18, wherein the NMR apparatus is arranged to perform plural scans, and to detect said absorption measurement by detecting an absorption measurement averaged over the plural scans.

20. A sensor system according to any one of claims 16 to 19, wherein the one or more types of nuclei include hydrogen nuclei.

21. A sensor system according to claim 20, wherein the analysis unit arranged to analyse the measurements by deriving a measure of the amount of water in the sample from said measurements of the amount of hydrogen nuclei.

22. A sensor system according to any one of claims 17 to 19, wherein:

the one or more types of nuclei include hydrogen nuclei;

the NMR apparatus is arranged to perform at least one scan by:

performing at least one second scan through resonance of hydrogen nuclei after an interval since any previous scan through resonance of hydrogen nuclei that is sufficiently short to prevent spin relaxation of hydrogen nuclei in water but sufficiently long to allow spin relaxation of hydrogen nuclei in hydrocarbons;

the NMR apparatus is arranged to detect an absorption measurement by:

detecting a first absorption measurement of the energy of the electromagnetic field absorbed by the sample during said at least one first scan, and detecting a second absorption measurement of the energy of the electromagnetic field absorbed by the sample during said at least one second scan; and

the analysis unit is arranged to analyse the measurements by deriving a measure of the amount of water in the sample from the first and second absorption measurements.

23. A sensor system according to any one of claims 16 to 22, wherein the one or more types of nuclei include chlorine nuclei and/or sodium nuclei.

24. A sensor system according to claim 23, wherein the analysis unit is arranged to analyse the measurements by deriving a measure of the salinity of the sample from said measurements of the amount of chlorine nuclei or sodium nuclei.

25. A sensor system according to claim 24, wherein the analysis unit is arranged to analyse the measurements by deriving a measure of the salinity of the sample from said measurements of the amount of chlorine nuclei or sodium nuclei and from a measure of the amount of water in the sample.

26. A sensor system according to any one of claims 16 to 25, wherein the one or more types of nuclei comprise plural types of nuclei.

27. A sensor system according to any one of claims 16 to 26, wherein the NMR apparatus arranged to perform continuous wave nuclear magnetic resonance on a sample that is a fluid sample.

28. A sensor system according to claim 27, wherein the NMR apparatus arranged to perform continuous wave nuclear magnetic resonance on a sample that comprises water and hydrocarbons.

29. A sensor system according to claim 27 or 28, wherein the NMR apparatus arranged to perform continuous wave nuclear magnetic resonance on a sample that is a flowing fluid sample.

30. A sensor system according to any one of claims 16 to 29, wherein the NMR apparatus arranged to perform continuous wave nuclear magnetic resonance on a sample that further comprises gas.