Method and Apparatus for Nuclear Magnetic Resonant Measurement of Flow Velocity

This apparatus and method are directed to the problem of measuring flow velocity. In measuring velocity of a material having an element of interest, magnetic resonance techniques are utilized to obtain signals resulting from polarization of nuclei or electrons in an element of interest. For example, the element of interest may be hydrogen. Carbon compounds may be especially rich in free electrons subject to EMR measurement. It is possible to align the nuclei of hydrogen in a material. Other elements can be selected. The material flows through a pipeline at a velocity to be determined. The velocity of the flow is to be measured without intrusion into the pipeline. To this end, the pipeline is constructed with at least a section of pipe made of non-ferromagnetic material to enable the flowing materials (and hence the element of interest) to be exposed to a transverse magnetic field of specified intensity. In conjuction with this, first and second coils forming RF fields at right angles to the magnetic field are concentric around the pipe. In one embodiment of this apparatus, a first coil is used to achieve magnetic resonance between the imposed magnetic field intensity, the frequency of RF energy transmitted from the coil into the flowing material, and the particular element. The element is involved by either aligning the nuclei in NMR or unpaired electrons of the element in EMR. The velocity of this element can then be determined with the method and apparatus which are set forth herein.

It is only required that the element be available in sufficient quantity to provide a suitable amplitude output signal. That is, increases in the element population as might occur with increases in density or changes in the fill factor do not alter the data which is obtained through the use of the present method and apparatus. Rather, this method and apparatus works in conjunction with a relatively uniform magnetic field H.sub.o between opposing poles of a magnet of this field strength. The magnetic field first polarizes the element of interest to set the stage for obtaining an NMR or EMR response from the element. The magnet has a relatively long length. The field has a length which is sufficiently long to accommodate two spaced concentric coils. This magnetic field cooperates with a first coil and second coil. It is desirable that the polarized nuclei or electrons remain in the magnetic field for a period of time more or less comparable to T.sub.1 to obtain suitable polarization which occurs at a rate determined by T.sub.1. If this initial polarization is accomplished for the maximum velocity, then there is assurance that lesser flow velocities will be accompanied by even greater preliminary polarization. As will be understood, a modest degree of polarization (alignment of a small portion of the element of interest) achieves a sufficient output that data can be observed. Greater polarization of the element of interest, even to the extent of 100% thereof, does not change the velocity measurement so long as the output signal is sufficient to exceed noise levels and observable. In most petrochemical fluids and also including coal pipelines, the element of interest is hydrogen.

It has been discovered in this invention that the use of two coils, spaced from one another, is advantageous. The first coil is labelled a tag coil. This is a coil that first imposes an RF signal on the flowing element of interest. The frequency of this signal has a relation to the magnetic field intensity. This signal is periodically switched off and on. The RF signal causes the magnetically polarized nuclei to be rotated away from alignment with the magnetic field, ending polarization. Since the RF is switched off and on, this creates polarized and depolarized cylindrical slugs of material flowing through the pipe. As the slugs move through the pipe, the tagged (depolarized) hydrogen nuclei arrive at the detection coil downstream. The detected NMR signal at the second coil decreases abruptly on arrival of a tagged slug. The NMR signal changes again at the next interface between adjacent polarized and depolarized slugs of flowing material. By measuring the time between the application of the tag signal and the arrival of the tagged nuclei (depolarized) at the detection coil, the flow velocity can be measured. It will be understood that travel time is observed both when the tag is switched on and when the tag is switched off. In other words, each interface can be observed in transit.

An alternate embodiment of this apparatus utilizes a feedback loop to control the tag signal duty cycle. The tag modulation frequency then becomes proportional to flow velocity.

To achieve better results with this arrangement, the optimum is obtained by making the coils relatively short, by imposing a maximum separation distance between the tag coil and the detection coil. Preferably, the magnetic field is uniform across a span determined by a coil dimensions and coil spacing and includes an upstream portion having a length related to the maximum permissible flow velocity. On the other hand, as the flow velocity decreases, there is a problem of diffusion of the flowing material, and also the requirement that the transit time for the lowest speed of interest must be less than T.sub.1 to prevent loss of the tagging effect. In other words, the tagged (depolarized) element will be repolarized if the flow between coils is sufficiently slow that the transit to the second coil is great compared to T.sub.1. For relatively high resolution, the NMR pulse repetition rate must be relatively high compared to the flight time. This however degrades the signal to noise ratio. Accordingly, in a given spacing between the coils, the high end of the velocity measurement range is limited by the repetition rate which assures acceptable resolution, and the low end of velocity is limited by obtaining transit time less than T.sub.1. Variations in the coil spacing are difficult; it is preferable to vary frequencies as opposed to moving coils.

Coil field dimensions limit slug definition and hence resolution. If the coil field is narrow, the slug is sharply defined and able to be more easily observed. The axial length of the flowing material polarized by the field of the coil can be effectively reduced by using a magnetic field gradient. In this manner, the effective length of the slug of flowing material is shorter than the physical length of the coil.

With the foregoing in view, the apparatus of this disclosure is defined as a magnetic resonance system for resonating either the nuclei or electrons of selected elements. NMR responses can be obtained from hydrogen. EMR responses can be obtained alternately from free electrons such as those associated with carbon in coal. In either case, the apparatus utilizes a relatively wide magnet which forms a fairly uniform magnetic field across a non-ferromagnetic pipe. This field is transverse to the flow of material of interest. First and second coils are located in this magnetic field. The first is a tag coil. It imposes an RF energy burst on the element of interest. The magnetic field polarizes or aligns the element of interest. The short RF burst from the tag coil depolarizes the element which then flows in a transit time to the second coil, a detection coil. Changes in NMR response at the second coil indicate variations in transit time and hence velocity.

So that the manner in which the above recited features, advantages and objects of the present invention are attained and can be understood in detail, more particular description of the invention, briefly summarized above, may be had by reference to the embodiments thereof which are illustrated in the appended drawings.

It is to be noted, however, that the appended drawings illustrate only typical embodiments of this invention and are therefore not to be considered limiting of its scope, for the invention may admit to other equally effective embodiments.

FIG. 1 is a schematic diagram of the preferred embodiment illustrating a magnet, tag coil and detection coil arranged relative to a pipeline for measuring velocity in the pipeline;

FIG. 2 is a timing chart showing the movement of tagged flowing material;

FIG. 3 is a schematic of a magnet and coils showing a gradient arrangement; and

FIG. 4 is a timing chart of an interlaced signal sequence.

Attention is first directed to FIG. 1 of the drawings. There, the flow velocity measuring system is identified by the numeral 10. A pipe 11 delivers a flowing fluid of interest. The fluid is a mix of compounds including an element of interest. The element of interest may be hydrogen. An alternate is free electrons attributed to carbon. Other elements can be selected. The elements are of the sort to respond to magnetic resonance, either by polarizing the nuclei or free electrons of the element. The numeral 12 identifies a first slug of the element in the pipe 11. This refers to a moving slug of material in the pipe. It is moving from left to right as viewed in FIG. 1. A similar slug 13 is is also identified. Another slug 14 is also included. The slugs 12, 13 and 14 will be described respectively hereinafter as the polarized slug, tagged slug and detected slug.

A magnet 15 is used to impose a relatively uniform field on the pipe 11. The width of the magnet 15 is broken into four distances marked on FIG. 1. The distance D.sub.1 should be sufficiently great that a substantial population of the element of interest is polarized. The distance D.sub.2 is relatively narrow, a narrow coil dimension being desirable. The distance D.sub.3 is preferably not so long that the transit time in that distance approaches T.sub.1. The distance D.sub.4 is sufficient in length that the detected slug 14 can be measured and a NMR response obtained.

A tag coil 16 fits within D.sub.2. There is also a detection coil 17 at D.sub.4. The tag coil 16 is connected with a tag transmitter 19. It is switched off and on by a tag control circuit 20. The coil 17 is connected with a transmit and receive circuit 22. This circuit forms a controlled sequence of operation for the coil 17 so that it is periodically used to transit and periodically to receive. The TR circuit 22 serves as a input circuit for a receiver 24. The receiver 24 is connected to a sample and hold amplifier 25. The amplifier 25 forms an output to an integrator 26. The integrator is connected to an indicator circuit 27 which is either a meter movement, other meter device, strip chart recorder or the like. FIG. 1 further discloses a pulsed transmitter. The preferred frequency is in the range of three to ten megahertz. It is pulsed off and on. The transmitter 28 is controlled by the sequencer circuit 30.

The sequencer circuit 30 controls the transmitter 28 and the receiver 24. It also controls operation of the sample and hold amplifier 25 and the indicator 27. Additionally, it controls the control circuit 20. That circuit in turn operates the tag transmitter 19 to form signals for the coil 16 in a controlled sequence. Sequencing is understood best by referring to FIG. 2 of the drawings. In FIG. 2, the wave form 40 identifies the control signal which is applied to the tag control circuit 20. The wave form 40 is typically in the range of 5 to 20 hertz. When tagging occurs, it is accomplished in bursts at 1600 hertz; during a 10 hertz "on" cycle, 80 RF bursts occur to tag the flowing material. While one RF burst might suffice, several RF bursts are used to enable the output NMR signal to be integrated and thereby improve the signal to noise ratio. The circuit 20 switches the tag transmitter 19 off and on for repeated bursts to tag the flowing material and to enable signal improvement by integration. Assume, that the element of interest is hydrogen, and further assume that a three megahertz signal matches the field strength formed by the magnet 15 to thereby initiate NMR response from the element of interest. This NMR relationship can be best be understood by comparing the three slugs of material in the pipe 11. First of all, the material at 12 is flowing in the magnetic field having an intensity of H.sub.o Which relates to the three megahertz signal. The material in the slug 12 becomes polarized. That is, it is polarized by the magnetic field. This achieves an initial alignment of the element nuclei (or electrons as the case may be) of interest, perhaps all of it, but at least a sufficient portion thereof to form an output signal. This is accomplished after the nuclei or unpaired electrons of the element of interest have been in the magnetic field H.sub.o for a duration approaching T.sub.1. It is not necessary that complete polarization be achieved so long as adequate polarization is obtained. Adequate polarization refers to polarization which is sufficient to provide adequate signal output for the operation of this device. It may be less than 100% polarization.

If the coil 16 is off, then the flowing material simply becomes even more completely polarized as it moves across the face of the magnet 15 and passes through the distances D.sub.1, D.sub.2 and D.sub.3 added together. This slug of material at 12 may become completely polarized, but again, total polarization is not essential. Subsequently, that slug is within the confines of detection coil 17 and is able to provide an NMR (or EMR as the case may be) response to RF pulses transmitted from the detection coil 17.

Assume for the moment that the polarized slug 12 eventually encounters the tag coil 16 at an instant when the tag coil 16 is provided with RF pulses and hence transmits a RF energy into the slug 12. When this occurs, the element of interest in the slug 12 is depolarized. The slug then flows down the pipe 11 to the position 13. At this point, the slug is described as tagged, referring to the fact it is now depolarized. The RF pulse transmitted from the transmitter 19 through the coil 16 accomplishes depolarization or tagging. The tagged material is now exposed to the magnetic field H.sub.o and repolarization is initiated. This requires an interval of time which is a function of T.sub.1. Assuming for the moment that repolarization is not accomplished, the slug travels on through the pipe and eventually encounters the detection coil 17. At the detection coil 17, the depolarized or tagged material changes the NMR response of the material to RF pulsed interrogation.

There are front and rear interfaces for the slug 14. The two interfaces are assumed to be sharply defined. In practice, they are not so sharply defined and some diffusion may occur due to turbulence in the flow. Even so, there is in fact an adequate change in response at the detection coil 17 observing the tagged and untagged flowing material in the pipe 11. Thus, when the tagged slug comes into the vicinity of the coil 17, it provides an NMR response which is different from that of untagged material. The NMR response is smaller in amplitude. With this in view, it will now be observed in FIG. 2 of the drawings that the positive going pulse 42 for the tag signal indicates the interval during which the tag transmitter 19 is operated to tag the flowing slug 13. It is tagged because the coil forms the RF signal which depolarizes the slug 13. Reference is made to the signal 44 from the detection coil. The transmitter 28 is operated for the interval 46. It is switched off for the interval 48. While it is on, it transmits a pulse to the detected slug 14. Assume that this slug is a tagged slug which was tagged earlier when it passed through the coil 16. As the slug 14 flows into the near vicinity of the detection coil 17, and is subjected to a transmitted RF pulse, a change will occur in the wave form 44 which is indicative of tagging. There is a change between the larger and smaller NMR signal outputs. When the flowing material is interrogated by transmitting a pulse, the NMR signal output from the detection coil 17 is maximum when the material then being interrogated has not been tagged. The opposite is also true, namely, the output signal is minimum because the material then being interrogated by the NMR transmitted pulses is tagged or inadequately polarized.

Assume as a matter of scale (depending on flow rates and the particular element of interest, physical spacing of the coils and other facts) that only 2% of the material is polarized after tagging while 63% of the material is polarized when it has not been tagged. In this instance, output amplitude will be approximately 30 times greater.

The test circuitry includes the following components, these components being described along with their sequence of operation. This circuitry is to be distinguished from the tag transmitter circuitry. Going now to FIG. 1, the control circuit 20 switches on the tag transmitter 19 for a specified instant in a timed sequence subject to the sequencer control circuit 30. Assume that it forms a pulse which is quite large, perhaps instantaneous power of 100,000 watts. As will be understood for purposes of exaggeration, the scale of the transmitted pulse to a received pulse is not particularly critical and is otherwise given merely as representative, not necessarily to constrain the practice of the invention. Thus, the tag transmitter circuitry in reality is relatively simple including the tag transmitter 19 and the transmitter coil 16. This transmits the pulse under control of the sequencer 30 described above.

First of all, when this large pulse is formed in the coil 16, there is always the possibly of directly coupled pulse formed in the coil 17. To this end, the TR network 22 is switched off at this instant. This avoids direct coupling from coil 16 to coil 17.

The interrogation or test circuit utilizes the pulse transmitter 28 to form a pulse for the coil 17. That transmitted pulse of perhaps 1,000 watts peak power is sent through the TR network 22 to the coil 17. When the TR network enables the transmitted pulse to pass, the receiver 24 is isolated or blanked off. The TR network then flips, switching off communication from the transmitter 28 to the coil 17; it simultaneously connects the coil 17 so that any signal in the coil 17 is supplied to the receiver 24. Assume that the receiver is switched for a few microseconds. During that interval, it will form an output signal for the sample and hold amplifier 25. This is perhaps more readily understood by viewing FIG. 4. The curve 70 shows the transmitter switched on; at this point, the receiver 24 is off. The receiver is turned on while the transmitter 28 is off, both happening under control of the TR network 22. Some peculiar wave-form such as the representative wave-form at 74 in FIG. 4 is received. A value is obtained therefrom by the sample and hold amplifier 25. Thus, the test circuit includes the pulse transmitter 28 which forms the signal protecting, the TR network 22 and the receiver. The receiver output of course is supplied to the sample and hold amplifier.

The NMR interrogate signal has a change at 50, and a later change at 52. The changes 50 and 52 describe transitions between the peaks 46 and 48. The two transitions are offset from the transitions of the pulse 42 in the tag signal 40. This time interval 54 is a function of flow velocity.

In FIG. 2, the numeral 56 identifies another wave form. The wave form 56 shows a transition at 58 corresponding to the transition 50. The time delay is greater; hence, the time interval 54 is much longer. This time interval 54 can be varied dependent on flow rate.

Consider as a possibility continuous scanning of the NMR output signal. That is, the output from the detection coil 17 is periodically tested for NMR amplitude. One technique of NMR output detection is to apply a three megahertz pulse of about three microsecond duration, best shown in FIG. 4 at 70. The pulse repetition rate is about 1600 pulses per second. This sampling rate tests the material 1600 times per second to provide sufficient data points to thereby fully describe the NMR response. Through the use of a receiver 24 designed for relative high sensitivity, low noise and rapid recovery and with a band width of about 300 kilohertz, the profile of the tagged and untagged material flowing through the detection coil 17 can be determined. The data points will have the shape of the wave forms 44 and 56, differing in time delay 54. This time delay 54 is indicative of the flow velocity. The several data points are output from the sample and hold amplifier 25 and are shown at 72 in FIG. 4. The data points are supplied to the integrator 26 which yields the approximate rectangular wave form 56 shown in FIG. 2. Compared with the sampling rate of 1600 hertz, a suitable tag rate is perhaps 5-20 hertz. The RF burst from the tag transmitter 16 thus shapes the wave form at 40.

There is the possibility of signal interference between the coils 16 and 17. They are physically aligned with one another and are relatively close. They have only a few turns. Long coils are generally undesirable. There is the risk of transmitting signals from the coil 16 to the coil 17. Longer coils increase the coil interference. An interlaced operation is preferable to avoid direct coupling between the two coils. It will be recalled that one suggested sampling rate at the detection coil 17 is 1600 samples per second, this being approximately three microsecond bursts occurring every 600 microseconds. Not only does this provide a light duty cycle, but this also enables the NMR receiver 24 connected to the detection coil 17 to be switched off and on in an interlaced fashion so that the receiver 24 is isolated from the transmitter. The receiver 24 is thus switched on to observe the expected NMR response 74 (FIG. 4) after each interrogate transmitter burst 70. With a cycle beginning with a three microsecond burst 70 from the transmitter 28, the receiver 24 is then operated for an interval such as perhaps 200 microseconds as shown at 76. Accordingly, the receiver 24 is switched on for 200 microseconds and is switched off for about 400 microseconds to comprise the 600 microsecond cycle shown at 76. During the 200 microsecond interval after a transmission when the receiver 24 is turned on, the tag transmitter 19 is switched off (see signal 78). During the remainder of the 400 microsecond interval, the tag transmitter 19 can be turned on (see signal 78). When the tag transmitter 19 is turned on for only a portion (not necessarily all) of the 400 microsecond interval, and this transmitter pulse is repeated for several pulses, the tag impact is integrated by the material. In other words, even though the tag transmitter 19 is switched on for only a part of the 600 microsecond period, the tagged slugs of material are integrated to become one slug having a length dictated by flow rate. The actual slug movement when the tag transmitter is off part of the 600 microsecond interval is insignificant, and is substantially zero. Thus the tag on signal 42 enables transmission of between 40 and 160 RF pulses for a tag signal envelope of between 5 and 20 hertz with RF bursts occurring at 1600 hertz. These rates are representative and may vary.

In summary, the receiver 24 is on when both transmitters are off. The transmitter 28 is on when the receiver 24 is off and also when the transmitter 19 is off. In fact, when either transmitter is on, the other transmitter is off.

There is a scale factor which comes into play depending on the velocity and diameter of the pipe. This also relates to the length of the coil. As will be understood, the portion of material tagged by the coil is not precisely defined within the confines of the coil geometry. There is some loss of resolution at the ends of the coil. To this end, maximum allowable spacing is set out by the T.sub.1 of the material. Attention is next directed to FIG. 3 of the drawings for a feature improving slug definition. In FIG. 3, a magnetic field with gradient is illustrated. The embodiment 100 shown in FIG. 3 incorporates a magnet with pole pieces 101 and 102 opposite a conduit 103. In the conduit 103, the element of interest is flowing. There exists for this particular element a relationship wherein the field strength H.sub.o is related to the frequency F.sub.o. A magnetic field different from H.sub.o does not enable resonance at the particular frequency F.sub.o to occur. Flowing material is thus exposed to a magnetic field H.sub.1 which different from the required field for resonance represented by the symbol H.sub.o. The magnet is constructed with tapered faces as a convenient means for obtaining a magnetic field gradient. The gradient passes through the value of H.sub.o. This value H.sub.o is achieved at the magnet face portion 104. This gradient exposes the flowing material to the requisite field intensity. The slug of material at 105 is exposed to the tag coil 106 for transmission of an RF tone burst having the requisite frequency F.sub.o to accomplish resonance with the species of interest coacting with the requisite field strength H.sub.o. The slug 105 is thus confined to a specific region. In light of this fact, the coil 106 can be longer then the slug 105; the slug 105 is narrow because the end portions of the coil 106 coact with flowing material which does not have the required initial polarization. Thus, the slug 105 is exposed to the magnetic field of proper intensity and RF tone bursts to accomplish the requisite polarization. The slug 105 is relatively narrow and is more sharply defined.

While there may be other intensities such as field strength H.sub.1 and H.sub.2, the interaction between the RF frequency and field intensity is not accomplished except at the slug 105 within the gradient magnetic field defining the slug 105.

The narrow width obtained by field gradient is true of the tag coil 106. This narrow width can also be obtained in the detector coil 107. While the coil 107 may be quite long, there is a region within the coil 107 where the magnetic field strength again is the requisite value H.sub.o. This defines a flowing slug 108 which is exposed to the interacting field strength and RF frequency. So to speak, the use of magnetic field gradients narrows or shortens the effective length of both coils. The incorporation of additional coil turns and the extension of the coils lengthwise along the pipe does not increase their size with the requisite decrease in accuracy or sensitivity.

Accordingly, FIG. 3 shows a system whereby magnetization across the pipe 103 is accomplished. In the region of the tag coil 106, there is a gradient in the magnetic field which provides the requisite field strength H.sub.o to interact with the coil frequency F.sub.o to shorten or narrow the flowing slug of material excited by the transmitted RF pulse. This shortened flowing slug of material in the field H.sub.o is the tagged material. Equally, the detector coil 107 can be quite long but its effective length is shortened by locating it in a magnetic field having a gradient passing through the requisite valve H.sub.o. A positive gradient is one wherein the flowing material is first exposed to a more intense field and flows through the pipe further to a less intense magnetic field. In like fashion, a negative gradient occurs when the flow is in the opposite direction. It will be observed that FIG. 3 includes similar positive gradient magnetic field areas. If desired, a negative gradient could also be used.

Through the use of gradient in the magnetic field at both the tag coil and the detector coil, system sensitivity can be sharpened, enabling more precisely defined measurements.

While the foregoing is directed to the preferred embodiment, the scope of the present disclosure is determined by the claims which follow.