Classifications

• • G—PHYSICS
  • G10—MUSICAL INSTRUMENTS; ACOUSTICS
  • G10K—SOUND-PRODUCING DEVICES; METHODS OR DEVICES FOR PROTECTING AGAINST, OR FOR DAMPING, NOISE OR OTHER ACOUSTIC WAVES IN GENERAL; ACOUSTICS NOT OTHERWISE PROVIDED FOR
  • G10K11/00—Methods or devices for transmitting, conducting or directing sound in general; Methods or devices for protecting against, or for damping, noise or other acoustic waves in general
  • G10K11/18—Methods or devices for transmitting, conducting or directing sound
  • G10K11/24—Methods or devices for transmitting, conducting or directing sound for conducting sound through solid bodies, e.g. wires
• • G—PHYSICS
  • G01—MEASURING; TESTING
  • G01F—MEASURING VOLUME, VOLUME FLOW, MASS FLOW OR LIQUID LEVEL; METERING BY VOLUME
  • G01F1/00—Measuring the volume flow or mass flow of fluid or fluent solid material wherein the fluid passes through a meter in a continuous flow
  • G01F1/66—Measuring the volume flow or mass flow of fluid or fluent solid material wherein the fluid passes through a meter in a continuous flow by measuring frequency, phase shift or propagation time of electromagnetic or other waves, e.g. using ultrasonic flowmeters
  • G01F1/662—Constructional details
• • G—PHYSICS
  • G01—MEASURING; TESTING
  • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
  • G01N2291/00—Indexing codes associated with group G01N29/00
  • G01N2291/02—Indexing codes associated with the analysed material
  • G01N2291/028—Material parameters
  • G01N2291/02836—Flow rate, liquid level

Definitions

• the present invention relates to ultrasonic flow measurement systems and particularly to wedges or buffer rods used to couple a transducer to a pipe or conduit wall for effecting a fluid measurement.
• the present invention provides a buffer waveguide for interposition between a source of shear waves and a fluid or the wall of a conduit containing the fluid.
• the buffer waveguide is solid but relatively thin, shaped in a side view somewhat like a hockey stick and characterized by an elongated lead-in or buffer portion, corresponding to the hockey stick handle, and a generally shorter seat/radiator portion corresponding to the blade of the stick.
• Energy traveling down the buffer is guided by or redirected off an internal surface and propagates without dispersion as shear waves in the blade to an exit face which seats against the pipe or fluid.
• the waveguide is characterized by a length X3, thickness XI and a core width X2 such that it has a core rectangular cross section of relatively high aspect ratio with preferably XI « X2 « X3.
• the thickness XI is comparable to the wavelength of a shear wave of the ultrasonic energy to be propagated therein, and is preferably large enough so that the exit face of the radiator portion, which is preferably both flat and rectangular, provides an orienting support surface allowing the blade to be clamped directly onto or welded to a pipe. Small topological features such as thermocouple mounting tabs may be added outside the core rectangle.
• the core width X2 is much greater than the shear wavelength therein, and preferably also greater than half the length of the shear wave packet.
• the buffer may include small flanges, weld beads, one or more reflective notches, or holes located outside the principal ray path along the buffer; it may also include a step in thickness to control depth of insertion of the seat portion though a pipe wall into which the device is to be welded.
• the principal bounding surfaces or major faces of the waveguide are planar and parallel, while narrow edge faces may be oriented obliquely and serve to reflect energy traveling down the handle to the exit face of the blade, or to deflect spurious path energy.
• several oblique faces in parallel planes reflect the incoming energy at a plurality of angles to result in an extended emission of energy along the blade portion.
• a continuously variable angle of incidence is provided along the exit face by an internal curved reflective surface.
• the exit face may further incorporate, or may rest in a slot containing, a low sound speed material;
• the active (piezoelectric or transducer) end of the buffer may include a tungsten-loaded plastic or silicone rubber attenuative material to allow high pulse repetition frequencies to be used.
• the waveguide is advantageously mounted in a spoolpiece either by welding into a slot in the spoolpiece, or by welding into a custom jig or onto a standard pipe riser clamp which is then clamped against the spoolpiece.
• Custom spoolpieces may be locally fabricated by the simple expedient of attaching appropriately oriented buffer/waveguide assemblies to a pipe segment.
• the buffer is formed of a sheet metal stamping folded about a central axis and having reflective faces formed by internal or recessed edges of the stamping.
• the buffer has a thickness of 6 millimeters with a width of 3-6 centimeters, a length of about 15 to 20 centimeters and a blade length of about 4 to 8 centimeters.
• the 6 millimeter wide exit face sits substantially flat against pipes having diameters in the range of 8-10 inches or more. Little preparation of the pipe surface is required to employ the buffer in a clamp-on mounting.
• a flat or slot is milled and the waveguide is welded to the pipe.
• the high-aspect-ratio rectangular cross section of the buffer results in a well propagated shear wave signal with little ringing or mode conversion.
• Energy leaving the exit face of the blade propagates to one or more well- defined receiving locations within the pipe interior along precise paths, thus enhancing measurement accuracy and allowing, for example, off-diameter paths to be interrogated for profile compensation or tomographic imaging.
• Slots or recesses may be milled and counter-milled to prepare sites in the conduit or spoolpiece wall for accurately seating the buffer and directing its output along a defined plane in the fluid.
• the buffer is a thin solid sheet which may be conceptualized as a body comprised of a triangle plus one or more polygonal extensions forming the standoff and the blade.
• the triangle is an integral wedge which receives shear energy from the buffer portion, and radiates the energy along its hypotenuse. Regions of the hockey stick blade outside the triangle may include holes positioned out of the signal path to accommodate clamps or hooks without impairing the wave function, and portions of the lead-in or buffer section outside the principal ray path may include isolated scattering features to provide reference echoes for calibration and dynamic temperature compensation.
• echoes from a first reference reflector that is normal to the path centerline and from a second reflector at the toe of the blade contain time and amplitude information from which the sound speed in the blade, as well as the efficacy of coupling to the pipe or fluid, are calculated.
• Fig. 1 shows an elongated nondispersive directive waveguide of the present invention for installation on a pipe, conduit or other solid body, to measure flow, liquid presence or level, or other fluid characteristic
• Fig. 1A shows another embodiment similar to that of Fig. 1, having a piezoelectric element of multi-element construction schematically divided into a combination of a waveguiding triangular wedge foot portion, and a polygon-bounded waveguiding extension to the piezomaterial;
• Fig. IB is a conceptual drawing showing a range of angles contemplated between the buffer portion relative to the wedge or foot portion;
• Fig. 1C shows a multi-angle embodiment in which three angles of incidence are provided in a single waveguide sensor
• Fig. ID shows another multi-angle construction in which a curved reflecting surface generates a continuum multiplicity of angles of shear-wave incidence
• Fig. IE shows a system wherein a plurality of identical waveguides are placed in a row to increase the system tolerance to a wide and unpredictable range in sound speeds in the fluid, and to increase the aperture;
• Fig. IF shows an embodiment where the buffer joins the blade to form nonplanar curved and sharply bent waveguides
• Fig. 1G shows a waveguide embodiment in which primary and secondary reflecting surfaces are protected against fouling to reduce exposure to debris;
• Fig. 1H shows a sheet metal stamping having first and second faces, in which half-cavities are formed on the first face in a pattern symmetrical about a fold centerline, such that after folding sharply about that centerline through an angle of 180°, the half- cavities become adjacent as the two halves of the first face touch one another, and thereby provide shielded reflecting surfaces functionally similar to those in Fig. 1G, invisible from the second face, which now bounds the folded configuration.
• Fig. II illustrates an embodiment with attenuative potting near a capped end to damp reverberations therein
• Fig. 1J shows system having a combination of waveguides of the invention at five locations on a pipe
• Fig. IK shows the end view of a pipe that contains a fluid with scatterers, illustrating a reflection measurement with two buffer waveguides of the invention
• Fig. IL shows a differential mass flowmeter for organic liquid that combines Rao's Rule for organic liquids and two flow velocity measurements, on hot and cold pipes with waveguides of the invention
• Fig. 1M shows an embodiment for performing downhole measurements
• Fig. 2 shows several waveguides of the invention welded to a pipe section, for permanent installation at accurately controlled depths and locations;
• FIG. 2 A, B show views of a spirometer system using waveguides of the present invention
• FIG. 2C illustrates processing for fast response with the structure shown in
• Fig. 2D shows a buried conduit measurement system
• Fig. 2E shows waveguides of the invention temporarily coupled to a pipe in planes parallel and perpendicular to the pipe axis by pipe riser clamps;
• Fig. 2F shows several hybrid constructions that provide (a) removability; (b) explosion proof design for low sound speed pipes, achieved by utilizing a composite waveguide; and (c) refraction angles in the contained fluid beyond limits imposed by the pipe material's sound speed relative to that in the fluid;
• Fig. 3 is a perspective view showing outside of the front, and inside of part of the back, of a Gauss-Chebyshev (GC) multipath flowcell in which off-diameter chordal measurements are made in well-defined symmetrically-located inboard and outboard planes using eight waveguides as shown in Fig. 1 welded in a minimally-intrusive manner into narrow oblong slots all on one (front) side of the spoolpiece;
• GC Gauss-Chebyshev
• Fig. 3A is a corner view of the spoolpiece of Fig. 3, more clearly showing four reflectors mounted in the wall opposite the transducers and in their planes;
• Fig. 3B shows a partly sectioned side view of a GC spoolpiece showing several clamped-on transducers at the top and bottom;
• Fig. 3C shows a system in which a sheet stamping is rolled or formed into a cylinder with features aligned to form a GC and diameter path spoolpiece made out of a low-cost sheet metal;
• Fig. 4 is a monolithic embodiment having a plastic tube conveying symmetrical waves, a thin section conveying mode-converted shear waves, and permanent or removable transducers, flange-mounted for illustrative purposes of removability;
• Fig. 4A shows a mode-converting embodiment which is part tubular and part sheet, the tubular part comprising the major part of the extension portion of the assembly, said tubular part being of thin wall and therefore capable of conveying with virtually no dispersion, compressional waves in a quasi-extensional or quasi-S 0 symmetrical plate wave lowest-order mode, where these waves are launched from an easily-coupled, easily removable longitudinal wave transducer, and where these waves leak off into the sheet as shear waves, vertically polarized, at an angle of incidence similar to that of the waveguide of Fig. 1;
• Fig. 4B is another monolithic embodiment, which is geometrically similar to the plastic embodiment of Fig.
• Fig. 4C is another mode-converting structure which, when partly wrapped around the pipe, is particularly well-suited to launching horizontally-polarized shear waves around a section of pipe of the same material or same sound speed as the foot portion, and in addition, in different pipes, to launching vertically-polarized shear waves from a surface-coupled transducer of Rayleigh waves that can be coupled to pipes of lower sound speed than the waveguide, e.g., to brass pipe from SS waveguide;
• Fig. 4D is another mode-converting waveguide having a rod buffer
• Fig. 5 shows a buffer waveguide embodiment configured to introduce two different modes into a pipe, shear waves at oblique incidence and longitudinal waves at normal incidence;
• Fig. 6 illustrates differential signal processing adapted to one or more pairs of waveguides;
• Fig. 7 shows noninvasive, externally mounted embodiments configured to measure transit time with the wind and across the wind but not against the wind, as a means of measuring flow by a transit time method when conditions do not allow for strong enough signal propagation against the direction of flow;
• Fig. 7A is a wetted-transducer version of Fig. 7;
• Fig. 8 shows an asymmetric duty cycle where interrogation time spent against the wind is greater than with the wind responsive to the signal to noise ratio
• Fig. 8A illustrates a system with two identical buffer waveguides of the invention, one clamped to the outside of the pipe at the six o'clock position, while it is "wetted" mate, at twelve o'clock, passes through an oblong hole.
• the present invention is a waveguide which acts as a buffer to provide a remote coupling face for signals from an ultrasonic source or transducer, a body for non- dispersive propagation of the signals, and an exit face for coupling signals out of the body, either into a pipe wall, or directly into a fluid, generally at a defined refraction angle along a controlled path, where the refraction angle depends in part on the soundspeed in the fluid medium.
• the waveguide is thin, and has major opposed faces, so that it is plate-like. It has roughly the shape of a hockey stick.
• Various embodiments exploit this geometry for structural features, for flow-compatible intrusions into the pipe and for signal enhancement in particular situations.
• the buffer thickness is preferably not less than about half the pipe wall thickness, for welded high-pressure installations.
• the buffer were too thin compared to the pipe wall, the relatively thick weld required by codes or normal welding practice would degrade signal and increase short circuit. It has been found in tests on SS ten-inch schedule 40 pipe having a pipe wall about 3/8-inch or 10- mm thick, that standard fillet welds yield satisfactory signals when the thickness of the waveguide is about 6 mm.
• the force on the weld is less in pounds than the numerical value of the fluid pressure, e.g., 1500 psig or kg/sq inch, force -1000 pounds or kg.
• the same small area means a modest coupling force of 1000 pounds or kg yields a high coupling pressure, 1500 psi or kg/sq inch, respectively.
• the new design is appropriate for both clamp- on and wetted applications.
• Table 1 provides a summary overview of structural features which may be incorporated in different embodiments of the invention, and systems or processing protocols implemented with these waveguides, and different methods of making the waveguides.
• Material of Waveguide Usually steel or stainless steel; the standoff can 1, 1M, be all or partly plastic, ceramic. 2A
• Temperature Range for Typical Application Normally -40 to +125°C; 1 extremes from -196 to +300°C, depending on the length of extension.
• Temperature Compensation Derived from T measurement or from reference echo at time t w ; other echoes can be generated for additional UG, information on T distribution. Temperature sensor tabs can be placed 1H outside the main beam without degrading the SNR.
• Axial Flow (V z ) is main objective.
• Secondary Flow is secondary objective, but can be essential to achieving high accuracy in V z .
• Secondary flow includes swirl, swirl angle ⁇ , circulation Y, and the crossflow vector V x which is the resultant of orthogonal components V ⁇ ⁇ 2E and V x 2- Present invention can be combined with prior art or co-pending new measurement techniques for measuring sound speed c and/or secondary flow components in a plane perpendicular to the pipe axis.
• c can be measured at high temperature using a tube waveguide, a bundle of small-diameter wires in a tube, or a marginally-dispersive waveguide at each end of the diameter.
• 1J Monolithic: formed in one piece by casting; by welding a cap onto the buffer extension; folding a sheet metal preform into a cylinder and wedge; 1, 1H (injection) molding.
• Folded sheet metal can be welded together at or near 2A, the edge bounding two adjacent sheet sections of like shape and 2B dimensions, as a step to simplify subsequent seal-welding into a slot, or to add rigidity.
• Rigidity can also be enhanced by adding ribs, or "flanges" at the edges, mainly outside the main beam, said flanges creating a local cross section resembling a small I-beam.
• Pipe material usually steel or SS; can be plastic or other material.
• Plastic pipe allows use of off-diameter GC paths without the transducers 2, 2F, penetrating the pipe, in which case flat- ended narrow reflectors, intrusive 3, 3A, or recessed in axially-elongated slots in are preferably used to reliably 3B, 3C reflect and thereby utilize that portion of a spreading beam that lies in the preferred GC planes.
• Regions of the pipe corresponding to slots may be milled part way through the wall from the outside and inside, leaving a pressure-bearing window that locates and mates with its respective transducer on the outside, and radiates into the fluid inside the pipe.
• the pipe can be made of sheet metal, preferably reinforced by flanges or other cutouts to maintain the sought cross section, usually round.
• Piezoelements reciprocity; multiplicity of piezoelements or standoffs. 1A, IE, 5, 6, 7,
• Fig. 1 shows an elongated nondispersive directive waveguide 10 suitable for installation on a pipe, conduit or other solid body, for measuring flow, liquid presence or level, or other fluid characteristic.
• Installation as described below is nonpenetrating and noninvasive, such as clamp-on, braze-on or weld-on, or a wetted mounting (penetrating the pressure boundary) but not necessarily protruding beyond the inner surface of that boundary, and where even the wetted mounting may include clamp-on or clamp-in, braze-in and weld-in.
• the waveguide may be removably clamped or permanently bonded to a window formed in a pipe at specific regions according to the desired paths of integration. (An illustration of this detail appears in Fig. 2F, where removable hockey stick lOrmv is about to be clamped against window lOw.)
• the waveguide 10 includes an elongated buffer 12 having an ultrasonic source, namely transducer 15 coupled to one end, leading to a shoe or wedge 14 at the other end, which has an exit face 16 for launching the ultrasonic energy received from the source.
• the buffer propagates the energy non-dispersively as a horizontally-polarized shear wave along a straight and non-mode-converting path, while a face 18 internally redirects the energy traveling along the buffer 12 so that it strikes exit face 16 at a defined angle of incidence.
• a plurality of features 13 located away from the principal ray path along the buffer provide reference temperature signals of known spatial origin, allowing direct calibration of transit time in the buffer. These are illustrated in the composite schematic Fig. 1A described below.
• the buffer 12 and lower portion 14 will be generally referred to herein as the "handle” and “blade”, by analogy to their hockey-stick shape. Both have a rectangular or plate-like cross-section of high aspect ratio. This results in a relatively large surface area of the buffer portion, making it particularly effective as a passively cooled thermal-stand-off, and results in a relatively small surface area of the edge face 16, making it an easily coupled and highly efficient emitter with a wide footprint relative to the signal wavelength in the fluid.
• the small area of the edge face 16 has a numerical value preferably less than about one-tenth of the square of its length. This small area minimizes heat transfer from the fluid or pipe into the stand-off.
• the high aspect ratio yields a wide yet directive beam in the fluid in one plane, and an initially narrow beam that exhibits greater divergence laterally.
• the blade portion receives energy along oblique relatively shallow paths near the exit face 16 from the redirecting face 18.
• This region, bounded by face 16 which contacts the conduit may be substantially isothermal so that the ray paths from transducer to face 18 and from face 18 to face 16 are each substantially linear.
• the buffer 12 may be oriented generally perpendicular to the source of heat at the conduit surface so that propagation in the buffer also proceeds without refraction.
• FIG. 1 the drawing shows one commonly used form of this device.
• a number of early models were machined out of 1/4-inch-thick carbon steel or SS (stainless steel) type 316 plate, to form a generally-hockey-stick shaped waveguide 10 having a narrow thickness x, an overall width z + w and a height v.
• the waveguide which shall occasionally be referred to as a hockey stick, may be said to be comprised of a blade or foot portion 14 and an extension portion 12 joined in such a way that a reflecting surface 18 reflects without mode conversion substantially all the incoming shear waves incident upon it.
• the dimension w is also the approximate width at the top of the waveguide where the piezoelement 15 is mounted.
• piezo end Preferred shapes for high temperature applications are those for which y»z»x.
• the height of the foot portion may be about twice w.
• x 6.4 mm
• z 70 mm
• v 200 mm
• w 35 mm.
• Early steel hockey stick waveguide models built approximately according to these dimensions utilized shear wave piezoelements 6.4-mm wide x 25.4-mm long by 0.5, 1.0 and 2.0 MHz thick.
• the piezo element thickness is the resonant thickness at the fundamental frequency.
• Preferred values for v depend primarily on the anticipated temperature range, and the thermal and acoustical properties of the material chosen for the waveguide.
• Waveguide materials include metals such as Ni- plated CS (carbon steel), SS (stainless steel), Ti, brass, aluminum; ceramics; plastics including polyamide-imide or phenolic plastics for high temperature, CPVC or PEEK (poly ether ether ketone) for more ordinary temperatures, and graphite or composites, for example.
• metals such as Ni- plated CS (carbon steel), SS (stainless steel), Ti, brass, aluminum; ceramics; plastics including polyamide-imide or phenolic plastics for high temperature, CPVC or PEEK (poly ether ether ketone) for more ordinary temperatures, and graphite or composites, for example.
• Piezoelement 15 is a shear mode piezoelectric material such as PZT, with particle motion parallel to major surfaces of waveguide 10 and perpendicular to the shear wave path in extension 12, and dimensioned as mentioned above.
• the angle of reflecting surface 18 makes an angle of 30 deg to the normal to the pipe, said normal being erected perpendicular to the pipe and passing through the midpoint of the bottom surface 16 of the hockey stick. That surface will sometimes be referred to herein as the radiating hypotenuse, a terminology discussed further below in connection with Fig. 1A.
• a shear wave coming from the piezoelement at the piezo end and striking surface 18 reflects or is guided and subsequently strikes surface 16 at an angle of incidence of 60 deg. This is known to be a desirable angle of incidence, for obtaining ultrasonic flow measurements by clamp-on contrapropagation applied to liquids in many standard carbon steel or SS pipes.
• the top end of the hockey stick is terminated in a SS316 externally-pipe-threaded adapter cap.
• a standard BNC or other electrical connector is installed at the end of that adapter cap.
• a short electrical lead wire connects from a terminal on the electrical connector to the ungrounded electroded side of the piezoelement 15.
• the cap is typically fillet- welded to seal around the top of the hockey stick's extension 12, before the piezoelement is installed. Applicant has found that welding around that joint does not degrade the ultrasonic shear waves guided within and through that welded region.
• the adapter cap is simplified to a standard one-inch 3000# SS316 cap, for example.
• the cap is oblong-slotted at one end to match the waveguide cross-section to which it is subsequently welded.
• a standard one-inch SS316 hex head plug is then threaded into the cap. The plug is tapped at its hex head end to accept the BNC or other electrical connector.
• the cap region in preferred designs remains relatively cool, under 125°C, even though the opposite end is coupled to a hot pipe at temperature, >260°C.
• the cool end ensures that the electrical properties of the piezoelement and adjacent bonds or electrical tuning elements, e.g., an inductor, remain substantially constant.
• a thermoelectric cooler is installed within the cap (202TE in Fig. II), to further limit the temperature excursions of the piezo or electrical elements.
• the thermoelectric cooler turns on only if the cap region thermocouple senses that the cap temperature reached a T setpoint, for example, 99°C.
• the cooler runs until the cap T drops, e.g. to 95°C.
• a hole 17 of about 12.7-mm diameter is drilled through the foot 14. This serves several purposes. During manufacture it holds the stock in place, for machining. In application it provides a place to apply high coupling pressure, -1000 psi or more. It also locates the clamp-on version within a yoke having a corresponding hole, as illustrated in Figs. 3A and 3B, below. Above the hole, on the surface opposite the radiating hypotenuse, there is a small drill point or indent 17a, which is another point for application of coupling pressure, or for alignment.
• Another topological feature is the small shallow groove 17b which is oriented parallel to the shear wavefronts incident upon it. Its reflecting edge begins where the hole centerline intersects the bottom surface 16.
• This groove about 1/4- to 1/2- mm deep, and 3-mm wide, is positioned so the time for the shear wave to travel along the path centerline from the piezoelement 15 to the groove 17b and back to the piezoelement 15, and denoted t w , is indeed the time it would take for ultrasound to travel along the centerline path between two piezoelements, if their identical hockey sticks were butted at the radiating hypotenuse ends and with point symmetry about the centerlines of aligned hole(s) 17.
• t w hole Another hole (of which two examples are shown in Fig. IB) can be drilled or otherwise machined parallel to hole 17, through approximately the midpoint of the shallow groove, to increase the amplitude of the t w reference echo.
• This "t w hole” may be round or square and would typically be approximately the same size as the width of the shallow groove 17b, namely, about 3 mm in the present example. Only a small percentage of the shear energy reflected off the surface 18 is reflected by the shallow groove. Most of the energy reflecting internally off the bottom 16 bounces off the end surface which may be rounded to scatter this energy. However, a relatively strong end echo may nevertheless be detectable by the piezoelement 15, and its arrival time after the t w echo can be interpreted in terms of sound speed and temperature in the foot 14.
• the ratio of amplitudes of the two echo signals, from the groove or hole and from the rounded end, is a normalized measure of the coupling efficiency into the adjacent medium to which the radiation is directed.
• the amplitude of the groove echo also is a measure of the electroacoustic strength of the piezoelement (k31) and the attenuation along the path in the waveguide. Ringdown in the waveguide, however, may also be used as a measure of waveguide attenuation. At all events, these echo signals provide reference points for correcting time or amplitude values of measurements and protocols in the system.
• a weld bead may optionally be provided as a shallow bead whose end nearest the bottom ends rather precisely, achieved by machining, at a height appropriate to the depth that the bottom portion 14 is to be inserted into a slot. This simplifies weld-in installation. This same function is achievable by thinning one or both sides of the foot by .005 inches per side to the appropriate height, or by center punching at that height.
• rounded tabs 13 projecting outward from the right-side bounding surface of extension 12. That surface preferably is rounded to scatter ultrasound incident thereon, just like the end surface 18a drawn parallel to reflecting surface 18 (Fig. 1).
• the rounded tabs are alongside yet out of the main beam, so each can be tapped e.g. #8-32 for installation of a short (3/16-inch length) standard panhead SS screw to hold thermocouples or other temperature sensing devices. In high temperature applications this is useful to keep track of the temperature distribution along the extension.
• a similar blind threaded hole can be tapped in the cap, to attach a thermocouple there to monitor that the cap region does not exceed say 100 deg C.
• the rounded surface note that while it runs generally parallel to the flat edge surface between reflecting surface 18 and the piezo end, its contour creates asymmetrical boundary conditions with respect to the beam centerline path. This helps discourage interfering multipath propagation.
• Fig. 1A shows a composite schematic view of a waveguide similar to that of Fig. 1, except that the piezoelectric element is shown in the form of a multi-element construction 15a, 15b. This is done to achieve ease of fabrication, or for multi- frequency operation, or for creating a wide beam within one sensor, leading to a large aperture at the radiating surface 16. Lettered points indicate important vertices and other important points along the ultrasonic path.
• the device can be understood to be a combination of a waveguiding triangular wedge portion 14' in the foot of the device indicated by letters A, B, C, and a polygon-bounded waveguiding extension 12 delineated by letters BHDEFC which extends outward to where the piezomaterial is located.
• the phantom buffer BCE'D' illustrates the "virtual" position of the buffer with respect to the reflective face 18 centered at principal reflection point Q5.
• the extension portion may actually occupy the position indicated by 12', so that the entire waveguide provides a straight path to the exit face. In this case the edge along side D'B guides but does not deflect the energy.
• the invention contemplates that the extension portion may occupy a number of different angular orientations when used in an embodiment that guides wave energy to a redirecting face, and the angle may be selected based on the desired angle of incidence at the exit face for the conduit or the fluid/velocity involved.
• Fig. 1A also illustrates the right triangle wedge ABC (shaded) and a waveguiding polygonal extension generally defined by or including points CBDE, together with certain topological features for sensing or controlling shear-wave transducer position, shear-wave propagation, and positioning or coupling the assembly on an adjacent pipe.
• Leg BC of the wedge is parallel the incoming shear wavefronts 15e.
• the transducer S is a source (or receiver) of horizontally polarized shear waves having particle motion (indicated by a double-ended arrow) perpendicular to the path and parallel to the first and second major faces ABHDEFJA and AA...JJ,AA.
• Topological features are formed in the waveguide. These preferably include a small blind hole reflector at Q4. The hole at Q4 is drawn under and approximately tangent to the plane that is perpendicular to Q1Q2 that passes through Q2.
• Another topological feature is the small t w reference reflecting groove which is drawn as a shallow slot between Ql and Q3. The term t w is the round-trip transit time along the centerline path between S and Ql . This groove is oriented parallel to the incoming shear wavefronts
• shear ray paths are drawn as dashed line segments 15c,d parallel to standoff centerline 15f.
• Optional projections TCI, ... TC 5 may be formed in the body 12 as small tapped tabs alongside the signal propagation path but more or less out of the way of the main beam. Their purpose is for attaching thermocouple, resistance temperature detector or other temperature-sensing means, from which the temperature distribution and the corresponding sound speed distribution along the path may be determined. In one embodiment, by intentionally drilling the tap drill holes extra deep and thus slightly into the path of the main beam, reference echoes are generated along the standoff.
• the triangular wedge region ABC may be filled with a material having special waveguide properties needed adjacent the radiating hypotenuse edge AB.
• a material having special waveguide properties needed adjacent the radiating hypotenuse edge AB For example, in one embodiment an assembly mostly made of SS316, and having a shear- wave velocity of about 3100 m/s, has a cavity of width about 90% of the thickness X ⁇ , milled, cast or otherwise formed in the ABC region. This cavity is then filled with silver (Ag) or other material to provide a much lower vertically polarized shear- wave sound velocity (1610 m/s) at the angle of incidence appropriate to radiating at a large refracted angle ⁇ 3 into an adjacent fluid, when the assembly is used as a
• the hockey stick construction of the present invention may operate at several frequencies. It is a broadband waveguide.
• Two (or more) piezoelements could be installed at one end, and launch, for example, pulses having quite different center frequencies such as 0.5 and 1 or 1 and 2 MHz. These pulses may be launched simultaneously or sequentially, and will propagate toward the radiating hypotenuse along the thin blade at the shear wave phase velocity, about 3100 m/s for the commonly used 300-series alloys of stainless steel.
• one may launch pulses having center frequencies of 2 and 4 MHz in some sequence, e.g., alternately frequency hopping from one frequency to the other and then back again.
• Another modulation scheme contemplated with these waveguides is to simultaneously amplitude-modulate each of two frequencies according to the square of a sine or cosine function, respectively.
• This can produce a composite wave which, at the start, could be 100 percent formed of the lower frequency component (say 2 MHz).
• the amplitude of this component then diminishes from an initial value down to zero according to a cosine squared envelope over signal interval, e.g., five microseconds corresponding, by way of example to ⁇ /4.
• the higher frequency component say 4 MHz
• Fig. IB is a conceptual drawing showing the range of angles that the extension or leg portion 12 can make relative to the wedge or foot portion 14.
• the foot is coupled through or to the outside of the pipe wall Wto project a signal through fluid F.
• the shallow reflecting groove 17' is shown across the incoming wave path to reflect a reference signal back up the standoff portion, and an alternative or additional round through hole 17rdh or square hole 17sqh are shown. These may augment the amplitude of reflected energy, or may be used separately instead of a groove.
• the leading edge of these holes is positioned tangent to the leading edge of the shallow groove.
• Fig. 1C shows details of a multi-angle construction in which the redirecting face 18 is stepped to provide three different faces 18a, 18b, 18c reflecting rays a,b,c at three angles of incidence in a single waveguide sensor. This introduces the shear wave at different angles of incidence, in anticipation of different fluid sound speeds, which may for example compensate for such differences and yield particular angles of refraction.
• Another use of this feature is to create a plurality of precisely-defined beams a,b,c which intersect in the fluid, to make reflection-mode measurements, for example, when the fluid is of a scattering nature.
• Fig. ID shows another multi-angle construction in which the reflecting surface 18 is a curved reflecting surface that generates a continuum multiplicity of angles of shear-wave incidence ⁇ from 30° to 80°.
• Point B corresponding to point B in Fig. 1A, is at the bottom, on the extension of line E-EE. This means E-EE-B is a straight line.
• Fig. IE shows a plurality of n identical waveguides placed in a row to increase the system tolerance to a wide and unpredictable range in sound speeds in the fluid. This also increases the aperture by at least n times, according to the gap between them being zero to the order of the radiator length X4, i.e. the length of the exit face 16.
• the illustrated mounting of identical waveguides at top and bottom positions as transmitters and receivers exploits the reciprocity capability of piezo transducers. That is, each piezoelectric element can be used as a transmitter, or a receiver, or serve both functions in alternation, effectively transducing (i.e., transforming) vibrational to electrical signal energy in either direction.
• the magnetostrictive transducers discussed below with respect to Fig. 4D are also reciprocal (send/receive) devices.
• At the right is shown a monolithic assembly in which a single body has three standoffs each holding one or more transducers.
• Fig. IF shows a spoolpiece in which a plurality of waveguides in accordance with the invention are attached to interrogate various chordal or diametral paths. These include waveguides which are sharply bent and welded 10a or clamped 10a' to the pipe, waveguides which are gently curved and either welded or clamped 10b, 10b', straight, i.e. planar, waveguides 10, and penetrating waveguides 10c for directly launching a signal along a direction into the fluid.
• the midradius circle MRC is shown at a radius equal to half the inside radius of the pipe.
• FIG. 1G shows a waveguide 100 in which secondary reflectors 18', 18" limit the active length 16' of the exit face 16, and the primary and secondary reflecting surfaces are protected against fouling by covering, concealing or otherwise reducing their exposure to possible sources of debris.
• Elongated angled quadrilateral slots may form reflective cavities only part way into the sheet to form these faces.
• the hole 20 goes through.
• the "toe” and “heel” of the foot portion can be parallel. This makes it easy to fit to and install into a slot similar to those in the wetted versions (e.g., Fig. IF, transducer assembly 10c, or the devices of Fig. 3, 3 A or 3B.
• Fig. 1H shows a sheet-metal stamping to implement the embodiment of Fig. 1G.
• the stamping has first and second faces parallel to the plane of the drawing, in which half-cavities can be formed illustratively on a first face in a pattern symmetrical about a fold centerline, such that after folding sharply about that centerline through an angle of 180°, the half-cavities become adjacent as the two halves of the first face touch one another, and thereby provide shielded reflecting surfaces functionally similar to those in Fig. 1G, invisible from the second face, which now forms the outer surface and bounds the folded configuration.
• the resulting external seam can be welded shut.
• the two adjacent halves can be spot welded together at various points (denoted by asterisks) outside the ray paths indicated by lines L, L'.
• Fig. II illustrates a sectional view through a waveguide 200 of the present invention, in which a cap assembly 202 at the transducer end covers attenuative potting such as tungsten-loaded epoxy to damp reverberations near the capped end of a buffer.
• the distance from the radiating hypotenuse 16 to the potting, X p is about equal to buffer length X3 minus buffer width XI.
• Fig. 1J shows a representative system 210 employing a combination of waveguides at five locations on a pipe 1.
• the pipe 1 has a horizontal section and a vertical riser section V connected thereto by a swept elbow la.
• the horizontal section of pipe is shown in partial section showing a conduit only partly full of liquid.
• a pair of weld-on waveguides lOd and 10d' are welded at a first location underneath the pipe.
• a residue lOr may be present on the bottom of the pipe. When the residue lOr is absent and the liquid is relatively clean, one may expect to be able to measure the liquid level by timing the echo from the liquid surface S, provided one has a reasonable estimate of the sound speed in the liquid. If there is much residue, it may be more accurate and more reliable to measure the position of the surface S by interrogating from above by means of a pair of waveguides lOe and 10e'.
• the horizontal section H may be full and the liquid may rise into the vertical riser section V.
• a third pair of waveguides are clamped in a horizontal plane, i.e., the waveguides lOf, lOf are clamped on the side of the riser in a plane perpendicular to the axis of the pipe in that region.
• These waveguides may be variously oriented according to whether they are to sense propagation of a zigzag shear wave in the wall according to a method explained in Lynnworth, Seger and Bradshaw, U.S. Patent No. 4,320,659 (March 23, 1982), or are to sense along a chordal path.
• a valid measurement of liquid level in FIG. 1J is straightforward if one has a reasonable estimate of sound speed c in the liquid. More generally to cover the case where c is not known, applicant utilizes multipaths having liquid endpoints either at the pipe ID or at the radiating hypotenuse of a wetted hockey stick, or both. The basic idea is to interrogate the liquid surface using acoustic paths which differ in length in the liquid, where the difference in path lengths is known or calculable from the geometry of the pipe and the transducer properties and locations. Then one solves for the two unknowns, sound speed c and liquid level H, based on transit times tp and tp2 measured over two independent paths PI and P2.
• the path is refracted where the ultrasonic wave enters and leaves the liquid, but this minor complication is dealt with using Snell's Law.
• the speed of sound in the hockey stick is determined from the material, e.g. 3200 m/s for 316SS near room temperature, or if desired, by timing echoes from the reference notch to the free end parallel to that notch and spaced a known distance away, e.g., 25 mm.
• Liquid presence at particular levels can be determined from the presence of a signal when the waveguides are placed at appropriate axial or circumferential positions.
• the GC spoolpiece of FIG. 3, 3 A or 3B may be seen as a specific example of how to use different paths whose path difference is very well defined, to measure H when c is unknown.
• One can additionally execute further measurements between upstream and downstream clamp-on transducers at or near the pipe bottom, said clamp- on transducers having appropriate spacings achieved via side-by-side arrangements or via different axial spacing as indicated in FIG.
• a solution for c, Hand then, from the difference in transit times upstream minus downstream, the flow velocity V.
• Fig. 1J Still referring to Fig. 1J, to accommodate such measurements there are provided at the left and right ends of the horizontal run of pipe, an additional set of side-by side waveguides lOh, 10h', and an overlapped arrangement of waveguides lOg, 10g' of the present invention. These views are related to the ones in Fig. IF.
• the axial spacing and exit angles of the two different waveguide pieces lOd, 10d' and lOg, 10g' provide two fluid paths PI and P2 of differing length for determining c as described above, illustrated with reflection or double-reflection paths off the fluid surface.
• the waveguides may be circumferentially adjacent, or axially adjacent with a slight circumferential offset, and still effectively utilize the illustrated sending/receiving paths due to the directional but spreading beam that emanates into the fluid on either side of the central plane of each wedge portion.
• the long radiating hypotenuse yields a directive beam in the fluid or adjacent solid in the end-fire sense. This means the beam spread is relatively small about the refracted angle ⁇ 3 calculated by Snell's Law, for the obliquely incident shear wave.
• a consequence of the thin hockey stick is that there will be much larger beam spread side to side. This is used to advantage in the side to side arrangements.
• variable liquid level provides a reflecting surface within the common beams.
• the scatterers can be detected by two side-by-side hockey stick assemblies, both of which point in the same general direction, e.g., into the flow.
• a further consequence of the spreading beam about the hockey stick midplane is that even members of a pair that are not side-by-side but are somewhat separated circumferentially, can nevertheless communicate with one another. For example, skipping ahead briefly to the circumferentially spaced assemblies shown in Fig. 3A, below, and implied in Fig. 3C, there is detectable and usable signal between what we will call out-of-plane transducers. This communication can be via scatterers or via reflection from the opposite wall, or via direct transmission if transducer assemblies are positioned on opposite sides of the pipe. What this amounts to is a new way of providing with only a few transducers, a large number of interrogation paths.
• the principal axially-oriented slots may be denoted A and AA, utilizing reflector R A , slots B and BB using their reflector R B , etc.
• a transducer installed at A can communicate with not only AA but also BB, CC and DD. This provides a way to cat scan the flow profile and image the flow by tomographic reconstruction. Similar remarks apply to reflection mode measurements.
• a pulse transmitted into the fluid by A can be received at a number of receiving sites B, C, or D, as well as back at A.
• the invention provides new and useful opportunities for additional interrogation paths in a multi-plane spoolpiece.
• Fig. IK shows the end view of a pipe 1 that contains a fluid with scatterers s, e.g., water plus air bubbles, and depicts a reflection measurement of scatterer velocity during the quiet time t ⁇ between the waveguide end-to-end echoes A & B (bottom trace on the oscilloscope).
• the change in transit time ⁇ t$c is obtained by stroboscopically timing echoes (top trace) from the ensemble of scatterers within the scattering cell v, using two (or more) transmissions, according to the method of Jacobson et al., U.S. Patent No. 4,787,252 (November 29, 1988).
• the other principal components in the system are the commercially-available pulser/receiver and a preamplifier which may be tuned.
• the scattering cell or interrogation volume v is defined by the beam intersections and the range gate timing windows.
• FIG. IL shows a differential mass flowmeter for organic hydrocarbon (HC) liquid that combines Rao's Rule for organic liquids and two flow velocity measurements, on the hot and cold pipes for a diesel engine.
• the cold incoming and hot outgoing flows are of the same composition but differ in temperature (T ), density (p), sound speed (c) and volumetric flow rate (Q). Temperature does not need to be measured er se, as the fractional density difference ⁇ p/p can be accurately estimated as (1/3) Ac/c, one-third the fractional sound speed difference, based on relationship due to M. R. Rao, Indian J. Physics, Vol. 14, p. 109 (1940), cited in A. B. Bhatia, Ultrasonic Absorption, p. 24, Dover Publ. (1967).
• the ratio of mass flow rates simplifies to a simple function of V j , V2, c j and C2 provided the meter factors Kj, K2 can be estimated accurately enough from the interrogation geometry, cj and C2, or perhaps from other information.
• the c's may indicate temperature accurately enough to determine the kinematic viscosity v, from which K and Re are ordinarily determined.
• the attenuation coefficient ⁇ is used to estimate Re, again leading to K. Similar arguments apply to organic hydraulic fluid, where a AMp would be attributed to a leak.
• Fig. 1M shows the hockey stick construction adapted to a downhole tool 400.
• hockey stick waveguides 410a, 410b are mounted to extend through the wall 401 of the tool and their blades 414a, 414b are welded with through- welds thereto, while their shafts 412a, 412b extend within the tool interior.
• the downhole environment is one in which thermal control is difficult or not feasible, and this problem is addressed by a further aspect of the invention by forming the shafts of a first, metal portion 4', and a second portion 4" formed of ceramic and having a low thermal conductivity (as well as a low coefficient of thermal expansion), which extends from the metal trunk and mounts to the transducer.
• the ceramic part 4" is clamped to part 4' at a flanged metal to ceramic joint JMC.
• the blades 414 are thus positioned to radiate energy radially outward normal to the tool surface and detect signals indicative of conditions in the annular drilling mud region surrounding the tool, to determine one or more of viscosity and density, or a function thereof, of the surrounding material.
• a vacuum- jacket VJ insulates the electronics module EM.
• Fig. 2 shows a spring loaded internal clamp jig 150 to hold or set the insertion depth for several waveguides to be welded onto or into a pipe section, for permanent installation at controlled locations, particularly at accurately controlled depths of insertion in slots.
• a pair of spring biased cylindrical bodies 151, 152 rest tangent to the inner wall of the pipe, and waveguides 10 or reflectors R are inserted to abut the jig before they are welded.
• Fig. 2 A, B show two views of the waveguides cast permanently into the spoolpiece 250, which in this case is a spirometer compactly configured to measure breathing dynamics.
• the device may be fabricated as a monolithic plastic flowcell, shown in side view, end view and electrical schematic view in Figs. 2A, B&C, respectively.
• the spirometer consists of a square duct and four or six transducers for simultaneous measurements in orthogonal planes and/or in one plane. Extension portions of this multipath spirometer may be separable at J-J to accommodate sterilization procedures.
• the spirometer flowcell body, or at least the snorkel mouthpiece, may be a low-cost expendable item. It connects to the mouthpiece end 250b. The opposite end is the hose end 250c, which terminates in a standard hose connection 250d.
• Transducers preferably are reusable items for use with another spirometer body, unless the transducers themselves can be made so cheap that they can be treated as expendable too.
• one preferred fast-response construction uses only four transducers and places all four in one plane.
• the four transducers including their oblong caps, fit between planes formed by the extensions of the outer walls of the square section. These boundary planes are indicated by the dashed ( — ) extensions BB in the end view, Fig. 2B.
• the wedge plus extension transducer assembly functions like a clamp-on airflow transducer, constructed in monolithic fashion, with one or more portions optionally removable and re-couplable.
• Axial interaction length L is kept short so that c is essentially constant along that length, despite patient's exhaled breath not necessarily remaining at a constant temperature, humidity or composition, over time.
• Fig. 2C are label P ⁇ and P D , respectively.
• the "footprints" 288, 289 of the optional side transducers 250e,250f are shown as rectangles in the side view.
• the hockey stick material has shear wave velocity within approximately 10 percent of that of water near room temperature, i.e., roughly 1350 to 1650 m/s, and has a relatively low attenuation coefficient at 500 kHz, less than 3dB/cm as well as a relatively low density, less than twice that of water, i.e., less than ⁇ 2g/cc with the flowcell conduit having a substantially rectilinear or square cross section such that the corners tend to block acoustic cross talk between transducers on adjacent and on diametrically opposed sides of the conduit.
• the low density and low shear wave sound velocity in combination with corner-blocking of crosstalk allow one to carry out high signal to noise ratio ultrasonic flow measurements of air at ordinary temperature and pressure utilizing a relatively large refracted angle in the air within the conduit.
• the refracted angle ⁇ 3 is set so sin ⁇ 3 exceeds 0.2, i.e., exceeds the ratio of sound speed in air to that in water, or, roughly, ⁇ 3 is greater than 10°.
• a phenolic plastic such as is employed in the manufacture of brake linings, has been found to be a suitable material and performs well for this purpose.
• the oscilloscope is preferably a two-channel oscilloscope as illustrated with delayed time base triggered by the transmit pulse, with the received signals fed through preamplifiers, which may be tuned preamplifiers.
• Fig. 2D shows another useful embodiment wherein the waveguides form a permanent construction for measuring flow in a buried conduit.
• a pair of axially-spaced waveguides are clamped to interrogate paths in a non-vertical plane to avoid reflection from the generally wet or dirty bottom surface.
• the clamp may be a unit such as shown in U.S. Patent 4,286,470.
• Fig. 2E shows waveguides temporarily coupled to a pipe in planes parallel or perpendicular to the pipe axis by pipe riser clamps in an economical embodiment appropriate for ordinary clamp-on applications.
• the waveguide passes through a slot in the clamp and is welded thereto.
• the waveguide is welded on the outside surface of the clamp, and a rectangular weld build-up on the inner surface provides a continuation of the metal path between the transducer and the pipe.
• the clamp passes through and is welded to a slot or cut-out in the body of the waveguide.
• the clamp and waveguide are assembled by butt-welding, or by through- welding in a slot, so as to provide the path equivalent of an uninterrupted waveguide.
• Fig. 2F shows several hybrid constructions adapting waveguides of the invention in a system to provide one or more of: (a) removability; (b) explosionproof design for low sound speed pipes, achieved by utilizing a composite waveguide in which the wedge or foot portion, and part of the extension, are of low sound speed. This yields a large refracted angle in plastic pipe.
• the piezo end of the buffer extension portion is of relatively high sound speed typical of metals and this metal is threaded, capped, sealed and otherwise assembled to make an explosionproof assembly.
• the use of low sound speed material in the foot provides refraction angles in the contained fluid beyond limits imposed by the pipe material's sound speed relative to that in the fluid.
• Fig. 2F also shows a removable waveguide 1 Ormv which is secured by a clamp (not shown) in a slot milled into the pipe surface so that it is aimed precisely along a vertical (as shown) plane.
• a corresponding countermilled flat is formed on the pipe interior surface, leaving a window lOw of uniform thickness in the conduit wall as the floor of the receiving slot. This assures that the signal is launched along a precise direction into the fluid, and also maintains a pressure boundary which allows removal and replacement of the waveguide in the field.
• Fig. 3 is a perspective view showing the eight slots on the front side, and the four slots for accommodating reflectors on the back side of a specially slotted section of pipe designed for use as a Gauss-Chebyshev multipath flowcell in which off-diameter chordal measurements are made in well-defined symmetrically-located inboard and outboard planes.
• Eight waveguides of the type shown in Fig. 1 are to be welded in a minimally- intrusive manner into the illustrated narrow oblong slots that are all located on one side of the spoolpiece and which, for convenience, is referred to as the "front" side.
• the spoolpiece is of a size that provides opportunities to mount additional sensors to independently measure swirl, circulation, sound speed or crossflow. This creates an integral spoolpiece configured for sensing both axial and secondary flow components. This construction advantageously places the passive reflectors in one location, which may, for example, be inaccessible once the spoolpiece is installed in a plant.
• the invention also contemplates forming internally-milled transmitting flats with blind flat-bottom externally-milled slots to provide pressure-bounding windows several mm thick (not shown). There windows locate the transducers, which may be removably clamped or epoxied thereto.
• Fig. 3B shows a partly sectioned side view of a GC spoolpiece showing several clamped-on transducers at the top and bottom. Yokes Y are shown explicitly, near the top centerline and at the bottom.
• the clamped-on waveguides may then be calibrated using GC multipath waveguides as the reference, at high temperature and/or disturbed flow conditions.
• the enhanced directional beam-forming structure of the present waveguide may be used with simple sheet structures to provide highly accurate measurements.
• Fig. 3C shows the layout for one such GC and diameter path spoolpiece made out of a low-cost sheet metal stamping. Slots are shown for transducers at locations A, AA, B, BB, ... X, XX, X', XX'. Reflector slots are shown at positions RA, RB, etc. When formed into a right circular cylinder, the slots accommodate sheet metal or other transducers and reflectors, to provide GC and diameter paths similar to those shown in the spoolpiece embodiment of Fig. 3 A.
• the top seam joins edges T, T of the stamping with centerline slots X, X' thus positioned along a line at the bottom. All transducers are positioned on one side (left) of the conduit, with all reflectors on the other side (as viewed from the end, along the conduit axis).
• Fig. 4 shows another monolithic embodiment comprised of a plastic tube PT conveying symmetrical waves, a thin section T conveying mode-converted shear waves, and permanent or removable transducers Tx, flange-mounted for illustrative purposes of removability.
• Figure 4B is another monolithic version, made of SS. It is geometrically similar to the plastic version, but due to a smaller Poisson's ratio for SS, the cylindrical portion is tilted back more, to achieve a similar angle of incidence near 60 degrees. The sine of the refracted angle equals .707 times the square root of [Poisson's ratio plus one].
• Fig. 4 shows another monolithic embodiment comprised of a plastic tube PT conveying symmetrical waves, a thin section T conveying mode-converted shear waves, and permanent or removable transducers Tx, flange-mounted for illustrative purposes of removability.
• Figure 4B is another monolithic version, made of SS. It is geometrically
• FIG. 4A shows another mode-converting structure which is part tubular and part sheet, the tubular part comprising the major part of the extension portion of the assembly.
• This tubular part is of thin wall and therefore capable of conveying with virtually no dispersion, compressional waves in a quasi-extensional or quasi-S 0 symmetrical plate wave lowest-order mode.
• These waves may be launched from an easily-coupled, easily removable longitudinal wave transducer. These waves leak off into the sheet as shear waves, horizontally polarized with respect to the sheet's major faces, yet vertically polarized with respect to the pipe to which it will be coupled, leaking occurring at an angle of incidence similar to that obtained in the waveguide of Figure 1.
• the tube may be replaced by a thin solid rod, which may additionally be magnetostrictive to allow direct activation as shown in Fig. 4D.
• the solid rod adjacent the blade is a source of ultrasonic energy.
• the rod diameter is somewhat larger than the "thin rod" condition of about a tenth the wavelength of a compressional wave therein, an initially broadband pulse will disperse (smear) with the lower frequencies traveling faster than higher frequencies.
• This behavior is described by the Pochhammer and Chree thin-rod equations and is analogous to that of symmetric plate waves of zero order, analyzed by Lamb, and accordingly part of the group of plate waves called Lamb waves. This consideration leads us to two modes of operation for the device of Fig. 4D.
• the solid rod when operated in or near the long wavelength limit, provides a broadband pulse to the blade.
• a pulse which may be relatively broadband when launched at the solenoid, gradually smears, so that its lower frequency components reach the blade first, followed by higher frequency components. This is symbolically indicated by the waveforms drawn alongside the rod in Fig. 4A.
• the refracted angle of its leaky wave in the blade will be less than the refracted angle for the higher frequencies. Applicant uses this dispersion to create a spreading wave in the blade, much as occurs with the curved reflector shown in Fig.
• Fig. 4A illustrates a monolithic assembly made from one piece of sheet metal, e.g., SS316, approximately 1 to 2 mm thick.
• the central "fuselage” is curved into a cylindrical shell that guides quasi-extensional ( ⁇ S 0 ) waves.
• the "wings" on each side function as the wedge plus a short extension in this embodiment.
• the invention also contemplates another mode-converting structure which is partly wrapped around the pipe, and is particularly well-suited to launching horizontally-polarized shear waves around a section of pipe of the same material or same sound speed as the foot portion.
• the waveguide additionally conveys Rayleigh waves that can be coupled to pipes of lower sound speed than the waveguide, as occurs, for example with a SS waveguide and a brass pipe.
• Fig. 4C Such an embodiment is shown in Fig. 4C.
• the Rayleigh wedge (Fig. 4C) launches a wave propagating at velocity C R which is coupled into the wall W and launched in the fluid F.
• Fig. 5 shows another system utilizing hockey stick waveguides of the invention to interrogate the fluid in a particularly effective protocol.
• This protocol introduces two different modes into the pipe, shear waves at oblique incidence and longitudinal waves at normal incidence, generated independently and waveguided external to the pipe independently over most or all of their paths, to measure at the intersection of their paths in the fluid, the flow of a medium containing scatterers in a first mode of application.
• the common volume of intersection may be further restricted by range gating.
• flow is determined by a three-step process where in step 1 , flow is estimated to be conventional tilted diameter contrapropagation using a pair of oblique shear wave transducers of the type shown in Fig.
• Another refinement of the meter factor K is achieved using the waveguides of the present invention by developing a measure of conduit surface roughness from the spectrum or amplitude of the zig-zag vertically polarized shear wave that is internally reflected and travels through or around the conduit wall between the waveguide mounting positions.
• the equation of Pai which appears for example in Ultrasonic Measurements for Process Control of L. Lynnworth (Academic Press 1989) p.250, provides a relation of conduit surface roughness ⁇ personally to average flow profile.
• Analyzing the roughness/reflectivity relationship at the pipe/liquid interface(s) generally requires that one consider the frequency- and angular- dependence of scatter .
• the apparent rate of energy decay, or ringdown characteristic in the wall, the fluid, or both may yield ⁇ g to an accuracy which is sufficient to calculate K to within one percent.
• the short circuit "noise" associated with the early arrival of vertically polarized zig-zag shear may be analyzed according to this aspect of the invention to detect excess attenuation of a spectral component indicative of the surface roughness.
• This measurement can be refined by first measuring the zig-zag transmission characteristics when the conduit is empty.
• the zig-zag signals also contain spectral information that indicates the pipe schedule or wall thickness.
• the radial extent of the window is typically two to four times longer for the two-beam intersection (see K03) than for a single beam normal to the pipe (see AR j_). Therefore, the normal beam in principle can yield two to four times better spatial resolution of the flow profile, per microsecond of range gate width. This is especially important near the wall, where the profile is maximally curved and where more fluid is represented per microsecond than deeper into the flowing cross-section, away from the wall.
• the validity or accuracy depends not only on the backscattered intensity from turbulence or other sources of scatter in the fluid, but also on how well the radial unsteady flow component of turbulence can be related to the time-averaged axial component of flow velocity at that ylR.
• the waveguides provide a precise structure for aiming beams, which in conjunction with a programmable intervalometer/flow meter can accurately assemble a set of fluid measurements which fully characterize the flow.
• This property of accurately aiming the transmitted beam and range-gating (windowing) and filtering the receiver allows a number of novel configurations adapted to adverse measurement conditions or special measurement situations such as vortex shedding measurements.
• Fig. 6 illustrates a system employing differential signal processing adapted to one or more pairs of waveguides, of the present invention. These are mounted in a short pipe segment and are symmetrically located to equalize common mode noise (crosstalk) as much as possible. Then, rather simple signal processing is applied either to parallel, oblique or to crossed paths and this suffices for flow measurement or flow switch applications.
• the difference signal AS surviving a subtraction and filtering process is substantially responsive to one or more of: beam drift, turbulence induced attenuation ( a j - ot£)), vortex shedding.
• one set of preferred paths in the fluid are perpendicular to the plane passing through the bluff body and the pipe axis, as indicated in the view looking into the pipe after it has turned through 90°.
• a fast response measurement system is illustrated in the upper portion of Fig. 6 employing pulse-echo interrogation of electrically paralleled transducer assemblies A, B, C and D which have intentionally dissimilar stand-off lengths allowing signal reception in a single-channel instrument during quiet times between buffer reverberations. Standoffs A and B are made longer than the pair C and D to assure time-separation of simultaneously-launched signals that cross in the fluid.
• a 1.25-inch differential in total length separates the received signals shown on the oscilloscope delayed time base at time t AB and t CO by about 10 microseconds plus or minus the effect of flow.
• Each received signal (at t CD and at / AB ) widens in proportion to the flow velocity, the leading edge arriving earlier, the trailing edge arriving later.
• the pulser/receiver (P/R) electronics assembly 607 may be the Panametrics Model 5055PR or similar commercially available equipment, operable via its front-panel toggle switch or knob in pulse-echo or through-transmission mode.
• the transmitting junction point 607a (Fig. 6) is disconnected from receiving junction point 607b.
• Receiving junction 607b is then connected instead to the input jack of the P/R 607, so that the changes in time-separated signals at times / CD and / AB are observable on the oscilloscope.
• the electronic excitation source 607e drives one or two transducers 607t, 607t' according to whether switch 605sw is open or closed, respectively.
• the acoustic central rays, drawn as centerlines through each transducer and continuing through the wall 607p of the pipe propagate as refracted rays 604a, 604b, preferably parallel.
• the bluff body or strut 601s may be similar to struts 602a or 602b. Alternatively it may be of a shape which is symmetrical about a plane transverse to the pipe whose wall is denoted 607p, or is otherwise shaped so that it sheds whether the flow direction is into the transverse plane, or out of the plane. By placing acoustic sensors fore and aft of such a bidirectional shedder, a bidirectional vortex shedding flowmeter is obtained.
• the present invention also contemplates a strut that contains two different cross-sections.
• the first cross-section is bluff with respect to flow from a first direction
• the second cross-section is bluff with respect to flow from a second direction.
• These cross sections may be positioned in alternation at intervals of .2 - .5 D along the diameter dimension of the strut.
• the ultrasonic beams are directed to interact unambiguously with wakes from one or the other of these cross-sections.
• the invention also contemplates a strut which sheds at a first well-defined edge for flow from a first direction, and sheds from a second well-defined edge for flow from a second direction. These edges may be the top and bottom edges of the strut, which in this case may have a uniform cross- section along the entire diametral dimension of the strut.
• the two flow directions are typically forward and reverse flow directions.
• FIG. 7 illustrates a system using a perpendicular (crossflow V ⁇ ) path to measure transit time / and fluid sound speed 3, and using an oblique path, inclined at ⁇ 3 in the direction of flow to measure the incremental contribution to sound speed (Fsin ⁇ 3) caused by the flow velocity V.
• Heavy arrows indicate the direction of interrogation perpendicular to or oblique to flow. If flow direction reverses, the direction of interrogation along the oblique path would be reversed.
• This two-path or differential path method avoids the need to transmit against the direction of flow, and thereby avoids the higher attenuation aj j associated with that direction as V approaches or exceeds Mach 0.1.
• Dashed locations V ⁇ and V x " represent alternative positions for crossflow transducers, that put the crossflow path closer to the oblique (hypotenuse) path.
• lines with arrows at both ends indicate longitudinal (compressional) or shear (transverse) particle motions in buffers. In these buffers, the sound speed near the pipe is eg and c , respectively.
• Open arrow labeled V symbolizes flow in the axial direction, left to right.
• the inclined path length in the fluid P D/COS ⁇ 3.
• the refracted angle, ⁇ 3 is calculable from Snell's Law and knowledge of the geometry and c , the shear- wave velocity in the wedge portion adjacent the pipe.
• V (3000 m/s) - 1/4.33 tf, with the units of tf being seconds.
• a wetted version of the nondispersive directive shear- wave buffer can be combined with a compressional wave transducer to measure transit times along two different paths, neither of which is directed against the flow, as shown in Fig. 7A.
• the launch point P c is the same for perpendicular and oblique rays.
• the triple midradius path in a plane perpendicular to the axis could also be used to measure C3 perpendicular to the axis, and similarly be combined with a one-way with-the-wind oblique measurement to determine flow velocity V.
• Fig. 8 shows a time diagram for a duty cycle where interrogation time spent against the wind is greater than with the wind, and where that duty cycle is therefore asymmetrical and responsive to the SNR (signal to noise ratio) to enhance the measured signal. Weighting can be responsive to signal amplitude, or be set to yield approximately equal numbers of successful transmissions in each direction per unit time.
• Fig. 8A illustrates another novel construction wherein two identical HS buffers 801a, 801b differ in that one is clamped to the outside of the pipe at the six o'clock position, while the other is "wetted" at twelve o'clock and passes through an oblong hole to radiate directly into steam at the ID.
• the wetted guide is acoustically isolated by an attenuative seal.
• two clamp-on waveguides each identical to guide 801b result in an effective clamp-on steam flowmeter.
• annular gaps and recesses from the ID are drawn exaggerated for clarity.
• the bundle construction described therein can be modified to operate marginally dispersively or slightly dispersively as described in connection with Figs. 4A and 4D, namely, such that a chirp pulse is produced having a change in period (or frequency) of a factor of two in a few cycles, e.g., five cycles.
• This is done by forming the bundle out of relatively large-diameter rods.
• a bundle of some 30 rods each of diameter 3 mm can be fitted within a tube of 25.4 mm (1 ") OD, depending on the ID of the tube.
• This dispersive chirping bundle has an interesting application in measuring the sound speed and/or flow velocity of steam or other gas which at times flows rapidly but at other times flows very slowly.
• the leading cycles are of low frequency and hence are likely to be preserved despite beam drift and despite scattering by turbulence, even if the higher frequency cycles do not survive.
• the time resolution afforded by the low frequency cycles is adequate when flow is high. At low flow that low frequency time resolution may be inadequate, but the later cycles which are of high frequency content readily survive the journey across the less-turbulent, lower-flow-velocity lower-Mach-number gas (or fluid) stream and afford the necessary higher resolution of transit time.
• the core width X2 having a numerical value typically around 25 or 30 mm, readily accommodates the chirp without introducing interference problems in that early part of the pulse used for timing.
• Still another feature of this propagation-produced chirp is that each of the leading cycles is progressively narrower in time, according to the dispersive character of the chirp-producing portion of the waveguide buffer.
• This provides a relatively easy way to avoid the classic problem of cycle skipping, when trying to correlate upstream and downstream transmissions, each of which may suffer different attenuations and waveform distortions.
• we have an anti-skip feature With the chirp, one can be quite confident that a particular cycle is indeed the third cycle, for example, and not the second or fourth cycle, by virtue of its period.
• By adjusting the rod length and rod diameter to wavelength ratio one can predetermine just how much dispersion will be obtained, i.e., how much the cycle-to-cycle variation will be.
• the period of each cycle of interest preferably differs from its neighbors by five or ten percent.
• the difference is controllable by the method shown in U.S. Patent 5,139,838, by selecting the operating point along the curve of figure 1 of that '838 patent.
• the crystal is mounted or energized to produce a pulse having a low-frequency leading edge and contains distinguishing characteristics that are preserved as they are needed. Note that the change in period occurs in a monotonically- decreasing manner, as a consequence of the dispersion relation depicted in figure 1 of the '838 patent, starting at the leftmost or thin rod portion of the function graphed therein.
• the low frequency leading portion provides a coarse measure of transit time, refinable by using the subsequent higher-frequency components for higher resolution of transit time.
• the propagation characteristics of the fluid, and the electronics utilized will determine how much of the chirped pulse is usable. The greater the portion of the spectrum that survives, the better is the resolution.
• the chirp features described here have been demonstrated using a commercially available 500-kHz broadband NDT contact transducer and titanium rods, confirming that implementation is practical using devices and materials manufacturable to commercial tolerances.
• the shear wave crystal or piezoelement is not perfectly aligned with particle motion parallel to the major surfaces of the buffer, there will be some spurious unwanted generation of Lamb waves.
• the frequency-thickness product be such that the A 0 wave is dispersive, or that multimodes are present such as A T , A2 , A3 etc.
• a limitation to this approach is that one does not want the thickness to be so thick that multipaths become a nuisance as they were in prior art devices.
• Graphs of the phase velocity of Lamb waves vs the frequency -thickness product for steel reveal that the A 0 mode is strongly dispersive below about 1 MHz-mm, while the next asymmetric plate mode, AT , is strongly dispersive between about 1.5 and 2.5 MHz-mm. While there appear to be small regions of frequency - thickness products (e.g. 1 to 1.5 MHz-mm) having little dispersion for the A 0 mode, in fact it would be rather difficult to operate with broadband pulses in such regions and not excite symmetrical modes (e.g. the S 0 mode) which are dispersive in those same regions.
• symmetrical modes e.g. the S 0 mode
• the minimum thickness for clamp-on is limited in a practical sense by factors such as buckling under pressure coupling; mismatch between a very thin blade and a thick- walled pipe; or beam spread in a thick-walled pipe which would occur from a line source. Buckling is prevented by limiting the pressure available for coupling, and/or by stiffening, for example, by making the extension or wedge in the form of an I-beam.
• a very thin blade extracts only a part of the energy available from a piezoelement whose width may typically be on the order of at least 1 mm up to perhaps 10 mm.
• GC Global System-Chebyshev
• Spoolpieces are typically constructed by welding standard flanges to both ends of a section of pipe.
• the GC spoolpiece of Fig. 3B shows 10-inch 150# weld neck flanges butt welded to the ends of a 16-inch long section of so-called 10-inch schedule 40 pipe.
• the OD of that flange is 16 inches, compared to the 10.75-inch OD for the pipe.
• the annular difference in this example is 2.63 inches.
• Figs. 3A and 3B it is not always possible to satisfy the constraint X () L ⁇ FPAD, for a one-piece solid extension.
• One remedy is to hinge the extension as indicated by the hinges H in Fig. 2F and arrange for a shear wave coupling to be introduced at the butted joint, e.g., pressure coupling.
• An alternate remedy is to make the extension in two sections, one short enough to satisfy the just-mentioned inequality where the short extension is interpreted as X ⁇ L-
• the means the transducer is removable, with part of the extension being removable at the same time.
• the joint may be vee or saw-toothed when viewed from a major face. Vee grooves at 45 deg and having a pitch on the order of one mm are appropriate. One large vee also suffices. Coupling can utilize known liquid couplants (according to Arenberg, 1948, pp. 20 and 22 or other known references cited in Applicant's 1989 book, Ultrasonic Measurements for Process Control, Academic Press, section 3.3).
• Detailed calculations lead to graphs such as those in Lynnworth and Chen, pp. 575-578, in the IEEE Proc. Ultrasonics Symposium, 1975.
• Fig. 8A One spoolpiece that utilizes the transducers of this invention, and specifically designed for measuring steam flow, is shown in Fig. 8A.
• the special mounting gasket shown within the dashed circle 801c in that Figure allows pipe penetration to directly launch and receive signals in the steam, but provides a secure seal and a high degree of acoustic isolation against pipe-borne short circuit noise and ringing.
• the short circuit is prevented by acoustic isolation between the transducers, or the transducer and pipe, for example by installing soft attenuating gasket material or O-rings under flanged sections, as described in applicant's US patent 5,515,733, issued May 14,1996.
• the weight of each pulse utilized can be its SNR.
• a more discriminatory weighting formula is to assign weights according to a power of the SNR, say the SNR squared. This means a very strong, high SNR pulse strongly influences the average and for that reason, if strong enough, immediately can be taken as the probably valid reading. In that case the processor can immediately produce a final reading, i.e. not take further interrogations in that direction or attempt to develop a measurement within the standard deviation requirement. The disregard for a large number of interrogations based on their poor "quality" (low SNR) is part of this SNR- discriminating method.
• this weighting method as Darwinian SNR weighting, in the sense that only the strongest signal survives to influence the subsequent calculation of flow.
• An analogy from optics might be, when driving in a fog, the driver sets and corrects his/her course based on visual clues picked up during moments of relative transparency.
• the contrapropagation method can be modified to measure along a diameter path perpendicular to the axis, a first transit time, from which the sound speed c is determined. Then, along the usual inclined or tilted diameter path, but only in the direction of flow, the velocity of which is V, a second transit time is measured which is a measure of the fluid path P divided by c + V cos ⁇ where ⁇ is the angle of the path relative to the axis. The is then calculated by an equation of the form
• K Kc ⁇ A t/L
• L axial projection of P in the flowing fluid
• K is a meter factor that is intended to take flow profile into account.
• K is not a constant, in general, and depends on the pipe roughness, Reynolds number Re and whether the flow is steady or unsteady. Uncertainties in K in fact are probably the major motivating factor behind the use of multipath flowcells using midradius chords or the GC chords having axial components.
• One system of the present invention further enhances flow detection by adding tag cross correlation, and by detecting and analyzing noise due to the chaotic turbulent medium, and/or noise due to particles such as loose scale carried down the pipe.
• the general implementation of a four-mode flowmeter with these four modes is disclosed in the commonly-owned U.S. 4,787,252 (1988) of Jacobson et al.
• the waveguides of the present invention offer particularly advantageous transducer mountings for implementing such a system, and, as applied to fluids such as steam, certain advantages are achieved as follows.
• the drifted beam can be picked up by axially-displaced transducers, which advantageously may be buffered by the two methods mentioned above, namely, the hockey stick waveguides for oblique incidence and detection, and bundled assemblies of U.S. Serial No. 477,761 for signal paths at normal incidence.
• the detected signals are analyzed by the tag cross correlation method, and under these conditions, the high velocity of steam carries the eddies or other characteristic signature of the random medium several diameters downstream before the signature decorrelates and becomes illegible. Jitter may be intentionally increased by using high frequencies and/or by allowing the beam to traverse the steam two or more times before being received.
• various acoustic short circuit or crosstalk noise contributions may be reduced by any of several techniques, the selection of a technique depending on the measurement mode.
• a differential mode technique may be used to cancel stationary noise; averaging may be used to reduce random noise; and quadrature may be employed to reduce crosstalk in the tag cross correlation measurement according to a method of Jacobson et al. published in Ultrasonics (May,1985) pp. 128-132.
• crosstalk pickups straddling the receiving transducer are averaged, weighted and phase shifted to yield a signal very similar to the crosstalk between them and which appears intertwined or combined with the signal at the middle receiving transducer.
• Another useful technique is to enhance measurements by chaos analysis according to the method of Wayne Hill of the Foster-Miller company, described in U.S. patent 5,600,073.
• the use of the waveguides described above enables well defined signals to be launched and reliably received in the steam conduit.
• the shear wave velocity in some plastics is within a factor of 2 to 4 of c in the gas.
• the shear wave velocity in some common plastics is: acrylic, 1100 m/s; nylon 6-6, 1070 m/s; (low-density) polyethylene, 540 m/s; teflon, 501 m/s.
• the refracted angle approaches the angle of incidence, typically 60 deg for most of our examples herein.
• the refracted angle would be -23 deg, if the incident angle were 60 deg. This yields ray paths resembling those for clamp-on with steel and water.
• ⁇ is 0.4.
• ⁇ g is -36 deg.
• the assembly can be bonded or molded to a short section of plastic pipe, or implemented in a molded monolithic version. Further illustrations of methods of achieving removability appear in Fig. 4D.
• the tube is replaced by a thin solid rod, which may be magnetostrictive.
• the actuating solenoid for the magnetostrictive rod is preferably removable.
• the hockey stick buffer extends such early concepts to new uses in high temperature fluid measurement systems.
• the invention also contemplates improvements in the aforesaid tag cross correlation method.
• a jitter signal as the difference between two successive readings over a given path, in the same direction.
• the path may be inclined and the direction of interrogation may be inclined in the direction of flow.
• the stationary coherent noise due to leakage around the pipe shown in a phasor diagram in the aforesaid Jacobson et al. 1995 article
• the jitter then represents the nonstationary effects of turbulence and scatterers.
• the tag cross correlation method can then be executed using jitter observed at the two axially-displaced locations, such as hockey stick locations denoted UU and DD in Fig.
• the axial displacement, or axial distance may be on the order of the pipe diameter, e.g., 1/2 to 2 pipe diameters.
• This tag method with suppression of stationary noise may be particularly well suited to measuring steam flow by clamp-on assemblies (FIG. 8 A), where flow velocities of a compressible fluid often exceed Mach 0.1, and make it very difficult to transmit reliably against the direction of flow.
• This method is advantageous because it uses interrogation paths having no component against the direction of flow. The drift due to the high Mach number is common to both paths and so drift does not degrade the cross-correlation. At very low flow velocities the very same inclined paths can be bidirectionally interrogated to yield the flow velocity with a contrapropagation geometry.
• the thin radiator constituting the exit face of the waveguide also offers new opportunities to delineate measurements essentially in narrow planes of measurement, and this in turn provides a way to explore improving the accuracy of the beam drift amplitude-based flow measuring method, which has now been practically abandoned because of its general inaccuracy compared to transit time measurements.
• strut When the strut is cast, welded or otherwise formed as an integral part of the meter body, and also including the transducer wedge and extension assemblies, we have another version of a monolithic flowmeter flowcell.
• One application of the above described measurements is to provide a flow meter having a neural network method programmed to quickly detect the signal quality in each of a number of modes, and select the best use of the available multi-interrogation modes based, for example, on SNR and/or standard deviation (statistical) criteria, so that the most likely correct answer is provided by the flowmeter.
• the special case of a fluid at high temperature that can support 0.5, 1 or 2 MHz ultrasonic waves, that normally is conveyed at high pressure, such as steam at 250 deg C and a pressure of some 30 bar, or -450 psig may be more amenable to measurement.
• the fluid instead of some three orders of magnitude below water density, the fluid is only 1.5 orders of magnitude down on the density scale.
• the speed of sound in hot steam, @250 deg C is some 50% faster than in ordinary air.
• refracted angles, while still small, will be -50% greater than in air, for the same incident conditions.
• a means for canceling a large percentage of acoustic short circuit noise, or crosstalk In order for a clamp-on transducer system to work in pressurized steam, it will be desirable to have at our disposal, a means for canceling a large percentage of acoustic short circuit noise, or crosstalk.
• the two-path (X-configuration) of Fig. 2C (or that of Fig. 6 but with the four clamp-on U and D oblique transducers repositioned into the x- configuration) is thus appropriate, operated in a differential mode.
• the two pipe-borne metal paths (which shall be denoted AB and CD) are nearly identical, and to the extent that they are identical, subtracting the signals at the two receiving transducers tends to cancel crosstalk.
• the residual signal is the difference between upstream and downstream propagations in the steam.
• the amplitude of that difference signal is related to the steam flow velocity. If the steam-borne signals are sinusoidal, the phase difference increases with flow velocity and the difference signal will likewise increase with flow, not quite linearly, but in a predictable manner approximately according to the trigonometric identity:
• ⁇ here as the phase shift due to flow, in each direction of interrogation.
• the other angle, ⁇ depends on the transit time across the steam path that would occur if flow were zero.
• ⁇ depends on the transit time across the steam path that would occur if flow were zero.
• c and V in transit time fiowmeters. If flow velocity is small enough, this method applies, but it would fail for large flows or changing sound speed c.
• FIG. 6 A better differential path method that still relies on amplitude measurement and still cancels common mode crosstalk, and common mode flow noise, appears in Fig. 6.
• This configuration responds to flow as sensed in planes symmetrically displaced very slightly about a plane perpendicular to the axis, and utilizes the physically-narrow yet directive aperture of each buffer so that a small drift produces a relatively large change in the differential signal amplitude.
• the paths are not across the diameter [and thereby differ from Petermann, US patent 2,874,568 (1959)] but actually across chords, typically many chords that all generate about the same transit time. A sector of the flow is thus compared at each receiving transducer, and the difference provides a measure of the flow velocity.
• the DA differential amplifier
• the DA can be replaced by a summing device, in the simplest case consisting of a joint between the wires coming off the piezoelements, one flipped 180 deg relative to the other.
• the 8A is constructed with a flange larger than the slot in the pipe or in its retaining member.
• acoustic isolation is achieved as in applicant's US patent 5,515,733 (May 14, 1996), and transit time is measured alternately upstream and downstream.
• Sound speed c may be measured separately across a diameter path using a bundle buffer similarly isolated, said bundle buffer being constructed, for example, according to the construction shown in co- pending US patent application serial number 08/477,761 filed June 7, 1995.
• the diameter path buffer may be a marginally dispersive waveguide of the type shown in applicant's US Patent 5,159,838 issued Nov. 3, 1992.
• the larger flange prevents expulsion of the buffer.
• the other may be welded into the spoolpiece.
• high repetition frequencies are required. This implies damping of the buffer. This can be most easily done near the cooler end, under or within the cap that encloses the piezoelement, usually at a temperature under 100 deg C.
• the Reynolds number Re depends on gas density
• the turbulent-flow meter factor K for a tilted diameter path varies 1%> per order of magnitude change in Re, it often suffices, with respect to estimating K, to determine the density only approximately.
• the strength of the received signal can be taken as a measure of gas density.
• inboard and outboard GC paths may be utilized in this regard, or first and triple transits, if echoes are detectable for one or multiple diameter traverses.
• the buffer as described in connection with Fig. 2E is useful with respect to measuring secondary flow components such as swirl and circulation in a plane perpendicular to the conduit axis.
• secondary flow components such as swirl and circulation in a plane perpendicular to the conduit axis.
• transmission of ultrasound in a plane perpendicular to the axis can be interpreted in terms of liquid present or absent (i.e., liquid level at a discrete point), and viscosity (based on attenuation of the liquid-borne signal).
• the fluid medium liquid, high-pressure gas, two-phase medium
• the gas signal strength compared to the conduit noise level can yield information on gas pressure.
• Another application is measuring the frequency of vortices shed by a bluff body or other shedder in the stream. In the past, this has been done using one or a pair of transducers just downstream of the shedder, as exemplified by Joy and Colton, U.S. Patent No. 3,680,375 (1972). However, the transducers apparently had to be wetted, especially if the fluid is a gas such as ordinary air, or else the receiver would be overwhelmed by crosstalk.
• Fig. 6 shows several forms of differential measurement where the conduit noise is largely canceled. The small difference between the upstream and downstream
• the strut would be welded into the pipe, either at the factory or in the field via a hot tap procedure.
• the transducers would preferably be totally outside the conduit pressure boundary, i.e., clamped on. This avoids the usual electrical feedthroughs from pressure or temperature sensors that are ordinarily used to sense/
• both inclined paths are interrogated in the direction of flow, rather than against the flow.
• the two inclined paths may be interrogated simultaneously (the preferred way, if practical) or else interrogated one at a time in fairly close succession, - 1 to 10 ms apart.
• the prior vortex shedder ultrasonic flowmeter art includes plastic bodies containing a cast-in bluff body and external diametrically opposed ultrasonic transducers, both of which are downstream of the shedder.
• plastic bodies containing a cast-in bluff body and external diametrically opposed ultrasonic transducers, both of which are downstream of the shedder.
• Tokico Ltd of Kawasaki City, Japan Model GS-F3030E ultrasonic vortex flowmeter, PEEK series, made of poly ether ether ketone resin.
• Additional prior art includes H. S.
• FIG. 6 A more symmetrical and therefore better way to compare upstream vs downstream circulation about the strut 601 s, is depicted in the bottom right-most portion of Fig. 6.
• upstream and downstream measurements are each executed in planes perpendicular to the axis.
• the exit faces 603a,b are contoured to match the pipe curvature.
• the cw and ccw interrogations (604a & b, respectively) at upstream ( ⁇ U) and downstream ( ⁇ D) locations tend to cancel any swirl effects.
• the switch 605sw when closed, causes the cw and ccw interrogations to be launched simultaneously.
• the differential mode then more completely removes common mode noise and yields a difference signal indicative of the vortex shedding frequency.
• the contoured clamp-on transducers can be rotated circumferentially with respect to the strut (bluff tee) to determine the best angle of interaction with the shed vortices, from 0 to 90 degrees.
• the receive window can be preset to toggle between the two expected ranges of inboard and outboard transit times as each anticipated resonance is utilized. Or the window can be wide, covering all paths, and the frequency selection (resonance) of each pair of transducers can be relied upon to separate information for each path in succession.
• the approximately 6-mm width of the hockey stick transducers (all of identical shape, except for slightly dissimilar wedge thinning to regulate the depth of insertion into the oblong GC slots) satisfies the broadband propagation requirements for 0.5 to 4 MHz.
• This system may also be implemented by using a set of clamp-on transducer assemblies which are identical in all respects except for the frequencies of their piezo crystals.
• the weighted average temperature can be determined by measuring local temperature Tat the several tabs TCI, TC2, ...illustrated in Fig. 1 and 1A. If the gradient is not too severe one can simply take the linearly weighted average
• Plastics will have a greater fractional coefficient of sound speed with temperature, (l/c)dc/dT.
• Ceramics and refractory alloys will have a lesser (l/c)dc/dT, near zero in some cases, e.g., fused quartz.
• Graphite also has a near-zero (l/c)dc/dT.
• Fig. 2F Let us first consider as the objective, interrogation along a spiraling triple midradius path.
• Transducer sites are generally in the axially displaced positions marked Z ] and Z - Because of limitations on refraction imposed by Snell's Law, that path is unreachable by clamp-on if the pipe is thick and of substantially greater sound speed than the liquid. For example, the shear wave velocity in steel is more than twice the longitudinal wave speed in water, and refracted waves at 30 deg are not practical.
• the hybrid solution to this problem is depicted in Fig. 2F, with introduction of a plug having a shear wave sound speed C2 nearly equal or identical to c j ,the sound speed in the liquid.
• Some forms of graphite, phenolic and tin, have shear wave velocities near 1500 m/s.
• the longitudinal wave radiating into the liquid is indicated by the arrow to be along the v-axis, pointing downward, in this end view. After three midradius helical traverses the ray would approach the pipe wall at 30 deg to the x- axis.
• a plug and removable member again solve the refraction problem, but they are installed rotated 60 deg clockwise relative to the first orientation.
• the two transducers on the left side are mounted by hinge asssemblies H to oriented buffer/waveguides on the spoolpiece.
• Each hinge H includes a removable hinge pin lOhp so that the transducer/hinge plate outer member can be completely separated and removed.
• sealing gaskets may be provided to seal around the mating surfaces to preserve their cleanliness. Screws hold the parts together and provide coupling pressure for effective transmission of shear waves.
• the crosstalk isolation method using silicone O- rings may be advantageously adapted to the rectangular cross-section of the present thin waveguides, installed through flanged nozzles.
• the sheet metal GC spoolpiece made from the stamping shown in Fig. 3C is another way to achieve accuracy at minimal cost. Applicant expects that as data are gathered and analyzed, along with advances in CFD (computational fluid dynamics), it may become practical to use just tilted diameter paths and compute or look up the meter factor K as a function of nonideal inlet conditions. This would allow further cost reduction, as off-diameter paths for profile compensation would become unnecessary and secondary flow components would not need to be measured directly.
• One way to acquire such data would be to use spoolpieces similar to those illustrated, wherein tilted diameter and other paths are included. Again referring to the sheet metal stamping of Fig.
• diameter paths are to be obtained by installing transducers (hockey sticks) through ports X, XX, X', XX', and off-diameter paths are to be obtained using the other oblong ports A, AA, ...D, DD.
• Reflectors installed within slots R subscripted according to the transducer ports A...XX' are also to be used in conjunction with their respective transducer pairs.
• FIG. 4A and 4B A particularly low-cost method of making one monolithic version of this invention is shown in Fig. 4A and 4B.
• the extensional portion provides for easy removability of the actual electroacoustic transducer.
• the excitation coil may be a removably mounted coil.
• a Rayleigh buffer may be envisioned as an extended source of waves at the Rayleigh velocity C ⁇ .
• the Rayleigh buffer does not yet appear to be as generally useful for industrial systems and process measurements as the shear buffer waveguide described herein, but there are occasional situations where it may be advantageously combined with the shear buffer.
• the Rayleigh buffer wraps more easily around the circumference of a pipe, than does the shear buffer. It can also be used an extended source in the axial direction, making line or arc contact. As an extended source or extended receiver, it might replace several separate hockey sticks for applications where the exact arrival point is unpredictable.
• the Rayleigh buffer must generally be made of a material different from that of the adjacent pipe, so that its velocity would exceed that of shear waves in the pipe wall by some 5 to 15%, to couple efficiently to shear waves at useful refracted angles.
• SS Rayleigh buffers could work with brass pipe; fused silica buffers could work with SS pipe.
• could work is meant, the necessary velocity ratios are achievable with such materials combinations. Applicant has made preliminary tests with these particular combinations at frequencies around 1 MHz.
• the modulations imposed on ultrasonic pulses occur in time (or phase) and amplitude.
• a corollary apparently not recognized nor used previously, is that if the signals transmitted across turbulent flow vary from instant to instant, there must be certain times at which the signals propagate with less modulation, i.e., less attenuation caused by turbulence, than at other times.
• a special case is when the turbulent eddies of interest are specifically those created intentionally by a bluff body, and where the modulation occurs at a frequency proportional to the flow velocity. This too has a corollary, and again, apparently not recognized nor used previously. Between the vortex sheets there are lulls when the ultrasonic signal can get across the path with little or no rotation, scattering or, in general, with little attenuating modulation effect.
• ultrasonic flowmeters have a performance characteristic where if the error magnitude were plotted along the y axis, and the square root of time were plotted along the x axis, the performance would be approximated by a hyperbola in the first quadrant, asymptotic to x and v axes. Roughly analogous to the Heisenberg Uncertainty Principle, if we specify time too short, the uncertainty in flow must be large. Conversely, if we require exceedingly small error, a very long time is required to obtain it.
• the present concept of varying the measurement or interrogation time spent in each of two different directions with respect to the direction of flow, as a function of the SNR, is an attempt to optimize the joint product of accuracy and response time with respect to user requirements.
• the signal amplitude S may be taken as the amplitude of the strongest signal within a predetermined time window, provided certain other requirements are met (e.g., frequency, signature).
• the noise may be taken as the waveform, usually its amplitude N, at a point or region just prior to the onset of the signal.
• the "onset" of the signal (t onse t) m turn may be taken as the point in time at which the signal's envelope, extrapolated backward in time, intersects the time axis.
• the "envelope” may be piecewise linear, Gaussian or other common shape associated with analog- or digital-filtered signals.
• the "noise region" just prior to the signal may be taken as an interval containing one to ten or more cycles at the center frequency of the pulse. That interval will preferably be adjusted according to pulse shape and other conditions imposed by the application. It may be advantageous to set that interval equal to the width of the transmitted pulse. Another convenient choice is to set the noise- sensing interval equal to the time it takes for the received pulse to build up to its maximum value. Determining tj j & tj by the MRI/S [Maximum Reliable (SNR-Weighted) Information per Second] Method
• U is taken as a measure of the uncertainty in the measurements.
• Method may control or adjust transmitted power, frequency, spectrum, coding or modulation method, and receiver processing modes associated with such transmissions, as a function of observed average or peak SNR, signal strength S, ratio of number of acceptable received signals NQJ ⁇ XO the number of attempts N TR ⁇ , that is, N 0 ⁇ IN TRY, in upstream (U) and downstream (D) directions.
• Kp may equal 0.5 initially but be adjusted as a function of Sp ) Sp Sj j p j j , or as a function of SNRp /SNR jj .
• K j) can be set to the value SNRj /SNRf . Report /£>, ⁇ p, ⁇ SNRfj>, PSD (power spectral density) information, and/or other significant statistical information.
• the value for may be made responsive to changing propagation conditions by setting it equal to the average of, say, the three highest SNR's in the three most recent measuring intervals.
• the U-D process may proceed in symmetrical or random fashion until all paths and directions have been interrogated. Then the process repeats.
• scattering depends in general on the ratio of the scatterer cross-section to the ultrasonic wavelength, and also on the direction of incident and scattered beams. If we represent the size of the scatterer by ⁇ x and the ultrasonic wavelength by ⁇ , scattering will depend on some power of their ratio ⁇ x / ⁇ .
• the effective wavelength is shorter into the wind by 10%), and longer with the wind by 10%o.
• the difference is 20%. It seems plausible that scattering into the wind vs with the wind will be different, higher into the wind, because the shorter wavelength is scattered more, just like a 20% higher frequency is scattered more in the polycrystalline solid. (12 MHz vs 10 MHz in Babikov' s numerical example for steel, with the absorption being about 1 dB per cm, or 100 dB/m for normally annealed steel, for these two frequencies.)
• the scatterer size in turbulent fluids is the scale of turbulence. The shape of the scatterer, however, may be more important than the difference in wavelengths in the two directions.
• the code may have to be modified to reduce bandwidth, especially if one strives to correlate U and D interrogation signals. This poses a dilemma, because a broadband code containing lots of information is ordinarily the preferred way to distinguish a signal buried in noise.
• One compromise solution may be to use a narrow band code transmitted twice, at two close but different carrier frequencies to avoid cycle skip ambiguity to which narrow band signals are more susceptible.
• Integer relations may take advantage of piezoelectric transducer resonances, and/or provide opportunities to properly interpret arrival times for under sampled waves.
• the attenuation would be influenced largely, although not entirely, by the wavelength corresponding to the carrier frequency.
• the bandwidth also influences the attenuation and pulse distortion, due to the frequency dependence of attenuation by scattering or other attenuative processes.
• the waveguides of the present invention are well adapted by their signal propagation and coupling properties to both active lauching of ultrasonic signals, and to reception of lauched signals or even passive monitoring of intrinsic noise in a fluid system.
• an individual waveguide may perform more than one function, and indeed will commonly operate both as a sending and a receiving transducer, and may be paired with one or more other such transducers to within its beamspan.

Landscapes

• Physics & Mathematics (AREA)
• Engineering & Computer Science (AREA)
• Acoustics & Sound (AREA)
• Multimedia (AREA)
• Electromagnetism (AREA)
• Fluid Mechanics (AREA)
• General Physics & Mathematics (AREA)
• Measuring Volume Flow (AREA)

Priority Applications (3)

Applications Claiming Priority (4)

Publications (1)

Family

ID=26704777

Family Applications (1)

Country Status (5)

Cited By (15)

Families Citing this family (130)

Citations (3)

Family Cites Families (11)

• 1997
  • 1997-06-20 US US08/879,690 patent/US6047602A/en not_active Expired - Lifetime
  • 1997-10-27 WO PCT/US1997/019238 patent/WO1998019296A1/en not_active Ceased
  • 1997-10-27 DE DE69732300T patent/DE69732300T2/de not_active Expired - Lifetime
  • 1997-10-27 EP EP97912897A patent/EP0935798B1/en not_active Expired - Lifetime
  • 1997-10-27 JP JP52059898A patent/JP4107685B2/ja not_active Expired - Fee Related

Patent Citations (3)

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events