FIELD OF THE INVENTION

This invention relates to monitoring parameters of solid particles in multiphase flows, such as the particle size and size distribution of larger particles, in turbulent process flows.

BACKGROUND TO THE INVENTION

When processing certain types of process gas or process fluid carrying solid particles, there is a need for measuring the mass or size of particles flowing in such a process fluid. Examples of such flows are found in pneumatic conveying of granules as well as mineral slurry transfers by pipes or launders.

Where the particles’ solid densities are in a limited range, particle mass is sometimes used as a proxy for particle size. In one particular example, e.g., in the mining mineral process industry, particle diameter size and mass are thus closely related. For example, for two particles with the same shape but with diameters differing with a factor of 2, their volume would differ by a factor of 8, much more than the typical mineral density variation. Therefore, particle mass is commonly used as a useful proxy for particle size in the mining industry. This can be seen where classification equipment that depends on particle mass, like cyclones, are used in the place of screens, which would have been more accurate size classifiers.

Faults in either type of such classification equipment may lead to oversize particles being channelled to a process stream designed for fines, or vice versa, undersize particles being channelled to process stream designed for coarse particles. These faults cause process problems and financial loss downstream when the wrong sized particles end up in the wrong stream. As an example, it may cause financial loss if a valuable size fraction is discarded to a tailing disposal due to a classification fault. The problem can be addressed if quick measurements of particle size or mass is available. There is a need for a solution to this problem in the art. RELATED ART

Current methods for particle size or particle mass measurement include: gravitational sedimentation methods; microscope image analysis methods; stream scanning methods including electrical sensing zone, light blockage and light scattering; field scanning methods like light diffraction; speed of sound and ultrasonic attenuation; photon correlation spectroscopy methods; sieving methods; elutriation methods; and physical/mechanical/acoustic impact/impingement/collision methods.

The last group of physical impact methods proved most practical in the past for realtime in-line measurements as they can be used on the full density process stream as- is, without taking samples or dilution of the stream. Prior art methods require an impact surface that is interposed at a normal orientation, or least partly normal, relative to an incoming flow direction from which particles approach the impact point. Therefore, the prior art requires: (1) that a sensor’s impact surface extends or protrudes into the process flow, (2) that the impact surface should be on an inside wall of an elbow or sharp bend in a pipe, or (3) a constriction in a pipe. Disadvantages of these techniques include:

• Impedance of process flow with increased risk of blockages, especially in cases

(1) & (3);

• Segregation, therefore a non-representative measurement, especially for case

(2) due to the forced direction change in slurry flow;

• Additional wear of measuring surfaces in all three cases; and

• Precluding the option of using anti-wear elbows, like blind tees for case (2).

GB 2431991 A and WO 2007/052022 A1 are examples of prior art using impact surfaces on elbows (case (2) as described above) to measure acoustic noise due to particle impacts. The need is expressed to install sensors for particle size measurement with impact surfaces at an angle to predominant flow and if this cannot be done on an angle, it is proposed that an intrusive probe with an angled head can be used (case (1) as described above). The sensors with angled impact surfaces relative to predominant flow can be supplemented with cylindrical sensors that serve to monitor metal loss (erosion and corrosion), and/or to serve as reference of flow conditions for the sensors with angled impact surfaces on elbows or probes, but the cylindrical sensors do not measure particle size or mass.

The present invention seeks to provide quick, preferably real-time measurement of particle size or mass, while avoiding or ameliorating the disadvantages of the prior art, mentioned above.

SUMMARY OF THE INVENTION

According to a first aspect of the present invention there is provided apparatus for monitoring particle size of solid particles in a multiphase stream, said apparatus comprising: a wall defining a wall surface of an elongate flow passage for housing the multiphase stream, said wall surface being parallel to a flow direction of the multiphase stream in the flow passage; and an impact sensor embedded in the wall and exposed to the flow passage, said impact sensor having an impact surface that is flush with the wall surface; wherein the flow passage is straight, at least on an upstream side of the impact sensor.

The term "flow passage" is used herein to denote a space through which a fluid can flow, relative to reference such as an adjacent body, so the includes a flow passage inside a body and also a flow passage along an outside of a body.

The term "flow direction" is used herein to denote the overall direction in which a fluid flows and if the fluid flow is turbulent or if there are inconsistencies in flow direction within a flowing fluid, then its "flow direction" is the average or predominant direction of flow of the whole fluid.

The impact sensor may include an abrasion resistant impact body defining the impact surface and the abrasion resistant impact body may be in the form of an abrasion resistant surface layer, a force sensor disposed immediately adjacent the impact body, and preferably a weight on a side of the force sensor, that is opposite from the impact body. The impact body, force sensor and weight may form a unit with an internal natural frequency that is higher than the inverse of the longest particle contact time that is expected for the solid particles according to the Hertz theory of impact.

The unit comprising the impact body, force sensor and weight may be embedded in a material that is substantially softer and has a lower natural frequency than the unit.

The wall may comprise a thin body housing the impact sensor, said body being attachable to a substrate with a high ratio between a dimension of the body in the flow direction compared to a thickness of the body by which the body protrudes from the substrate into the flow passage.

According to another aspect of the present invention, there is provided a method for monitoring particle size of solid particles in a multiphase process stream using the apparatus described herein above, said method comprising: causing the process steam to flow in the flow passage with turbulence, so that the solid particles in the flow stream impact on the impact surface of the impact sensor; measuring the impact of the solid particles on the impact surface to generate a signal; and calculating particle size of the solid particles from the signal.

According to a further aspect of the present invention, there is provided a method for differentiating between solid particle materials in a multiphase stream using apparatus comprising an impact sensor exposed to the multiphase stream flow, wherein the impact sensor has an internal natural frequency that is higher than the inverse of the longest particle contact time that is expected for the solid particles according to the Hertz theory of impact; by calculating the coefficient of restitution of each particle impact from the ratio between first and second slopes of a measured pulse.

According to a further aspect of the present invention, there is provided a method for determining any three parameters selected from: particle mass, size, density, hardness and velocity of solid particles, in a multiphase stream using apparatus comprising an impact sensor exposed to the multiphase stream flow, wherein the impact sensor has an internal natural frequency that is higher than the inverse of the longest particle contact time that is expected for the solid particles according to the Hertz theory of impact; by, calculated from the width, height and first slope of the measured pulse where the other two parameters are known or of a limited range.

BRIEF DESCRIPTION OF THE DRAWINGS

For a better understanding of the present invention, and to show how it may be put into effect, the invention will now be described by way of non-limiting example, with reference to the accompanying drawings in which:

Figure 1 shows a schematic representation of the position of an impact sensor according to the present invention, in a wall of a flow passage;

Figure 2 shows schematic representation of velocity vectors of a solid particle that impacts the sensor of Figure 1 ;

Figure 3 shows a diagrammatic view of a first embodiment of a sensor according to the present invention, used on a flow passage;

Figure 4 A shows a side view and top view, respectively, of the apparatus of Figure 3; Figure 4B shows a longitudinal sectional view of the apparatus of Figure 4A, taken at B-B;

Figure 4C shows a cross-sectional view of the apparatus of Figure 4A, taken at C-C;

Figure 5 shows a typical signal from a solid particle impact, measured in the apparatus of Figures 4A-4C;

Figure 6 shows a diagrammatic representation of method of using the apparatus of Figures 3 and 4;

Figure 7 shows a typical output of the method of Figure 6;

Figure 8 shows three typical signals obtained from measurements made with the apparatus of Figures 3 and 4;

Figure 9 shows a diagrammatic view of a second embodiment of a sensor according to the present invention, used on a flow passage; Figure 10 shows typical signals that resulted from use of the apparatus of Figure 9; and

Figure 11 shows, a diagrammatic view of a third embodiment of a sensor according to the present invention, used to measure the flow of particles passing it.

DETAILED DESCRIPTION OF THE DRAWINGS

The invention encompasses a technique and apparatus for measuring solid particle size (for which solid particle mass serves as a proxy) and other parameters in a turbulent multiphase flow using an impact sensor surface that is parallel to the average incoming flow direction preceding the impact point, as shown in Figure 1 . In Figure 1 , a multiphase flow medium or process stream 10 flows in a general flow direction 12 in a flow passage that is bound by a wall 14 and an impact sensor 16 is embedded in the wall so that an impact surface 18 of the sensor is generally flush with an inner surface of the wall.

In contrast to prior art, the impact surface 18 does not extend or protrude into the process flow and the flow passage is straight on an upstream side of the sensor 16 and the impact surface would therefore not normally experience impingement or impact from solid particles in the process stream 10 during laminar flow. The process stream 10 could include solid particles in a flow medium of either liquid or gas. The former case may or may not contain gas bubbles; the latter may or may not contain liquid droplets.

Figure 2 shows the skew velocity vector 20 of a typical solid particle in the process stream 10, where the velocity vector is imparted to the solid particle by a turbulent flow of the process stream. This velocity vector 20 can be described as the sum of two orthogonal vectors; a first vector 22 that is parallel to the average flow direction 12 and a second vector 24 that is perpendicular to the average flow direction.

According to the present invention, the impact force of individual solid particles, perpendicular to average/mean flow direction 12, i.e. impact forces resulting from the vector 24, are measured over time. As the particle impacts the sensor, only its perpendicular component is measured.

Where prior art treated turbulence at best as source of random disturbance that hinders measurement, it was found in the present invention that the distribution of particle velocity 24 orthogonal to the flow direction 12 as caused by turbulence and measured as the impact of the individual solid particles on the impact surface 18 parallel to the flow direction, is sufficiently repeatable, structured and systematic to serve as source of valuable information on useful particle parameters such as: mass distribution including top mass, size distribution including top size, density distribution, or hardness distribution of the particles.

A schematic diagram of a first embodiment of measurement apparatus according to the present invention is shown in Figure 3 and is generally identified by reference sign 26, while less schematic views of the apparatus 26 are shown in Figures 4A-4C.

Figure 3 shows upward flow of process stream in the form of a slurry 10 in a pipe. The term "pipe" is not limited in this context to a hollow cylindrical body, but extends to denote any hollow body that defines an elongate flow passage, bounded by walls 14. A sideways opening is provided in the pipe wall 14 with a flange 28 fitted to the pipe. A sensor body 30 that is preferably of stainless steel, is inserted into the flange 28 until its leading end/nose/tip is flush with the inside surface of the pipe wall 14. A metal or ceramic impact body “C”, embedded in abrasion-resistant Rubber “R”, in contact with a piezoelectric force sensor “P” which is backed by a steel weight “S” and with the sensor’s output wires connected to a sensitive analogue pre-amplifier “A”.

Referring to Figure 4, the apparatus 26 is shown without the process stream, but reference sign 10 is also used to identify the flow passage in which the process stream would flow, when the apparatus is in use. The same reference signs are used in Figure 4 as in Figure 3, with the addition of the sensor nose or tip 32, which is the leading part of the apparatus 26 that is installed to be flush with the inner surface of the pipe wall 14. Further, Figure 4 does not show the pre-amplifier of Figure 3, but shows the wires 34 leading from the sensor body 30 to the pre-amplifier. In tests conducted with the apparatus 26 of Figures 3 and 4, it was found that the amplitude and frequency components measured in this manner provides information on the presence, mass distribution and nature of solid particles in the flow or process stream 10, similar to the prior art where impacting surfaces were provided that were not parallel to the flow direction of the process stream approaching the impact point.

Fig 5 shows a typical signal as measured in the apparatus 26 when an individual solid particle in the process stream 10 impacts the impact body C. This signal is similar to signals in the prior art related to particle size measurement by acoustic emission (AE). It shows the original impact followed by a decaying ringing, similar to signals found in the prior art, containing more than one frequency component.

In the present invention, it was found that some of the frequency components do not change between impacts from different solid particles and are characteristic of the sensor’s construction natural oscillation frequencies. It was also found that a further frequency component changes from particle to particle and is probably caused by the impact contact time of the particle, a function of its velocity, size and hardness, as is described in the Hertz theory of impact. In addition, it was found that the square of the signal amplitude is approximately proportional to the kinetic energy of the particle in a direction orthogonal to the general flow direction. It was found that the value of such particle parameter information can be improved by combining the measured amplitude and frequency components with the separately measured instantaneous flow rate of the multiphase flow, all serving together as inputs to a lookup table or machine learning model.

Figure 6 shows an application example of such a technique or method where amplitude of the impact signal of a solid particle in the process stream 10 on the impact body C of the apparatus 26 and flow rate of the process stream are used. This specific example is a solution for determining the top size fraction of solid particles in a slurry flow 10 for the purpose of detecting screen wear or tearing. The analogue Pre-amplifier “A” is followed by a band-pass filter “BPF” to remove DC offset, low-frequency interference and to provide an anti-aliasing filter for the analogue-to-digital converter “ADC”. “Abs” calculates absolute values, followed by a low-pass filter “LPF”.

Fig 7 shows a typical output of ADC as “y1” and its corresponding output of LPF as “y2”. The LPF signal goes through a simple peak-detector function “Peak”. The next function “Hist” gathers the peak values found during a time period of a few seconds in a histogram and cumulatively adds it from the highest to the lowest bin.

Fig 8 shows three typical results to illustrate the effect of top particle mass on the cumulative histogram. “y1” is the signal for a water slurry containing a range of sand particle sizes up to 1.5mm where the sand comprises 26% of the slurry by mass. “y2” is the signal when a further 1 % by mass sand particles of 1 mm to 1.5mm is added. “y3” is the signal when a further 1 % by mass sand particles of 3mm to 3.5mm is added. The graph shows that for a specific slurry flow rate, the particle top size can be determined by the impact intensity where the cumulative counts trends to zero and that this part of the graph is not influenced by an increase in smaller particles.

The last function in Fig 6, “LUT”, is a lookup table that maps maximum impact intensity to particle size as a function of slurry flow rate. An alternative to a lookup table can be a supervised machine learning model.

In general, the most consistent results were obtained when the impact body C was made of a hard material, preferably harder than most of the particles in the stream. Sapphire is a ceramic that worked particularly well as an impact body C, being both hard and tough.

It was further found that the clearest signals could be measured with the least ringing when the impact body C, piezo-electric force sensor P and steel body S were all as small, hard, dense, and close to the impact zone 18, as possible. Figure 9 shows a second embodiment of measurement apparatus 36 according to the present invention, which includes the same components as the first embodiment shown in Figures 3 and 4, except that the components have been arranged to provide better results.

In the apparatus 36 of Figure 9, a process stream in the form of a slurry 10 flows in a pipe with pipe walls 14 with an opening to the side and with a flange 28 fitted to the pipe, a sensor body 30 is inserted into the flange until its nose/tip 32 is flush with the inside surface of the pipe wall 14. A thin sapphire ceramic impact body “C” in contact with a piezoelectric force sensor “P” which is backed by a tungsten carbide weight “W”. All three of these hard components C,P,W are embedded in abrasion-resistant relatively soft rubber “R”, and with the sensor’s output wires connected to a sensitive analogue pre-amplifier “A”.

Fig 10 shows typical signals that resulted from the apparatus 36 or sensor of Figure 9, arranged without a pipe 14 so that the nose/tip 32 had an upward orientation and tested by dropping two sizes of solid particles from two heights onto its sapphire impact body face.

“y1” shows the larger particle dropped from the higher height,

“y2” shows the larger particle dropped from the lower height,

“y3” shows the smaller particle dropped from the higher height and

“y4” shows the smaller particle dropped from the lower height.

If Figure 10 is compared to Figure 5 (bearing the 10x timescale difference in mind), practically no ringing occurs in Figure 10, as the internal natural frequency of the combined impact body C, piezoelectric sensor P and backing weight W was very high by design, plus they are suspended in soft rubber R with a very low natural frequency.

The signal thus achieved in the apparatus 36 of Figure 9 is considerably clearer than in prior art and shows the collision force that the relatively soft particle exerts on the relatively hard sapphire face while it distorts, come to a standstill and departs. Very little evidence of acoustic waves are present, only direct force over the collision contact time, conveying considerably more information than what was achieved in the prior art. The first slope “S1” is proportional to the particle hardness and arrival velocity, the maximum pulse height closely proportional to the square root of its kinetic energy and the pulse width proportional to contact time, a function of particle size, density and hardness. It was found that the pulse width for this configuration is surprisingly insensitive to particle velocity. A lower departure slope “S2” than S1 implies a lower departure velocity, therefore loss of energy, therefore a more inelastic collision with a lower coefficient of restitution and therefore a solid particle of a different material.

Pulse slope S1 , proportional to velocity v and pulse height, proportional to square root of particle energy %mv ² , allows the particle’s mass m to be estimated. Where particle hardness is known or can be assumed to be in a limited range, contact pulse width indicates particle mass. Likewise, where particle density is known or can be assumed to be in a limited range, particle mass m indicates particle size.

Testing of these techniques of the present invention showed that sideways/orthogonal velocities caused by turbulence and as measured with this technique is not as random as expected but shows a sufficiently repeatable, structured and systematic distribution of velocities sideways/orthogonal to the flow direction 12, with a clear limit in maximum velocity for each slurry flow rate and for each particle size. This unexpected result means that a solid particle sensor surface or impact surface 18 can be used that is parallel to a multiphase flow, as long as the flow is turbulent, with the benefits of zero impairment of process flow with no risk of blockages, a representative measurement due to no forced direction change in slurry flow and accompanying segregation, and minimised wear of measuring surface.

Measurements of particle size or mass have been made in the prior art using impact surfaces that with a component of normal orientation relative to the flow direction - with the well-known disadvantages mentioned above. Further, measurements of other characteristics of flowing process streams, such as measurements of corrosion, ultrasonic measurement of flow rates, have been conducted with instruments oriented parallel to the flow directions, but this orientation is not considered in the prior art for particle mass or size measurement, because it is presumed that the impact of solid particles on a surface that is parallel to the flow direction, would be negligible and insufficient to serve as basis for reliable measurement of particle size or mass. The present invention solves the problems caused by impact sensors with orientations that are normal relative to flow directions, in a way that persons skilled in the art would not have considered.

In addition, owing to its flush location and protection in a flow wall 14, no acoustic wave conductor is required and the impact body C, sensor P and backing weight W can be manufactured small and hard. This lifts its natural mechanical frequencies to higher than the inverse of particle contact time and therefore does not interfere with the pure force signal. In addition, the location enables these components to be embedded to “float free” in a relatively large soft rubber volume, which natural frequencies are much lower than the inverse of particle contact time and therefore, again, does not interfere with the pure force signal.

This technique to measure solid particle parameters is not limited to water slurries or pipes. It is only limited by the existence of turbulent flow. It is expected also to work for particles transported by oil and other liquids, as well as air and other gas. It is also expected to work in the floor and sides of launders, chutes, channels and river beds, in the walls of buildings, and outsides of vehicles and aircraft.

Figure 11 shows a third embodiment of apparatus 38 according to the present invention in an example for such an application, where it is not possible to embed a sensor into the wall of a channel or structure but with a requirement to measure the flow of particles passing it. The invention enables the insertion of a flat body housing the sensor with a high ratio between the body's dimension along the direction of the flow 12 versus its dimension by which the body protrudes into the flow, from a substrate, causing minimal flow disruption and with its sensing surface 18 is still parallel to the flow direction 12.