Method and device for detection of particles in flowing media
TECHNICAL FIELD
The described device comprises an instrument for the con¬ tinuous detection of particles in flowing media. The detec¬ tion relates to particles above a certain size or within a certain size interval, the sensitivity increasing with in¬ creased size of the particles. A typical field of appli¬ cation is the measurement of the erosion capacity of par¬ ticles, also called erosiveness, in flue gases for gas tur¬ bines .
THEORY
When parallel beams of monochromatic light fall upon a dust particle, light hitting the particle is deflected. Light is scattered in all directions. A greater part of the scatt¬ ered light is deflected to a very marginal extent from the original forward direction. A screen placed in the path of the light beam, at a relatively large distance beyond the particle, is illuminated strongly in that surface of the circle where the light beam meets the screen. Outside of this surface of the circle, the screen is illuminated slightly by light which has been deflected by the dust particle. This deflected light is expected to decrease continuously in intensity with increasing radial distance from the centre of the beam. Since the light is of mono¬ chromatic nature, however, adjacent deflected beams will influence one another by interference. This results in concentric ring-shaped dark areas on the screen, where light waves extinguish one another. What can then be discerned on the screen are a number of concentric light rings with dark rings in between. The described phenomenon is an example of Fraunhofer diffraction of light . The pattern created is called diffraction pattern. The magnitude of the angles through which the light is deflected is determined by
the size of the particles, which in this case is conside¬ rably greater than the light wavelength. At a larger dis¬ tance, the pattern is large compared with the cross-section area of the original beam.
If a lens is placed in the path of the light after the particle and a screen is mounted in the focal plane of the lens, the non-deflected light will be concentrated at a focal point on the optical axis of the light. The light deflected by a dust particle forms conical surfaces of light which extend from the dust particle and light up the screen with light rings centered around the focus, separated by dark rings.
Movement of the light-deflecting particles does not influence the diffraction pattern, since parallel light beams are always focussed on the axis and a given conical angle of deflection always results in the same radial displacement (s) in the focal plane of the deflected light. For small deflections, the angle of deflection is given by s/f, where f is the focal distance of the lens.
There are different methods for drawing conclusions about particle sizes by recording the light intensity of the different light rings in the diffraction pattern. For recording the intensity of the light, suitable types of light-sensitive elements are used.
BACKGROUND ART
The state of the art for measuring the size of dust particles is based, inter alia, on the use of a laser beam which is caused to pass the region where a study of the occurrence of particles and particle size is to be carried out. With knowledge of the property of small particles to scatter light and form a diffraction pattern, a plate with photo-sensitive diodes is placed in the focal plane of a lens which collects the light to the plate. Fraunhofer
diffraction occurs at a large distance from the scattering source, but with the aid of a lens the pattern created may be studied at close hand. Particles of a certain size give rise to light rings with known radii and intensities. The light intensities in different rings are measured with photosensitive diodes. The collected measured values then provide information about the occurrence of particles, the size and quantity of the particles. Known equipment aims at measuring either the total mass or the size distribution of the particles in a flowing medium. Measurement of the total mass of particles in many cases provides insufficient infor¬ mation about the composition of particles.
In the known devices for particle measurement, a large number of photosensitive diodes are used on the mentioned plate in the focal plane. These photosensitive diodes may be arranged in rings, called detector rings, each of which measures light for a specific radius of the light rings generated by the deflected light. By measuring the inten¬ sity of the light in the different angles of deflection and supplying the collected values to a computer for calcu¬ lations, the desired particle properties and particle distribution may be determined. The mathematics on which these calculations are based is dealt with in an article entitled "A Laser Diagnostic Technique for the Measurement of Droplet and Particle Size Distribution" from AIAA 14th Aerospace Science Meeting, Washington, D.C., January 26-28, 1976, by J. Switϊtenbank et al. From this article it is clear how a light power received by a detector ring can be calculated with the formula
πd
P = P.. LN
' — - J0 ( αSl } + J, ( -Sl } ■ Jo 0 ( -S2 f) ~ Ji 1 s2 > ] π - d λ
where Prj = net light power emitted from the laser L = length of the measuring volume N = particle density (number of identical particles/unit of volume) d = particle diameter λ = light wavelength si — inside radius of detector ring/focal length of lens S2 — outside radius of detector ring/focal length of lens Jo and Ji = Bessel functions of the 0-th and the
1st order, respectively.
To calculate inhomogeneous particle distributions with varying particle sizes within the measuring volume, a large number of detectors and very comprehensive calculation work are needed.
Another technique is to make use of square detectors, available on the market, with a large amount of photodiodes collected in a matrix. By collecting the measured values from photodiodes which are located at the same mutual distance from the centre of the detector plate, the same function can be obtained as with the ring detectors described.
The instruments mentioned above as well as available instruments of related types are primarily to be regarded as laboratory instruments and are not adapted for use in an industrial environment. The measuring distance is also greatly limited and renders impossible installation in, for example, flue gas channels in industrial plants . An instrument for measuring particles in such plants, if designed in accordance with the principles used so far, would become very costly.
Simple, reliable instruments for continuous measurements over longer distances, for example wide flue gas channels,
which are designed to give alarm on the occurrence of particles above a certain size or within a certain size range, do not exist. In addition, it is desirable to have instruments with a higher sensitivity to larger than to smaller particles in certain applications, for example in the field of application of this invention. No general particle measurement instrument is required here, but a device which is able to warn when too high contents of relatively large particles occur.
A variety of instruments for particle measurement of various kinds exist, but nothing is known which has the great size dependence required by instruments in this invention. What is desired is a device in which the electrical output signal therefrom is dependent on the occurrence of particle diame¬ ters with at least the factor diameter raised to the sixth power. The reason for this is that the dependence of the erosiveness on particle diameters in a gas turbine appro¬ ximately follows this functional relationship, that is, the extent to which the wear of the components of the gas turbine depends on the sizes of particles flowing there¬ through.
SUMMARY OF THE INVENTION
Gas turbines may be used for extracting residual energy from flue gases in power plants fired with solid fuels . To rapidly determine whether particles which act in an erosive manner on the turbines start arising in the flue gases, an instrument has been produced. This instrument is to warn when large particles above a certain size or within a certain size interval are detected.
The measured light power in one single detector ring is a function of the diameter of a light scattering particle according to the formula above. In principle, one ring would be sufficient as such to provide information as to the size of the analyzed particle. The problem is that it is
impossible to determine whether the measured light power emanates from one single large particle or from many minor dust grains, each of which scatters light and cooperates to form the measured value of the light power.
When the measured power of the light which falls upon different detector rings is transformed into electrical signals, tests have established that the difference in signal level between the output signals from two detector rings with diameters of different sizes gives a difference signal which grows more rapidly than the signal level of one single detector ring upon recorded increasing particle sizes. A more detailed study of the difference signal shows that this signal is dependent on the particle size above an extended size interval with a factor d raised to the sixth power. This is a fact which has not been previously mentioned.
This behaviour of the difference signal is due to the fact that, in the expression for the light power, dominating terms in dependence on particle diameters of the fourth order will be subtracted. The Bessel function JQ (X) in the light power formula above can be expanded in an even power series, see, for example, Handbook of Mathematical Functions, National Bureau of Standards, 1.964, section 9.4, with the following appearance
Jθ (x) = A + Bx2 + Cx4 + ...
where A, B, C are constants
whereas the Bessel function Ji (x) has an odd power series development according to
Jl(x) = Dx + Ex3 + Fx5 + ...
where D, E, F are constants.
In the formula for the received light power, the Bessel functions are included as square factors, and therefore a power series expansion of the received light power contains only even Squares of arguments and may, after simplification, be written as
Pi = Aid4 + Bid6 + Cid8 + ...
where P is light power d is particle diameter
Ai, Bi, Ci etc. are constants
for a first detector ring, provided that the light wave¬ length and the geometry of the instrument are considered to be fixed. For a second detector ring the received light power may be written as
P2 = A2d4 + B2d6 + C2d8 + ...
For, inter alia, the particle sizes which occur in flue gases, each preceding term is much greater in magnitude than the succeeding term. Thus, the light power value Pi is dominated by the term Aid4 and the light power value for the second detector ring P2 by the term A2d . If the detector rings are formed so that all sub-factors in the constants Ai and A2 are influenced to make Ai and A2 equal, the diffe¬ rence in received light power will be
Pi - P2 = (Bi - B2) ■. d6 + (Ci - C2) • d8 + ...
where the first term (Bi - B2) • d6 is dominating and from this follows that Pi - P2 is almost proportional to dβ, which is the sought size dependence. The proportionality is very difficult to verify quantitatively but is supported by experimental evaluations.
The difference signal formed from the two detector rings with the value Pi - P2 is dependent both on the total number
and the size distribution of the particles in a manner which makes the measured value of the difference signal a good measure of the total erosiveness of the particles in, for example, a gas turbine. For certain purposes there may be a need for a measure of only the size distribution of the particles independently of the total number of particles within the measuring volume. As will be clear from the appearance of the power formula above, the dependence on the particle density N within the length L of the measuring volume will disappear upon a division of the difference signal Pi - P2 by one of the absolute signals, for example Pi. The result
Pi - P2
Pi
will then be an approximate measure of the mean volume of the particles .
The invention is based partially on known technique. A laser is mounted on the side of a flue gas channel. The beam from the laser passes a windowed opening into the flue gas channel, extends perpendicular to the flue gas flow and continues at the other side surface of the flue gas channel out through a similarly windowed opening in the channel. The light then hits a lens provided with a small oblique mirror at the front in the centre. The mirror deflects the concentrated, unbroken laser beam to the side, whereas the lens focusses light deflected by the dust particles onto a plate provided with photosensitive diodes, the task of which is to detect incident light powers of the light falling on the diodes. If dust particles get into the path of the beam, light will be deflected from the optical axis of the laser. Light diffraction arises with the above-described light rings falling onto the detector plate. Measurement • takes place in a very limited selection of these light rings simultaneously by the photosensitive diodes being placed in the shapes of rings on the plate in only some of the positions where those light rings or parts of those light
rings, whose light power is desired to be recorded, fall upon the respective detector ring. The light power in the selected light rings are recorded and transformed into electrical signals, which are forwarded to an electronic device for signal processing, for example a computer. This device delivers a difference signal, which emanates from differences in light power of that light which falls upon the respective detector ring. The difference signal then controls instruments which provide concrete information about the total erosiveness of the flue gas particles.
TECHNICAL ADVANTAGES
The novelty in this invention resides in the fact that measurement is performed in only a small number of detector rings, which may be achieved by utilizing a difference signal according to the above. The method used provides rapid information about the occurrence of the sought particle size. It utilizes relatively simple components, therefore becomes comparatively inexpensive, and is simple to handle compared with other equipment available on the market. By the limitation to a small number of detector rings, the useful measuring distances for the instrument can be made several times longer than what is technically and economically reasonable with prior art methods.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 shows the relationship between recorded light power as a function of particle diameter, partly for a detector ring, partly for the difference power from detector rings positioned in pairs, all shown on a logarithmic scale.
Figure 2 show curves which illustrate the relationship between received light power as a function of particle size for two detector rings with different diameters but with the same area.
SUBSTITUTE SHEE" .SA SE
Figure 3 shows a schematic arrangement of all the components included in the device. In addition, the path of the light through the units is also shown.
Figure 4 shows an arrangement of the measuring device, in which the output power from the laser is measured in an alternative manner.
DESCRIPTION OF THE EMBODIMENTS
By making computer calculations on a selection of narrow detector rings with the aid of the above power equation for a large number of particle sizes, curves showing the received light power as a function of the particle diameter have been obtained. Figure 1 shows the resultant curves. From these curves it can be deduced that the read-off light power for a single detector ring is dominated by a dependence on the fourth power of the particle diameter. However, the difference in light power between two detector rings varies with a dominating term which is dependent on the sixth power of the particle diameter.
The main principle utilized in this invention is explained in Figure 2. This clearly shows how the light power increases with the particle size for each separate detector ring. As will be clear, an even greater size dependence has the difference which arises between read-off light powers for the two detector rings with increasing size of the light scattering particles . By the design of an instrument which reads this difference in light powers, a continuous super¬ vision of particle sizes in the flowing medium may be obtained.
A preferred embodiment of a device for detection of undesired particles in flue gases will be described with reference to Figure 3.
The light from a laser 1 first hits a spatial filter with a beam expander 2 which expands the beam and focusses this at a desired distance beyond the far edge of the flue gas channel, where the beam passes out from the channel through a window 3. A beam stop 4 is applied at the centre of the window 3 or, alternatively, at the centre of the lens 6, to divert the central light beam. The beam stop is formed as a mirror, which deflects the beam for further analysis . A light scattering dust particle 5 diverges the beams which fall upon it. The scattered light is collected in an achromatic double lens 6, passes through an interference filter 7 and thereafter illuminates the detector rings . The generated signals are forwarded to an electronic unit 9 for forming a difference signal, which provides information about the particle size.
The somewhat oblique window 10 on the laser side, which window is antireflex-coated only on the flue gas side, reflects light to a grey filter 11 which removes 99% of the incident light . After passage through an interference filter 12, this light is allowed to fall on a detector 13, which then provides information about the transmitted light power.
Grains of dirt on the glass 10 may also give backscattering. Such light is collected by an interference filter 14 with a lens 15 and is measured with a detector 16. Any fouling of the glass 10 may then be recorded here.
Transmitted light power from the laser can also be measured by an alternative arrangement of the equipment according to Figure 4. A beam splitter 21 in the form of a thin glass window is placed in the path of the laser beam at an angle of 45° to this. The reflex from the window, one side of which is antireflex-coated, is allowed to fall on a photodetector 22, which then provides information about the optical output power of the laser. In this case, the beam splitter is placed ahead of the spatial filter 2, any
disturbances of the laser light from the oblique glass in the beam splitter then being filtered off before it is allowed to penetrate the measuring volume. With this method, two advantages are gained, namely, that it is not necessary to tilt the window 10 nearest the measuring volume to obtain suitable laterally reflected light, and that the window 10 is removed while maintaining control of the optical output power. Otherwise, an oblique window 10 may have a negative influence on the laser light through the measuring volume.
The concentrated light beam which has been deflected by the beam stop 4 is first allowed to hit an oblique grey filter 17, which reflects a small and variable part of the light to a fiber-optic cable 18. The light from the fibre optics 18 is returned to one of the detector rings 8 to give both of these rings an equal amount of background light. The light which penetrates the grey filter 17 illuminates, after passage through an interference filter 19, a quadrant detector 20. To suppress the light power and adapt this to the photodiodes in detector 20, the grey filter has filtered off 99.9% of the original light. The detector 20 gives information about transmitted power in combination with detector 13 or, alternatively, detector 22. Deficient centering and alignment can be read out with a detector 20.
Calibration of the measuring equipment is carried out with the aid of a calibration glass which is placed at a definite distance in front of the lens, for example at the numeral 5, in Figure 3. Each calibration glass has been coated with particles of a definite size. This has been achieved by- dropping identical particles of a known size, suspended in a liquid, out onto glass plates. Four or five different sizes of particles which well cover the measuring range are suitable to utilize. The calibration glasses are placed in holders . When specific calibration glasses with known particle sizes are placed in the laser beam, the measuring equipment is trimmed for the best correspondence between
measured values and the known particle sizes of each calibration glass.
Instead of the above-mentioned components for realizing the instrument, it may, of course, be realized in a number of other ways . The grey filter may be replaced by a beam splitter, a semi-transparent mirror, etc. As an alternative to a fibre-optic cable, mirrors may be used, other light sources than lasers may be utilized, and it is also possible to replace photodiodes with other light-sensitive elements, such as photo-multipliers.
To increase the reliability of the instrument, it may, of course, incorporate more than one pair of detector rings. The difference signal from, for example, different pairs of detector rings may control electronic units separately, the output signals thereof then being compared or weighed together.
With suitable adjustments, the device may be used also for other applications than what has been described here, for example for the detection of suspended particles in liquids or in other fields where it is desired to monitor a corre¬ sponding presence of dust particles .
By adapting the diameters and design of the detector rings to the signal-processed electronic unit, detection of the desired particle sizes may be performed.