AU2005205846B2 - Method and apparatus for detecting particles in a gas flow - Google Patents

Description

P/00/011 Regulation 3.2

AUSTRALIA

Patents Act 1990 COMPLETE SPECIFICATION STANDARD PATENT Invention Title: Method and apparatus for detecting particles in a gas flow The following statement is a full description of this invention, including the best method of performing it known to us: 004919654 2

O

O

Method and apparatus for detecting particles in a gas flow IND Field of the invention 0 The invention relates to a method for detecting particles entrained in a gas flow and to an IND apparatus capable of carrying out such a method. More particularly the invention relates to a OO 5 triboelectric emission monitor and an optical dynamic emission monitor.

O

Background of the invention 0 0 There are several known means of measuring the level of a particulate material entrained in a gas flow.

These can be split into a number of categories including those based on electrostatic or triboelectric phenomena on the one hand, and those based on optical phenomena on the other.

As will be appreciated by a person skilled in the art, known particulate emission monitors have various drawbacks, such as. limited sensitivity to light particles, a tendency to have display velocity dependence in their measurements, non linearity between particle concentration and detector signal and limited diagnostic capabilities, to name a but few.

The applicant does not concede that the prior art described herein is part of the common general knowledge at the priority date of the application.

Summary of the invention In a first aspect, the present invention provides a method of improving the linearity of a probe signal produced by a triboelectric particle detector including the steps of: determining the instantaneous number of particles detected by said particle detector; associating said instantaneous number of particles detected by said particle detector with a value on a characteristic curve; and using said associated value to calculate the particle density from the detected probe signal.

Preferably, the step of calculating the instantaneous number of particles detected by said particle detector includes the additional steps of: determining the numerical density of particles in the gas flow; 004919654 3

O

O

estimating the volume over which a particle will produce a detectable signal by the particle detector; and

C/)

IND dividing the numerical density of particles in the gas flow by the estimated volume over

O

which a particle will produce a detectable signal by the particle detector to obtain the ,O 5 instantaneous number of particles detected by said particle detector.

S The characteristic curve is preferably determined empirically for said particle detector by a method including the steps of: S- maintaining all environmental and particle detector conditions substantially fixed;

O

determining the probe signal at a first known particle density; varying said particle density in a controlled manner until said particle density is equal to a second known particle density, and while varying said particle density; and determining said probe signal for the range of particle densities between said first and second known particle densities.

Preferably, the characteristic curve is substantially a square law for high particle densities.

The characteristic curve is preferably substantially linear for low particle densities.

Brief description of the drawings A number of embodiments of the present invention will now be described by way of example only, with reference to the accompanying drawings in which: Figurel shows a triboelectric probe according to an embodiment of the present invention.

Figure 2 is an end view of the probe of figure) showing the air flow around the probe and the mechanism for current production for the heaviest particles (produces both DC and AC current).

Figure 3 is an end view of the probe of figure 1 showing air flow around the probe and the mechanism for AC current production for medium weight particles and for heavy particles which are further from the probe axis.

Figure 4 is an end view of a probe according to an aspect of the present invention showing air flow around the probe and the mechanism for AC current production for light particles.

Figures 5A to 5E shows a series of exemplary probe shapes in accordance with the present invention.

004919654 4

O

O

S Figures 6A to 6J show a further series of exemplary probe cross sectional shapes in accordance with the present invention.

IND Figure 7 show a graph of the detector signal versus particulate density for a typical prior art 0 triboelectric probe, and also a model linear probe output.

Figure 8 shows a graph of amplifier gain plotted against frequency, for the frequency response of S the integrating pass band of a probe according to an embodiment of the present invention, and also an alternative method using two separate pass-bands.

Figure 9 shows a schematic circuit diagram for an emission monitor according to an embodiment of the present invention.

Figures 10OA to 1 OH show a number of examples configurations of optical dynamic opacity detectors.

O

O

Figure I I shows an emission monitor according to an embodiment of the invention having a plurality of optical dynamic opacity detectors.

ND Detailed description of the embodiments Figure 1 illustrates a detection means being an embodiment of to the present invention. The detection 5 means 70 is adapted to detect particulate material entrained in a gas flow, the direction of which is indicated by 00 arrow-. The detection means 70 includes a probe shaft 40 with a generally square cross-section having four faces.

On each of the faces of the probe shaft 40 there is a separate and electrically isolated probe electrode 42. 44, 46.

48. Probe electrodes 46 and 48 are located on the upstream side 43 of the probe shaft 40 and probe electrodes 42 and 44 are located on the downstream side 45 of the probe shaft 40. Detection means 12 also includes an insulating sleeve 72 to prevent current leakage from the probe electrodes 42, 44, 46, 48 to the duct (not shown).

When the detection means 70 is placed in a gas flow, particulate material which is suspended in the gas flow interacts with the probe electrodes 42, 44, 46, 48 to produce a signal which can be used to determine the particulate flow rate. Heavy and/or large sized particles interact primarily with the probe electrodes 46, 48 on the upstream side 43 of the probe shaft 40. Whereas lighter and/or smaller particles interact primarily with the probe electrodes 42, 44 on the downstream side 45 of the probe shaft 40. By this means a detector signal will be produced by the upstream electrodes 46, 48 which is related to the quantity of large particles and a separate signal is produced by the downstream electrodes, 42, 44 which is related to the quantity of small particles in the gas flow. As will be described below these signals may be processed singly or together to calculate such qualities as: Mass Concentration; Mass Flow Rate; Numerical Concentration: Volume Concentration: and Class Concentration.

Additionally, the signals can be processed together to determine the particle size distribution by empirical methods.

The principle of operation of the detector is as follows.

If the gas flow is smooth then all particles suspended in the gas flow regardless of the particle size are able to substantially follow streamlines in the gas flow. However, if the gas flow is disturbed, for example by placing an object in the gas flow. some particles will continue to move in a straight paths, ie their inertia will cause them to continue in their original direction as if the disturbance had not occurred, and collide with the obstruction, as depicted by path 41 of figure 1. whilst other particles will continue to follow the streamlines of the gas as depicted by path 47 of figure 1.

At velocities commonly encountered in bag houses etc, where a detection means such as described herein may be used, the gas flow will become turbulent on the downstream side 45 of a detector. Tests have shown that the downstream detectors eg 42. 44 are particularly sensitive to light particles. It is believed that the increased sensitivity of the downstream detector to light particles is due to the tendency of light particles to be drawn into the IN turbulent flow downstream of the probe, where they circulate back toward the downstream detectors 42, 44 thereby increasing the intensity of the resultant signal. The path 49 is that followed by the lightest particles in the gas flow.

Figure 2, 5 and 6 show an end view of a probe 40 similar to that shown in figure I with the streamlines of the gas IN flow depicted in figures 2, 3 and 4 by dashed lines. As can be seen, the gas flow remains laminar streaming around 00 the contour of the upstream side 43 of the probe 40 until the point where the gas must make a sudden change in direction to follow the contour of the probe 40. At this point turbulence in the form of vortices or eddies is ,1 produced in the gas flow.

As shown in figure 2 the heaviest particles will continue to move in a straight line even though the gas Sflow has diverted to flow around the probe 40. These particles will collide with the electrodes 46. 48 on the upstream side 43 of the probe 40 producing a detector current as described above. However, some particles will follow the streamlines of the gas flow to a sufficient extent to avoid collisions with the probe. Figure 3 shows the situation that occurs with medium weight particles, which follow the gas flow and are diverted around the probe initially. However, they are too heavy to follow the eddies created on the downstream side 45 of the probe where the gas flow becomes turbulent. Once the particle is no longer entrained in the gas glow it will continue to move generally downstream from the probe 40 in a substantially straight path. Thus the larger particles interact with the electrodes 46, 48 on the upstream side 43 of the probe 40 by inducing a current in the electrodes 46, 48 or by transferring charge by contact with the electrodes 46, 48.

Conversely as shown in Figure 4, particles of low mass will follow the stream lines of the gas flow and be drawn into the turbulent flow on the downstream side 45 of the probe, bringing the smaller particles within the detection range of the downstream electrodes 42, 44. By having a detector on the downstream side 45 of the probe the probability that lighter particles will be detected is increased.

Advantageously, the probe 40 acts as an air flow deflection means, deflecting the airflow around its contour and causing turbulent flow on its downstream side.

In the embodiment shown in Figures 1, 2, 4 and 4 there are four probe electrodes in total, namely electrodes 46, 48 on the upstream side 43 of the probe shaft 40 and electrodes 42, 44 on the downstream side 45 of the probe shaft 40. It is advantageous to have pairs of detectors on the upstream 43 and downstream sides 45 of the probe shaft 40, as it allows the two upstream electrodes 46, 48, and the two downstream electrodes 42, 44 to be coupled differentially to the electronics module of the emission monitoring system to cancel out electrical interference common to both detectors in the pair, such as the electric fields generated by a non-earthed duct or an electrostatic precipitator.

It should be noted that the shape of the probe 40 and the number of electrodes on the surface can be changed to suit the application. The probes of figures 1 to 4 are of square or lozenge shape in cross-section.

However, figure 5 and figure 6 illustrate alternative probe cross sections. If desired a probe can have a single upstream electrode and single downstream electrode, or more than two upstream and two downstream electrodes.

Figures 5A to 5E show end views of five additional embodiments of detection means. In each figure, the direction of the airflow is shown by arrow 50 and the airflow pattern around the airflow deflection means is shown in broken lines.

O

Figure 5A shows an end view of a triboelectric probe 40 with a generally circular cross-section having two detectors 80, 90. The probe 40 is integral with the airflow deflection means.

The airflow pattern around a probe with a circular cross section is dependent on the velocity of the airtlow 00 f and the size of the probe. Since it will usually be the case that the velocity range for the airflow during normal operation of the equipment on which the probe is installed will be known, the cross sectional size of the probe can be chosen so that vortices are created on the downstream side of the probe. As will be appreciated by the person skilled in the art, the position of vortex separation for a probe with a smoothly curved cross section varies depending on the velocity of the airflow. Therefore, it is preferable to have a corner or some other surface feature on the deflection means, or probe surface, to act as a deflection means, or to have a separate deflection means, to induce vortex separation at a predetermined position. Knowing the position of vortex separation greatly simplifies the task of positioning the detector adjacent to, or in, the turbulent flow.

Returning to figure 5A, the upstream detector 80 will detect particles following the laminar airflow as they are deflected around the front part of the probe 40. The downstream detector 90 will detect any particles that are drawn into the vortices on the downstream side 45 of the probe Figure 5B shows an end view of a triboelectric probe 40 with a generally rectangular cross section and a detector located on the downstream side 45 of the probe shaft. The probe shaft 40 again acts as the airflow deflection means for this probe. In this probe 40 there is no upstream detector, thus only particles light enough to be drawn into the vortices on the downstream side 45 of the probe shaft 40 will be detected by the detector 90. A probe of this configuration may be particularly suited to applications where only particles below a certain size are to be measured.

Figure 5C shows a detection means including two elongate probes 40, 40A. The first probe shaft 40 has a detector 80 on its upstream face. The first probe 40 acts as an airflow deflection means causing turbulence on its downstream side 45. A second probe shaft 40A is located within the turbulent flow downstream of the first probe shaft 40 and has a detector 90 on its downstream side for detecting small particles entrained in the turbulent flow.

The positioning of the downstream detector 90 on the second probe shaft 40A can be chosen to increase the likelihood of interaction with small particles, as can the position of the second probe shaft 40A relative to the first probe shaft 40. As discussed above, the upstream detector 80 will tend to detect heavy particles, whereas the downstream detector 90 will detect lighter particles.

Figure 5D illustrates a detection means including two elongate probes 40, 40A and an airflow deflection member 82. The first probe 40 has a detector 80 on its entire surface. This first elongate probe 40 is effectively a standard prior art triboelectric probe. The second probe shaft 40A is placed in the turbulent flow downstream of the airflow deflection member 82 and has two detectors 90 on its surface for detecting particles drawn into the turbulent flow on the downstream of the deflection member 82. The positioning of the downstream detectors 90 on the second probe shaft 40A can be chosen to increase the likelihood of interaction with small particles dependent on the airflow pattern, as can the position of the second probe shaft 40A relative to the airtlow deflection means 82.

As discussed above a detector in the laminar flow will have a greater sensitivity to larger particles, whereas the second probe shaft, downstream from the deflection means, will be more sensitive to lighter particles.

O

Figure 5E shows a triboelectric probe 40 with a generally rhombic or lozenge shaped cross section with a detector 90 located on the downstream side of the probe shaft 40. The shape of probe shaft 40 enables the probe shaft 40 to act as the airflow deflection means for the detector 80. In this embodiment there is no upstream detector 0 thus only particles light enough to be drawn into the vortices on the downstream side of the probe shaft 40 will be

V')

detected by the downstream detector 80. A probe of this configuration may be particularly suited to applications Swhere only particles below a certain size are to be measured.

O

S 10 Figures 6A to 6J show cross sectional views of a further ten probes according to embodiments of the present invention. The probes in each of the figures 6A to 6J are elongate probes adapted to be placed in a gas flow moving in the direction shown by arrow In Figures 6A to 6J like parts are labelled alike. Each of the probes shown in Figures 6A to 61 have two upstream electrodes 46, 48 and two downstream electrodes 42, 44 attached to a probe shaft 40. In these embodiments the probe shaft 40 is also acting as the airflow deflection means with the corners 51, 53 acting as vortex separation means. It will be noted that each of the embodiments of probes shown in figures 6A to 6J are symmetrical about an axis parallel to the direction of gas flow 50. Whilst this feature is not essential to the operation of the system, by virtue of the fact that the symmetrical shape of the probe, and symmetrical positioning of the upstream and downstream detectors results in approximately equal signals being produces by each of the upstream, and each of the downstream detectors, cancellation of interference if the upstream and downstream detectors can more easily be effected when the detectors are connected to the measuring means differentially.

Figure 6A shows a probe with a square cross section similar to that shown in figure 1. The corners 51 and 53 provide a vortex separation means which induce turbulence in the airflow 50 downstream of their position, and in particular, cause turbulent flow adjacent to the downstream detectors 42, 44. The elongate nature of this crosssectional shape Figure 6B depicts a probe 40 of rhombic or lozenge shaped cross section. The corners 51 and 53 provide a vortex separation means which induce turbulence downstream of their position, and in particular, cause turbulent flow adjacent to the downstream detectors 42, 44. An elongated shape or streamlined shaped probe can be advantageous in conditions where the abrasive or corrosive nature of the particulate matter causes the probe 40 to be eroded away. By streamlining the probe the number of particles that collide with the probe 40 are reduced.

thereby reducing the effect of the particles of the probe Figure 6C also depicts a probe 40 of rhombic or lozenge shaped cross section. The corners 51 and 53 provide a vortex separation means which induce turbulence downstream of their position, and in particular, cause turbulent flow adjacent to the downstream detectors 42, 44. The relative sharpness of the corners 51 and 53 in comparison to the probe shown in figure 6B will cause grater turbulence downstream of their position.

Figure 6D shows a probe 40 having a generally triangular cross section. The upstream detectors 46, 48 are positioned side by side on the flat front face of the probe shaft 40. The downstream detectors 42. 44 are positioned on respective downstream faces of the probe shaft. Again, the relative sharpness of the corners 5 I, 53 will cause 0 greater turbulence and can increase the sensitivity of the detector to light particles. Although the upstream detectors 46, 48 are on a single surface it is advantageous to have two separate detectors so that they can be coupled N differentially to the measuring means in order to cancel out interference.

00 Figure 6E shows a probe 40 with irregular polygonal cross section, which is symmetrical about an axis Slying parallel to the direction of gas flow 50. Each of the upstream faces are of the same length and meet along the iaxis of symmetry of the probe 40 at a right angle. Each of the downstream faces are of the same length and meet 0 10 along the axis of symmetry of the probe 40 at an acute angle. The angles of the corners 51 and 53 are equal. In N general the geometry of the probe cross section can be chosen to maximise the contact between the downstream detectors 42, 44 and the turbulence created by the corners 51, 53.

Figure 6F shows a triangular shaped probe 40. The probe cross section is that of an isosceles triangle with its base being on the downstream side of the probe. Thus, the upstream detectors 46, 48 are angled relative to the direction of the gas flow, whereas the downstream face of the probe 40 on which the downstream detectors are mounted lies in a plane perpendicular to the gas flow. The sharp cuttoff of the probe 40 at the corners 51. 53 will cause greater turbulence and can increase the sensitivity of the detector to light particles.

Figure 6G to 6J show probes 40 with cross sections having curved surfaces. The shape of the curves can be chosen to maximise turbulence produced by the corners 51, 53, or to ensure that the downstream detectors 42, 44 are positioned adjacent to the turbulent flow, or to aid in the deflection of larger particles around the probe in order to reduce collisions and hence probe erosion.

Figure 6G shows a probe 40 having a cross section of a quadrant of a circle, with the upstream detectors 46, 48 mounted on respective radial sides of the quadrant shape and the down stream detectors 42. 44 being positioned side by side on the circumferential side of the quadrant. The probe of Figure 6J has the same cross sectional shape as the embodiment of figure 6G with the exception that the probe is oriented such that the vertex between the two radial sides of the quadrant shaped cross section lies on the downstream side of the probe 40. Thus the curved surface forms the upstream side of the detector and supports the upstream detectors 46, 48.

The embodiments of probes shown in figures 6H and 61 have two flat faces and two concave surfaces, and are generally diamond shaped in cross section. In Figure 6H the upstream side faces of the probe 6H are concave.

whereas in figure 61 the downstream faces are concave.

As discussed above the corners 51 and 53 in figures 6A to 61 induce the separation of the gas flow from the contour of the probe and produce turbulent flow downstream of their position, and particularly adjacent to the downstream detectors 42, 44. The concave surfaces on the upstream side of figure 6H ensures strong vertex separation at the corners 51, 53.

The method according to the invention may also be applied to a detection means, which is not an elongate body such as a probe. For example a detector may be mounted on the downstream side of a mass suspended in the

O

O

gas flow and shaped in an appropriate fashion to cause turbulent flow adjacent to the detector in order to detect light particles in the gas flow. Clearly an upstream detector may also be used with a detection means of this configuration in order to detect longer particle sizes. Application to a ring type detection means is also possible.

The configuration of the detection means depicted in the accompanying figures are merely shown by way of example and are not intended to limit scope of the invention to the configurations depicted.

In the embodiments described above the detectors mounted on the detection means are of the triboelectric 00 type, however the invention is also applicable to detection means utilising other detector types.

Figures 10A to 10H show a series of configurations of prior art optical emission monitors mounted in a conduit. The block marked 1201 is in each figure a transmitter and the block marked 1202 is in each case, a receiver.

Figures 10A to 10H show a series of configurations of optical emission monitors mounted in a conduit.

The block marked 1201 is in each figure a transmitter and the block marked 1202 is in each case, a receiver.

Figure 10 A shows a simple detector with the transmitter 1201 placed diametrically opposite the receiver 1202. Light is transmitted across the conduit through the gas flow to the receiver 1202.

Figure O10B shows a detector with the transmitter 1201 placed adjacent to a receiver 1202 with a mirror 104 placed diametrically opposite the receiver 1202 and transmitter 1201. Light is transmitted across the conduit through the gas flow to the mirror 104 and reflected back to the receiver 1202. This type of detector provides approximately twice the optical path length of the detector of figure A meaning that the active detection volume is increased. However, a detector of this sort is more difficult to set up due to the difficulty in aligning the optics and is also difficult to keep contamination free.

Figure O10C shows a detector with the transmitter 1201 placed adjacent to a receiver 1202 with a prism 106 placed diametrically opposite the receiver 1202 and transmitter 1201. Light is transmitted across the conduit through the gas flow to the prism 106 and reflected back to the receiver 1202. Again this set up has a high active detection volume while being a little easier to set up than the detector described in 10B, however keeping the prism contamination free is again a difficulty.

Figure 10D shows a detector with the transmitter 1201 placed adjacent to a receiver 1202 with a mirror 104 mounted on a mounting means 108 opposite the receiver 1202 and transmitter 1201. Light is transmitted across the conduit through the gas flow to the mirror 104 and reflected back to the receiver 1202. This type of detector can be installed in the conduit 18 as a single article in order to increase the robustness of the detector allowing the alignment of the optics to be fixed in the factory.

Figure 10E shows a detector with the transmitter 1201 placed adjacent to a receiver 1202 with a prism 106 mounted on a mounting means 108 opposite the receiver 1202 and transmitter 1201. Light is transmitted across the conduit 18 through the gas flow to the prism 106 and reflected back to the receiver 1202. This type of detector can also be installed in the conduit 18 as a single article in order to increase the robustness of the detector. Alignment of the optics is still important but again this can be fixed in the factory.

Figure 10F shows a detector with the transmitter 1201 placed adjacent to a receiver 1202 with a optic fibre 110 mounted on a mounting means 108 opposite the receiver 1202 and transmitter 1201. The optic fibre or optic N fibres 110 are housed within in a protective shield. Light is transmitted across the conduit 18 through the gas flow to the optic fibre 110 and then guided back to the receiver 1202. This type of detector provides can also be installed in the conduit 18 as a single article in order to increase the robustness of the detector. However alignment of the optics is still important.

00 0Figure O10G shows a detector with the transmitter 1201 placed adjacent to a receiver 1202 with a J shaped optic fibre(s) 100 extending from the receiver 1202. The optic fibre(s) 100 are housed within in a protective shield.

f Light is transmitted across the conduit 18 through the gas flow to the optic fibre(s) 100 and guided back to the receiver 1202 by the optic fibre(s) 100. This type of detector can also be installed in the conduit 18 as a single Carticle in order to increase the robustness of the detector. Alignment of the optics in this case is simplified, as there is only one air-optics interface to align.

Figure 10H shows a detector with the transmitter 1201 placed adjacent to a receiver 1202 with a J shaped optic fibre(s) 100 extending from the transmitter 1201. The optic fibre(s) 100 are housed within in a protective shield. Light is guided via the optic fibre(s) to a position in the centre of the gas flow, then transmitted through the gas flow to receiver 1202. This type of detector can also be installed in the conduit 18 as a single article in order to increase the robustness of the detector. Alignment of the optics in this case is also simplified, as there is only one air-optics interface to align.

Any one of the configurations shown in prior art figure 10A to 10H can be modified or adapted to be an embodiment of the present invention, however the configurations depicted in figure 10G and 10H are preferable as they are mechanically robust and offer features particularly advantageous for the application described herein as will be apparent to a person skilled in this art.

Figure 11 shows a detection system using an optical detector of the configuration shown in figure combined with the detection means of figure I. Similar to the embodiment shown in figure 1 the detection means shown in figure I I has a probe shaft 40 with a rhombic or lozenge shaped cross section. It has two upstream faces 46, 48 and two downstream faces 42, 44 each with associated detectors. The optical fibres, which are acting as light guides for the detectors in this detection system, are housed within the probe shaft 40. A curved elbow portion 102 as shown in figure 10 protrudes beyond the probe shaft in order to transmit a beam of light 104 along each face of the probe shaft. Each detector additionally has a receiver 106 mounted on the opposite end of the probe shaft to the curved elbow 102. Thus. the detector produces a light beam substantially parallel to each surface 46, 48, 42, 44 of the probe shaft 40. The downstream detectors will be more likely to detects smaller particles which are entrained in the turbulent flow on the down stream side of the probe shaft 40. Modifications of the optical detection system which are necessary to suit optical dynamic detection of particles, rather than triboelectric detectors, will be known by a person skilled in the art.

In a detection means with multiple optical detectors as shown in figure I I the detectors can share a light source eg. a laser diode or other suitable light source, or have one light source for each detector.

It is also possible for the light beam to be shaped so as to increase the likelihood of detecting particles. By shaping the light beam as a ribbon the active detection volume is increased. Further, the distance between the probe 1 shaft and the light beam can also be adjusted to ensure detection of the maximum number of particles depending on the probe geometry and velocity of the gas flow.

As discussed above, an effective "closed loop" system can be realised for an optical dynamic detection monitor by modulating the amplitude of the light source, then detecting only that portion of the received signal which is related to that modulation, thus excluding the effect of the interfering signal.

Turning now to the processing of the detector signal(s), the currents induced in the detectors (eg 42, 44, 46, 48 of figure The detector signal(s) and are processed by the electronics module 16 of the emission S 10 monitoring system (see figure Generally, the electronics module 16 can include either analog or digital signal processing circuitry, therefore in this description unless otherwise stated no distinction will be made between these two possible implementations.

The electronics module of an emission monitor is shown in figure 9 as block 16. In general terms the electronics module contains all electronic and signal processing components of the emission monitor with the exception of the probe itself and the measuring means 52.

The AC component of the signal probe is preferably used for measuring particulate emissions. The DC component is considered unreliable for emission monitoring however it can be used for diagnostic purposes and testing. The chosen cut off frequency between DC and AC frequencies is about 0.1 Hz. However, this can be increased if the velocity of the particles is always high. The upper cut off frequency is limited to below 50 or 60 Hz (as appropriate) in order to avoid interference from the mains power frequency. In most applications the upper cut off frequency of the pass band for the signal processing path of the electronics module is approximately 20 Hz. As will be clear to a person skilled in the art, the upper and lower cut off frequencies of the pass band are chosen by taking into account practical circuitry considerations and possibly the need to restrict total circuit capacitance if the device is to be applied in intrinsically safe situations.

In a digital version the detector signal may be sampled at a rate of 55 Hz. rendering the mains frequency a residual signal at 5 Hz and multiples of 5 Hz. This signal can then be removed using a comb filter that rejects frequencies that are a harmonic of 5 Hz.

Integrator 54 is configured to have a frequency response that decreases as the detector signal frequency increases, as shown by plot 1 of figure 9. In Figure 9, the gain of the integrator is plotted against the detector signal frequency on a log scale. In short, the integrator 54 is less sensitive to higher frequencies and more sensitive to lower frequencies. It can be seen that over a substantial portion of the graph that the frequency response of the integrator is linear with a drop of approximately 6 dB per octave. This is done in order to make the output of the integrator 54 substantially independent of velocity.

If the velocity of the gas flow and particulate material is doubled all mechanical and electrical processes that occur will happen in half the time, ie the process of a particle approaching and receding from the probe will take half the time it did at half the velocity. Consequently the duration of the signal produced by each event will be halved and the number of events that occur will double, since the total charge transferred remains the same, the current (which is charge per unit time) will double for any given particle density. Accordingly. the gain of the integrator is reduced in an approximately inversely proportional manner to the amplitude of the detector signal as O this frequency changes, rendering the processed signal and hence particle measurement substantially independent of the particle velocity.

In order to partially compensate for frequencies that have been cut off either above or below the pass band 0 of the integrator an increase in gain is provided for signals near the extreme ends of the pass band.

An alternative method of providing a substantially velocity independent measurement of the density of fparticles entrained in the gas flow can be performed by splitting the detected signal into a high band and low band

O

S 10 detector signal which is then independently processed and smoothed. The ratio of the two values is then used for determining an estimate of velocity by comparison with an empirical relationship. This velocity estimate can then be used to determine the particulate mass concentration as shown below. Clearly empirical testing is required in order to determine what ratio of high band to low band frequency relates to a particular velocity of flow.

Plot 2 and 3 of figure 9 represent the low band gain and high band gain respectively for an implementation of this method. The low band signal is centred at 1.0 Hz and can be processed by an upward rate limiter to reduce the low band signal sensitivity to interference from low frequency noise caused by, for example, human movement, damper actuation and isokinetic testing, etc. This interference can cause significant energy in the signal at the low frequency end of the pass band that is not directly related to the detector signal. The high band signal is centred at 12 Hz.

As can be seen from figure 9, the relationship between the peak gain for the high band and low band signal is approximately inversely proportional to their central frequency. The line joining the peak amplitude points of the two pass bands on a Bode plot would fall at approximately 6dB per octave which is approximately equal to the response of the integrator in the previous method. In testing it has been found that frequencies above the upper cut off of the low band should be attenuated at approximately 12 decibels per octave more than frequencies below the lower cut off of the high band signal. In order that the high band filter adequately rejects mains interference, it may be designed with higher (lower damping) than the low band filter.

The ratio between the detected high band signal and detected low band signal are then applied to an empirical function generator to determine an indicated particle velocity. The velocity can then be used to calculate the mass concentration of particulate material in the gas flow as described above. Or alternatively this approach can be combined with the previous method by using the velocity estimate, to remove any remaining velocity dependence from the emission measurements.

Alternatively. a weighted sum of the high band signal and the low band signal can be used to determine the mass concentration of the particular matter in the gas flow directly.

The key quantities to be measured by an emission monitoring system according to the current embodiment include, inter alia, the following: Mass Concentration: Mass Flow Rate; Numerical Concentration; Volume Concentration; and Class Concentration.

As will be appreciated by a person skilled in the art, Mass Concentration is the most commonly required 00 output in dynamically monitoring airborne industrial discharge, whereas, Mass Flow Rate is the most commonly required output in assessing the total airborne discharge from industry over a period of time. Numerical if Concentration may be found useful if as expected, future standards place greater importance on smaller particles than present standards. Furthermore, Class Concentration provides a numerical particle concentration weighted I 10 with particle size so as to match common clean room specification. For example, an output value of 1000 means that the gas stream just meets the particle size distribution specified for class 1000.

Alternatively Volume Concentration may be an alternative to Mass Concentration if particle density is not known.

The other deduced values Gas Velocity, Particle Size and Probe Capacitance are useful for other purposes as described elsewhere.

A process for determining each of these quantities will now be described in which: Input parameter K is a calibration constant for linearity; K may be adjusted in proportion to relative triboelectric activity of the particulate being measured, and reduced to account for any shielding effects; Input parameter N is the calibration constant for final scaling: N accounts for many factors; Input parameter D represents the compressed solid particle density in kg/litre, or specific gravity: (A default value of 3 may be used).

Input parameter C is a clean room class factor used to produce weighted output in particles per cubic foot; C is determined by a single test under controlled conditions; and Input parameter A (effective duct area in m 2 default I: may be adjusted to scale True Mass Flow Rate output) As discussed above, it is advantageous for the detector signal to be proportional to the particle density. As each particle that passes the detector induces a small current in the detector consisting of a negative part and a positive part, when many particles are exposed to the detector at any one time, the total signal will be the sum of a number of positive parts and a number of negative parts induced in the detector; this results in a degree of cancellation or masking. As the signals from these particles are essentially unrelated to one another, the sum of the all these individual signals will be a noise signal whose total power is related to the sum of the powers of the constituent signals. As current and voltage are related to the square root of power, then the measurement of detector current or voltage will produce a square-root characteristic, as shown in Figure 7 by plot Y3. Mathematically. the total signal is said to be the RMS (root mean square) sum of the constituents, which is far smaller than the arithmetic sum of the individual amplitudes.

SIf the numerical concentration of particles in the gas stream is low enough only one particle is detected at a time, so there is no cancellation, and the detected signal at the measurement means is linearly related to the average particulate mass concentration. as shown in Figure 7 plot Y2. In any real measurement the characteristic will vary from that of Y2 at very low signals to that of Y3 at very high signals, as shown in Figure 7 plot Y 1.

00 In practice, observations on test facilities have indicated that for a 10:1 increase in particulate concentration there will be observed an increase in detected signal commonly between 5:1 and 7:1. By way of comparison, a 10:1 increase in particulate concentration would cause an increase in detected signal of 10:1 if the 0 10 system were linear (Y2 in Figure or 3.16:1 if the system were true square root (Y3 in Figure Since the observed non-linearity is clearly in the transition region between these two extremes, the non linearity characteristic could move towards one extreme or the other as the number of particles exposed to the detector varies. Therefore in order to achieve linearity, an indicator of the number of particles exposed to the detector is found and used to linearise the detector signal, rather than assuming a fixed characteristic for this non-linearity correction.

If a particle detection system includes means to measure gas velocity, mean particle size and probe capacitance as proposed elsewhere, then those various signals may also be used to provide the means to deduce the number of particles exposed to the detector, and thence to correct for linearity errors. If any one or more of these signals is not available, then a nominal or estimated constant may be substituted eg two methods of measuring gas velocity are described below.

In determining the number of particles exposed to the detector, Probe Capacitance is used as an indication of the effective volume around the probe in which particles can be detected, since the two are related to a sufficient accuracy for these purposes. In the calibration of Probe Capacitance, the reference (zero) condition should include any inactive portion of the probe. For example if the probe projects through a tunnel to the duct, then the reference measurement should be made with a short probe attached that projects through the tunnel just to the duct, but no further. The use of this parameter in the calculation may provide the important benefit of automatically adapting to different probes without otherwise recalibrating, even for reasons such as probe breakage, erosion or high buildup.

In order to determine the quantities of interest the following calculations can be performed: Measure Concentration, Gas Velocity, Particle Size and Probe Capacitance as detailed elsewhere.

Convert Concentration into ConcentrationA by using Gas Velocity and Particle Size to remove dependency on Gas Velocity or Particle Size, as detailed elsewhere.

Calculate Particle Volume ParticleSize Calculate ConcentrationB ConcentrationA Probe Capacitance Particle Volume.

Calculate ConcentrationC ConcentrationB K.

Calculate Numerical Concentration ConcentrationC (I ConcentrationC) N Probe Capacitance (or other relationship which generally reverses the nonlinearity of Figure 7 plot Y1.) 16 Calculate Volume Concentration Numerical Concentration Particle Volume.

Calculate Class Concentration Volume Concentration C Particle Size'.

Calculate Mass Concentration Volume Concentration D.

Calculate Mass Flow Rate Mass Concentration Gas Velocity A.

5 Clearly A and D should be adjusted to suit the application. K. N and C may be set initially to any values Swhich suit the system used, but subsequently will be set to specific installation default values. Finding the default value of K requires two initial isokinetic tests at substantially different concentrations. Preferably this is performed with minimal variation in the other parameters, so that any residual dependencies on those other parameters will not influence the calibration. After such calibration, K will become fixed for that particulate material and conditions, CI 10 and a further simple individual test may be carried out to determine the values of the scaling factors N and C.

The accuracy of all these calculations is preferably, but need not be, better than one percent, but as the dynamic range required may be very large, so it may be found more convenient to represent all values in logarithmic form. Logarithmic representation of such values additionally has the benefit of allowing faster execution on smaller processors which may lack a multiply facility, since the simple addition of logarithmic values is equivalent to a multiplication operation.

In any given application, individual on-site isokinetic testing can still further improve accuracy by fine adjustment of K, particularly if conditions are not optimal. By recording the results of all such individual on-site testing, a database of K values will eventually be compiled for all relevant dust materials in various forms and conditions, for example in turbulence, or after an electrostatic precipitator or wet scrubber. It is clear from the foregoing that any attempt made in the past to compile such a database would be severely flawed if it failed to account for these factors.

An exemplary initial calibration process is as follows: Run the plant on which the emission monitor is mounted with a high particulate mass concentration. Once the flow has stabilized the first of the isokinetic tests used to calculate K as described above can be run.

Average ConcentrationB as Concl and simultaneously accumulate a sample isokinetically; at the end of Test 1, determine from that sample the total average mass concentration Isol.

Run the plant with a low particulate mass concentration. Once the flow has stabilised run Test 2 as described above, averaging ConcentrationB as Conc2 and simultaneously accumulating a sample isokinetically: at the end of Test 2, determine from that sample the total average mass concentration Iso2.

Calculate G Conc2 Isol (Concl Iso2), G represents the required additional gain increase over the tested range to make dector the output linear.

Calculate K for that material and those conditions using the (Conc I G*Conc2) (G I).

A new isokinetic test at any typical concentration under the same conditions can be conducted to determine values of N and C such as to rescale the Mass Concentration and Class Concentration outputs to their correct values. These values need not be further adjusted, and they can then be fixed for that implementation of the C/ system.

N If the invention is equipped with a serial data interface such as a network port. then all these deduced 0 values may be made available in registers. If only a 4-20mA interface is provided, then only one value may be available at any one time.

In order to check that the signal processing means of the electronics module 16 is functioning correctly, a validation signal can also applied to the probe via input 58 (figure The validation signal is an AC signal with a frequency that lies within the pass band of the signal processing means. The validation signal is combined with the detector signal and processed by the signal processing means simultaneously.

'i 10 As the validation signal is at a known frequency, it can be filtered out using an appropriate narrow band filter and analysed separately to the detector signal. For example, the validation signal is generated at 10 Hz and once processed can be removed from the detector signal using a filter.

In the digital version above the comb filter used to remove mains frequency from the signal will also remove the validation signal from the measurement signal if a 10Hz validation signal is used. As already discussed, the combined detector and validation signal are sampled at a frequency of 55 Hz thus leaving the validation signal and mains frequency noise as a residual signal at either 5Hz or a harmonic of 5 Hz. It is also possible to apply a comb filter to the detector signal to separate the measurement signal from the validation signal and the mains interference to obtain further diagnostic information about the functioning of the emission monitor. It will be clear to the person skilled in the art that due to the nature of the detector signal an insignificant amount of the measurement signal is lost by filtering our the validation signal and mains interference.

The processed validation signal once separated from the measurement signal can then be compared to a reference signal. The reference signal is produced by applying the validation signal to the signal processing means via input 58 during calibration of the instrument (ie. not under normal operating conditions), and measuring the processed signal produced at the input to the measurement means. The processed reference signal provides an indication of the operation of the signal processing path on the known reference signal under known conditions.

This signal is compared to the processed validation signal under normal operating conditions in order to detect any variations in the operation of the signal processing means. The comparison may be carried out by the measurement means, or if one is available, a data acquisition system or the SCADA system attached to the emission monitoring system. If a data acquisition system or the SCADA system is used a profile of the operation of the apparatus can be built up over time, which can assist in the validation of the emission monitoring system.

Additional information can be gained about the operation of the probe using the DC component of the detector signal. It has been observed that under some conditions, for example when moisture bridges across the insulation between probe and duct, the DC component of the detector current may rise substantially with reference to the AC component. The ratio of the DC component to the AC component can therefore be calculated and compared with a threshold value. When this ratio exceeds the threshold, an alarm may be raised to indicate abnormal process conditions.

A further diagnostic feature can also be provided for a triboelectric emission monitor of this sort. A known electrical excitation signal, known as an admittance measuring signal 62, can be applied to the detectors. and the IN detector voltage produced, measured by admittance measurement unit 64. The admittance measuring signal should be at a sufficiently high frequency to allow a lpF variation in the capacitance of the detector to be easily measured.

for example, a 40kHz signal is suitable. The higher the admittance measuring frequency, the more circuit N admittance will be tolerable, and it may be prudent to use significantly higher frequencies.

00 In order to detect changes in the admittance of the detector 12 of the previous embodiment, a calibration measurement must be made for a number of known admittances, for example, with no additional admittance i.C attached across the detector, and with a single known reference resistor attached across the detector, eg a I M2 S 10 resistor. The steps taken in calibrating the admittance measuring system are as follows: All calculations and measurements are made in terms of complex (vector) values.

A known complex voltage 62 is applied to the detector with no other impedance attached to the detector 12 and the detector voltage produced is measured. This measurement provides a zero point value for the calculations.

The known impedance is then connected across the detector 12 and the known complex voltage is again applied to the detector 12. The voltage produced from the detector 12 is then measured, thus providing a second reference point for calibration.

The calibration factor Cal for the admittance measurement can now be calculated using the formula: 1000 Cal 1000 (Yr Yo) where Y, is the complex admittance signal voltage at 62 divided by the complex admittance signal voltage at 12 produced with the 1 M! resistor is connected across the detector, Y, is the complex admittance signal voltage at 62 divided by the complex admittance signal voltage at 12 measured when no connection is made across the detector, and 1000 is the admittance of the I MQ resistor in nS.

Once the Cal value is determined as above, the following formula can be used to calculate the admittance Y of the detector terminal: Y Cal wherein equals the complex admittance signal voltage at 62 divided by the complex admittance signal voltage at 12 and the other quantities are as defined above.

The real part of the admittance measurement signal Y calculated above is a measure of the conductance at the detector and can be used to indicate whether the probe's insulation has been bridged by conductive material and current is either leaking to. or from, the conduit.

The imaginary part of the admittance Y can be used as a measurement of the susceptance of the detector or multiplied by 2nt.(frequency of the admittance measurement signal) to give a capacitance measurement. The capacitance measurement can be used to determine if the probe's geometry has changed. for example by build up on the probe, breakage or severe bending or the probe. or loss of connection to the probe.

IN In order to allow emission monitoring and admittance measurement to be conducted simultaneously. It is necessary for the input amplifier 54 to have a low input impedance at lower frequencies, for example less than kQ at I Hz and a higher input impendence at an input frequency equal to the admittance measuring frequency such as 5 MD at 40kHz. These two preferred characteristics sound contradictory, however the high input impedance at 00 14~) admittance excitation frequency and low impendence over the measurement pass band can be achieved by using an integrator as shown in figure 9 by reference numeral 54.

The output of integrator 54 is fed back to the detector via high value resistors 66, 68. At high frequency the integrators gain approaches zero so the input impedance of the detector is defined only by the circuit resistors whereas, at low frequency the input impedance is defined by the following equation: R Fedback 1 Gain Integr,,r where RFeedhaiik is the value of the feedback resistor and GainIntegra ir is the integrator gain. It should be noted that the integrator gain is very high at these frequencies and is negative in value, meaning that the quantity 1 Gainintegra,,, is a large positive value, giving a low impedance at low frequency.

The admittance measuring signal is applied to the detector continuously during normal operation thereby allowing simultaneous measurements of particulate emission and detector parameters. The processing of the admittance signal is undertaken by a separate admittance measurement unit 64.

It should be noted that where in the specification or claims the terms "comprised" or "comprising" are used those terms should be interpreted inclusively rather than exclusively.

It will be understood that the invention disclosed and defined herein extends to all alternative combinations of two or more of the individual features mentioned or evident from the text or drawings. All of these different combinations constitute various alternative aspects of the invention.

The foregoing describes embodiments of the present invention, and modifications obvious to those skilled in the art can be made thereto, without departing from the scope of the present invention.

Claims (6)

1. A method of improving the linearity of a probe signal produced by a triboelectric particle detector including the steps of: 00 determining the instantaneous number of particles detected by said particle detector; Sassociating said instantaneous number of particles detected by said particle detector with S a value on a characteristic curve; using said associated value to calculate the particle density from the detected probe signal.

2. A method as claimed in claim 1 wherein the step of calculating the instantaneous number of particles detected by said particle detector includes the additional steps of: determining the numerical density of particles in the gas flow; estimating the volume over which a particle will produce a detectable signal by the particle detector; and dividing the numerical density of particles in the gas flow by the estimated volume over which a particle will produce a detectable signal by the particle detector to obtain the instantaneous number of particles detected by said particle detector.

3. A method as claimed in claim 1 wherein said characteristic curve is determined empirically for said particle detector by a method including the steps of: maintaining all environmental and particle detector conditions substantially fixed; determining the probe signal at a first known particle density; varying said particle density in a controlled manner until said particle density is equal to a second known particle density, and while varying said particle density; and 004919654 21 determining said probe signal for the range of particle densities between said first and second known particle densities.

4. A method as claimed in claim 1 wherein the characteristic curve is substantially a square law for high particle densities. 00 5

5. A method as claimed in claim 1 wherein the characteristic curve is substantially linear for low particle densities.

6. A method of improving the linearity of a probe signal produced by a triboelectric particle detector, substantially as herein described with reference to the accompanying drawings. Dated 6 September 2005 Freehills Patent Trade Mark Attorneys Patent Trade Mark Attorneys for the Applicant: Goyen Controls Co Pty Ltd