Addressability in particle detection

专利摘要:
A method of determining at least one point of entry of smoke into a smoke detection system, the system having a sampling pipe network including at least one sampling pipe and a plurality of sampling inlets through which an air sample can enter the at least one sampling pipe of the 5 smoke detection system for analysis by a particle detector, said method including: determining a volume of sample air that has passed through at least part of the smoke detection system since a predetermined event or a value corresponding to said volume; and determining through which sampling inlet of the plurality of sampling inlets the smoke entered the smoke detection system based, at least in part, on the determined volume or value. Systems for implementing such a 0 method and related methods are also described. Doc. 2005127028 16 i 22 +- 21
公开号:AU2013200353A1
申请号:U2013200353
申请日:2013-01-21
公开日:2014-05-01
发明作者:Kemal Ajay；Brian Alexander；Karl Boettger；Ron Knox；Peter Mundy；Thor NORTH；Rajiv Singh
申请人:Xtralis Technologies Ltd Bahamas；
IPC主号:G01F1-66

专利说明:
P/00/01 I Regulation 3.2 AUSTRALIA Patents Act 1990 COMPLETE SPECIFICATION STANDARD PATENT Invention Title: Addressability in particle detection The following statement is a full description of this invention, including the best method of performing it known to us: 2 Addressability in particle detection Field of the invention The present invention relates to particle detection. For illustrative purposes only, the preferred embodiment of the present invention will be described in relation to a smoke detection system, 5 but the invention should not be considered to be limited to that exemplary use. Background of the invention Air sampling or aspirated smoke detection systems operate by drawing air samples through a sampling network, to a central high sensitivity particle detector. The sampling network typically includes one or more sample pipes with a number of air sample inlets in the form of sampling 0 holes or sampling points located along the length of the pipe(s). In such an arrangement, a single detector may be fed with air originating from many distinct geographical locations at which the air sample inlets are located. Thus a single such detector can monitor for the presence of smoke at many distinct locations simultaneously. One recognised difficulty with air sampling systems as described above is that they do not 5 identify through which air inlet smoke enters the system. If the air inlet is known, the geographical location of the source of the smoke may be inferred. This allows investigation of the likely site of the fire including allowing a person to be directed to the location of the smoke, so that they may investigate and possibly intervene and prevent further growth of the fire, or shut down equipment in the area. Alternatively, an appropriate fire suppression system may be 20 deployed in a localised way, limiting damage caused by the system, as well as expense. There have been attempts to provide air sampling particle detection systems capable of determining the geographical location at which smoke is detected, for example Jax, 'Method and Device for locating accumulations of pollutants', U.S. 5,708,218 and Hekatron Vertriebs GmbH, 'Verfahren und Vorrichtung zur Erkennung eines Brandes', EP 1811478. 25 Each of these systems measures the elapsed time between two instants at which measurements are made to infer where along sampling pipe (i.e. through which sample inlet) the detected smoke entered the system. However, this inferential process is often unreliable.
3 The Jax system measures the elapsed time between detection of a first smoke level, and a second smoke level. The time between detection of a first, lower level of smoke, and a second, higher level of smoke indicates the distance along the collection line at which smoke entered the system. However, this process may be inaccurate. For example, systems employing this 5 approach rely upon the actual level of smoke detected at the first point of entry remaining approximately constant for the period of time beginning from the point at which smoke is first detected until the contribution from the second point of entry can be reliably detected. More specifically, an increase in smoke level, such as that caused by a fire of growing size, may result in an inaccurate estimate of the geographical location from which air has been drawn. 0 In Hekatron, a first air-sampling detection unit detects the presence of smoke. Responsive to detection of smoke, a second air-sampling detection unit is engaged, the air sampling unit drawing air along the pipe network. The time elapsed between initial detection by the first air sampling unit and detection by the second air-sampling unit is measured. Ideally, the time elapsed indicates the location from which smoke filled air has been drawn. To ensure accuracy, 5 such a system requires the aspiration system to operate in a highly consistent manner, each time it is operated. However, this is difficult to achieve as various features influence the operation of the fall, e.g. degradation of the aspiration system over time and variations in operational and environmental conditions e.g. air density, or the constriction of sampling points by dirt over time, will change the airflow characteristics within the system, and make the .0 inference of the smoke address based on elapsed time potentially unreliable. In some schemes, airflow may be temporarily reversed, introducing clean air to the sampling network, before redrawing air for detection. The idea in such schemes is to flush substantially all smoke particles from the system, before redrawing air through the sampling network and measuring the delay before detecting smoke. In theory, a longer delay indicates that the 25 particles entered the sampling network at a point farther from the detector. However, these schemes suffer a drawback in that during the phase that clean air is introduced to the sampling network, smoke particles within the monitored environment may be displaced in the area surrounding the air inlets, since clean air is being expelled from the inlets. When air is subsequently drawn through the system, there may be an additional delay before smoke 30 particles are once again drawn into the inlet.
4 It is therefore an object of the present invention to provide a particle detection system that addresses at least some of the aforementioned disadvantages. An alternative object of the invention is to provide the public with a useful choice over known products. Reference to any prior art in the specification is not, and should not be taken as, an 5 acknowledgment or any form of suggestion that this prior art forms part of the common general knowledge in Australia or any other jurisdiction or that this prior art could reasonably be expected to be ascertained, understood and regarded as relevant by a person skilled in the art. Summary of the invention In a first aspect of the present invention there is provided a method of determining at least one 0 point of entry of smoke into a smoke detection system, the system having a sampling pipe network including at least one sampling pipe and a plurality of sampling inlets through which an air sample can enter the at least one sampling pipe of the smoke detection system for analysis by a particle detector. The method includes: determining a volume of sample air that. has passed through at least part of the smoke detection system since a predetermined event or a 5 value corresponding to said volume; and determining through which sampling inlet of the plurality of sampling inlets the smoke entered the smoke detection system based, at least in part, on the determined volume or value. The predetermined event could be, for example, a smoke detection event; or a change in an air sample flow characteristic in the smoke detection system. 20 In some embodiments the method includes continuously determining a flow rate of the air sample passing through at least part of smoke detection system. Alternatively the method includes commencing determination of the volume of sample air or a related value upon the occurrence of the predetermined event. The volume of the air sample that has passed through at least part of smoke detection network 25 or a related value can be determined by accumulating a flow rate measurement over time. The rate of flow measurement is preferably a volumetric flow rate measurement. Most preferably the he flow rate measurement is determined using an ultrasonic flow sensor. The step of determining a volume of sample air that has passed through at least part of the smoke detection system since a predetermined event or a value corresponding to said volume, 30 can include determining any one of more of: a mass; a length; a pressure; a temperature, a 5 second volume; or an accumulated count of volume-related events, or other parameter that that relates to a volume of sample air that has passed through at least part of the smoke detection system since the predetermined event. The method can include collecting all or a proportion of the sample air that has passed 5 through at least part of the smoke detection system since the predetermined event. The method can further include changing an air sample flow characteristic in response to a first smoke detection event. For example, changing an air sample flow characteristic in the smoke detection system can include one or more of the following: * opening a valve; 0 * closing a valve; e changing a direction of an air sample flow in at least part of the smoke detection system; * changing a rate of air sample flow in at least part of the smoke detection system; * starting an aspiration system; and * stopping an aspiration system. 15 In a second aspect of the present invention, there is provided an apparatus for determining at least one point of entry of smoke into a smoke detection system of the type having a particle detector in fluid communication with an air sampling network, the air sampling network having at least one sampling pipe and a plurality of sampling inlets through which an air sample can enter the at least one sampling pipe of the smoke detection system for analysis by the particle 20 detector, and an aspirator for drawing the air sample through the air sampling network to the detector. The apparatus includes: means for determining a volume of sample air that has passed through at least part of the smoke detection system since a predetermined event or a value corresponding to said volume; and means for identifying at least one point of entry of particles into the sampling network based on the detected volume or value.
6 The apparatus preferably identifies one or more of said points of entry by reference to one or more corresponding sampling inlets through which smoke determined to have entered the system. The means for determining a volume of sample that has passed through at least part of the 5 particle detection system, or value related to said volume, preferably includes a flow sensor. Most preferably the flow sensor comprises an ultrasonic flow sensor. The apparatus is preferably configured to perform a method in accordance with the first aspect of the present invention. In a third aspect of the present invention, there is provided a smoke detector including a particle 0 detection chamber to detect particles in an air sample, an inlet to receive an air sample from an air sampling network, said the sampling network having at least one sampling pipe and a plurality of sampling inlets through which a sample can enter the at least one sampling pipe for analysis by the particle detection chamber, and an aspirator for drawing the sample through the air sampling network to the detector, the detector further including a processor configured to: 5 identify at least one point of entry of smoke into the sampling network based, at least in part, on a volume of sample air that has passed through at least part of the smoke detector or sampling network since a predetermined event, or a value corresponding to said volume. The smoke detector can include a flow sensor, e.g. an ultrasonic flow sensor, configured to detect rate of flow of sample air passing through at least a part of the smoke detector. 20 The processor is preferably configured to cause the smoke detector to perform a method in accordance with the first aspect of the present invention. Also disclosed herein is a method of determining the point of entry of particles into a particle detection system, said particle detection system including a particle detector and a sampling network in fluid communication with the particle detector, the sampling network including a 25 plurality of inlets through which a fluid is drawn, the particle detection system further including means for drawing fluid through the sampling network to the detector. The method includes: comparing a first particle detection profile to a second particle detection profile; determining an offset between the particle detection profiles at which the profiles match to a predetermined degree; and, determining a location of entry of particles into the detection system on the basis 30 of that offset.
7 In some embodiments, the offset is a time offset. In other embodiments, the offset is a volume offset. In some embodiments, the comparison involves calculation of a cross-correlation between particle detection profiles. 5 In some embodiments, a maximum value of the calculated cross correlation is determined, and an offset between particle detection profiles corresponding to the maximum value is determined. In some embodiments, the calculated cross correlation function is determined and compared to a predetermined value. Preferably, the fluid is air, and the means for drawing fluid through the sampling network to the 0 detector is an aspirator. One embodiment includes determining that at least a first predetermined particle detection criteria has been met on the basis of a first particle detection profile being a comparison of the first and second particle detection particles. The method can include continuously storing a first and/or second particle detection profile. 5 Alternatively one of the profiles may be stored only after at least one predetermined criteria has been fulfilled. The method can include changing an air flow characteristic in at least part of the particle detection system prior to beginning a comparison of the first and second particle detection profiles. 20 In one form the step of changing an air flow characteristic in the particle detection system includes one or more of the following: " opening a valve; * closing a valve; " changing a direction of an air flow in at least part of the particle detection system; 8 e changing a rate of air flow in at least part of the particle detection system; * starting an aspiration system; and " stopping an aspiration system. Further disclosed herein is an apparatus for determining the point of entry of particles into a 5 particle detection system of the type having a particle detector in fluid communication with an air sampling network, the air sampling network having a plurality of inlets through which air may enter the air sampling network, and an aspirator for drawing air through the air sampling network to the detector, the apparatus including means for determining a volume of air passing through at least a part of the particle detection system, said apparatus including: means for 0 receiving a signal representative of the volume of air passing through at least a part of the particle detection system; means for determining a location in the air sampling network at which air carrying particles entered the network on the basis of the determined volume. Also disclosed herein is a device for determining the point of entry of particles into a particle detection system through one or more of a plurality of air inlets. The device includes means for 5 determining a volume of air flowing through at least part of the particle detection system and means for determining a point of entry of the particles based upon the measured volume. Preferably, the apparatus for determining the point of entry of particles into the particle detection system identifies the source of particles by reference to at least one inlet through which particles are likely to have entered. 20 Further preferably, the apparatus for determining the point of entry of particles into the detection system identifies the source of particles by providing an indication of the distance of along the sampling network at which particles entered the air sampling network. Further disclosed herein is a method of determining the point of entry of particles into a particle detection system having a sampling pipe network with a plurality of sampling points through 25 which particles can enter the particle detection system. The method includes, determining the volume of air passing through at least part of particle detection system and determining through which sampling hole of the plurality of sampling points the particles entered the particle detection system.
9 The method can include, detecting a first particle detection event and a second particle detection event, and measuring the volume of air passing through at least part of particle detection network between the particle detection events. The method can include continuously measuring the volume of air passing through at least part 5 of particle detection network. Alternatively the method can include activating the volume measurement upon the occurrence of a predetermined condition. The volume of air passing through at least part of particle detection network is preferably measured by summing a rate of flow measurement over time. Preferably the rate of flow measurement is a volumetric flow rate measurement. Most preferably it is determined using an 0 ultrasonic flow sensor. Further disclosed herein is a particle detection system including a particle detector, a sampling network in fluid communication with the particle detector, and means for drawing fluid through the sampling network to the detector. The sampling network includes a plurality of inlets, the inlets being arranged into a plurality of location groups. Each location group has. an address 5 defined by the presence or absence of an inlet connected to each of a plurality of sampling pipes. The particle detector is configured to draw air along each sampling pipe and in the event that smoke is detected, determine the address of the location group through which particles entered the detector based upon both the presence and absence of particles in each of the sampling pipes. .0 Also disclosed herein is a method of determining a single point of entry of particles into a particle detection system. The particle detection system includes at least one particle detector, a sampling network in fluid communication with a or the particle detector, and means for drawing fluid through the sampling network to a or the detector. The sampling network includes a plurality of sample communication paths along which a sample can be drawn and in which the 25 presence of particles can be independently detected by at least one of the detectors, wherein each sample communication path includes at least one sample inlet. Each of said inlets further belongs to one of a plurality of location groups defined by the physical location of the inlet. The particle detection system being configured to determine whether particles are been detected on an air sample from each sample communication path. The method includes: 30 determining a location group of inlets at which particles entered into the particle detection system uniquely on the basis of whether particles have or have not been detected on each sample communication path.
10 In one embodiment, the sampling network comprises a plurality of pipes that respectively correspond to a sample communication path, and the step of determining that particles have been detected at a location group comprises determining that particles have or have not been detected in fluid drawn through each of the plurality of pipes. 5 Further disclosed herein is an apparatus for determining the point of entry of particles into a particle detection system of the type having at least one particle detector in fluid communication with a sampling network, and aspiration means for drawing fluid through the sampling network to the or a particle detector, the sampling network including a plurality of sample communication paths in which particles can be separately detected. The sampling network includes a plurality 0 of sample inlets, each inlet being a member of a location group at one of a plurality of physical locations; the apparatus further including means for determining a location at which particles are present on the basis of whether particles have or have not been detected on each sample communication path. Also disclosed herein is a method in a particle detection system having: 5 at least one particle detector; and a sampling system including a sampling pipe with a plurality of sampling inlets, said sampling system being arranged to convey a sample to be analysed from an environment surrounding the sampling inlet via the sampling pipe to the at least one particle detector; a flow inducer arranged to cause an air sample to flow in the sampling system to the at 20 least one particle detector; the method including: measuring a first particle concentration in a sample arriving form the sampling system; varying a sampling parameter at a subset of the sampling inlets; measuring a second particle concentration in a sample arriving form the sampling 25 system; 11 measuring a particle concentration in a sample arriving form the sampling system; on the basis of the first and second particle concentrations and the varied sampling parameter. The sampling parameter that is varied can be flow rate through the first subset of sampling inlets. The variation can be triggered by opening or closing valves or using a fan or other flow 5 inducer to increase (or decrease) flow through the subset of sampling inlets. In this case the varied sampling parameter used to determine the measuring a particle concentration in a sample arriving form the sampling system can be a flow rate through the subset of sampling inlets. In some embodiments the sampling parameter that is varied is the particle concentration drawn 0 through the first subset of sampling inlets. The variation can be triggered by adjusting a filtering parameter applied to the first subset of sampling inlets, e.g. by interposing or removing a filter in the flow path of air entering through the sampling inlets. In this case the varied sampling parameter used to determine the measuring a particle concentration in a sample arriving form the sampling system can be a sample concentration the subset of sampling inlets. 5 In some embodiments the first subset of sampling inlets is the same as the second subset of sampling inlets. The first or second subsets if sampling inlets may include a plurality of inlets, or may be a single inlet. Also disclosed herein is a method for detecting contaminant(s) in air samples from a plurality of air intake paths, the method including: 20 varying the flow balance between the multiple paths by increasing or partially reducing the flow in one or more of the plurality of air intake paths to create a plurality of different flow patterns; measuring the contaminant level of the combined air intake paths for each of the plurality of different flow patterns; and 25 determining the contaminant level of each air intake path by using known, predetermined or measured values of flow rate in each air intake path for each of the plurality of different flow patterns, 12 wherein the number of different flow patterns created and the number of contaminant level measurements taken are sufficient to determine the contaminant level in each air intake path. Varying the flow balance is preferably achieved over the plurality of different flow patterns by 5 partial flow reduction in each of the air intake paths, in turn. In other words, if there are four air intake paths, a first subset of the air intake paths (e.g. three paths) are partially closed while the remaining intake path(s) remain open while the contaminant level is measured. Next, that first subset air intake path is reopened and a second different subset of air intake paths is partially closed while the remaining air intake path(s) remain open and a second measure of the 0 contaminant level is made. This is continued until four different flow patterns are created while four measurements of the contaminant level are taken. The partial reduction in flow is preferably achieved by partially closing valves in the air intake paths. So, each valve is partially closed in turn while the other valves remain open. In this arrangement, the flow rate through each air intake path may not be known. Therefore, it may 5 be necessary to measure the flow rate in each air intake path, for each of the plurality of different flow patterns. In an alternative form, the step of varying the flow balance may be achieved by having rnoveable baffles within the air intake paths. For example, the moveable baffles may be in the form of rotatable discs movable to a number of selectable positions. The discs have openings a0 which, depending upon the selected position, create a predetermined flow rate. Thus, in this arrangement, flow rate measurements may not be required. In a third alternative method of varying the flow balance, each air intake path may be vented in turn while the other pipes remain unvented. Compared to the other two methods described above, this will result in an increase in air flow through each vented air intake path in turn and 25 may also affect the flow rate in the other air intake paths. In a preferred form, there are as many flow patterns created as there are air intake paths. Given that there are as many measurements of contaminant level as there are flow patterns, this means the number of measurements of contaminant level equal the number of flow paths too. This will provide enough information to determine the contaminant level in each air intake 13 path, provided the flow rate in each air intake path is also known/predetermined or measured for each flow pattern. In some arrangements, the flow rate is measured in each air intake path. This is preferably achieved by a flow rate sensor having a reasonably high degree of accuracy. In a most 5 preferred form, flow rate is measured by ultrasonic flow rate sensors, one in each air intake path. Preferably, with the measured contaminant levels for each flow pattern and the known/predetermined or measured flow rates in each path for each flow pattern, a series of equations may be solved as follows: 0 C 1 = X 1 Fj 1 /(Fj 1 + F 12 + ... F 1 n) + X 2
F
12 /(Fj 1 + F 12 + ... F 1 n) ... + Xn F 1 n/(Fij + F 1 2 + ... F 1 )
C
2 = X 1
F
21
/(F
21 + F 22 + ... F 2 n) + X 2
F
22
/(F
21 + F 22 + ... F 2 n) + ... Xn F 2 n/(F 21 + F 22 + ... F 2 n) Cn = X 1 Fn 1 /(Fn 1 + Fn2 + ... Fn) + X 2 Fn 2 /(Fn 1 + Fn2 + ... Fn) + ... XnFn/(Fn 1 + Fn2 + ... Fn) where .5 X 1 ... Xn = concentration in air intake paths 1 to n
C
1 ... C 2 = measured contaminant level of the combined air intake paths
F,
1 ... Fn1 = flow rate in pipe 1 for flow patterns 1 to n
F
1 2 ... Fn 2 = flow rate in pipe 2 for flow patterns 1 to n Fin... Fn = flow rate in pipe n for flow patterns 1 to n 20 In a preferred form, the air intake paths may be in the form of air sampling pipes. Each air sampling pipe may feed into a respective intake port on a detector unit. The flows may be merged in a manifold, in the detector unit prior to being fed to the detector.
14 The step of measuring, whether for the contaminant level or the flow rate may involve multiple readings from which an average is taken. Alternatively, any other statistical calculation may be made to determine the central tendency of the multiple readings. Also disclosed herein is a sensing system for detecting contaminants in air samples from a 5 plurality of air intake paths, the system including: a control system for controlling flow control means in each of the air intake paths to increase or partially reduce the flow in one or more of the air intake paths to create a plurality of different flow patterns; a detector to measure the contaminant level of the combined air intake paths, the control 0 system controlling the detector to measure the contaminant level for each of the plurality of different flow patterns; the control system being further operable to determine the contaminant level of each air intake path using known, predetermined or measured values of flow rate in each air intake path for each of the plurality of different flow patterns; and 5 the control system being operable to create a sufficient number of different flow patterns and to control the detector to take a sufficient number of measurements to determine the contaminant level of each air intake path. The sensing system may be in the form of a sensing unit which includes air intake ports corresponding to the number of air intake paths. Each air intake port may be coupled to a 20 respective sampling pipe. Each of the flow control means may be disposed within the sensing unit or alternatively may be disposed in a respective sampling pipe. Preferably, the control system is able to control the measurement of flow rate. Also disclosed herein is a sampling point for an environmental sampling system of the type having a at least one elongate sampling duct defined by a peripheral wall and having plurality of 25 sampling inlets located along the duct's length and extending through the wall to allow the ingress of a sample, said environmental sampling system being configured to draw a sample from the environment through the sampling inlets into the duct and to convey the samples 15 through the duct to an analysis device, the sampling point including a sample injection inlet extending into an interior of the duct inward of the peripheral wall thereof. The sample injection inlet can include a pipe extending through the peripheral wall of the duct. Most preferably the pipe has an outlet at or near the centre of the duct, away from the 5 peripheral wall of the duct. The sample injection inlet can have its outlet facing in a downstream direction of flow in the duct. In a preferred form the sample injection inlet is an L-shaped pipe, with a first inlet end for drawing a sample from the environment and a second, outlet end located within the duct and having an outlet facing in a downstream direction of flow in the duct. Also disclosed is a method 0 in an environmental sampling system of the type having a at least one elongate sampling duct defined by a peripheral wall and having plurality of sampling inlets located along the duct's length and extending through the wall to allow the ingress of a sample, said environmental sampling system being configured to draw a sample from the environment through the sampling inlets into the duct and to convey the samples through the duct to an analysis device, the 5 method including: providing a structure to ameliorate diffusion of at least a front of a discrete sample portion, along the duct, as the sample portion travels down the duct. The structure can be a sampling point including a sample injection inlet extending into an interior of the duct as described above. The structure could also be a structure that creates 20 turbulence within the duct configured to prevent laminar flow within the duct in use. For example, the structure could be a contoured or textured wall of the duct; a turbulator; a passive or active rotating element or the like. Also disclosed herein is a sampling system for an environmental analysis system, said sample system including at least one elongate sampling duct defined by a peripheral wall and having 25 plurality of sampling inlets located along the duct's length and extending through the wall to allow the ingress of a sample into the duct, said environmental sampling system being configured to draw a sample from the environment through the sampling inlets into the duct and to convey the samples through the duct to environmental analysis system, the sampling system further including means to ameliorate diffusion of at least a front of a discrete sample portion, 30 along the duct, as the sample portion travels down the duct. The structure can be a sampling 16 point including a sample injection inlet extending into an interior of the duct as described above. The structure could also be a structure that creates turbulence within the duct configured to prevent laminar flow within the duct in use. For example, the structure could be a contoured or textured wall of the duct; a turbulator; a passive or active rotating element or the like. 5 The structure could extend substantially the whole length of the duct, or be localised, e.g. at or near, one or all, of the sampling inlets. Further disclosed herein is a method in an environmental sampling system of the type having at least one elongate sampling duct having plurality of sampling inlets located in series along the duct's length to allow the ingress of a sample from the environment, said environmental 0 sampling system being configured to draw a sample from the environment through the sampling inlets into the duct and to convey the samples through the duct to an analysis device, the method including: changing the airflow characteristic in the duct to alter a local sample concentration at or near at least one particular sampling inlet to increase the local sample concentration towards the sample concentration in the atmosphere surrounding the particular 5 sampling inlet. Changing the airflow characteristic can include stopping or reversing a direction of flow in the duct to so that a portion of a sample adjacent the particular sampling inlet is expelled from the sample inlet. The method then includes drawing an additional sample from the environment via the particular sample inlet. The steps of stopping or reversing a direction of flow in the duct to so 20 that a portion of a sample adjacent the particular sampling inlet is expelled from the sample inlet, and drawing an additional sample from the environment via the particular sample inlet can be repeated one or more times. The method can include oscillating the direction of flow in the duct such that a repeated process of expulsion and re-sampling the environment occurs. 25 The method can then include transporting the contents of the duct to the analysis device. This transportation is preferably performed with minimal dilution of the sample within the duct, or mixing between longitudinally positioned portions of the sample of the duct. For example the method can include; one or more of the following: closing one or more of the sampling inlets prior to transportation, 17 opening duct at an upstream position to provide a low flow impedance; blowing the sample along the duct from an upstream position. An environmental sampling system of the type having at least one elongate sampling duct having at least one sampling inlet located along the duct's length to allow the ingress of a 5 sample from the environment, said environmental sampling system being configured to draw a sample from the environment through the or each sampling inlet into the duct and to convey the samples through the duct to an analysis device, The system further including sample amplification arrangement to ameliorate dilution of the sample by air flow in the duct. The sample amplification arrangement could include a device to reverse flow direction in at 10 least a portion of the duct. The device to reverse flow direction is preferably arranged to cause multiple reversals of flow direction to promote mixing of an air sample at or adjacent a sampling inlet. The device to reverse flow could be, for example, a reversible fan, bellows, reciprocating piston, vibrating membrane, or the like. Also disclosed herein is an environmental sampling system of the type having at least one 5 elongate sampling duct having plurality of sampling inlets located in series along the duct's length to allow the ingress of a sample from the environment that is configured to perform the above method. The environmental sampling system can include one or more of the following: One or more valves to control flow along the duct and/or through one or more of the sampling inlets; 20 fans, blowers or other flow inducing means to control flow along the duct and/or through one or more of the sampling inlets. A particle detection system, and preferably a smoke detection system, is also provided that includes an environmental sampling system of the above type to deliver air samples for analysis from a plurality of locations. 25 Also disclosed herein is a method in an environmental sampling system of the type having at least one elongate sampling duct having plurality of sampling inlets located along the or each duct's length to allow the ingress of a sample from the environment, said environmental sampling system being configured to draw a sample from the environment through the sampling 18 inlets into the duct and to convey the samples through the duct to an analysis device to detect the presence of a threat substance in the sample, the method including: operating in a detection mode in which the presence and or concentration of the threat substance is being monitored, and in the event at least one criterion is met, the system 5 performs the step of: operating in a localisation mode to determine which of the sampling inlets the threat substance entered the system. The method can include operating in a training mode to characterise a sample flow through the at least one sampling duct to the analysis device so as to enable determination of which of the l0 sampling inlets the threat substance entered the system in the localisation mode. The localisation mode can include a sample amplification phase transportation phase. Brief description of the drawings Illustrative embodiments of the invention will now be described by way of a non-limiting example with reference to the accompanying figures. In the figures: 15 Figure 1 shows a particle detection system including an air sampling network; Figure 2 shows a particle detection system employing two particle detectors to enable determination of the location at which smoke enters an air sampling network; Figure 3 shows a particle detection system employing a single particle detector coupled to an air sampling network having two branches separated by a valve; 20 Figure 4 shows a particle detection system employing two particle detectors coupled to a single air sampling pipeline; Figures 5 and 6 graphically illustrate a timing of events as measured at respective detectors (or branches) of a particle detection system; 19 Figure 7 illustrates another embodiment of a particle detection system that is used to determine a location particles entering the system; Figure 8 illustrates a particle detection system including a sampling system including a plurality of valves, for altering a sampling parameter of the sampling system [to implement an 5 embodiment of one aspect of the invention]; Figure 9A illustrates a particle detection system including a sampling system including a plurality of filters which are configured to alter a sampling parameter the sampling system [to implement an embodiment of one aspect of the invention]; Figure 9B illustrates a filter and valve arrangement used in the system of figure 9A; o Figure 10A is a schematic diagram of a particle detection system according to a preferred embodiment of the present invention; Figure 1GB is a schematic diagram of a portion of the particle detection system of Figure 1 0A; Figure 10C is a schematic view of the portion of the particle detection system as per Figure 10B, except with one of the valves in a partially closed position; and 5 Figure 10D is a schematic view of the portion as per Figure 10C, except that one of the other valves is partially closed; Figure 1IA illustrates a particle detection system; Figure 11 B is a graph illustrating diffusion of a front of a sample portion as the sample portion travels down a duct; 20 Figure 11 C illustrates a flow speed profile within the sample duct of figure 11 A; Figure 12 illustrates 3 sampling points according to different embodiments of the present invention, that may ameliorate the affect of the diffusion illustrated in figure 11 B; Figures 13A to 13D are examples of turbulators that may ameliorate the affect of the diffusion illustrated in figure 11 B; 20 Figure 14 illustrates a particle detection system including an air sampling network that is connected to bellows that can be used to oscillate the direction of sample flow within the air sampling duct to counteract sample dilution by other sampling inlets within the particle detection system; 5 Figures 14A to 14E illustrate an exemplary system that uses a vibrating membrane to perform sample amplification in a manner analogous to that of Figure 14; Figure 15 illustrates a particle detection system including an air sampling system that has an upstream fan that can be used to counteract sample dilution by other sampling inlets within the particle detection system. 0 Figure 16 illustrates a particle detection system having an air sampling system including a valve upstream of the sampling inlets that can be used to open the end of the sampling duct to enhance transport of sample in the duct to the particle detector for analysis; Figure 17 illustrates a variant of the system of figures 14A to 14E; Figure 18 illustrates a particle detection system including an air sampling network that has a 5 sample amplification arrangement comprising a plurality vibrating membranes; and Figure 19 illustrates another particle detection system including an air sampling network with branched sampling pipes and which has a sample amplification arrangement comprising a plurality vibrating membranes. Detailed description of the embodiments 20 Figure 1 shows a particle detection system including a particle detector 11 in fluid communication with a sampling network 28. The sampling network includes a plurality of inlets 29 through which air is drawn. An aspirator 16 draws air into the sampling network 28 through inlet 21 and along into a particle detection chamber 14. Air sample exits the detection system through outlet 22. 25 The detector includes a flow sensor 24. In a preferred embodiment of the present invention, an ultrasonic flow sensor as described in WO 2004/102499 is employed. This sensor enables volumetric flow measurements to be made. The flow sensor 24 provides an indication of the 21 volume of air flowing into the particle detector 10 from the sampling network 28 per unit time. The output of the flow sensor 24 may be used to infer, for example, when flow faults e.g. a blockage of the sampling network 28 or reduced aspirator performance, has occurred. The system 10 also includes a controller 40 for determining the level of particles in the air 5 sample based on the detector's 14 output and apply alarm and fault logic to the detector output alert a user to the presence of particles and the operating state of the system. A typical installation of a Vesda or ICAM smoke detector, from Xtralis Pty Ltd. would be an example of a system of this type. Such a detection system can be applied in an embodiment of the present invention to 0 additionally determine the point of entry of particles into the air sampling network 28. Figure 2 shows two particle detectors 202 and 204, each particle detector being of the type illustrated in Figure 1. Each detector is connected to a respective pipe of sampling network 203 and 205 respectively. The sampling networks 203 and 205 are effectively parallel and configured to monitor the same area. Each detector is also connected to a control unit 207, 15 containing a microcontroller 209. Pipe 203 has a plurality of air inlets 206-216. Similarly, pipe 205 has a plurality of air inlets 218-230. Each air inlet from pipe 203 can be paired with an inlet from its parallel air pipe 205. At the time of installation, each inlet from pipe 203 is positioned to be close to a corresponding inlet from pipe 205. The inlets are therefore arranged in pairs. For example, air inlet 206 of pipe 203 and air inlet 218 of pipe 205 are together labelled air sampling 20 inlet pair 232, because air inlet 206 and air inlet 218 are placed in close physical proximity. For example each pair of inlets may be located in the same room of a row of offices, or even be attached to a common sampling point. In normal operation, the aspirator of particle detector 202 draws air pipe 203. The aspirator of particle detector 204 draws air through pipe 205. As each particle detector draws air, the 25 scattered light or "smoke level" is measured, and reported to the control unit 207. The microcontroller 209 of the control unit 207 stores the reported smoke levels in its internal memory. In the event that smoke enters the air sampling network at air sampling inlet pair 232, the distance that smoke must travel to reach particle detector 202 from air inlet 206 is much smaller 30 than the distance that smoke must travel to reach particle detector 204 from air inlet 218.
22 Accordingly, particle detector 202 will register an increased smoke level due to smoke entering air sampling inlet pair 232 before particle detector 204. When the detected smoke level of one of the detectors 202,204, say particle detector 202, surpasses a predetermined threshold (which may also be an alarm threshold or not) , the 5 microcontroller begins to monitor the volume of air that has been drawn through one or both of the detectors. Because the smoke introduced at air inlet 218 must travel along the length of sampling pipe 205 before it can be detected at detector 204. After the particle detector 204 has drawn some volume of air, particle detector 204 will record an increased smoke level similar to that seen by particle detector 202. When this increased smoke level is recorded, the 10 microcontroller 209 finishes monitoring the volume of air that has been drawn through detector 204. This final volume can be used to determine the sampling hole through which the smoke entered the air sampling pipe. Because the flow sensor e.g. 24, outputs volumetric rate of flow, the volume of air passing through the detector is determined by integrating the output of the flow sensor over time. For 15 example, the flow rate may be output one or more times per second by the sensor. These volumes can be accumulated either in the detector itself or at the microcontroller 209 to determine the total volume of sample air that has flowed. The microcontroller 209 then uses the determined volume of air drawn by detector 204 to infer the sampling inlet pair through which the smoke particles were introduced. In one embodiment, 20 the microcontroller achieves this by consulting a lookup table such as the one below: Volume Air Inlet Pair -5L Pair 1 -3L Pair 2 -1L Pair 3 1L Pair 4 3L Pair 5 5L Pair 6 The lookup table contains measured volumes mapped back to a corresponding sampling hole pair. Each volume corresponds to the volume of air that is drawn through the second detector before particles are detected by it. The negative and positive values indicate which detector of 23 the pair 202 or 204 measure the volume. In this case a negative value indicates that the detector 202 measures volume. For example, the microcontroller 209 may measure a volume of 112 mL of air drawn through detector 204 in the time between a smoke detection event by detector 202 and a subsequent 5 detection event by detector 204. The row of the table that has a volume most closely corresponding to the volume is the fourth row, and corresponds to Pair 4. Pair 4, in turn, corresponds to air inlet pair 238. Had the measured volume instead been -1 12mL, the closest table row would have been the entry for -100mL, and Pair 3 (air inlet pair 236) would have been determined as the point at which smoke entered the system. 0 As will be appreciated, instead of measuring volume directly a value that corresponds to volume could be used in other embodiments of the present invention. For example the amount of air sample that has passed through the system can be determined by measuring a parameter other than volumetric flow rate, for example, if a mass flow sensor is present in the detector the output of such a sensor is able to be used in an embodiment of the present invention as it is related to 5 volume by a correction factor that corrects for the temperature or density of the fluid. Other physical parameters may also be used, including but not limited to as length, pressure or temperature or a count of volume-related events. For example, the speed of the sample flow can be measured (e.g. in ms 1 ) at location and accumulated (eg. summing or integration etc.) to determine an amount of air that has passed through the system in the form of a "length". 20 Volume could also be represent as a "length" by using the air sample (or known proportion of it) to displace a piston. The total displacement of the piston by the collected sample (or fixed proportion thereof) will represent a measure of the amount of air that has passed through the system, alternatively for a small cylinder size the a number of cycles of the piston could be counted to yield an numerical value corresponding to the volume of air sample that has passed 25 through the system. To give an example in which the physical parameter being used to determine an amount of air passing through the system is pressure or temperature, consider a system in which the air sample (or a known proportion of the air sample volume) is captured in a first chamber of a closed system, the actual volume V 1 (or pressure if volume if fixed) of this amount of air may 30 never be known. However if the temperature T 1 and pressure P 1 (or volume if pressure is fixed) of the captured sample is measured. The captured sample is then moved to a second camber 24 of known, volume V 2 and the new temperature T 2 and pressure P 2 are related to the initial volume by Boyle's law. By controlling one the either pressure or temperature to be held constant during the transfer of the sample (or sample portion) to the second chamber a temperature or pressure can be used as an amount that relates to volume of sample air that has 5 passed through the system. If a measurement of a value , such as mass, pressure, temperature and length, is used in place of volume, the look-up table may alternatively map those other physical parameters directly to the air inlet pair number, without having to undertake the intermediate step of calculating the volume. 0 Once the air inlet pair number has been determined, the air inlet pair number can then be communicated to a secondary device, such as a Fire Alarm Control Panel (FACP) or displayed to the user, to enable the localisation of the fire. The lookup table can be created during the commissioning of the system, for example, by introducing smoke to each sample inlet pair and measuring the volume of air drawn before 5 detection. As will be appreciated, if smoke has entered at sampling pair 232, there will be a very large volume of air drawn by detector 204 in the period after detection by detector 202 while detector 204 waits to detect the increased smoke level. Conversely, if smoke entered the system through sampling pair 242, detector 204 would detect an increased smoke level before detector 202, detector 202 drawing a very large volume of air while waiting to detect the 20 increased smoke level. If smoke were to enter the sampling network toward the middle, for example at sample pair 236, although detector 202 would detect an increased smoke level first, the volume of air drawn before detection by detector 204 would be relatively smaller than in either of the first cases, since by the time of detection by detector 202, smoke will have already been drawn a substantial distance toward detector 204. 25 A person skilled in the art will appreciate that in the present configuration, where the sampling pipe network length is large, and transport time of particles through the sampling network is large, it will be possible to detect the presence of smoke before determining the location of smoke. For example, in the event that smoke is introduced at sampling inlet pair 232 of Figure 2, smoke entering sampling hole 206 will quickly proceed to detector 202, and be detected. 30 Detector 202 can immediately raise an alarm, despite the fact that smoke has not yet been detected by detector 204. Accordingly, where regulations prescribe the time by which smoke 25 introduced to a sampling hole must be detected, this particular configuration is capable of detecting and reporting upon the presence of fire upon detection of smoke particles. Determination of the geographic location of the fire can then proceed in the manner previously described using a threshold level that is not an alarm level. 5 Accordingly, in a preferred form, the threshold used for determining an addressing event for each detector is higher than the lowest alarm (eg: a pre-alarm) threshold. A preferred embodiment waits until a higher level of particles is detected before attempting addressing. In one embodiment, instead of employing a lookup table, the volume offset is multiplied by a constant to determine the distance along the sampling network at which smoke particles 0 entered the system. In another embodiment, the volume offset is used as a variable in a function, which when evaluated, yields an estimate of the distance along the sampling network at which particles entered. In yet another embodiment, the volume offset is used as an index into a lookup table, the resulting lookup value being an estimate of the distance along the pipe. In preferred embodiments, the multiplicative constant, function, or lookup table described 5 immediately above is determined at the time of commissioning by introducing smoke to each sampling hole pair and measuring the resulting volume offset to generate calibration data. As a person skilled in the art will appreciate, it may be possible to infer results for a subset of sampling holes by introducing smoke to another subset of holes, and relying upon the known distribution of sampling pairs in the sampling network. 20 As a person skilled in the art will appreciate, modifications of the invention can be adapted to determine, for example, the spread of a fire. The information reported by the system may be a distance along the sampling network at which particles appear to have entered, although this distance may not correspond to a sampling inlet pair. The calculated distance or air inlet may be presented directly to an end user. The calculated 25 distance or air inlet may also be communicated to another system, such as a fire alarm control panel (FACP). Where a fire alarm control panel has been designed to receive data from a system of addressable point detectors rather than a single aspirated smoke detector having multiple sampling points, the present system may communicate the calculated distance or inlet to the fire alarm control panel in a way which mimics a system of addressable point detectors, 30 thereby utilising the FACPs understanding of geographic location of fires without actually utilising individual addressable point detectors.
26 Figure 3 illustrates an alternative embodiment of the invention that employs a single particle detector attached to an air sampling network comprising two pipes 303 and 305 and a valve 304. In normal operation, air is drawn through pipe 303. When smoke detector 202 detects smoke above a predetermined threshold, valve 304 is moved to obstruct pipe303, and to allow 5 air to flow through pipe 305, and the microcontroller 309 begins to record the volume of air drawn through detector 302. When smoke particles are detected by detector 302, microcontroller 309 finishes recording the volume of air drawn though detector 302. The volume of air passing through air sampling network 305 and into particle detector 302 prior to again detecting particles is then used to infer the point at which smoke particles enter pipe 305, 0 using any of the methods herein described. Figure 4 shows yet another approach which employs two particle detectors attached to a single air sampling network. Initially, smoke detector 402 operates and smoke detector 404 is inoperative. Smoke enters the system through air inlet 408. The smoke is drawn through the air sampling network, and detected by smoke detector 402. The determination of a smoke 5 detection event triggers smoke detector 402 to become inoperative, smoke detector 404 to become operative, and microcontroller 409 to begin recording the volume of air drawn through detector 404. The aspirator of smoke detector 404 draws air along air sampling network 403 in a direction opposite to the initial flow direction caused by the aspirator of smoke detector 402. If smoke enters only through a single air inlet 408, smoke detector 404 cannot detect smoke until '0 smoke from air inlet 408 reaches it. According to the present invention, the volume of air drawn by detector 404 after the initial detection by smoke detector 402 and up until subsequent detection of smoke by detector 404 is used to determine the air inlet through which smoke particles enter air sampling network 403, using any of the methods herein described. The inventors have realised that it can be advantageous to use the volume of air drawn through 25 the system or corresponding values to determine the point of entry of particles into the air sampling system. Moreover, by measuring volume rather than time, certain disadvantages or problems associated with reliance on measurement of time may be ameliorated. For example, it is known that with usage the sampling inlets gather dirt and get constricted, resulting in greater pressure drop and less flow of air. This means changing transport time for air samples 30 over the life of the system. However the volume of air displaced to get a sample to the detector is relatively constant over time which makes the correlation between displacement volume and address more stable than transport time. Moreover if there are delays in opening a valve or beginning an aspirator, or the fan starts more slowly than expected the volume of air drawn 27 through the system before particles are detected a second time is likely to be relatively unchanged, as compared to time based systems. Advantageously volume-based addressing systems may be able to be operated independent of the flow rate or over a range of variable flow speeds, enabling techniques such as those described below, in which the system opens up 5 an end cap to speed up the flow of a sample to the detector. Other types of flow sensor can be used in embodiments of the invention, for example a mass flow sensor, which provides an indication of the mass of air moving past the sensor over time. However, because mass flow sensors are insensitive to the density of the air they measure, other information such as the temperature of the air is required in order to determine the volume 0 of the air moving past them. A further difficulty that can arise in implementing embodiments of the above invention and that of the prior art is the potential difficulty in reliably determining that two equivalent smoke detection events has occurred, for example noise introduced prior to conversion of a signal from analogue to digital form may frustrate the process of determining when smoke is detected by 5 detector 202, or detector 204. The inventors have devised an improved process that avoids or ameliorates this drawback. A smoke detection system such as that of Figure 2 produces two distinct data sets or "particle detection profiles". One data set is drawn from particle detector 202. The second data set is drawn from particle detector 204. Each data set contains a series of measured smoke levels. 20 The data set may also contain information regarding the volume of air flowing through the detector, or a time at which a particle smoke level was measured. (n the following example, we will describe a system that monitors smoke levels over time. A person skilled in the art would appreciate that the method can be adapted to measuring smoke levels compared to the volume of air drawn by the system (as described above), however for 25 illustrative purposes, we presently describe the system in relation to a series of measured smoke levels taken at various times. Figure 5 illustrates a particle detection profile. Detected smoke level is represented along its vertical axis. Time is measured along the horizontal axis. The smoke levels are those measured by detector 202 of Figure 2. Figure 6 shows a second particle detection profile. It is 30 similar to that of Figure 5, except that it relates to smoke levels measured by detector 204.
28 Comparing the figures, detector 202 detected a smoke level that reached a maximum at time 200, at which time it was deactivated and the particle detection output returned substantially to zero. Detector 204 detects a maximum smoke level at time 300. The different times are at least partially attributable to the additional distance along the sampling network 205 that smoke 5 reaching detector 204 must travel. It would be possible to use the difference between the time of each maximum or the difference in time at which each profile crosses some predetermined threshold e.g. a smoke level of 150 on the vertical axis (which may be different to the alarm thresholds in use), to estimate the air inlet through which smoke entered the particle detection system. However, more preferably a cross correlation can be calculated using the data 0 illustrated in Figures 5 and Figure 6. For real and continuous functions f and g, the cross-correlation is calculated according to the formula: (f * g)(t) = f (g(t + rdr A person skilled in the art will appreciate that this equation can be adapted for use with discrete 5 measurements, such as the smoke levels detected in the present systems. For example, such a system can be implemented in hardware by temporarily storing a particle detection profile of each detector data in a respective buffer, e.g. a ring buffer. The buffers may be chosen so as to store data such that the longest possible offset measurable by the system can be accurately calculated. The cross correlation at a point can then be calculated by multiplying each pair of 20 data elements in turn, and adding them, as described by the equation above. This process can then repeated for each possible offset t, to determine the overall cross-correlation function. The cross correlation function can then be used to estimate of the time offset between two particle detection events. This can in turn be used to infer through which inlet pair the particles entered the sampling pipe network. In some embodiments, information from the cross 25 correlation function is used to locate further geographic locations at which smoke may have entered the system. In one embodiment, multiple peaks of the cross-correlation function are identified. A list of time offsets is calculated based upon the location of each peak and its corresponding cross correlation value. The time offsets are used to infer the geographic location of the source of 30 smoke. This can be used to potentially infer multiple locations at which fire occurs.
29 Figure 7 illustrates a detector particle detection system 700 that includes a particle detector 702 in fluid communication with an air sampling network in the form of pipes 704, 706, 708 and 710. Each pipe includes a plurality of inlets, arranged into sampling inlet groups 712 to 740. Each sampling inlet group corresponds to a physical address, eg: a room or location that is serviced 5 by the detector. Each sample inlet group includes between one and four air inlets. The particle detector is connected to each pipe, and configured to provide an indication to a controller whether particles have been detected in fluid drawn through each pipe. The detector 702 could for example be four VESDA smoke detectors (from Xtralis Pty Ltd) detectors coupled to a central controller or a detector capable of independently detecting smoke on up to 4 pipes. 0 Each of sampling inlet groups 712 to 740 comprises one, two, three or four individual sampling inlets. The inlets are arranged into groups such that the same pattern does not occur twice. For example, sampling inlet group 730 includes an inlet on each pipe but no other group includes an inlet on each pipe. Sampling inlet group 712 includes an inlet only on pipe 710, but no other sampling inlet group includes only a hole on pipe 710. In the example of Figure 7 5 the inlets are arranged in groups corresponding to a 4-bit Gray code. Consistent with the discussion previously in relation to Figure 2, at the time of installation, the inlets from each group are positioned close to one another. In the event that smoke enters the sampling network at a particular inlet, smoke should enter each of the pipes for which there is an inlet present in that group. For example, if smoke enters the sampling network near the 0 location of sampling inlet group 730, one would expect smoke to enter each of the four pipes 704, 706, 708 and 710 at that location. Conversely, if smoke enters the sampling network at sampling inlet group 712, one would expect smoke to only enter pipe 710, since at that location, no other pipe includes an inlet. Upon detection of particles in the samples drawn into the individual pipes 704, 706,708, 710, the particle detection system is able to determine the point 25 of entry of smoke into the sampling network based upon the pattern of detection across the pipes 704, 706,708, 710. The table of Figure 7 more completely illustrates the possible combinations of particle detection states across the four pipes and their corresponding particle detection locations. It is useful to begin by defining a nomenclature for expressing the indicated smoke levels. For present 30 purposes, we will use a four binary bits to correspond to the detected smoke levels for each of pipes 704, 706, 708, and 710 respectively. For example, the indication '1111' corresponds to 30 detection of smoke at some threshold level, in air drawn from each of pipes 704, 706, 708, and 710. The indication '1100' would refer to detection of smoke in air drawn from each of pipes 704 and 706. The indication '1010' would refer to detection of smoke in air drawn from each of pipes 704 and 708. Accordingly, each of these four bit indications can be treated as an 5 address that corresponds to a location. There are fifteen non-zero four bit numbers. Accordingly, these fifteen numbers can be used to distinguish fifteen separate locations. The table of Figure 7 lists each of the possible fifteen non-zero binary numbers in the column 'Gray Code' address. Alongside each binary number is one of 15 locations in the 'Location' column. The 'Smoke Detected' column shows whether smoke had been detected at the assigned 0 threshold level at pipe. There is a large number of possible ways of allocating addresses to each location. For example, in some embodiments, each successive location from 1 to 15 may take a subsequent binary number, in a manner similar to ordinary counting. Accordingly to this scheme, location 1 would have the address '0001' (which is a binary representation for the decimal number '1') 5 and location 2 would have the address '0010' (which is a binary representation for the decimal number '2'). In this scheme, location 15 is given the binary address '1111', which is a binary representation for the decimal number 15. However, the illustrated embodiment uses a different method of allocating addresses, called a 'Gray code'. In the illustrated gray code of Figure 7, the location 1 is given the address '0001'. .0 Location 2 is given the address '0011' (which corresponds to the binary for the decimal number '3'). Location 3 is given the address '0010' (which corresponds to the binary for the decimal number '2'). This sequence of numbering has a special property when each of the binary representations is considered. In particular, each pair of adjacent locations has a binary representation that differs by precisely one bit. For example, location 4 has the address '0110', 25 whereas location 5 has the address '0111', and so only the fourth bit of each number differs. Similarly, location 11 has the address '1110' whereas location 12 has the address '1011', and so these also differ by their second bit only. The way in which addresses are chosen may influence performance in the presence of detection errors. In particular use of a Gray code scheme may be, more robust to addressing 30 errors than a straight "counting" address scheme in which successive locations are addressed by successive binary numbers. To illustrate this point, in a system that adopts the gray code numbering as described in figure 7, there is roughly a fifty percent chance that for a single bit 31 error the determined location of the fire will be a location adjacent to the actual location of smoke, since the address of each adjacent location differs by a single bit only. A person skilled in the art would appreciate that judicious selection of the sample inlet groups and increasing of the number of pipes feeding the detector can result in increased redundancy 5 for the purpose of the localizing decision. In practical terms, the introduction of this redundancy may be such that, for example, simultaneous entry of smoke at multiple sample inlets can be distinguished, or alternatively, such a system may simply provide greater resilience to error. Figures 8 and 9 show two embodiments of a further mechanism for providing addressability within an aspirated particle detection system of the type described in Figure 1. 0 Turning firstly to Figure 8, which shows a particle detection system 800, including a particle detector 11, coupled to an air sampling system 26. The air sampling system 26 includes a sampling pipe 28, including five sample points 29. As described in relation to Figure 1, the aspirator of the particle detector 11 draws air samples in through the sample inlets 29, which then travel along the pipe 28 and into the detector 11 for analysis. In this embodiment, each 5 sampling hole 29 additionally includes a valve 802. Each valve 802 is independently able to adjust flow through its respective sampling hole 29. The valves are controlled by the central controller of the detector 11, and are configured to be opened and closed under the control of detector 11. The purpose of the valves 802 on each sampling inlet 29 is to enable the smoke detector 11 to 20 vary one of its systems' sampling parameters in order to assist in determining which of the sampling inlets 29 particles of interest have entered the system 800 through. Upon an initial detection of particles of interest by the detector 11, at a predetermined threshold level, the detection system 800 goes into the localisation routine. In this routine, the detector 11 causes the valves 29 to vary a sampling parameter, in this case flow rate, of air entering the sampling 25 inlets. This variation may be performed on an inlet by inlet basis, or in groups of multiple inlets. After each variation in flow rate, a new particle concentration measurement is made. The initial particle concentration measurement and the second particle concentration measurement along with a variation parameter can then be used to determine which of the sample inlets particles of interest entered through.
32 This works because the particle level detected at the detector 11 is a weighted sum of particle concentrations and flow rates of the sample flow at each individual inlet 29. By varying the smoke level or flow rate through the sampling inlets, it is therefore possible to solve the set of simultaneous equations to determine the particle level entering any one sample inlet or group of 5 inlets. To illustrate a simple example, consider a smoke detection system including a smoke detector and a sampling network having a pipe with two sample inlets. In this example, the level of smoke detected when all valves are open is given by the following equation: 0 DetectorSmokeAllValvesOpen = Smoke1* fowl + Smoke2* fow2 fowl + flow2 Where, DetectorSmokeAllValvesOpen is the total smoke detected by the smoke detector; Smoke is the smoke level in the sample entering sample inlet 1; flow is the flow rate of the sample entering through sample inlet 1; Smoke2 is the smoke level entering the sample inlet 2; and 15 flow2 is the flow rate through sample inlet 2. Now, when the first sample inlet is closed by its valve, the weighted sum of smoke arriving at the detector becomes: DetectorSmoke Valvesl Closed = Smokel * 0 + Smoke2* flow2 0 + flow2 It will be noted that this weighted sum is identical to equation 1, except that flow = 0, because 20 the valve on sample inlet 1 has been closed fully. We are now in a situation where we can solve these equations for Smoke1, to determine the amount of smoke that has entered through sample inlet 1, as follows: 33 Smokel = DetectorSmokeAllValvesOpen(flowl + flow2) - DetectorSmokeValveslClosed(0 + flow2) fowl Thus, if we know flow, flow2 and the change in flow, we can solve the equation and determine what smoke level entered at sample inlet 1. This principle also works in the event that the 5 valves 802 only partially restrict flow through their respective sampling hole when they are closed, so long as it is possible to determine the flow rate at each sampling inlet 29. In order to allow flow rate to be detected, the system 800 includes a flow sensor 804 at each sample inlet 29. The flow sensor 804 could be a high sensitivity flow sensor, such as an ultrasonic flow sensor or a lower cost thermal flow sensor of the type which will be known to those skilled in the 0 art. In some embodiments, the valves 802 will not reduce the flow rate through their respective sample inlet to 0, but will only reduce it by some fraction. The following equation demonstrates how in a two hole system, as described in relation to the last example, smoke level through sample inlet 1 (Smokel) may be calculated if valves are used to reduce the flow rate through 5 their respective sampling holes to half their previous flow rate. Smokel = DetectorSmokeAllValves Open(flowl + flow2) - DetectorSmoke ValveslClosed(O.5 fowl + flow2) 0.5flowl In a further embodiment of the present invention, instead of varying flow rate through the sample inlet to solve the simultaneous equations, it is possible to vary the level of smoke entering each of the inlets. This can be achieved by selectively interposing a filter into the flow 20 path through each of the sample inlets 29. An example of such a system is shown in Figures 9A & 9B. The system of figure 9A 900, includes a detector 11 connected to a sampling network 26, Which includes sampling pipe 28, into which air samples are drawn through plurality of sample inlets 29. Each sample inlet additionally includes a selectable filter arrangement 902, which is shown in more detail in Figure 9B. The selectable filter arrangement 902 presents an air sample 25 inlet 904 (equivalent to inlet 29) at one end, and a sample outlet 906 at the other. The air sample inlet 904 is open to the environment, and allows an air sample from the environment to be drawn into the selectable filter arrangement 902. The sample outlet 906 is connected to the sampling pipe 28. Inside the selectable filter arrangement 902 are two flow paths, one path, 908, which is unfiltered, and another 910 which includes a filter 912. The selectable filter 30 arrangement 902 additionally includes a valve 914. The valve 914 is moveable between the first 34 position in which it blocks the filtered flow path 910, and a second position in which it blocks the unfiltered flow path 908. After smoke has initially been detected by the detector 11, at a threshold level, and the detector goes into its localisation mode, in which it attempts to determine which sample inlet 29 particles have entered the system from, the valve 914 is 5 triggered to switch between the first position in which particles drawn in through the inlet 904 are allowed to pass through to the outlet 906, into a second position, in which any particles entering the inlet 904 are removed from the airflow passing out of the outlet 906 by the filter 912. In a preferred form, the filter 912 is a HEPA filter or other high efficiency filter which will remove substantially all particles from the airflow. 0 The sampling point 29, and in this case the selectable filter arrangement 902 includes a flow sensor 916 to measure flow rate entering the sampling point 29. As will be appreciated, a similar set of equations to that described in connection with the first example, can be applied to the system of the type illustrated in Figure 9A and 9B. For a two hole system, as discussed above, the level of smoke arriving at the detector when all 5 sample inlets have their input unfiltered can be expressed with the following equation: DetectorSmokeAllUnfiltered = Smokel * fowl + Smoke2 * flow2 fowl + flow2 Where, DetectorSmokeAllUnfiltered is the level of smoke received at the detector when all flows are unfiltered, and all other terms are as described above in connection with equations 1 through 4. 20 After the selectable filter arrangement of the first sampling hole is moved into its filtered mode, the weighted sum expressing the level of smoke received at the detector is expressed as follows: DetectorSmokeFilteredl 0 * flowl + Smoke 2* flow2 fowl + flow2 Where, DetectorSmokeFilteredl is the level of smoke received at the detector when the flow 25 through sample inlet 1 is fully filtered.
35 Solving these equations simultaneously yields the following equation, from which the level of smoke arriving at sample inlet 1 can be determined. Smokel = DetectorSmokeAllUnfiltered(flowl + flow2) - DetectorSmokeFilteredl(flowl + flow2) fowl In order to handle increasing or decreasing smoke levels which may change reliability of this 5 type of localisation process, the sequence of taking measurements in a first state and a second state can be repeated, and equivalent states averaged over a number of cycles. For example, the first measurement with all valves open can be taken followed by a smoke level measurement with the varied parameter, followed again by an equivalent initial reading with all valves open again. The two valve open measurements can then be averaged and used in 0 subsequent calculations. Further variation on the present systems can be implemented where instead of constricting or reducing the flow through each of the sampling points, the flow rate at the sampling points is increased, either by opening a valve, to increase the size of the sampling hole to decrease its flow impedance, and thereby increase the proportion of the total airflow from the system which 5 is drawn through that sampling point, or by putting a fan at each sampling point and actuating or varying the speed of the fan to either increase or decrease the flow through the sampling point by a known amount. The above embodiment has been described with a simple two inlet system. However, as will be appreciated, an as described in Figures 8 and 9A, systems are likely to have more than two 20 sampling inlets. In such systems it is possible to scan through each of the inlets individually and vary the sampling parameter at only one inlet at a time. However, it may be beneficial to perform the variation in a grouped manner in which a subset of the total number of inlets have their sampling parameters adjusted in each measurement cycle. In some cases it may be possible to vary the sampling parameters of all sampling inlets by a differential amount in order 25 to determine the contribution of each. As will be appreciated, the more inlets in the system that there are, the more times the process of varying sampling parameters and remeasuring particle concentration needs to be performed in order to collect sufficient data to solve the necessary set of equations.
36 The concept described in connection with figures 8, 9A and 9B can be extended more generally to a method for detecting contaminant(s) in air samples drawn from a plurality of air intake paths and determining the contaminant level in each. For example the methods could be applied to an aspirating particle detector that is coupled with a sampling network having a plurality of air 5 sampling pipes feeding to the single detector, where the contribution from each pipe or branch of the sampling system is to be determined. Figure 7 describes a system in which this type of 'per pipe' localisation or addressing is used. In the example of Figure 7 the multi-pipe air sampling system may feed into a single contaminant detector such that it requires sampling of one pipe at a time, in order to determine 0 which of the pipes has the contaminant in the air stream. This can be achieved by sealing all but one of the pipes and allowing a sample to enter the detector from one pipe at a time while the detector measures the contaminant level. This is repeated for each of the pipes in the multi pipe air sampling network. The sealed pipe must be fully sealed against air flow in order to obtain accurate measures of the contaminant level in the open pipe. However, complete 5 sealing is very difficult to achieve in low or reasonable cost valves. However by using a method similar to that described in connection figures 8, 9A and 9B the requirement of complete sealing can be avoided. Figure 10A schematically illustrates a sensing system 1010 having and a sampling pipe network 1011 comprised of a total of two sampling pipes 1012, 1014. Each sampling pipe 1012, 1014 !0 defines an air intake path therethrough. The air intake paths are combined at manifold 1016. The manifold 1016 may include suitable baffles to assist with combining the air flows. Air is drawn through the sensing system 1010 through the use of the fan 1018. A subsample from the combined air flows is drawn through detector loop 1020 in which a filter 1022 and a particle detector 1024 are provided. Once the air flow has passed through detector loop 1020, it rejoins 25 the main air flow path 1019. A flow sensor 1026 may optionally be provided prior to the outlet 1028 of the system 1010. As will be appreciated the sensing system 1010 is equivalent to the detector 11 of figure 1. Each of the sampling pipes 1012, 1014 has a valve such as a butterfly valve or another type of flow modifier 1030, 1032. Additionally, each sampling pipe 1012, 1014 includes an ultrasonic 30 flow sensor 1013 and 1015.
37 It should be noted that, although the valves 1030, 1032, flow sensors 1013, 1015 and manifold 1016 are illustrated as forming part of the sampling network 1011, they may equally be physically located within the housing of the sensing system 1010 and thus form part of the sensing system 1010 without changing operation of the present invention. 5 A method according to the present invention will now be described in connection with figures 10B to 10D. In normal operation, each valve 1030, 1032 is fully open as shown in Figure 10B. However, when the particle detector 1024 detects the presence of a contaminant in the sampled air flows at a predetermined level, the scanning method according to the present invention is undertaken. Firstly, the first sampling pipe 1012 is partially closed as shown in Figure 10C. In 10 this condition, the particle detector 1024 takes a measure of the contaminant (C 1 ). Additionally, the flow rate is measured in the sampling pipes 1012, 1014 (Fmp, where F is the flow, m is the measurement number and p is the pipe number. Thus, the flow rate measurements will be F 11 and F 12 ) with flow sensors 1013 and 1015 respectively. In the next step, the other sampling pipe 14 is partially blocked by moving the butterfly valve to 15 the position illustrated in Figure 10D. In this condition, the particle detector measures the contaminant level (C 2 ). Additionally, flow rate measurements are taken (F 2 1 , F 22 ). Assuming that the amount of contaminant (or relative amount of contaminant between pipes) is not changing significantly during the scanning period, the individual contaminant measurement for a pipe can be calculated from the following set of simultaneous equations: 20 C 1 = X 1
F
1
/(F
1 + F 12 ) + X 2
F
12
/(F
1 + F 12 )
C
2 = X 1
F
2 1
/(F
21 + F 22 ) +X 2
F
22
/(F
21 + F 22 ) where X 1 is the actual contamination in pipe 1 and X 2 is the actual contamination in pipe 2. Advantageously, embodiments of the present invention enable cross-talk between the sample 25 pipes, caused by imperfect sealing of the sample pipes, for a given species of contaminant to be eliminated without costly, precision valving. Instead, low-cost butterfly valves or other types of flow modifiers are sufficient to accurately eliminate the cross-talk, and allow pipe addressability to be achieved.
38 As noted above, the instead of using valves to partially close the pipes, a filter could be selectively interposed into the pipes to reduce the contaminant level in each pipe temporarily by a known amount (preferably to 0) and the method adjusted to solve for Contaminant level as described above for hole addressing. 5 In the various embodiments described herein, a common step which is performed, is an initial detection of particles at a detector and more particularly an attempt to accurately identify the receipt of the smoke from a particular sampling inlet of the sampling system. In particular, the event which is most commonly sought to be detected is an arrival of a smoke front that is propagating down sampling pipe, and which represents smoke which entered a particular 0 sample inlet after a change in the operation in the sample network, e.g. opening or closing of valves or flushing the pipe network with clean fluid, or reversing flow direction or the like. Figures 1 1A and 11 B illustrate this concept. Figure 1 1A illustrates a particle detection system 1100, which includes the detector 1102, and a sampling pipe network 1104. Sampling network 1104 has three sample inlets, 1106, 1108 and 5 1110. A smoke plume 1112 is located adjacent to sampling inlet 1108. Take for example a situation in which the direction of flow in the sampling network 1104 is reversed and the detector 1102 is attempting to determine the time of arrival of smoke entering the system from sampling hole 1108. A graph of determined smoke concentration against time is illustrated in Figure 11 B. Initially, for some period, 1020 low smoke level is detected as the sample fluid arriving at the .0 detector only contains sample fluid from sample inlet 1106. At time T1, an increase in smoke is detected. Over the next time period 1022, when the sample from inlet 1108 begins arriving the detected smoke level ramps up until time T2, when approximate steady state level is detected. In the graph of Figure 11B, the ramp-up 1022 is not due to an increase in smoke level, but due to a smearing or diffusion of the smoke front of sample entering sampling hole 1108. If the entry 25 of particles from the environment into the sampling network was even and instantaneous, there would be a step change in the smoke level detected by the detector 1102, at T1 when the sample from hole 1108 arrives at the detector 11. The present inventors believe that there are a range of factors contributing to the diffusion of the smoke front, representing the arrival of the sample portion that includes an air sample drawn 30 through a particular one of the sample inlets of the sampling system. Chief amongst these is suspected to be the existence of a flow speed gradient across the cross-section of the air sampling duct. Figure 11 C illustrates a cross section through an air sampling duct 1130 such as 39 pipe 1104. Arrows 1132 indicate that flow rate in the central portion of the duct 1130 is greater than the flow rate near the walls of the duct. The belief is that it takes some amount of time for a sample being drawn in through a sample inlet, e.g. 1134 to break into the fast flowing central region of the flow in the duct 1130, and 5 therefore the smoke front is smeared out when it arrives at the detector. This mechanism however has competing factors, namely initially a sample will be introduced into the slow flowing peripheral air within the duct which will delay its arrival at the detector. However over time part of the sample will find its way into the fast flowing central region which will minimise its transport time to the detector. 0 The inventors have proposed that a physical structure can be placed in the duct of the sampling network (i.e. in the pipe of the sampling network) to ameliorate this problem. In a first family of solutions, the inventors propose a sample injection inlet which extends inward from the wall 1131 of the pipe 1130, towards the centre 1133 of the pipe 1130, so as to deliver the sample in the faster flowing region of the sample flow. Three examples of such a sample injection inlet are 5 shown in Figure 12. In Figure 12, a duct forming part of an air sampling system in the form of pipe 1200 is illustrated. The pipe 1200 is defined by a wall 1202. Three sample injection inlets 1204, 1206 and 1208 are also illustrated. The first sample injection inlet 1204 is a short tube 1210, which extends from the side wall into the pipe 1200, towards its centre 12-12. Sample injection inlet 1206, is similar e0 to inlet 1210 but terminates on its inside end 1214 with a Chamfered tip. The tip has the effect of functionally making the outlet 1216 point in a downstream direction with respect to the flow within the pipe. Finally, sample injection inlet 1208 takes the form of an inverted L shaped tube 1220. Its inlet is external to the duct 1200, and its outlet 1222 faces in a downstream direction and is aligned 25 with the centre of the duct 1200, thus injecting samples, drawn into the sample inlet 1208, at the centre of the pipe in the fastest flowing fluid flow. These three examples take advantage of the faster flowing central region of flow within the pipe to minimise smearing of samples drawn in through the sample inlet. An alternative to this injection method is illustrated in 13A to 13D. This series of examples uses 30 a structure which creates turbulence within the duct of the sampling system to prevent or disrupt 40 laminar airflow within the sampling duct, to thereby minimise flow gradient of the type illustrated in Figure 11C. Figures 13A to 13D each illustrate a segment of duct 1300, 1310, 13,20 and 1330 respectively. In Figure 13A, the inside wall 1302 of the duct 1300 is used as a turbulator. The wall 1302 has 5 been roughened or given surface contour or texture such as ribs, lines, bosses, or other, to create a rough surface that disrupts flow across it. In Figure 13B the turbulator is a series of turbulence causing protrusions 1312 extending inward from the wall 1310 of the pipe, and are used to caused disruption of laminar flow within the pipe 1310. 0 Figure 13C illustrates an example in which a plurality of turbulence causing members extend the full breadth of the pipe 1320. In this example the turbulators are in the form of open mesh elements 1322. The open mesh elements 1322 have a hole size sufficiently large that they will tend not to clog over time but will cause turbulence to be created in the pipe 1320. As will be appreciated by those skilled in the art, a range of different shaped turbulators which span 5 across the interior of a sampling duct can be devised. Figure 13D illustrates a further example in which a moving turbulence causing element 1332 is placed inside the pipe 1330. In this case, a series of fans 1334 and 1336 are supported in the pipe 1330. The fans may be actively driven or passively rotating, but serve to stir the air or cause turbulence therein, as the air flows past them. 20 In this example, it has been convenient to describe the turbulence causing structure in a region of the duct which is an adjacent sampling inlet, however it should be noted that there is no particular reason why this should be done and the turbulence causing structure could be placed away from sampling inlets. As will be appreciated with the four examples described above, the purpose of the turbulence 25 causing structure is to break down the flow profile across the air sampling duct such that the air entering from a sampling inlet will travel along the sampling duct to the detector like a 'packet', rather than having part of it travel relatively faster or slower than another part and thereby smear out the arrival of the sample front at the detector.
41 Alternatively, or in addition to the techniques described above, the present inventors have identified that additional improvements in detecting which sample inlet of a plurality of sample inlets, smoke is received from by at least partially ameliorating the effect of dilution on air samples drawn into the sampling network. Consider a particle detection system such as that 5 illustrated in Figure 11A. In such a system, the air sample drawn into sampling pipe 1108 will be drawn into the sampling pipe 1104, where it mixes with, and is diluted by a sample drawn from sampling point 1110. Similarly, the air sample drawn from sample inlet 1106 is diluted by samples drawn from all up-stream sample inlets. Thus, by the time air samples arrive at the detector 1102, the actual concentration of particles which is detected will be greatly diluted 1o compared to the sample concentration in the atmosphere surrounding the particular sampling inlet through which the particles entered the sampling network. The present inventors have determined that certain modifications to the systems described herein can be performed to ameliorate this problem, either by increasing the concentration of samples drawn into the sampling pipe, such that they more closely reflect the actual concentration of particles in the 5 atmosphere surrounding this sampling point and/or by providing mechanisms for delivering samples to the detector with minimal additional dilution. Figure 14 illustrates a first exemplary system 1400 which implements such a technique. The system 1400 includes a detector 11, and an air sampling network 26 including a sampling pipe 28 having five sample inlets 29 at the far end 1402 of the air sampling pipe 28, the detector !0 system 1400 includes a sample amplification arrangement in the form of bellows 1404, which are driven by an actuation means 1406. The bellows 1404 perform the function of blowing or sucking air along or from the sampling pipe network in a manner to be described below. As will be appreciated by those skilled in the art, a wide variety of systems could be used to replace the bellows structure, for example, a reciprocating pneumatic piston, or reversible fan or pump or 25 other like air movement device could be used in place of the bellows 1404. Operation of system 1400 will now be described. Initially, once particles at a threshold level have been detected by the detector 11, the system 1400 enters a localisation mode in which the location of particles in the system will be determined. In this mode, the primary air movement system, e.g. the aspirator 16 of the detector 11 is stopped and the system enters a sample 30 amplification phase in which the controller communicates via communications channel 1408 with the actuation device 1406 of the bellows 1404. With the fan stopped, or alternatively with a valve at the detector end of the sampling network 26 closed, the sampling pipe 28 contains a fixed volume of air, in use the bellows 1404 is used to increase and decrease the volume of air 42 contained within the sampling pipe network 26. When the bellows is expanded the volume increases and additional sample fluid is drawn into each of the sampling inlets 29. When the bellows is contracted some portion of the air within the sampling network 26 is expelled from the sampling inlets 29. By expanding and contracting the volume of air within the sampling pipe 5 network, air is repeatedly pumped into and out of each of the sampling inlets creating a localised sample portion within the sampling pipe 28, surrounding each of the sampling inlets 29, which more closely reflect the level of particles of interest in the environment directly adjacent each of the sampling inlets 29, than would be the case with the continually drawn and continually diluted sample stream. o Consider the situation at a single one of the sampling inlets 29, the air sample drawn into the sampling inlet enters the sample pipe network and mixes with the existing flow within the pipe 28. The existing air flowing past the sampling inlet dilutes the sample with samples drawn from all upstream sampling inlets. When the flow in the pipe 28 is stopped by closing a valve 1410 at the detector end of the pipe 28 or possibly by stopping the aspirator of the detector 11, then the 5 bellows 1404 are contracted and then, some portion of air within the sampling pipe 28 surrounding the sampling point 29 is expelled from the sampling point 29, as air is pushed along the sampling pipe 29 by the bellows. However, the air which is expelled from each sampling point includes the diluting samples from the upstream sampling points. Suction is again applied to the pipe network 28 by expanding the bellows 1404 and an additional air sample is drawn !0 into each sampling point. Whilst this sample is also diluted by the fluid which already exists within the sampling pipe adjacent the sampling point, part of this diluting air is the air sample which was previously drawn into the sampling point of interest. Therefore, the total concentration after the second sampling is increased compared to the first. With repeated cycles of expelling and sampling via a sampling inlet, the proportion of air within the pipe 28 in a 25 portion of the sample surrounding the sampling inlet begins to approach increases and the particle level begins to approach that in the atmosphere surrounding sampling inlet. Using this method, discrete sample portions within the sampling pipe 29 are formed which represent, more closely, the environment surrounding the sampling inlets. Because dilution is reduced, the methods described above which rely on detection of the onset of a smoke level increase i.e. a 30 smoke front to determine the location of entry of particles along the sampling network can be improved. Once the sample amplification phase is completed the system enters a transportation phase and moves the sampled air, now including sample packets which are relatively localised, back to the detector for analysis.
43 Figures 14A to 14E illustrate an exemplary system that uses a vibrating membrane, e.g. a speaker to perform sample amplification. The system 1420 includes a particle detector 11 coupled to an air sampling network 26. The air sampling network 26 includes a sampling pipe 28 having a plurality of air sample inlets 29. The air sampling network is coupled to the detector 5 via a sample amplification arrangement 1422 and aspirator 1424. The aspirator 1424 operates to draw samples into the sampling network and push them to the detector 11 for analysis in a manner that will be described in more detail bellow. The sample amplification arrangement 1422 performs a similar job to the bellows of figure 14 in that it causes oscillation of the flow direction in the air sample system to promote mixing of air in the region of surrounding each sample inlet 0 29 and air in the sampling pipe 28. In this example the sample amplification arrangement 1422 includes a membrane 1426 that is mounted within a housing 1428 and driven back and forth in reciprocating motion by an actuator. The actuator and membrane can be provided by a loudspeaker. Preferably the membrane is made to oscillate at a subsonic frequency, and most preferably at between 2 and 10 Hz. 5 In ordinary operation the aspirator 1424 runs at a first speed setting that is sufficient meet sample transport time requirements and draws air samples to the detector 11. Once particles are detected in the sample flow, the system 1420 enters a localisation mode beginning with a sample amplification phase. In this phase, illustrated in figure 14A, the fan enters a low speed operation and the sample amplification arrangement 1422 is activated. The membrane 1426 .0 oscillates and agitates the air in the pipe 28 to cause mixing with air nearby the entrance to each sample inlet 29. Because the fan is running at low speed, a mixed air sample that more closely approaches the true particle concentration in the air surrounding the sampling network 26 enters each sampling inlet 29 and slowly builds a packet of air downstream of each inlet. In figures 14B to 14D the agitation is continued as the fan 1424 runs slowly and builds the sample 25 packet 1430. Next in figure 14E, the system 10 enters transportation phase. In this mode the fan 1424 increases speed, and the membrane 1426 is stopped. The sample packets, e.g. 1430 are then drawn back to the detector 12 with the fan running in fast mode. As described below, various techniques (e.g. by blocking sampling inlets, opening the end of the pipe etc.) can be employed 30 to minimise mixing or smearing of the sample packets to thereby increase the reliability the localisation techniques applied. Figure 15 illustrates a second embodiment of a system 1500 which performs a similar method to that described in connection to Figure 14.
44 In Figure 15 like features have been like numbered with respect to Figure 14 and the earlier embodiments and for brevity will not be re-explained. In this example, the sampling network 26, at its distal end 1502 includes a fan 1504, and a valve 1506. Optionally at the end 1508 of the sampling pipe 28 which is closest to the detector 11, there may additionally be a second valve 5 1510. In this example, the valve 1506 is normally closed while valve 1510 is open during ordinary operation of the detector 11. Once the detector goes into its localisation mode however, the position of the valves 1510 and 1506 is changed and valve 1510 is closed and valve 1506 is opened. The fan 1504 is then used to perform the same function as the bellows 1404 of Figure 14. In this regard the fan 1504 is used to either blow some of the contents of the 0 sampling pipe 26 from the sampling points 29, or suck samples in through the sampling points 29 as described above. As will be appreciated, this oscillation of between sucking and blowing samples can be performed by the primary aspirator of the particle detector 11. However, by putting the fan 1504 at the far end of the sampling pipe network 28, an additional advantage can be gained, namely that the fan 1504 can be used at the end of this process to push the contents 5 of the sampling pipe 126 to the detector 11, rather than using the aspirator of the detector 11 to suck air samples down the sampling pipe 28. The advantage of using a blower fan 1504 at the end of the pipe 28 is that the sampling pipe 28 becomes positively pressurised and thus during the transportation phase does not draw any additional air samples from the environment surrounding the sampling points 29. In this way, a relatively undiluted column of sampling air 0 containing packets/portions of sample air corresponding to each sampling inlet 29 is delivered to the detector 11 such that the 'packets' of sample which were formed by the oscillation process can be distinctly detected by the detector 11. As will be further appreciated the oscillation of between sucking and blowing samples during sample amplification can be performed by using the primary aspirator of the particle detector 11 and the fan 1504 operating 25 in concert. For example, both fans may be set to operate synchronously, i.e. moving air in one direction and then the other to enhance localised mixing of samples around their respective sampling holes, or alternatively the fans can be set to alternately apply suction to their respective ends of pipe 26 to draw the sample fluid along the pipe in one direction. Thus rather than using the bellow-like push/pull on the sample flow from one end of the pipe 26 an 30 alternating pull/pull mechanism from two ends of the pipe is used. At the non-pulling end a valve can be closed (or partially closed) to control the amount of sample flow entering the pipe's 26 end. Advantageously this mechanism allows the system to increase the concentrating effect of bellows action. It also allows the sample packet to be formed on both upstream and downstream of the sample inlet position. The increased concentrating effect also enables the 35 system to cut down on the number of flow oscillation cycles for any given concentration 45 increase or mixing increase, relative to a system that acts at one end. This scheme may also average out (and possibly neutralise) the effect that fires closer to the detector end up with a higher slug concentration. As will be described below in connection with figures 18 and 19 a double ended flow modulation can be advantageously used to selectively perform sample 5 amplification. Figure 16 illustrates a further example of a system implementing the oscillation method and a mechanism for reduced dilution of delivery of final, increased concentration, air samples to the detector. The system 1600 includes a detector 11, and sampling pipe network 26, as described in connection with Figures 14 and 15, and similar features have been labelled with the same 0 reference numerals. In this example, the process of oscillating between sucking and blowing samples is performed by the primary aspirator of the detector 11. The sampling network 26 is additionally provided with a valve 1602 located upstream of the final sampling inlet 29. After sample concentration has been increased, as described above, using the main aspirator of the detector 11, the valve 1602, which is coupled to the controller of the detector 11 by 5 communications channel 1604, is opened. The valve 1602 is configured to open the end of the sampling network to the atmosphere such that it approximates an open pipe which has substantially less flow impedance than any one of the sampling inlets 29. When the aspirator of the detector 11 then applies suction to the sampling network 26, drawing air is preferentially drawn into the end of the sampling pipe 28, and the sample packets already within the pipe 28 .0 are drawn along to the detector 11. Because the open pipe end has low flow impedance, the level of air drawn into each of the sampling inlets 29 is greatly reduced, thus greatly reducing dilution of the samples as they are delivered to the detector 11. The reduced tendency for air to be drawn into the sampling inlets 29, when the valve 1602 is opened will also reduce the modification of the sample packets by smoke in the environment at or near the location of other 25 sampling holes. The reduced flow into the sampling holes 29 when the valve 1602 is open will also make the calculation of the smoke source position less dependent on the flow at the sampling holes. As described above, the system is initially trained to determine which hole a sample packet has arrived from based on how much air is drawn through the sampling network once the localisation phase has been entered. However, because the sampling holes may block 30 in a variable way over time the reliability of volume or time measurements based on the initial training may vary over time. By opening the valve 1602 the sample inlets 29 become less influential in the flow in the sampling pipe 28 and consequently the effect of differential blocking of the sampling inlets 29 over system life will be reduced. Finally opening the valve 1602 will 46 reduce flow impedance and the transportation phase faster. e.g. 40 sec for a 100 m pipe at 50 L/min rather than 110 sec with the end of the pipe closed. In some embodiments the valve 1602 or sampling network 28 beyond the last sampling inlet 29 can be provide with a filter, e.g. a HEPA filter through which air is drawn. This assists the 5 sample packet from the last sample inlet 29 in standing out from the air being drawn into the end of the pipe which might also contain particles or interest or even dust. As will be appreciated in the examples given herein, valves could additionally be applied to each of the sampling inlets 29 to further facilitate the effect of the flow control mechanisms (eg. bellow, fan, valve and equivalent structures) applied to the end of the pipe. For example, each 0 of the sampling inlets 29 can be provided with a valve which is controlled in concert with the pipe end flow control system to optimise its performance. In order for the above techniques to work reliably in the field, it is necessary to calibrate or train the system to as to the volume of air moved before an air sample entering a each sampling inlet arrives at the detector (or each detector), thus effectively characterising the system. Most 5 preferably the system is trained while the air is being moved through the system in the same way as during the system's localisation mode. For example, if the system uses a pusher fan method, described below in connection with figure 15, a significant localisation error we is likely to occur if the system is trained using normal detection operation when the pusher fan is not running. In one form, in which the system has a single air moving device, e.g. fan or the like or 20 there is no mechanism to dramatically change the flow impedance or flow path through the detection system when changing from the detection mode to localisation mode, a relatively simple, but time consuming process can be implemented in training mode. In this case the system can be trained as follows. With the system operating normally, the system measures the volume of air moved starting from the time at which a test smoke e.g. smoke spray is 25 dispensed, until the smoke arrives at the detector. This measurement is made for each sampling inlet. However this is can be time consuming as the training sequence needs to be performed for each inlet separately and the system may need to be left to return to normal operation between each cycle. Preferably the training mode uses a modified behaviour to reduce training time. 30 In other embodiments, e.g. a system which has an open valve plus a filter at end of the pipe during its transportation phase, the training mode involves opening the valve at the end of the 47 pipe for a period of time. Smoke can then be selectively administered to each sample hole (or to multiple holes in selected patterns) so that the system will still suck smoke through the holes. In training mode the system operates as follows: a. The system then opens valve at the end of the pipe. 5 b. User then inputs to the detector when smoke is administered at a sampling inlet. c. The detector measures the volume of air moved starting from the indicated time until smoke is detected for each sample inlet. In embodiments with a pusher fan (and preferably a valve and filter at the end of the pipe) it is more difficult to simulate smoke entering a sampling pipe. For example, it is not possible to get 0 spray smoke into a sampling inlet with the pusher fan continuously running. Therefore an alternative method is needed. Such as: a. Replicate the standard bellows operation, but with introduced smoke, including: i. Run the system normally; ii. Enter the calibration process; 5 iii. Activate the bellows as if particles had been detected, and indicate to a user that this process has begun; iv. User applies spray smoke at the sampling inlet under test. v. Deactivate the bellows and turn on the pusher fan to go into the transportation phase as normal, and record the volume of air transported before the smoke arrives at the 20 detector. vi. System indicates that the hole has been calibrated. vii. System closes valve and turns off pusher fan. viii. Other sample inlets are then calibrated in the same way.
48 b. A Special training mode: i. System running normally. ii. User puts the system into the test mode. iii. The system continues to draw air in normally and the user applies spray smoke at hole 5 and indicates this to the system. iv. The system then immediately turns on pusher fan. v. The system then records volume of air through flow sensor between indication of "spray start" and smoke being detected. vi. The system then indicates that a sample inlet has been calibrated. o vii. System closes valve and turns off pusher fan. viii. The next hole is then calibrated using the same process. c. Special smoke injector. This method is faster for the user but the user needs special equipment. This method involves use of an injection device which allows smoke to be sprayed into a sample inlet in a manner 15 that other positive pressure in the pipe. One way of doing this involves use of a test smoke generator unit that seals around the sample inlet and sprays smoke into the inlet. For example the smoke generator can have an outlet that includes a foam gasket which clamps around the sample inlet so air is not coming out the sample hole. Once fitted and a sample is injected into the sampling inlet the user inputs to tithe system that this smoke was sprayed. The system 20 records the volume of air moved before the smoke pulse arrives at the detector. Rather than empirically testing the behaviour of the system a simulator can be used. The simulator is similar to Aspire (from Xtralis Pty Ltd) The simulator works out the expected volumes per hole during the transportation phase based on the actual system hole dimensions and distances.
49 In the above testing methods a user can either interact with the detector directly to communicate inputs to it, e.g. to enter training mode, or indicate when test smoke has been sprayed etc. However in a preferred embodiment the detector system includes an interface, preferably wireless, by which the detector communicates with a user device, e.g. a portable 5 computer, tablet computer, smart-phone or the like, and the user device runs an application that allows the detector to be controlled to operate as described. Figure 17 illustrates a variant of the system of figures 14A to 14E. The system 1700 is identical in all respects to the system of figures 14A to 14E and operates in the same manner, with the exception that the sample amplification arrangement 1702 is located at the upstream end of the 0 sampling pipe 28. This simplifies the detector end of the sampling network 26 and facilitates retrofitting of a sample amplification arrangement 1702 to a legacy detection system that was originally installed without such a capability. Figure 18 illustrates a particle detection system including an air sampling network that has a sample amplification arrangement comprising a plurality vibrating membranes. Essentially this 5 system 1800 is a double ended version of the systems of figure 14A to 14E and figure 17. In this embodiment the two pistons 1802, 1804 (formed from the vibrating membranes of loudspeakers) act together to form the sample amplification arrangement. These can be operated in concert as described in connection with opposing fans of figure 15. However, being loudspeakers (or other similar air movement device capable of causing rapidly oscillating air 0 flow) these pistons 1802, 1804 offer new the ability to selectively perform sample amplification at one or more sample inlets 29 along the sampling pipe 28. This can be achieved by oscillating the pistons with a selected phase difference between them. This causes selective reinforcement or cancellation of the sample amplification action at different places along the sampling pipe 28. Figure 19 illustrates another particle detection system including an air sampling network with 25 branched sampling pipes and which has a sample amplification arrangement comprising a plurality vibrating membranes. The system 1900 includes a particle detector 11, coupled to an air sampling system 26. The air sampling system 26 is branched such that it has sampling pipes28A and 28B each of which includes a plurality of sample inlets 29A and 29B arranged in series along their length. At the upstream ends of the pipes 28A and 28B are located pistons 30 1902, 1904. A common piston 1906 is placed at the downstream end of the sampling network 26. The sample amplification arrangement comprising the pistons 1902, 1904, 1906 can be operated to selectively cancel its oscillation effect by choosing appropriate phase differences 50 between oscillation of the pistons in the sample amplification phase. For example in the example the downstream piston 1906 is operated in phase with the upstream piston 1902 on the pipe 28A, but out in anti-phase to the upstream piston 1904 on the pipe 28B. The result is that sample amplification occurs only on the sample inlets 29A but not on inlets 29B. 5 This process can be extended and combined with the method described in connection with figure 18. In this regard, greater selectivity can be achieved by operating the downstream piston 1906 is with a selected phase difference to the with the upstream piston 1902 on the pipe 28A, and no oscillation of piston 1904. Most preferably, if a node in the oscillation pattern coincides with the junction between the pipes 28A and 28B sample amplification will be minimised (or 0 possibly eliminated) on pipe 28B and selective sample amplification can be achieved along the length of pipe 28A. As will be appreciated the double-ended sample oscillation techniques described in connection with figures 18 and 19 could also be implemented with other types of air flow movement devices, e.g. bellows, fans (as illustrated in figure 15) or the like. 5 As will be appreciated from the foregoing, a number of techniques have been described within this document to improve addressing in aspirated particle detection systems which include centralised detector and a plurality of sample inlets placed along a duct or pipe of an air sampling system. It will be apparent to those skilled in the art that elements of each of the systems could be combined to further enhance system performance. To give but one example, .0 the pipe network work system of Figures 14, 15 or 16 could be used to increase smoke concentrations within the pipe network to deliver a clearer smoke concentration front to the detector for use in the cross-correlation method described in connection with Figures 5 and 6. Moreover, instead of using time based correlation, volume based correlation could be used as described above. Other combinations will be readily apparent to those skilled in the art. 25 It will be appreciated that the present invention, although described in relation to the detection of smoke, can equally be applied to any other material that can be usefully detected by a sampling system, including gases, dust, vapour, or biological materials. It will be understood that the invention disclosed and defined in this specification extends to all alternative combinations of two or more of the individual features mentioned or evident from the 51 text or drawings. All of these different combinations constitute various alternative aspects of the invention.

权利要求:
Claims (20)
[1] 1. A method of determining at least one point of entry of smoke into a smoke detection system, the system having a sampling pipe network including at least one sampling pipe and a plurality of sampling inlets through which an air sample can enter the at least one sampling pipe 5 of the smoke detection system for analysis by a particle detector, said method including: determining a volume of sample air that has passed through at least part of the smoke detection system since a predetermined event or a value corresponding to said volume; and determining through which sampling inlet of the plurality of sampling inlets the smoke entered the smoke detection system based, at least in part, on the determined volume or value. 0
[2] 2. The method according to claim 1, wherein the predetermined event is any one or more of: a smoke detection event; a change in an air sample flow characteristic in the smoke detection system.
[3] 3. The method according to claim 1 or 2, which includes continuously determining a flow 5 rate of the air sample passing through at least part of smoke detection system.
[4] 4. The method according to claim 1 or 2, wherein the method includes commencing determination of the volume of sample air or a related value upon the occurrence of the predetermined event.
[5] 5. The method according to any one of the preceding claims, wherein the volume of the air 20 sample that has passed through at least part of smoke detection network or a related value is determined by accumulating a flow rate measurement over time.
[6] 6. The method according to claim 5, wherein the rate of flow measurement is a volumetric flow rate measurement.
[7] 7. The method according to claim 6, wherein the flow rate measurement is determined 5 using an ultrasonic flow sensor. 53
[8] 8. The method according to any one of the preceding claims wherein the step of determining a volume of sample air that has passed through at least part of the smoke detection system since a predetermined event or a value corresponding to said volume, includes: determining any one of more of 5 a mass; a length; a pressure; a temperature, a second volume; 0 an accumulated count of volume-related events, or other parameter that that relates to a volume of sample air that has passed through at least part of the smoke detection system since the predetermined event.
[9] 9. A method as claimed in any one of the preceding claims which includes collecting all or a proportion of the sample air that has passed through at least part of the smoke 15 detection system since the predetermined event.
[10] 10. The method according to any one of the preceding claims, wherein the method further includes changing an air sample flow characteristic in response to a first smoke detection event.
[11] 11. The method according to claim 10, wherein the step of changing an air sample flow characteristic in the smoke detection system includes one or more of the following: 20 o opening a valve; * closing a valve; " changing a direction of an air sample flow in at least part of the smoke detection system; e changing a rate of air sample flow in at least part of the smoke detection system; 54 * starting an aspiration system; and * stopping an aspiration system.
[12] 12. An apparatus for determining at least one point of entry of smoke into a smoke detection system of the type having a particle detector in fluid communication with an air sampling 5 network, the air sampling network having at least one sampling pipe and a plurality of sampling inlets through which an air sample can enter the at least one sampling pipe of the smoke detection system for analysis by the particle detector, and an aspirator for drawing the air sample through the air sampling network to the detector, the apparatus including: means for determining a volume of sample air that has passed through at least part of 0 the smoke detection system since a predetermined event or a value corresponding to said volume; and means for identifying at least one point of entry of particles into the sampling network based on the detected volume or value.
[13] 13. The apparatus according to claim 12, wherein the apparatus identifies one or more of 5 said points of entry by reference to one or more corresponding sampling inlets through which smoke determined to have entered the system.
[14] 14. The apparatus according to claim 12 or 13, wherein the means for determining a volume of sample that has passed through at least part of the particle detection system, or value related to said volume, includes a flow sensor. 20
[15] 15. The apparatus according to claim 14, wherein the flow sensor comprises an ultrasonic flow sensor.
[16] 16. The apparatus according to any one of claims 12 to 15, wherein the apparatus is configured to perform a method in accordance with any one of claims 1 to 11.
[17] 17. A smoke detector including a particle detection chamber to detect particles in an air 25 sample, an inlet to receive an air sample from an air sampling network, said the sampling network having at least one sampling pipe and a plurality of sampling inlets through which a sample can enter the at least one sampling pipe for analysis by the particle detection chamber, 55 and an aspirator for drawing the sample through the air sampling network to the detector, the detector further including a processor configured to: identify at least one point of entry of smoke into the sampling network based, at least in part, on a volume of sample air that has passed through at least part of the smoke detector or 5 sampling network since a predetermined event, or a value corresponding to said volume.
[18] 18. The smoke detector as claimed in claim 17 wherein the smoke detector includes a flow sensor configured to detect rate of flow of sample air passing through at least a part of the smoke detector.
[19] 19. The smoke detector as claimed in claim 18 wherein the smoke detector includes an 0 ultrasonic flow sensor.
[20] 20. The smoke detector as claimed in claimed in any one of claims 17 to 19 wherein the processor is configured to cause the smoke detector to perform a method as claimed in any one of claims 1 to 11.

类似技术:

---

公开号 | 公开日 | 专利标题

---

US11002579B2|2021-05-11|Addressability in particle detection

---

US10161837B2|2018-12-25|Aspirated particle detection with various flow modifications

---

JP6475882B2|2019-02-27|Filter bypass

---

EP2438360B1|2016-10-19|Gas detector apparatus

---

US8949037B2|2015-02-03|Method and system for detecting and monitoring emissions

---

CA2675173C|2021-11-16|Method and system for detecting and monitoring emissions

---

US7724150B2|2010-05-25|Environmental sensor, particle counting system having an environmental sensor, and methods of operating the same

---

AU2013200353B2|2015-03-19|Addressability in particle detection

---

JP2017512309A|2017-05-18|Improvement of suction sampling system

---

CN103513003B|2015-03-25|CO on-line detection device based on constant-flow-velocity smoking machine

---

CN209327301U|2019-08-30|A kind of pico-stations monitor

---

CN110967479A|2020-04-07|Wound infection testing system and method

---

CN111033216A|2020-04-17|Airflow control for particle sensors

同族专利:

---

公开号 | 公开日

---

AU2013200353B2|2015-03-19|

引用文献:

---

公开号 | 申请日 | 公开日 | 申请人 | 专利标题

---

AU2003902318A0|2003-05-14|2003-05-29|Vision Fire And Security Pty Ltd|Improved Sensing Apparatus And Method|

---

US7375642B2|2004-08-24|2008-05-20|Wagner Alarm- Und Sicherungssysteme Gmbh|Method and device for identifying and localizing a fire|

---

GB2430027A|2005-09-09|2007-03-14|Kidde Ip Holdings Ltd|Fibre bragg temperature sensors|

---

AT517407T|2007-05-16|2011-08-15|Siemens Ag|DETECTION AND LOCATION OF A FIRE|

---

US20100194575A1|2009-01-30|2010-08-05|Carlos Pedrejon Rodriguez|Dual channel aspirated detector|

---

法律状态:
2015-07-16| FGA| Letters patent sealed or granted (standard patent)|

---

2017-02-02| PC| Assignment registered|Owner name: GARRETT THERMAL SYSTEMS LIMITED Free format text: FORMER OWNER(S): XTRALIS TECHNOLOGIES LTD |

---

2021-08-19| MK14| Patent ceased section 143(a) (annual fees not paid) or expired|

优先权:

---

申请号 | 申请日 | 专利标题

---

AU2012904516||2012-10-16||

---

AU2012904516A|AU2012904516A0||2012-10-16|Addressability in particle detection|

---

AU2012904854A|AU2012904854A0||2012-11-02|Addressability in particle detection|

---

AU2012904854||2012-11-02||

---

AU2013200353A|AU2013200353B2|2012-10-16|2013-01-21|Addressability in particle detection|AU2013200353A| AU2013200353B2|2012-10-16|2013-01-21|Addressability in particle detection|

---

JP2015535938A| JP6438399B2|2012-10-16|2013-10-16|Addressability in particle detection|

---

CN201380054236.7A| CN104718435B|2012-10-16|2013-10-16|Particulate detection addressing technique|

---

US14/433,201| US10545041B2|2012-10-16|2013-10-16|Addressability in particle detection|

---

CA2886675A| CA2886675A1|2012-10-16|2013-10-16|Addressability in particle detection|

---

KR1020157010624A| KR20150068963A|2012-10-16|2013-10-16|Addressability in particle detection|

---

PCT/AU2013/001201| WO2014059479A1|2012-10-16|2013-10-16|Addressability in particle detection|

---

AU2013332261A| AU2013332261A1|2012-10-16|2013-10-16|Addressability in particle detection|

---

TW106132458A| TWI665437B|2012-10-16|2013-10-16|Method and apparatus for determining at least one point of entry of smoke into a smoke detection system, and smoke detector|

---

TW102137299A| TWI609175B|2012-10-16|2013-10-16|Method and apparatus for determining at least one point of entry of smoke into a smoke detection system, and smoke detector|

---

EP13848049.6A| EP2909588A4|2012-10-16|2013-10-16|Addressability in particle detection|

---

HK15109128.9A| HK1213630A1|2012-10-16|2015-09-17|Addressability in particle detection|

---

AU2017254906A| AU2017254906A1|2012-10-16|2017-11-01|Addressability in particle detection|

---

JP2018151548A| JP2019007967A|2012-10-16|2018-08-10|Addressing ability in particle detection|

---

AU2019203934A| AU2019203934A1|2012-10-16|2019-06-05|Addressability in particle detection|

---

US16/706,408| US11002579B2|2012-10-16|2019-12-06|Addressability in particle detection|