GB2269665A - Optical beam smoke sensor - Google Patents

Abstract

A sensor which comprises emitters 2a to 2g which can emit narrow beams of infra-red light via optical elements 3a to 3g, each beam being angled in a different direction. Each emitter may be driven individually via electronic means 4a to 4g under the control of a microcomputer 5, such as to control the direction and the energy of the beam. Light receivers 8 and optical elements 9 have a field of view which includes the total emission field. For any combination of emitters selected the quantity of light received may be analysed. Used in a singlepath optical beam smoke sensor, control of the beam permits a reflective target to be imaged and the sensor to be automatically aligned, improving the performance and reliability. The sensor may also be used without a target by utilising reflections from the walls of the room in which it is installed, or by utilising an auxiliary receiver/emitter unit which detects the pulses of light and emits a return pulse after a short time delay. <IMAGE>

Description

OPTICAL BEAM SMOKE SENSOR This invention relates to optical beam smoke sensors primarily for use in detecting fires.

The detection of smoke by sensing the obscuration of a beam of infra-red light is a well known technique. Most existing optical beam detectors comprise two separate units, a transmitter and a receiver, installed on opposite walls with a clear line of sight through the volume to be protected, preferably within 0.6m of the ceiling. The effective beam is the line of sight between the optics of the units, and therefore the separation equals the path length. The maximum path length used is typically lOOm. The method is most commonly employed in larger internal volumes, since a low concentration of smoke caused by a small fire must be present in several metres of the path length in order to provide a sufficient level of obscuration. Generally the ceiling height should be greater than 3m, since the light path must not be obstructed by people or obstacles.

Although they detect fires well, existing optical beam smoke detectors suffer from a number of disadvantages, many of which relate to the signal to noise ratio. This needs to be as high as possible, since only a small proportion of the emitted light typically reaches the receiver unit even at minimum range, and the signal dependence on the path length generally follows an inverse square law. The receiver must be able to sense small reductions in the signal level to detect the presence of smoke, and must also sense larger reductions within a shorter time interval to identify attenuations caused by birds and other intervening objects.

To help maximise the signal to noise ratio, consistent with minimising the size, power consumption and cost of.the detector, the transmitter's angle of emission and the receiver's field of view should ideally be as small as practicable (eg +or- 1 degree), and this is a solution adopted with many detectors.

A major problem in using such small angles is the setting up and the maintenance of an accurate alignment between the transmitter and receiver. Alignment is generally by a mechanical adjustment, and the operation can be a time consuming two-man job. Even when the detector is well mounted, the beam intensity can vary and cause false alarms, because of temporary misalignments of the building structure caused by vibrations, wind, or temperature gradients. To counter this, experience has shown that the angles should be at least +or-3 degrees using traditional alignment methods.

Angles of approximately +or- 5 degrees in the vertical axis and +or- 10 degrees in the horizontal axis would enable a detector requiring no alignment procedures. However, such large angles would significantly degrade the signal to noise performance, and also increase the risk that a receiver may see the output from any adjacent transmitters.

Optical beam smoke detectors which combine a transmitter and a receiver within the same unit are also known, as disclosed for example in GB 2183825A. Such combined units are generally installed with a retro-reflector. This makes the effective path length double the separation between the detector and the retro-reflector. Advantages of this arrangement are that the effective sensitivity to fires is increased, and that the cost of the detector and its installation may be significantly reduced. In principle, the alignment problems can be minimised, since only one unit needs to be adjusted.

However, the signal to noise ratio of a combined detector and retro-reflector is strongly constrained by the size and dispersion angle of the retro-reflector, so in practice narrow angles of emission and fields of view are still used.

Some of the resulting loss may be recovered because the received pulse may be accurately synchronised with the emitted pulse, a feature which with normal optical beam smoke detectors requires electrical interlinking between the transmitter and the receiver. However, even if a broader field of view could be employed, this would increase the risk that the receiver could see spurious reflections from objects in the near field and mistake these for the signal from a retro-reflector installed in the far field.

One example of a combined unit used with a retro-reflector, disclosed in GB 2111278A, has the emitter and receiver disposed at the respective focii of a bi-focal Fresnel lens.

Although the Fresnel optics splits the emitted beam into a 'pencil of rays' these effectively comprise a single beam, which is returned to the sensor via the retro-reflector.

Although claimed to make alignment easier, the optical arrangement described would offer no practical advantages over the other known techniques. The optics of the sensor would also be likely to scatter too much light from the emitter to the receiver to be usable in practice.

The prior art described above relates to so called singlepath optical beam smoke detectors. Sensing the presence of smoke by imaging a scene in two dimensions is also known.

One way in which this can be done is by sensing the obscuration along multiple paths, and there are significant potential advantages in detecting fires using this approach.

GB 1439325 discloses a detector which uses a single continuous light source and a plurality of photo-electric receivers which monitor the signals from retro-reflectors within the field of view of the light source. A second photo-electric receiver is used to compensate for variations in the intensity of the light source. This technique suffers from a number of disadvantages, for example the use of a relatively expensive receiver array, the inefficient dispersion of the light energy into whole scene, and the difficulty of aligning the receiver array to receive the signals from the retro-reflectors.

The principle of using a multi-beam emission pattern in a volumetric detector is also known, but not for smoke sensing. For example EP 0188748 (US 4760381), relating to an intruder detector, describes an unit which emits a multibeam pattern of infra-red light, and uses one or more receivers within the same unit to detect changes in the scattered or reflected light caused by people moving within the protected volume. Various optical methods are described to emit the beam patterns, and to receive the light. In all of these a multiplicity of beams is emitted simultaneously, and these beams are widely spread into areas or sectors of the protected volume. In some implementations the beam pattern is duplicated, and the sensor switches automatically between the two sets of beam patterns, requiring a coincidence of signals in order to register an alarm.

Some of the arrangements diclosed in EP 0188748 superficially resemble the imaging volumetric optical beam smoke sensor embodiment of the present invention. However, a key feature of EP 0188748 is the simultaneous emission of many beams, which is adapted to sense the signal fluctuations caused by the movement of people in and out of the beams. It would not be suitable for quantitively imaging specific parts of the volume, which is necessary for sensing the gradual and small changes in obscuration caused by smoke from a slowly developing fire. In particular the methods disclosed would not work for smoke sensing because the signals resulting from light scattered from those beams which strike surfaces in the near field of the sensor would tend to swamp the signals from surfaces in the far field.

An object of the present invention is to improve the performance, reliability and ease of installation of singlepath optical beam smoke detectors. A second object is to permit the realisation of an imaging volumetric optical beam smoke sensor.

According to the present invention there is provided a sensor capable of sensing the presence of smoke by analysing changes in the amount of light received at the sensor caused by variations in the obscuration of a beam of light wherein the sensor emits a.single beam of pulsed light with the majority of the light beam energy being concentrated within an angle of less than 10 degrees in at least one axis, characterised in that the sensor includes at least one light receiver and more than one light emitter and that the predominant direction of emission of the said light beam may be altered by the selective use of one or more of the said light emitters under the control of a microcomputer or equivalent means in conjunction with suitable electronic and optical means, and wherein for each direction of emission the-quantity of light received may be analysed by the said microcomputer or equivalent means.

The invention will now be described by way of example with reference to the accompanying drawing, where figure 1 shows the general arrangement of the sensor and figure 2 illustrates a typical beam pattern for a single-path optical beam smoke sensor when used with a retro-reflector or auxiliary receiver/emitter unit.

The sensor 1 comprises an array of light emitters 2a to 2g which are capable of emitting light via an array of optical elements 3a to 3g. An array of electronic means 4a to 4g permits each of the light emitters 2a to 2f to be driven individually to output pulses of light under the control of a microcomputer 5. The light emitters 2a to 2g and the optical elements 3a to 3g are disposed such that a plurality of narrow beams of pulsed light 6a to 6g may be emitted, each angled to take a different path from the sensor.

On the other side of a light barrier 7 from the light emitters 2a to 2g there are disposed one or more light receivers 8 and suitable optical elements 9 to provide a field of view which at least includes all of the angles of emission of the light beams 6a to 6g. The light receivers 8 are connected to means 10, which amplify the signal and provide pulse sampling and/or analogue to digital signal conversion such as are necessary to input signals proportional to the light received into the microcomputer 5.

The microcomputer 5 is so arranged that for the emission of pulsed light from any combination of the light emitters 2a to 2g the quantity of light received may be analysed.

In a first preferred embodiment of the invention, adapted for use in a single-path optical beam smoke sensor, an array of seven emitters 2a to 2g are employed. Each emitter in conjunction with optical elements 3a to 3g emits a beam of light having the majority of light concentrated within an angle of emission in the vertical axis of'approximately +or5 degrees and in the horizontal axis of approximately +or- 2 degrees. The beams are disposed in the horizontal plane with angles between their centres of approximately 3 degrees. The image formed by each beam therefore overlaps that formed by the adjacent beam. If all the light emitters were to be made to emit at the same time the total emission field would form a bar pattern, approximately 10 degrees in the vertical axis by 20 degrees in the horizontal axis.The field of view of the receiver is arranged to approximately match this total emission field.

The sensor may be used in either of two modes. The first mode is in conjunction with a reflector, which could either by a mirror, or preferably a prismatic type retro-reflector.

Such retro-reflectors, moulded from transparent plastic are well known. They work by the beam being totally internally reflected from the surfaces of the prisms, which have defined angles in order to return the beam substantially back along the path from whence it came.

An alternative, particularly appropriate where the path length. is very long, or where the near field of the sensor is cluttered with reflecting objects is to use an auxiliary receiver/emitter unit. Such a unit may be similar in design to the sensor itself, or may be simpler in construction. It detects the emitted pulse and re-emits a new pulse, preferably after a short time delay. This delay allows the sensor to distinguish the pulse emitted by the unit from spurious reflections of its own emissions. In its simplest form the unit amplifies the received pulse and emits a pulse proportional to that received.

Figure 2 shows a sensor 1, a typical pattern of beam images 6a to 6g and a reflector or auxiliary receiver/emitter unit 11. The sensor, under software control initially emits from each emitter in turn and the optimum beam, or set of two or three adjacent beams (for example 6c and 6d in figure 2) is selected, and subsequently used. During on-going operation, the re-selection process may be automatically repeated at appropriate intervals.

In the single-path optical beam smoke sensor the prime purpose of the selectable beam emission is to concentrate the pulse energy as far as is practical in one part of the total emission field. This helps overcome the conflict between having a high signal to noise ratio and a large angle of emission, whilst at the same permitting the alignment to be set up electronically and thereafter automatically maintained. If the units are installed in measured positions on reasonably parallel walls the total emission field will be large enough to require no special alignment in the horizontal axis, and probably a check with a spirit level in the vertical axis.

When used with a retro-reflector, where there is nd measurable time shift between the transmitted and the received pulse, there is an additional important advantage in the selectable beam emission. By imaging across the total emission field the sensor software may not only determine the position of the retro-reflector in the emission field, but also the approximate relative amplitude of the signal which is being returned from the retroreflector, as compared with the signal caused by general background scatter. The retro-reflected signal amplitude may be used by the software to automatically establish the threshold alarm sensivity, and to warn the user if insufficient retro-reflected signal is present.

In a second preferred embodiment of the invention, adapted for use in an imaging volumetric optical beam smoke sensor, an array of seven emitters 2a to 2g are employed. Each emitter in conjunction with optical elements 3a to 3g emits a beam of light having the majority of light concentrated within an angle of emission of approximately +or- 2 degrees in both the vertical axis and in the horizontal axis. The beams are disposed in the horizontal plane with angles between their centres of approximately 15 degrees. The optical elements 9 provided for the one or more receivers 8 are arranged to give a received fields of view which approximately matches the emission fields of the emitters.

The imaging volumetric optical beam smoke sensor is mainly intended for use in medium sized rooms, especially those with complicated ceiling profiles. The sensor is mounted on a high wall, and the beams are spread in a wide horizontal plane, just below the ceiling level. It uses only the natural scatter and reflections from the walls, and therefore to achieve the maximum operational path length the light energy in each beam is concentrated in as narrow an emission angle as is practical. Because only natural scatter is used, no alignment or set up is normally necessary.

The sensor, under software control initially emits from each emitter in turn and then selects those beams which return an acceptable signal. During on-going operation all of the selected beams are used in an appropriate sequence. The software of the sensor analyses the crude spatial image to determine whether one or more light beams are being obscured by the presence of smoke.

For this type of sensor the use of multiple independently controllable light beams is essential in order to: - ensure good geometric coverage of the entire potential hot gas and smoke layer within the room; - increase the probability of receiving acceptable signals; - reduce the risk of the obscuration of an individual beam by a person or an object causing a false alarm.

Advantages of such a sensor over other known techniques for detecting fires, particularly over the widely used ceiling mounted point fire detectors, include the following.

- Improved fire detection, since the effect of the smoke concentration in most of the hot gas layer may be analysed. This is very important where the air movement in a volume may be high and where there is a risk that smoke will not reach a given point.

- The installation cost may be lower, since smaller numbers of wall-mounted sensors may need be used, and no alignment operations are necessary.

- It is suitable for use with uneven or beamed ceilings.

- It is not ceiling mounted, which reduces the visual impact in aesthetically important applications (eg.

historic buildings), and is more practical for maintenance where access to the ceiling may be blocked by objects cluttering the floor area.

By a further feature of the invention the electronic means 4a to 4g may be capable of varying the electrical power input to the light emitters 2a to 2g under the control of the microcomputer 5, preferably over a range of at least two orders of magnitude. Dependent on the path length and other factors related to the application, the quantity of light received will vary greatly for a given amount of light emitted. The control of the power input to each emitter therefore helps to maintain a more constant signal level at the receiver. It also permits the mean power consumption of the sensor to be reduced, since the microcomputer only need emit sufficient light along a given path as is needed to achieve an appropriate signal to noise ratio. The control of the power input may be achieved by a number of known techniques.

A preferred method of controlling the power input, which is capable of providing a high level of stability where light emitting diodes are used as emitters, is to separately charge a capacitor associated with each emitter with a known quantity of charge. This known quantity of charge is then discharged through the emitter during the period when it is pulse driven. A typical pulse period to provide a optimum signal to noise ratio in this type of sensor lies in the range from 10 to 50 microseconds. The stored charge may be conveniently controlled by the microcomputer by varying the time for which each capacitor is charged from a constant current source, with a maximum time significantly longer than the pulse period, eg 10 milliseconds.

The preferred wavelengths in which the light emitters 2a to 2g emit are in the infra-red between 800nm and 1000nm.

These are outside of the visible band, are known to provide a good sensitivity to smoke, and efficient light emitters and receivers are readily available. By a further feature of the invention one or more of the light emitters may emit at a predominant wavelength which is different from one or more of the other light emitters. The obscuration caused by certain types of smoke is known to be strongly dependent on wavelength. With a suitable disposition of the light emitters 2a to 2g and the optical elements 3a to 3g this permits the microcomputer to compare the light obscuration along similar paths at different wavelengths.

Suitable light emitters 2a to 2g would be 880nm GaAlAs light emitting diodes, available from a number of suppliers, eg the SFH484-2 from Siemens AG (Germany). Suitable optical means 3a to 3g and 9 would be specially formed lenses, injection moulded from acrylic plastic. The microcomputer 5 may conveniently be a single-chip CMOS microcomputer with an on-chip analogue to digital convertor, such as the MC68HC05B6 supplied by Motorola Inc (USA). Suitable light receivers 8 would be infra-red photodiodes, such as the BPW41D supplied by Zetex Ltd (UK). The electronic means 2a to 2g and 10 would be constructed from appropriate electronics components to perform the functions described such as would be obvious to one skilled in electronics design. The light barrier 7 could be formed from a sheet of appropriate metal.

It will be understood that the foregoing is given by way of example only. The number of emitters employed, their angles of emission, and the relative directions in which the axes of the light beams are disposed may be varied to suit the particulars of the product design and the fire detection application. The principles of operation described are not fundamentally dependent on the details of the electronics circuitry, or on the specification of the components, and could be realised using a variety of means, for example the functions of the microcomputer 5 could be carried out by an application specific integrated circuit. Other features may also need to be included in the sensor designs as disclosed in order to realise practical fire detectors, for example additional optical, electronic and software functions to further enhance the signal to noise ratio and to provide for interfacing to an appropriate system.

It will be obvious that the sensors, although primarily intended for sensing the presence of smoke will be capable of sensing other factors which may change the intensity of the emitted beams. This may include the obscuration of the beams and/or scattering or reflection of the beams by intervening objects, animals, or people. In may also include the sensing of refractive index discontinuities in the air through which the beams pass, for example caused by thermal effects.

Claims (8)

1. A sensor capable of sensing the presence of smoke by analysing changes in the amount of light received at the sensor caused by variations in the obscuration of a beam of light, wherein the sensor emits a single beam of pulsed light with the majority of the beam's energy being concentrated within an angle of less than 10 degrees in at least one axis, characterised in that the sensor includes at least one light receiver and more than one light emitter and the predominant direction of emission of the said beam may be altered by the selective use of one or more of the said light emitters under the control of a microcomputer or equivalent means in conjunction with suitable electronic and optical means, and wherein for each direction of emission the quantity of light received may be analysed by the said microcomputer or equivalent means.

2. A sensor according to claim 1 in which the light intensity of the emitted beam may be altered under the control of the microcomputer or equivalent means in conjunction with suitable electronic means, in order to help maintain a more constant signal level at the light receiver and/or to minimise the power consumption of the sensor.

3. A single-path optical beam smoke sensor according to claim 1 or claim 2 wherein the received light signals which are used to sense the presence of smoke result from a pulsed beam of light emitted by the sensor which is reflected by a retro-reflector positioned in the neighbourhood of the sensor, and wherein the direction of the beam is controlled in order to estimate the quantity of light received from the retro-reflector and/or in order to maximise the quantity of light received by the sensor.

4. A single-path optical beam smoke sensor according to claim 1 or claim 2 wherein the sensor is used in conjunction with an auxiliary optical receiver/emitter unit, and the received light signals which are used to sense the presence of smoke result from a pulsed beam of light emitted by the auxiliary receiver/emitter unit in response to the reception of a pulsed beam of light received from the sensor, and wherein the direction of the beam is controlled in order to maximise the quantity of light received by the auxiliary receiver/emitter unit.

5. An imaging volumetric optical beam smoke sensor according to claim 1 or claim 2 wherein the received light signals which are used to sense the presence of smoke result from pulsed beams of light separately emitted by the sensor which are reflected or scattered from the internal surfaces of the volume in which the sensor is installed, and wherein the direction of the emitted beam is controlled in order to analyse changes in the light obscuration along paths through different parts of the said volume.

6. A sensor according any one of claims 1 to 5 wherein the light emitters emit in the infra-red at wavelengths between 800nm and 1000nm.

7. A sensor according claim 5 wherein one or more of the light emitters emits in the infra-red at a predominant wavelength which is different from one or more of the other light emitters, in order to permit the light obscuration along similar paths to be compared for different wavelengths.

8. A sensor substantially as herein described with reference to figure 1 and/or figure 2 of the accompanying drawing.