EP0940789B1 - Method for fire detection - Google Patents

Abstract

The fire alarm method involves transmission of ultrasound signals from one transmitter and microwave- or optical- signals, or a combination of all three different wave-types from an additional transmitter. Doppler signals reflected from objects are received by receivers assigned to the transmitters, and the received signals are evaluated separately and in combination, so that a varied reception signal is detected for one type of wave, in the case that a reception signal for one or more other types of wave shows no movement, smoke, particles or variation in the gas composition in the detection zone.

Description

The invention relates to a method for Fire alarm, whereby emitted signals between a transmitter unit and a receiving unit the area to be monitored happen and be reflected.

1.) Anfangsphase (kleine unsichtbare Rauchpartikel),

2.) Schwelphase (größere sichtbare Rauchpartikel),

3.) Flammenphase (elektromagnetische Energie und Schallenergie, Gase),

4.) Hitzephase (Wärme, Gase).

Automatic fire detectors are of fundamental importance in security systems in private households and in public facilities. In order to initiate extinguishing measures quickly, it is important that a source of fire is detected and localized as soon as possible. A fire can be divided into the following phases, whereby different detectable materials are released, [1],:

1.) initial phase (small invisible smoke particles),

2.) Smoldering phase (larger visible smoke particles),

3.) flame phase (electromagnetic energy and sound energy, gases),

4.) Heat phase (heat, gases).

Depending on the type of fire, the four phases run in succession or almost in parallel. For optimal fire detection with high sensitivity and at the same time high Security against false alarms is therefore a detection as many fire characteristics as possible, such as smoke, Gases, heat, necessary because each component is taken on its own does not necessarily have to be caused by a fire (e.g. cigarette smoke, Solar radiation, radiator heat). For this This is why conventional fire detectors combine different systems for the detection of smoke and gases and for temperature determination [2].

The majority of building fires start due to the high Plastic part of the facility, with strong smoke and Flue gas development. Play for early fire detection Smoke detectors play an important role. Here are mainly optical detectors and ionization detectors used [3]. Ionization detectors are subject to radioactive emissions α-radiation have strict requirements and are, among other things, in the cannot be used in private areas. With optical detectors the evaluation of air samples in a size up to 4 cm Smoke chamber where the air samples have to be taken. For optimal room surveillance, therefore, favorable air circulation or the number of The detector must be large and distributed evenly throughout the room. At the same time, as mentioned, smoke detection is for one Fire detection is not always sufficient, so that for heat detection additional temperature sensors must be attached, however, the rules of attachment are those of the smoke detectors oppose (not near radiators and just not near ventilation). Another disadvantage Optical detectors are susceptible to false ones Alarms: From a non-constant signal level of the transmitted Signals, from temperature fluctuations, signs of aging or various diffraction and diffraction phenomena results in a weakening of the light beam, from which wrongly on smoke on the way from the transmitter to the receiver is closed.

From the patent US-A-4 625 199 is a motion alarm system known that from a combination of microwaves and Ultrasound systems exist to determine the movement of certain Detect targets and exclude the other targets. The microwave and ultrasound reception signals from the monitoring room are processed individually to produce ultrasonic motion detection signals and microwave detection signals receive.

Furthermore, from the European patent application EP-A-0 103 375 the use of a sensor for fire detection by linking infrared signals and ultrasound signals known.

The aim of meaningful early fire detection is to: Sensitivity to characteristics in the early phase of the Increase fire and at the same time the possibility of one minimize false alarms.

This task is solved by the combination of features of claim 1.

In the invention, a combined fire detector on microwave and ultrasound base, or in combination with Light described, which according to the invention microwave and Emits ultrasound signals and the interference of normally, d. H. in the case of non-fire, given Doppler frequency - coherence [4] between the microwave reception signals and the Uses ultrasound reception signals as a recognition criterion. With the reversal of the principle of coincidence, the new procedure be referred to as an "anti-incidence" procedure.

The combination of the acoustic detector with a microwave sensor has the advantage that microwave signals as Reference can be used because of smoke or smoke Heat phenomena on the propagation path are hardly touched; while the signals reflected by objects in space, if they are about to be monitored from the same place Be emitted space, coherent Doppler frequencies own, i.e. they are reciprocal to each Ultrasonic and microwave wavelength. So kick in Space fluctuating in the ultrasonic signals which are not observed in the microwave signal, so is this an indication of corresponding fluctuations in the propagation medium, which is typical of an emerging fire are like rising air and combustion gases. The exploitation of the microwave-ultrasound coincidence principle sensitive and reliable fire detection. moreover becomes an additional important motion detector [5; 6; 7] Use of the microwave ultrasound Doppler frequency principle proposed.

Advantageous configurations can be found in the subclaims become.

1) Rauch auf dem Übertragungsweg vom Sensor zum Empfänger ruft einen Pegelverlust der Schallwelle aufgrund von Absorption und Diffraktion hervor,

2) eine Wärmeentstehung bewirkt eine veränderte Schallgeschwindigkeit und Dopplerverschiebungen durch die Luftturbulenzen auf dem Übertragungsweg,

3) Flammen verursachen charakteristische rhythmische Veränderungen der Dopplerfrequenz (Flackerbewegungen).

Thanks to intelligent signal evaluation, several fire characteristics can be extracted from the sound signal alone:

1) Smoke on the transmission path from the sensor to the receiver causes a level loss of the sound wave due to absorption and diffraction,

2) heat generation causes a change in the speed of sound and Doppler shifts due to the air turbulence on the transmission path,

3) Flames cause characteristic rhythmic changes in the Doppler frequency (flickering movements).

By evaluating the signal amplitude and the spectrum is a limited fire detection with only one Realizable sensor type. The high probability of false alarms however, makes an acoustic fire detector unreliable. The reliability of the fire detection is determined by the parallel use of a second sensor on another physical principle, significantly increased. This can be, for example, a microwave detector.

In the combined fire detector types of waveforms can (CW, FMCW, pulse; C ontinuous W ave F requence M odulated-CW) are used. If unmodulated continuous wave signals (CW signals) are used, basically only movements can be detected. The Doppler shifts may by determining the signal phase or by means of spectral analysis with FFT (F ast F individual Fourier ransformation T) are evaluated. The significance of the statement is increased if the alarm criterion for several successive evaluation areas must be met before a fire is concluded.

Different types of fire have typical spectra, which in stored in a database and used for comparison can. Neural networks can be used for classification or methods of fuzzy logic can be used advantageously [8th]. The classification of the signal parameters of the ultrasound and Microwave receive signals for decision making, for example Signal level, propagation damping, flicker frequency, Detected object distances is often only taken into account certain general rules and their processing using "fuzzy" logic or a trained one Decision network possible. With the use of frequency-modulated continuous wave signals (FMCW signals) the distance to objects can be determined. A periodically frequency-modulated signal is transmitted reflected on an object and reaches the recipient: The frequency of the difference signal from the transmit and receive signals is then proportional to the object distance. If ultrasound and microwave signals to a common reflector can be aligned, for example by the distance from the ultrasonic sensor at a different distance the reflector detected by the microwave signal become. So the distance to the respective object must be beforehand not be known. As compared to ultrasonic FMCW systems a large relative bandwidth to microwave systems is feasible and the distance resolution, for example by interpolation of the spectrum, easily in the range of a few Millimeters can be done by tracking the distance a strong reflection object, for example one vertical wall, which by the microwave reference the change in the speed of sound was recognized as fixed be tracked due to heat generation. If flicker frequencies are detected in the spectrum at the same time, which in turn are missing in the microwave signal, can Fire can be closed as the cause. A fuzzy evaluation, the use of one database or the integration over several Evaluation areas (sweep intervals) also increase the significance here the statement. With FMCW signals also a Localization of the source of the fire over a distance of several Meters away possible. A delimitation of the coverage area is done with simple means such as Low pass, reached.

Can be distinguished due to their distance and realize in particular in the context of the present invention, in which the objects a fluctuating interference is present at distance resolution method such as run-time method, FMCW, correlation process with PN code (P seudo- N oise) objects. This enables localized sources of fire to be located. Separate ultrasonic and microwave sensors even enable two-dimensional detection.

The tools of the described evaluation, such as quadrature demodulation, arctangent calculation, phase tracking, threshold evaluation, possibly Hilbert transformation and FFT, are used in the same way for motion detection, so that the two alarm detectors, fire and intrusion alarm, are combined without additional hardware to form a universal, powerful room monitoring system can be. The evaluation principle described can advantageously implement with digital signals, preferably igitalen on a microcontroller, or D S ignal p ROCESSOR (DSP).
The fire detector described has the character of a route sensor: the quantities to be sensed influence the emitted signal on the propagation path. The sensitivity range of the sensor thus corresponds to its detection range, which results from the range and the opening angle of the radiation. This means that significantly fewer fire detectors are required to cover the entire room area than with conventional point detectors, such as optical detectors and ionization detectors. An optimal attachment of the sensors can be determined by smoke tests with test fires, which results in a specific sensitivity profile for the room. In normal-sized rooms, one detector is often sufficient, which also includes the function of the motion detector.

In combination with an active or passive IR detector (I nfra r ot) with fan beams a lateral assignment of the seat of the fire becomes possible, thus its localization is improved. With the use of an active IR detector, a distance measurement to spatially positioned objects is also possible. In extreme cases, it can be used to calculate how the microwave signals should look - this results in a "virtual microwave Doppler sensor".

An active IR detector could therefore be used instead of the microwave sensor take over the "control" of the ultrasonic sensor or used in addition to increased security against interference signals become.

Fig 1 zeigt ein Blockschaltbild eines kombinierten Mikrowellen- Ultraschall-Brandmelders mit Fuzzy-Auswertungauf einem Mikrocontroller und Fuzzy-Kompressor,

Fig 2 zeigt ein Blockschaltbild eines kombinierten Mikrowellen- Ultraschall-Brandmelders mit Auswertung auf einem DSP (Digitaler Signal Prozessor) mit FMCW-Signalen,

Fig 3 zeigt den Phasengang der Ultraschall- und Mikrowellen-Empfangssignale für vier verschiedene Scenarien und

Fig 4 zeigt das Verhalten des Melders bei offenem Buchenholzbrand.

Exemplary embodiments are described below with the aid of schematic figures:

1 shows a block diagram of a combined microwave-ultrasound fire detector with fuzzy evaluation on a microcontroller and fuzzy compressor,

2 shows a block diagram of a combined microwave-ultrasound fire detector with evaluation on a DSP (digital signal processor) with FMCW signals,

3 shows the phase response of the ultrasound and microwave reception signals for four different scenarios and

Fig. 4 shows the behavior of the detector when the beechwood fire is open.

1 shows the block diagram of a microcontroller-based (eg 8-bit microcontroller SAB 80537) combined fire detector with fuzzy evaluation. The transmission channel SK of the ultrasonic sensor comprises a signal generator G and modules for signal conditioning of the transmission signal sus (t) (transmission amplifier V, adaptation A). After the receive converter EM, the receive signal e _us (t) is divided into two orthogonal components i _us (t) and q _us (t) after amplification and bandpass filtering BP by quadrature demodulation QDM. The ultrasound transmitter transducers SW and reception transducers EW are preferably ultrasound transducers with high quality and sensitivity, such as piezoceramic bending transducers. The microwave Doppler sensor (e.g. 2.5; 5.8; 10; 24 GHz) is usually subject to relatively low requirements because the monitored distances and the resolution requirements are moderate.

The demodulated received signals of the ultrasonic sensor and the microwave sensor are read in 4 channels (2 ultrasonic signals _ius , q _us , 2 microwave signals i _mw , q _mw ) alternately in blocks of, for example, 256 points via the internal A / D converter. For 40 kHz ultrasound and 24 GHz microwave continuous wave signals with a monitored speed range of 2 cm / s to 2 m / s, Doppler frequencies of approximately 5 to 480 Hz for ultrasound and of approximately 3 to 320 Hz for microwave occur. The movements caused by fire appear in the range up to approx. 200 Hz. Accordingly, a sampling frequency of approx. 1 kHz can be used.

The amplitude of the received signals is tracked continuously and performed a Doppler evaluation. The phase of the complex Ultrasonic and microwave signals are calculated using an arctangent certainly. The Doppler frequency results from the derivation of the phase, the direction of movement from the Sign. By tracking the signal phase in successive Sampling points is a detection of flicker (pendulum) movements easily possible. The extracted features Doppler frequency, flickering movement and intensity profile are used for alarm decision over several consecutive Evaluation intervals followed.

Alternatively, the Doppler evaluation can only be carried out if a change is registered that is greater than is a set tolerance range. As another option results in the detection of "irregularities" Keep ultrasound sensor continuously active and the microwave sensor to switch on as a control only if there are any significant ones Attenuation and fluctuations on the way of propagation of the sound signal can be registered.

For a fuzzy evaluation, algorithms for fuzzification, Rule processing and defuzzification provided become. This could be done with a pure software implementation or can be achieved with a hardware implementation. As Mixed form is the use of a fuzzy logic coprocessor (e.g. SAE 81C199) who runs these processes independently processed and thus relieves the load on the microcontroller. The membership functions and the rules of a fuzzy set are written in address-coded form in so-called knowledge bases and in EPROMs filed.

2 shows the block diagram of a system with FMCW evaluation on a DSP (eg DSP56000). In addition to the information from the intensity and Doppler shift of the received signals, the deviation of the distance-proportional frequencies of the difference signals from ultrasound and microwave can be used to infer interference on the propagation path and the source of the fire can be localized. The evaluation can largely be done on the DSP to simplify the hardware requirements. The ultrasound signal S _us (t) can be generated by software and output via a D / A converter. The ultrasound received signal e _us (t) and the downmixed microwave difference signal d _mw (t) are read in via an A / D converter board. The difference signal d _us (t) is formed in the receiver by software multiplication, so that the ultrasound sensor now only consists of the ultrasound transmitter transducer SW and the receiver transducer EW (preferably piezoceramic bending oscillators). A sensor with good linearity in frequency modulation is preferably used as the microwave FMCW module. Using an FFT over blocks of, for example, 1024 sampling points, the (real) spectrum of both received signals is formed and examined for maxima.

With the combined Doppler sensor, measurements are available for examination the detection properties of the combined fire detector Have been carried out. In Figure 3 is an evaluation feature the phase development of the received signals is plotted, which indicates the sensitivity of the sensor to changes demonstrated the transmission route. In Fig 3a) are minimal Recognize phase fluctuations due to background noise. Fig. 3b) shows the influence of air movements on the Phase of the ultrasound signal. In contrast, smoke causes strong Fluctuations in the ultrasound phase, as shown in Fig. 3c). In contrast, the microwave phase remains almost unaffected. Under The detection takes into account the amplitude information additional support for typical fire phenomena. Compared for this purpose in Fig. 3d) is the evaluation of the received signals shown when moving people.

The detection capability was then determined under test conditions in accordance with EN 54/7 of the combined sensor based on a test fire TF 1, beech wood fire - small bright particles, examined. The sensor was placed under the ceiling, its distance to the source of the fire was 3m. 4a) and 4b) are the amplitudes of the received signals over a period of 5 minutes. The ultrasound signal shows considerable amplitudes, similar to that caused by movement of people could be. The enables a clear distinction Comparison with the microwave signal together with an evaluation fluctuations in Doppler frequency; see Fig 4c).

The results confirm the potential of a combined Ultrasonic microwave sensor for fire detection. For a classification of the fire characteristics (signal level, propagation damping, Flicker frequency) can be advantageous neural Networks or methods of fuzzy logic are applied, comparisons with cataloged spectra of different Enable types of fire, see [4].

This can be done by connecting to a house installation bus System integrated in a complex room surveillance system become.

literature

[1] K. Bartels: "Fire and intrusion detection systems", RPB paperbacks No. 5, Franzis'-Verlag 1980

[2] W. Friedl (ed.): "Minimize false alarms: fire and Intrusion alarm systems ", VDE publishing house 1994

[3] R. Rivoir: "Smoke detection", First European School on Sensors (ESS94), September 12-17, 1994, Lecce, Italy, pp. 120-129

[4] W. Heywang, M. Guntersdorfer, P. Kleinschmidt: "The Doppler shift of reflection from electromagnetic and intrusion alarm system evaluating ultrasound radiation ", Patent DE2613845, 10/26/78

[5] P. Kleinschmidt: "Device for burglar protection", invention report v. 19/6/78

[6] V. Mágori: "Procedures for the suppression of pigeons and Direction evaluation Brei Doppler motion detectors ", invention report v. 11/7/78

[7] H. Ruser, M. Vossiek, V. Mágori, H.-R. Tränkler: "Combined ultrasonic microwave sensor for the reliable Detection of presence and movement ", Sensor'97, May 13-15, 1997, Nuremberg, vol. 1, pp. 229-234

[8] H.C. Müller, A. Fischer: "A robust fire detection algorithm for temperature and optical smoke density using fuzzy logic ", Proc. 1995 Carnahan Conf. on Security Technology, Sanderstead, UK, pp. 197-204, 1995

Claims (13)

1. Method for a fire alarm,
   in which

   ultrasound signals are transmitted from at least one transmitter and microwave or light signals are transmitted from at least one further transmitter, or a combination of all three different types of waves are transmitted,

   Doppler signals which are reflected from objects are received by receivers associated with the transmitters,

   the received signals are evaluated individually and in combination, so that, if severe fluctuations are detected in the amplitude and frequency of the output signals from the receivers for the ultrasound sensors while at the same time the amplitude and frequency fluctuation in the output signals from the receivers of the microwave sensors is small, it is deduced that this is caused by a fire.
2. Method according to Claim 1,
   in which microwave and ultrasound signals are used with in each case at least one transmitter and one associated receiver.
3. Method according to one of the preceding claims,
   in which, if the frequencies of the microwave and ultrasound output signals from the receivers do not match, within a tolerance band, it is deduced that the movement of a person or object is the cause of the external influences.
4. Method according to one of the preceding claims,
   in which the features of the output signals from the receivers are extracted on the basis of threshold evaluation.
5. Method according to one of the preceding claims,
   in which range-selective transmission signals, such as pulses or frequency-modulated signals, are used in addition to the amplitude and the Doppler frequency to determine the distance to powerful reflectors in the common detection area of the sensors.
6. Method according to Claim 5,
   in which changes to the signal transmission path which are typical of a fire are deduced from reflectors being detected at a different distance by an ultrasound or light sensor than by a microwave sensor.
7. Method according to Claims 5 or 6,
   in which a source of a fire is localized by evaluation of the range information in the output signals from the ultrasound or light sensors.
8. Method according to one of the preceding claims,
   in which the additional use of active or passive infrared detectors with beams allows lateral association with the source of the fire, so that its localization is improved.
9. Method according to one of the preceding claims,
   in which typical features of microwave and ultrasound output signals from different external influences are stored as patterns in a databank, and are compared with the features of the actual output signals in order to decide on whether a fire has occurred.
10. Method according to one of the preceding claims,
    in which active infrared detectors are used to carry out distance measurements to spatially positioned objects in a calibration phase, in order to calculate how the microwave signals should appear, and pattern signals are produced in this way, without using microwave sensors, for evaluation of the output signals from the ultrasound sensors with regard to a fire situation.
11. Method according to one of the preceding claims,
    in which typical features of external influences are used for definition of feature spaces, corresponding association functions and association rules, and the features of the current output signals from the association are used to decide on whether a fire situation has occurred, using fuzzy logic methods.
12. Method according to one of the preceding claims,
    in which the features of the current output signals are assessed using neural networks.
13. Method according to one of the preceding claims,
    in which the combination of a movement alarm and fire alarm, and hence a universal room monitoring system, is produced using the same hardware and signal processing methods, but by reversing the evaluation principle.

Images