EP0654770A1 - Device for early detection of fires - Google Patents

Abstract

The arrangement contains a plurality of detectors which are connected to a central station and of which some are fitted with at least two sensors (1, 2) for monitoring different fire characteristics. One sensor (1) is preferably a thermal sensor and the other sensor (2) is an optical sensor. In addition, the arrangement contains means for processing the signals of the sensors. These means are arranged in a decentralised manner in the detectors and they contain a microcontroller (MCU) for conditioning the sensor signals and for signal processing for the purpose of obtaining danger signals. The danger signals are obtained in a neuronal network (NN). <IMAGE>

Description

The present invention relates to an arrangement for the early detection of fires, with a plurality of detectors connected to a control center, some of which are equipped with at least two sensors for monitoring different fire parameters, and with means for processing the signals from the sensors.

As the individual sensors monitor different parameters in detectors with multiple sensors, the response behavior of the detectors can be better balanced and the error alarm rate per detection point can be significantly reduced. In addition, the redundancy associated with multiple monitoring increases reliability and leads to a balance between the weak and strong points that occur with single detectors.

The invention now aims to further reduce the error alarm rate per detection point and, at the same time, to obtain the earliest possible detection capability.

This object is achieved according to the invention in that the means mentioned for processing the signals from the sensors are arranged decentrally in the detectors and have a microcontroller for processing the sensor signals and for signal processing for the purpose of obtaining hazard signals, and in that the hazard signals are obtained in a neuronal manner Network.

In the arrangement according to the invention, the signal processing is thus shifted from the control center into the detectors and is decentralized, as a result of which the limitation of the communication bandwidth of the usual connections between the control center and detectors has no influence. In addition, the observation length of the signals is not subject to any restrictions and the possibility of overloading the control center is practically excluded. The high redundancy of the system also has the The advantage that the detectors can trigger an alarm themselves if the main processor fails at the control center.

The use of the neural network has the advantage that the reliability of the detector function is generally improved in that there is a wide range of possibilities for linking the different signal signatures, that is to say the detection patterns, and can also be optimally used in the neural network.

Fig. 1

ein Übersichtsdiagramm der Signalverarbeitung im Melder,

Fig. 2a, b

ein Schema der beiden Signalpfade der Signalverarbeitung; und

Fig. 3

ein Diagramm des neuronalen Netzwerks der Signalverarbeitung.

The invention is explained in more detail below with the aid of an exemplary embodiment and the drawings; shows:

Fig. 1

an overview diagram of signal processing in the detector,

2a, b

a schematic of the two signal paths of signal processing; and

Fig. 3

a diagram of the neural network of signal processing.

Fig. 1 shows an overview of the signal processing in the detector, which can be divided into five stages S1 to S5. The first stage S1 consists of the sensor hardware and essentially contains a thermal sensor 1 formed by an NTC sensor, an optical sensor 2 formed by a light pulse transmitter and a light pulse receiver, a bias network 3 for the thermal sensor 1 and an ASIC 4. About the sensor -Hardware also includes an A / D converter 5 of a microcontroller MCU.

In a known manner, the MCU has a ROM mask which contains the operating system and the sensor software of the detector and thus controls all processes at the functional level, that is to say the sensor control, the signal processing, and the addressing and communication with the control center. The ASIC 5 contains all amplifiers and filters for the signal of the light pulse receiver, a one-chip temperature sensor, the control electronics for the light pulse transmitter, a quartz oscillator and the start / power management as well as the line monitoring for the MCU. There is a bidirectional, serial data bus and various control lines between the MCU and the ASIC 4.

The signals are processed in the second stage S2 following the A / D converter 5, with various compensations being used to try to obtain the most accurate possible representation of the real measurement variables. In the third stage S3, signal signatures or criteria are extracted, which are then condensed in the fourth stage S4 in a neural network NN to form a scalar hazard signal and are assigned to a hazard level. In the fifth stage S5, the decision about the definitive danger level is finally made in a verification stage 6 and forwarded together with the functional state or status to the communication interface of the MCU.

1, the first three stages S1 to S3 are passed through separately from the signal from the thermal sensor 1 and from the signal from the optical sensor 2, which is symbolized in the figure by two signal paths, a "thermal" and an "optical" path, which are then combined in the fourth stage S4, that is to say in the neural network. The signal flow of the two paths through the stages S1 to S3 is shown in FIGS. 2a and 2b, and the neural network NN is shown in detail in FIG. 3.

The thermal and then the optical signal path will now be described in more detail below: The NTC temperature sensor 1 is operated in a pulsed manner via the biasing network 3 and the NTC voltage is fed to the A / D converter 5. The NTC temperature data are subsequently analyzed in a stage 7, interruptions and short circuits being recognized. In step 7, the influence of small driver voltage changes on the measured value is also compensated to increase the measuring accuracy. Any interference peaks are removed in the following “anti-EMI” algorithm 8. This limits the signal change from one measurement to the next to certain values stored in the data memory of the MCU. Normal fire signals pass this algorithm unchanged.

The output signal of the A / D converter is then converted into a temperature value in an linearization stage 9 using an interpolation table in accordance with the characteristics of the NTC sensor. Then in a block 10 the heat dissipation through connecting wires and plastic wall and in a block 11 the Heat capacity of the NTC sensor 1 compensated. The output signals of blocks 10 and 11 then pass through a digital filter bank 12 and are finally linked in a stage 13 with parameters. At the output of stage 13 and thus at the end of the thermal path there are then several signature signals or criteria S1 to Sm which are dependent on the NTC signal and thus on the temperature.

In the optical signal path, a pulse generator 14, which generates a current pulse of almost 100 μs every 3s, drives an infrared light-emitting diode 15 which forms the light pulse transmitter and sends a light pulse into the optical scattering space. The light scattered by any smoke is collected by a lens and directed to a receiver photodiode 15 '. The resulting photocurrent is integrated by an integrator 16 in synchronism with the transmission pulse. The subsequent, still differential voltage amplifier 17 offers several selectable gain settings. The coarse detector adjustment is then carried out. A so-called AMB filter 18 eliminates DC components and low-frequency interference from the signal. High-frequency interference has already been eliminated by integrator 17. A single unipolar signal appears at the output of the AMB filter 18 and is further amplified by a voltage amplifier 19.

The output signal of the amplifier 19 is converted into digital data in the A / D converter 5, with which the software signal processing begins (FIG. 1, stage S2). The effective signal swing is now determined by forming the difference in a stage 20 between a light and a dark measurement. This arrives in a block 21 and, thanks to the availability of the ASIC temperature, can be corrected there in such a way that the temperature drops of the optoelectronic components are largely compensated for. The software-based fine adjustment, which is also carried out in block 21, serves as the last and practically continuous adjustment of the signals to a desired size. In the next block 22, tracking eliminates those signal components which are caused by very slow environmental influences (for example dustiness) and which would produce a false smoke signal over time and thus change the sensitivity.

The result of the previous processing steps is a quantity that represents the effective, filtered, adjusted, temperature-compensated and updated smoke value and forms the immediate reference for determining the hazard level. The last link (block 23) in the optical signal processing are algorithms controlled by various parameter sets, which assess the temporal behavior of the variable representing the smoke value. At the end of the optical signal processing path, the signature signals Sm + 1 to Sn are then available.

The signature signals S1 to Sn of the thermal and the optical path form the input level L0 of a layered, neural network NN, which is shown in FIG. 3. It can be seen from the representation of the neural network NN in FIG. 1 that these input variables are dependent either on the temperature signal (T) or on the optical signal (O) or both. In addition to the input level L0, the network also has other levels L1 to L5 with so-called neurons or nodes. In these, the input variables weighted with parameters are subjected to an addition and a maximum and / or minimum linkage. The addition takes place in the neurons labeled A and the maximum and / or minimum linkage in the neurons labeled M.

die nach dem Prinzip "alles gehört dem Stärksten" arbeitet.The maximum linkage is the nonlinear network function:

$\text{yi = max (w1 * x1, w2 * x2, ..., wn * xn), [xi = input value, yi = output value]}$

who works on the principle that "everything belongs to the strongest".

Zwischen den Neuronen sind grundsätzlich alle Verbindungen möglich. In einer Lernphase während der Entwicklung des Melders kann das Netzwerk in eine Lernumgebung eingebunden werden. Dabei werden sich durch den Lerneffekt des Netzwerks bestimmte Verbindungen als bevorzugt erweisen und sich verstärken und andere werden gleichsam verkümmern. Alternativ kann das Netzwerk auch ohne Lernphase konstruiert werden. In beiden Fällen werden aus Sicherheitsgründen im Betrieb die Gewichte des Netzwerks eingefroren.The addition is the dot product:

$\text{yi = Σ wi * xi, [xi = input value, yi = output value].}$

In principle, all connections are possible between the neurons. In a learning phase during the development of the detector, the network can be integrated into a learning environment. Due to the learning effect of the network, certain connections will prove to be preferred and strengthen, and others will, at the same time, atrophy. Alternatively, the network can also be used without a learning phase be constructed. In both cases, the weights of the network are frozen during operation for security reasons.

Between the input and output levels L0 and L5 of the neural network NN, the respective input variables are concentrated on a single output variable, which represents a scalar danger signal. In a quantization stage 24, the hazard signal is assigned to one of several, for example at least three, security levels, and this signal assigned to one of the security levels is the output signal GS of the neural network NN.

Finally, the verification of the final danger level takes place in the verification level 6 downstream of the neural network. The corresponding output signal GSdef, together with the functional state (FIG. 1, "Status"), is communicated to the control center via the MCU communication interface.

• Die Messung der aktuellen ASIC-Temperatur mit Hilfe eines Einchip-Temperatursensors wurde bereits erwähnt. Diese Messung, die periodisch erfolgt, liefert einen Temperaturwert, mit dem die Temperaturgänge der optoelektronischen Bauteile softwaremässig kompensiert werde, so dass auch bei extremen Temperaturen zuverlässige Rauchdichtemessungen vorgenommen werden können.
• Die Funktionsweise der Signalnachführung wurde ebenfalls bereits erwähnt. Das Rauchdichtesignal wird von sehr niederfrequenten Anteilen befreit, um Einflüsse der Umwelt auszufiltern, die signifikant langsamer sind als Brandphänomene (beispielsweise Verstaubung). Damit wird eine sehr gute Langzeitkonstanz der Rauchempfindlichkeit erreicht.
• Regelmässig wird automatisch ein Selbsttest auf gewisse Fehler durchgeführt, der den Melder einer detaillierten Diagnose unterzieht.

Finally, some particularly advantageous properties and additional functions of the smoke detector described should be mentioned:

• The measurement of the current ASIC temperature using a single-chip temperature sensor has already been mentioned. This measurement, which is carried out periodically, provides a temperature value with which the temperature responses of the optoelectronic components are compensated for by software, so that reliable smoke density measurements can also be carried out at extreme temperatures.
• The functionality of signal tracking has also already been mentioned. The smoke density signal is freed from very low-frequency components in order to filter out environmental influences that are significantly slower than fire phenomena (e.g. dustiness). This ensures a very good long-term consistency in smoke sensitivity.
• A self-test for certain errors is carried out automatically, which subjects the detector to a detailed diagnosis.

Although the shift of signal processing from the control center to the detectors and the use of a neural network for signal processing for detectors with multiple sensors is particularly advantageous, these multiple sensors should not be understood as restrictive. Of course, detectors with only one sensor can also be designed in the manner described. It should also be mentioned that the neural network NN represents a very special type related to a fuzzy logic and could therefore also be replaced by a fuzzy logic.

• Diese neuronalen Netzwerke wären eine Art von Transversalfilter und hätten ein wesentlich geringeres Gedächtnis als rekursive Filter;
• am Ausgang jedes dieser neuronalen Netzwerke wäre nur je eine Signalsignatur pro Brandphänomen (Rauch, Temperatur) erhältlich, wogegen die vorgeschlagene Lösung S1 bis Sm Signalsignaturen für das Brandphänomen Temperatur und Sm+1 bis Sn Signalsignaturen für das Brandphänomen Rauch zur Verfügung stellt. Diese Mehrzahl von Signalsignaturen ist aber für die sichere Funktion des neuronalen Netzwerks NN (Fig. 3) sehr wichtig, weil man dieses dann so ausbilden kann, dass seine Funktionen voll verständlich und überblickbar sind. Und letzteres ist in einem Sicherheitssystem unbedingt erforderlich.

A very essential feature of the present arrangement is formed by the digital filter bank 12 and the block 23 (FIG. 1), it being possible in particular for the digital filter bank to contain recursive filters. If one were to use a neural network instead of this filter bank and / or block 23 and to supply this time pattern of the sensor signals sequentially, then one would have two major disadvantages compared to the proposed solution:

• These neural networks would be a type of transversal filter and would have a much smaller memory than recursive filters;
• at the output of each of these neural networks, only one signal signature per fire phenomenon (smoke, temperature) would be available, whereas the proposed solution S1 to Sm provides signal signatures for the temperature fire phenomenon and Sm + 1 to Sn signal signatures for the smoke fire phenomenon. However, this plurality of signal signatures is very important for the safe functioning of the neural network NN (FIG. 3), because it can then be designed in such a way that its functions are fully understandable and clear. And the latter is absolutely essential in a security system.

Claims (17)

Arrangement for early detection of fires, with a plurality of detectors connected to a control center, some of which are equipped with at least two sensors for monitoring different fire parameters, and with means for processing the signals from the sensors, characterized in that the said means for processing the signals from the sensors (1, 2) are arranged decentrally in the detectors and have a microcontroller (MCU) for processing the sensor signals and for signal processing for the purpose of obtaining hazard signals, and that the hazard signals are obtained in a neural network (NN) . Arrangement according to claim 1, characterized in that the signal processing for each of the two sensors (1, 2) has a separate path, and that the two paths are merged at the input of the neural network (NN). Arrangement according to claim 1 or 2, characterized in that the neural network (NN) has a plurality of levels (L1 to L5) with nodes (A, M) in which the input variables weighted with parameters are subjected to addition and maximum and / or minimum linkage become. Arrangement according to claim 3, characterized in that the microcontroller (MCU) has a mask with the operating system and the sensor software of the detector and a data memory, and that the microcontroller is assigned an ASIC (4), the amplifier and filter for the signal of the receiver of the optical sensor (2), a temperature sensor, the control electronics for the transmitter of the optical sensor and a quartz oscillator. Arrangement according to claim 3, characterized in that the thermal path a first stage (S1) with a bias network (3) for the operation of the thermal sensor (1) and with an A / D converter (5), a second stage (S2 ) for processing the signals and for possible compensations and a third stage (S3) for obtaining signal signatures, which form input variables for the neural network (NN). Arrangement according to claim 5, characterized in that the second stage (S2) has a block (7) for analyzing the output signals of the A / D converter (5) for possible errors and / or for compensating the influence of changes in the driver voltage on the measured value and / or a block (8) for removing interference peaks, a block (9) for converting the measured value into a temperature value and / or a block (10 or 11) for compensating the heat dissipation and / or the heat capacity. Arrangement according to claim 6, characterized in that in the block (8) for removing interference peaks, the signal change from one measurement to another is limited to specific values. Arrangement according to claim 5, characterized in that the third stage (S3) contains means for linking the output signals of the said elements, so that at the end of the thermal path various signature signals derived from the temperature signals are available. Arrangement according to claim 3, characterized in that the optical path has a first stage (S1) with a pulse generator (14) for driving the transmitter (15) and with an integrator (16) for the signal of the receiver (15 ') of the optical sensor (2), as well as with an A / D converter (5), a second stage (S2) for carrying out any compensation, and a third stage (S3) for obtaining signal signatures, which input variables for the neural network (NN) form. Arrangement according to claim 9, characterized in that the integrator (16) is followed by a voltage amplifier (17) for rough adjustment and this a filter (18) for selective detection of the received light pulse while suppressing interference signals. Arrangement according to claim 10, characterized in that the filter (18) performs a calculation of the signal pulse values before, after and during a light pulse. Arrangement according to claim 9 or 10, characterized in that the second stage (S2) has a block (20) for determining the signal swing, a block (21) for compensating the temperature drops of the optoelectronic components and / or for fine adjustment, and / or has a block (22) for compensating the background signal and for eliminating signal components composed of slow environmental influences, so that the output signal of the second stage represents a balanced, temperature-compensated and tracked smoke value. Arrangement according to Claim 9, characterized in that the third stage (S3) contains a block (23) for assessing the temporal behavior of the smoke value delivered by the second stage (S2) by means of filtering, and in that the smoke value signal thus filtered is a signature signal of the optical Path forms. Arrangement according to claims 5 and 9, characterized in that a concentration of the input variables takes place in the nodes (A, M) of the neural network (NN) and that a scalar danger signal is available at the output level (L5) of the network and in a quantization stage (24) is assigned to one of several danger levels. Arrangement according to claim 14, characterized in that the neural network (NN) is followed by a verification level (6) for verifying the definitive danger level. Arrangement according to claim 1, characterized in that the neural network (NN) is preceded by a digital filter bank (12) to which the signals of at least one type of sensor (1) are fed, and which have several signal signatures or at their output for the neural network Provides criteria (S1 to Sm) for the fire phenomenon in question. Arrangement according to claim 16, characterized in that the digital filter bank (12) contains recursive filters.

Images