EP0070449B1 - Method and device for increasing the reaction sensitivity and the disturbance security in a hazard, particularly a fire alarm installation - Google Patents

Abstract

A method and apparatus for increasing the response sensitivity and the interference resistance in an alarm system such as a fire alarm system which cyclically samples a plurality of alarm units in the system for obtaining a series of measured values from each alarm unit, the measured values being utilized to form a quiescent value which is stored in a quiescent value memory. With each sampling cycle a current comparison value is formed from the alarm measured value, the stored quiescent value, and a comparison value from a previous sampling cycle stored in a comparison value memory. The current comparison value is then written in the comparison value memory as the new comparison value. The current comparison value is compared with a rated limiting value, and if the comparison value is greater than or equal to the rated limiting value, a display unit is activated indicating an alarm. If the comparison value is less than the rated limiting value, a new quiescent value is formed from the measured value and the stored quiescent value and written into the quiescent value memory.

Description

The invention relates to a method for increasing the sensitivity and interference immunity in a hazard, in particular fire alarm system according to the preamble of claim 1 and an arrangement for performing this method.

Automatic alarm systems, for example fire alarm systems, in which certain parameters of a fire, such as smoke density, temperature, radiation, are evaluated and evaluated in order to detect an alarm or a fault, must also have a high level of immunity to interference with a high sensitivity. For example, each detector can have a threshold circuit which, when the specified fire parameter (threshold) is exceeded, emits an alarm signal to the control center. To increase interference immunity, timers were provided in the detectors or during an evaluation in the control center, which only display an alarm when the absolute threshold has been exceeded for a predetermined time. Several thresholds for changing a fire parameter above which the alarm is to be triggered have also been defined. A central evaluation of the detector signals has also improved because the alarm threshold could be adapted more easily to the respective requirements.

DE-OS 21 47 022 describes a circuit arrangement for achieving greater sensitivity in a fault value signaling system with fluctuating noise levels, in which the individual fault value detectors are queried one after the other and an average noise level value is formed from the signals emitted. The signals coming from the fault value detectors are compared with the mean interference levels, and a signal output device is applied when the signal is exceeded or fallen short of by an adjustable amount. In this known arrangement, a single mean interference level value is formed for all fault value detectors, which is used as a comparison value adapted to the environment (e.g. solar radiation) for the response of a fault value detector. This arrangement does not eliminate the individual interference level fluctuation of the detector queried in each case. With this arrangement, it is not possible for each detector to have a relevant mean, i.e. H. to create a respective detector idle value that can be updated over time.

A change in the idle signal of a detector, e.g. B. due to component aging, contamination, wetness etc. lead to changes in sensitivity due to the fixed evaluation thresholds and in the borderline cases to false response or ineffectiveness of the detector concerned. The object of the invention is therefore to provide a method for detector evaluation, in which a high level of interference immunity is ensured over a very long time with high sensitivity. Aging of the components and soiling of the detectors should not have an adverse effect on the sensitivity of the detectors.

This object is achieved according to the invention with respect to the method with the characterizing features of claim 1 and with respect to the arrangement with the characterizing features of claim 4.

With this method, an average detector measurement value is formed in a central evaluation device for each detector. This is derived as the detector idle value from the respective previous detector measured values and stored in a memory provided for this purpose as the current idle value. For each interrogation cycle, the difference between its current measured value and its last stored idle value is formed for each detector. These differences are used to form a current comparison value, which is stored in a comparison value memory provided for this purpose. This current comparison value is compared in a comparison device with a predetermined limit value. If this current comparison value is smaller than the specified limit value, a new rest value is formed from the current detector measured value and the stored rest value. This is written into the idle value memory for the next processing cycle. If the current comparison value is equal to or greater than the predetermined limit value, the comparison device controls a display device which displays an alarm or fault or another event.

With the help of the individual transmitted detector measurement values, a rest value is formed for each detector, either when the system is switched on or on request, e.g. B. during revision or maintenance, can be formed again. The rest value is expediently automatically updated with a large time constant of, for example, one day.

Instead of the absolute measured value, the difference between the detector measured value and the idle value is used for the detection of events. This difference is constant, e.g. B. at intervals of a few seconds or with each query cycle, newly determined, evaluated according to their size and evaluated. A comparison value is expediently derived from these differences, which controls a display device when a specified limit value is exceeded. The respective current comparison value is determined from the difference between the current measured value, the stored rest value and the stored comparison value, the difference being reduced by a constant value, so that smaller fluctuations in the measured value, which are below the constant value, do not lead to an event display. Because this result will integrated into a sum signal, ie the result is added to the last stored comparison value. The sum signal obtained in this way corresponds to the current comparison value. In order to limit this comparison value downwards, it is compared with zero in a comparison stage. If the result is above zero, it is saved directly in the comparison value memory for calculation in the next cycle. Otherwise zero is saved, just as the comparison value is zero in the first query cycle.

A detector idle value is expediently formed from the detector measured values with the aid of arithmetic logic units, which can be stored in a memory provided for this purpose, the first detector measured value corresponding to the idle value in the first interrogation cycle. It is used to form the comparison value. Using a parameter EPS (0 <EPS <1), the detector sensitivity for the formation of the resting value can be influenced by evaluating the difference (MW-RWA) with the specifiable parameter (EPS).

• Fig. 1 ein Meldermeßwertdiagramm für herkömmliche Brandmelder,
• Fig. ein Diagramm der Meldersignale nach dem erfindungsgemäßen Verfahren,
• Fig. 3 ein Ausführungsbeispiel im Blockschaltbild,
• Fig. 4 und 5 Details dieses Blockschaltbildes nach Fig. 3.

The mode of operation and an arrangement for carrying out the method are explained in more detail below with the aid of diagrams and a circuit example. It shows

• 1 is a detector measurement diagram for conventional fire detectors,
• 1 shows a diagram of the detector signals according to the method according to the invention,
• 3 shows an exemplary embodiment in a block diagram,
• 4 and 5 details of this block diagram of FIG. 3rd

The diagram shown in FIG. 1 a shows the course of a detector measured value MW as a function of time T. The diagram shows an alarm threshold, designated AISW, which runs parallel to the time axis. The detector itself has a rest value, which is drawn as a theoretical value as a straight line that rises slightly and is designated RW. In parallel, a disturbance threshold STSW is drawn at a constant distance CON. Around time T1, the detector measured value MW has increased significantly compared to its idle value RW. However, this increase in the measured value is not so great that it reaches the alarm threshold AISW, and therefore no alarm is displayed. If the detector idle value RW changes in the direction of the alarm threshold ALSW, an identical event would erroneously generate an alarm at time T2. The detector has become more sensitive by itself. The rise in the detector measured value MW, which is not greater than at time T1, now causes the alarm threshold ALSW to be exceeded, so that a false alarm is generated. Such a false alarm message is avoided with the method according to the invention, as will be explained in more detail later.

1 b, the detector value MW is also plotted over time T. The ALSW alarm threshold is shown parallel to the time axis. Likewise, the detector idle value RW is drawn as a straight line, which, however, tends towards the time axis, i. H. the detector idle value RW changes contrary to the alarm threshold ALSW. Parallel to the rest value RW, a straight line is drawn above it at a constant distance CON, which represents the interference threshold STSW. This diagram shows that the detector becomes less sensitive over time. At around time T1, a detector measured value MW occurs, which deviates considerably from the idle value RW. The detector measured value MW is so large that it exceeds the alarm threshold ALSW and is therefore recognized on alarm. Approximately at time T2, there is an increase in the measured value of the detector, which is approximately as large as at time T1, based on its idle value RW. However, the increase in the detector value is not large enough to reach or exceed the alarm threshold ALSW, so that no alarm is detected at time T2. In a conventional fire alarm system, the alarm is no longer recognized at time T2 because the idle value RW has developed away from the alarm threshold ALSW. With the fire alarm system according to the invention, this lost alarm is also recognized.

In order to ensure that the detectors respond significantly more reliably, the detector sensitivity should remain constant over a very long time using the method according to the invention. Therefore, the difference between the measured value and the rest value is considered instead of the absolute measured value. As mentioned at the beginning, the comparison value VW of the detector M is determined from the respective current detector measurement value MW, based on its idle value RW, and from its previously stored comparison value VWA and only then compared with a predetermined limit value GRW. In the exemplary embodiment, this is explained in more detail with reference to FIGS. 3 to 5 for the alarm case.

2a shows a measured detector value MW over time T, the time axis corresponding to the rest value RW. An interference threshold STSW is shown at a constant distance above the idle value RW. The previously defined alarm threshold for the detector measurement value MW is drawn in by a line parallel to the rest value RW at a corresponding height. Corresponding to the diagram in FIG. 2a, the sum signal SUS of the detector is shown as a function of the time T in FIG. 2b. The limit value for the sum signal SUS, at which an alarm is detected, is designated GRW. The diagrams for three typical measured signal images are explained below.

The normal detector measured value curve (MW over time T) is shown in FIG. 2a, and the sum signals SUS derived therefrom, which lead to the alarm detection, are shown below. According to the method according to the invention, both the size of the alarm evaluation and the detection of faults Detector measured value as well as the duration of the detector measured value are decisive. The detector measured value signal is evaluated with each sampling cycle. The difference (MW-RWA) is formed from the respective current detector measured value MW and the stored idle value RWA and is continuously determined, for example with every sampling cycle. This difference is related to a fixed value, namely a disturbance threshold STSW, in order not to add up smaller fluctuations in measured values which lie below this disturbance threshold to form an alarm signal.

The sum signal SUS according to FIG. 2b detects when it reaches or exceeds the predetermined threshold, limit value GRW, on alarm. In FIG. 2a, the measured value suddenly rises above the alarm threshold ALSW at time T1 and falls again below the alarm threshold ALSW before time T2. In conventional systems, this event 1 would already result in an alarm if the alarm was not checked again before the alarm was given. According to FIG. 2b, if the sum signal SUS is considered, the method according to the invention shows no increase in the sum signal SUS beyond the limit value GRW. So there is no alarm. At time T3, the detector measured value MW falls below the interference threshold STSW (FIG. 2a), which means that when the sum signal SUS (FIG. 2b) is formed, it is added as a negative signal. In order to prevent integration of the sum signal SUS into the negative range, a comparison with 0 is carried out when the comparison value VWB is formed, as will be explained later with reference to an exemplary embodiment. This is illustrated in this diagram at time T4. The sum signal SUS is not integrated again until time T5. At time T6, the detector measured value MW (Fig. 2a) reaches the alarm threshold ALSW (event 2). However, the sum signal SUS has not yet been integrated up to the predetermined limit value GRW. It is only at time T7 that the sum signal SUS reaches the limit value GRW and triggers an alarm AL up to time T8. So an alarm is only given when the detector measured value is not only large enough, but also pending for a certain time.

Another typical measured value signal image (event 3) shows a slow increase in the measured detector value MW in the direction of the alarm (FIG. 2a). A conventional fire alarm system would not yet recognize an alarm, since the measured value MW had not yet reached the alarm threshold ALSW at time T11. According to the method according to the invention, however, the detector measured value MW, based on the idle value RW after it has exceeded the interference threshold STSW (FIG. 2a), is integrated (FIG. 2b) from time T10 and the sum signal SUS already reaches the limit value at time T11 GRW and triggers the alarm AL. In this case, a steady increase in the detector measured value in the direction of the alarm threshold is detected at an early stage.

3 shows an exemplary embodiment for alarm detection in the block diagram. It can be seen from the example of a detector M that the detector measured values MW reach the control center Z from the detector M via the detection line L. On the one hand, the measured value MW reaches a comparison value generating device VWB and a rest value forming device RWB. A memory VWSP is assigned to the device for forming the comparison value VWB, in which the current comparison value VWN is stored. A memory RWSP is assigned to the device for rest value formation RWB, in which the current rest value RWN is stored. For each interrogation cycle, a new comparison value VWN (VWB) is formed for each detector from its current measured value MW, its last stored idle value RWA and its last stored comparison value VWA. This current comparison value VWN is stored on the one hand for the next processing cycle in the comparison value memory VWSP, and on the other hand compared with a predetermined limit value GRW, in the exemplary embodiment for alarm, in the comparison device VGE, which is arranged downstream of the two devices. If the current comparison value VWN is greater than or equal to the limit value for alarm GRW, an alarm AL is displayed in the display device ANZ connected downstream of the comparison device VGE. If the current comparison value VWN does not exceed the limit value GRW, the new detector measured value MW can be used together with the old idle value RWA from the idle value memory RWSP to calculate a new idle value RWN, which is used to rewrite the idle value memory RWSP. The block diagram (Fig. 3) illustrates the detection of alarm. In a similar way, faults can be recognized and displayed.

4 shows the device for forming the comparison value VWB in more detail. The detector value MW passes from the detector to the control center Z and to a first arithmetic logic unit ALU1. There, the old idle value RWA is subtracted from the idle value memory from the detector measured value MW. A predeterminable constant value CON is subtracted in a second arithmetic logic unit ALU2, which is connected downstream of the first ALU1. The second arithmetic logic unit ALU2 is followed by a third arithmetic logic unit ALU3, which adds the result of the ALU2 to the last (stored) comparison value VWA. The comparator K1 connected downstream of the ALU3 with an assigned demultiplexer D1 only compares the result from the ALU3 (sum signal SUS) with the value 0 in order to achieve a downward limitation of the sum signal (SUS according to FIG. 2b). If the value is less than 0, the multiplexer D1 outputs 0 at its output. Is the value however, greater than 0, the sum signal SUS is at the output of the multiplexer D1 as the current comparison value VWN. This output leads to the comparison device VGE, in which the new comparison value VWN is compared with the limit value GRW using a further comparator K2.

The second demultiplexer D2 connected downstream of the second comparator K2 controls the display device ANZ when the comparison value VWN is greater than or equal to the limit value GRW (VWN> GRW). If the comparison value VWN is smaller than the limit value GRW (VWN <GRW), the second demultiplexer D2 controls the rest value formation device RWB and enables the formation of a new rest value RWN, as will be explained in more detail with reference to FIG. 5.

5 shows the circuit arrangement for forming the idle value RWB. It has a first multiplier MU1, which is followed by an adder AD1 with a first input. It has a subtractor SU1 which is supplied with a constant value EPS (0 <EPS <1). This constant EPS can be used to influence the difference (MW-RWA) from the respective measured measured value (MW) and the stored detector idle value (RWA) for the formation of the idle value. This constant value EPS is given to the first input, the detector measured value MW to the second input of the first multiplier stage MU1. The output signal (1-EPS) of the subtractor SU1 reaches the second multiplier stage MU2, to which the last stored idle value RWA comes from the idle value memory RWSP. The output of the second multiplier stage MU2 leads to the second input of the adder stage AD1, which, controlled by the comparison device VGE, forms the current idle value RWN via the enable input E if VWN <GRW. The current reporting value MW is multiplied by the constant value EPS in the first multiplier M1. The old idle value RWA from the idle value memory RWSP is multiplied by the value (1 - EPS) in the second multiplier MU2. The adder AD1 then delivers the new idle value RWN (EPS - (MW-RWA) + RWA = RWN) at the output.

With the inventive method, a slow change in the detector, z. B. compensated for by component aging or contamination. The sensitivity of the detectors remains constant for a very long time. Different applications can usually be served with uniform detectors and evaluation programs. In addition, slowly developing fires as well as rapidly spreading fires are detected at the earliest possible point in time, whereby malfunctions and deceptions of the alarm system are largely prevented.

Claims (6)

1. A method of increasing the response sensitivity and the resistance to interference in an alarm installation, in particular a fire alarm installation, with a central control unit (Z) to which a plurality of automatic monitors (M) are connected and cyclically interrogated, and in which the measured monitor values (MW) are analysed to form a mean value from the individual monitor signals and compared with the monitor signals in question, and in the event that the mean value is overshot or undershot by an adjustable amount, a single device is actuated, characterised in that from the current measured monitor values (MW) for each monitor (M), a mean measured monitor value is formed as a monitor rest value (RWB), and is stored for the monitor in question in a rest value store (RWSP) as current rest value (RWN), that the difference (MW-RWA) between the current measured monitor value (MW) and the stored rest value (RWA) is formed, that in order to obtain a current and storable (VWSP) comparison value (VWN), these differences are integrated to form a sum signal (SUS), and that the current comparison value (VWN) is compared (VGE) with a predetermined limit value (GRW) and in the event that it is overshot (VWN a GRW), the signal display device is actuated for an alarm and interference display device (ANZ), and in the event that it is undershot (VWN < GRW), the current rest value (RWN) is formed.

2. A method as claimed in Claim 1, characterised in that the difference (MW-RWA), which ist formed (ALU1) with each interrogation cycle is reduced (ALU2) by a predeterminable constant (CON) and integrated (ALU2) to form the sum signal (SUS), that with each interrogation cycle the sum signal (SUS) is compared with the value zero (K1, D1) and, when it is greater than or equal to zero (0), is input as a current comparison value (VWN) into a comparison value store (VWSP), where the sum signal (SUS) is set at zero during the first interrogation cycle.

3. A method as claimed in Claim 1, characterised in that for the formation (RWB) of the current rest value (RWN), the current measured monitor value (MW) and the last rest value (RWA) are added (AD1) to be influenced (MU1, SU1, MU2) by a predeterminable constant value (EPS) and input as current rest value (RWN) into the rest value store (RWSP) in dependence upon the current comparison value (VWN < GRW) where, during the first interrogation cycle, the rest value corresponds to the first measured monitor value.

4. An arrangement for the execution of the method claimed in Claim 1, characterised in that the central control unit (Z) is supplied with measured monitor values (MW), a device (RWB) being arranged for the formation of a monitor rest value with an assigned rest value store (RWSP), and a device (VWB) for the formation of a comparison value with an assigned comparison value store (VWSP), and these two devices are followed by a comparison device (VGE), which acts in dependence upon the current comparison value (VWN) in relation to the predetermined limit value (GRW) to operate a following display device (ANZ) or facilitate a new rest value formation (RWB).

5. An arrangement as claimed in Claim 4, characterised in that the device for comparison value formation (VWB) possesses a first substractor (ALU1) supplied with the current measured value (MW) and the stored rest value (RWA), whose output is connected to a second subtractor (ALU2) fed with a constant value (CON), whose output is connected to an adder (ALU3) supplied with the stored comparison value (VWA) and leading to a comparison device which consists of a comparator (K1) having a second input, supplied with the value zero (0), and a following demultiplexer (D1).

6. An arrangement as claimed in Claim 4, characterised in that the rest value formation device (RWB) possesses a first multiplier (MU1) supplied with the current measured monitor value (MW) and a constant value (ESP), and a subtractor (SU1) supplied with a constant value (ESP) and a value »1« and whose output is connected to a second multiplier (MU2) whose second input is supplied with the stored rest value (RWA), and that the output of the first multiplier (MU1) and the output of the second multiplier (MU2) lead to an adder (AD1) by which, in the event that the current comparison value (VWN) undershoots the limit value (GRW), it is possible by means of the enable input (E) to form a current rest value (RWN) which can be input into the rest value store (RWSP).

Images