FI103368B - Smoke alarms - Google Patents

Description

103368

Fire alarm

TECHNICAL FIELD The present invention relates to a fire alarm system, and in particular to a fire alarm system, wherein fire detection is performed by calculating fire source parameters or information such as heat generation rate, smoke generation rate or gas formation rate using fire simulation mathematics.

10 Prior Art

In a conventional fire alarm system, fire detection for the purpose of giving an alarm is basically performed by comparing fire-related secondary sensor-detected parameters, such as temperature, smoke concentration, or CO gas concentration, or their temporal change rates, with pre-set values. The simplest system detects the occurrence of a fire and gives an alarm when the values detected by the sensors exceed the specified thresholds of en-20. A system is also known in which fire occurrence is assumed when the time change rate obtained by derivatizing the detector value: exceeds a preset value or when future fire prediction changes are predicted by first or second order .25 orders of magnitude changes from past detector values.

* ···, The standard fire alarm system is thus basic • · II! seemed to have the principle that the heat, temperature, smoke, or gas generated by the secondary secondary fire is detected by the sensors and that the occurrence of the fire is assumed directly on the basis of these secondary parameters.

• · · ·. * * However, the fire development process is variable //; va, and combustion products or environmental conditions contribute to the fire propagation process in a complex manner.

Thus, the condition of the secondary heat (temperature), smoke 2 103368, or gas produced will also vary differently with changes in environmental conditions such that many difficulties are encountered when the occurrence of fire must be accurately and promptly detected by these variable sensor data or para-5 meters.

DISCLOSURE OF THE INVENTION It is a primary object of this invention to provide a novel fire alarm system in which the occurrence of a fire can be detected by calculating primary fire source parameters such as heat generation rate, smoke generation rate or gas generation rate on fire source based based on gas concentration using a fire simulation mathematical model.

Another object of the present invention is to provide the aforementioned fire alarm system, wherein the reliability of detection is further improved by correlating a plurality of primary fire source parameters with basic support.

To achieve the above objects, the present invention in the broadest sense is a fire alarm system comprising: a sensor member disposed within a fire detection area that detects a physical phenomenon associated with a fire, such as temperature, smoke concentration, or CO gas concentration; a fire source parameter calculator, which computes primary fire source parameters, such as heat-30 generation rate, smoke generation rate or gas formation rate at the fire source, in combination with detection information from the sensor bodies and a preset arithmetic operation formula; and a fire occurrence inferring means for detecting fire source parameters based on the rate of change of fire source parameters calculated by the fire source parameter calculator 3 103368.

In the fire alarm system of the present invention, the arithmetic-logical program of a mathematical model of a fire condition in a room simulating a fire condition in a room applies the above formula in a fire source parameter calculator, and primary fire source parameters, such as secondary parameters such as temperature, smoke concentration, or CO gas concentration, by inverse calculation from the above formula, and accurate fire detection can then be made based on the rate of change of these primary fire source parameters to provide an alarm.

These primary parameters of the fire source, i.e.. The rate of heat generation, the rate of smoke generation or the rate of gas formation are, by their nature, uniquely determined without the influence of combustion products or environmental conditions, and contribute to improving the accuracy of fire parameter calculation, which in turn significantly improves the reliability of fire detection.

According to the present invention, the source of the fire itself. ... j rate of heat generation, rate of smoke generation, or gas. The rate of formation can be calculated from secondary fire-related secondary phenomena, such as the detection of sensors by heat. ' patient, smoke concentration, or CO gas concentration • · • · **; of fire, by inversely calculating a mathematical model to simulate a fire »» · * · * *, and fire can be detected by the rate of change of the primary fire source • · parameters. Thus it is: the risk that a false alarm source can be inferred from a fire,. : Can be minimized, which significantly improves the reliability of fire detection.

The foregoing and other features and advantages of the present invention will become more apparent from the following description and the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 shows a block diagram of an embodiment of a fire alarm system according to an embodiment of the present invention.

Figure 2 shows a two-layered zone model 10 used in a fire detection algorithm in the present embodiment.

Figure 3 is a flowchart of the fire detection algorithm of the present embodiment.

Figures 4a, 4b and 4c show the sensor response during burning of a wooden chair as well as the change in heat generation rate AQ, the change in smoke generation rate ACs and the rate of gas formation rate AG as a function of time according to the present invention.

Figures 5a, 5b and 5c show the sensor response during cooking as well as the change in heat generation rate AQ, the change in smoke generation rate ACs and the change in gas generation rate AG as a function of time according to the present invention.

Figures 6a and 6b show a sensor response of wood; during the burning of the chair in several rooms of different sizes. 25 as well as changes in the heat-cure rate AQ as a function of time according to the present invention.

'IV.' Figures 7a and 7b show the correlation coefficient R derived by weighting the derived correlation coefficient R from the heat generation rate and the rate of smoke generation determined by the burning of a wooden chair as shown in Figures 4a, 4b and 4c. · *. ·. * Temporal changes of the quality value RD and its derivative • · · V · dRD / dt.

Figures 8a and 8b show the rate of heat generation and the rate of smoke generation obtained during cooking. 5a, 5b, and 5c, the temporal changes in the correlation value R 'obtained from the weighting of the correlated coefficient R and the derivative dRD / dt of its derivative.

BEST MODE FOR CARRYING OUT THE INVENTION In Figure 1, a fire alarm system according to an embodiment of the present invention comprises a plurality of sensors disposed on the ceiling or other such room, namely temperature sensor 10, smoke concentration sensor 12 and CO gas concentration sensor 10 14. , the smoke concentration Cs and the CO gas concentration G in an analogous manner and give detection signals proportional to the values expressed. The fire alarm system also comprises a sampling circuit 16 which receives these detection signals from sensors 10, 15 12 and 14. The detection signals are sampled at fixed intervals in the sampling circuit 16, and converted by the AD converter in the circuit to output digital signals.

In the present embodiment, each of the sensors 10, 12 and 14 is disposed within each fire detector area.

20 However, if necessary, two or more corresponding sensors of the same type may be located in each fire detection area. As for the signal transfer system between the sensors and the sampling circuit: 16, although ·; ·· in this embodiment is directly wired. 25 system, which directly transfer the detected analogiasig- • · · signal via a signal line, so any of the cell [11 ^ pivaa signal transmission system, such as kiertokyselyjär- • · Y \ system, comprising performed a sampling • · * · * "sensor-side poll and response signal transmission of the detection signal, may be applied directly instead of the wired system.

The present system also includes a sampling system. : in circuit 16 aliases, to provide different initial values of ____: Fire Source Parameters ____: Unit 18 and Setpoint Setting Unit 20 for Fire Source Parameters 6 103368 units 18 for performing a fire simulation mathematical model. The arithmetic-logical program of the mathematical model simulating the fire is pre-installed in the fire source parameter computation unit 18. The reverse 5 mathematical model calculation calculates the heat generation rate, the smoke generation rate and the gas formation rate from the outputs. temperature Θ, smoke concentration Cs and CO gas concentration G

Various initial values for performing a fire simulation mathematical model from unit set-up unit 20 are provided to calculator 18, which responds to the set-point according to the conditions in the fire notification area where the above-mentioned sensors are installed and calculates primary fire source parameters from the sensor detection data.

The fire alarm system also comprises a fire occurrence inference unit 22, which receives the source parameters calculated by the calculating unit 18, i.e.. a heat reduction rate of 20, a rate of smoke generation and a rate of gas formation, and an alarm display unit 24 which responds to the alarm signals from the fire occurrence terminator 22 and provides an alarm in optional form ;;; such as acoustic and / or visual forms.

. 25 Fire Outbreaking Unit 22 Detected:. see fire occurrence based on the rate of change of the fire source parameters above the preset baseline or performing arithmetic operations to predict the fire when the baseline has been exceeded, according to a first 30 or second order function using the fire source parameters obtained up to that time. When the fire occurrence inference parameters based on the fire source's para- ·· »·. · · Meters are obtained from the fire occurrence inference unit 22, the fire detecting output, i. the alarm signal, 7 103368 is provided to the alarm display unit 24, from which the alarm is issued.

The principle of the arithmetic operation of the fire source parameters 5 performed by the fire source parameter calculating unit 18 of Fig. 1 will be explained in detail below.

Several mathematical models based on physical sciences have been proposed for use in analyzing the properties of a fire in a room. These 10 can be classified into field equation models and zone models.

With these mathematical models, the flow state of the smoke concentration or temperature in the room is obtained as a solution of the diffe-15 equation based on the heat or smoke volume generated by the fire source. In the field equation model, a closed space with all entrances, exits, doors, and windows closed is used as a reference space, and the inside space is divided into hundreds of small subspaces, each cubic in length, each 20 pages long. The mass persistence equation, the impulse retention equation, the space equation, and the boundary condition are applied to each subspace to determine the temperature or smoke concentration in the room. The field equation is characterized by the fact that, since the concentration in the smaller subspace of the curve is calculated in detail, phenomena such as temperature or smoke concentration during a fire can be accurately determined.

As the computation in the field equation model is performed in each of these hundreds of subspaces, the computation time 30 increases and causes a problem in real-time processing, as well as the difficulty of parsing the arithmetic operation. ·, · 'Meter values can only be hardly changed.

. ·. : In the zone model, closed room space is taken as the basic * '; reference mode and is divided into two or more times. 35 rock. The zone model is characterized in that it determines the average temperature or average smoke concentrations in the upper floor of the room. Because it is a simple model, computing time can be reduced and real-time processing is possible on a personal computer.

The zone model has the additional advantage that parameters such as room dimensions (ceiling area and height), ambient temperature, heat dissipation rate or rate of heat generation per unit of time can be set or changed freely, the height distance to the boundary layer, i.e. upper and lower the interface between the lower floor can be determined, and that the status of the hazardous floor of the room can be roughly determined. However, the zone model cannot be said to have a better accuracy than the field model because the number of terms using the arithmetic operation of the finite differences and the number of terms using the arithmetic operation is reduced to improve the computation time.

Thus, the fire source parameter computation unit 18 of the present embodiment can use a field equation model whenever a detailed and accurate arithmetic operation is necessary, while the zone model can be used whenever real-time processing is necessary.

The following explanation is made for the case of using a zone model that can perform a real-time operation to detect the onset of a fire, although the accuracy • · is slightly reduced because the available information to detect the occurrence of a fire is · 30 sensor output.

• · ·. · .1 Various methods have been used to get the zone model utilized. However, there is no such example. : ki utilization that is theory independent.

• «· • • • •» I · • · «· · · • 9 103368

In the present embodiment, the two-layered model ASET (Available Safe Egress Time) -B (2), which is one of the mathematical model programs analyzed by L. Y. Cooper (1), is being applied and is currently being developed by W.D. Dalton.

References: (1) Cooper, L.Y., A Mathematical Model for Estimating Available Safe Egg Time in Fires, Fire and Materials, Vol. 6, Nos. 3 and 4, p. 135-144; 1982, September / December.

(2) Walton, W.D., ASET-B, A Room Fire Program for Personal Computers, National Bureau of Standards (U.S.), NBSIR 85-3144; 1985, April, p. 1-35.

15

Thus, in the present embodiment, the inverse arithmetic operation of the mathematical model is performed on the basis of detection information obtained from multi-quality sensors to determine changes in the amount of heat, smoke or gas in the fire source and to detect fire occurrence based on the arithmetic operation.

• 1 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 10 103368

Figure 2 is a schematic representation of a zone model computed by the Fire Source Parameter Calculation Unit 18 as a fire simulation calculation model.

The zone model shown in Figure 2 is a two-layer zone model. Since it is a simple model adapted to determine the average temperature 9h or the average smoke concentration CSh of the upper layer 28, the conditions are set as follows.

Assume that all openings, except for minor leaks from the floor surface, i.e. entrances and exits, doors or windows, are closed so that atmospheric pressure in the room is constant and that any increase in atmospheric pressure in the room can be neglected due to leakage from the floor surface.

15 It is also assumed that a fire occurs in a fire on the floor. The heat or smoke released from this fire foci is raised by the buoyancy and reaches the roof surface. Thereby, the evolving smoke cloud 26 rises sideways into the ambient cold air 20 and the hot air stream reaching the roof diffuses and reaches the side wall surface to form a hot layer, i. the upper layer 28. Between the upper layer and the lower air layer 30, the interface 32 thus developed gradually descends towards the floor surface as the fire *: * 25 progresses over time.

• * · · ^: ***: It is further assumed that in such a double-layer • · ·; 'j *; in this zone model, the temperature and smoke concentration are constant both in the hot upper layer 28 and,. in the lower layer 30, which is at ambient temperature, • · * I. *. ^ 30 and that the heat exchange between the layers takes place in a smoke cloud *. 26.

The simulation determines, from the previously obtained rate of heat generation per unit time 11,103,368, the temperature 0h of the upper layer 28 and the distance Z from the firebox to interface 32.

Thus, the distance Z from the focal point to the interface 32 of the upper layer 28, the average temperature 0h of the upper layer 28 and the smoke concentration Csh can be determined by solving the differential equations mentioned below. The initial conditions set by the setpoint setting unit 20 are shown together with the differential equations. As for the C0 gas concentration Gh, the same equations as for the smoke concentration Csh apply.

Equation (1): r -Cl * AQ - C2 * ZiQ1 / 3 · Z1 ^ 3 (when 0 <Z <Zo) 15 dz / dt = <-Cl * Δ0 (when -F <Z <0) 0 (when Z = -F) 20 Equation (2): f 0h [C1 * AQ - (0h / 0o-1) * C2 * AQ1 / 3-z5 / 3] / (Zo-Z) J (when 0 <Z <Zo ): d0 / dt = <·: · 25. ···. 0h * Cl - AQ / (Zo + Z) (kun -F £ Z <0) • · \ • · · · · · · · · ·

Equation (3): • · 30 / 0h / 0o (ACs-Csh-C2 * AQ1 / 3-Z5 / 3) / (Zo-Z) »*! * · • · · (when 0 <Z s Zo) dCsh / dt = \ 0h / 0o * ACsh / (Zo-Z) (when -F «SZ <0) 12 103368

Initial conditions (t = 0): Z = Zo = H-F; 5 0h = 0o (1 + Zo_5 / 3-AQo2 / 3 * C1 / C2);

Csh = ACs / ACso * Zo-5/3 · AQo-1/3 / C2 (where ACs / ACso-1) 10 d0h / dt = 0o (C1 * AQo2 / 3 / C2) [2 AQf / AQo + 5 ( Cl · Δ QO + C2 * ΔΟο1 / 3 * Zo5 / 3)] / (6 · Zo8 / 3); dCsh / dt = (Cl * AQo2 / 3 / C2) [5+ {Zo / (Cl ^ Qo)} (3-ACsfMCso χ 5 AQf / Δ0ο + 5 ♦ C2 -Δ0ο1 / 3 * Zo2 / 3)] / ( 6 * Zo8 / 3); where:

Cl = (1-LC) / (p * CP * eo * S); C2 = (0.21 / S) [(1-LR) g / (p-CP * 0 0)] 1/3; 20 AQf = dAQo / dt;

Csf = dAcso / dt, (t = 0); Δ0: rate of heat generation per unit time; ACs: smoke generation rate per unit time; ΔΟο: initial heat generation rate; ··· 25 AC50: initial smoke generation rate; · · M · 1: floor area of the enclosed space; • «· H: ceiling height; • · 4 F: height of firebox; CP: specific heat of air; 30 LR: heat loss rate of radiation; • · · * · * LC: heat loss rate of convection; 0o: ambient temperature; g: gravitational acceleration; p: density of air.

103368 In such a two-tiered zone model, changes in temperature or smoke concentration are defined as changes in heat generation rate per unit time or changes in smoke generation rate per unit time 5 and, as a distance at interface 32, changes in rate of heat generation per unit time. In addition, if the advanced Euler equation is used to solve differential equations, its arithmetic operations can be performed at a higher rate.

In the present embodiment, the temperature Θ indicated by the temperature sensor 10 and the smoke concentration indicated by the smoke concentration sensor 12 are treated similarly, as in the two-layer model, the average temperature 0h and the average smoke concentration Csh of the upper layer 28 when calculating the heat generation rate and time.

Figure 3 is a flowchart of a fire detection algorithm based on the fire source parameter estimation used in the embodiment of Figure 1.

In Figure 3, before the system is activated, the setpoint setting unit 20 sets the setpoint values to the fire source parameter calculator 18 in step SI. During this initial set-up, all approximate the initial ··· 25 values, except for Cl, C2, AQL, ACsL, the rate of heat build per unit time AQ and the rate of smoke formation per unit time ACs shown in the differential equations ( 1), (2) and (3) as initial conditions in a double-layer model, are fed or set by an internal arithmetic • ♦ · M 30 operation.

• i * * · * * After the initial values are set in step 1, the system is in the active state and the program proceeds to; "· step S2, where the values obtained from temperature sensor 10, expressed as temperature Θ, from the smoke concentration sensor 14 103368 12 the resulting smoke concentration Cs and the CO gas concentration concentration G from the sensor 14 are sampled at fixed intervals in the sampling circuit 16.

Then, in steps S3 to S6, the change in the rate of heat generation-5 per unit time AQ is determined. Thus, in step S3, the initial value of AQ is set. In step S4, the average temperature of the upper layer 0h and the distance to the smoke layer Zh at this time are calculated by the above-mentioned ASET-B arithmetic operation. Then, in step S5, the absolute value of the difference between the average temperature 0h calculated by ASET-B and the temperature Θ detected by the temperature sensor 10 is compared to a predetermined value e, and steps S3 through S5 are repeated until this difference becomes equal to or less than e. 0.001 to increase the value initially set at 15 AQ. In step S6, the value of AQ when the condition of step S5 is satisfied is set to the value of the rate of heat generation at this time.

The program then proceeds to step S7, where the smoke generation rate ACs and the gas generation rate AG are set.

At step S8, the temperature already set at this time is used to perform the arithmetic operation of ASET-B on the smoke concentration Csh and the gas concentration Gh. the determination. In step S9, it is checked whether the absolute value of the difference between Cs and • · · 25 Csh and the it- · * * ··· 'difference between G and Gh is less than a predetermined value ε, e.g.

• · · * 0.001. If this condition is not met, steps S7 and S9 are repeated to incrementally increase AG and ACs. The values of ACs:: and A G at the time when the condition 30 of step S9 is fulfilled are set to the smoke generation rate and the gas -. *. the rate of formation at the instant in question (step '· sio).

The program then proceeds to step S11, where it is analyzed whether the change in heat generation rate AQ set in step S6 and the change in smoke generation rate ACs and the rate of gas formation rate Ag set in step S10 exceed the previously set airflow level of the fire (alarm). If the analysis of the results of the arithmetic operation in step S11 indicates that these results exceed the alarm output level, the program proceeds to step S12 where the predictive arithmetic operation is performed using the heat generation rate change AQ, smoke generation rate change ACs and the gas formation rate. For example, such a predictive arithmetic operation may use Newton's regressive interpolation formula. In addition to the aforementioned predictive operation, the arithmetic operation of fire detection in step S12 may also determine changes in the first order difference and / or the second order difference up to a time corresponding to a predetermined number of sampling times prior to the current time of the arithmetic operation. the result exceeds the baseline, 20, or fire detection by arithmetic operation In step S12, the correlation and / or weighted correlation value between corresponding results of arithmetic operations can also be determined.

In step S13, the occurrence of fire is deduced from the result obtained in step S12.

Figures 4a, 4b and 4c are graphs showing the time evolution of the heat generation rate AQ, the time evolution of the smoke generation rate ACs and the time evolution of the gas formation rate AG obtained by this direct function. 30, the fire source parameter calculation unit 18 *, together with the sensor detected temperature Θ, the distance L to the upper layer interface, the detected smoke concentration Cs and the detected gas concentration G, as determined by the fire embodiment 16 of the present embodiment, where the chair (materials: fabric, urethane foam and wood) burns in the middle of the room floor with a room floor area of 6.7 x 4.3 l by 28.81 m 2 and a ceiling 5 height of 2.5 m.

Figures 5a, 5b, and 5c are similar patterns in the fault alarm test, in which cooking in the kitchen is exemplified by roasting nine servings of fish in the same room as in Figure 4.

Comparing the results from the source of the fire of Figure 4 and the fault alarm of Figure 5, it is seen that the temporal change in heat generation rate Aq in the case of fire in Figure 4a shows a sharp peak at that time when temperature Θ suddenly rises as fire progresses. Conversely, no such peak is detected in the change in heat generation rate Aq for the source of the fault alarm shown in Figure 5a. Thus, the occurrence of a fire can be detected from the correx 20 as both the temperature Θ and the rate of heat generation AQ increase linearly. In the event of a fire, the change in smoke generation rate ACs shown in Figure 4b and the change in gas formation rate AG shown in Figure 4c increase correlatively with the peak values relative to the heat generation rate change · · · · ... · AQ so that a more accurate fire detection By correlating these three parameters, namely the heat generation rate AQ, the smoke generation rate ACs, and the gas generation rate AG, between at least two parameters 30.

Conversely, in the case of the source of the fault alarm shown in Fig. 5, the change in smoke generation rate between ACs and the change in gas formation rate AG on the one hand and the change in heat generation rate AQ 'on the other hand.

There is no correlation from which the source of the fault alarm can be accurately distinguished. It is also possible to detect the source of the fire and fault alarm by the fact that the smoke generation rate ACs and the gas generation rate 5A G in the case of a fault alarm are similar to the change mal in the case of fire shown in Fig. 4, but less extensive in the case of a fire alarm.

Figures 6a and 6b show the results of similar experiments in which the temporal variations AQ of the heat generation rate are plotted as a function of temperature huon for different room sizes. The results show satisfactory agreement with the calculated changes in the rate of heat generation AQ, despite the change in room size. It is seen that the same change in heat generation rate AQ can be obtained by the present embodiment in the case of the same fire regardless of the size of the room. This also applies to the smoke generation rate ACs and the gas generation rate AG.

1, a specific embodiment of the fire end termination unit 22 shown in Figure 1 will be described below.

In the present embodiment, the occurrence of fire is expressed by a correlative arithmetic operation in the fire derivation unit 22 using two of the parameters obtained from the fire. from the source parameter calculation unit 18, namely the heat generation rate AQ, the smoke generation rate ACs and the gas generation rate AG.

• · 30 The correlation factor R is first defined by the following 1 formula (4): 18 103368 R = Sxy / VSx-Sy (4) where Sxy, Sx and Sy Expressed by the following formulas: 5 m2

Sx = Σ (X ± - X) 2 i = ml 10 m2

Sy = Σ (Yi - Ϋ) 2 (5) i = ml 15 m2

Sxy = Σ Xi * Yi - η · Χ * Ϋ i = ml 20 where X, Y are any combination of two of Aq, ACs and AG; X, Ϋ mean time-averaged values; n represents the number of data used (= m2 - ml + 1).

The correlation coefficient R calculated from equation (4) is multiplied by 25 for weighting by the absolute value of the combination vector D of the two values used in the correlation calculation | D | to determine the weighted correlation value RD. The absolute value | DI of the composite vector used in the weighting can be defined by the following equation: 30 | D | = | Ui + Vj | (6) • · · ... i where U and V denote any two fire sources • · · · ... * of the parameters AQ, ACs, and AG, respectively, modified by optimized scaling • «· ·. ♦, and i and j are unit vectors of corresponding dimensions.

35 As the correlation factor R and the composite vector • · change over time, the following formula expresses the weighted correlation value RD as a function of time: RD (t) = R (t) + | D (t) | (7)

Thus, the value of the coefficient R of the correlation statement-40 calculated at a given time is weighted by the two detected values U and V

19 103368 of the eigenvalue of the recombinant vector | d | and correlation value RD, which is a correlation factor R more heavily weighted at higher values of U and V, is determined.

Figure 7a shows the weighted correlation value RD

time changes as determined from Equations (4) and (7) using the heat generation rate AQ and the smoke generation rate Acs shown in Figures 4a and 4b during a fire. In Figure 7a, a sharp change in peak 10 occurs in the correlation value RD. A fire can thus be detected when the correlation value Rd exceeds a preset threshold RL.

Figure 7b shows the derivative data of the correlation value Rd shown in Figure 7a. Significant changes sufficient to detect a fire also occur in this derivative information.

Fig. 8a shows a correlation value RD determined according to formulas (4) to (7) with respect to the heat generation rate AQ and the smoke generation rate ACs in the case shown in Figures 5a and 5b when no fire occurs. In this case, the correlation value RD remains below the threshold R1 so that the source of the fault alarm can be detected. Figure 8b shows the temporal changes in the derivative values of the correlation value RD shown in Figure 8a.

• · · ...: 25 In the above-described embodiment, the rate of heat generation • * ·:., R per unit time AQ, the rate of smoke generation • «« 'per unit time ACs and the CO gas generation rate per unit AG are calculated as the primary fire source: Y: parameters. However, since fire flames produce • ·. · *. ·. 30 ions, the change in fire output per unit time can be calculated from the information detected by the ion sensor in the same way as the primary fire source parameters, so it can be used as an additional 35 fire detection information.

Claims (5)

1. A fire alarm system, characterized in that it comprises: a sensor means located within the fire monitoring area and detecting a physical phenomenon in connection with fire, such as temperature, smoke concentration or CO gas concentration; calculating means for the fire source parameters, which calculates the primary parameters of a fire source's heat generation rate, smoke evolution rate or gas evolution rate on the basis of detection data from said sensor means and a preset arithmetic calculation formula; and fire assessment means, which distinguishes a fire for initiating an alarm on the basis of the rate of change in the fire source parameters calculated by said fire source calculation means. Fire alarm system according to claim 1, characterized in that the fire source calculation means comprises an arithmetic counting means for calculating the fire source primary parameters by a reverse arithmetic operation from secondary temperature, smoke concentration or CO gas concentration data. centering detected by the sensor means, wherein the inverse arithmetic operation is based on said formula according to a fire simulation mathematical model for analyzing the location of a fire in the monitored room. • · · 3. Fire alarm system according to claim 1,. Characterized in that the assessment body for • f. the fire includes correlation calculation means for separating a fire by correlation calculation using at least two kinds of quantities of primary parameters obtained from the fire source calculating means. 103368 23 Fire alarm system according to claim 3, characterized in that the correlation calculator is arranged to discern a fire by calculating results weighted on the basis of the absolute value 5 for a combination vector D, which is determined by means of the at least two quantities used in the correlation calculation. A fire alarm system according to claim 1, characterized in that the sensor means includes sensors for temperature, smoke concentration and CO gas concentration. · «1 II ·! • · «• · tl · •« · • · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·