JP3069465B2 - Fire alarm system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fire alarm system using an analog fire detector.

[0002]

2. Description of the Related Art Conventionally, an analog sensing output from an analog fire sensor such as an analog photoelectric smoke sensor or an analog ion sensor is returned to a receiver or a repeater using time division multiplex transmission or the like. In the receiver or repeater, the returned analog sensing output level is a preset alarm level (forecast level, fire report level, interlock report level)
There is a fire alarm system which counts a predetermined time (hereinafter referred to as accumulation time) from the time exceeding the alarm time, and issues an alarm if the analog detection output level exceeds the alarm level at the end of the counting of the accumulation time.

[0003]

In the above-described fire alarm system, the alarm level can be set for each fire detector, or the alarm level can be switched between daytime and nighttime. It cannot be changed depending on the situation. Therefore, there has been a problem that the so-called non-fire report is frequently generated because the detected output level exceeds the set alarm level due to a non-fire factor, and the time during which the detected output level exceeds the set alarm level often exceeds the accumulation time.

The present invention has been made in view of the above-mentioned problems, and an object thereof is to reduce a non-fire report by changing an alarm level in accordance with a situation, and to provide an early warning of an actual fire. It is to provide a fire alarm system that can be found.

[0005]

According to the first aspect of the present invention, when the analog detection output level of the analog fire sensor reaches a preset alarm level, the accumulation time is counted. In a fire alarm system that determines whether or not to issue an alarm based on the analog detection output level at the end of counting, a time series until the analog detection output level of the analog fire detector reaches a preset standard alarm level Means for obtaining a plurality of feature amounts from the data; means for obtaining a certainty factor of a fire and a certainty factor of a non-fire based on a determination rule set composed of one to a plurality of feature values of the feature amount group; Alarm level determining means for changing the alarm level according to the given confidence level, wherein the changed alarm level is lower than the standard alarm level Counts the time obtained by subtracting the time from the time when the analog sensing output level reaches the changed alarm level to the current time from the accumulation time as the remaining accumulation time, and the changed alarm level becomes higher than the standard alarm level. If it is high, the counting of the accumulation time is started when the analog sensing output level reaches the changed and set alarm level.

A second aspect of the present invention is an embodiment of the first aspect of the present invention, in which the alarm level is changed and set stepwise.

[0007]

According to the present invention, since the alarm level is changed based on the judgment of non-fire / fire, the alarm level can be lowered at the time of fire judgment, and the start time of the accumulation time can be expedited. Early detection is possible. Conversely, at the time of non-fire determination, the alarm level can be increased to delay the start of counting of the accumulation time, which can prevent the alarm from being caused by a transient increase in the sensing output level and reduce the number of non-fire alarms. I can do it.

[0008]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 shows the overall configuration of an embodiment of the fire alarm system of the present invention. The receiver (or repeater) 2 of this system receives the time series data of the analog detection output of the analog fire detector 1 through the processing unit 3 by polling performed through a transmission system using time division multiplex transmission or the like. The processing unit 3 receives the analog detection output level and sets a warning level for a standard forecast, a fire report, an interlock report, etc., which is set in advance corresponding to the sensor (the fire report level will be described below). ) Are obtained from the time-series data up to).

Here, the characteristic amount obtained by the processing unit 3 is, for example, in the case of the analog fire detector 1 comprising a smoke detector,
Output value, slope (or time to execute a specific section), average value (or integration value), difference of average value, number of data, difference value, sum of absolute value of difference, etc. seeking analog sensor output level at a particular point in time, for example, as that standard fire report level L ₀ to beyond analog sensor output level after a predetermined time from, in the case of gradient, for example, from beyond the forecast level standard The slope of the analog sensing output level in a predetermined section is determined, such as the slope of the analog sensing output level until the fire alarm level L ₀ is reached,
Further, in the case of the average value, for example, the standard fire alarm level L ₀
Calculate the average value of a predetermined section, such as the average value of the analog sensing output level for one minute until it exceeds, and in the case of the difference value, for example, the analog sensing output level when the fire alarm level is exceeded and the fixed time The difference between the analog sensed output level at a predetermined time and the analog sensed output level after a predetermined time, such as the difference between the analog sensed output level after that, is obtained. The difference between the analog sensing output levels sampled in a predetermined section is obtained and the sum of the absolute values of the differences is obtained, such as the sum of the absolute values of the differences from when the level exceeds the level to when the standard fire alarm level L ₀ is reached. It is.

The certainty determining unit 4 determines a certainty of a fire and a certainty of a non-fire according to a certainty rule set including one or a plurality of feature amounts in the above-described feature amount group. That is, the certainty determining unit 4 defines a fuzzy variable in order to determine whether a fire / non-fire occurs based on whether the characteristic amount (variable) is large or small. Here, the fire / non-fire decision rule consists of an antecedent part (condition part) and a consequent part (conclusion part). The antecedent part is composed of AND combination of multiple fuzzy variables, and the consequent part is the antecedent. A certainty factor of a phenomenon assumed when a part is established is assigned. As a method of inference, first, the fitness of each fuzzy variable in the antecedent part is calculated, and since the condition is AND-join, the product of the fitness of those variables is calculated, and the value of this product is the confidence of the phenomenon defined in the consequent part. And That is, the result of the inference becomes this certainty factor.

FIG. 2 shows a calculation process in the case of obtaining a certainty factor for cigarettes. FIG. 2 (a) shows a fuzzy variable in a case where the larger the feature amount X1, the larger the degree of conformity, and FIG. 2 (b). Indicates a fuzzy variable in the case where the smaller the feature amount X2, the higher the fitness level. In the illustrated example, the fitness levels are 0.7 and 0.8, respectively, and the product of these fitness levels is 0.7 × 0. .8, a confidence level of tobacco of 0.56 is obtained.

Inference is usually performed by calculating each set of decision rules for three phenomena of fire, tobacco, and water vapor in parallel, and determining the certainty factor of each phenomena. When there are a plurality of rule sets for one phenomenon, the larger one of the certainties obtained in each rule set is determined as the certainty of the phenomenon. Next, a determination rule set when the analog fire detector 1 including a smoke detector is used will be described with a specific example.

[0013] In the case where first and the second one fire report level with standard fire report level L _0, when the sensing smoke density of the smoke detector exceeds a first type of fire report level (5% / m) The variables defined as described below are used as the variables of the feature amounts used for determining the certainty factor. One minute until the fire alarm level (5% / m) is exceeded
Average concentration for 3 minutes before exceeding the average fire alarm level (5% / m) during 1 minute before exceeding the alarm level (5% / m)
Difference between the average concentration of the half and the latter half Example for 3 minutes until the fire report level (5% / m) is exceeded
Example, if the number of data in 0.1 to 2.5% / data number fire report level in m (5% / m) to 3 minutes to over e.g. 2.5 to 5.0% / m Then, the fire determination A set of rules, tobacco determination rules, and water vapor determination rules is composed of a combination of these variables. For example, two rules are used as fire determination rules.
One fire decision rule is made up of a variable, thus obtaining a confidence fire from fit of these two variables.

FIGS. 3A and 3B show the relationship between the degree of conformity of a variable and a feature value obtained from time-series data. Another fire judgment rule consists of a variable. FIG. 4 (a),
(B) shows the relationship between the degree of conformity of the variable and the feature amount obtained from the time-series data.

The larger one of the certainties obtained from the two fire judgment rules is determined as the certainty of the fire. Here, the former fire judgment rule expresses the characteristic that the smoke density output continues in response to a fire that is characterized by a gradual rise in smoke density and the continued detection output. In addition, the average density for one minute: high and the average density for three minutes: high are defined as a condition part.

According to the latter rule, the difference between the average density in the first half and the average density in the second half is large in order to indicate that the smoke density is increasing in response to a fire characterized by a gradual increase in smoke density. . In order to grasp such features, 1
Average density per minute: large and difference between the average density of the first half and the second half of 1 minute: large is defined as a condition part. Meanwhile tobacco decision rule is made up of a variable, to determine the confidence of tobacco from fit of these two variables. FIGS. 5A and 5B show variables ,
The relationship between the degree of matching and the feature amount obtained from the time-series data is shown.

In the case of cigarettes, after the smoker smokes tobacco at the place and the low concentration state continues, the smoke concentration suddenly rises due to accidental inflow of cigarette smoke into the sensor. Is a characteristic, so the cigarette judgment rule is 3
Number of data at 0.1 to 2.5% / m per minute: large and 3
The number of data within 2.5 to 5.0% / m per minute: small is defined as a condition part. Further steam decision rule is made up of a variable, to determine the confidence of the water vapor from the fit of these two variables. FIGS. 6A and 6B show the relationship between the degree of conformity of a variable and a feature amount obtained from time-series data.

[0018] For steam, because once the steam is characterized to exhibit a sharp upward trend smoke density emitted into the air, the water vapor decision rule using the definition of "small average density of 1 minute" I have. Also, in order to express that the smoke density rises from near 0 at once, "0.1
数 2.5% / m, the number of data is small ”. In other words, the average density for one minute: small and the number of data within 0.1 to 2.5% / m for three minutes: small are defined as the condition part.

In the case where the two fire alarm levels are set to the standard fire alarm level L ₀ , if the smoke density detected by the smoke detector exceeds the two fire alarm levels (for example, 10% / m), it is convinced. The variables of the feature amount used for determining the degree include those defined as follows. Beyond the forecast level (5% / m), the fire report level (10
% / M to reach) the slope fire report average of example 1 minute to over level concentrations fire report levels to over example 1 up to more than the average density fire report levels of 3 minutes minutes first and second halves of the flat
For example 0.1 to 3 minutes until it exceeds the difference fire report level equalizing concentration
Data number is in the 2.5% / m Then, the fire decision rule, tobacco decision rule, steam decision rule is a combination of these variables, two of the fire decision rule used when beyond the forecast level There are rules . One fire decision rule is made up of a variable, thus obtaining a confidence fire from fit of these two variables.

FIGS. 7 (a) and 7 (b) show the relationship between the degree of conformity of a variable and a feature value obtained from time-series data. Another fire judgment rule includes a variable and a variable . FIG. 8 (a)
(B) shows the relationship between the degree of conformity of the variable and the feature amount obtained from the time-series data. The larger one of the certainties obtained from the two fire judgment rules is determined as the certainty of the fire.

Meanwhile tobacco decision rule is made up of a variable, to determine the confidence of tobacco from fit of these two variables. FIGS. 9A and 9B show the relationship between the degree of conformity of a variable and a feature amount obtained from time-series data. Further, the water vapor judgment rule is composed of variables, and the certainty factor of the water vapor is determined from the fitness of these variables. FIGS. 10A and 10B show the relationship between the degree of conformity of a variable and a feature amount obtained from time-series data.

The alarm level is determined based on the certainty of each phenomenon obtained by the certainty determining unit 4 as described above, that is, the certainty of fire and the certainty of non-fire (the certainty of cigarette, the certainty of water vapor ...). The deciding unit 6 decides a fire / non-fire, and decides a fire alarm (alarm) level L to be changed based on the judgment result. Note that the feature amounts used may be other than those described above, and the above example is merely an example.

Next, determination of fire / non-fire and determination of the fire alarm level L will be described. First, the alarm level determining unit 6 sets the fire certainty α to, for example, 0 among the fire certainty and the non-fire certainty (cigarette certainty, steam certainty,...) Obtained by the certainty certainty determining part 4. If it is greater than or equal to 5, it is determined that there is a fire, and if the certainty factor α of the fire is less than or equal to 0.5 and the maximum value β of the certainty factors of non-fire is greater than or equal to 0.5, it is determined that there is no fire. In the case of other conditions, a judgment such as judging a fire is made.

Based on the results of these determinations, the fire alarm level L is determined as follows. For example, when it is determined that a fire has occurred, the standard fire alarm level L _{0 is changed} from the lowest fire alarm level L _min according to the certainty factor α (1.0 ≧ α ≧ 0.5).
Take the value up to. The following equation shows an example of a calculation equation in this case. Fire alarm level L = L _min +2 (1−α) × (L ₀ −L
_min ) In addition, when it is determined that there is no fire, the confidence β (1.0 ≧
According to (β ≧ 0.5), a value between the maximum fire alarm level L _max and the standard fire alarm level L ₀ is taken. The following equation shows an example of a calculation equation in this case.

Fire report level L = L _max -2 (1-β) ×
(L _max -L ₀₎ when further determines that fire, fire report level L = L ₀
And Here, when the fire report level L is finally determined after the calculation as described above, a method of determining the calculated numerical value as it is as the fire report level L, or a method of deciding stepwise by rounding down, rounding up, or rounding off Any appropriate method may be used.

When the fire alarm level L is determined to be changed as described above, if the maximum fire alarm level _Lmax is fixed, a non-fire alarm such as a place where the frequency of non-fire alarms is low or no nighttime person is present. _May be determined to be a fixed maximum fire alarm level _Lmax even in a case where the possibility of the alarm is low, and in this case, there is a risk of a false alarm. Therefore, in the present embodiment, the maximum fire alarm level _Lmax is determined in advance for each analog fire sensor 1 or each alert zone in which a plurality of analog sensors 1 are installed based on external factors.

The data of external factors include the frequency of non-fire alarms in the sensor or the caution area, or the operation of a key management system that monitors the air-conditioning equipment or lighting equipment or the lock / unlock provided at the installation location. The alarm level setting unit 7 and the false forecast frequency update storage unit 5 provide data of these external factors to the alarm level determination unit 6. The decision unit 6 lowers the maximum fire alarm level _Lmax for a sensor with a low possibility of non-fire alarm or a sensor provided in the warning area based on the data from these, and conversely, The maximum fire alarm level _Lmax is increased for a sensor having a high possibility or a sensor provided in a guard area.

The air conditioner operation detecting section 8 shown in FIG. 1 detects the on / off state of the air conditioner and supplies a detection signal to the alarm level setting section 7. It is provided in. The presence / absence sensor 9 detects whether or not a person is present in the alert area, and provides a detection signal to the alarm level setting unit 7, and is provided in each alert area. The warning level setting portion 7 sends to the alarm level determination section 6 of these detection signals as data for maximum fire report level L _max settings.

On the other hand, the false forecast frequency update storage unit 5 updates and stores the certainty determination contents of the certainty determination unit 4 and the frequency of false reports when a forecast is actually issued. _It is sent to the alarm level determination unit 6 as data for setting _max . Thus, analog fire detector 1
The processing unit 3 of the receiver 2 which has received the analog sensing output from the CPU 2 obtains the characteristic amount of the time-series data of the analog sensing output,
The rule set determines the confidence determining unit 4 respectively corresponding to the fire report level L ₀ based on the feature amount obtained by the processing section 3 when it exceeds the fire report level ₀ of the alarm level serving standards analog sensor output levels, fires , Tobacco, water vapor, etc., that is, the certainty of fire or non-fire.

The alarm level determination section 6 determines a fire alarm level L, which is an alarm level at which the counting of the accumulation time is to be started, based on the determined certainty factor, using the above formula. The maximum fire alarm level _Lmax used in the calculation for determining the fire alarm level L is the maximum fire alarm determined by external factors as described above corresponding to the analog fire detector 1 or the alert area to be targeted. The change of the fire alarm level is determined within the maximum fire alarm level _Lmax using the alarm level _Lmax .

When the analog detection output level reaches the fire alarm level L determined by the alarm level determination unit 6, the fire determination unit 10 starts counting a predetermined accumulation time. That it changes determined fire report level L is for counting the storage time dating back to the time beyond the fire report level L which analog sensor output level is determined changed if it is lower than the standard fire report level L ₀ , And actually starts counting the remaining storage time obtained by subtracting the elapsed time from that point to the present time from the storage time. Conversely, when the level is changed to a level higher than the standard fire alarm level L ₀ , the counting of the accumulation time is started from a point in time when the analog sensing output exceeds the changed fire alarm level L.

By doing so, in the former case, the end point of the counting of the accumulation time is the standard fire alarm level L _0.
As compared with the case where the counting of the accumulation time is started from the time when the time exceeds the alarm time, the time of the alarm is earlier, which leads to the early detection of the actual fire. In the latter case delay alarm time as compared with the case of starting the counting of the storage time from the time when the alarm time exceeds the fire report level L ₀ of a standard, non-fire report can handle that amount transient event Is reduced.

When the counting of the accumulation time based on the fire alarm level L changed as described above is completed, the fire determination unit 10 determines whether the analog detection output level taken in through the processing unit 3 exceeds the fire alarm level L. A determination is made as to whether or not it is exceeded, and if it is exceeded, a fire report is issued through the fire report unit 13. If the analog sensing output level is lower than the fire alarm level L at this time, the storage functions such as the certainty determining unit 4 and the storage time determining unit 6 are reset at this time, and then the analog sensing output level becomes lower. It goes into a waiting state until it exceeds the fire alarm level again.

By the way, in this type of conventional fire alarm system, if the analog sensing output level falls below the fire alarm level L before the accumulation time T is exceeded as shown in FIG. Therefore, as shown in FIG. 11B, when the analog detection output level exceeds the fire alarm level L again after the reset, the second accumulation time T ′ is set from that point, and as a result, Although the alarm is delayed, in the system of the present invention, even if the analog detection output level once falls below the fire alarm level L during the accumulation time T as shown in FIG. When the fire alarm level L exceeds the fire alarm level L again within the time T, an alarm can be issued at the end of the counting of the accumulation time T because the initially set accumulation time T is retained. It has become. In other words, it leads to early detection of fire.

The fire judging section 10 counts the accumulation time T shown in FIG. 12. At the same time, the bar graph indicator 12 shown in FIG. 13 is driven to visually display the remaining time of the accumulation time T. This display informs the monitoring staff of the situation up to the occurrence of the fire report and makes it easier to respond. Instead of the bar graph, a concentric indicator 13 shown in FIG. 14 (a) or a color display as shown in FIG. 14 (b) in which the display color density or display color is changed, for example, from yellow to orange to red The container 14 may be used. Further, display means for changing the length, frequency, and sounding pattern (for example, the interval between successive buzzers) of the buzzer, and display means for changing the blinking interval and the number of times of the lamp may be used. Not something.

The fire judgment unit 10 judges the fire alarm at the end of the accumulation time T. However, if the time until the counting of the accumulation time T ends is long, the alarm may be delayed. Therefore, for example, the determination of a fire may be made when a predetermined time elapses after the count of the accumulation time T is started beyond the fire alarm level.

In this case, for example, the average density for a predetermined time until the fire alarm level is exceeded, the analog sensing output level after a certain time after the fire alarm level is exceeded, and the sampling value between a certain time after the fire alarm level and a certain time. The judgment of the fire is made by the judgment rule set using the difference value or the like as the feature variable. If this method is used, it is possible to issue a fire report without waiting for the end of the counting of the accumulation time T.

Further, as described above, if the fire report level used as a standard is different from one type or two types, the judgment rule set used as described above is different, and the data set used is naturally different from the time series data. The features are also different. FIG. 16 shows the standard fire alarm levels L ₀₁ and L ₀₁ .
The relationship between ₀₂₂ and the data sets D ₁ and D ₂ to be used is shown.

When the fire report level used as a standard is different as described above, it is necessary to newly adjust the parameters of the judgment rule set. Is adjusted in advance, and when a fire report level to be used as a standard is set to a value near the fire report level, the same judgment rule set is applied. It is possible to work on the safe side with respect to issuing a.

In other words, the judgment rule is described according to whether the feature variable is large or small. If an appropriate feature is used at this time, the fire report level used as a standard increases and the data set used shifts. The value of the feature amount can be changed in a certain direction (that is, whether the value increases or decreases). For example, if the average value of a certain section until the standard fire information level is exceeded is selected as the characteristic amount, the value of the characteristic amount changes in a direction to increase when the standard fire information level used increases. Therefore, when the standard fire alarm level to be used rises, the feature amount is selected and the judgment rule is configured so that the confidence of non-fire changes in the direction of decreasing and the confidence of fire increases in the direction of increasing. Therefore, even if the standard set of fire alarm level of a certain standard is applied when using a slightly increased level of fire alarm level as a standard level, the set of judgment rules can be considered to be on the safe side (fire confidence increases, It works in the direction that the confidence of the fire decreases, and the accumulation time is set shorter.)

Here, as a standard standard fire report level,
When the fire alarm levels L ₀₁ , L ₀₂ , and L ₀₃ are set such that L ₀₁ <L ₀₂ <L _03, the judgment rule set R ₁ and the fire alarm level L which are parameter-adjusted to the fire alarm level L ₀₁ are set.
₀₂ previously provided in advance and a determination rule set R ₂ of the parameter adjustment relative actually the size of a standard fire report level L ₀ that is set corresponding to the analog fire detector 1, as described below All you have to do is select a decision rule set.

For example, when a smoke detector is used as the analog fire detector 1 to be used, the standard fire report level _L01 as a reference is changed to the above-described one type of fire report level (5% / m).
And another standard fire report level L ₀₂
If the fire alarm level of the species is 10% / m and the standard fire alarm level _L03, which is another standard, is 15% / m,
The actually set standard fire alarm level L ₀ is L ₀₁ (5%
/ M) ≦ L ₀ <L ₀₂ (10% / m), the determination rule set corresponding to one type of fire alarm level L ₀₁ described above is used, and L ₀₂ (10% / m) ≦ L ₀ < L ₀₃ (15% / m)
If you will select the rule set determines which corresponds to two fire report level L ₀₂ as described above. And L ₀₀ <
Not applicable if L ₀₁ (5% / m) or L ₀ ≧ L ₀₃ (15% / m).

[0043] Now when the analog sensing output level reaches against fire report level L ₀ of a standard in the description of the above embodiment, but was for the case of determining change fire report level, in fact, fire There is a forecast level as a lower level than the forecast level, and a standard level is also set in advance for this forecast level, and when the analog sensing output level reaches the standard forecast level, a decision to change the forecast level may be made. Of course it is good.

For example, when the two types of fire alarm levels are set as fire alarm levels, the forecast level is 1
The species fire report level may be set as the forecast level. In this case, the certainty determination unit 4 determines the certainty using the above-described determination rule set corresponding to one type of fire report level as the determination rule set corresponding to the forecast level.

Similarly, a standard interlocking report is set for the interlocking alarm level for controlling the smoke prevention equipment, and when the analog sensing output level reaches the standard interlocking alarm level, the interlocking alarm level change is determined. Of course, it is good to do so. It is needless to say that when making a decision to change the forecast level, the fire report level, and the interlocking report level, the rule of the level change is determined so that the report is issued in the order of the forecast, the fire report, and the interlocking report.

[0046]

According to the present invention, when the analog detection output level of the analog fire detector reaches a preset alarm level, the accumulation time is counted and an alarm is issued based on the analog detection output level at the end of the counting. Means for determining a plurality of feature quantities from time-series data until the analog detection output level of the analog fire detector reaches a preset standard alert level, and Means for obtaining a certainty factor of a fire and a certainty factor of a non-fire based on a determination rule set including one or a plurality of feature values of a group; and an alarm level determining means for changing an alarm level according to the determined certainty factor. If the changed alarm level is lower than the standard alarm level, the analog detection output level is changed. The time obtained by subtracting the time from the arrival time to the present time from the accumulation time is counted as the remaining accumulation time, and if the alarm level changed and set is higher than the standard alarm level, the analog sensing output level is changed and set. When the fire level is reached, the accumulation time starts counting, so when the fire is judged, the alarm level can be lowered and the start time of the accumulation time can be started earlier, which enables early detection of actual fires. On the contrary, when a non-fire judgment is made, the alarm level can be set higher than the above-mentioned standard alarm level to delay the start of counting of the accumulation time, so that a non-fire caused by a transient increase in the sensing output level, etc. Information can be reduced.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a fire alarm system of the present invention.

FIG. 2 is an explanatory diagram of determining a certainty factor according to the first embodiment;

FIG. 3 is an explanatory diagram of a fire judgment rule in the judgment rule set according to the first embodiment.

FIG. 4 is an explanatory diagram of another fire determination rule in the above determination rule set.

FIG. 5 is an explanatory diagram of a cigarette determination rule in the above determination rule set.

FIG. 6 is an explanatory diagram of a water vapor judgment rule in the judgment rule set according to the first embodiment.

FIG. 7 is an explanatory diagram of a fire judgment rule in another judgment rule set of the above.

FIG. 8 is an explanatory diagram of another fire determination rule in another determination rule set of the above.

FIG. 9 is an explanatory diagram of a cigarette judgment rule in another judgment rule set of the above.

FIG. 10 is an explanatory diagram of a water vapor judgment rule in another judgment rule set of the above.

FIG. 11 is an explanatory diagram of a conventional example for resetting a storage function.

FIG. 12 is an explanatory diagram for resetting a storage function of the fire alarm system of the present invention.

FIG. 13 is a front view of the indicator for use in the embodiment.

FIG. 14 (a) is a front view of another indicator used in the embodiment. (B) is a front view of another indicator used for the same as the above.

[Description of Signs] 1 Analog fire detector 2 Receiver 3 Processing unit 4 Confidence determination unit 5 False forecast frequency update storage unit 6 Alarm level determination unit 7 Alarm level setting unit 8 Air conditioning equipment operation detection unit 9 Person presence / absence sensor 10 Fire judgment unit 11 Fire notification unit

────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-5-54271 (JP, A) JP-A-2-263299 (JP, A) JP-A-2-263298 (JP, A) JP-A-2-263 245896 (JP, A) (58) Field surveyed (Int. Cl. ⁷ , DB name) G08B 17/00

Claims (2)

(57) [Claims] When the analog detection output level of the analog fire detector reaches a preset alarm level, the accumulation time is counted, and it is determined whether or not an alarm is issued based on the analog detection output level at the end of the counting. In the fire alarm system, means for obtaining a plurality of feature values from time-series data until the analog detection output level of the analog fire detector reaches a preset standard alert level, and one of these feature value groups A means for obtaining a certainty factor of fire and a certainty factor of non-fire based on a judgment rule set composed of a plurality of feature values; and an alarm level determining means for changing an alarm level according to the determined certainty factor. If the set alarm level is lower than the standard alarm level, the analog detection output level reaches the changed alarm level. From time obtained by subtracting from the accumulated time period until the present time counted as the remaining storage time,
If the changed alarm level is higher than the standard alarm level, the fire alarm system starts counting the accumulation time when the analog sensing output level reaches the changed alarm level. 2. The fire alarm system according to claim 1, wherein the alarm level is changed and set stepwise.