JPH0720503B2 - Fire determination device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial field of application) The present invention determines a fire from the time change of a position vector based on the detection signals of a plurality of sensors that detect different physical changes specific to the fire. It relates to a fire determination device.

(Prior Art) Conventionally, the applicant of the present application discloses a device for determining a fire based on a vector locus of an nth-order space based on two or more types of physical change amounts peculiar to a fire. Is proposed.

In this device, for example, the temperature and smoke density are detected by a sensor, and the two detected values or their predicted values are given as 2
A vector in the dimensional space is calculated, and a fire is determined when the vector breaks through a closed curve that gives a preset danger level in the secondary space.

(Problems to be Solved by the Invention) However, in a device for judging a fire from a vector of an n-dimensional space based on such n kinds of detected values or predicted values, the detected value from the sensor is used as it is. In this case, the detected value includes a time fluctuation peculiar to the sensor as a noise component, and an erroneous fire determination is made.

Therefore, in the conventional device, for example, so-called digital filter processing is performed to obtain a moving average of sampled data to reduce the noise component. The fluctuation of the vector locus could not be suppressed sufficiently, and there was a risk that an incorrect fire judgment would be issued.

(Means for Solving Problems) The present invention has been made in view of such problems in the related art. Even if the detected value of the sensor includes a noise component, n
An object of the present invention is to provide a fire judging device capable of making more accurate fire judgment by eliminating temporal fluctuation of a vector locus in a dimensional space.

In order to achieve this object, in the present invention, a plurality of sensors for detecting two or more n kinds of different physical change amounts peculiar to a fire; and n kinds of detection values detected by the sensors or the detection values Vector calculation means for calculating a vector of an n-dimensional space determined by a predicted value calculated based on the value at regular time intervals; and a barycentric position in the n-dimensional space calculated from a plurality of position vectors calculated by the vector calculation means. A center of gravity calculating means for controlling the center of gravity; and a fire determining means for determining a fire whether the vector given by the center of gravity of the center of gravity calculating means is inside or outside a closed curved surface giving a preset danger level in the n-dimensional space; Is provided.

(Operation) In the fire determination device of the present invention having such a configuration, the detected values from the plurality of sensors are not individually filtered, but the n-dimensional space obtained from the detected values at regular time intervals. Calculate centroid position of multiple position vectors in 1
Alternatively, the centroid processing is performed multiple times, and the temporal fluctuation of the vector locus due to noise can be suppressed by this centroid processing, and the temporal change of the vector locus in the n-dimensional space due to the progress of the fire can be suppressed. It can catch properly and make a more accurate fire judgment.

(Embodiment) FIG. 1 is a block diagram showing an embodiment of the present invention.

First, the configuration will be described. 1a, 1b, ... 1n are sensors, and the sensors 1a to 1n detect n kinds of different physical changes such as temperature, smoke concentration, and CO gas concentration, which are peculiar to a fire. , Outputs an analog signal according to the detected amount. 2a, 2b, ... 2n are transmission devices provided for each of the sensors 1a to 1n, and the transmission devices 2a to 2n
2n has a function of converting an analog detection signal from the sensors 1a to 1n into a digital signal and digitally transmitting it to the receiving side.
Of course, A / D conversion may be performed on the receiving side.

It should be noted that the sensor 1a that detects different n types of physical change amounts
~ 1n are installed in the same warning area, and to enable fire detection under the same conditions,
It is installed in close proximity.

A reception control unit 3 includes a data sampling unit 4, and the data sampling unit 4 is connected to output lines from the transmission devices 2a to 2n on the sensor 1a to 1n side. Here, as the digital transmission between the transmission devices 2a to 2n and the data sampling unit 4, a polling method in which the data sampling unit 4 sequentially calls the transmission devices 2a to 2n to transmit the digital data, the transmission devices 2a to 2n side Appropriate digital transmission methods such as a method of sequentially transmitting digital data together with an address code and a method of connecting the transmission devices 2a to 2n to the data sampling unit 4 via independent signal lines can be used.

The sampling data of the data sampling unit 4, which has received the data from the sensors 1a to 1n by such a digital transmission method, is given to the centroid computing unit 5.

The center-of-gravity calculation unit 5 has a buffer memory that stores n types of sensor data output from the data sampling unit 4 in a predetermined sampling cycle. The center-of-gravity calculation process in the center-of-gravity calculation unit 5 is performed as follows. Become.

Now, if three kinds of sensor data Xi, Yi, and Zi are given from the data sampling unit 4 at time ti, it is possible to obtain 3
When a position vector in a three-dimensional space is calculated and, for example, m (where m ≧ 3) position vectors are obtained, The center of gravity position Gn (Xi, Yi, Zi) is calculated based on

In the equation (1), n indicates the number of times of the center of gravity processing.

FIG. 2 is an explanatory diagram showing a calculation process of the center of gravity processing when n = 4 in the equation (1).

First, in FIG. 2 (a), the first center-of-gravity processing for n = 1 is
It shows the three-dimensional space, and three position vectors B
If the coordinate positions of 1, B2, B3 (vector tip positions) are D1, D2, D3, the center of gravity G1 of the triangle in the three-dimensional space connecting these three coordinate positions D1 to D3 is calculated.

FIG. 2B shows coordinate positions D2, D when coordinate position D4 indicated by a new position vector is obtained in addition to FIG. 2A.
Find the first center of gravity G2 of the triangle in the three-dimensional space connecting 3 and D4.

Further, in FIG. 2 (c), when a new coordinate position D5 is obtained in addition to FIG. 2 (b), the first center of gravity G3 of the triangle in the three-dimensional space connecting the three coordinate positions D3, D4, D5 Then, the second center of gravity G4, which is the center of gravity processing for n = 2, is further calculated from the triangle connecting the reception positions G1, G2, G3.

Although not shown below, when the coordinate position D6 is obtained in this way for n = 3, the first center of gravity is calculated from the coordinate positions D4, D5, D6.
Ask for G4. Find the second center of gravity G22 from the first center of gravity G2, G3, G4. Further, when the coordinate position D7 is obtained, the coordinate positions D5, D6,
Determine the first center of gravity G5 from D7. Find the second center of gravity G23 from the first center of gravity G3, G4, G5.

When the coordinate position D8 is obtained for n = 4, the first gravity center G6 is obtained from the coordinate positions D6, D7, D8. Obtain the second center of gravity G24 from the first center of gravity G4, G5, G6. Determine the third center of gravity G32 from the second center of gravity G22, G23, G24. When the coordinate position D9 is obtained, the coordinate position D
First, find the center of gravity G7 from 7, D8, D9. 1st center of gravity G5, G6, G
From the 7th time, find the center of gravity G25 for the second time. Second center of gravity G23, G24, G25
From the third time, find the center of gravity G33. Obtain the fourth center of gravity G41 from the third center of gravity G31, G32, G33.

Of course, the center-of-gravity processing by the center-of-gravity calculation unit 5 is the same as that shown in FIG.
Not limited to a dimensional space, a two-dimensional space when there are two types of sensor data and a four-dimensional space when there are four types of sensor data, such as a four-dimensional space. It will start processing.

Referring again to FIG. 1, the vector given by the barycentric position in the nth-order space obtained by the barycentric processing of the barycentric computing unit 5 is input to the comparing unit 6 and the risk level preset by the closed surface setting unit 7 is set. It is determined whether it is inside or outside the closed curved surface that gives

FIG. 3 shows a closed curved surface that gives a danger level set by the closed curved surface setting unit 7 by taking a two-dimensional space for detecting temperature and CO gas concentration as sensor data as an example. A closed curve f (x) ₀ that gives a danger level for fire judgment is set inside the danger level,
The comparison unit 6 has a position vector at time t ₀ having the position of the center of gravity obtained by the center of gravity processing of the unit of calculation of the center of gravity 5. Is inside the closed curve f (x) _0, the comparison output is not output, and the vector by the processing of the center of gravity obtained at the subsequent time t1 Occurs outside the closed curve f (x) ₀ as shown in the figure, a comparison output is generated.

Referring again to FIG. 1, the output of the comparison unit 6 is given to the control unit 8, and the comparison unit 6 outputs the position vector of FIG. As shown in, the fire alarm signal is output in response to the comparison output when the closed curved surface f (x) ₀ is exceeded.

Next, the operation of the embodiment of FIG. 1 will be described with reference to the flowchart of FIG.

In the flowchart of FIG. 4, first, as shown in block 40, the data sampling unit 4 provided in the reception control unit 3 transmits the detection data from the sensors 1a to 1n transmitted by the transmission devices 2a to 2n at predetermined sampling intervals. Is sampled, and in the next block 42, the position vector in the n-dimensional space is calculated based on the n types of sensor data.

Subsequently, the process proceeds to the calculation of the center of gravity of the block 44. In this case, since n = 4 in the equation (1) is set, the data sampling and the vector calculation of the blocks 40 and 42 are shown in FIG. 2 (a). As described above, for example, in the first center-of-gravity processing in which n = 1 in the three-dimensional space, the first center-of-gravity calculation is performed when the coordinate positions D1, D2, D3, which are three vector data, are obtained. , And finally n = 4 4
The center of gravity calculation of the 9th time is the coordinate position of 9 vector data.
The process is performed when D1 to D9 are obtained, and thereafter, each time new vector data is obtained, the centroid calculation is performed using the newly obtained data.

Next, the process proceeds to decision block 46 and the position vector at time t obtained by the gravity center calculation And a closed surface f (x) ₀ that gives a predetermined risk level are compared, and a position vector based on the centroid calculation is compared. Is inside the closed surface f (x) ₀ , the process again returns to the data sampling in block 40, while the position vector calculated by the center of gravity calculation Is beyond the closed curved surface f (x) ₀ , the process goes to block 48 to output an alarm signal and then returns to the data sampling of the belloc 40 again.

FIG. 5 is a graph showing the vector locus in the two-dimensional space by the center of gravity processing in the embodiment of FIG. 1 obtained in the fire combustion experiment using the temperature and CO gas concentration as the sensor detection values together with the vector locus by the sensor data. It is a figure.

That is, FIG. 5 shows the temperature on the horizontal axis and the CO gas concentration on the vertical axis. As a fire experiment, normal heptane of liquid fuel was burned to detect the temperature and the CO gas concentration. As shown by the curve B, the vector locus by the CO gas concentration draws a greatly changed locus.

On the other hand, the vector locus based on the center-of-gravity processing with n = 4 is as shown by the curve A, and the vector locus A based on the center-of-gravity processing advances to the center of the variation of the vector locus B of the detected value. As a result, n = It has been confirmed that a sufficient filter effect can be obtained by the gravity center processing of 4.

FIG. 6 is a block diagram showing another embodiment of the present invention. In this embodiment, the sensor data sampled by the data sampling unit 4 in the reception control unit 3 is different for each sensor data in the moving average calculation unit 9. After the moving average is obtained with the frequency band, the center of gravity calculation unit 5 performs the center of gravity processing from the moving average data, and the other configuration is the same as that of the embodiment of FIG.

In the embodiment of FIG. 6, since the moving average calculator 9 obtains moving average data having different frequency bands for each sensor data from the sampling data, the number of centroid processing n in the centroid calculator 5 is n = 1. By doing so, a sufficient filter effect can be obtained.

FIG. 7 is a flow chart showing the operation of the embodiment of FIG. 6, and is the same as the flow chart of FIG. 4 except that the moving average calculation of the sampling data is performed in block 50 following the data sampling in block 40. .

FIG. 8 shows a vector locus of the temperature and the CO gas concentration in the two-dimensional space when the center of gravity processing is performed from the moving average of the sampling data in the embodiment of FIG. 6, the detected temperature and the detected CO gas concentration. The vector locus by is B,
The vector locus based on the moving average of the sensor detection values is as shown by C indicated by a broken line, but the fluctuation of the vector locus can be seen only with the moving average. Therefore, the vector locus A is obtained by subjecting the sensor data obtained as the moving average to the centroid processing with n = 1. The vector locus A obtained by this centroid processing is that of the centroid processing with n = 4 shown in FIG. Almost the same as in the case, the vector locus B of the sensor detection value advances along the center of variation, and a sufficient filter effect is obtained.

Further, in the embodiment of FIG. 6, by performing the center of gravity processing based on the moving average of the sensor data, the number of times of the center of gravity processing is only one, so the arithmetic processing itself can be simplified.

In the above embodiment, the position vector subjected to the center of gravity processing in the n-dimensional space is obtained from the sensor detection data itself. However, the sensor data which will be obtained in the future is predicted based on the sensor data and is used as a prediction vector. The fire may be determined based on the processing of the center of gravity based on it.

That is, the position vector obtained from the sensor data at the current time t ₀ is Then, the position vector after t time is By first order approximation, or It is also possible to determine the fire by performing a quadratic approximation, and perform the gravity center processing of the prediction vector approximated in this way.

The predictive calculation for determining a fire from the predictive vector after such a predetermined time is described in detail in JP-A-61-49279.

Further, in the above embodiment, when the sensor data is used as it is, the barycentric process becomes n = 4, and when the moving average of the sensor data is obtained, the barycenter process becomes n = 1. The number of times n of the center of gravity processing can be set appropriately so that a sufficient filter effect can be obtained.

(Effect of the Invention) As described above, according to the present invention, the center-of-gravity processing is performed to find the center-of-gravity position of a plurality of vectors in the n-dimensional space obtained from the detection values of the sensor sampled at regular intervals Since the fire is judged based on the vector obtained by the processing, the noise component that is a temporal fluctuation included in the sensor data is sufficiently suppressed, and the vector locus of the vector locus in the n-dimensional space accompanying the progress of the fire is suppressed. Accurate fire decisions can be made by appropriately creating temporal changes.

Further, as an effect peculiar to the embodiment, by performing the center of gravity processing after obtaining the moving average of the sensor data, it is possible to reduce the number of calculations of the center of gravity processing.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, and FIG. 2 is an explanatory view showing the calculation process of the center of gravity calculation with n = 4 in the embodiment of FIG. 1 in a three-dimensional space. Is a graph showing the setting of a closed curved surface that gives a danger level by taking a two-dimensional space of temperature and CO gas concentration as an example, FIG. 4 is a flow chart showing the operation of the embodiment of FIG. 1, and FIG. FIG. 6 is a graph showing a vector locus obtained by the center of gravity calculation and a vector locus obtained by sensor data in a two-dimensional space of temperature and CO gas concentration according to the embodiment of the figure, and FIG. 6 is a block showing another embodiment of the present invention. FIG. 7 is a flow chart showing the operation of the embodiment of FIG. 6, and FIG. 8 is a vector locus obtained by the centroid calculation according to the embodiment of FIG. 6, a vector locus of sensor data and a vector of moving average data. Trajectory in two-dimensional space of temperature and CO gas concentration It is the vector illustration. 1a to 1n: Sensors 2a to 2n: Transmission device 3: Reception control unit 4: Data sampling unit 5: Center of gravity calculation unit 6: Comparison unit 7: Closed curved surface setting unit 8: Control unit 9: Moving average calculation unit

Claims (1)

[Claims] 1. A plurality of sensors for detecting two or more different types of physical changes that are peculiar to a fire; and n types of detection values detected by the sensors, or a prediction calculation based on the detection values. Vector calculation means for calculating a vector of an n-dimensional space determined by the predicted value at regular time intervals; n from a plurality of vectors calculated by the vector calculation means
A barycenter calculating means for calculating the barycentric position in the dimensional space; a fire depending on whether the vector given by the barycentric position of the barycentric calculating means is inside or outside the closed curved surface giving a preset danger level to the n-dimensional space A fire judging device comprising: a fire judging means for judging the situation;