JP2982660B2 - Offset estimation method and device for smoke detection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for estimating an offset in a smoke detector, and more particularly, to outputting a signal corresponding to the concentration of smoke particles contained in a gas to be measured, and outputting the output signal. The present invention relates to a method and apparatus for estimating an offset in a smoke detection device that calculates a concentration of smoke in a gas to be measured by performing a predetermined process as an input.

[0002]

2. Description of the Related Art Conventionally, a light emitting section and a light receiving section are arranged in a dark box, and a light shielding member for preventing light from the light emitting section from being directly guided to the light receiving section is arranged. When smoke particles enter, scattered light caused by the smoke particles is received by the light receiving unit, and the smoke particles are detected based on an output signal from the light receiving unit (an output signal substantially corresponding to the scattered light intensity). Such a photoelectric smoke sensor has been proposed.

In such a photoelectric smoke sensor, the intensity of the scattered and reflected light caused by the smoke particles is significantly weaker than the intensity of the output light from the light emitting section. It must be minimized that the light is reflected by each part and finally received by the light receiving part.
However, in such a photoelectric smoke sensor, it is necessary to introduce the gas to be measured including smoke particles into the dark box, so that smoke particles adhere to each part of the dark box as the measurement operation is performed. The light from the light emitting unit is scattered and reflected by the attached smoke particles, and a part of the scattered reflected light is received by the light receiving unit. Hereinafter, an output from the light receiving unit due to such scattered reflected light is referred to as an offset. This offset is output from the light receiving unit even if the density of the smoke included in the measurement target definition is 0, so that the measurement accuracy is reduced.

In order to solve the problem caused by the offset, a first capacitor for holding the peak value of the output corresponding to the latest pulse light beam obtained by the light receiving unit, and a first capacitor obtained one time before obtained by the light receiving unit. A second capacitor that holds the peak value of the output corresponding to the pulsed light beam, and a differential amplifier in which the voltage held in each capacitor is applied to each input terminal. Smoke detectors that are not responsive to slow changes such as changes (see Japanese Patent Publication No. 56-6600);
An integration control circuit is provided which integrates the amount of fluctuation of the output of the light receiving unit from a constant reference value for a long time and generates a control signal when this value exceeds a specific threshold value. A photoelectric smoke sensor in which the output amplification factor is controlled so that the fluctuation value from the reference value of the light receiving unit during monitoring becomes 0, so that a report or a failure is not generated due to the influence of a gradually changing external factor. Tableware (Japanese Patent Publication No. 56-1281)
No. 3) and a dirt sensitivity correction circuit obtains an average value per hour of the sensing signal from the analog sensor, and obtains the minimum value per hour in the average value per hour. Remember for several days, and 1
By finding the minimum value per average per hour and correcting the dirt sensitivity according to this minimum value, the dirt detection sensitivity can be corrected to compensate for the dirt contamination in the construction stage, etc. There has been proposed a system (see Japanese Patent Application Laid-Open No. Hei 5-325058) that can prevent a dirt alarm from occurring frequently in stages.

[0005]

[Problems to be Solved by the Invention] Japanese Patent Publication No. 56-6600
The smoke detector having the configuration disclosed in Japanese Patent Laid-Open Publication No. H10-27064 does not respond to a slow change such as aging due to contamination by a pair of capacitors and a differential amplifier. Is impossible, and furthermore, since a slow change is neglected, there is a disadvantage that it is not possible to perform detection with an absolute value.

The photoelectric smoke detector having the structure disclosed in Japanese Patent Publication No. 56-12813 integrates the amount of fluctuation of the output of the light receiving unit from a constant reference value, and when the integrated value exceeds a specific threshold value, emits light. Since the light emission output of the light receiving unit or the output amplification factor of the light receiving unit is controlled so that the fluctuation value from the reference value of the light receiving unit during monitoring becomes zero, it is impossible to estimate the offset, and Since the variation value is set to 0 only when the integral value exceeds the threshold value, there is an unpredictable variation value at other times, and there is a disadvantage that the smoke density detection sensitivity cannot be increased much.

The system having the structure disclosed in Japanese Patent Application Laid-Open No. Hei 5-325058 corrects the dirt sensitivity by using the minimum detection signal as an offset level. For example, this system automatically controls an air purifier ( When used for control such as stopping the air purifier when the dust concentration falls below a certain level, it becomes impossible to use an offset level lower than the offset level. Further, since the adopted offset level is equal to or higher than the actual offset level, there is a disadvantage that the detection sensitivity of the dust concentration cannot be increased much.

When measuring the concentration of smoke, there is a high possibility that the gas to be measured contains not only smoke particles but also particles having a diameter larger than that of the smoke particles. Scattering and reflection occur even when the diameter of the smoke is larger than that of the smoke detector, so that any smoke detector cannot accurately detect the concentration of only smoke.

[0009]

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems, and an offset estimating method and apparatus for a smoke detection device capable of highly accurately estimating an offset affected by contamination inside a dark box. It is intended to provide.

[0010]

According to a first aspect of the present invention, there is provided a method of estimating an offset in a smoke detection apparatus, wherein three or more estimated offset points are extracted from an output signal from a smoke sensor and correspond to the extracted estimated offset points. This is a method of calculating an offset estimation value based on a noise removal signal value. Claim 2
The method of estimating the offset in the smoke detection device of the present invention extracts three or more offset estimation points from the output signal.
In this method, a noise component is removed from an output signal to output a noise removal signal, and three or more offset estimation points are extracted from the noise removal signal.

According to a third aspect of the present invention, there is provided a method for estimating an offset in a smoke detection apparatus, comprising the steps of:
This is a method of extracting an estimated offset point from a section between the point at which the moving average is minimized. An offset estimation method in the smoke detection device according to claim 4 extracts four or more offset estimation points, and calculates three or more based on the extracted offset estimation points.
Obtaining a set of one or more offset estimation points, calculating an offset temporary estimation value based on the noise removal signal value corresponding to the offset estimation point belonging to each set, calculating an average value of the plurality of offset temporary estimation values, In this method, the average value is used as an offset estimation value.

According to a fifth aspect of the present invention, there is provided an offset estimating apparatus for a smoke detection apparatus, comprising: filter means for removing a noise component from an output signal from a smoke sensor to output a noise removal signal;
The apparatus has an offset estimating unit including extraction means for extracting one or more offset estimation points, and calculation means for calculating an offset estimation value based on a noise removal signal value corresponding to the extracted offset estimation points.

[0013]

According to the first aspect of the present invention, a signal corresponding to the concentration of smoke particles contained in a gas to be measured is output from the smoke sensor, and the output signal from the smoke sensor is input to calculate the concentration. In performing predetermined processing in the section to calculate the concentration of smoke in the gas to be measured, three or more offset estimation points are extracted from the output signal from the smoke sensor and correspond to the extracted offset estimation points. Since the offset estimation value is calculated based on the noise removal signal value to be calculated, the offset can be estimated with high accuracy. Then, the concentration of the smoke particles contained in the gas to be measured can be accurately detected due to the fact that the offset is estimated with high accuracy. Here, as the smoke sensor, a photoelectric smoke sensor may be employed, but a smoke sensor other than the photoelectric sensor may be employed.

More specifically, it is known that the smoke sensor detects not only smoke particles contained in the gas to be measured, but also dust, and outputs a signal corresponding to the concentration of the smoke particles and dust. I have. Here, the smoke particles are 0.01 to 1
While pollen has a particle size of about several tens of μm and cotton dust has a particle size of about 100 μm, while dust has a particle size of μm, dust has a particle size of about 100 μm. It can be seen that the particles are larger than the smoke particles. As a result, the output from the smoke sensor includes a sharp impulse-like waveform due to the dust. That is, the concentration of smoke particles attenuates over several hours, while the concentration of dust attenuates in about 5 seconds.

Further, the inventor of the present application has found that when only the smoke particles are contained in the gas to be measured, the smoke decay curve is substantially proportional to the exponential function exp (-αt) (α is a constant, t is time). I found to do. This point will be further described. Without ventilation, it is considered that the smoke in the air is attenuated by adhering to walls and the like. Assuming that the ratio of smoke particles adhering to the wall when it hits the wall is constant, the smoke density D and the time t are considered to have the following relationship. D = D0exp (−αt) where D0 indicates the output voltage of the smoke sensor at time t = 0.

For this reason, the output VD of the smoke sensor having a certain offset β is expressed by the following equation. VD = VD0exp (−αt) + β where VD0 indicates the output voltage of the smoke sensor at time t = 0. At this time, β is equal intervals t3, t2, t1
It can be calculated as follows using the data VD1, VD2, and VD3 of three points collected at (t3-t2 = t2-t1 = td). VD1 = VD0exp (−αt1) + β VD2 = VD0exp (−αt2) + β = VD0 ｛exp (−αt1) * exp (−αtd)｝ + β VD3 = VD0 [exp (−αt1) * ｛exp (−αtd)｝ ² ] + Β where VD0 * exp (−αt1) = a, VD0 * e
Assuming that xp (−αtd) = b, VD1 = a + β, VD2 = ab + β, VD3 = ab ² +
It can be seen that VD1, VD2, and VD3 are obtained by adding β to the geometric progression of the first term a common ratio b.

Therefore, VD3-VD2 = ab (b-1) VD2-VD1 = a (b-1) (VD3-VD2) / (VD2-VD1) = b β = VD1- (VD2-VD1) / (b −1), which shows that the offset β can be estimated from the data of three points.

After the offset β is estimated,
By subtracting the offset β from the output voltage from the smoke sensor, a relationship of D = D0exp (−αt) can be obtained, and the smoke density D can be obtained from this relationship. However, even when the offset is estimated based on the output including the impulse-like waveform and the exponential function, the estimated offset includes an error caused by the impulse-like waveform.

In the offset estimating method according to the first aspect, three or more offset estimating points are extracted from the output of the smoke sensor, and the offset is estimated based on the corresponding signal value of the noise removal signal and the exponential function. In addition, the offset can be estimated with high accuracy, and the concentration of smoke particles contained in the gas to be measured can be accurately detected.

According to a second aspect of the present invention, in extracting three or more offset estimation points from an output signal, a noise component is removed from the output signal to output a noise removal signal. Since three or more offset estimation points are extracted from the noise removal signal,
The influence of the noise component can be eliminated, and the offset can be estimated with higher accuracy.

According to the offset estimating method in the smoke detecting device of the present invention, the starting point of the duration of the noise-removed signal when smooth attenuation continues for a certain time or more, and the point at which the moving average is minimized. Since the offset estimation point is extracted from the section between the above, the accuracy of the offset estimation can be improved. According to the offset estimation method in the smoke detection device of claim 4, four or more offset estimation points are extracted, and three or more sets of offset estimation points are obtained based on the extracted offset estimation points. Since the provisional offset estimation value is calculated based on the noise removal signal value corresponding to the belonging offset estimation point, the average value of the plurality of provisional offset estimation values is calculated, and this average value is adopted as the offset estimation value. The accuracy of offset estimation can be further improved.

According to the present invention, a signal corresponding to the concentration of smoke particles contained in a gas to be measured is output from a smoke sensor, and an output signal from the smoke sensor is input to calculate a concentration. In calculating the concentration of smoke in the gas to be measured by performing predetermined processing by the section,
An offset estimating unit included in the density calculating unit removes a noise component from the output signal by a filter unit to output a noise removal signal, extracts three or more offset estimated points by an extracting unit, and extracts the extracted offset estimated points. The calculation unit calculates an offset estimation value based on the noise removal signal value corresponding to. Therefore, the offset can be estimated with high accuracy, and the concentration of the smoke particles contained in the gas to be measured can be accurately detected due to the highly accurate estimation of the offset. Here, as the smoke sensor, a photoelectric smoke sensor may be employed, but a smoke sensor other than the photoelectric sensor may be employed.

[0023]

Embodiments of the present invention will be described below in detail with reference to the accompanying drawings. FIG. 1 is a plan view showing an internal mechanism of an embodiment of a photoelectric smoke sensor applied to the present invention, and FIG. 2 is a schematic front view. This photoelectric smoke sensor includes a substrate 1 on which a light emitting unit 1a including a light emitting diode and a first light receiving unit 1b including a photodiode are mounted, and a dark box (case body) 2.
Here, as the substrate 1, for example, a printed wiring board made of glass epoxy can be exemplified, and as the dark box 2, for example, a box formed using an acrylonitrile butadiene styrene copolymer (hereinafter abbreviated as ABS) The body can be exemplified.

The dark box 2 is a cubic box as a whole. A wall member extending upward is formed inside the dark box 2 to form a light-emitting portion accommodating space 2a and a first light-receiving portion accommodating space 2b. A lens system 2d is provided in the first light receiving unit accommodation space 2b.
Is provided. Then, the optical axis of the light emitting section 1a and the lens system 2
The light emitting section 1a and the lens system 2d are arranged such that the optical axis of the light d forms a predetermined angle. In addition, a hole for allowing the light emitting unit 1a to enter is formed in the inner part of the light emitting unit housing space 2a, and on the optical axis of the lens system 2d,
A hole is formed in the inner part of the first light receiving unit housing space 2b to allow the first light receiving unit 1b to enter. However, a lens system may be provided in the light emitting unit housing space 2a.

The light emitted from the light emitting section 1a is prevented from being guided directly to the first light receiving section 1b through the lens system 2d in the intermediate portion between the light emitting section housing space 2a and the first light receiving section housing space 2b. A light shielding member 2e is provided. In addition, a noise light shielding member 2f for preventing light that may act as noise light from being guided to the first light receiving portion 1b through the lens system 2d includes a light emitting portion housing space 2a, a first light receiving portion housing space 2b, and a light shielding portion. It is provided at a position substantially facing the member 2e. A plurality of noise light shielding members 2f are provided, and noise light shielding members 2f1 and 2f2 adjacent to each other are provided.
This constitutes a noise light attenuation chamber. The light introduced into the noise light attenuating chamber is reflected a plurality of times on the inner wall surface of the chamber where the absorbance is high, so that its intensity is sufficiently attenuated. A noise light shielding member (first plate member) 2f1 to which the direct light from the light emitting unit 1a is irradiated is formed with a hole 2j at a predetermined position for directly passing the light, and the noise light shielding member 2f1. The light trap 2k is made up of a plate member defining the first light receiving unit housing space 2b and the outer wall member of the dark box 2. The noise light shielding member 2
The noise light shielding member 2f2 facing f1 and the outer wall member of the dark box 2 located between the noise light shielding members 2f1 and 2f2 correspond to the second plate member.

Further, a slit 2g for releasing stress is formed at a predetermined position substantially opposite to the light shielding member 2e, and a predetermined position opposite to the light shielding member 2e across the base of the light shielding member 2e and the slit 2g. The substrate 1 is screwed with screws 2h through these holes. Further, an opening 2i for introducing a fluid is provided at a predetermined position near the tip of the light shielding member 2e. Needless to say, an opening (not shown) for introducing a fluid is also provided at a corresponding portion of the substrate 1.

The operation of the optical smoke sensor having the above configuration is as follows. The substrate 1 on which the light emitting unit 1a and the first light receiving unit 1b are mounted is brought into contact with the lower surface of the dark box 2, and in this state, the substrate 1 and the dark box 2 are integrally connected by a pair of screws 2h, thereby providing optical communication. Complete the smoke sensor assembly. In this state, the light emitting unit 1a, the light shielding member 2
e, the relative positions of the lens system 2d and the first light receiving section 1b are set as expected.

By arranging the photoelectric smoke sensor thus assembled at a predetermined position, the concentration of smoke particles in the atmosphere to be measured can be measured as follows. If the atmosphere to be measured does not contain any smoke particles, the light emitted from the light emitting portion 1a propagates straight as it is, and is not irradiated at all to the lens system 2d due to the presence of the light shielding member 2e. As a result, the first light receiving unit 1
The output signal from b remains small. At least a part of the light propagating linearly is introduced into the optical trap 2k through the hole 2j and is almost completely attenuated by being trapped. Further, the remainder of the light propagating linearly is sufficiently attenuated by multiple reflection by the noise light shielding members 2f1 and 2f2 and the outer wall member of the dark box 2.
Therefore, the noise light caused by the direct light from the light emitting unit 1a hardly affects the first light receiving unit 1b.

FIG. 3 is a flow chart for explaining an embodiment of the offset estimating method according to the present invention. Step S
At P1, data input from the photoelectric smoke sensor and filtering are performed, and data is extracted for every ten. A divergence prevention process is performed in step SP2, and step SP3
Performs the process of detecting the start point of the offset estimation period,
The processing from step SP1 to step SP3 is repeated until it is determined in step SP4 that the start point has been detected.

Then, in step SP5, the data extracted in step SP1 is packed by skipping five data (50
The data is converted into data extracted for each of them, and data input and filtering are performed in step SP6 to extract data for every 50 data. In step SP7, a divergence prevention process is performed, and in step SP8, it is determined whether the data extracted immediately before is smaller than the current data (whether the current data is the end point of the offset estimation period or not). Steps SP6 and SP7 are repeated until it is determined that the data extracted immediately before is smaller than the current data.

Thereafter, in step SP9, it is determined whether or not there are three or more data in the offset estimation period.
If it is determined that there are three or more pieces of data, an offset estimation value is calculated in step SP10, and an offset calculation value is calculated in step SP11. Then, when the offset calculation value is calculated in step SP11, or when it is determined in step SP9 that there are only two or less data in the offset estimation period, the processing in step SP1 is performed again.

FIG. 4 shows a data input and filtering process (steps SP1 and S2 in the flowchart of FIG. 3).
It is a flowchart explaining P6). In step SP1, the number of skips is set to 0, data is input in step SP2, and filtering is performed in step SP3. In step SP4, the number of skips is increased by one, and in step SP5
The processing from step SP2 to step SP4 is repeated until it is determined that the skipping number is equal to the predetermined number (the number set in steps SP1 and SP6 in the flowchart of FIG. 3). Then, in step SP6, the last data returned in the filtering process in step SP3 is written to the first FIFO memory.

By performing this processing, the amount of memory and the calculation time in the computer can be reduced. FIG. 5 is a flowchart illustrating the filtering process (the process of step SP3 in the flowchart of FIG. 4). In step SP1, the sum of the previous input data and 30 times the average data is divided by 31 to obtain new average data. In step SP2, the previous input data is larger than the average data, and It is determined whether the previous input data is larger than the input data. If it is determined that the previous input data is larger than the average data and the previous input data is larger than the input data, the previous input data is converted to the average data in step SP3. Rewrite to Then, when the processing in step SP3 is performed, or in step SP2, it is determined that the previous input data is not larger than the average data and / or that the previous input data is not larger than the input data. If so, the average data is returned to the higher-level processing (the processing in the flowchart of FIG. 4).

By performing this process, it is possible to remove pulse noise and random noise caused by dust having a larger particle size than smoke particles. FIG. 6 is a flowchart illustrating the divergence prevention processing (the processing of steps SP2 and SP7 in the flowchart of FIG. 3). In step SP1, it is determined whether or not the calculated offset value is larger than the current data. If the calculated offset value is larger, in step SP2, the value obtained by dividing the sum of 20 times the calculated offset value and the current data by 21 is used. Is a new offset calculation value. If the processing in step SP2 is performed, or if it is determined in step SP1 that the offset calculation value is equal to or smaller than the current data, the absolute value of the difference between the immediately preceding data and the current data is determined in step SP3. And 20 of the data indicating the roughness of the data
The value obtained by dividing the sum of the times and 21 by 21 is used as data indicating new data roughness, and it is determined in step SP4 whether the data indicating data roughness is smaller than 0.1 mV. If it is determined that the data indicating the data roughness is smaller than 0.1 mV, the value obtained by dividing the sum of 100 times the offset calculation value and the current data by 101 in step SP5 is used as a new offset calculation. Value.
When the processing in step SP5 is performed, or in step SP4, the data indicating the data roughness is 0.1 mV.
If it is determined that this is the case, the process returns to the original flowchart. In the flowchart of FIG. 6, with respect to the data indicating the data roughness, the absolute value of the difference between the immediately preceding input data is obtained each time the input data is input, and the absolute value of the data indicating the data roughness at the corresponding point in time is obtained.
The sum of the doubling and the division is divided by 21 and the contribution of the absolute value is reduced to 1/21 and averaged, thereby obtaining data indicating new data roughness. Therefore,
The data indicating the roughness of the data can be regarded as the average value of the absolute values of the data changes.

In the processing of the flowchart of FIG. 3, since the offset is estimated from the attenuation curve, a large error may occur depending on the shape of the attenuation curve, and the error may gradually increase. However, in the process of the flowchart in FIG. 6, when the offset calculation value is lower than the true offset, the process of bringing the offset calculation value close to the noise removal signal value is performed on the condition that the attenuation curve is close to horizontal. When the calculated offset value is higher than the calculated offset value, the process of lowering the calculated offset value is performed on condition that the value of the noise removal signal is lower than the calculated offset value, so that a large error can be prevented from occurring.

FIG. 7 is a flowchart for explaining the process of detecting the start point of the offset estimation period (the process of step SP3 in the flowchart of FIG. 3). In step SP1, it is determined whether the number of data in the first FIFO memory is 21 or more. If the number of data is 21 or more, in step SP2, the current data, the data before 21 and the intermediate data are added. A change in the slope of the noise removal signal is calculated based on the calculated value, and is stored in the second FIFO memory. Conversely, if the number of data is less than 21, the number of data in the second FIFO memory is set to 0 in step SP3.
If the processing in step SP2 or step SP3 has been performed, it is determined in step SP4 whether or not the number of data in the second FIFO memory is equal to the size of the second FIFO memory. The data indicating that there is no start point of the offset estimation period is returned to the processing of the original flowchart. Conversely, if both are determined to be equal, the second FIFO is determined in step SP5.
The average of the change in the inclination in the memory is calculated, and step SP
In step SP6, the variable i is set to 1, and in step SP7, the average of the changes in the three slopes i to i + 2 from the front of the second FIFO memory is calculated, and in step SP8, the variable i is incremented. And step SP9
, 0.7 times the average calculated in step SP5 is smaller than the average calculated in step SP7, and 1.3 times the average calculated in step SP5 is larger than the average calculated in step SP7. Judge whether or not
0.7 times the average calculated in step SP5 is smaller than the average calculated in step SP7, and
If it is determined that 1.3 times the average calculated in P5 is larger than the average calculated in step SP7, in step SP10, it is determined whether the variable i points to the second from the end of the second FIFO memory. And if it points to something other than the second from the end, step S again
The process of P7 is performed. In step SP9, 0.7 times the average calculated in step SP5 is used in step S9.
Step S5 is greater than or equal to the average calculated in P7, or 1.3 times the average calculated in step SP5.
If it is determined that it is smaller than or equal to the average calculated in P7, the data indicating that there is no start point of the offset estimation period is returned to the processing of the original flowchart. If it is determined in step SP10 that the variable i points to the second from the end of the second FIFO memory,
The data indicating that the start point of the offset estimation period has been detected is returned to the processing of the original flowchart.

In this flowchart, the change in the slope means that the value of the noise removal signal (voltage) sampled at equal intervals in time is, as shown in FIG.
2, d3, then (d3-d2) / (d2-d
Given in 1). As shown in FIG. 9, the attenuation curve of the noise removal signal does not change smoothly at the beginning of the attenuation and at the end of the attenuation, and local unevenness is recognized. However, the processing of the flowchart of FIG. 7 is performed. By doing
The time point at which the local unevenness is recognized at the beginning of the attenuation is set as a start point (see a in FIG. 9), and the influence of the local unevenness can be eliminated. Further, in order to exclude a range where local unevenness is recognized at the end of the attenuation, the process of step SP8 in the flowchart of FIG. 3 is performed.

FIG. 10 is a flowchart for explaining the process of calculating the estimated offset value (the process of step SP10 in the flowchart of FIG. 3). In step SP1, the variable sp is set to 0, and in step SP2, the first
The sp-th data from the beginning of the FIFO memory, the last data (current data), and data intermediate between the two are extracted, and the change in the slope is calculated in step SP3.
In step SP4, the change in the inclination is larger than 0,
In addition, it is determined whether or not the difference is smaller than 1. If it is determined that the change in the slope is larger than 0 and smaller than 1, an offset is calculated from the extracted three data in step SP5. In step SP6, it is determined whether or not the offset is within the range of the offset calculation value ± 30 mV. If it is determined that the offset is within this range, this offset is added to the total offset in step SP7, In step SP8, the variable sp is increased by 2.

If it is determined in step SP4 that the change in inclination is equal to or less than 0 or 1 or if it is determined in step SP6 that the offset is out of the range of the offset calculation value ± 30 mV, The processing of step SP8 is performed as it is. Thereafter, in step SP9, it is determined whether or not the variable sp indicates the end of the first FIFO memory. If it is determined that the variable sp indicates the end, in step SP10, the average of the offsets is obtained from the total offset to estimate the offset. Then, the process returns to the original flowchart. Conversely, the variable sp
Does not indicate the end of the first FIFO memory, the processing of step SP2 is performed again.

In this flowchart, the offset is calculated by calculating the values of the noise removal signals (voltages) sampled at equal intervals in time as shown in FIG.
When d3, d3-[(d3-d2) / ｛(d3
-D2) / (d2-d1) -1}] {(d3-d2) /
(D2-d1)}. FIG. 11 is a flowchart illustrating the process of calculating the offset calculation value (the process of step SP11 in the flowchart of FIG. 3).

In step SP1, it is determined whether or not the calculated offset value is larger than the estimated offset value. If it is determined that the calculated offset value is larger than the estimated offset value, the calculated offset value is doubled in step SP2. A value obtained by dividing the sum of the offset estimated value and the offset estimated value by 3 is set as a new offset calculated value. Conversely, if it is determined that the calculated offset value is smaller than or equal to the estimated offset value, a value obtained by dividing the sum of 30 times the calculated offset value and the estimated offset value by 31 in step SP3 is used as the new calculated offset value. Value. And step S
When the processing of P2 or step SP3 is performed,
The process returns to the original flowchart.

It should be noted that the constants shown in the above flowcharts are merely examples, and
Of course, other values can be used. When the above-described offset estimation method is adopted, three or more pieces of data are extracted from a region where no local unevenness is generated from a noise-removed signal from which noise has been removed by filtering, and the offset is accurately determined. Can be estimated. Then, between the output o of the smoke sensor and the smoke density x, o = ax
Since there is a relationship of + b (where a is a coefficient and b is an offset), the relationship of ob−ax can be obtained by subtracting the estimated offset from the output of the smoke sensor, and based on this relationship, The smoke density D can be calculated.

Although the embodiment using the photoelectric smoke sensor as the smoke sensor has been described above, it goes without saying that a smoke sensor other than the photoelectric sensor may be employed. FIG.
FIG. 1 is a block diagram showing an embodiment of an offset estimating apparatus according to the present invention. The offset estimating apparatus includes a data input from a smoke sensor, a data input for filtering, a filtering unit 11, a data input, a divergence prevention processing unit 12 for performing divergence prevention processing by using data from the filtering unit 11 as an input, A start point detection unit 13 that performs a detection process of a start point of an offset estimation period using data from the input and the filtering unit 11 as an input, an end point detection unit 14 that performs a detection process of an end point of the offset estimation period, A data extraction unit 15 for extracting the above data, an offset estimation value calculation unit 16 for calculating an offset estimation value based on the extracted data, and an offset calculation value calculation unit 17 for calculating an offset calculation value
And

The operation of each component is the same as that of the corresponding step in the flow chart, and a detailed description thereof will be omitted. Therefore, the offset can be estimated with high accuracy by extracting three or more pieces of data from a region where no local unevenness has occurred in the noise removal signal from which noise has been removed by the filtering. Since there is a relation of o = ax + b (where a is a coefficient and b is an offset) between the output o of the smoke sensor and the smoke density x,
The relationship of ob = ax can be obtained by subtracting the estimated offset from the output of the smoke sensor, and the smoke density D can be calculated based on this relationship.

[0045]

According to the first aspect of the invention, the offset can be estimated with high accuracy, and the concentration of smoke particles contained in the gas to be measured due to the fact that the offset is estimated with high accuracy can be improved. Can be accurately detected. The invention of claim 2 has a specific effect that the accuracy of offset estimation can be improved.

The invention of claim 3 has a specific effect that the accuracy of estimating the offset can be further improved.
According to the invention of claim 4, the offset can be estimated with high accuracy, and the concentration of smoke particles contained in the gas to be measured can be accurately detected due to the fact that the offset is estimated with high accuracy. This has a specific effect that the possible configuration can be simplified.

[Brief description of the drawings]

FIG. 1 is a plan view showing an internal mechanism of an example of a photoelectric smoke sensor.

FIG. 2 is a schematic front view of the same.

FIG. 3 is a flowchart illustrating an embodiment of an offset estimating method according to the present invention.

FIG. 4 is a flowchart illustrating in detail a process of steps SP1 and SP6 in the flowchart of FIG. 3;

FIG. 5 is a flowchart illustrating in detail a process of step SP3 in the flowchart of FIG. 4;

FIG. 6 is a flowchart describing in detail the processing of steps SP2 and SP7 in the flowchart of FIG. 3;

FIG. 7 is a flowchart illustrating in detail a process of step SP3 in the flowchart of FIG. 3;

FIG. 8 is a schematic diagram illustrating a process of calculating a change in the slope of the noise removal signal.

FIG. 9 is a schematic diagram showing an attenuation curve of a noise removal signal together with a start point and an end point of an offset estimation period.

FIG. 10 is a flowchart describing in detail a process of step SP10 in the flowchart of FIG. 3;

FIG. 11 is a flowchart illustrating in detail a process of step SP11 in the flowchart of FIG. 3;

FIG. 12 is a block diagram showing an embodiment of an offset estimating apparatus according to the present invention.

[Explanation of symbols]

Reference Signs List 11 data input and filtering unit 15 data extraction unit 16 offset estimated value calculation unit 17 offset calculated value calculation unit

Continuation of the front page (58) Field surveyed (Int.Cl. ⁶ , DB name) G08B 17/02-17/12 G01N 21/53

Claims (5)

(57) [Claims] 1. A smoke sensor for outputting a signal corresponding to the concentration of smoke particles contained in a gas to be measured, and a predetermined process performed by using an output signal from the smoke sensor as an input to determine the concentration of smoke in the gas to be measured. In the smoke detection device having a density calculating unit for calculating, three or more offset estimation points are extracted from the output signal, and an offset estimation value is calculated based on a noise removal signal value corresponding to the extracted offset estimation point. An offset estimating method in a smoke detector. 2. Extracting three or more offset estimation points from an output signal, removing a noise component from the output signal, outputting a noise removal signal, and extracting three or more offset estimation points from the noise removal signal. The offset estimation method in the smoke detection device according to claim 1, wherein the extraction is performed. 3. An offset estimation point is extracted from a section between a start point of a duration when smooth attenuation continues for a predetermined time or more and a point at which a moving average is minimized in the noise removal signal. An offset estimating method in the smoke detector according to claim 1 or 2. 4. Extracting four or more offset estimation points,
A set of three or more offset estimation points is obtained based on the extracted offset estimation points, a temporary offset estimation value is calculated based on a noise removal signal value corresponding to the offset estimation point belonging to each set, and a plurality of offset estimation points are calculated. 4. The method according to claim 1, wherein an average value of the provisional estimated values is calculated, and the average value is used as an offset estimated value. 5. A smoke sensor for outputting a signal corresponding to the concentration of smoke particles contained in a gas to be measured, and a predetermined process is performed by using an output signal from the smoke sensor as an input to determine the concentration of smoke in the gas to be measured. A smoke detector having a density calculator for calculating, wherein the density calculator removes a noise component from the output signal and outputs a noise removal signal.
Extracting means (15) for extracting three or more estimated offset points; calculating means (16) (17) for calculating an estimated offset value based on a noise removal signal value corresponding to the extracted estimated offset points; An offset estimating device for a smoke detecting device, comprising: an offset estimating unit including: