JP3264112B2 - Earthquake discrimination inference device and earthquake discrimination inference method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an earthquake discriminating inference apparatus and an earthquake discriminating inference method which are attached to, for example, a household gas meter and are suitable for shutting off gas supply when an earthquake occurs.

[0002]

2. Description of the Related Art Conventionally, as this kind of earthquake discriminating inference device, ON / OFF from a seismic sensor built in a gas meter is used.
Some signals are measured for a predetermined period, and when a predetermined condition is satisfied, an earthquake is determined.

[0003]

However, in such a conventional example, after storing ON / OFF signal data from the seismic sensor for a predetermined period of time, for example, three seconds, the data is stored on the basis of the data. Since the discriminating process is performed, there is a disadvantage that the discriminating process cannot be performed until the predetermined period for storing data has elapsed.

[0004] The present invention has been made in view of the above points, and has as its object to provide an earthquake discrimination inference apparatus and an earthquake discrimination inference method which are fast and have few erroneous discriminations .

[0005]

In order to achieve the above-mentioned object, the present invention is configured as follows.

That is, an earthquake discriminating and inference apparatus according to claim 1 of the present invention provides an ON / OFF signal output means for outputting an ON / OFF signal corresponding to a vibration waveform, and an ON / OF signal.
Measured over a predetermined period by the F signal output means
An ON cycle based on the obtained ON / OFF signal,
A feature amount extraction unit that extracts three feature amounts of a total ON time and a maximum value of the ON time, and credibility for determining whether the ON period data extracted by the feature amount extraction unit is credible. For the data of the ON period, the data determined to be credible is used, and for the feature amounts other than the ON period, the data of the feature amount extracted by the feature amount extracting unit is used for the earthquake. And inference means for inferring whether or not.

[0007]

According to the above configuration, ON period, the three feature amounts of the ON time sum total and ON time maximum value out extraction, moreover,
For the ON cycle data, its credibility is determined, and based on the ON cycle data determined to be credible and other feature data, it is inferred whether or not an earthquake has occurred. Compared to a conventional example in which discrimination processing is performed after accumulating ON / OFF signal data, high-speed judgment is possible, and data of an ON cycle corresponding to a vibration cycle is inferred after judging credibility. ,
Misjudgments due to shocks other than earthquakes can be greatly reduced.

[0008]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a block diagram of one embodiment of the present invention.

The earthquake discriminating and inference apparatus 1 of this embodiment includes a seismic sensor 2 as an ON / OFF signal output means j for outputting an ON / OFF signal by vibration and a microcomputer 3
The microcomputer 3 has an ON period, a total ON time, and a maximum ON time value of 3 as described later.
Feature amount extracting means 4 for extracting two feature amounts, and O
A data credibility determining unit 6 for determining whether or not the ON cycle data is credible based on the N time ratio; and the extracted feature amount data and the ON cycle data are determined to be credible. data using, for other feature amounts, using the data of the feature amount extracted by the feature extraction means 4, file whether the seismic each oN cycle
It has a function as inference means 5 for performing inference.

Note that these characteristic amounts are based on data sampled from the time when the first ON signal is input to each time when the ON signal is sequentially input. Is to make an inference based on the feature amount after determining its authenticity. That is, each time the ON signal is input, the feature amount is sequentially updated based on data from the time when the first ON signal is input.

The seismic sensor 2 is shown in FIG. 2B with respect to the vibration waveform shown in FIG. 2A, with a point exceeding a predetermined threshold value as an ON point. ON like
/ OFF signal.

The feature amount extracting means 4 turns on the seismic sensor 2.
The ON period, the total ON time (= ON time 1 + ON time 2 + ON time 3+...) And the maximum value of ON time are sampled and extracted from the / OFF signal every time the ON signal is given. Particularly, in this embodiment, the ON cycle (short, medium, long)
Is the feature amount.

That is, in this embodiment, the following five feature values are extracted.

ON time sum ON time maximum value ON cycle (short) occurrence frequency = ON cycle is 140 msec
Number of occurrences less than ON cycle (medium) occurrence frequency = ON cycle is 140 msec
Occurrence count less than 400 msec ON cycle (long) occurrence frequency = ON cycle is 400 msec
The number of occurrences as described above. The extraction of the ON time total (total of ON time), the ON time maximum value, and the ON cycle occurrence frequency as the feature amounts is based on the following reason.

First, the total ON time is an index indicating the duration of the vibration, and the ON time output from the seismic sensor is used.
The time can be referred to as a speed equal to or higher than a predetermined acceleration. At this time, as shown in the following equation, the total ON time has a correlation with the vibration energy. Is an index that represents

E = (１ ／) mV ² (E: energy
m: mass V: velocity) E∝V ^{2 The} maximum value of the ON time represents the intensity of shaking per half cycle. In the case of a direct earthquake, the duration of the vibration is relatively short. Since the energy per half cycle is large, the feature amount is used.

Further, it is possible to globally grasp a frequency band (period band) higher than a predetermined acceleration from the ON period outputted from the seismic sensor.

Since this ON cycle is の of the oscillation cycle, the ON cycle is shorter than 140 msec (short).
Corresponds to a band with a vibration cycle of less than 0.28 seconds (frequency = 3.6 Hz or more), and this band is a band in which the seismic wave and the shock wave partially overlap, and an ON cycle with an ON cycle of 140 msec or more and less than 400 msec ( Middle) is the vibration cycle =
0.28 seconds to 0.8 seconds (frequency = 1.25 Hz to 3.5
Hz), and the seismic wave has the highest frequency, ie, the central frequency band of the seismic wave. The ON cycle (long) with an ON cycle of 400 msec or more has a vibration cycle of 0.8 seconds (frequency = 1. 25Hz) or more,
It is the band of the response period of seismic waves on soft ground and the response cycle (frequency) of skyscrapers.

By the way, as shown in FIG.
When an impact vibration due to a disturbance as shown in (1) is applied, an excessive acceleration that cannot be generated by an earthquake will be applied.Therefore, in a steel ball type seismic sensor, the steel ball jumps. It cannot be turned off at the time when it should be turned off because of the jumping, and as shown by a in the output waveform of the seismic sensor in FIG. In the case of the damped vibration, the irregular ON signal is output as shown by b, for example, due to the variation of the product of the seismic sensor itself. Therefore, an erroneous ON cycle is detected, and the May be determined.

Therefore, in this embodiment, among the feature quantities extracted by the feature quantity extracting means 4, for the ON cycle, the data credibility judging means 6 judges whether or not the data is authentic as follows. Then, the frequency of occurrence is counted for only the ON period data determined to be authentic and used for inference.

The determination of the authenticity is based on the O cycle with respect to the ON cycle.
This is performed according to the ratio of N times, that is, the ON time ratio.

That is, in this embodiment, when the ON time ratio satisfies, for example, the following condition, it is determined that there is credibility.

0.2 ≦ ON time / ON cycle ≦ 0.8 The lower and upper limits of 0.2 and 0.8 are examples obtained experimentally, respectively. The lower limit value is erroneous ON cycle data due to the output of an irregular ON signal generated by damping vibration near the threshold of the seismic sensor, and the upper limit value is excessive acceleration due to disturbance. OFF when
This is for preventing data of an erroneous ON cycle due to lack of a signal from being used for inference as having no credibility. The above conditional expression for determining authenticity can be transformed into a form of power of 2 that is easy to perform digital processing.

OFF time ≦ 4 × ON time and ON time ≦ 4 × OFF time However, OFF time = ON cycle−ON time Next, fuzzy inference based on the feature amount will be described.

FIG. 4 shows a membership function of the above-mentioned five feature values.

FIG. 3A shows the membership function of the ON total time, and includes three functions of “short”, “medium” and “long”.
"Short" means that the total ON time is 1
“Medium” corresponds to a total ON time of 140 msec or more and less than 400 msec,
“Long” corresponds to a total ON time of 400 msec or more. For example, assuming that the total ON time is less than 140 msec, the fitness of the label “short” is “1”, and the fitness of the labels “medium” and “long” is “0”.

FIG. 3B shows a membership function of the frequency of ON period (short) occurrence, which has two labels of "small" and "many". Short)
Corresponds to three or less occurrences, and “many” corresponds to four or more occurrences of the ON cycle (short). For example, ON
Assuming that the frequency of occurrence of the cycle (short) is three or less, the fitness of the label “small” is “1”, and the fitness of the label “many” is “0”.

FIG. 9C shows a membership function of the frequency of occurrence of the ON cycle (medium), and has two labels of "small" and "many". During)
Corresponds to seven or less occurrences, and “many” corresponds to eight or more occurrences of the ON cycle (medium). For example, ON
Assuming that the frequency of occurrence of the cycle (medium) is a value of 7 or less, the fitness of the label “small” is “1”, and the fitness of the label “many” is “0”.

FIG. 3D shows membership functions of the frequency of occurrence of ON cycle (long), "small" and "many".
"Small" corresponds to three or less occurrences of the ON cycle (long), and "many" corresponds to ON
This corresponds to four or more occurrences of the period (medium). For example,
Assuming that the ON cycle (long) occurrence frequency is a value equal to or less than three times, the fitness of the label “small” is “1”, and the fitness of the label “many” is “0”.

FIG. 5E shows a membership function of the maximum value of the ON time, which has two labels of “short” and “long”. 20
“Long” corresponds to the ON time maximum value of 200 msec or more. For example, if the maximum value of the ON time is less than 200 msec, the fitness of the label “short” is “1” and the fitness of the label “long” is “0”.

FIG. 5 is a diagram showing an inference rule for earthquake discrimination.

In this embodiment, it is determined whether or not an earthquake has occurred according to the following seven rules.

Rule 1. if sum of ON time is short and ON cycle (short) occurrence frequency is small andON cycle (medium) occurrence frequency is large then earthquake Rule 2. if sum of ON time is short and ON cycle (short) occurrence frequency is small andON cycle (long) occurrence frequency is many then earthquake Rule 3. 3. If ON time sum is medium and ON cycle (medium) occurrence frequency is low then earthquake Rule 4. If the total ON time is medium and the ON cycle (short) occurrence frequency is low and the ON ON cycle (medium) occurrence frequency is high then earthquake Rule 5. 5. If total ON time is medium and ON cycle (short) occurrence frequency is small and andON cycle (long) occurrence frequency is large then earthquake Rule 6. 6. If sum of ON time is long and ON cycle (long) occurrence frequency is many then earthquake rule 7. if ON cycle (short) occurrence frequency is large and ON cycle (medium) occurrence frequency is large and ON time maximum value is long then earthquake That is, in this embodiment, the fitness of 1,0 obtained by the above membership function When a logical AND operation is performed on (grade) in accordance with these rules and the result is 1, the earthquake is determined. When the result is 0, it is determined that the earthquake is not an earthquake.

Next, what kind of earthquake is assumed for these rules will be described below.

First, rules 1 and 2 are based on the assumption of an earthquake on relatively soft ground. Although the total ON time is short, the occurrence frequency of the ON cycle (medium) is large, so that a typical earthquake is generated. It is considered that the vibration has a frequency component. In addition, in order to avoid a malfunction, a logical AND with a small frequency of ON cycle (short) occurrence is taken.

Rule 3 assumes a direct earthquake and has a low ON cycle (medium) frequency and a medium total ON time. Therefore, although the frequency is small, the total ON time is medium. Therefore, it is considered that the swing per unit cycle is strong and dangerous.

Rule 4 and rule 5 assume a strong earthquake with a relatively long epicenter, and have a long oscillation period because the oscillation period is medium to long and the total ON time is medium. Is considered a dangerous earthquake.
When the epicenter is far away, the vibration lengthens due to the strength of the reflected wave and shaking. Further, in order to avoid a malfunction, a logical AND with a small frequency of ON cycle (short) occurrence is taken.

Rule 6 is based on the assumption of an earthquake on soft ground and an earthquake shake on a high-rise building, and is a vibration of very large energy due to a long total ON time and a large frequency of ON cycles (long). it is conceivable that.

Rule 7 is based on the assumption that an earthquake occurs on a hard ground and an earthquake that is extremely damaged. The oscillation component of vibration occurs over a wide band, and the maximum value of the ON time is very large. It is considered a big earthquake.

Next, the operation of the earthquake discriminating inference apparatus having the above configuration will be described with reference to the flowchart of FIG. 6 and the timing chart of FIG.

In this earthquake discriminating and inference device, every time an ON signal is given from the seismic sensor 2, ie, every ON cycle,
Data is sampled to extract the above-mentioned features and ON
For the data of the cycle, the data authenticity judgment means 6
The presence or absence of credibility is determined according to the above conditional expression, and inference is performed based on ON cycle data determined to be credible and other feature amounts. That is, every time an ON signal is given, the feature is sequentially updated based on data from the time when the first ON signal is input, and inference is performed based on the updated feature.

Further, in this embodiment, when five seconds have passed since the first ON signal was supplied, data sampling and inference are terminated to set the data in the clear state, and the processing is restarted from the beginning. And finally O
If a new ON signal is not provided after 2.5 seconds or more after the N signal is provided, the inference is terminated.

First, when the seismic sensor 2 is turned on and the first ON signal is given (step n1), the 5-second timer is started (step n2), and the 5-second timer or 2.5-second timer is started. It is determined whether the timer has expired (step n3). If the time has expired, it is determined that the earthquake has not occurred. If not, it is determined whether the seismic sensor has been turned on (step n4). ,
When the seismic sensor 2 is turned on and an ON signal is given, the timer of 2.5 seconds is updated (step n5), and data sampling is performed at a sampling cycle of, for example, 10 msec or less (step n6). The feature amount is extracted (step n7), and inference is performed as described above based on the data of the ON cycle that is extracted and determined to be authentic and the data of the other extracted feature amount (step n8). It is determined whether or not it is an earthquake (step n9). If it is an earthquake, a corresponding output is given. If it is not an earthquake, the process returns to step n3. The shutoff valve is closed and the supply of gas is shut off based on the output indicating that the earthquake has occurred.

As described above, every time the seismic sensor 2 is turned on, data sampling is performed, and a feature quantity based on data from the time when the first ON signal is input is extracted and inference is performed. As described above, the discrimination can be performed faster than the discrimination performed after data sampling from the ON / OFF signal of the seismic sensor for a predetermined period, for example, three seconds.

Further, focusing on the predominant period of the seismic wave / shock wave, and using the ON period corresponding to the vibration period as a characteristic amount, erroneous discrimination due to shocks other than earthquakes can be reduced. Since the ON time total value corresponding to the ON time sum and the vibration intensity per half cycle described in the above is used as the characteristic amount, dangerous vibration can also be determined.

Further, as for the ON cycle, only data determined to be credible is used for inference, so that incorrect ON cycle data generated due to damped vibration near the threshold value of the seismic sensor or excessive due to disturbance. Erroneous ON cycle data generated when an acceleration is applied is not used for inference, whereby erroneous discrimination due to an impact other than an earthquake can be greatly reduced.

In order to evaluate the embodiment, as a result of changing the length of a pipe of a gas meter having a built-in seismic discriminating inference device and hitting it with a steel ball to give an impact or knocking down a gas cylinder to give an impact, It was confirmed that this earthquake discrimination and inference device did not make a misjudgment due to an impact other than an earthquake, and was extremely effective.

[0049]

In the above-described embodiment, only the data of the ON cycle determined to have no credibility is not used for inference. However, as another embodiment of the present invention, the ON cycle data determined to have no credibility is used. For the ON time corresponding to
It may not be used as the data of the ON time sum and the ON time maximum value.

In the above embodiment, the inference has been made for each ON cycle, but the present invention is not limited to each ON cycle.

[0052]

As described above, according to the present invention, the ON period,
High-speed and high-precision discrimination can be performed based on the three feature amounts of the ON time sum and the ON time maximum value, and the credibility of the ON cycle data corresponding to the vibration cycle can be improved. Since it is used for the inference after the judgment, the erroneous judgment due to the impact other than the earthquake can be greatly reduced.

[Brief description of the drawings]

FIG. 1 is a block diagram of one embodiment of the present invention.

FIG. 2 is a diagram illustrating an output of a seismic sensor corresponding to a vibration waveform.

FIG. 3 is a diagram showing a disturbance vibration waveform and an output waveform of a seismic sensor.

FIG. 4 is a diagram showing a membership function of each feature amount.

FIG. 5 is a diagram showing rules of inference.

FIG. 6 is a flowchart for explaining the operation.

FIG. 7 is a timing chart for explaining the operation.

[Explanation of symbols]

 2 Seismic device 3 Microcomputer 4 Feature extraction means 5 Inference means 6 Data authenticity judgment means

Claims (4)

(57) [Claims] An ON / OFF signal output means for outputting an ON / OFF signal corresponding to a vibration waveform, and the ON / OFF signal output means outputs the ON / OFF signal over a predetermined period.
Based on the measured and the ON / OFF signals obtained I, 3 ON period, ON time sum total and ON time maximum value
Feature amount extracting means for extracting three feature amounts; credibility determining means for determining the presence or absence of credibility of the ON cycle data extracted by the feature amount extracting means; and credibility for the ON cycle data. The inferring means for inferring whether or not the determined data is an earthquake using the feature amount data extracted by the feature amount extracting means for the feature amount other than the ON period. Earthquake discrimination inference device. 2. The earthquake discriminating and inference apparatus according to claim 1, wherein the credibility determining means determines the presence or absence of credibility based on a ratio of an ON time to an ON cycle. 3. An on / off signal corresponding to a vibration waveform
ON / OFF signal output means to output
Based on the ON / OFF signal obtained
Then, ON cycle, ON time total and ON time maximum value
A feature amount extracting step of extracting the three feature amounts of: and a signal of the ON cycle data extracted in the feature amount extracting step.
The credibility determination step of determining the presence or absence of credibility, and the ON cycle data was determined to be credible.
For data other than the ON period,
Using the feature data extracted by the quantity extraction means,
And an inference step of inferring whether or not the earthquake is inferred. 4. The method according to claim 1, wherein the credibility determination step is performed for an ON cycle.
Judgment of credibility based on the ratio of ON time
4. The method according to claim 3, wherein the method comprises: