Detector for detecting seismic motion and prediction device for predicting intensity of seismic motion based on detection result

Abstract

To provide a seismic motion detection / prediction system capable of further quickly and highly precisely detecting P-waves and estimating intensity of S-waves based on the detected P-waves.SOLUTION: A P-wave detector in a seismic motion detection / prediction system includes a plurality of seismometers, a calculation part for calculating a determination value for determining detection of P-waves, a determination part for determining whether or not to detect the P-waves based on the determination value, and a communication part for transmitting P-wave data for use in prediction of S-waves to the outside when the determination part determines that the P-waves are detected. An S-wave prediction device in the seismic motion detection / prediction system includes a communication part for receiving the P-wave data transmitted from the P-wave detector, and an S-wave prediction part for predicting intensity of incoming S-waves based on the P-wave data.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] The present invention relates to seismic motion detection and prediction technology. More specifically, the present invention relates to a seismic motion detection and prediction system that detects P-waves that generate initial tremors and predicts S-waves that generate main tremors based on the detected P-waves, the system comprising a P-wave detection device that detects P-waves using a plurality of seismometers, and an S-wave prediction device that predicts the intensity of S-waves based on the P-waves detected by the P-wave detection device. [Background technology]

[0002] Predicting earthquakes is extremely difficult. Therefore, earthquake early warning systems have been proposed that calculate the time of occurrence, epicenter location, and magnitude of an earthquake based on earthquake observation data from numerous seismic observation points immediately after an earthquake occurs, and then notify areas that have not yet been reached of the predicted arrival time and intensity of the main seismic tremors based on this information.

[0003] For example, Patent Document 1 (Japanese Patent Publication No. 2015-25714) discloses a disaster warning interlocking system that issues warnings based on information from seismometers installed in buildings in areas near the epicenter, and is linked to on-site warnings for earthquake response according to the seismic intensity and damage situation. This system is configured to include various home appliances installed in buildings and connected to seismometers in a communicative manner, and a central server that is connected to seismometers throughout the country in a communicative manner and aggregates earthquake information from those seismometers throughout the country.

[0004] Furthermore, the applicant of this application has proposed an earthquake warning system disclosed in Patent Document 2. This earthquake warning system is configured to install three or more seismometers on the premises of a building or business establishment, to detect P-waves based on the correlation coefficient between the data signals measured by these seismometers, and to predict the intensity of S-waves based on the P-wave data received from the P-wave detection point, and to issue a warning based on this. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2015-25714 [Patent Document 2] Patent No. 6887310 [Overview of the project] [Problems that the invention aims to solve]

[0006] However, in conventional systems, including the system disclosed in Patent Document 1, it is assumed that each seismometer is installed at a considerable distance apart. Therefore, even if information from these seismometers is aggregated in a central server, it is difficult to quickly and accurately detect the P-waves that generate the initial tremors.

[0007] Furthermore, even if conventional systems aggregate information from multiple seismometers into a central server and predict the arrival of S-waves that generate the main seismic waves based on the information from these multiple seismometers, the processing of the vast amount of seismometer data takes time, making it impossible to issue an emergency warning.

[0008] The technology proposed in Patent Document 2 is a useful technology that can detect P-waves quickly and with high accuracy compared to conventional systems, including the system disclosed in Patent Document 1. However, in recent years, considering the need to respond to direct-hit earthquakes, which are a growing concern, particularly in the Tokyo metropolitan area, there is a demand for even faster and more accurate earthquake motion prediction technology.

[0009] The present invention aims to provide an earthquake motion detection and prediction system that can detect P-waves more quickly and accurately, and predict the intensity of S-waves based on the detected P-waves. [Means for solving the problem]

[0010] The present invention provides a P-wave detection device, which is used in an earthquake motion detection and prediction system that detects P-waves that generate initial tremors and predicts S-waves that generate main tremors based on the detected P-waves, and is placed at a P-wave observation point to detect P-waves. The P-wave detection device comprises a plurality of seismometers, a calculation unit that calculates a determination value for determining whether P-waves have been detected, a determination unit that determines whether P-waves have been detected based on the determination value, and a communication unit that transmits P-wave data used to predict S-waves to an external source when the determination unit determines that P-waves have been detected. The determination value for determining whether P-waves have been detected is calculated using a plurality of real-time measurement data for each predetermined time interval (referred to as "primary interval" in this specification) measured by each of the plurality of seismometers.

[0011] In one embodiment, the judgment value is determined using the standard deviation of multiple real-time measurement data. In another embodiment, the judgment value is determined using the mean deviation of multiple real-time measurement data. In yet another embodiment, the judgment value is determined using the correlation function of multiple real-time measurement data between multiple seismometers. The judgment may also be determined using a combination of these methods.

[0012] Preferably, the P-wave detection device further includes a pre-rupture filter that detects pre-rupture ground motion associated with the P-wave that occurs immediately before the P-wave and distinguishes the pre-rupture ground motion from the P-wave. Furthermore, preferably, the P-wave detection device further includes a minute-event filter that detects minute events unrelated to the P-wave that occur immediately before the P-wave and distinguishes the minute events from the P-wave.

[0013] The present invention also provides an S-wave prediction device for predicting S-waves based on detected P-waves, which is used in an earthquake motion detection and prediction system that detects P-waves that generate initial tremors and predicts S-waves that generate main tremors based on the detected P-waves. The S-wave prediction device comprises a communication unit that receives P-wave data transmitted from the above-mentioned P-wave detection device and an S-wave prediction unit that predicts the intensity of the arriving S-waves based on the P-wave data.

[0014] The S-wave prediction device can also be a mobile device, in which case it is preferable to further include a position measurement unit that measures the position of the S-wave prediction device, and a grace period calculation unit that calculates the grace period until the S-wave arrives based on the measured position and the position of the P-wave detection device that detected the P-wave. [Effects of the Invention]

[0015] The seismic motion detection and prediction system according to the present invention is configured to detect and determine P-waves based on multiple seismometers installed inside buildings or on the premises of business establishments, and to predict the intensity of S-waves based on P-wave data received from P-wave detection points. Therefore, the seismic motion detection and prediction system according to the present invention makes it possible to quickly and accurately detect P-waves that generate initial tremors, and to quickly issue emergency warnings regarding main tremors. Furthermore, the seismic motion detection and prediction system according to the present invention is equipped with a filter to distinguish between P-waves and minor events unrelated to earthquakes or pre-earthquake ruptures that occur before an earthquake, thereby preventing missed or underestimated P-waves and enabling more accurate detection and determination of P-waves. [Brief explanation of the drawing]

[0016] [Figure 1] This is a schematic diagram showing an example of the installation of the main system in an earthquake motion detection and prediction system according to one embodiment of the present invention. [Figure 2] This diagram shows an installation image of the main system in the earthquake motion detection and prediction system. [Figure 3A] This is a block diagram illustrating the overview of the earthquake motion detection and prediction system, where each of the main systems functions as a P-wave detection device and an S-wave prediction device. [Figure 3B] This is a block diagram showing the functions of the data processing unit of the main system. [Figure 4] These are flowcharts of the basic processing in the earthquake motion detection and prediction system. (a) shows the flowchart when the main system performs P-wave detection processing, and (b) shows the flowchart when the main system performs S-wave prediction processing. [Figure 5] An example of measurement data by a seismometer is shown. [Figure 6] In the case where the main system of the ground motion detection and prediction system functions as a P-wave detection device, it is a flowchart of a process for performing P-wave detection determination using the standard deviation as an index. [Figure 7] In the case where the main system of the ground motion detection and prediction system functions as a P-wave detection device, it is a flowchart of a process for performing P-wave detection determination using the mean deviation as an index. [Figure 8] In the case where the main system of the ground motion detection and prediction system functions as a P-wave detection device, it is a flowchart of a process for performing P-wave detection determination using the correlation coefficient as an index. [Figure 9] It is a flowchart showing the process of the precursor destruction filter in the P-wave detection device. [Figure 10] It is a flowchart showing the process of the micro event filter in the P-wave detection device. [Figure 11] It is a flowchart showing a specific example of the P-wave detection process in the P-wave detection device. [Figure 12] In the S-wave prediction device, it is a flowchart showing the process of predicting the intensity of the S-wave. [Figure 13] It is a block diagram showing an overview of a mobile S-wave prediction device according to another embodiment of the present invention.

Embodiments for Carrying Out the Invention

[0017] Hereinafter, embodiments of the present invention will be described in detail while referring to the drawings. <X The present invention provides an earthquake motion detection and prediction system that rapidly and accurately detects P-waves, which generate initial tremors, and predicts the intensity of S-waves, which generate the main tremors, based on the detected P-waves. The earthquake motion detection and prediction system rapidly and accurately detects P-waves (Primary waves), which are the earthquake motions that generate initial tremors, using multiple seismometers, and can predict S-waves (Secondary waves), which are the earthquake motions that generate the main tremors, based on the detected P-waves.

[0018] The seismic motion detection and prediction system can also detect seismic motion caused by preceding ruptures related to the earthquake, which may occur immediately before the rupture that causes the earthquake, and can distinguish this seismic motion from P-waves. Furthermore, the seismic motion detection and prediction system can also detect seismic motion caused by unrelated minor events that may occur independently immediately before the earthquake, and can distinguish this seismic motion from P-waves. In this way, by detecting seismic motion from preceding ruptures and minor events, and distinguishing these seismic motions from P-waves, the detection of P-waves can be made more reliable.

[0019] [Overview of the Earthquake Motion Detection and Prediction System] Figure 1 is a schematic diagram showing an example of the installation of the main system in an earthquake motion detection and prediction system according to one embodiment of the present invention. In the earthquake motion detection and prediction system 1 according to the embodiment of the present invention, three seismometers are installed inside a building or on the premises of a business, and P-wave detection and determination are performed based on measurement data from these three seismometers. In this embodiment, as shown in Figure 1, an example is shown in which three seismometers consisting of a first seismometer 101, a second seismometer 102, and a third seismometer 103 are installed, but the number of installed seismometers may be multiple, such as two or four or more. However, when a majority vote determination is made by multiple seismometers for P-wave detection, it is preferable that the number of seismometers be odd.

[0020] Furthermore, in this embodiment, the first seismometer 101 and the second seismometer 102 are shown to be installed inside the building, and the third seismometer 103 is installed on the site to which the building belongs, but the installation method is not limited to this. However, it is preferable that the three seismometers be installed on the site to which the building belongs (including inside the building) at a certain distance apart. If the seismometers are placed close together, for example, multiple seismometers may measure vibrations from trucks traveling on nearby roads, which may interfere with P-wave detection. The appropriate distance between seismometers is set appropriately depending on the building and site to which the seismometers are installed, but for example, a distance of about 30m to 100m is assumed.

[0021] In the earthquake motion detection and prediction system 1, the first seismometer 101, the second seismometer 102, the third seismometer 103, and the system that processes the data measured by these seismometers are referred to as the main system 100. Figure 2 is a diagram showing an installation image of the main system 100 of the earthquake motion detection and prediction system 1 according to an embodiment of the present invention, and shows the main system 100 installed at points A, B, C, ...

[0022] In this embodiment, the seismic motion detection and prediction system 1 is configured such that the main systems 100 are deployed in various locations as described above, and each main system 100 can communicate with each other via a network. Therefore, a main system 100 that detects a P-wave can transmit the measured P-wave data to other main systems 100. At this time, each main system 100 performs P-wave detection determination using the measurement data measured by the three seismometers 101, 102, and 103, making it possible to perform rapid and highly accurate P-wave detection.

[0023] Figure 3A is a block diagram showing an overview of the seismic motion detection and prediction system 1 according to the present invention, and shows, for example, a state in which multiple main systems 100A, 100B, 100C, ..., 100X are connected to each other via a network N. In Figure 3A, the main systems installed at points A, B, C, ..., and X are designated as 100A, 100B, 100C, ..., and 100X, respectively. Since these configurations are similar, only one main system 100A will be described here.

[0024] The main system 100A used in the earthquake motion detection and prediction system 1 according to the present invention has a first seismometer 101, a second seismometer 102, and a third seismometer 103 installed at an appropriate distance apart at one location. The data measured by the first seismometer 101, the second seismometer 102, and the third seismometer 103 (the measurement data signals of each seismometer are denoted as S1, S2, and S3) is transmitted to the data processing unit 110. The data processing unit 110 in the main system 100A is a general-purpose information processing device consisting of a CPU, a ROM that holds programs that run on the CPU, and RAM which is the CPU's work area. The data measured by the first seismometer 101, the second seismometer 102, and the third seismometer 103 is processed by the data processing unit 110.

[0025] Figure 3B is a block diagram showing the functions of the data processing unit 110. The data processing unit 110 includes a calculation unit, a determination unit, and a prediction unit, and further includes a minute event filter and a pre-destruction filter. The detailed functions of each will be described later.

[0026] The data processing unit 110 operates in cooperation with the various components connected to it as shown in the diagram. Furthermore, various control processes in the earthquake motion detection and prediction system 1 according to the present invention are realized by the CPU executing programs stored in storage means such as ROM and RAM within the data processing unit 110.

[0027] A storage unit 120, such as a hard disk or solid-state drive, is connected to the data processing unit 110. This storage unit 120 can store an S-wave prediction function F used to predict the intensity of S-waves. Furthermore, the storage unit 120 can store programs necessary for the operation of the earthquake motion detection and prediction system 1, initial data, intermediate processing data, facility information (facility name, location information (latitude, longitude), etc.), site information (ground amplification, average depth of direct-hit earthquake, etc.), and the data processing unit 110 can refer to various data. The S-wave prediction function F used to derive the intensity of the S-wave waveform based on P-wave measurement data is dependent on each location, so the storage unit 120 can store the respective S-wave prediction function F for each location. For example, if location A has an S-wave prediction function FA for location A, if location B has an S-wave prediction function FB for location B, and so on. The S-wave prediction function F for each location is prepared in advance and can be stored in the storage unit 120.

[0028] Furthermore, a communication unit 150 is connected to the data processing unit 110, enabling communication with the outside world via wireless or wired means. The communication unit 150 is capable of transmitting data transferred from the data processing unit 110 to main systems 100B, 100C, ..., 100X (which are not its own machine) via an external network N. The communication unit 150 is also capable of receiving data transmitted from main systems 100B, 100C, ..., 100X (which are not its own machine) via the external network N, and transmitting the received data to the data processing unit 110.

[0029] [Processing in the seismic motion detection and prediction system] Next, the processing in the earthquake motion detection and prediction system 1 according to the present invention will be described. The seismic motion detection and prediction system 1 can measure seismic motion, process for determining whether P-waves have been detected, and process for predicting the intensity of S-waves based on P-wave data. When determining whether P-waves have been detected, it can also detect seismic motion caused by minor events and / or preceding ruptures and process them to distinguish these seismic motions from P-waves.

[0030] Figure 4 is a flowchart showing the processing of the seismic motion detection and prediction system 1 according to an embodiment of the present invention. Figure 4(a) shows the processing of one of the multiple main systems 100, which is a main system 100 that detects P waves (for example, main system 100A), and Figure 4(b) shows the processing of a main system 100 that predicts the intensity of S waves based on P wave data from the main system 100 that has detected P waves (main systems 100B, 100C, ..., 100X). All of the main systems 100A, 100B, 100C, ..., 100X that constitute the seismic motion detection and prediction system 1 according to the present invention can perform the processing shown in Figure 4. When performing the processing shown in Figure 4(a), the main system 100 functions as a P wave detection device that detects P waves, and when performing the processing shown in Figure 4(b), the main system 100 functions as an S wave prediction device that predicts S waves.

[0031] (P-wave detection determination) In Figure 4(a), for example, the P-wave detection process is started in the main system 100A, which is closest to the epicenter (S4a-1). Next, in S4a-2, the ground motion is measured in each of the three seismometers 101, 102, and 103. As shown in Figure 5, the ground motion is detected as follows: the waveform of the first seismometer 101 is S1, the waveform of the second seismometer 102 is S2, and the waveform of the third seismometer 103 is S3. Generally, each seismometer detects a waveform of random noise, for example, as shown in the waveform of S1, then the waveform of the P-wave is detected as the earthquake occurs, and then the S-wave is started. In the present invention, the generation of P-waves can be detected with high accuracy at an earlier time, and the intensity of S-waves can be estimated based on the P-wave data. The sampling frequencies of data S1, S2, and S3 from each seismometer 101, 102, and 103 can generally be 100 Hz or 200 Hz, but are not limited to these, and can be any frequency that can be used for P-wave detection and determination in the present invention. In this specification, the respective measurement data S1, S2, and S3 sampled at any frequency are referred to as "real-time measurement data".

[0032] In each seismometer 101, 102, and 103, random noise is measured under normal conditions, and when an earthquake occurs, P-waves are measured. In the main system 100A, in s4a-3, P-wave detection is determined based on the measured values ​​of each seismometer 101, 102, and 103. P-wave detection can be determined using a determination value calculated using a predetermined index, and the predetermined index can be the standard deviation, mean deviation, correlation coefficient, or a combination thereof. The index is calculated using multiple real-time measurement data at predetermined time intervals. In this specification, this predetermined time interval is called a "primary interval". The size of one primary interval is preferably 0.05 seconds to 0.2 seconds from the viewpoint of measurement reliability and P-wave determination reliability, but is not limited to this. If the size of the primary interval is small, the number of real-time measurement data to be averaged will be small, and fluctuations in individual real-time measurement data will be more likely to affect the accuracy of P-wave detection. On the other hand, if the size of the sub-interval is large, the number of real-time measurement data points included in the sub-interval increases, which lengthens the processing time and may make it difficult to quickly confirm the detection of P-waves.

[0033] For example, if the size of the sub-interval is 0.1 seconds and the seismometer's sampling frequency is 100 Hz, the number of real-time measurement data points in one sub-interval is 10. If the size of the sub-interval is 0.05 seconds and the seismometer's sampling frequency is 100 Hz, the number of real-time measurement data points in one sub-interval is 5. The method for determining P-wave detection using each indicator will be described later.

[0034] Thus, in this invention, instead of using the real-time measurement data from the seismometer itself, multiple real-time measurement data acquired over a single sub-interval are averaged and processed, and further averaged and processed from the data of multiple sub-intervals. This reduces the variability in predicted seismic intensity that may occur due to data fluctuations when processing real-time measurement data as is, thereby enabling faster and more accurate predictions.

[0035] If it is determined that a P-wave has been detected, the main system 100A transmits the measured P-wave data to the main systems 100B, 100C, ..., 100X via the network N (s4a-6). The P-wave data includes at least the raw data of the measured P-wave. The transmitted P-wave data may, but is not limited to, data measured by one of the three seismometers 101, 102, and 103 whose noise level during calm periods is at an intermediate value. The data transmitted from the main system 100A is not limited to P-wave data and may also include, for example, the S-wave intensity predicted by the main system 100A. The S-wave intensity included in the transmitted data is preferably the intensity predicted by a method similar to that described later in this specification, but is not limited to that, and may be the intensity predicted by a known method.

[0036] Furthermore, if the seismic motion detection and prediction system 1 includes a pre-rupture filter and a minor event filter, it is preferable that the provisional detection of seismic motion (s4a-4) and the confirmed earthquake (s4a-5) are performed after the P-wave detection determination (s4a-3). Details of these will be described later.

[0037] (Decision based on standard deviation) In one embodiment for determining the detection of P waves, the standard deviation of real-time measurement data is used as the indicator for determination. Figure 6 is a flowchart of the process for determining P wave detection using the standard deviation as an indicator. In this embodiment, the detection of P waves can be determined based on a determination value obtained as the difference between the standard deviation σ[primary interval] of multiple real-time measurement data in a primary interval and a first threshold LT1. The first threshold LT1 is calculated by multiplying the average value σ01, which is obtained by averaging the standard deviations of multiple real-time measurement data obtained for each primary interval under normal conditions, over multiple primary intervals, by a constant K.

[0038] Random noise is measured under normal conditions in each of the three seismometers 101, 102, and 103, and the measured random noise data is sent to the data processing unit 110. The calculation unit in the data processing unit 110 uses the random noise data from each seismometer to calculate the average value σ01, which is obtained by averaging the standard deviations of multiple random noise data obtained for each elementary interval over multiple elementary intervals. σ01 is calculated, for example, by averaging the standard deviations of multiple random noise data obtained for each elementary interval over a period of time that is not limited to, for example, 100 seconds (corresponding to 1000 elementary intervals if the size of the elementary interval is 0.1 seconds). Preferably, σ01 is updated at all times.

[0039] The threshold value LT1 is obtained by multiplying the above average value σ01 by a constant K, i.e., LT1 = K × σ₀₁ This can be done. The constant K is an empirical value obtained from past earthquake data, and for example, a value between 1.5 and 2.0 can be used, but is not limited to this range. The value of K can be arbitrarily determined by the system administrator.

[0040] When an earthquake occurs, the P-wave detection and determination process is started (s6-1). The calculation unit uses multiple real-time measurement data sent from each seismometer 101, 102, and 103 to calculate the standard deviation σ[primary interval] of multiple real-time measurement data in one elementary interval for each seismometer 101, 102, and 103. It is preferable that drift correction is performed on the data from each seismometer using a well-known method before calculating σ[primary interval]. The calculation unit calculates the difference D1 between this σ[primary interval] and the normal threshold LT1. D1 = σ[primary interval] - LT1 The difference data D1 is calculated (s6-2). The difference data D1 is sent to the determination unit provided in the data processing unit 110.

[0041] The determination unit generates a determination value of 1 if the difference data D1 is positive or negative, and a determination value of 0 if the difference data D1 is negative. The determination unit performs a majority vote determination using the determination values ​​of the three seismometers 101, 102, and 103 (s6-3). That is, if one of the three seismometers 101, 102, and 103 has a determination value of 1, it is determined that no P-waves were detected in that elementary interval, and the same determination is made in the next elementary interval (NO in s6-3). On the other hand, if two or more of the three seismometers 101, 102, and 103 have a determination value of 1, it is determined that P-waves were detected in that elementary interval (s6-4). If it is determined that P-waves have been detected (as shown in Figure 4(a), after provisional detection and confirmation of earthquake status as necessary), the measured P-wave data can be transmitted via network N to the main systems 100B, 100C, ..., 100X as described above (s4a-7 in Figure 4).

[0042] (Decision based on mean deviation) Next, in another embodiment for determining the detection of P waves, the mean deviation of real-time measurement data is used as the indicator for determination. Figure 7 is a flowchart of the process for determining P wave detection using the mean deviation as an indicator. In this embodiment, the determination of P wave detection can be made based on a determination value obtained as the difference between the average value Z[primary interval] of multiple real-time measurement data in a primary interval and a second threshold LT2. The second threshold LT2 is calculated using the average value Z0 obtained by averaging the average values ​​of multiple real-time measurement data obtained for each primary interval under normal conditions over multiple primary intervals, and the value obtained by multiplying the standard deviation σ02 of multiple real-time measurement data obtained for each primary interval under normal conditions by a constant K.

[0043] The calculation unit of the data processing unit 110 calculates an average value Z0 by averaging the average values ​​of multiple real-time measurement data (random noise data) obtained for each elementary interval at each seismometer during normal operation, over multiple elementary intervals. Z0 is calculated, for example, by averaging the average values ​​of multiple random noise data obtained for each elementary interval over a period of 100 seconds (for example, if the size of the elementary interval is 0.1 seconds, this corresponds to 1000 elementary intervals). The calculation unit also calculates σ02 using the random noise data from each seismometer. It is preferable that σ02 is constantly updated.

[0044] The threshold LT2 is calculated using the above Z0, σ02, and constant K. LT2 = Z0 + K × σ0² This can be done. The constant K is as described above.

[0045] When an earthquake occurs, the P-wave detection and determination process is started (s7-1). The calculation unit uses multiple real-time measurement data sent from each seismometer 101, 102, and 103 to calculate the average value Z[primary interval] of multiple real-time measurement data in one elementary interval for each seismometer 101, 102, and 103. It is preferable that drift correction is performed on the data from each seismometer using a well-known method before calculating Z[primary interval]. The calculation unit calculates the difference D2 between this Z[primary interval] and the normal threshold LT2. D2 = Z[primary interval] - LT2 The difference data D2 is calculated (s7-2). The difference data D2 is sent to the determination unit provided in the data processing unit 110.

[0046] The determination unit generates a determination value of 1 if the difference data D2 is positive or negative, and a determination value of 0 if the difference data D2 is negative. The determination unit performs a majority vote determination using the determination values ​​of the three seismometers 101, 102, and 103 (s7-3). The majority vote determination is as described above with reference to Figure 6. If it is determined that a P wave has been detected (s7-4) (after provisional detection and earthquake confirmation determination as necessary, as shown in Figure 4(a)), the measured P wave data can be transmitted to the main systems 100B, 100C, ..., 100X via the network N, as described above (s4a-7 in Figure 4).

[0047] (Decision based on correlation coefficient) In yet another embodiment for determining P-wave detection, a correlation function of multiple real-time measurement data at predetermined time intervals between multiple seismometers is used as an indicator. In this embodiment, P-wave detection is determined based on a determination value obtained as the difference between the correlation coefficients r12, r23, and r31 of multiple real-time measurement data in a primary interval obtained between seismometers 101, 102, and 103 and a third threshold LT3. The third threshold LT3 is determined based on the correlation coefficients C012, C023, and C031 of multiple real-time measurement data during normal conditions obtained between seismometers 101, 102, and 103. The threshold LT3 can be a value of, for example, 0.5 to 1.0, but is not limited to this, and can be arbitrarily determined by the system administrator based on past earthquake data or empirical values. Figure 8 is a flowchart of the process for determining P-wave detection using correlation coefficients as an indicator.

[0048] The calculation unit of the data processing unit 110 uses multiple random noise data from the three seismometers 101, 102, and 103 to determine the correlation coefficients C012, C023, and C031 between two seismometers, respectively. Specifically, it calculates the normal correlation coefficient C012 between seismometer 101 and seismometer 102, the normal correlation coefficient C023 between seismometer 102 and seismometer 103, and the normal correlation coefficient C031 between seismometer 103 and seismometer 101. The normal correlation coefficients are calculated, for example, using 100 seconds of random noise data.

[0049] When an earthquake occurs, the P-wave detection and determination process is started (s8-1). The calculation unit uses multiple real-time measurement data sent from each seismometer 101, 102, and 103 to determine the correlation coefficients r12, r23, and r31 of multiple real-time measurement data in one elementary interval for each pair of seismometers 101, 102, and 103. It is preferable that drift correction is performed on the data from each seismometer before determining the correlation coefficients. The calculation unit then calculates the difference between these correlation coefficients r12, r23, and r31 and the normal threshold LT3. D12 = r12 - LT3 D23 = r23 - LT3 D31=r31-LT3 The result is calculated (s8-2). The difference data D12, D23, and D31 are sent to the determination unit provided in the data processing unit 110.

[0050] The determination unit generates a determination value of 1 if the difference data D12, D23, and D31 are positive or negative, and generates a determination value of 0 if they are negative. The determination unit performs a majority vote using the three determination values ​​(s8-3). If only one of the three determination values ​​is 1, it is determined that no P-waves were detected in that elementary interval, and the same determination is made in the next elementary interval ("NO" in s8-3). On the other hand, if two of the three determination values ​​are 1 ("YES" in s8-3), it is determined that P-waves were detected in that elementary interval (S8-4). If it is determined that P-waves have been detected (after provisional detection and earthquake confirmation determination as necessary, as shown in Figure 4(a)), the measured P-wave data can be transmitted to the main systems 100B, 100C, ..., 100X via the network N, as described above (s4a-7 in Figure 4).

[0051] (Pre-emptive destruction filter) It is known that small ruptures (precursor ruptures) occur in the Earth's crust immediately before an earthquake occurs. While the ground motion caused by these precursor ruptures is related to the main shock, it can be an obstacle from the standpoint of P-wave detection and determination. Specifically, when such precursor ruptures occur, there is a risk of misidentifying the ground motion caused by the precursor rupture as true P-waves, or underestimating the magnitude of P-waves, which can lead to an underestimation of S-waves. Therefore, it is preferable that the ground motion detection and prediction system 1 includes a precursor rupture filter to distinguish and remove precursor rupture ground motion, which is the ground motion caused by the precursor rupture, from P-waves.

[0052] The pre-rupture filter uses a pre-rupture index to determine whether the measured seismic motion is a pre-rupture seismic motion. As the pre-rupture index, the average value of multiple real-time measurement data in each sub-interval can be calculated, the difference deviation between the average values ​​of two sub-intervals can be calculated, and the value obtained as a moving average of the two difference deviations can be used. The predetermined number of sub-intervals for determining pre-rupture (the first number) is not limited and can be set as appropriate, taking into account the speed and reliability of the pre-rupture determination.

[0053] Figure 9 is a flowchart of the process for determining pre-rupture using the pre-rupture filter. First, when seismic motion is detected by each seismometer 101, 102, and 103, the pre-rupture filter detects seismic motion in the same way as the P-wave detection determination using one of the methods described above (the method explained using Figures 4(a) to 8). In a sub-interval, if two or more of the three seismometers 101, 102, and 103 have a determination value of 1 (majority vote determination), it is determined that seismic motion has been "provisionally detected" in that sub-interval. This process is carried out over a predetermined number of sub-intervals (first number).

[0054] If it is determined that seismic motion has been provisionally detected consecutively in a predetermined number of sub-intervals (the first number), that seismic motion becomes a "candidate earthquake" (s9-2), and the following processing is performed. While processing is performed in the predetermined number of sub-intervals, the pre-rupture filter calculates a pre-rupture index using real-time measurement data for each sub-interval (s9-3), and determines whether negative values ​​of the pre-rupture index have appeared in a predetermined number of sub-intervals (the second number) or more in the latter half of the predetermined number of sub-intervals (referred to as the "judgment interval") (s9-4). If negative values ​​appear consecutively, the pre-rupture filter determines that the seismic motion is due to pre-rupture (s9-5). Note that the number of judgment intervals and the number of sub-intervals for determining pre-rupture seismic motion are not limited. These numbers of sub-intervals can be set as appropriate, taking into consideration the speed and reliability of the pre-rupture determination.

[0055] (Minor event filter) In large earthquakes, micro-earthquakes often accompany the main earthquake. In particular, in the case of so-called inland earthquakes where the epicenter is close to the point where P-wave detection and judgment are performed, small earthquakes with a magnitude of a few gals and a duration of a few seconds occur before the main earthquake, and these ground motions can interfere with the detection and judgment of P-waves. Specifically, when such micro-earthquakes occur, there is a risk of mistaking the ground motion caused by these micro-earthquakes for true P-waves, or underestimating the magnitude of P-waves, which can lead to an underestimation of S-waves. Therefore, it is preferable that the earthquake motion detection and prediction system 1 includes a micro-event filter to distinguish and remove these micro-event ground motions, which are the ground motions caused by these micro-earthquakes, from P-waves.

[0056] Figure 10 is a flowchart of the process for determining minute events using a minute event filter. The minute event filter uses data from a predetermined number of prime intervals to determine minute events. The predetermined number can be the same as the number of prime intervals (first number) used to determine the preceding failure described above. The predetermined number of prime intervals used to determine minute events is not limited and can be set appropriately considering the speed and reliability of the minute event determination.

[0057] First, as in the case of the preceding rupture described above, when seismic motion is detected by each seismometer 101, 102, and 103, seismic motion is detected in the same manner as the P-wave detection determination using one of the methods described above (the method explained using Figures 4(a) to 8). In a sub-interval, if two or more of the three seismometers have a determination value of 1 (majority vote determination), it is determined that seismic motion has been "provisionally detected" in that sub-interval. This process is carried out over a predetermined number of sub-intervals (first number).

[0058] If seismic motion is tentatively detected continuously in a predetermined number of sub-intervals (the first number), that seismic motion becomes a "candidate earthquake" (s10-2), and the following process, namely "earthquake confirmation determination," is performed. First, the minute event index value Z10 used for determining minute events is calculated using one or all three seismometers, which are arbitrarily selected from the three seismometers (s10-3). The minute event index value Z10 is the average value of multiple real-time measurement data in each of the predetermined number of sub-intervals (the first number), shifted for each sub-interval and averaged (i.e., the moving average value of the sub-intervals). The seismometer to be selected can be, for example, one of three seismometers whose normal random noise level is at the median, but is not limited to this.

[0059] The difference D4 between this minute event index value Z10 and the minute event threshold ST, D4=Z10-ST The following is calculated (s10-4). The minute event threshold ST may be a predetermined value provided to the seismic motion detection and prediction system 1, or it may be made possible for the administrator to change it as needed.

[0060] The micro-event filter generates a judgment value of 1 if the difference data D4 is positive, and a judgment value of 0 if it is 0 or negative (s10-5). When a judgment value of 1 is generated, it is determined that this seismic motion is a true P-wave, i.e., a P-wave has been detected (earthquake confirmed) (s10-7). If it is determined that a P-wave has been detected, the measured P-wave data can be transmitted via network N to the main systems 100B, 100C, ..., 100X (s4a-6 in Figure 4). On the other hand, when a judgment value of 0 is generated, i.e., when the difference data D4 is 0 or negative, it is determined that this seismic motion is due to a micro-event and is not a P-wave (s10-6).

[0061] As described above, the seismic motion detection and prediction system 1 according to the present invention uses an index calculated from data measured by multiple seismometers and distinguishes minute vibrations from true P-waves to perform faster and more accurate detection. The detected P-wave measurement data is transmitted to another main system 100 to contribute to the prediction of S-wave intensity in the other main system 100.

[0062] (Example of a typical earthquake detection process) Figure 11 shows a flowchart of a typical P-wave detection and determination process that combines the methods described above. In this example, real-time measurement data measured in 0.01-second intervals is used. After checking for the presence or absence of real-time measurement data in a given elementary space, the elementary interval average of the real-time measurement data is calculated. In this example, the size of the elementary space is 0.1 seconds.

[0063] Next, the elementary interval determination is performed using the elementary interval average. As described above, the elementary interval determination can be performed using one of the following methods or a combination thereof: comparing the standard deviation of real-time measurement data with a threshold, comparing the mean deviation of real-time measurement data with a threshold, or comparing the correlation coefficient of real-time measurement data between the three seismometers with a threshold. It is determined whether the index exceeds the threshold for each of the three seismometers 101, 102, and 103. If two or more of the three seismometers 101, 102, and 103 exceed the threshold, it is determined that a P-wave was tentatively detected in that elementary interval.

[0064] The process of determining the initial interval is repeated, and when a provisional detection is determined for 10 consecutive initial intervals, the seismic motion becomes a candidate for an earthquake. At this point, in the 10th initial interval, it is determined whether the candidate earthquake is a pre-emptive rupture. The determination of pre-emptive rupture is performed as described above using a pre-emptive rupture index (moving average of differential deviations). Since pre-emptive rupture is performed only once between the provisional detection and the determination of pre-emptive rupture, if pre-emptive rupture has already been performed, the process proceeds to the next step. If the current candidate earthquake is determined to be a pre-emptive rupture, the process starts again from the first provisional detection.

[0065] If the current earthquake candidate is determined not to be a pre-rupture event, or if a pre-rupture event has already been determined, it is determined whether the earthquake candidate is a minor event. The determination of whether it is a minor event is performed as described above using a minor event index (moving average of 10 prime intervals). If the current earthquake candidate is determined not to be a minor event, it is determined to be a true earthquake, and the earthquake is confirmed. When an earthquake is confirmed, the measurement result of the seismometer with the midpoint value among the three seismometers 101, 102, and 103 is taken as the P-wave intensity. Note that the number of seismometers is not limited to three; for example, if two seismometers are used, the measurement result of the seismometer with the larger value can be taken as the P-wave intensity, and if one seismometer is used, the value of that seismometer can be taken as the P-wave intensity. Even if four or more seismometers are used, the P-wave intensity can be determined by an appropriate method. The P-wave intensity is the average value of the real-time measurement data in the 10th prime interval.

[0066] (S-wave intensity prediction) Next, we will describe the processing related to S-wave intensity prediction performed by main systems 100 other than the main system 100 that detected the P-wave. Figure 4(b) is a flowchart of the processing performed by main systems 100 (main systems 100B, 100C, ..., 100X) that predict S-wave intensity based on P-wave data from main system 100A (here, main system 100A functions as a P-wave detector), as described above. These 100B, 100C, ..., 100X function as S-wave prediction devices that predict S-waves. Hereafter, we will explain assuming that main system 100B is the S-wave prediction device that predicts S-waves.

[0067] In Figure 4(b), the S-wave detection process begins in s4b-1. Next, in s4b-2, the main system 100B receives P-wave data transmitted from the main system 100A, which detected the P-wave. When the main system 100B receives the P-wave data, in s4b-3, the prediction unit of the data processing unit 110 of the main system 100B predicts the S-wave intensity using the received P-wave data and the distance attenuation formula. If the predicted S-wave intensity is above a threshold, the main system 100B issues an alarm in s4b-4 according to the intensity level. If the predicted S-wave intensity is below the threshold, the process ends without issuing an alarm (s4b-5).

[0068] Figure 12 is a flowchart of the process for predicting the intensity of S-waves and determining whether or not to issue an alarm. In Figure 12, when the process starts (s12-1) and P-wave data is received in s12-2, the main system 100B obtains a predetermined S-wave prediction function F at point A where the P-wave was detected in s12-3. This S-wave prediction function F is used to predict the intensity of S-waves from the measured P-wave data and is not limited to this function, but can be obtained, for example, as the ratio of P-wave intensity to S-wave intensity for each point in the main system 100 by analyzing past earthquake data. The S-wave prediction function F may be transmitted from the main system 100A along with the P-wave, or it may be read from data previously stored in the storage unit 120 of the main system 100B. For example, when the main system 100A predicts S-waves from P-waves detected by its own seismometer, it can predict S-waves using the S-wave prediction function stored in its own system. Alternatively, if the main system 100B predicts the S wave based on the P wave received from the main system 100A via the network, it can predict the S wave using the S wave prediction function transmitted along with the P wave detected by the main system 100A. In s12-4, the main system 100B multiplies the acquired S wave prediction function F by the P wave measurement data to determine the S wave intensity S at the P wave detection point A. A To predict.

[0069] The main system 100B further grasps how the S wave predicted to have an intensity S at point A affects point B from the distance attenuation of ground vibrations from point A to point B. Therefore, in s12-5, the main system 100B calculates the distance attenuation between point A and point B using a distance attenuation formula well-known to those skilled in the art, which is predetermined and preferably incorporated into a program, and in s12-6, based on the calculated distance attenuation and the intensity S A it calculates the intensity S of the S wave at point B. A B

[0070] When the intensity S B is predicted, in s12-7, the main system 100B compares the obtained intensity S B with a predetermined threshold THS. If the intensity S B is less than the threshold THS, the process ends without issuing an alarm notification (s12-9). If the intensity S B is greater than or equal to the threshold THS, the main system 100B notifies an alarm at the notification unit 130 according to the level of the intensity S B (step s12-8).

[0071] [Embodiment in which the S-wave prediction device is mounted on a moving body] In the ground motion detection and prediction system 1 according to the present invention, further, by mounting the S-wave prediction device on a moving body to form a moving body type S-wave prediction device, it becomes possible to more effectively reduce the damage at the time of an earthquake. For example, when suddenly attacked by a large shake (S wave) during rush hour, particularly large damage is predicted in large cities, and when an earthquake of seismic intensity 6 or 7 arrives on the elevated railway or highway, damage such as derailment and overturning of trains, falling of trains and vehicles from the elevated structure, and multiple collisions of vehicles is assumed. By mounting the S-wave prediction device on a moving body and quickly reporting earthquake information, the disaster can be reduced. Here, the moving body includes not only trains and automobiles but also portable electronic devices such as smartphones, tablets, and smartwatches.

[0072] Figure 13 is a block diagram illustrating an overview of a mobile S-wave prediction device 200 according to another embodiment of the present invention. The mobile S-wave prediction device 200, mounted on a mobile body, includes, for example, a communication unit 151 for receiving various data via the Internet, a position measurement unit 161 for determining its own position using a GPS system or the like, a storage unit 121 for storing various programs, initial data and prediction results, and, if necessary, an S-wave prediction function F, and a notification unit 131 for issuing an alarm based on the prediction results. When, for example, the main system 100A detects a P-wave, the mobile S-wave prediction device 200 receives P-wave data from the main system 100A via the communication unit 151. The mobile S-wave prediction device 200 can predict the S-wave intensity at its own position (the point where the P-wave data was received) based on the received P-wave data, the S-wave prediction function F at the point where the main system 100A detected the P-wave, and the distance attenuation from point A to the mobile device calculated using a predetermined, preferably programmed, distance attenuation formula well known to those skilled in the art. The S-wave prediction function F and the distance attenuation formula are as described in the section on S-wave intensity prediction in the main system described above.

[0073] The mobile S-wave prediction device 200 further includes a grace period calculation unit in the data processing unit 111 that calculates the grace period until the S-wave arrives. The grace period calculation unit uses its own position measured by the position measurement unit 161 and the position of the P-wave detection device (for example, the main system 100A) that detected the P-wave to determine the predicted grace period until the S-wave reaches the location of the mobile S-wave prediction device 200. The grace period is expressed by the following formula, using the distance L between its own position and the position of the P-wave detection device that detected the P-wave. T = L / 4 - L / 7 - a Here, the propagation speed of P-waves is 7 km / s, the propagation speed of S-waves is 4 km / s, and a is the time required for the P-wave detection device to detect the P-waves. The mobile S-wave prediction device 200 issues an alarm via the notification unit 131, which includes the predicted S-wave intensity and the grace period.

[0074] In yet another embodiment, the mobile S-wave prediction device 200 can also be configured to receive earthquake early warnings from the Japan Meteorological Agency. When an earthquake early warning is received, the notification unit 131 issues an alarm based on the earlier of the time the S-wave intensity prediction is completed in the S-wave prediction unit and the time the earthquake early warning is received. Also, when an earthquake early warning is received, the grace period calculation unit uses its own position measured by the position measurement unit 161 and the location of the epicenter included in the earthquake early warning to determine the predicted grace period T' until the S-wave reaches the location of the mobile S-wave prediction device 200. The grace period T' can be determined as described above using the distance L' between its own position and the location of the epicenter included in the earthquake early warning. The mobile S-wave prediction device 200 compares the grace period T calculated based on the position of the P-wave detection device that detected the P-wave with the grace period T' calculated based on the position of the epicenter included in the earthquake early warning, and issues an alarm with a shorter grace period via the notification unit 131.

Claims

1. In a seismic motion detection and prediction system that detects P-waves that generate initial tremors and predicts S-waves that generate main tremors based on the detected P-waves, a P-wave detection device is provided at a P-wave observation point to detect P-waves, A plurality of seismometers installed within a predetermined range at an observation point; a calculation unit that calculates a determination value for determining whether a P wave has been detected, using a plurality of real-time measurement data for each elementary section that is a predetermined time interval measured by each of the plurality of seismometers; a determination unit that determines whether a P wave is detected based on the determination value; a communication unit that transmits P-wave data used to predict S-waves to an external device when the determination unit determines that a P-wave has been detected; Equipped with A preceding rupture filter is provided for determining whether or not a seismic motion is a preceding rupture seismic motion caused by a preceding rupture when the seismic motion is detected, by using a preceding rupture index, and for distinguishing and removing the preceding rupture seismic motion from a P wave. P wave detector.

2. When it is determined that seismic motion has been detected in the determination unit, the determination of the detection of seismic motion is repeated for a first number of elementary sections, and if seismic motion is detected (provisionally detected) continuously in the first number of elementary sections, The preceding destruction filter is calculating an average value of a plurality of real-time measurement data in each of the first number of elementary sections, calculating a difference deviation between the average values ​​of two elementary sections, and calculating the preceding destruction index, which is a value obtained as a moving average of the two difference deviations; When the preceding fracture indicator having a negative value appears consecutively in a second number or more of elementary sections in any judgment section within the first number of elementary sections, the detected earthquake motion is judged to be a preceding fracture earthquake motion.

2. The P wave detection device of claim 1.

3. The method further includes a small event filter that detects a small event unrelated to a P wave that occurs immediately before a P wave and distinguishes the small event from a P wave.

3. A P wave detection device according to claim 1 or 2.

4. When it is determined that seismic motion has been detected in the determination unit, the determination of the detection of seismic motion is repeated for a first number of elementary sections, and if seismic motion is detected (provisionally detected) continuously in the first number of elementary sections, The small event filter includes: calculating a moving average of an average value of a plurality of real-time measurement data in each of the first number of prime intervals; Calculating the difference between the moving average value and a predetermined micro-event threshold value; When the difference is positive, a P wave is determined to be detected.

4. The P wave detection device according to claim 3.

5. The determination value is determined using any one of a standard deviation of a plurality of real-time measurement data, an average deviation of a plurality of real-time measurement data, and a correlation function of a plurality of real-time measurement data between the plurality of seismometers, or a combination of these. A P wave detection device according to any one of claims 1 to 4.

6. the determination value is calculated based on a difference between a standard deviation of a plurality of real-time measurement data in the elementary section and a first threshold value; the first threshold value is calculated using an average value obtained by averaging standard deviations of a plurality of real-time measurement data calculated for each of the elementary intervals under normal conditions over a plurality of the elementary intervals; 6. The P wave detection device according to claim 5.

7. the determination value is calculated based on a difference between an average value of a plurality of real-time measurement data in the elementary section and a second threshold value; the second threshold value is calculated using an average value obtained by averaging the average values ​​of the plurality of real-time measurement data obtained for each of the elementary intervals during normal times over the plurality of elementary intervals, and a value obtained by multiplying the standard deviation of the plurality of real-time measurement data obtained for each of the elementary intervals during normal times by a constant.

6. The P wave detection device according to claim 5.

8. the determination value is calculated based on a difference between a correlation coefficient of a plurality of real-time measurement data of the elementary section calculated among the plurality of seismometers and a third threshold value; the third threshold value is determined using a correlation coefficient of a plurality of real-time measurement data under normal conditions obtained among the plurality of seismometers; 6. The P wave detection device according to claim 5.

9. In a seismic motion detection and prediction system that detects P-waves that generate initial tremors and predicts S-waves that generate major tremors based on the detected P-waves, an S-wave prediction device for predicting S-waves based on the detected P-waves, comprising: A communication unit that receives P-wave data transmitted from the P-wave detection device according to any one of claims 1 to 8; an S wave prediction unit that predicts the intensity of an arriving S wave based on the P wave data; An S wave prediction device comprising:

10. The S wave prediction unit is and receiving a predetermined S-wave prediction function for predicting an intensity of an S-wave at a point where the P-wave data is transmitted; predicting an intensity of an S wave at a point where the P wave data is received based on the P wave data, the S wave prediction function, and distance attenuation between a point where the P wave data is transmitted and a point where the S wave prediction device is located; The S wave prediction device according to claim 9.

11. A position measurement unit that measures the position of the S wave prediction device; a grace period calculation unit that calculates a grace period until the arrival of the S wave based on the measured position and a position of a P wave detection device that detects the P wave; The S wave prediction device according to claim 9 or 10, further comprising:

Images