Seismic intensity estimation system
The seismic intensity estimation system uses smart devices with acceleration and magnetic sensors to correct waveforms for movement, ensuring accurate seismic intensity estimation at specific points, overcoming the limitations of existing systems.
Patent Information
- Application Number
- JP2024068131
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-16
- Filing Date
- 2024-04-19
- Publication Date
- 2024-11-28
AI Technical Summary
Existing earthquake intensity estimation systems, such as those using smart devices, struggle to accurately determine seismic intensity at specific points due to movement of the devices during an earthquake, which affects the accuracy of acceleration waveforms.
A seismic intensity estimation system using smart devices with built-in acceleration and magnetic sensors to correct acceleration waveforms by identifying movement sections and adjusting the waveforms based on magnetic sensor data to account for changes in device posture during an earthquake.
The system accurately estimates seismic intensity at any point by correcting acceleration waveforms for movement-induced errors, enhancing accuracy without the need for additional seismometers, and reducing costs.
Smart Images

Figure 2024166089000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to a seismic intensity estimation system using a smart device having an acceleration sensor and a magnetic sensor built in. [Background technology]
[0002] When an earthquake occurs, the seismic intensity is estimated at each of a number of points located within a wide area. For example, Patent Document 1 discloses a seismic intensity information providing device comprising: a database including a memory unit that stores multiple tile image files; an earthquake information table that stores earthquake ID type codes, basic earthquake data, and storage locations of tile image files; an estimated seismic intensity distribution table that stores earthquake ID type codes, measured seismic intensity, and data representing the block for which the measured seismic intensity is estimated; and an observation point seismic intensity table that stores earthquake ID type codes, measured seismic intensity, and data representing the observation point of the measured seismic intensity; a tile image file generation unit that generates multiple tile image files and expands them in the memory unit when the estimated seismic intensity distribution table or the observation point seismic intensity table is updated; an earthquake search unit that outputs data representing the earthquake ID type code and the storage location; and a seismic intensity search unit that outputs data representing the earthquake ID type code and measured seismic intensity. In Patent Document 1, the tile images are colored with a color that represents the seismic intensity value estimated for the range of each tile image. When the range of the tile image includes an observation point where the seismic intensity was observed, a figure including a number that represents the seismic intensity value is drawn in the position corresponding to the observation point. The tile image representing the seismic intensity is displayed on the user's terminal, superimposed on a map image.
[0003] When an earthquake occurs, for example, when trying to estimate the damage situation of a building, it is desirable to obtain the seismic intensity at the point where the building is located more accurately. In this regard, in the configuration of Patent Document 1, even if the point where the building is located is input, the seismic intensity of the entire large mesh including the point is obtained, not that point. Therefore, in the configuration of Patent Document 1, a sufficiently accurate value of the seismic intensity to be used for estimating the damage situation of the building cannot be obtained. On the other hand, in order to accurately estimate the seismic intensity at any point, it is possible to install many seismometers at high density. However, installing many seismometers is costly. Therefore, it is being considered to increase the density of earthquake observation by using acceleration sensors installed in smart devices such as smartphones and tablet terminals, which are owned by many people, instead of seismometers and configuring smart devices to detect earthquakes.
[0004] For example, Patent Document 2 discloses an evaluation system that evaluates the earthquake shaking of an evaluation target. The evaluation system includes an acquisition means that acquires first information indicating the earthquake shaking of the evaluation target, identified by a terminal device provided in the evaluation target, and second information indicating the earthquake shaking around the evaluation target, identified by a terminal device provided around the evaluation target, and an evaluation means that evaluates the earthquake shaking of the evaluation target based on the first information and the second information acquired by the acquisition means. The terminal device is, for example, a smartphone or a tablet terminal.
[0005] Patent document 3 also discloses a configuration including a number of altitude alarms that are arranged vertically within a structure with multiple floors and report their respective height positions to nearby mobile terminal devices, a number of smartphones equipped with a sensor that detects the vibrations of the smartphone and a position information acquisition unit that acquires the current location and altitude of the smartphone, an observation control unit that detects the attitude of the smartphone and starts vibration detection by the sensor based on the duration of that attitude, an information collection unit that acquires the number of smartphones that have started vibration detection and their respective current locations and collects the detection results by the sensor of each mobile terminal, and a determination unit that determines that the vibration is caused by an earthquake when the number of smartphones that have detected vibration exceeds a predetermined number based on the detection results collected by the information collection unit.
[0006] However, when an earthquake occurs, the smart device may not be able to observe seismic motion with high accuracy. For example, if an earthquake occurs while the smart device is placed on a desk, if the earthquake is small, the friction between the smart device and the desk causes the smart device to move along with the desk and not move relative to the desk, so that the smart device can observe seismic motion with high accuracy. However, if the earthquake is large enough that a force exceeding the friction between the smart device and the desk acts on the smart device, the smart device may move relative to the desk and slide, or may fall off the desk. In such a case, the smart device cannot observe seismic motion with high accuracy. [Prior art documents] [Patent documents]
[0007] [Patent Document 1] JP 2017-133996 A [Patent Document 2] JP 2020-193935 A [Patent Document 3] International Publication No. 2018 / 174296 Summary of the Invention [Problem to be solved by the invention]
[0008] The problem to be solved by the present invention is to provide a seismic intensity estimation system that can accurately estimate the seismic intensity at any point. [Means for solving the problem]
[0009] In order to solve the above problems, the present invention employs the following means: That is, the present invention provides a seismic intensity estimation system using a smart device with a built-in acceleration sensor and a magnetic sensor, comprising: a movement determination unit that determines whether the smart device has moved from a stationary state due to an earthquake based on a measurement result by the magnetic sensor, and if it is determined that the smart device has moved, specifies a movement section that is a time section during which the smart device has moved, an acceleration waveform correction unit that corrects a portion of the acceleration waveform observed by the acceleration sensor that corresponds to the movement section when it is determined that the smart device has moved, and generates a corrected acceleration waveform, and a seismic intensity class estimation unit that estimates a seismic intensity class based on the corrected acceleration waveform. A smart device that is placed on a desk or the like and is stationary may slide relative to the surface if the earthquake is large enough to cause a force acting on the smart device that exceeds the static friction between the smart device and the surface on which the smart device is placed. Alternatively, the smart device may bounce up and down on the surface, or may slide on the surface and then fall off the surface. In this way, when a smart device moves from a stationary state due to an earthquake, the direction in which the smart device is placed, i.e., its posture, changes at least a little. When a smart device exhibits this type of behavior, the behavior of the smart device does not follow the behavior of the surface on which it is installed, and therefore the acceleration waveform observed by the acceleration sensor installed in the smart device does not properly observe an earthquake. In the above configuration, the measurement results of the magnetic sensor installed in the smart device are used. If the smart device moves from a stationary state due to an earthquake and its posture changes, the behavior is reflected in the measurement results of the magnetic sensor. Therefore, using the measurement results of the magnetic sensor installed in the smart device, it is possible to determine whether the smart device has moved and to identify the movement section, which is the time section during which the smart device moved. The movement section identified in this way is a section in which the behavior of the smart device does not follow the behavior of the installation surface, in other words, a time section in which the earthquake was not normally observed by the acceleration sensor. In the above configuration, the portion of the acceleration waveform in the movement section is corrected to generate a corrected acceleration waveform, and the seismic intensity scale is estimated based on this. In this way, the acceleration waveform of the time section in which the earthquake was not normally observed is appropriately corrected, so that the seismic intensity scale at the location where the smart device is installed can be more accurately estimated. Many people own smart devices. Therefore, there is a high possibility that a person who owns a smart device is located at or near a location where seismic intensity estimation is required. Therefore, when a certain location is specified, it is possible to provide a seismic intensity estimation system that can accurately estimate the seismic intensity at any location by using the measurement results from a smart device located at the specified location.
[0010] In one aspect of the present invention, the movement determination unit determines that the smart device has moved when the change in the average value of the magnetic flux density before and after an earthquake in the time history waveform of the magnetic flux density, which is the measurement result by the magnetic sensor, is greater than or equal to a movement determination threshold, and when it determines that the smart device has moved, it calculates the standard deviation within a predetermined time interval using the time as a base point while advancing time in the time history waveform of the magnetic flux density, and when the standard deviation becomes greater than or equal to a movement section determination threshold, sets the end time of the time interval using the time as a base point as the start time of the movement section, and when the standard deviation becomes smaller than the movement section determination threshold, sets the start time of the time interval using the time as a base point as the end time of the movement section. According to the above configuration, the change in the average value of the magnetic flux density before and after the earthquake is calculated in the time history waveform of the magnetic flux density measured by the magnetic sensor. If the change is large, it means that the attitude of the smart device has changed significantly before and after the earthquake. Therefore, by comparing the change in the average value of the magnetic flux density before and after the earthquake with a movement determination threshold and determining that the smart device has moved if the change is equal to or greater than the movement determination threshold, it is possible to appropriately determine whether the smart device has moved. If it is determined that the smart device has moved, the movement section is further specified. Since the magnetic flux density value varies greatly while the smart device is moving, the movement section can be accurately specified by calculating the standard deviation within a predetermined time interval in the time history waveform of the magnetic flux density and comparing it with a threshold value. More specifically, the standard deviation is calculated within a predetermined time interval with the time as a base point while advancing the time in the time history waveform of the magnetic flux density. If the standard deviation becomes equal to or greater than the moving section determination threshold at a certain time after the standard deviation has remained smaller than the moving section determination threshold, it is considered that the time has advanced from a state in which the entirety of the predetermined time interval was in a region of the time history waveform where the variation in the magnetic flux density was small, and the most advanced part of the predetermined time interval with the time as the base point has entered a region of the time history waveform where the variation in the magnetic flux density began to increase, and the standard deviation has begun to increase. Therefore, by identifying the end time of the time interval with the time as the base point, the start time of the moving section can be accurately identified. Furthermore, if the standard deviation becomes equal to or greater than the movement section determination threshold at a certain time, it is considered that the standard deviation becomes smaller as time advances from a state in which the entirety of a given time interval was in a region of the time history waveform where the magnetic flux density variation is large, and the entirety of the given time interval starting from that time, including the part that is the latest in time and closest to the start time of the time interval, reaches a region of the time history waveform where the magnetic flux density variation is small. Therefore, by identifying the start time of the time interval starting from that time, the end time of the movement section can be accurately identified. In this way, it is possible to properly determine whether the smart device has moved and to properly calculate the section of movement for which the acceleration waveform observed by the acceleration sensor is to be corrected, thereby properly correcting the acceleration waveform and, as a result, improving the accuracy of the seismic intensity calculation.
[0011] In another aspect of the present invention, the movement determination unit determines whether the smart device has moved horizontally based on the measurement results of the magnetic sensor of components in each of a first direction and a second direction that are perpendicular to each other in a plane parallel to the screen of the smart device, and if it is determined that the smart device has moved horizontally, identifies a horizontal movement section that is the movement section in the horizontal direction, and determines whether the smart device has moved in a vertical direction based on the measurement results of the magnetic sensor of components in an orthogonal direction that is perpendicular to the screen, and if it is determined that the smart device has moved in the vertical direction, identifies a vertical movement section that is the movement section in the vertical direction. When an earthquake occurs and a smart device placed on an installation surface slides on the installation surface, the posture of the smart device changes to some extent. As a result, a change appears in the measurement results of the components in the first direction and the second direction that are perpendicular to each other in a plane parallel to the screen of the smart device measured by the magnetic sensor. In the above configuration, whether or not the smart device has moved in the horizontal direction can be determined based on the measurement results of the components in the first direction and the second direction measured by the magnetic sensor, so that the determination can be made appropriately. In addition, if an earthquake occurs and the smart device placed on the installation surface bounces up and down on the surface or falls off the surface, the measurement results of the components in the orthogonal direction perpendicular to the screen by the magnetic sensor will change. In the above configuration, whether the smart device has moved in the up and down direction can be determined appropriately based on the measurement results of the components in the orthogonal direction by the magnetic sensor.
[0012] In another aspect of the present invention, when it is determined that the smart device has moved in a horizontal direction, the acceleration waveform correction unit generates the corrected acceleration waveform in the first direction and the second direction by complementing a portion of the acceleration waveform in the horizontal movement section where the amplitude of each of the components in the first direction and the second direction is constant by a curve fitting process using a sine wave, and the acceleration waveform correction unit calculates a component composite acceleration by combining the components in the first direction, the second direction, and the orthogonal direction of a waveform based on the corrected acceleration waveform at each time, and calculates a component composite acceleration when the time is within the vertical movement section and the component composite acceleration falls. If the acceleration is greater than or equal to the lower judgment threshold, the smart device is deemed to be falling at that time, and the components of the waveform based on the corrected acceleration waveform in the first direction, the second direction, and the orthogonal direction are set to 0 to recalculate the component resultant acceleration; if the time is within the up / down movement section and the component resultant acceleration is smaller than the fall judgment threshold, the smart device is deemed to be bouncing up and down at that time, and the component of the waveform based on the corrected acceleration waveform in the orthogonal direction is set to 0 to recalculate the component resultant acceleration, and the seismic intensity class estimation unit estimates the seismic intensity class based on the component resultant acceleration. When an earthquake occurs while the smart device is placed on the installation surface, if the seismic force exceeds the maximum static friction force between the smart device and the installation surface and the smart device slides on the installation surface or moves horizontally on the surface while bouncing up and down, the acceleration waveform often tends to have parts with constant amplitude in each of the components in the first and second directions. In such a case, in the above configuration, the parts of the components in the first and second directions of the acceleration waveform that have constant amplitude within the horizontal movement section are complemented by a curve fitting process using a sine wave to generate corrected acceleration waveforms in the first and second directions. Therefore, the corrected acceleration waveforms in the first and second directions can be generated appropriately. In addition, in the vertical movement section, the smart device is in either a bouncing state where the smart device bounces up and down on the installation surface, or a falling state where the smart device falls from the installation surface. Here, the magnitude of acceleration acting on the smart device is different between the bouncing state and the falling state. Therefore, in the above-mentioned configuration, for a waveform based on the corrected acceleration waveform, the components in the first direction, the second direction, and the orthogonal direction are combined at each time to calculate a component composite acceleration, and the component composite acceleration is compared with a fall determination threshold to determine whether the smart device is in a bouncing state or a falling state at each time in the vertical movement section. When the smart device is in a bound state, the smart device floats in the air, falls, and lightly hits a surface. For this reason, the acceleration waveform shows excessive acceleration in the orthogonal direction that would not be observed by a fixed seismometer. If the seismic intensity scale is estimated based on such an acceleration waveform, the seismic intensity scale may be estimated to be higher than the actual one. Therefore, in the above configuration, at the time when the smart device is determined to be in a bound state, the first and second directions are corrected by the curve fitting process using a sine wave as described above, and the component in the orthogonal direction is adjusted to be 0 to recalculate the component composite acceleration, and the seismic intensity scale is estimated based on this component composite acceleration. This reduces the influence of the excessive acceleration observed in the orthogonal direction in the bound state on the estimation of the seismic intensity scale. Furthermore, when the smart device is in a falling state, it is possible that a large acceleration is observed in the smart device, similar to when the smart device is in a bounding state. Furthermore, since the smart device is completely suspended in the air, no external force due to the earthquake acts on the smart device, and no waveforms due to the earthquake are obtained. For this reason, in a falling state, it is difficult to assume that reliable acceleration waveforms that can be used to estimate the seismic intensity scale are obtained in all directions, the first direction, the second direction, and the orthogonal direction, and if the seismic intensity scale is estimated based on such acceleration waveforms, the estimation accuracy of the seismic intensity scale is reduced. Therefore, in the above-mentioned configuration, in a falling state, the component composite acceleration is recalculated by adjusting the components to be 0 in all directions, the first direction, the second direction, and the orthogonal direction, and the seismic intensity scale is estimated based on this component composite acceleration (i.e., the component composite acceleration is 0 at the relevant time). This reduces the influence of acceleration waveforms that are not considered to be reliable in a falling state on the estimation of the seismic intensity scale.
[0013] In another aspect of the present invention, the seismic intensity estimation system further includes a seismic intensity class reliability evaluation unit that evaluates, based on the acceleration waveform observed by the acceleration sensor, each of a placement status evaluation index, which gives a better result the more horizontally the smart device is placed during an earthquake, a drop presence / absence evaluation index, which gives a better result if the smart device is not dropped during an earthquake, and a displacement amount evaluation index, which gives a better result the smaller the displacement of the smart device during an earthquake, and evaluates the reliability of the seismic intensity class estimated by the seismic intensity class estimation unit based on the evaluation results of the placement status evaluation index, the drop presence / absence evaluation index, and the displacement amount evaluation index. Even if an acceleration waveform is obtained from a smart device as described above and corrected to estimate a seismic intensity scale, if the smart device is not placed horizontally, for example on a desk, when the acceleration waveform is obtained, if it has been dropped, or if it has moved significantly, the estimated seismic intensity scale may not be sufficiently reliable. In contrast, according to the above-mentioned configuration, the seismic intensity class reliability evaluation unit evaluates each of the placement status evaluation index, which gives a better result the more horizontally the smart device is placed during an earthquake, the drop presence / absence evaluation index, which gives a better result if the smart device does not fall during an earthquake, and the displacement amount evaluation index, which gives a better result the smaller the displacement of the smart device during an earthquake.Then, the seismic intensity class reliability evaluation unit evaluates the reliability of the seismic intensity class estimated by the seismic intensity class estimation unit based on the evaluation results of each of the placement status evaluation index, the drop presence / absence evaluation index, and the displacement amount evaluation index.Therefore, based on this evaluation result, it is possible to determine whether the estimated seismic intensity class has sufficient reliability. Effect of the Invention
[0014] According to the present invention, it is possible to provide a seismic intensity estimation system capable of estimating seismic intensity at any point with high accuracy. [Brief description of the drawings]
[0015] [Figure 1] 1 is a block diagram of a seismic intensity estimation system according to an embodiment of the present invention. [Diagram 2] FIG. 1 is an explanatory diagram of a smart device. [Diagram 3] FIG. 11 is an explanatory diagram showing an example of changes over time in magnetic flux density measured by a magnetic sensor of a smart device when an earthquake occurs. [Figure 4] FIG. 11 is an explanatory diagram showing an example of the change over time in the standard deviation of the magnetic flux density measured by a magnetic sensor in a smart device when an earthquake occurs. [Diagram 5] 11 is an explanatory diagram showing an example of an acceleration waveform and a corrected waveform obtained by correcting the acceleration waveform when an earthquake occurs and the smart device is in a sliding or bounding state. FIG. [Figure 6] FIG. 13 is an explanatory diagram showing discrimination between a bound section and a falling section based on component resultant acceleration. [Figure 7]This figure compares the component resultant acceleration based on the acceleration waveform observed by a seismometer fixed on the installation surface and the component resultant acceleration based on the acceleration waveform observed by the acceleration sensor of a smart device. [Figure 8] FIG. 11 is an explanatory diagram showing adjustment of a waveform in a bound section. [Figure 9] FIG. 13 is an explanatory diagram showing adjustment of a waveform in a falling section. [Figure 10] 2 is a flowchart of a seismic intensity estimation method according to an embodiment of the present invention. [Figure 11] FIG. 11 is a diagram showing experimental results relating to the embodiment. [Figure 12] FIG. 11 is a block diagram of a seismic intensity estimation system according to a modified example of the above embodiment. [Figure 13] 13 is a graph showing the relationship between seismic intensity and the maximum displacement of a fixed seismometer and an unfixed smart device during an earthquake. [Figure 14] 13 is a graph showing verification results regarding the above modified example. [Figure 15] 13 is a flowchart of a part for evaluating the reliability of an estimated seismic intensity scale in the seismic intensity estimation method according to the above modified example. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0016] The present invention is a seismic intensity estimation system using a smart device with an acceleration sensor and a magnetic sensor built in. In the seismic intensity estimation system, based on the measurement results by the acceleration sensor, when the smart device slides horizontally, bounds vertically, or falls, the acceleration waveform observed by the acceleration sensor is corrected, and the corrected acceleration waveform is used to estimate the seismic intensity scale. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings. Fig. 1 is an explanatory diagram of a seismic intensity estimation system according to the present embodiment. In order to increase spatial density, the seismic intensity estimation system 1 of this embodiment uses a smart device 10 to estimate the seismic intensity scale at the location where the smart device 10 is located. A large number of smart devices 10 are in circulation, owned by many people, and usually distributed at high density over a wide area. Therefore, if seismic motion can be observed with sufficient accuracy on each of them, it becomes possible to observe seismic motion at any location with high accuracy.
[0017] FIG. 2 is an explanatory diagram of a smart device. The smart device 10 is, for example, a smartphone, a tablet terminal, etc. The smart device 10 is formed in a flat plate shape and has a screen 10a on one surface thereof. As shown in FIG. 1, a smart device 10 includes an acceleration sensor 11, a magnetic sensor 12, a storage unit 13, and a transmission unit 14 built therein. The acceleration sensor 11 acquires, for example, acceleration in each of a first direction and a second direction perpendicular to each other in a plane parallel to the screen 10a of the smart device 10, and an orthogonal direction perpendicular to the screen 10a. In this embodiment, as shown in FIG. 2, the width direction of the smart device 10 is the first direction X, the length direction is the second direction Y, and the thickness direction is the orthogonal direction Z. When the smart device 10 is placed on a horizontal surface such as a desk, the surface forming the screen 10a of the smart device 10 or the surface located on the opposite side, i.e., the rear side, is often provided facing the installation surface. In this case, the screen 10a of the smart device 10 is provided horizontally. Therefore, when an earthquake force acts on the smart device 10 placed on the installation surface, the acceleration sensor 11 observes and acquires the component of the acceleration acting on the smart device 10 in a first direction X along a horizontal plane, a component in a second direction Y perpendicular to the first direction X in the horizontal plane, and a component in the perpendicular direction Z, i.e., the vertical direction. The observed acceleration waveform is stored in the memory unit 13.
[0018] Similar to the acceleration sensor 11, the magnetic sensor 12 acquires magnetic flux density in each of the first direction X, the second direction Y, and the orthogonal direction Z. The magnetic flux density is an index expressing the orientation of the smart device 10, that is, in which direction the smart device 10 is facing. When an earthquake force acts on the smart device 10 placed on the installation surface, the magnetic sensor 12 observes and acquires each of the components of the magnetic flux density in a first direction X along a horizontal plane, a component in a second direction Y perpendicular to the first direction X in the horizontal plane, and a component in the orthogonal direction Z, i.e., the vertical direction. The observed magnetic flux density is stored in the memory unit 13.
[0019] The acceleration and magnetic flux density acquired by the acceleration sensor 11 and the magnetic sensor 12 are stored in the memory unit 13 at predetermined short time intervals or continuously. If the smart device 10 does not have a function of storing the acceleration and magnetic flux density in the memory unit 13 at predetermined time intervals or continuously, an application that realizes the function may be created and added to the smart device 10. In this way, when an earthquake occurs, the smart device 10 acquires the components of the acceleration waveform, i.e., the time history waveform of acceleration, in the first direction X, the second direction Y, and the orthogonal direction Z as the observation result by the acceleration sensor 11. Furthermore, when an earthquake occurs, the smart device 10 acquires the components of the first direction X, the second direction Y, and the orthogonal direction Z of the time history waveform of the magnetic flux density as the measurement results of the magnetic sensor 12.
[0020] The storage unit 13 stores the results of observation and measurement by the acceleration sensor 11 and the magnetic sensor 12 as described above. The transmission unit 14 transmits the acceleration waveform and the magnetic flux density waveform stored as time history waveforms in the memory unit 13 to the seismic intensity estimation device 20 described below via the network 100. The network 100 is, for example, a public communication network that allows wireless or wired communication between the smart device 10 and the communication unit 23 of the seismic intensity estimation device 20.
[0021] The seismic intensity estimation device 20 includes a location input unit 21, a smart device search unit 22, a communication unit 23, a movement determination unit 24, an acceleration waveform correction unit 25, a seismic intensity scale estimation unit 26, and a memory unit 27. The point input unit 21 accepts input of a point for which the seismic intensity class is to be estimated by the seismic intensity estimation system 1. More specifically, for example, a user inputs a point for which the seismic intensity class is to be estimated using an input device (not shown). The user may specify a point for which the seismic intensity class is to be estimated by inputting latitude and longitude using, for example, a keyboard. Alternatively, the user may specify a point for which the seismic intensity class is to be estimated by selecting a desired point on a map displayed on an output device (not shown), such as a display, using a touch panel, a mouse, or the like.
[0022] The storage unit 27 stores information about all smart devices 10 that are set to be able to communicate with the seismic intensity estimation device 20. The smart device search unit 22 searches the information stored in the storage unit 27 for the smart device 10 that was located closest to the location input by the location input unit 21 when the earthquake occurred. The communication unit 23 communicates with the transmission unit 14 of the smart device 10 searched for by the smart device search unit 22, acquires the acceleration waveform and magnetic flux density waveform at the time of the occurrence of an earthquake that are stored in the memory unit 13 of the smart device 10, and stores them in the memory unit 27.
[0023] Here, when an earthquake occurs, the smart device 10 may be being carried by the user, or may be placed on an installation surface and in a stationary state. When the smart device 10 is being carried by the user, the acquired acceleration waveform contains acceleration caused by transportation other than seismic motion, and is therefore not suitable for use in estimating the seismic intensity scale. On the other hand, when the smart device 10 is placed on an installation surface and in a stationary state, it is considered that basically only seismic motion is reflected in the acceleration waveform. Therefore, the movement determination unit 24 determines whether the smart device 10 is placed on an installation surface. Specifically, the movement determination unit 24 determines whether the acceleration of the smart device 10 in the orthogonal direction Z coincides with the gravitational acceleration. If the acceleration in the orthogonal direction Z coincides with the gravitational acceleration, it is determined that the smart device 10 is placed on a horizontal plane and only seismic motion is reflected in the acceleration waveform.
[0024] When the acceleration in the orthogonal direction Z does not match the gravitational acceleration, the movement determination unit 24 determines that the smart device 10 is not placed on a horizontal plane, and therefore the acceleration waveform may reflect some acceleration other than seismic motion. In this case, the smart device search unit 22 searches for the smart device 10 located next closest to the point input by the point input unit 21 when the earthquake occurred. The communication unit 23 communicates with this smart device 10 located next closest, acquires the acceleration waveform and the magnetic flux density waveform stored in the memory unit 13 of the smart device 10 when the earthquake occurred, and stores them in the memory unit 27. Then, the movement determination unit 24 determines whether the acceleration in the orthogonal direction Z matches the gravitational acceleration with respect to this acceleration waveform. When the acceleration in the orthogonal direction Z does not match the gravitational acceleration, the smart device search unit 22 searches for the smart device 10 located next closest to the point input by the point input unit 21 when the earthquake occurred. In this way, the seismic intensity estimation device 20 acquires the acceleration waveform and the magnetic flux density waveform of the smart device 10, which was placed on a horizontal surface and in a stationary state at the time when the earthquake occurred.
[0025] For example, if an earthquake occurs while the smart device 10 is placed on an installation surface, and the earthquake is small, the smart device 10 moves along with the installation surface due to the friction between the smart device 10 and the installation surface, and does not move relative to the installation surface. In this case, the acceleration waveform observed by the acceleration sensor 11 is considered to be a correct observation of the earthquake. However, if the earthquake is large enough that a force exceeding the frictional force between the smart device 10 and the installation surface acts on the smart device 10, the smart device 10 may move relative to the installation surface, such as sliding on the installation surface, bouncing up and down on the installation surface, or dropping from the installation surface. In this case, as will be described later, the acceleration waveform observed by the acceleration sensor 11 may not be an accurate observation of the earthquake. In contrast, the seismic intensity estimation system 1 of the present embodiment determines whether the smart device 10 has moved relatively from the installation surface, and if it is determined that the smart device 10 has moved, adjusts and corrects the acceleration waveform during that time, and estimates the seismic intensity scale based on the result. In order to perform such processing, the seismic intensity estimation system 1 needs to accurately determine whether the smart device 10 has moved relatively to the installation surface on the installation surface.
[0026] Basically, some force, including seismic force, acts on the smart device 10 throughout the entire time an earthquake occurs. Therefore, the acceleration data observed by the acceleration sensor 11 has a time history waveform in which the acceleration changes in a complex manner, even if the smart device 10 is placed on the installation surface. Based on such data whose values change in a complex manner over time, it is not easy to accurately determine, for example, at what time the smart device 10 starts to move relative to the installation surface. Thus, when an earthquake force is applied, the acceleration changes in a complex manner. On the other hand, the attitude of the smart device 10 does not change basically when the smart device 10 is not moving relative to the installation surface, and changes significantly only when the smart device 10 starts to move relative to the installation surface. Therefore, in the seismic intensity estimation system 1 of this embodiment, the change in the attitude of the smart device 10 is observed to determine whether the smart device 10 is moving relative to the installation surface on the installation surface. In this embodiment, the attitude of the smart device 10 is determined based on the magnetic flux density measured by the magnetic sensor 12.
[0027] FIG. 3 is an explanatory diagram showing an example of the change over time in magnetic flux density measured by a magnetic sensor in a smart device when an earthquake occurs. For example, assume that the magnetic sensor 12 obtains a waveform as shown in Fig. 3 as the component of the first direction X of the time history waveform of the magnetic flux density. In Fig. 3, time T1 is the time when the earthquake occurs, and time T2 is the time when the earthquake ends. In such a case, the magnetic flux density is almost constant until time T1, starts to change significantly after time T1, ends this change a little before time T2, and then becomes almost constant again. In this case, the magnetic flux density continues to change significantly from time T1 to time T2, and the attitude of the smart device 10 continues to change during this period, and therefore, it is considered that the smart device 10 is moving relative to the installation surface. In this way, if the posture of the smart device 10 changes, the value of the magnetic flux density also changes accordingly. Therefore, based on the time history waveform of the magnetic flux density, it is possible to determine whether the smart device 10 has moved relative to the installation surface from a stationary state on the installation surface due to the earthquake, and if it is determined that the smart device 10 has moved, it is possible to identify the movement section, which is the time period during which the smart device moved.
[0028] 3, the magnetic flux density measured by the magnetic sensor 12 changes slightly even when no earthquake has occurred. For this reason, it is not preferable to determine whether the smart device 10 has moved by simply comparing the magnetic flux density values before and after an earthquake. In this embodiment, the average magnetic flux density is calculated before and after an earthquake, and the changes in this average value are compared. Specifically, the average value of the magnetic flux density before the earthquake occurred is taken as B for the component of the first direction X of the time history waveform of the magnetic flux density. norm The average magnetic flux density after the earthquake is B endIn this case, the movement determination unit 24 calculates the amount of change ΔB in the average value of the magnetic flux density before and after the earthquake using the following formula (1).
number
[0029] The movement determination unit 24 also calculates the amount of change ΔB in the average value of the magnetic flux density before and after the earthquake in the second direction Y of the time history waveform of the magnetic flux density, using the above formula (1). Then, the movement determination unit 24 compares the amount of change ΔB in the first direction X and the second direction Y with a movement determination threshold, respectively. When the amount of change ΔB is equal to or greater than the movement determination threshold in at least one of the first direction X and the second direction Y, the movement determination unit 24 determines that the smart device 10 has moved in the horizontal direction. The movement determination unit 24 also calculates the amount of change ΔB in the average value of the magnetic flux density before and after the earthquake in the orthogonal direction Z of the time history waveform of the magnetic flux density, using the above formula (1). Then, the movement determination unit 24 compares the amount of change ΔB in the orthogonal direction Z with the movement determination threshold. When the amount of change ΔB in the orthogonal direction Z is equal to or greater than the movement determination threshold, the movement determination unit 24 determines that the smart device 10 has moved in the up-down direction.
[0030] In this way, the movement determination unit 24 determines that the smart device has moved based on the measurement results by the magnetic sensor 12, more specifically, when the change in the average value of the magnetic flux density before and after the earthquake in the time history waveform of the magnetic flux density measured by the magnetic sensor 12 is greater than or equal to the movement determination threshold. The movement determination unit 24 determines whether the smart device 10 has moved in the horizontal direction based on the measurement results of the components in each of the first direction X and the second direction Y by the magnetic sensor 12. The movement determination unit 24 determines whether the smart device 10 has moved in the up-down direction based on the measurement result of the component in the orthogonal direction Z by the magnetic sensor 12.
[0031] Next, if it is determined that the smart device 10 has moved as a result of determining whether the smart device 10 has moved, the movement determination unit 24 specifies a movement section which is a time section during which the smart device 10 has moved. 3, the magnetic flux density value changes and varies significantly while the smart device 10 is moving. Therefore, in this embodiment, the degree of variation in the magnetic flux density value is observed by calculating the standard deviation, and the movement section is identified based on the standard deviation. Figure 4 is an explanatory diagram showing an example of the change over time in the standard deviation of the magnetic flux density measured by the magnetic sensor of a smart device when an earthquake occurs. In Figure 4, the standard deviation of the magnetic flux density is plotted as a graph below the time history waveform of the magnetic flux density shown in Figure 3.
[0032] When the movement determination unit 24 determines that the smart device 10 has moved in the horizontal direction, the movement determination unit 24 advances time in the time history waveform of the magnetic flux density and calculates the standard deviation within a predetermined time interval starting from that time as a base point. In this embodiment, the movement determination unit 24 calculates, for each time, a component of the first direction X of the time history waveform of the magnetic flux density, the time t being the base point, and a predetermined time interval W B More precisely, during a certain time interval W from the time t B The time tW just before B A given time interval W from the time t to the time t B For the magnetic flux density data within, the standard deviation σB is calculated using the following formula (2). t Calculate.
number
number
[0033] The movement determination unit 24 calculates the standard deviation σB of the magnetic flux density at each time point for the component in the first direction X of the time history waveform of the magnetic flux density by the above formula (2). t Calculate the moving section determination threshold τ σB The movement determination unit 24 compares the standard deviation σB t is the moving section determination threshold τ σB The time when it becomes equal to or greater than the first time T σs n , standard deviation σB t The time when becomes smaller than the moving section determination threshold is called the second time T σe n As a result of a single earthquake, the smart device 10 may repeatedly move and stop. In this case, the standard deviation σB t The moving section determination threshold τ σB As a result of the comparison with the first time T σs n , second time T σe n In the above, the subscript n is a serial number assigned to the multiple sections. For example, when an earthquake occurs, the time of the first movement of the smart device 10 is the first time T σs 1 , second time T σe 1 In addition, for the movement section in which the smart device 10 moves for the second time, these times are expressed as the first time T σs 2 , second time T σe 2 It is expressed as:
[0034] In the time history waveform of the magnetic flux density, when the standard deviation within a predetermined time interval is calculated with the time as a base point while advancing the time, the standard deviation is determined to be within a moving section determination threshold value τ σBAfter the state continues to be smaller than the threshold value τ σB If this is the case, the specified time interval W B The time advances from a state in which the entire area of the time history waveform is in a region where the variation in magnetic flux density is small, and a predetermined time interval W B Among them, the time interval W B This is considered to be the case where the most advanced part of the time history waveform, closest to the end time of the test, is approaching the area where the variation in magnetic flux density begins to increase, and the standard deviation begins to increase. B By identifying the end time of the travel section, the start time of the travel section can be accurately identified. In this embodiment, the time interval W B Since time t is used as the base point and the period from that time t to the previous time is set, the time interval W B The end time of the period coincides with the base time t. Here, the standard deviation σB t is the moving section determination threshold τ σB The time when this occurs is the first time T σs n Therefore, in this case, the time interval W B The end time of the first time T σs n Therefore, the movement determination unit 24 determines whether the first time T σs n The start time T of the movement section in the first direction X Bs n and set as shown in the following equation (4).
number
[0035] In addition, once the standard deviation is σB After that, at a certain time, the moving section determination threshold τ σB If the time interval W BThe time advances from a state in which the entire area of the time history waveform is in a region where the magnetic flux density variation is large, and a predetermined time interval W B Among them, the time interval W B The time interval W that contains the part of the time that is closest to the start of B It is considered that the standard deviation is small because the whole of the time history waveform has reached a region where the variation in magnetic flux density is small. Therefore, the time interval W B By identifying the start time of the travel section, the end time of the travel section can be accurately identified. In this embodiment, the time interval W B Since time t is used as the base point and the period from that time t to the previous time is set, the time interval W B The start time of the time interval W is a given time interval from the base time t. B Here, the standard deviation σB t is the moving section determination threshold τ σB The time when it becomes smaller is the second time T σe n Therefore, in this case, the time interval W B The start time of the second time T σe n From a given time interval W B Therefore, the movement determination unit 24 determines the second time T σe 1 From a given time interval W B The previous time T σe n -W B The end time T of the movement section in the first direction X Be n and set as shown in the following equation (5).
number
[0036] In the above, for a certain time, a predetermined time interval starting from the time t is defined as a predetermined time interval W BHowever, the time interval between the time t and the time interval W B The time interval may be set as a time interval just after (in the future). In this case, the start time T Bs n is the first time T σs n For a given time interval W B The time is the sum of the above. In this case, the end time T Be n is the second time T σe 1 The time will match.
[0037] If it is determined that the smart device 10 has moved horizontally, the movement determination unit 24 calculates the start time and end time of the second direction Y of the time history waveform of the magnetic flux density in the above manner. Then, the movement determination unit 24 identifies an overlapping time section between the start time and the end time in each of the first direction X and the second direction Y as a horizontal movement section, which is a movement section in the horizontal direction. In addition, if it is determined that the smart device 10 has moved in the vertical direction, the movement determination unit 24 calculates the start time and end time of the orthogonal direction Z of the time history waveform of the magnetic flux density in the orthogonal direction Z as described above. Then, the movement determination unit 24 specifies the time section between the start time and the end time in the orthogonal direction Z as a vertical movement section, which is a movement section in the vertical direction.
[0038] In this way, when the movement determination unit 24 determines that the smart device 10 has moved, the movement determination unit 24 advances time in the time history waveform of the magnetic flux density and calculates a predetermined time interval W B Standard deviation within σB t Calculate the standard deviation σB t is the moving section determination threshold τ σB If this is the case, the time interval W B End time Tσs n Let T be the start time of the travel section. Bs n and the standard deviation σB t is the moving section determination threshold τ σB If it is smaller than the time, the time interval W B Start time T σe n -W B The end time of the travel section T Be n Set as. Furthermore, when the movement determination unit 24 determines that the smart device 10 has moved in the horizontal direction, it specifies a horizontal movement section that is a movement section in the horizontal direction. Furthermore, when the movement determination unit 24 determines that the smart device 10 has moved in the up-down direction, it specifies an up-down movement section that is a movement section in the up-down direction.
[0039] When the movement determination unit 24 determines that the smart device 10 has moved, the acceleration waveform correction unit 25 determines in what manner the smart device 10 has moved. When an earthquake occurs, the behavior of the smart device 10 placed on the installation surface is considered to be roughly in four types of states: a stationary state, a sliding state, a bounding state, and a falling state. The stationary state is a state in which the frictional force between the smart device 10 and the installation surface is greater than the force acting due to the earthquake, and the smart device 10 is stationary relative to the installation surface, vibrates in accordance with the installation surface, and does not move relative to the installation surface. The sliding state is a state in which an earthquake causes a horizontal force acting on smart device 10 that is greater than the static friction force between smart device 10 and the installation surface, causing smart device 10 to slide against the installation surface, thereby moving relative to the installation surface. The bouncing state occurs when a force equal to or greater than the acceleration of gravity acts on the installation surface in the vertical direction, causing the smart device 10 to bounce up and down. In this case, the smart device 10 often also moves in the horizontal direction. The dropped state is a state in which the device falls downward from the installation surface and then rolls widely on the floor surface below.
[0040] Among the above states, the posture of the smart device 10 does not change in a stationary state, but changes in a sliding state, a bouncing state, and a falling state. Therefore, the movement determination of the smart device 10 by the movement determination unit 24 determines whether the smart device 10 is in a stationary state, a sliding state, a bouncing state, or a falling state. The acceleration waveform correction unit 25 corrects and adjusts the acceleration waveform observed by the acceleration sensor 11 depending on whether the smart device 10 is in a sliding state, a bounding state, or a falling state.
[0041] First, when the movement determination unit 24 determines that the smart device 10 has moved in the horizontal direction and a horizontal movement section is identified, the acceleration waveform correction unit 25 corrects the components in the first direction X and the second direction Y of the acceleration waveform observed by the acceleration sensor 11 in the horizontal movement section. FIG. 5 is an explanatory diagram showing an example of an acceleration waveform and a corrected waveform obtained by correcting the acceleration waveform when an earthquake occurs and the smart device is in a sliding or bounding state. When the movement determination unit 24 determines that the smart device 10 has moved horizontally and a horizontal movement section is identified, it is considered that there is a high possibility that the smart device 10 is in a sliding or bouncing state in the horizontal movement section. In the sliding state, no force greater than the static friction force acts in the horizontal direction on the smart device 10. Therefore, as shown in Fig. 5, in the component W1 in each of the first direction X and the second direction Y of the acceleration waveform, the actual acceleration waveform WR is not observed, and the amplitude becomes constant, resulting in a plateau portion WP, in which the acceleration data is missing. Similarly, in a bound state, a plateau portion WP may occur in each of the components W1 in the first direction X and the second direction Y of the acceleration waveform.
[0042] In such a case, the acceleration waveform correction unit 25 complements the portions WP of the acceleration waveform in which the amplitude of the component W1 in each of the first direction X and the second direction Y is constant by curve fitting using a sine wave, thereby generating corrected acceleration waveforms WC in each of the first direction X and the second direction Y. First, the acceleration waveform correction unit 25 calculates the acceleration A sat The acceleration waveform correction unit 25 calculates the acceleration A sat is calculated, for example, by multiplying the static friction coefficient and the gravitational acceleration. When the smart device 10 is in a sliding state, the amplitude is constant and the amplitude when the smart device 10 hits a plateau is the acceleration A calculated as above. sat On the other hand, when the smart device 10 is in a bounding state, the impact of the smart device 10 bouncing up and down and colliding with the installation surface may appear not only in the orthogonal direction Z component of the acceleration waveform, but also in the first direction X and second direction Y components. For this reason, the peaking acceleration A sat becomes unclear, and the acceleration reaches a plateau A sat The above values are recorded in large quantities in the acceleration waveform. Therefore, the components of the acceleration waveform in the first direction X and the second direction Y contain the peak acceleration A. sat If the above value is observed, in this embodiment, it is considered that this value is unreliable because it may have been recorded during a bounding state, and the head-on acceleration A sat The above values are calculated by applying a constant acceleration A sat The waveform is corrected assuming that it has reached a plateau. The acceleration waveform correction unit 25 uses two pieces of acceleration data, one before and one after the portion where the acceleration peaks out, as shown by point P in Fig. 5, and performs a curve fitting process (curve fitting) using a sine wave to generate a corrected acceleration waveform WC. For example, the nonlinear least squares method can be used for the curve fitting.
[0043] In this manner, the acceleration waveform correction unit 25 corrects the portion of the acceleration waveform observed by the acceleration sensor 11 in each of the components W1 in the first direction X and the second direction Y in the horizontal movement section identified by the movement determination unit 24, and combines the corrected waveforms of the components W1 in the first direction X and the second direction Y with the component in the orthogonal direction Z of the acceleration waveform observed by the acceleration sensor 11 to obtain a corrected acceleration waveform. In a case where the components W1 in the first direction X and the second direction Y of the acceleration waveform observed by the acceleration sensor 11 are not corrected, in this embodiment, for the sake of simplicity of the following explanation, the components in the first direction X, the second direction Y, and the orthogonal direction Z of the acceleration waveform observed by the acceleration sensor 11 will be collectively referred to as a corrected acceleration waveform. When the smart device 10 is in a bounding state or a falling state, some acceleration is observed as the orthogonal direction Z component of the acceleration waveform observed by the acceleration sensor 11. In this case, this acceleration value may be a factor that may reduce the estimation accuracy when the seismic intensity scale is estimated later by the seismic intensity scale estimation unit 26. The adjustment of the orthogonal direction Z component in the bounding state or the falling state will be described later.
[0044] Next, as a preparation for calculating the seismic intensity, the acceleration waveform correction unit 25 performs a Fourier transform on each of the components of the corrected acceleration waveform in the first direction X, the second direction Y, and the orthogonal direction Z. The acceleration waveform correction unit 25 applies a filter to correct the effect of the period of the seismic wave. For example, three types of filters can be applied: a low-cut filter, a high-cut filter, and a filter that represents the effect of the period. The low-cut filter is expressed, for example, by the following equation (6).
number
number
number
[0045] Next, when the movement determination unit 24 determines that the smart device 10 has moved in the vertical direction and identifies an up-down movement section, the acceleration waveform correction unit 25 respectively identifies a section within the up-down movement section in which the smart device 10 was in a bouncing state and a section in which it was in a falling state. As already explained, in the bounding state, the smart device 10 bounces up and down in small movements, floating in the air, and behaving as if it is about to collide with the installation surface due to free fall. In the air, no downward force acts on the smart device 10 other than the acceleration of gravity. Therefore, the acceleration waveform in the orthogonal direction Z tends to peak out at about 1 G in the downward direction and not peak out in the upward direction. In addition, in a falling state, the smart device 10 falls downward from the installation surface and rolls widely on the floor surface below after falling, so that large accelerations tend to be recorded in all components of the acceleration waveform in the first direction X, the second direction Y, and the orthogonal direction Z.
[0046] For this reason, in this embodiment, the components in the first direction, second direction, and orthogonal direction of the waveform based on the corrected acceleration waveform, which has been subjected to Fourier transform, filtering, and inverse Fourier transform as described above, are vectorially combined at each time to calculate a component resultant acceleration, and based on this, it is determined whether each time in the vertical movement section corresponds to a bounding state or a falling state. FIG. 6 is an explanatory diagram showing the discrimination between the bounding section and the falling section based on the component resultant acceleration. In Fig. 6, the component resultant acceleration calculated as described above is shown as a time history graph. In this waveform, a section identified as a vertical movement section by the movement determination unit 24 is indicated as a section UDA. In a section SA that is not a vertical movement section UDA even though the component resultant acceleration is a value other than 0, the smart device 10 is not moving in the vertical direction and is in a sliding state. In this way, the acceleration waveform correction unit 25 determines that the smart device 10 is in a sliding state in a section that is not included in the up-down movement section UDA among the sections identified by the movement determination unit 24 as the movement section through which the smart device 10 has moved. In such a section SA, the seismic intensity scale estimation unit 26, which will be described later, estimates the seismic intensity scale based on the component composite acceleration already calculated for the waveform based on the corrected acceleration waveform.
[0047] In this embodiment, the acceleration waveform correction unit 25 judges whether each time in the vertical movement section UDA is in a bound state or a falling state based on the magnitude of the component resultant acceleration. As described above, in a falling state, a large acceleration tends to be recorded. Therefore, if the component resultant acceleration at each time in the vertical movement section UDA is smaller than the fall determination threshold, the acceleration waveform correction unit 25 considers the smart device 10 to be in a bound state at that time. Also, if the component resultant acceleration at each time in the vertical movement section UDA is equal to or greater than the fall determination threshold, the acceleration waveform correction unit 25 considers the smart device 10 to be in a falling state at that time. In FIG. 6, a section in which the smart device 10 is determined to be in a bouncing state is indicated as section BA, and a section in which the smart device 10 is determined to be in a falling state is indicated as section FA.
[0048] Figure 7 shows a comparison of the component resultant acceleration based on the acceleration waveform observed by a seismometer fixed on the installation surface and the component resultant acceleration based on the acceleration waveform observed by the acceleration sensor of a smart device. Figure 7 was obtained as a result of a shaking table experiment. In FIG. 7, when the component resultant acceleration is smaller than 2G, the component resultant acceleration of the acceleration waveform observed by the acceleration sensor 11 of the smart device 10 shows a value close to the component resultant acceleration of the acceleration waveform observed by a seismometer fixed on the installation surface. On the other hand, when the component resultant acceleration is larger than 2G, the component resultant acceleration of the acceleration waveform observed by the acceleration sensor 11 of the smart device 10 is significantly different from the value of the component resultant acceleration of the acceleration waveform observed by a seismometer fixed on the installation surface. Here, when the smart device 10 is in a bound state, it is considered that the smart device 10 behaves more like a seismometer fixed on the installation surface, and is closer to the state where the smart device 10 is placed on the installation surface, than when it is in a falling state. Therefore, the case where the component resultant acceleration is 2G or less, which shows a behavior close to the state where the smart device 10 is placed on the installation surface, is considered to correspond to the bound state. For this reason, in this embodiment, a value of 2G is used as the fall determination threshold.
[0049] As described above, if the component composite acceleration is smaller than the fall determination threshold at a time within the vertical movement section, the acceleration waveform correction unit 25 considers that the smart device 10 is in a bounding state at that time. In this case, as already described, the first direction X and second direction Y components of the acceleration waveform are subjected to a curve fitting process using a sine wave to generate an appropriate corrected acceleration waveform. However, in the bounding state, the smart device 10 bounces up and down and collides with the installation surface, so that the impact is reflected in the orthogonal direction Z component of the acceleration waveform as a large acceleration that is not observed in the sliding state, and as a result, a large acceleration is also observed in the orthogonal direction Z component of the corrected acceleration waveform. Such acceleration data reflecting factors other than earthquakes can be a factor that can reduce the estimation accuracy when the seismic intensity scale is estimated later by the seismic intensity scale estimation unit 26. 8, in this embodiment, when the acceleration waveform correction unit 25 regards the smart device 10 as being in a bounding state, it recalculates the component resultant acceleration at that time by setting the acceleration value of the orthogonal direction Z component of the waveform based on the corrected acceleration waveform to 0. That is, it ignores the component value of the acceleration in the orthogonal direction Z in the bounding state, and calculates the component resultant acceleration by regarding the smart device 10 as being in a sliding state rather than a bounding state.
[0050] In addition, if the component composite acceleration is equal to or greater than the fall determination threshold at a time within the vertical movement section, the acceleration waveform correction unit 25 considers the smart device 10 to be in a falling state at that time. In a situation in which the smart device 10 is suspended in the air for a long time, such as in a falling state, the smart device 10 is in a state in which external forces due to earthquakes do not act on the smart device 10. In addition, since the smart device 10 rolls significantly on the floor surface below after falling, large accelerations tend to be recorded in all components of the acceleration waveform in the first direction X, the second direction Y, and the perpendicular direction Z, but it is difficult to think that such accelerations are properly observed external forces due to earthquakes. Such acceleration data reflecting factors other than earthquakes can be a factor that can reduce the estimation accuracy when the seismic intensity scale is estimated later by the seismic intensity scale estimation unit 26. 9, in this embodiment, when the acceleration waveform correction unit 25 considers that the smart device 10 is in a falling state, it recalculates the component resultant acceleration at that time by setting the acceleration values in the first direction X, the second direction Y, and the orthogonal direction Z components of the waveform based on the corrected acceleration waveform to 0. Based on the idea that the acceleration of any component is not useful in a falling state, the component resultant acceleration is set to 0, and the acceleration data in the falling state is not taken into consideration when estimating the seismic intensity scale.
[0051] The seismic intensity class estimation unit 26 estimates the seismic intensity class based on the corrected acceleration waveform, more accurately, based on a waveform based on the corrected acceleration waveform obtained by applying Fourier transform, filtering, and inverse Fourier transform to the corrected acceleration waveform. In this embodiment, the seismic intensity class estimation unit 26 estimates the seismic intensity class based on the component composite acceleration at each time point calculated as described above based on the corrected acceleration waveform. Specifically, the seismic intensity class estimation unit 26 calculates a value a so that when the total time during which the absolute value of the component resultant acceleration calculated as above is equal to or greater than a certain value a, is calculated, this value a is exactly 0.3 seconds. Then, the seismic intensity class estimation unit 26 substitutes the value a calculated as described above into the following equation (9), rounds the obtained value to two decimal places, and discards the first decimal place to calculate the measured seismic intensity I. I = 2log(a) + 0.94 … (9)
[0052] The seismic intensity scale estimating unit 26 further calculates a seismic intensity scale based on the measured seismic intensity. The seismic intensity scale estimating unit 26 sets the seismic intensity scale to 0 when the measured seismic intensity is less than 0.5, to 1 when the measured seismic intensity is 0.5 or more and less than 1.5, to 2 when the measured seismic intensity is 1.5 or more and less than 2.5, to 3 when the measured seismic intensity is 2.5 or more and less than 3.5, to 4 when the measured seismic intensity is 3.5 or more and less than 4.5, to 5 when the measured seismic intensity is 4.5 or more and less than 5.5, to 6 when the measured seismic intensity is 5.5 or more and less than 6.5, and to 7 when the measured seismic intensity is 6.5 or more. The seismic intensity scale estimation unit 26 displays the estimated seismic intensity scale on an output unit such as a display (not shown).
[0053] Next, a method for estimating seismic intensity using the above-mentioned seismic intensity estimation system 1 will be described with reference to Figures 1 to 9 and 10. Figure 10 is a flowchart of the method for estimating seismic intensity. For example, when an earthquake occurs, a user inputs a location for which the seismic intensity scale is to be estimated, using an input device (not shown) (step S1). The smart device search unit 22 searches the information stored in the storage unit 27 for the smart device 10 that was located closest to the point input by the point input unit 21 when the earthquake occurred (step S2). The communication unit 23 communicates with the transmission unit 14 of the smart device 10 searched for by the smart device search unit 22, acquires the acceleration waveform and magnetic flux density waveform at the time of the earthquake stored in the memory unit 13 of the smart device 10, and stores them in the memory unit 27 (step S3).
[0054] Based on the measurement results by the magnetic sensor 12, more specifically, when the change in the average value of the magnetic flux density before and after the earthquake in the time history waveform of the magnetic flux density measured by the magnetic sensor 12 is equal to or greater than the movement determination threshold, the movement determination unit 24 determines that the smart device has moved (step S4). When the movement determination unit 24 determines that the smart device 10 has moved in the horizontal direction, it specifies a horizontal movement section that is a movement section in the horizontal direction. When the movement determination unit 24 determines that the smart device 10 has moved in the up-down direction, it specifies an up-down movement section that is a movement section in the up-down direction.
[0055] The acceleration waveform correction unit 25 determines whether or not the movement determination unit 24 has determined that the smart device 10 has moved in the horizontal direction (step S5). When it is determined that movement has occurred in the horizontal direction and a horizontal movement section has been identified (Yes in step S5), the acceleration waveform correction unit 25 corrects the components in the first direction X and the second direction Y of the acceleration waveform observed by the acceleration sensor 11 in the horizontal movement section (step S6). The process then proceeds to step S7. If it is not determined that the object has moved in the horizontal direction (No in step S5), the acceleration waveform correction unit 25 transitions to step S7.
[0056] The acceleration waveform correction unit 25 performs a Fourier transform on each of the components of the corrected acceleration waveform in the first direction X, the second direction Y, and the orthogonal direction Z. The acceleration waveform correction unit 25 applies a filter to correct the influence of the period of the seismic wave. The acceleration waveform correction unit 25 applies an inverse Fourier transform to the result of applying the above filter, and generates a waveform based on the corrected acceleration waveform (step S7). Then, the acceleration waveform correction unit 25 determines whether or not the movement determination unit 24 has determined that the smart device 10 has moved in the up or down direction (step S8). When it is determined that movement has occurred in the vertical direction and the vertical movement section has been identified (Yes in step S8), the acceleration waveform correction unit 25 determines whether the component composite acceleration is greater than or equal to the fall determination threshold for each time (step S9). If the component resultant acceleration is equal to or greater than the fall determination threshold (Yes in step S9), the smart device 10 is deemed to be falling at that time, and the acceleration waveform correction unit 25 sets the acceleration values of the components in the first direction X, the second direction Y, and the orthogonal direction Z of the waveform based on the corrected acceleration waveform to 0, and recalculates the component resultant acceleration at that time (step S10). The process then proceeds to step S12. If the component resultant acceleration is smaller than the fall determination threshold (No in step S9), the smart device 10 is deemed to be in a bound state at that time, and the acceleration waveform correction unit 25 sets the acceleration value of the orthogonal direction Z component of the waveform based on the corrected acceleration waveform to 0, and recalculates the component resultant acceleration at that time (step S11). The process then proceeds to step S12. If it is determined that the acceleration waveform correcting section 25 has not determined that the object has moved in the up or down direction (No in step S8), the process proceeds to step S12.
[0057] The seismic intensity scale estimation unit 26 calculates the value a for the component resultant acceleration calculated as above, and calculates the measured seismic intensity (step S12). The seismic intensity scale estimating unit 26 calculates the seismic intensity scale based on the measured seismic intensity (step S13). The seismic intensity scale estimation unit 26 displays the estimated seismic intensity scale on an output unit such as a display (not shown).
[0058] Next, the effects of the above-described seismic intensity estimation system 1 will be described. The seismic intensity estimation system 1 as described above is a seismic intensity estimation system 1 using a smart device 10 incorporating an acceleration sensor 11 and a magnetic sensor 12, and is equipped with a movement determination unit 24 that determines whether or not the smart device 10 has moved from a stationary state due to an earthquake based on the measurement results by the magnetic sensor 12, and if it is determined that the smart device 10 has moved, identifies a movement section, which is the time section during which the smart device 10 has moved; an acceleration waveform correction unit 25 that, if it is determined that the smart device 10 has moved, corrects the portion of the acceleration waveform observed by the acceleration sensor 11 that corresponds to the movement section to generate a corrected acceleration waveform; and a seismic intensity class estimation unit 26 that estimates a seismic intensity class based on the corrected acceleration waveform. A smart device 10 that is placed on a desk or the like and is stationary may slide relative to the installation surface if the earthquake is large enough that a force exceeding the static friction force between the smart device 10 and the installation surface on which the smart device 10 is installed acts on the smart device 10. Alternatively, the smart device 10 may bounce up and down on the installation surface, or may slide on the installation surface and then fall from the installation surface. In this way, when the smart device 10 moves from a stationary state due to an earthquake, the direction in which the smart device 10 is placed, i.e., its posture, is changed at least a little. When smart device 10 exhibits such behavior, the behavior of smart device 10 does not follow the behavior of the surface on which it is installed, and therefore the acceleration waveform observed by the acceleration sensor installed in smart device 10 does not normally observe an earthquake. Here, in the above configuration, the measurement results by the magnetic sensor 12 provided in the smart device 10 are used. If the smart device 10 moves from a stationary state due to an earthquake and its posture changes, the behavior is reflected in the measurement results by the magnetic sensor 12. Therefore, using the measurement results by the magnetic sensor 12 provided in the smart device 10, it is possible to determine whether the smart device 10 has moved and to identify the movement section, which is the time section during which the smart device 10 has moved. The movement section identified in this manner is a section in which the behavior of the smart device 10 does not follow the behavior of the installation surface, in other words, a time section in which an earthquake was not normally observed by the acceleration sensor 11. In the above configuration, the portion of the acceleration in the movement section is corrected to generate a corrected acceleration waveform, and the seismic intensity scale is estimated based on this. In this manner, the acceleration waveform of the time section in which an earthquake was not normally observed is appropriately corrected, so that the seismic intensity scale at the location where the smart device 10 is installed can be more accurately estimated. Many people own smart devices 10. Therefore, there is a high possibility that a person who owns a smart device 10 is located at or near a location where seismic intensity estimation is required. Therefore, when a certain location is specified, it is possible to provide a seismic intensity estimation system 1 that can accurately estimate the seismic intensity at any location by using the measurement results of the smart device 10 located at the specified location.
[0059] In particular, in the seismic intensity estimation system 1 as described above, the results of observation by other seismometers such as fixed seismometers are not required when correcting the acceleration waveform observed by the smart device 10. Therefore, when realizing the seismic intensity estimation system 1, the configuration is simple, and therefore it is possible to reduce costs.
[0060] In addition, the movement determination unit 24 determines that the smart device 10 has moved if the change in the average value of the magnetic flux density before and after an earthquake in the time history waveform of the magnetic flux density, which is the measurement result by the magnetic sensor 12, is equal to or greater than the movement determination threshold.When it determines that the smart device 10 has moved, it calculates the standard deviation within a predetermined time interval using that time as a base point while advancing time in the time history waveform of the magnetic flux density, and when the standard deviation is equal to or greater than the movement section determination threshold, it sets the end time of the time interval using that time as a base point as the start time of the movement section, and when the standard deviation is smaller than the movement section determination threshold, it sets the start time of the time interval using that time as a base point as the end time of the movement section. According to the above configuration, the change in the average value of the magnetic flux density before and after the earthquake is calculated in the time history waveform of the magnetic flux density, which is the measurement result by the magnetic sensor 12. If this change is large, it means that the attitude of the smart device 10 has changed significantly before and after the earthquake. Therefore, by comparing the change in the average value of the magnetic flux density before and after the earthquake with a movement determination threshold and determining that the smart device 10 has moved if the change is equal to or greater than the movement determination threshold, it is possible to appropriately determine whether the smart device 10 has moved. If it is determined that the smart device 10 has moved, the movement section is further specified. Since the value of the magnetic flux density varies greatly while the smart device 10 is moving, the movement section can be accurately specified by calculating the standard deviation within a predetermined time interval in the time history waveform of the magnetic flux density and comparing it with a threshold value. More specifically, the standard deviation is calculated within a predetermined time interval with the time as a base point while advancing the time in the time history waveform of the magnetic flux density. If the standard deviation becomes equal to or greater than the moving section determination threshold at a certain time after the standard deviation has remained smaller than the moving section determination threshold, it is considered that the time has advanced from a state in which the entirety of the predetermined time interval was in a region of the time history waveform where the variation in the magnetic flux density was small, and the most advanced part of the predetermined time interval with the time as the base point has entered a region of the time history waveform where the variation in the magnetic flux density began to increase, and the standard deviation has begun to increase. Therefore, by identifying the end time of the time interval with the time as the base point, the start time of the moving section can be accurately identified. Furthermore, if the standard deviation becomes equal to or greater than the movement section determination threshold at a certain time, it is considered that the standard deviation becomes smaller as time advances from a state in which the entirety of a given time interval was in a region of the time history waveform where the magnetic flux density variation is large, and the entirety of the given time interval starting from that time, including the part that is the latest in time and closest to the start time of the time interval, reaches a region of the time history waveform where the magnetic flux density variation is small. Therefore, by identifying the start time of the time interval starting from that time, the end time of the movement section can be accurately identified. In this way, it is possible to properly determine whether the smart device 10 has moved and to properly calculate the section of movement to be corrected for the acceleration waveform observed by the acceleration sensor 11, thereby properly correcting the acceleration waveform and, as a result, improving the accuracy of the seismic intensity calculation.
[0061] In addition, the movement determination unit 24 determines whether the smart device 10 has moved horizontally based on the measurement results of the components in each of the first direction X and the second direction Y that are perpendicular to each other in a plane parallel to the screen 10a of the smart device 10 by the magnetic sensor 12, and if it is determined that the smart device 10 has moved horizontally, it identifies a horizontal movement section that is a movement section in the horizontal direction, and determines whether the smart device 10 has moved up or down based on the measurement results of the components in the orthogonal direction Z that is perpendicular to the screen 10a by the magnetic sensor 12, and if it is determined that the smart device 10 has moved up or down, it identifies a vertical movement section that is a movement section in the vertical direction. When an earthquake occurs and the smart device 10 placed on the installation surface slides on the installation surface, the posture of the smart device 10 changes more or less. This causes a change in the measurement results of the components in the first direction X and the second direction Y that are perpendicular to each other in a plane parallel to the screen 10a of the smart device 10 by the magnetic sensor 12. In the above configuration, whether or not the smart device 10 has moved in the horizontal direction can be determined based on the measurement results of the components in the first direction X and the second direction Y by the magnetic sensor 12, so that the determination can be made appropriately. Furthermore, if an earthquake occurs and the smart device 10 placed on the installation surface bounces up and down on the surface or falls off the surface, a change appears in the measurement result of the component in the orthogonal direction Z perpendicular to the screen 10a by the magnetic sensor 12. In the above configuration, whether or not the smart device 10 has moved in the up and down direction can be determined appropriately based on the measurement result of the component in the orthogonal direction Z by the magnetic sensor 12.
[0062] Furthermore, when it is determined that the smart device 10 has moved in the horizontal direction, the acceleration waveform correction unit 25 complements the parts of the acceleration waveform in which the amplitudes of the components in each of the first direction X and the second direction Y are constant in the horizontal movement section by curve fitting processing using a sine wave to generate corrected acceleration waveforms in the first direction X and the second direction Y. The acceleration waveform correction unit 25 calculates a composite component acceleration by combining the components in each of the first direction X, the second direction Y, and the orthogonal direction Z of the waveform based on the corrected acceleration waveform at each time, and calculates a composite component acceleration when the time is within the vertical movement section and the composite component acceleration is greater than or equal to a fall determination threshold. If the above is true, the smart device 10 is deemed to be falling at that time, and the components of the waveform based on the corrected acceleration waveform in the first direction X, the second direction Y, and the orthogonal direction Z are set to 0 to recalculate the component composite acceleration. If the time is within the up-and-down movement section and the component composite acceleration is smaller than the fall determination threshold, the smart device 10 is deemed to be bouncing up and down at that time, and the component of the waveform based on the corrected acceleration waveform in the orthogonal direction Z is set to 0 to recalculate the component composite acceleration. The seismic intensity class estimation unit 26 estimates the seismic intensity class based on the component composite acceleration. When an earthquake occurs while the smart device 10 is placed on the installation surface, and the seismic force exceeds the maximum static friction force between the smart device 10 and the installation surface, causing the smart device 10 to slide on the installation surface or move horizontally on the surface while bouncing up and down, the acceleration waveform often tends to have a portion in each of the components in the first direction X and the second direction Y where the amplitude is constant. In such a case, in the above-mentioned configuration, the portion in each of the components in the first direction X and the second direction Y of the acceleration waveform where the amplitude is constant within the horizontal movement section is complemented by a curve fitting process using a sine wave to generate a corrected acceleration waveform in the first direction X and the second direction Y. Therefore, the corrected acceleration waveform in the first direction X and the second direction Y can be appropriately generated. In the vertical movement section, the smart device 10 is in either a bounding state where the smart device 10 bounces up and down on the installation surface or a falling state where the smart device 10 falls from the installation surface. The magnitude of acceleration acting on the smart device 10 differs between the bounding state and the falling state. Therefore, in the above configuration, for a waveform based on a corrected acceleration waveform, components in the first direction X, the second direction Y, and the orthogonal direction Z are combined at each time to calculate a component composite acceleration, and the component composite acceleration is compared with a fall determination threshold to determine whether the smart device 10 is in a bounding state or a falling state at each time in the vertical movement section. When the smart device 10 is in a bound state, the smart device 10 floats in the air, falls, and lightly hits a surface. For this reason, in the acceleration waveform, excessive acceleration is observed in the orthogonal direction Z, which is not observed by a fixed seismometer. If the seismic intensity scale is estimated based on such an acceleration waveform, a seismic intensity scale that is higher than the actual scale may be estimated. Therefore, in the above-mentioned configuration, at the time when it is determined that the smart device 10 is in a bound state, the first direction X and the second direction Y are corrected by the curve fitting process using a sine wave as described above, and the component composite acceleration is recalculated by adjusting the component in the orthogonal direction Z to 0, and the seismic intensity scale is estimated based on this component composite acceleration. This reduces the influence of the excessive acceleration observed in the orthogonal direction Z in the bound state on the estimation of the seismic intensity scale. Furthermore, when the smart device 10 is in a falling state, there is a possibility that the smart device 10 may observe a large acceleration similar to that in the bound state. Furthermore, since the smart device 10 is completely suspended in the air, the external force due to the earthquake does not act on the smart device 10, and no waveforms due to the earthquake are obtained. For this reason, in the falling state, it is difficult to consider that a reliable acceleration waveform that can be used for estimating the seismic intensity scale is obtained in all directions, the first direction X, the second direction Y, and the orthogonal direction Z, and if the seismic intensity scale is estimated based on such an acceleration waveform, the estimation accuracy of the seismic intensity scale is reduced. Therefore, in the above-mentioned configuration, in the falling state, the component composite acceleration is recalculated by adjusting the components to 0 in all directions, the first direction X, the second direction Y, and the orthogonal direction Z, and the seismic intensity scale is estimated based on this component composite acceleration (i.e., the component composite acceleration is 0 at the relevant time). This reduces the influence of the acceleration waveform, which is considered to be unreliable, on the estimation of the seismic intensity scale in the falling state.
[0063] The system further includes a location input unit that accepts input of a location for which the seismic intensity class is to be estimated, and a smart device search unit 22 that searches for a smart device 10 that was located at the location input by the location input unit 21 at the time of the earthquake.A movement determination unit 24 determines whether or not the smart device 10 has moved based on the measurement results by the magnetic sensor 12 of the smart device 10 searched for by the smart device search unit 22, and an acceleration waveform correction unit 25 corrects the portion of the acceleration waveform observed by the acceleration sensor 11 of the smart device 10 searched for by the smart device search unit 22 that is in the movement section. According to the above-mentioned configuration, when an earthquake occurs, a search is performed for a smart device 10 that was located at the location input by the location input unit 21, and processing is performed based on the measurement result by the magnetic sensor 12 of the smart device 10 and the acceleration waveform observed by the acceleration sensor 11. Therefore, it is possible to estimate the seismic intensity scale of any location.
[0064] (Experimental Example) Next, we will explain an experiment carried out on the above-described seismic intensity estimation system 1. A triaxial shaking table experiment was carried out with a smartphone placed on a desk. A seismometer was fixed to the desk, and the seismic intensity measured by this seismometer was taken as the true value. As an example, the seismic intensity was estimated by correcting the records observed by the acceleration sensor of a smartphone as described in the above embodiment. When correcting the observed records, various settings were set as follows: Movement detection threshold: 5μT Time interval W B :1s Movement section determination threshold τ σB :0.5 Acceleration A, which is the standard for reaching a plateau sat :100cm / s As a comparative example, the seismic intensity was estimated directly using the records observed by the acceleration sensor of a smartphone.
[0065] The experimental results are shown in Fig. 11. In Fig. 11, the results of comparing the true seismic intensity values with the results of the comparative example in the same experiment are plotted with x marks, and the results of comparing the true seismic intensity values with the results of the embodiment in the same experiment are plotted with ● marks. Compared to the comparative example, the embodiment is plotted closer to the diagonal line L extending from the lower left to the upper right in the figure. This confirmed that the seismic intensity can be estimated more accurately by performing correction.
[0066] (Modification of the embodiment) In the above embodiment, the seismic intensity estimation system 1 determines whether the smart device 10 is placed on the installation surface when an earthquake occurs, and if it is determined that the smart device 10 is placed on the installation surface, it further determines whether the smart device 10 has moved relatively from the installation surface, and if it is determined that the smart device 10 has moved, corrects the acceleration waveform acquired from the smart device 10 to estimate the seismic intensity scale. However, no matter how accurately the acceleration waveform is corrected to generate a corrected acceleration waveform, it is difficult to generate a corrected acceleration waveform with an accuracy that completely matches the correct acceleration waveform that should be obtained when the smart device 10 does not move relatively from the installation surface during the earthquake. Therefore, a difference is inevitably generated between the acceleration waveform that should originally be obtained by the smart device 10 and the corrected acceleration waveform. For this reason, an error may occur between the correct value of the seismic intensity scale at the position where the smart device 10 is installed and the seismic intensity scale estimated by the seismic intensity scale estimation unit 26. FIG. 12 is a block diagram of a seismic intensity estimation system according to this modification. In the seismic intensity estimation system 1A of this modified example, the seismic intensity estimation device 20A is equipped with a seismic intensity class reliability evaluation unit 30, which evaluates the reliability of the seismic intensity class estimated by the seismic intensity class estimation unit 26, that is, how trustworthy the class is.
[0067] To perform the above evaluation, the seismic intensity class reliability evaluation unit 30 first evaluates a placement condition evaluation index based on the acceleration waveform observed by the acceleration sensor 11, which produces better results the more horizontally the smart device 10 is placed during an earthquake. When it is determined that the smart device 10 has moved in the horizontal direction, the acceleration waveform correction unit 25 of the above embodiment generates corrected acceleration waveforms in the first direction X and the second direction Y by complementing the parts of the acceleration waveform in which the amplitudes of the components in each of the first direction X and the second direction Y are constant in the horizontal movement section by a curve fitting process using a sine wave. In other words, the acceleration waveform correction process by the acceleration waveform correction unit 25 is premised on the assumption that the smart device 10 is placed horizontally on a horizontal surface such as the surface of a desk. For this reason, the seismic intensity scale reliability evaluation unit 30 evaluates the placement status evaluation index based on the acceleration waveform observed by the acceleration sensor 11 so that the closer the smart device 10 is placed in a position close to horizontal during an earthquake, the better the result, and the worse the result is when the smart device 10 is placed at an angle from the horizontal.
[0068] Specifically, the seismic intensity scale reliability evaluation unit 30 evaluates the gravitational acceleration (9.80665 m / s 2 ) is calculated to evaluate how close to horizontal the smart device 10 is placed. For this purpose, first, the seismic intensity class reliability evaluation unit 30 calculates the average acceleration Gx in the first direction X, the average acceleration Gy in the second direction Y, and the average acceleration Gz in the orthogonal direction Z of the observed acceleration waveform for a certain period of time, for example, 5 seconds, immediately before the arrival of the seismic wave. Next, the seismic intensity class reliability evaluation unit 30 calculates the ratio PZ of the orthogonal direction Z to the resultant acceleration of the first direction X, the second direction Y, and the orthogonal direction Z by the following formula (10).
number
[0069] When the smart device 10 is placed in a completely horizontal position, the contribution of the gravitational acceleration in the orthogonal direction Z is calculated to be 1 (100%). In contrast, when the position is tilted, for example, by 10 degrees from the horizontal, the contribution of the gravitational acceleration in the orthogonal direction Z is cos10°=0.9848 (98.48%). Therefore, for example, the threshold value is set to 0.985 (98.5%), and when the calculated proportion PZ is equal to or greater than the threshold value, the smart device 10 is evaluated to be placed horizontally. In this modified example, the placement status evaluation index G S is set so that a smaller value indicates a better evaluation result. In this modified example, when the ratio PZ is equal to or greater than a threshold value, the seismic intensity class reliability evaluation unit 30 evaluates that the smart device 10 is placed horizontally, and sets the placement status evaluation index G S Let be, for example, 1. In addition, in this modified example, when the ratio PZ is smaller than the threshold value, the seismic intensity class reliability evaluation unit 30 evaluates that the smart device 10 is not placed horizontally, and sets the placement status evaluation index G S is set to a value, for example 5, which is greater than 1, which is the value when the device is evaluated as being placed horizontally.
[0070] Next, the seismic intensity class reliability evaluation unit 30 evaluates the drop presence / absence evaluation index based on the acceleration waveform observed by the acceleration sensor 11, which gives a good result if the smart device 10 has not fallen during the earthquake. In the above embodiment, the acceleration waveform correction unit 25 determines that the smart device 10 is in a falling state when the composite acceleration of the components of the acceleration waveform observed by the acceleration sensor 11 of the smart device 10 is greater than or equal to 2G, which is the fall determination threshold, and based on the idea that the acceleration of any component is not useful in a falling state, the acceleration waveform in the falling state is not taken into consideration when estimating the seismic intensity scale. Here, when the smart device 10 falls and hits a floor or the like, the smart device 10 is deformed by the impact, and the deformation may cause free vibration in the smart device 10. Then, in the acceleration waveform after the falling state is released (i.e., the waveform to be considered when estimating the seismic intensity scale), the waveform caused by this free vibration is mixed into the waveform caused by the earthquake, which may be a factor in reducing the accuracy of the estimated seismic intensity scale. For this reason, the seismic intensity scale reliability evaluation unit 30 evaluates the fall presence / absence evaluation index to have a bad value, based on the idea that if the smart device 10 falls during an earthquake, the waveform after the fall may contain a waveform such as free vibration that may be a factor in reducing the accuracy of the estimated seismic intensity scale. In this way, the seismic intensity scale reliability evaluation unit 30 evaluates the fall presence / absence evaluation index based on the acceleration waveform observed by the acceleration sensor 11, so that if the smart device 10 does not fall during an earthquake, the result is good, and if the smart device 10 falls during an earthquake, the result is bad.
[0071] Specifically, the seismic intensity scale reliability evaluation unit 30 determines the fall presence / absence evaluation index G based on the result of the determination by the acceleration waveform correction unit 25 as to whether or not the component composite acceleration of the acceleration waveform observed by the acceleration sensor 11 of the smart device 10 is equal to or greater than 2G, which is the fall determination threshold value. F Rate the following. In this modified example, the drop presence / absence evaluation index G F is set so that a smaller value indicates a better evaluation result. In this modified example, when the acceleration waveform correction unit 25 has never determined that the smart device 10 is in a falling state, the seismic intensity class reliability evaluation unit 30 evaluates that the smart device 10 has not fallen, and sets the fall presence / absence evaluation index G F Let be, for example, 1. In addition, in this modified example, when the acceleration waveform correction unit 25 determines that the smart device 10 is in a falling state even once, the seismic intensity class reliability evaluation unit 30 evaluates that the smart device 10 has fallen, and sets the fall presence / absence evaluation index G Fis set to a value greater than 1, which is the value when it is evaluated that the drop did not occur, for example, to a value of 3.
[0072] Furthermore, in order to perform the above evaluation, the seismic intensity class reliability evaluation unit 30 evaluates a displacement evaluation index based on the acceleration waveform observed by the acceleration sensor 11, which produces better results the smaller the displacement of the smart device 10 during an earthquake. As described in the above embodiment, when an earthquake occurs and the smart device 10 moves and is displaced from the original position, an accurate acceleration waveform cannot be obtained, for example, as shown by the plateau portion WP in Fig. 5. Therefore, the greater the displacement of the smart device 10, the more likely it is that the obtained acceleration waveform deviates from the acceleration waveform that should be obtained when the smart device 10 does not displace. For this reason, the seismic intensity scale reliability evaluation unit 30 evaluates the drop presence / absence evaluation index based on the acceleration waveform observed by the acceleration sensor 11 so that the smaller the displacement of the smart device 10 during an earthquake, the better the result, and the greater the displacement, the worse the result.
[0073] In order to make such a judgment, it is necessary to obtain a displacement waveform by integrating the acceleration waveform twice. Even if the acceleration waveform is obtained by a fixed seismometer, the calculation of the displacement waveform requires manual base axis correction, so it is difficult to perform this automatically with high accuracy. Furthermore, it is even more difficult to calculate a displacement waveform based on an acceleration waveform observed by a smart device 10 that is not fixed and may be displaced during an earthquake because the base axis changes intermittently, and it is even more difficult to perform this automatically with high accuracy. Here, when the change in the base axis is large, if the displacement waveform is calculated without base axis correction, the maximum displacement will be a large value. Also, there is a positive correlation between the seismic intensity and the maximum displacement, such that the maximum displacement of the smart device 10 increases as the seismic intensity increases. Therefore, in this modified example, the maximum displacement is calculated from the acceleration waveform of the smart device 10 without base axis correction, and the displacement of the smart device 10 is estimated based on the seismic intensity obtained from a fixed seismometer, and the displacement evaluation index is evaluated by calculating how far the calculated maximum displacement is from the estimated displacement.
[0074] FIG. 13 is a graph showing the relationship between the seismic intensity and the maximum displacement of a fixed seismometer and an unfixed smart device 10 during an earthquake. 13, a relationship between the seismic intensity measured by the seismic intensity estimation system 1 of the above embodiment and the maximum displacement of each smart device 10 was experimentally determined and plotted when vibrations were applied to both a fixed smart device 10 and an unfixed smart device 10 by a three-axis vibration table experiment. As a result, the following regression equation (11) was obtained.
number
[0075] The seismic intensity scale reliability evaluation unit 30 calculates the maximum displacement D based on the acceleration waveform obtained from the smart device 10. OThen, the seismic intensity scale reliability evaluation unit 30 calculates the maximum displacement D O The value log obtained by applying the common logarithm to 10 (D O ) is calculated by the seismic intensity estimation system 1 of the above embodiment, specifically, the measured seismic intensity I calculated by the formula (9), and JMA The value obtained when inputting log 10 (D R ) to get the standard deviation σ R By calculating how many times the distance is, the displacement evaluation index G D Specifically, the seismic intensity scale reliability evaluation unit 30 calculates the displacement evaluation index G D Calculate.
number
[0076] The seismic intensity scale reliability evaluation unit 30 calculates the placement status evaluation index G S , Fall Presence Evaluation Index G F , and the displacement evaluation index G D Based on the results of these evaluations, the reliability of the seismic intensity scale estimated by the seismic intensity scale estimation unit 26 is evaluated. Specifically, the seismic intensity scale reliability evaluation unit 30 calculates the placement status evaluation index G S , Fall Presence Evaluation Index G F , and the displacement evaluation index G D The maximum value of each of these is the overall evaluation index G C The calculation is as follows: Loading status evaluation index G S, Fall Presence Evaluation Index G F , and the displacement evaluation index G D Each of the following is a rating of 1 to 5 (Fall Presence / Absence Evaluation Index G F Since the overall evaluation index G C The overall evaluation index G C A smaller value indicates that the reliability of the seismic intensity scale estimated by the seismic intensity scale estimation unit 26 is higher, and a larger value indicates that the reliability of the seismic intensity scale estimated by the seismic intensity scale estimation unit 26 is lower. For example, the seismic intensity scale reliability evaluation unit 30 may calculate the overall evaluation index G C The values of 1, 2, 3, 4, and 5 may be associated with estimated grades A, B, C, D, and E, respectively, and the calculated estimated grades may be displayed.
[0077] Fig. 14 is a graph showing the verification results for this modified example. Fig. 14 plots the seismic intensity measured by a fixed seismometer and an unfixed seismometer, i.e., the smart device 10, when vibrations are applied to both. The closer the plotted points are to the 45-degree line, the higher the accuracy of the seismic intensity calculated by the smart device 10. In Figure 14, the higher the estimated grade is (A), the closer it is plotted to the 45-degree line, and it can be seen that the higher the estimated grade is, the higher the accuracy of the seismic intensity is.
[0078] Next, a method for evaluating the reliability of an estimated seismic intensity scale using the above-mentioned seismic intensity estimation system 1 will be described with reference to Fig. 10, Fig. 12 to Fig. 14, and Fig. 15. Fig. 15 is a flowchart of a part for evaluating the reliability of an estimated seismic intensity scale in the seismic intensity estimation method according to this modified example. When the calculation of the seismic intensity scale based on the measured seismic intensity by the seismic intensity scale estimating unit 26 is completed as shown in step S13 in FIG. 10, the seismic intensity scale reliability evaluating unit 30 calculates a placement status evaluation index G based on the acceleration waveform observed by the acceleration sensor 11, which is a better result when the smart device 10 is placed horizontally during the earthquake. Sis evaluated (step S21). Next, the seismic intensity scale reliability evaluation unit 30 determines a fall occurrence evaluation index G based on the acceleration waveform observed by the acceleration sensor 11. The fall occurrence evaluation index G is a good result if the smart device 10 has not fallen during the earthquake. F is evaluated (step S22). Then, the seismic intensity scale reliability evaluation unit 30 calculates a displacement evaluation index G based on the acceleration waveform observed by the acceleration sensor 11. The smaller the difference between the displacement of the smart device 10 during an earthquake and the equation (11), the better the result. D is evaluated (step S23). Finally, the seismic intensity scale reliability evaluation unit 30 calculates the placement status evaluation index G S , Fall Presence Evaluation Index G F , and the displacement evaluation index G D Based on the evaluation result, the reliability of the seismic intensity scale estimated by the seismic intensity scale estimation unit 26 is evaluated (step S24).
[0079] In the above configuration, the seismic intensity estimation system 1A calculates, based on the acceleration waveform observed by the acceleration sensor 11, a placement status evaluation index G, which gives a better result when the smart device 10 is placed horizontally during an earthquake. S If the smart device 10 does not fall during the earthquake, the result is good. F The smaller the displacement of the smart device 10 during an earthquake, the better the results. D and , and the placement status evaluation index G S , Fall Presence Evaluation Index G F , and the displacement evaluation index G D The seismic intensity scale reliability evaluation unit 30 evaluates the reliability of the seismic intensity scale estimated by the seismic intensity scale estimation unit 26 based on the results of each of the evaluations above. Even if an acceleration waveform is acquired from the smart device 10 as in the above embodiment and corrected to estimate the seismic intensity scale, if the smart device 10 is not placed horizontally, for example on a desk, when the acceleration waveform is acquired, if it has been dropped, or if it has moved significantly, the estimated seismic intensity scale may not be sufficiently reliable. In contrast, according to the above-described configuration, the seismic intensity scale reliability evaluation unit 30 calculates a placement status evaluation index G S If the smart device 10 does not fall during the earthquake, the result is good. F The smaller the displacement of the smart device 10 during an earthquake, the better the results. D The seismic intensity scale reliability evaluation unit 30 evaluates each of these placement status evaluation indexes G S , Fall Presence Evaluation Index G F , and the displacement evaluation index G D Based on each of the evaluation results, the reliability of the seismic intensity scale estimated by the seismic intensity scale estimation unit 26 is evaluated. Therefore, based on this evaluation result, it can be determined whether the estimated seismic intensity scale has sufficient reliability.
[0080] During an earthquake, it is possible that a large number of acceleration waveforms can be obtained from many smart devices 10. In such a case, by performing the above-mentioned process on each of the acceleration waveforms obtained from each smart device 10 and evaluating its reliability, it is possible to screen the data, for example, by using data obtained from smart devices 10 with an estimated grade of D or higher when a large amount of seismic intensity data is desired even if there is some error, or by using only data obtained from smart devices 10 with an estimated grade of A or higher when only highly reliable seismic intensity data is desired.
[0081] The seismic intensity estimation system 1 of the present invention is not limited to the above-mentioned embodiment and modified examples explained with reference to the drawings, and various other modified examples are conceivable within the technical scope thereof. For example, in the above embodiment, the point input unit 21 accepts the input of a point for which the seismic intensity scale is to be estimated by the seismic intensity estimation system 1, but this is not limited to the above. The point input unit may accept the input of an area having a certain range in the east-west and north-south directions, and estimate the seismic intensity scale for the entire area. In this case, the smart device search unit may search for a plurality of, for example, all, smart devices 10 located within the input area, and the movement determination unit 24 and the acceleration waveform correction unit 25 may determine the movement of these smart devices 10 and correct the acceleration waveform. The seismic intensity scale estimation unit 26 may estimate the seismic intensity corresponding to each of the plurality of smart devices 10, link this to the location information of the corresponding smart device 10, and create a seismic intensity map representing the entire specified area so that the seismic intensity corresponding to the smart device 10 is displayed at the location where the smart device 10 was located at the time of the earthquake.
[0082] In addition, it is possible to select and discard the configurations given in the above embodiments and modifications, or to change them to other configurations as appropriate. [Explanation of symbols]
[0083] 1, 1A Seismic Intensity Estimation System 24 Movement Judgment Unit 10 Smart device 25 Acceleration waveform correction section 10a Screen 26 Seismic Intensity Scale Estimation Section 11 Acceleration sensor 30 Seismic intensity scale reliability evaluation section 12 Magnetic sensor X 1st direction 20, 20A Seismic intensity estimation device Y Second direction 21 Point input section Z orthogonal direction 22 Smart Device Search Section
Claims
1. A seismic intensity estimation system using a smart device having an acceleration sensor and a magnetic sensor built therein, A movement determination unit that determines whether the smart device has moved from a stationary state due to an earthquake based on a measurement result by the magnetic sensor, and if it is determined that the smart device has moved, specifies a movement section that is a time section during which the smart device has moved; an acceleration waveform correction unit that corrects a portion of the acceleration waveform observed by the acceleration sensor in the movement section when it is determined that the smart device has moved, to generate a corrected acceleration waveform; a seismic intensity scale estimation unit that estimates a seismic intensity scale based on the corrected acceleration waveform; A seismic intensity estimation system comprising:
2. The movement determination unit is In a time history waveform of the magnetic flux density, which is the measurement result by the magnetic sensor, when a change in an average value of the magnetic flux density before and after the earthquake is equal to or greater than a movement determination threshold, it is determined that the smart device has moved; The seismic intensity estimation system of claim 1, characterized in that when it is determined that the smart device has moved, the standard deviation within a predetermined time interval is calculated using the time as a base point while advancing the time in the time history waveform of the magnetic flux density, and when the standard deviation is greater than or equal to a movement section determination threshold, the end time of the time interval using the time as a base point is set as the start time of the movement section, and when the standard deviation is smaller than the movement section determination threshold, the start time of the time interval using the time as a base point is set as the end time of the movement section.
3. The movement determination unit is Based on the measurement results of the components in a first direction and a second direction perpendicular to each other in a plane parallel to the screen of the smart device by the magnetic sensor, it is determined whether the smart device has moved in a horizontal direction, and if it is determined that the smart device has moved in the horizontal direction, a horizontal movement section that is the movement section in the horizontal direction is specified; The seismic intensity estimation system of claim 1 or 2, characterized in that it determines whether the smart device has moved in an up or down direction based on the measurement results of the component in an orthogonal direction perpendicular to the screen by the magnetic sensor, and if it is determined that the smart device has moved in an up or down direction, it identifies the up or down movement section that is the movement section in the up or down direction.
4. when it is determined that the smart device has moved in a horizontal direction, the acceleration waveform correction unit complements a portion of the acceleration waveform in which the amplitude of each of the components in the first direction and the second direction is constant in the horizontal movement section by a curve fitting process using a sine wave, thereby generating the corrected acceleration waveform in the first direction and the second direction; The acceleration waveform correction unit corrects, at each time, a waveform based on the corrected acceleration waveform, Calculating a component resultant acceleration by combining components in each of the first direction, the second direction, and the orthogonal direction; If the time is within the vertical movement section and the component resultant acceleration is equal to or greater than a fall determination threshold, the smart device is deemed to be in a falling state at the time, and the components of the waveform based on the corrected acceleration waveform in the first direction, the second direction, and the orthogonal direction are set to 0, and the component resultant acceleration is recalculated; If the time is within the vertical movement section and the component resultant acceleration is smaller than a fall determination threshold, the smart device is deemed to be bouncing up and down at the time, and the component in the orthogonal direction of the waveform based on the corrected acceleration waveform is set to 0, and the component resultant acceleration is recalculated; The seismic intensity scale estimation unit estimates the seismic intensity scale based on the component composite acceleration. The seismic intensity estimation system according to claim 3 .
5. Based on the acceleration waveform observed by the acceleration sensor, A placement status evaluation index that produces better results the more horizontally the smart device is placed during an earthquake; A fall presence / absence evaluation index that is a good result if the smart device does not fall during an earthquake; A displacement evaluation index in which the smaller the displacement of the smart device during an earthquake, the better the result; and evaluates the reliability of the seismic intensity scale estimated by the seismic intensity scale estimation unit based on the evaluation results of the placement status evaluation index, the fall presence / absence evaluation index, and the displacement amount evaluation index. The seismic intensity estimation system according to claim 1 .
Citation Information
Patent Citations
Seismic intensity information providing device and seismic intensity information providing method
JP2017133996A
Evaluation system and evaluation program
JP2020193935A
Earthquake observation system, earthquake observation program, and earthquake observation method
WO2018174296A1
Cited By
Ammunition reaction intensity multi-parameter factor quantitative evaluation method
CN120998324A
Method, apparatus, and electronic equipment for linking earthquake warnings and indoor occupancy identification.
JP2026150479A
Method, apparatus, and electronic equipment for linking earthquake warnings and indoor occupancy identification.
JP7859713B1
