Information processing device and information processing program

Abstract

The objective is to obtain an information processing device and information processing program that can detect abnormalities in acceleration sensors more efficiently than conventional technologies. [Solution] The information processing device 10 includes an acquisition unit 11A that acquires acceleration time history data obtained by an acceleration sensor of a seismometer installed in a building, and a conversion unit 11B that converts the acquired acceleration time history data into image data by aligning the time axis in one direction and converting the acceleration values ​​into pixel values.

Description

Technical Field

[0001] The present invention relates to an information processing apparatus and an information processing program.

Background Art

[0002] When a major earthquake occurs, in buildings expected to function as disaster prevention bases, a prompt determination of whether the building can continue to be used after the earthquake is required.

[0003] For example, in the Great East Japan Earthquake, the problem of people having difficulty returning home arose. For the building side to accept those having difficulty returning home, it is necessary to quickly confirm the safety of the facilities immediately after the earthquake. As one of the countermeasures, in recent years, the need for a structural health monitoring system that determines the soundness of a building after an earthquake has been increasing.

[0004] Currently, in a generally implemented structural health monitoring system, the deformation of a building is obtained by acceleration integration and mode analysis using information obtained from seismographs having acceleration sensors installed in multiple locations within the building, and the damage degree of each floor of the building is estimated from the maximum inter-story drift angle. This is the mainstream method.

[0005] By the way, for these systems to function when an earthquake occurs, it is a prerequisite that the installed seismographs can accurately record the shaking. However, due to the influence of factors such as malfunctions of the seismographs themselves, aging deterioration, and changes in installation conditions, the acceleration waveform may be affected by noise, shift (baseline deviation), drift (tilt), etc. In this case, there is a risk that the determination by the system may not be correct.

[0006] To address this problem, one method is to remove noise through filtering. However, it is difficult to predict how a malfunction in a seismometer will affect the acceleration waveform, and filtering alone is insufficient. Therefore, regular monitoring and inspection of whether the seismometer is measuring correctly is essential. During this process, in addition to the acceleration waveform, it is necessary for a person to visually check graphs of periodic characteristics such as Fourier spectra and discuss and evaluate whether there is a malfunction. As a result, inspecting numerous seismometers within a building, and even multiple locations, requires a great deal of manpower, making the demand for more efficient and automated detection of acceleration sensor malfunctions extremely high.

[0007] Conventionally, the following techniques have been used to detect abnormalities in acceleration sensors in seismometers.

[0008] Patent Document 1 discloses a seismometer that aims to easily determine whether or not equipment failure has occurred.

[0009] This seismometer has an acceleration sensor that measures acceleration due to shaking and outputs the measured acceleration data, and further comprises a vibrator that vibrates the acceleration sensor, acquires acceleration data output from the acceleration sensor when the vibrator vibrates, and determines equipment failure by comparing the original vibration wave obtained by the vibration of the vibrator with the vibration wave based on the acceleration data acquired from the acceleration sensor, the acceleration sensor is configured to measure acceleration along three mutually orthogonal measurement axes, the vibrator is arranged to vibrate the acceleration sensor in any direction of the three measurement axes, the vibrator is mounted on a substrate on which the acceleration sensor is arranged, and vibrates in at least two orthogonal axis directions, one of the directions of vibration is in the same direction as or inclined with respect to one of the measurement axes of the acceleration sensor, and the other direction of vibration is arranged to be inclined with respect to two measurement axes of the acceleration sensor that are orthogonal to one of the measurement axes.

[0010] Patent document 2 discloses an earthquake sensor aimed at reliably detecting failures in acceleration sensors.

[0011] This seismic sensor is for measuring acceleration caused by an earthquake and comprises four acceleration sensors that measure acceleration components in different measurement axis directions, and a support body that supports the four acceleration sensors, characterized in that the support body holds the four acceleration sensors in the measurement axis direction such that the vectors of the acceleration components in the measurement axis direction of each of the four acceleration sensors cancel each other out.

[0012] Patent Document 3 discloses a seismic motion measuring device aimed at determining the characteristics of a seismometer for preventing misjudgments based on its installation orientation, and enabling its use for preventing misjudgments.

[0013] This seismic motion measuring device comprises a first seismometer that is primarily responsible for measuring seismic motion, a second seismometer for preventing misjudgments, a seismic motion measuring unit that measures the time history of the first seismometer and the second seismometer, a characteristic determination unit that compares the time history of the first seismometer and the second seismometer measured by the seismic motion measuring unit to determine the characteristics of the second seismometer, and a misjudgment prevention unit that uses the characteristics determined by the characteristic determination unit to perform equalization processing of the first seismometer and the second seismometer to prevent misjudgments, wherein the second seismometer is an electrodynamic speedometer.

[0014] Non-patent document 1 discloses a method for detecting anomalies in seismometers within the domestic seismic observation network by training a neural network called an autoencoder with normal earthquake records. This method targets a single seismometer and uses Fourier transformed data as training data. [Prior art documents] [Patent Documents]

[0015] [Patent Document 1] Patent No. 6398137 [Patent Document 2] Patent No. 3808480 [Patent Document 3] Patent No. 6124253 [Non-patent literature]

[0016] [Non-Patent Document 1] Proceedings of the 34th Annual Conference of the Japanese Society for Artificial Intelligence, "Time-Series Anomaly Detection in Seismometers Using Autoencoders," Jianing Guo et al. (NTT Communications Corporation), Internet search. <URL:https: / / www.jstage.jst.go.jp / article / pjsai / JSAI2020 / 0 / JSAI2020_4L2GS1303 / _article / -char / ja / > [Overview of the project] [Problems that the invention aims to solve]

[0017] However, the technology disclosed in Patent Document 1 requires a vibrator to vibrate the acceleration sensor, the technology disclosed in Patent Document 2 requires four acceleration sensors, and the technology disclosed in Patent Document 3 requires a second seismometer to prevent false detections.

[0018] Therefore, the technologies described in these patent documents require special items in addition to the acceleration sensor that is the target of anomaly detection, and thus have the problem that the anomaly detection efficiency of the acceleration sensor is not necessarily high.

[0019] Furthermore, the technology disclosed in Non-Patent Document 1 targets a single seismometer and uses Fourier-transformed data as training data. This technology also has the problem that it cannot necessarily be said to have high efficiency in detecting anomalies in acceleration sensors.

[0020] The present disclosure has been made in view of the above facts, and an object thereof is to provide an information processing apparatus and an information processing program that can detect an abnormality of an acceleration sensor more efficiently as compared with the conventional technology.

Means for Solving the Problems

[0021] The information processing apparatus according to the present invention described in claim 1 includes an acquisition unit that acquires acceleration time history data obtained by an acceleration sensor provided in a building, and the acquired acceleration time history data is made to correspond to the time axis in one direction, and a conversion unit that converts the acceleration value into a pixel value to convert it into image data.

[0022] According to the information processing apparatus according to the present invention described in claim 1, the acceleration time history data obtained by the acceleration sensor provided in the building is acquired, and the acquired acceleration time history data is made to correspond to the time axis in one direction, and by converting the acceleration value into a pixel value to convert it into image data, it is possible to detect an abnormality of the acceleration sensor more efficiently as compared with the conventional technology without requiring complicated operations such as other articles or Fourier transform.

[0023] The information processing apparatus according to the present invention described in claim 2 is the information processing apparatus according to claim 1, wherein the acceleration time history data is data indicating the time history of acceleration for each of the X-axis direction, the Y-axis direction, and the Z-axis direction, which are directions that intersect each other, and the image data is color image data represented by three primary colors, and the conversion unit converts the acceleration value in the X-axis direction into the pixel value of the first primary color in the three primary colors of the image data, converts the acceleration value in the Y-axis direction into the pixel value of the second primary color in the three primary colors of the image data, and converts the acceleration value in the Z-axis direction into the pixel value of the third primary color in the three primary colors of the image data.

[0024] According to the information processing apparatus according to the present invention described in claim 2, the acceleration time history data is data indicating the time history of acceleration for each of the X-axis direction, Y-axis direction, and Z-axis direction, which are directions intersecting each other, and the image data is color image data represented by three primary colors. The conversion unit converts the acceleration value in the X-axis direction into the pixel value of the first primary color in the three primary colors of the image data, converts the acceleration value in the Y-axis direction into the pixel value of the second primary color in the three primary colors of the image data, and converts the acceleration value in the Z-axis direction into the pixel value of the third primary color in the three primary colors of the image data. Thus, even when applied to an apparatus that obtains three-dimensional acceleration time history data in the X-axis direction, Y-axis direction, and Z-axis direction as an acceleration sensor, abnormalities of the acceleration sensor can be detected more efficiently with a single image.

[0025] The information processing apparatus according to the present invention described in claim 3 is the information processing apparatus according to claim 1 or claim 2, wherein the acceleration sensors are provided on a plurality of floors of the building, and the conversion unit performs the conversion by associating each of the plurality of acceleration sensors provided on the plurality of floors with a direction intersecting the one direction of the image data.

[0026] According to the information processing apparatus according to the present invention described in claim 3, the acceleration sensors are provided on a plurality of floors of a building, and the conversion unit performs the conversion by associating each of the plurality of acceleration sensors provided on the plurality of floors with a direction intersecting the one direction of the image data. Thus, even for the plurality of acceleration sensors provided on the plurality of floors, abnormalities of the acceleration sensors can be detected more efficiently with a single image.

[0027] The information processing apparatus according to the present invention described in claim 4 is the information processing apparatus according to claim 1 or claim 2, further comprising a display control unit that controls a display unit to display an image indicated by the image data obtained by the conversion unit.

[0028] According to the information processing apparatus of the present invention as described in claim 4, by controlling the display unit to display the image shown by the image data obtained by the conversion unit, a person can visually detect the occurrence of an abnormality in the acceleration sensor.

[0029] The information processing program according to claim 5 of the present invention acquires acceleration time history data obtained by an acceleration sensor installed in a building, and causes a computer to perform a process that converts the acquired acceleration time history data into image data by aligning the time axis in one direction and converting the acceleration values ​​into pixel values.

[0030] According to the information processing program of the present invention described in claim 5, acceleration time history data obtained from an acceleration sensor installed in a building is acquired, and the acquired acceleration time history data is converted into image data by aligning the time axis in one direction and converting the acceleration values ​​into pixel values. This eliminates the need for other objects or complex calculations such as Fourier transforms, and allows for more efficient detection of abnormalities in the acceleration sensor compared to conventional techniques. [Effects of the Invention]

[0031] As explained above, the present invention makes it possible to detect abnormalities in acceleration sensors more efficiently compared to conventional techniques. [Brief explanation of the drawing]

[0032] [Figure 1] This block diagram shows an example of the hardware configuration of an information processing system according to the first embodiment. [Figure 2] This is a functional block diagram showing an example of the functional configuration of the information processing system according to the first embodiment. [Figure 3] This is a side cross-sectional view showing an example of the installation of the information processing device and seismometer according to the embodiment. [Figure 4] This figure shows an example of acceleration time history data and a converted image according to the embodiment. [Figure 5]This flowchart shows an example of the information processing flow according to the first embodiment. [Figure 6] This figure shows an example of a converted image display screen according to the first embodiment. [Figure 7] This block diagram shows an example of the hardware configuration of an information processing system according to the second embodiment. [Figure 8] This is a functional block diagram showing an example of the functional configuration of the information processing system according to the second embodiment. [Figure 9] This flowchart shows an example of the information processing flow according to the second embodiment. [Figure 10] This figure shows an example of an evaluation results screen according to the second embodiment. [Modes for carrying out the invention]

[0033] The following describes in detail an example of an information processing system to which the information processing apparatus and information processing program according to the present invention are applied.

[0034] [First Embodiment] First, the configuration of the information processing system 90 according to this embodiment will be described with reference to Figure 1. Figure 1 is a block diagram showing an example of the hardware configuration of the information processing system 90 according to this embodiment.

[0035] As shown in Figure 1, the information processing system 90 according to this embodiment includes an information processing device 10 and a plurality of seismometers 62 installed on each floor of the target building 60. Examples of information processing devices for the information processing device 10 include personal computers and server computers.

[0036] The seismometer 62 according to this embodiment is equipped with an acceleration sensor 64 having three sensors that individually detect acceleration in the X-axis, Y-axis, and Z-axis directions, which are directions that intersect with each other. Hereinafter, the sensor that detects acceleration in the X-axis direction will be referred to as the "X-axis sensor," the sensor that detects acceleration in the Y-axis direction will be referred to as the "Y-axis sensor," and the sensor that detects acceleration in the Z-axis direction will be referred to as the "Z-axis sensor."

[0037] In the information processing system 90 according to this embodiment, the information processing device 10 and a plurality of seismometers 62 are electrically connected. Although A / D (analog / digital) converters, communication devices, etc., are interposed between the information processing device 10 and each seismometer 62, these are not shown or described in order to avoid confusion.

[0038] The information processing device 10 can acquire time history data of acceleration in the X-axis direction detected by the X-axis sensor (hereinafter referred to as "X-axis acceleration time history data") from the acceleration sensors 64 provided on each of the seismometers 62. Similarly, the information processing device 10 can acquire time history data of acceleration in the Y-axis direction detected by the Y-axis sensor (hereinafter referred to as "Y-axis acceleration time history data") and time history data of acceleration in the Z-axis direction detected by the Z-axis sensor (hereinafter referred to as "Z-axis acceleration time history data"). In the following, when X-axis acceleration time history data, Y-axis acceleration time history data, and Z-axis acceleration time history data are described without distinction, they will be collectively referred to as "acceleration time history data".

[0039] The information processing device 10 according to this embodiment includes a CPU (Central Processing Unit) 11 as a computer, a memory 12 as a temporary storage area, a non-volatile storage unit 13, an input unit 14 such as a keyboard and mouse, a display unit 15 such as a liquid crystal display, a media read / write device (R / W) 16, and a communication interface (I / F) unit 18. The CPU 11, memory 12, storage unit 13, input unit 14, display unit 15, media read / write device 16, and communication I / F unit 18 are connected to each other via bus B. The media read / write device 16 reads information written on the recording medium 17 and writes information to the recording medium 17.

[0040] The storage unit 13 in this embodiment is implemented by an HDD (Hard Disk Drive), SSD (Solid State Drive), flash memory, etc. The storage unit 13, as a storage medium, stores an information processing program 13A. The information processing program 13A is stored (installed) in the storage unit 13 when the recording medium 17 on which the program 13A is written is set in the media read / write device 16, and the media read / write device 16 reads the program 13A from the recording medium 17. The CPU 11 reads the information processing program 13A from the storage unit 13 as appropriate, loads it into the memory 12, and sequentially executes the processes that the program 13A has.

[0041] Figure 2 is a functional block diagram showing an example of the functional configuration of the information processing system 90 according to this embodiment.

[0042] As shown in Figure 2, the information processing device 10 according to this embodiment includes an acquisition unit 11A, a conversion unit 11B, and a display control unit 11C. The CPU 11 of the information processing device 10 executes an information processing program 13A, and the CPU 11 functions as the acquisition unit 11A, the conversion unit 11B, and the display control unit 11C.

[0043] The acquisition unit 11A according to this embodiment acquires acceleration time history data obtained by an acceleration sensor 64 installed in the building 60.

[0044] Furthermore, the conversion unit 11B according to this embodiment converts the acceleration time history data acquired by the acquisition unit 11A into image data by aligning the time axis in one direction and converting the acceleration values ​​into pixel values.

[0045] Then, the display control unit 11C according to this embodiment controls the display unit 15 to display the image (converted image 70, which will be described in detail later (see also Figure 4)) indicated by the image data obtained by the conversion unit 11B.

[0046] In this embodiment, the information processing system 90 applies data showing the time history of acceleration in the X-axis direction, Y-axis direction, and Z-axis direction, which are directions that intersect each other, as acceleration time history data, i.e., the X-axis acceleration time history data, Y-axis acceleration time history data, and Z-axis acceleration time history data described above.

[0047] Furthermore, in the information processing system 90 according to this embodiment, color image data represented by three primary colors is applied as the image data.

[0048] The conversion unit 11B according to this embodiment converts the acceleration value in the X-axis direction into the pixel value of the first primary color in the three primary colors of the image data, the acceleration value in the Y-axis direction into the pixel value of the second primary color in the three primary colors of the image data, and the acceleration value in the Z-axis direction into the pixel value of the third primary color in the three primary colors of the image data.

[0049] Furthermore, the conversion unit 11B according to this embodiment performs the above conversion by corresponding each of the multiple acceleration sensors 64 provided on multiple floors of the building 60 to the direction that intersects the image data in the aforementioned direction.

[0050] In this embodiment, the color image data represented by the three primary colors of light, R (Red), G (Green), and B (Blue), is applied, but the embodiment is not limited to this. For example, image data represented by the three primary colors of pigment, C (Cyan), M (Magenta), and Y (Yellow), may be applied as the color image data. Furthermore, in addition to the RGB and CMY color spaces, image data from various other color spaces, such as the Lab color space and XYZ color space, may also be applied as the color image data.

[0051] Furthermore, in this embodiment, the horizontal direction of the image indicated by the image data is applied as the one direction, and the vertical direction of the image indicated by the image data is applied as the direction intersecting the one direction, but this is not limited to this. For example, conversely, the vertical direction of the image indicated by the image data may be applied as the one direction, and the horizontal direction of the image indicated by the image data may be applied as the direction intersecting the one direction.

[0052] Next, with reference to Figure 3, the installation status of the information processing device 10 and the seismometer 62 in the information processing system 90 according to this embodiment will be described. Figure 3 is a side cross-sectional view showing an example of the installation status of the information processing device 10 and the seismometer 62 according to this embodiment.

[0053] As shown in Figure 3, in the information processing system 90 according to this embodiment, seismometers 62 are installed on each floor from the basement floor (indicated as "B1F" in Figure 3) to the rooftop (indicated as "RF" in Figure 3) of the building 60. In the information processing system 90 according to this embodiment, one seismometer 62 is installed on each floor of the building 60 at approximately the same location in plan view. However, the system is not limited to this configuration, and it is also possible to install multiple seismometers 62 on each floor of the building 60, and their placement is not limited to approximately the same location in plan view.

[0054] Furthermore, as shown in Figure 3, in the information processing system 90 according to this embodiment, the information processing device 10 is installed on the basement floor of the building 60, but this is not the only configuration. For example, the information processing device 10 may be installed on another floor of the building 60, or it may be installed in a building different from the building 60.

[0055] Next, with reference to Figure 4, the method by which the conversion unit 11B according to this embodiment converts acceleration time history data into image data will be described in more detail.

[0056] The conversion unit 11B according to this embodiment associates acceleration in three directions: the two horizontal directions, the X-axis and Y-axis, and the vertical direction, the Z-axis, with R (red) for acceleration in the X-axis direction, G (green) for acceleration in the Y-axis direction, and B (blue) for acceleration in the Z-axis direction. Furthermore, the conversion unit 11B according to this embodiment converts the acceleration time history data measured by seismometers 62 on multiple floors of the building 60 into a single conversion image 70 by associating the vertical axis of the resulting conversion image 70, shown in the left diagram of Figure 4, with the acceleration sensors 64 on each floor of the building 60, and the horizontal axis with the time axis.

[0057] When creating the image for this converted image 70, if the image data representing the converted image 70 is 8-bit data, the pixel values ​​of R, G, and B will be in the range of 0 (zero) to 255. Therefore, the acceleration values ​​are standardized using the following equations (1) to (3) to derive the pixel values ​​(grayscale values) of R, G, and B. In equations (1) to (3), X, Y, and Z represent the acceleration values ​​in the X-axis direction, Y-axis direction, and Z-axis direction, respectively, and max(|X|), max(|Y|), and max(|Z|) represent the maximum absolute values ​​of the acceleration values ​​for each axis direction to be detected in the entire acceleration time history data (including all levels), respectively. In equations (1) to (3), the results of the calculations on the right-hand side are rounded to integers to obtain the pixel values ​​of R, G, and B.

[0058]

number

[0059] For example, if the acceleration values ​​are 0 in all three directions—X, Y, and Z—then [R, G, B] = [128, 128, 128] will be represented by the color gray.

[0060] Note that "127.5" in equations (1) to (3) is applied because the converted image 70 is an 8-bit image with a maximum pixel value of 255, while the acceleration time history data has positive and negative values ​​with 0 (zero) as the boundary. Therefore, half of 255 is applied. Consequently, "127.5" in equations (1) to (3) will vary depending on the bit configuration of the converted image 70.

[0061] Furthermore, in this embodiment, the conversion from acceleration time history data to pixel values ​​of the converted image 70 is performed by equations (1) to (3), but it is not limited to this. For example, if the priority of sensors for detecting abnormalities is predetermined, a calculation formula may be applied that calculates larger values ​​for pixel values ​​corresponding to sensors that detect acceleration in the axial direction with a higher priority.

[0062] The converted image 70 shown in Figure 4 is obtained by the conversion unit 11B according to this embodiment when the acceleration time history data detected by the acceleration sensor 64 is as shown in the right-hand figure of Figure 4. The example shown in Figure 4 is an example corresponding to the acceleration time history data in the three directions on each of the 19 floors from the 1st floor to the rooftop (R floor) of an 18-story building. As described above, the vertical axis of the converted image 70 represents the floor and the horizontal axis represents the time axis. The horizontal axis of the converted image 70 has 1000 pixels (=5 / 0.005) because the time step dt is 0.005 seconds and the recording is for 5 seconds, and the vertical axis is stretched to a length of 50 pixels per floor, with a pixel count of 950 (=19×50), resulting in an image with a size of 950 pixels × 1000 pixels.

[0063] Here, there are no particular restrictions on the pixel length in both the vertical and horizontal directions, but it is preferable to adjust them appropriately to match the number of seismometers 62 and the time period to be plotted, in order to facilitate visual evaluation or automatic evaluation by image recognition, etc.

[0064] Furthermore, the example shown in Figure 4 shows the results of intentionally adding white noise to the acceleration time history data in the X-axis direction on the 5th floor, and a baseline shift of 0.5 (gal) to the acceleration time history data in the Y-axis direction on the 10th floor. For convenience, the converted image 70 shown in Figure 4 is displayed in grayscale, which makes it somewhat difficult to see, but in the actual converted image 70, a band of a different color tone appears on the corresponding floor compared to other floors, making it easy to visually confirm that it contains anomalies.

[0065] In the converted image 70 shown in Figure 4, a series of slightly tilted stripes appears, which is because the stripes are gradually shifted horizontally due to the phase lag of the response waveforms, starting from the first floor. As the converted image 70 shows, in response to a steady-state seismic motion, the building continues to shake according to its vibration mode, and when the response waveforms of each floor are converted into images, it can be seen that a striped pattern with regularity in the vertical direction is observed.

[0066] If the acceleration sensors 64 installed in each seismometer 62 are measuring acceleration correctly, the pattern of the converted image 70 will show a certain regularity, unless the vibration mode changes due to building damage or the like. Therefore, if a floor is found in the converted image 70 that shows a change in color that deviates from this regularity, it can be assumed that some kind of abnormality has occurred in the acceleration sensor 64 of the seismometer 62 on that floor. Furthermore, it is possible to determine which direction the abnormality is in from the color tone. Specifically, if the color tone is an emphasized version of the primary color of R, G, or B that corresponds to the acceleration time history data, or the complementary color of that primary color, it can be determined that an abnormality has occurred in the sensor corresponding to that acceleration time history data.

[0067] Traditionally, the presence or absence of anomalies was confirmed by analyzing the records of individual seismometers, but it can be difficult to determine whether or not an anomaly exists by looking at only one record. In such cases, by comparing the records of seismometers within a building, it becomes possible to determine that a waveform that is clearly different from others is due to some kind of anomaly. The technology according to this embodiment can be said to be a method for efficiently performing this determination by converting acceleration time history data into a single image.

[0068] Next, the operation of the information processing device 10 according to this embodiment will be explained with reference to Figures 5 and 6. Figure 5 is a flowchart showing an example of the flow of information processing according to this embodiment. Information processing according to this embodiment is performed when the CPU 11 executes the information processing program 13A at the timing when an instruction input is received from the user via the input unit 14 to instruct the start of execution. However, it goes without saying that the timing at which the information processing is performed is not limited to this timing.

[0069] In step 100 shown in Figure 5, the CPU 11 acquires acceleration time history data from all seismometers 62 in the building 60 in the X-axis, Y-axis, and Z-axis directions, and for the most recent predetermined past period (5 seconds in this embodiment).

[0070] In step 102, the CPU 11 converts the acquired acceleration time history data into a converted image 70, as described above.

[0071] In step 104, the CPU 11 controls the display unit 15 to display a converted image display screen with a predetermined configuration using the converted image 70 obtained through the above process, and in step 106, the CPU 11 waits until predetermined information is input.

[0072] Figure 6 shows an example of the converted image display screen according to this embodiment. As shown in Figure 6, the converted image display screen according to this embodiment displays the converted image 70 along with a description of the displayed image.

[0073] This allows the user to visually and intuitively determine whether there is an abnormality in the acceleration sensor 64 of the seismometer 62 installed in the building 60, and in which axis direction the abnormality is occurring, by finding the characteristic region shown in Figure 6 from the displayed transformed image 70.

[0074] When the user refers to the converted image 70, they specify the exit button 15A displayed on the converted image display screen via the input unit 14. When the user specifies the exit button 15A, step 106 is determined to be positive and this information processing ends.

[0075] As described above, according to this embodiment, acceleration time history data obtained from an acceleration sensor installed in a building is acquired, and the acquired acceleration time history data is converted into image data by aligning the time axis in one direction and converting the acceleration values ​​into pixel values. Therefore, it does not require other objects or complex calculations such as Fourier transforms, and can detect abnormalities in the acceleration sensor more efficiently compared to conventional technology.

[0076] Furthermore, according to this embodiment, acceleration time history data is data showing the time history of acceleration in the X-axis, Y-axis, and Z-axis directions, which are directions that intersect with each other, and image data is color image data represented by three primary colors. The conversion unit converts the acceleration value in the X-axis direction to the pixel value of the first primary color in the three primary colors of the image data, the acceleration value in the Y-axis direction to the pixel value of the second primary color in the three primary colors of the image data, and the acceleration value in the Z-axis direction to the pixel value of the third primary color in the three primary colors of the image data. Therefore, even when an acceleration sensor that obtains acceleration time history data in three dimensions, namely the X-axis, Y-axis, and Z-axis directions, is applied, abnormalities in the acceleration sensor can be detected more efficiently using a single image.

[0077] Furthermore, according to this embodiment, acceleration sensors are provided on multiple floors of a building, and the conversion unit converts each of the multiple acceleration sensors provided on multiple floors to correspond to a direction intersecting one direction of the image data. Therefore, even for multiple acceleration sensors provided on multiple floors, abnormalities in the acceleration sensors can be detected more efficiently using a single image.

[0078] Furthermore, according to this embodiment, the display unit is controlled to display the image shown by the image data obtained by the conversion unit. Therefore, a person can visually detect the occurrence of an abnormality in the acceleration sensor.

[0079] [Second Embodiment] In the above embodiment, an example of a configuration in which the user is allowed to confirm the abnormality of the acceleration sensor 64 by displaying a converted image 70 was described. However, in this embodiment, an example of a configuration in which the abnormality of the acceleration sensor 64 is automatically detected using a learning model will be described.

[0080] In other words, by accumulating image data of building response acceleration records, i.e., converted image data 70, in addition to visual observation, it is possible to automatically detect anomalies when they occur by using machine learning on this data.

[0081] First, the configuration of the information processing system 90 according to this embodiment will be described with reference to Figure 7. Figure 7 is a block diagram showing an example of the hardware configuration of the information processing system 90 according to this embodiment. Components similar to those in the information processing system 90 according to the first embodiment shown in Figure 1 are denoted by the same reference numerals as in Figure 1, and their descriptions are omitted.

[0082] As shown in Figure 1, the information processing system 90 according to this embodiment differs from the information processing system 90 according to the first embodiment in that the storage unit 13 of the information processing device 10 stores a learning model 13B for evaluating whether or not there is an abnormality in the acceleration sensor 64 using the converted image 70.

[0083] Next, the functional configuration of the information processing system 90 according to this embodiment will be described with reference to Figure 8. Figure 8 is a functional block diagram showing an example of the functional configuration of the information processing system 90 according to this embodiment. Components similar to those in the information processing system 90 according to the first embodiment shown in Figure 2 are denoted by the same reference numerals as in Figure 2, and their descriptions are omitted.

[0084] As shown in Figure 8, the information processing system 90 according to this embodiment differs from the information processing system 90 according to the first embodiment in that an evaluation unit 11D is added to the information processing device 10.

[0085] In this embodiment, the evaluation unit 11D uses the learning model 13B to evaluate from the converted image 70 whether or not there is an abnormality in at least one of the sensors that detect acceleration in each axial direction in the acceleration sensor 64.

[0086] The learning model 13B according to this embodiment will be described in detail below.

[0087] Because it is difficult to predict in advance and pattern the effects on the output waveform from the acceleration sensor 64 due to the occurrence of an anomaly, supervised learning methods are not considered suitable as learning methods to be applied to the learning model 13B.

[0088] In contrast, an autoencoder method that applies unsupervised learning, commonly used in anomaly detection technology, is effective. For example, by training the system with the pattern of the converted image 70 when measurements are performed correctly, it becomes possible to automatically detect any anomaly when it occurs. A major advantage of this approach is that it does not require preparing converted image 70 when an anomaly occurs as training data.

[0089] In this case, the input information to the learning model 13B is the converted image 70, while the output information from the learning model 13B is a converted image 70 without abnormalities (hereinafter referred to as the "model output image"), regardless of whether the input converted image 70 has any abnormalities or not. For this reason, the evaluation unit 11D according to this embodiment evaluates that there is an abnormality in the acceleration sensor 64 corresponding to a given region if the difference in each region of each floor between the converted image 70 input to the learning model 13B and the model output image output from the learning model 13B in response to that input is greater than or equal to a predetermined threshold. Furthermore, if the evaluation unit 11D according to this embodiment evaluates that there is an abnormality in the acceleration sensor 64, it also identifies which axial direction of acceleration in the acceleration sensor 64 is abnormal based on the difference in color tone in the corresponding region of the floor between the converted image 70 and the model output image.

[0090] In this embodiment, the difference is calculated by applying the sum of the differences in pixel values ​​between corresponding pixels in each region of each floor between the converted image 70 and the model output image; however, it goes without saying that this is not the only possible method.

[0091] As described above, the information processing system 90 according to this embodiment uses an autoencoder as the learning model 13B, but is not limited to this. For example, a model based on a CNN (Convolutional Neural Network) may be used as the learning model 13B.

[0092] Next, the operation of the information processing device 10 according to this embodiment will be explained with reference to Figures 9 to 10. Figure 9 is a flowchart showing an example of the information processing flow according to this embodiment. The information processing according to this embodiment is performed when the CPU 11 executes the information processing program 13A at the timing when an earthquake of a predetermined seismic intensity or higher occurs. However, it goes without saying that the timing at which the information processing is performed is not limited to this timing. To avoid confusion, the case where the learning model 13B has already been trained will be explained here.

[0093] In step 200 shown in Figure 9, the CPU 11 acquires acceleration time history data from all seismometers 62 in the building 60, including all acceleration time history data in the X-axis, Y-axis, and Z-axis directions, for the most recent predetermined period (5 seconds in this embodiment).

[0094] In step 202, the CPU 11 converts the acquired acceleration time history data into a converted image 70, as described above.

[0095] In step 204, the CPU 11 inputs the transformed image 70 obtained through the above processing into the learning model 13B. By inputting the transformed image 70 into the learning model 13B, the learning model 13B outputs the model output image described above.

[0096] Therefore, in step 206, the CPU 11 acquires the model output image output from the trained model 13B.

[0097] In step 208, the CPU 11 performs an evaluation using the converted image 70 and the model output image, as described above.

[0098] In step 210, the CPU 11 controls the display unit 15 to display an evaluation result screen with a predetermined configuration using the evaluation results from step 208, and in step 212, the CPU 11 waits until predetermined information is input.

[0099] Figure 10 shows an example of the evaluation results screen according to this embodiment. The example shown in Figure 10 is an example where an abnormality is detected from any of the acceleration sensors 64.

[0100] As shown in Figure 10, the evaluation results screen according to this embodiment displays a message indicating that an abnormality in the acceleration sensor 64 is suspected, along with information that allows for the identification of the acceleration sensor 64, and information indicating the axis direction of the acceleration detected by the sensor suspected of having an abnormality.

[0101] Therefore, by referring to the displayed evaluation results screen, users can grasp this information and take action regarding the acceleration sensor 64 that is presumed to be malfunctioning.

[0102] Upon referring to the evaluation results screen, the user, after understanding the displayed content, selects the exit button 15A displayed on the evaluation results screen via the input unit 14. When the user selects the exit button 15A, step 212 becomes a positive determination and this information processing ends.

[0103] As described above, this embodiment uses a learning model to evaluate the converted image. Therefore, compared to displaying the converted image and having the user determine if the acceleration sensor 64 is abnormal, abnormalities in the acceleration sensor can be detected more efficiently.

[0104] In the embodiments described above, the case of converting acceleration time history data to color image data was explained, but this is not the only possible method. For example, the acceleration time history data may be converted to grayscale image data. In this method, an example can be given in which the acceleration time history data for one detection direction is converted into a single image data. In this method, in order to handle the acceleration time history data for the three directions described above, three image data sets corresponding to different detection directions will be created. In addition, separate from this method, the average value or sum of the acceleration values ​​in the acceleration time history data for each of the three directions described above may be calculated, and this value may be converted into a single image data.

[0105] Furthermore, while the above embodiments describe the case in which all acceleration time history data in the three directions are converted into a single transformed image 70, the invention is not limited to this. For example, one or any combination of the three directions may be converted into a single transformed image 70. In this form, examples include applying the same pixel value to all primary colors among R, G, and B that are not applied.

[0106] Furthermore, in the above embodiment, for example, the hardware structure of the processing unit that executes the acquisition unit 11A, the conversion unit 11B, the display control unit 11C, and the evaluation unit 11D can be any of the following types of processors. As mentioned above, these types of processors include a CPU, which is a general-purpose processor that executes software (programs) and functions as a processing unit, as well as programmable logic devices (PLDs), such as FPGAs (Field-Programmable Gate Arrays), which are processors whose circuit configuration can be changed after manufacturing, and dedicated electrical circuits, such as ASICs (Application Specific Integrated Circuits), which are processors with circuit configurations specifically designed to execute specific processes.

[0107] The processing unit may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the processing unit may consist of a single processor.

[0108] Examples of configuring a processing unit with a single processor include, firstly, a configuration where one or more CPUs and software combine to form a single processor, as is common in client and server computers, and this processor functions as the processing unit. Secondly, a configuration using a processor that realizes the functions of the entire system, including the processing unit, on a single IC (Integrated Circuit) chip, as is common in System-on-a-Chip (SoC) systems. Thus, the processing unit is configured, in terms of hardware structure, using one or more of the above-mentioned types of processors.

[0109] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits, which are combinations of circuit elements such as semiconductor devices. [Explanation of Symbols]

[0110] 10 Information Processing Devices 11 CPU 11A Acquisition Department 11B Conversion Unit 11C Display Control Unit 11D Evaluation Department 12 memory 13 Storage section 13A Information Processing Program 13B Learning Model 14 Input section 15 Display 15A Exit button 16. Media reading / writing device 17 Recording media 18 Communication I / F Section 60 buildings 62 Seismograph 64 Accelerometer 70 Converted Images 90 Information Processing Systems

Claims

1. An acquisition unit that acquires acceleration time history data obtained from acceleration sensors installed in the building, A conversion unit converts the acquired acceleration time history data into image data by aligning the time axis in one direction and converting acceleration values ​​into pixel values. Equipped with an information processing device.

2. The acceleration time history data is data that shows the time history of acceleration in the X-axis direction, Y-axis direction, and Z-axis direction, which are directions that intersect each other. The aforementioned image data is color image data represented by three primary colors. The conversion unit converts the acceleration value in the X-axis direction into the pixel value of the first primary color in the three primary colors of the image data, the acceleration value in the Y-axis direction into the pixel value of the second primary color in the three primary colors of the image data, and the acceleration value in the Z-axis direction into the pixel value of the third primary color in the three primary colors of the image data. The information processing apparatus according to claim 1.

3. The acceleration sensors are installed on each of the multiple floors of the building. The conversion unit performs the conversion by corresponding each of the multiple acceleration sensors provided on the multiple floors to the direction that intersects the image data in one direction. The information processing apparatus according to claim 1 or claim 2.

4. A display control unit controls the display unit to display the image shown by the image data obtained by the conversion unit. The information processing apparatus according to claim 1 or claim 2, further comprising the above.

5. We acquire acceleration time history data obtained from acceleration sensors installed in the building, The acquired acceleration time history data is converted into image data by aligning the time axis in one direction and converting the acceleration values ​​into pixel values. An information processing program that instructs a computer to perform a task.

Images