Building soundness monitoring system

Abstract

To evaluate, with high accuracy, the soundness of respective stories including a story not provided with an acceleration sensor.SOLUTION: A building soundness monitoring system comprises: an earthquake information acquisition unit 5 that acquires acceleration waveform data from an acceleration sensor 2; a section rigidity calculation unit 6 that calculates a primary characteristic frequency of a building and section rigidity; an analysis model generation unit 7 that calculates and sets story rigidity that is the rigidity of stories on the basis of the section rigidity calculated for sections and the weight of the stories, to generate an analysis model; an analysis model update unit 8 that updates the analysis model until the ratio between the primary characteristic frequency and a primary characteristic frequency calculated by executing characteristic value analysis on the analysis model becomes equal to or more than a lower limit threshold and equal to or less than an upper limit threshold, and repeats executing the characteristic value analysis; a maximum deformation angle estimation unit 9 that calculates a deformation waveform on the basis of a result of the last characteristic value analysis and the acceleration waveform data, and calculates the maximum deformation angle; and a structural performance estimation unit 10 that compares the maximum deformation angle with a deformation angle determination threshold to evaluate the soundness.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a building health monitoring system that evaluates the health of a building after an earthquake occurs. [Background technology]

[0002] Various building health monitoring systems have been proposed that can determine the structural performance, i.e., health of a building, such as the degree of damage to the building, after an earthquake occurs, without having to directly inspect the building. For example, Patent Document 1 discloses a configuration for detecting accelerations that occur in a structure due to an earthquake, calculating the inter-story deformation angle of the structure based on the detected acceleration, and diagnosing the structural performance of the structure based on the obtained inter-story deformation angle.

[0003] To accurately assess the soundness of a building, it is necessary to grasp with high precision the degree of deformation caused by the earthquake. To achieve this, it is conceivable to install multiple sensors per building. For example, Patent Document 2 discloses a configuration that includes a sensor that is placed on an observation layer of a building and detects the seismic intensity of an earthquake and the acceleration of vibrations applied to the building from the ground due to the earthquake, and a soundness determination unit that calculates a required shear force coefficient corresponding to the natural period of the building based on response information from the sensor and determines the soundness of the building, and transmits information indicating the determination result to a user's terminal.Patent Document 2 discloses a configuration in which the sensors include a first sensor that detects the acceleration of vibrations applied to the building from the ground and a second sensor installed on the top floor of the building. Patent Document 3 also discloses a building health assessment system that includes an earthquake detection unit that includes a building bottom sensor installed at the bottom of the building to detect acceleration and an earthquake early warning receiver that receives earthquake early warnings, multiple sensors installed at multiple locations on the building to measure the impact of an earthquake on the building at each location, and a health assessment unit that estimates and assesses the health of the building based on the measurement results.

[0004] However, in a building with many floors, such as a super-high-rise building, installing acceleration sensors on every floor may be impractical because it requires securing a location for an acceleration sensor on each floor and increases the total number of acceleration sensors, resulting in high installation costs. Therefore, acceleration sensors may be installed only on a limited number of observation floors among the multiple floors. During an earthquake, damage may occur not only to the lower floors but also to the middle floors of a building. In such cases, if acceleration sensors are installed only on a limited number of observation floors as described above, it may not be possible to accurately evaluate the soundness of floors other than the observation floors, even if there is a problem with the soundness of the floors. Therefore, it is desirable to be able to evaluate the soundness of each floor, including floors without acceleration sensors, with high accuracy from acceleration waveform data obtained from acceleration sensors installed on a limited number of floors of a building. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 2018-59718 [Patent Document 2] Japanese Patent Application Laid-Open No. 2016-75583 [Patent Document 3] Patent No. 6995792 Summary of the Invention [Problem to be solved by the invention]

[0006] The problem that the present invention aims to solve is to provide a building health monitoring system that can evaluate with high accuracy the health of each floor, including floors on which no acceleration sensors are installed, from acceleration waveform data obtained from acceleration sensors installed on a limited number of floors of a building. [Means for solving the problem]

[0007] The present invention employs the following means to solve the above problems: That is, the present invention is a building health monitoring system for evaluating the health of a building having multiple stories after the occurrence of an earthquake, comprising: an earthquake information acquisition unit that acquires acceleration waveform data at each of a plurality of observation stories from acceleration sensors provided at a plurality of observation stories among the plurality of stories; a section stiffness calculation unit that calculates, based on the acceleration waveform data, a first-order natural frequency of the building and a section stiffness that is the stiffness in each section between the plurality of observation stories; an analysis model generation unit that generates an analysis model by calculating and setting, for each of the plurality of stories, the section stiffness calculated for the section that includes that story and the weight of each of the plurality of stories; and a maximum deformation angle estimation unit that calculates displacement waveforms and calculates maximum deformation angles for all the floors based on the result of the final eigenvalue analysis performed by the analytical model update unit and acceleration waveform data for each of the plurality of observed floors. The present invention provides a building health monitoring system comprising: an analytical model updating unit that adjusts the story stiffness to update the analytical model and repeats performing the eigenvalue analysis on the updated analytical model until a frequency ratio, which is the ratio between a first-order natural frequency and the first-order natural frequency calculated by performing eigenvalue analysis on the analytical model, is equal to or greater than a lower threshold and equal to or less than an upper threshold; a maximum deformation angle estimation unit that calculates displacement waveforms and calculates maximum deformation angles for all the floors based on the result of the final eigenvalue analysis performed by the analytical model update unit and acceleration waveform data for each of the plurality of observed floors; and a structural performance estimation unit that compares the maximum deformation angle with a deformation angle judgment threshold to evaluate the health of each of the plurality of floors. According to the above configuration, acceleration waveform data for each of the multiple observation floors is acquired from acceleration sensors installed on a limited number of observation floors among the multiple floors. Based on this data, the first natural frequency of the building and the section stiffness, which is the stiffness of each section between the multiple observation floors, are calculated. Then, for each of the multiple floors, the story stiffness, which is the stiffness of that story, is calculated and set based on the section stiffness calculated for the section including that story and the weight of each of the multiple floors, thereby generating an analytical model. Eigenvalue analysis is performed on the analytical model generated in this way. If the frequency ratio, which is the ratio between the first natural frequency calculated by the section stiffness calculation unit and the first natural frequency calculated by performing eigenvalue analysis on the analytical model, is not equal to or greater than the lower threshold and equal to or less than the upper threshold, the story stiffness is adjusted to update the analytical model, and eigenvalue analysis is repeatedly performed on the updated analytical model until the frequency ratio is equal to or greater than the lower threshold and equal to or less than the upper threshold. In this way, since acceleration sensors are installed only on observed floors, stiffness values ​​are normally calculated only as section stiffness for the sections between observed floors. However, for each of multiple floors, including non-observed floors where acceleration sensors are not installed, the story stiffness of that floor is calculated based on the section stiffness to generate an analytical model. Then, the story stiffness is adjusted based on the results of an eigenvalue analysis of the analytical model, and the analytical model is updated repeatedly. In this way, the story stiffness of each floor is adjusted so that the frequency ratio is above the lower threshold and below the upper threshold. As a result, the story stiffness of each floor is considered to have a value close to the stiffness of the actual building. Furthermore, the results of the eigenvalue analysis performed on the last updated analytical model are, in other words, the results of the eigenvalue analysis performed on the analytical model after the story stiffness adjustment has been completed. Therefore, the results are likely to adequately reflect the actual building characteristics for all floors. Based on the results of this eigenvalue analysis, displacement waveforms and maximum deformation angles are calculated for all floors to evaluate their soundness, allowing for highly accurate evaluation of the soundness of all floors, including floors where acceleration sensors are not installed. In this way, a building health monitoring system can be provided that can evaluate with high accuracy the health of each floor, including floors where no acceleration sensors are installed, from acceleration waveform data obtained from acceleration sensors installed on limited floors of a building.

[0008] In one aspect of the present invention, the analysis model generation unit calculates, for each of the multiple floors, an earthquake story shear force distribution coefficient for that floor based on the weight of each of the multiple floors to calculate the story shear force at that floor, divides the story shear force at that floor by the average of the story shear forces of that floor in the section that includes that floor to calculate the story shear force share ratio at that floor, divides the story height of that floor by the average of the story heights in the section that includes that floor to calculate the story height share ratio at that floor, and calculates the story stiffness of that floor by multiplying the section stiffness calculated for the section that includes that floor by the story shear force share ratio at that floor and the story height share ratio at that floor. According to the above configuration, the analysis model generation unit first calculates the earthquake story shear force distribution coefficient for each of the multiple stories based on the weight of each of the multiple stories, calculates the story shear force at that story, and divides the story shear force at that story by the average story shear force of the stories in the section that the story is included in. As a result, for each of the multiple stories, the extent to which the story bears the story shear force acting on the section that the story is included in among the multiple observation stories is calculated as the story shear force share ratio for that story. Next, for each of the multiple floors, the analysis model generation unit divides the floor height of that floor by the average floor height of the section that that floor is included in. As a result, for each of the multiple floors, the extent to which the floor height of that floor contributes to the height of the section that includes that floor among the multiple observation floors is calculated as the floor height contribution ratio of that floor. The value calculated in this way by multiplying the story shear force contribution rate of each story by the story height contribution rate can be considered to be the stiffness contribution rate, which indicates the extent to which that story contributes to the stiffness in the section between multiple observation stories that includes that story.Based on this idea, the analysis model generation unit can provisionally determine the (initial value of) story stiffness of each story by multiplying the section stiffness calculated for the section that includes that story by the stiffness contribution rate, which is the product of the story shear force contribution rate of that story and the story height contribution rate of that story.

[0009] In another aspect of the present invention, the analysis model generation unit calculates, for each of the multiple floors, an earthquake layer shear force distribution coefficient for that floor based on the weight of each of the multiple floors, calculates the layer shear force at that floor, multiplies the layer shear force at that floor by the sum of the earthquake layer shear force distribution coefficients for the section that includes that floor to calculate a coefficient, and calculates the floor stiffness of that floor by multiplying the section stiffness calculated for the section that includes that floor by the coefficient for that floor. According to the above configuration, the analytical model generation unit first calculates the earthquake story shear force distribution coefficient for each of the multiple stories based on the weight of each of the multiple stories, calculates the story shear force at that story, and then calculates a coefficient by multiplying the story shear force at that story by the sum of the earthquake story shear force distribution coefficients for the section including that story. The coefficient for each story calculated in this way reflects the extent to which that story contributes to the stiffness of the section including that story among the multiple observation stories. Based on this concept, the analytical model generation unit can provisionally determine the story stiffness (initial value) of each of the multiple stories by multiplying the section stiffness calculated for the section including that story by the coefficient for that story.

[0010] In another aspect of the present invention, the frequency ratio is a value obtained by dividing the first natural frequency calculated in the section stiffness calculation unit by the first natural frequency calculated by performing the eigenvalue analysis on the analytical model, and when the frequency ratio is smaller than the lower threshold or larger than the upper threshold, the analytical model update unit adjusts the layer stiffness by multiplying each of the layer stiffnesses set in the analytical model by the frequency ratio. The hierarchical stiffness calculated by the analytical model generation unit is merely a provisional initial value and may not be an accurate value. Therefore, as already explained, the frequency ratio, which is the ratio between the first-order natural frequency calculated by the section stiffness calculation unit and the first-order natural frequency calculated by performing eigenvalue analysis on the analytical model, is calculated, and if the frequency ratio is not equal to or greater than the lower threshold and equal to or less than the upper threshold, the hierarchical stiffness is adjusted to update the analytical model, and eigenvalue analysis is repeatedly performed on the updated analytical model until the frequency ratio is equal to or greater than the lower threshold and equal to or less than the upper threshold. According to the above configuration, the analytical model update unit calculates the frequency ratio by dividing the first natural frequency calculated by the section stiffness calculation unit by the first natural frequency calculated by performing eigenvalue analysis on the analytical model. If this frequency ratio is smaller than the lower threshold or larger than the upper threshold, it is considered that the accuracy of the level stiffness set in the analytical model is insufficient, and as a result, the first natural frequency calculated by performing eigenvalue analysis on the analytical model deviates from the first natural frequency calculated by the section stiffness calculation unit as the correct value. Therefore, to improve the accuracy of the level stiffness, the analytical model update unit multiplies each level stiffness set in the analytical model by the frequency ratio calculated as described above. As a result, if the first natural frequency calculated by performing eigenvalue analysis on the analytical model is larger than the correct first natural frequency, the level stiffness is adjusted to decrease. If the first natural frequency calculated by performing eigenvalue analysis on the analytical model is smaller than the correct first natural frequency, the level stiffness is adjusted to increase. Furthermore, since this adjustment of the story stiffness is performed by multiplying the story stiffness of each of the multiple stories by the same value, i.e., the frequency ratio, adjusting the story stiffness does not change the story stiffness distribution ratio among the stories. This allows for appropriate adjustment of the story stiffness even when the accuracy of the story stiffness is not considered high. [Effects of the Invention]

[0011] According to the present invention, a building health monitoring system can be provided that can evaluate with high accuracy the health of each floor, including floors on which no acceleration sensors are installed, from acceleration waveform data obtained from acceleration sensors installed on limited floors of a building. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a block diagram of a building health monitoring system according to an embodiment of the present invention. [Figure 2] FIG. 2 is a diagram showing a schematic configuration of a building used as an example for explaining the processing in the embodiment. [Figure 3] 3 is a table showing the process for calculating the story shear force distribution ratio of each story for the building shown in FIG. 2. [Figure 4] This is a table showing the flow for calculating the share of floor height and the share of rigidity for each floor of the building shown in Figure 2. [Figure 5] FIG. 3 is an explanatory diagram of an analytical model for the building shown in FIG. 2. [Figure 6] FIG. 2 is a diagram showing a schematic configuration of a building used as an example for explaining the processing in a maximum deformation angle estimating unit in the embodiment. [Figure 7] This is an eigenmode as a result of eigenvalue analysis of the analytical model for the building shown in FIG. 6 after updating by the analytical model update unit. [Figure 8] 10 is a flowchart of a building health monitoring method according to the embodiment. [Figure 9] 3 is a graph showing relative displacement waveforms obtained from the sixth to tenth floors when the above embodiment is applied to the building shown in FIG. 2. [Figure 10] 3 is a graph showing relative displacement waveforms obtained from the first to fifth floors when the above embodiment is applied to the building shown in FIG. 2. [Figure 11] 3 is a table showing a flow of calculating the story stiffness of each story for the building shown in FIG. 2 in a building health monitoring system according to a modified example of the above embodiment. [Figure 12] 10 is a graph showing the layer stiffness calculated in the above modification. DETAILED DESCRIPTION OF THE INVENTION

[0013] The present invention is a building health monitoring system that calculates the response displacement and maximum deformation angle of floors where no acceleration sensors are installed, and evaluates the health of the building using the maximum deformation angle. In this invention, acceleration waveform data for each of the multiple observation stories is acquired from acceleration sensors installed on multiple observation stories among the multiple stories, section stiffness, which is the stiffness in each section between the multiple observation stories, is calculated based on the acceleration waveform data, and the story stiffness of each story is calculated from the section stiffness.An analytical model is then generated based on this story stiffness, eigenvalue analysis is performed on the analytical model, and the story stiffness is adjusted to update the analytical model, repeatedly, until the analytical model approaches a state that fully reflects the properties of the actual building. The above-mentioned story stiffness can be calculated using either the first story stiffness calculation method or the second story stiffness calculation method. Below, as an embodiment, an embodiment using the first story stiffness calculation method will be described, and then as a modified embodiment, an embodiment using the second story stiffness calculation method will be described.

[0014] (Embodiment: First hierarchical stiffness calculation method) Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Fig. 1 is a block diagram of a building health monitoring system according to an embodiment of the present invention, and Fig. 2 is a diagram showing a schematic configuration of a building used as an example for explaining the processing in this embodiment. The building health monitoring system 1 of this embodiment includes a plurality of acceleration sensors 2 and a computing device 4. The building health monitoring system 1 evaluates the health of a multi-story building 20 after an earthquake occurs. The building 20 is constructed on the ground G and has multiple floors in the vertical direction. In this embodiment, the building 20 is described as having 10 floors from the 1st floor to the 11th floor (the 11th floor is the RF, i.e., the rooftop floor), but the number of floors is not limited to this, and it goes without saying that the number of floors of the building 20 may be any number other than 10 as long as it is multiple.

[0015] The acceleration sensor 2 may be, for example, a wireless accelerometer. The acceleration sensor 2 is provided on each of a plurality of observation floors OF selected from among the plurality of floors. In the example of FIG. 2, the acceleration sensors 2 are provided on four floors: the first, fourth, eighth, and rooftop floors. Each of the acceleration sensors 2 may be provided on any floor, but it is preferable that they are provided dispersedly in the vertical direction. Each of the acceleration sensors 2 is provided on, for example, a floor slab of each observation floor OF. When an earthquake occurs, each of the acceleration sensors 2 detects (acquires) the acceleration occurring on the observation floor OF where the acceleration sensor 2 is provided as acceleration data, which is time history information of acceleration. Earthquake information detected by the acceleration sensor 2 is transmitted to a computing device 4 (described later) via an external network or the like, wirelessly or via a wired connection. In this embodiment, one of the acceleration sensors 2 is installed on the first floor, which is the lowest floor. This acceleration sensor 2 installed on the lowest floor indicates earthquake information of seismic waves that directly reach the building 20 during an earthquake, i.e., information on ground motion. From this perspective, the lowest floor of the multiple observation floors OF is particularly referred to as the ground motion observation floor BF. Among the multiple observation floors OF, floors other than the ground motion observation floor BF are referred to as response observation floors RF. In this embodiment, the fourth floor, eighth floor, and rooftop floor correspond to the response observation floors RF, and acceleration sensors 2 are installed on each of these floors. In this way, the multiple observation layers OF include one ground motion observation layer BF and multiple response observation layers RF. As described above, the building 20 of this embodiment is a 10-story building, so the first floor is the ground motion observation floor BF, but if the building 20 has a basement floor, the lowest basement floor may also be the ground motion observation floor BF.

[0016] As such, acceleration sensors 2 are not provided on all floors of building 20, but only on limited observation floors OF. Even in such a state, the building health monitoring system 1 of this embodiment calculates time history displacement waveforms relative to ground motion on all floors (except for the ground motion observation floor BF) and evaluates the building health based on these. To achieve this, when an earthquake occurs, the building health monitoring system 1 of this embodiment acquires earthquake information for each observation floor OF from the acceleration sensors 2, calculates the floor stiffness, which is the stiffness of each floor, based on this information to generate an analytical model, and performs eigenvalue analysis on the analytical model to estimate time history displacement waveforms relative to ground motion on each floor, including each of the non-observation floors NF that are not observation floors OF and do not have acceleration sensors 2. The calculation device 4 includes an earthquake information acquisition unit 5, a section stiffness calculation unit 6, an analysis model generation unit 7, an analysis model update unit 8, a maximum deformation angle estimation unit 9, and a structural performance estimation unit 10.

[0017] When an earthquake occurs, the earthquake information acquisition unit 5 receives and acquires acceleration waveform data detected by the acceleration sensors 2 at each of the plurality of observation levels OF via an external network or the like.

[0018] The section stiffness calculation unit 6 calculates the first natural frequency of the building 20 and the section stiffness, which is the stiffness in each section R between the plurality of observation stories OF, based on the acquired acceleration waveform data. This processing can be performed using a known system identification method. For example, if acceleration sensors 2 are installed on three response observation stories RF as shown in Figure 2, the equation of motion for a shear-type three-mass system is formulated, and displacement, velocity, and acceleration are calculated using the average acceleration method, and then rearranged to formulate a state equation. Next, an output equation is formulated based on the acquired acceleration waveform data. From the state space equation formulated in this way, the first-order natural frequency of the building 20, the section stiffness, which is the stiffness in each section R between the multiple observation stories OF, and the damping constant are calculated. Here, section R refers to the multiple floors extending from the floor immediately above the floor on which acceleration sensor 2 is installed to the floor on which the next acceleration sensor 2 is installed. For example, in the building 20 shown in Fig. 2, acceleration sensors 2 are installed on four floors: the first, fourth, eighth, and rooftop floors. Therefore, the building 20 has section R1, which includes the three floors from the second to fourth floors, section R2, which includes the four floors from the fifth to eighth floors, and section R3, which includes the three floors from the ninth floor to the rooftop floor. That is, in the building 20 shown in Fig. 2, the section stiffness of section R is calculated for each of sections R1, R2, and R3.

[0019] Next, the analytical model generation unit 7 generates an analytical model by calculating and setting the floor stiffness, which is the stiffness of each of the multiple floors, based on the section stiffness calculated for the section R that includes that floor and the weight of each of the multiple floors. In this embodiment, the analytical model generating unit 7 calculates the story stiffness of each of the multiple stories using the first story stiffness calculation method. For this purpose, the analytical model generating unit 7 first calculates the story shear force share ratio for each of the multiple stories. FIG. 3 is a table showing the process for calculating the story shear force share of each story for the building shown in FIG. The building health monitoring system 1 is given the weight Wi of each of the multiple stories. The weight Wi of each story can be obtained, for example, from structural design documents or design drawings. In this embodiment, the weight Wi of each story is the total weight of that story, including the weight of the structure of that story itself and the weight of items loaded on that story. In FIG. 3, weights Wi are set for 10 stories, from the second floor, which is the first story and one story above the ground motion observation story BF, to the rooftop, which is the 10th story and 10 stories above the ground motion observation story BF.

[0020] As shown in Fig. 3 as "sum of weights of floors above the current floor," the analytical model generation unit 7 calculates, for each floor, the sum Si of the weights Wi of the current floor and all floors above it. For example, for the second floor, the sum of the weights Wi of all floors from the second floor to the rooftop floor is calculated. For the rooftop floor, the weight Wi of the rooftop floor is the calculated sum. The analytical model generation unit 7 calculates a standardized weight αi for each story, as indicated by "standardized weight" in Fig. 3. For each story, the standardized weight αi is calculated by dividing the value of the sum of weights Si of the stories above that story by the total weight of the building 20. For example, for the third floor, the standardized weight αi is calculated by dividing the value of the sum of weights Si of all stories above the third floor, including the third floor, which is 3940, by the value of the sum of weights Si of all stories above the second floor, including the second floor, which is 4380.

[0021] The analytical model generation unit 7 calculates the earthquake story shear force distribution coefficient Ai for each story, as indicated as "earthquake story shear force distribution coefficient" in Fig. 3. In this embodiment, the analytical model generation unit 7 calculates the earthquake story shear force distribution coefficient Ai for each story using the following equation (1) based on the first-order natural frequency T obtained by system identification in the section stiffness calculation unit 6 and the normalized weight αi of the story.

number

[0022] The analytical model generation unit 7 calculates the story shear force Qi for each story, as shown as "Story Shear Force" in Figure 3. The analytical model generation unit 7 calculates the story shear force Qi for each story using the total weight Si of all stories above and including that story, and the earthquake story shear force distribution coefficient Ai for that story, using the following formula (2): Qi=Si×Ai×C0×Z×Rt...(2) In the above formula (2), C0 is the standard shear force coefficient, Z is the seismic zoning coefficient, and Rt is the vibration characteristic coefficient. The values ​​of the standard shear force coefficient C0, the seismic zoning coefficient Z, and the vibration characteristic coefficient Rt can be set by inputting them into the building health monitoring system 1 from outside.

[0023] As shown in Fig. 3 as "Average story shear force in section," the analytical model generation unit 7 calculates, for each section R, the average story shear force Qi of each story included in that section R. For example, in section R1, the average story shear force Qi from the second to fourth floors is calculated, in section R2, the average story shear force Qi from the fifth to eighth floors is calculated, and in section R3, the average story shear force Qi from the ninth floor to the rooftop floor is calculated. The analytical model generation unit 7 then calculates the story shear force sharing ratio SQi for each story, as shown in Figure 3 as "Story shear force sharing ratio." For each story, the analytical model generation unit 7 calculates the story shear force sharing ratio SQi for that story by dividing the story shear force Qi for that story by the average story shear force Qi for the section R that includes that story. For example, for the third floor, the story shear force sharing ratio SQi is calculated by dividing the story shear force Qi value for the third floor, 759.45, by the average story shear force Qi value for the section R1 that includes the third floor, 756.67. The story shear force sharing ratio SQi for each story calculated in this way indicates the extent to which the story contributes to sharing the story shear force compared to other stories in the section R that includes that story.

[0024] Next, the analytical model generating unit 7 calculates the share of the story height of each of the multiple stories. FIG. 4 is a table showing the flow for calculating the share of floor height and the share of stiffness for each floor of the building shown in FIG. The building health monitoring system 1 is given the story height Hi of each of the multiple stories. The story height Hi of each story can be obtained, for example, from structural design documents, design drawings, etc. In Fig. 3, story heights Hi are set for 10 stories, from the second floor, which is the first story and one story above the ground motion observation story BF, to the rooftop, which is the 10th story and 10 stories above the ground motion observation story BF.

[0025] As shown in Fig. 4 as "average floor height in section," the analysis model generation unit 7 calculates, for each section R, the average floor height Hi of each floor included in that section R. For example, in section R1, the average floor height Hi of the second to fourth floors is calculated, in section R2, the average floor height Hi of the fifth to eighth floors is calculated, and in section R3, the average floor height Hi of the ninth floor to the rooftop floor is calculated. The analytical model generation unit 7 then calculates the story height share ratio SHi for each story, as shown in Figure 4 as "Story height share ratio." For each story, the analytical model generation unit 7 calculates the story height share ratio SHi for that story by dividing the story height Hi for that story by the average story height Hi for the section R that includes that story. For example, for the third floor, the story height share ratio SHi is calculated by dividing the story height Hi value for the third floor, 3.5, by the average story height Hi value for the section R1 that includes the third floor, 4. The floor height share ratio SHi of each floor calculated in this way indicates how much higher the floor height of that floor is compared to other floors in the section R that includes that floor.

[0026] Furthermore, the analytical model generation unit 7 calculates the stiffness sharing rate SRi of each story, as shown as "stiffness sharing rate" in Fig. 4. For each of the multiple stories, the analytical model generation unit 7 multiplies the story shear force sharing rate SQi of that story, calculated as described above, by the story height sharing rate SHi of that story, to calculate the stiffness sharing rate SRi of that story. In this way, the stiffness share SRi, calculated by multiplying the story shear force share SQi and the story height share SHi for each floor, can be considered to be a value indicating the extent to which the floor in question shares stiffness in the section R that includes that floor between multiple observed floors OF.

[0027] Then, for each of the multiple floors, the analysis model generation unit 7 calculates the floor stiffness, which is the stiffness of the floor, by multiplying the section stiffness calculated by system identification in the section stiffness calculation unit 6 for the section R including that floor by the stiffness sharing rate SRi of that floor. Finally, the analytical model generation unit 7 generates an analytical model M that will be used later for eigenvalue analysis. FIG. 5 is an explanatory diagram of an analytical model for the building shown in FIG. The analytical model generation unit 7 generates an analytical model M having mass points in a number corresponding to the number of floors of the building 20. In the analytical model M, the weight of each mass point is set to the weight Wi of the floor corresponding to that mass point, as shown in FIG. 3. In the analytical model M, the stiffness of each mass point is set to the floor stiffness of the floor corresponding to that mass point, calculated as described above. In the analytical model M, the damping identified by system identification is set as the damping of each mass point. In addition, appropriate boundary conditions are set in the analytical model M. The boundary conditions are basically set to a fixed state, but in a building such as a warehouse, the boundary conditions can be set so as to simulate by pin indication. The analytical model M thus generated is subjected to eigenvalue analysis in the analytical model update unit 8, which will be described next.

[0028] Because acceleration sensors 2 are provided only on a limited number of observation stories OF among the multiple stories, in system identification, stiffness values ​​are obtained not for each story but for the entire section R, which is a collection of multiple stories. In contrast, by the above-described processing in the analysis model generation unit 7, the stiffness value obtained for the entire section R is divided and subdivided into the stiffnesses of each of the stories included in section R. In the analysis model M generated by the analysis model generation unit 7, the story stiffness of the story corresponding to each mass point is set as the stiffness of that mass point, which more closely reflects the characteristics and features of the story. This allows eigenvalue analysis to be performed more accurately than when the section stiffness is set as the stiffness value. The value of the hierarchical stiffness set in the analytical model M is adjusted and updated by the processing described next in the analytical model update unit 8, and the accuracy of the hierarchical stiffness and the analytical model M is improved.

[0029] In this way, the analytical model generation unit 7 generates an analytical model M by calculating and setting the hierarchical stiffness, which is the stiffness of each of the multiple hierarchical levels, based on the section stiffness calculated for the section R that includes that level and the weight Wi of each of the multiple hierarchical levels.

[0030] The analytical model update unit 8 performs eigenvalue analysis on the analytical model M generated by the analytical model generation unit 7. As a result of the eigenvalue analysis, the first natural frequency of the analytical model M is calculated. If the first natural frequency of the building 20 calculated with high accuracy by the section stiffness calculation unit 6 through system identification is taken as the correct value, then there is a possibility that the first natural frequency calculated from the analytical model M will differ from this correct value. If there is a large difference between the first natural frequency as the correct value calculated by the section stiffness calculation unit 6 and the first natural frequency calculated by performing eigenvalue analysis on the analytical model M, the analytical model update unit 8 adjusts the value of the story stiffness of the analytical model M, thereby adjusting the analytical model M so that the value of the first natural frequency obtained when eigenvalue analysis is performed again on the analytical model M will be closer to the correct value.

[0031] For this reason, after the eigenvalue analysis for the analytical model M is completed and the first natural frequency is calculated, the analytical model update unit 8 calculates a frequency ratio, which is the ratio between the first natural frequency as the correct value calculated by the section stiffness calculation unit 6 and the first natural frequency calculated by executing the eigenvalue analysis on the analytical model M. In particular, in this embodiment, the frequency ratio is calculated as a value obtained by dividing the first natural frequency calculated by the section stiffness calculation unit 6 by the first natural frequency calculated by executing the eigenvalue analysis on the analytical model M. The analytical model update unit 8 then compares the frequency ratio with a lower threshold and an upper threshold to determine whether the frequency ratio is equal to or greater than the lower threshold and equal to or less than the upper threshold. Ideally, the first-order natural frequency calculated by performing eigenvalue analysis on the analytical model M should be equal to the first-order natural frequency calculated by the section stiffness calculation unit 6, which is the correct value, so it is desirable that the frequency ratio be 1. Therefore, for example, values ​​of 0.9 and 1.1, or 0.95 and 1.05, etc., can be used as the lower threshold and the upper threshold.

[0032] If the frequency ratio is not equal to or greater than the lower threshold or equal to or less than the upper threshold, but is smaller than the lower threshold or larger than the upper threshold, the analytical model update unit 8 adjusts each of the hierarchical stiffnesses set in the analytical model M. More specifically, the analytical model update unit 8 adjusts the hierarchical stiffness by multiplying each of the hierarchical stiffnesses set in the analytical model M by the frequency ratio. For example, if the first natural frequency calculated by performing eigenvalue analysis on analytical model M is larger than the first natural frequency as the correct value, the frequency ratio will be a value smaller than 1, and this will be multiplied by the story stiffness to adjust the story stiffness so that it is reduced. Therefore, in analytical model M whose story stiffness has been adjusted in this way, the value of the first natural frequency calculated by eigenvalue analysis will be smaller than that of analytical model M before the update, and is likely to be a value close to the correct value of the first natural frequency. Conversely, if the first natural frequency calculated by performing eigenvalue analysis on analytical model M is smaller than the correct first natural frequency, the frequency ratio becomes a value greater than 1, and this is multiplied by the story stiffness to adjust the story stiffness so that it increases. Therefore, in analytical model M whose story stiffness has been adjusted in this way, the value of the first natural frequency calculated by eigenvalue analysis is larger than that of analytical model M before the update, and is likely to be closer to the correct first natural frequency. In this way, the layer stiffness of the analytical model M is adjusted and updated so that the value of the first natural frequency calculated by the eigenvalue analysis comes closer to the correct value.

[0033] In this embodiment, when determining whether the accuracy of the analytical model M has reached a certain level, only the first-order natural frequency is used, and the second-order and higher natural frequencies are not used. This is because, as shown in equation (1), when assuming the Ai distribution, the story shear force is assumed based on an evaluation of only the first order.

[0034] The analytical model updating unit 8 again performs eigenvalue analysis on the analytical model M whose hierarchical stiffness has been adjusted and updated as described above. The analytical model updating unit 8 again compares the frequency ratio calculated as a result with the lower and upper thresholds to determine whether the frequency ratio is equal to or greater than the lower threshold and equal to or less than the upper threshold. If the frequency ratio is not equal to or greater than the lower threshold and equal to or less than the upper threshold, the analytical model updating unit 8 again adjusts each of the hierarchical stiffnesses set in the analytical model in the above manner and updates the analytical model M. In this way, the analytical model update unit 8 adjusts the hierarchical stiffness and updates the analytical model M until the frequency ratio is greater than or equal to the lower threshold and less than or equal to the upper threshold, and repeats the process of performing eigenvalue analysis on the updated analytical model M.

[0035] When the eigenvalue analysis and adjustment of the analysis model M are completed, the frequency ratio is above the lower threshold and below the upper threshold, and the floor stiffness set in the analysis model M is considered to be close to the actual structure of the building 20 from which acceleration waveform data was acquired by the acceleration sensor 2. Furthermore, the results of the final eigenvalue analysis performed on the analysis model M, in which the frequency ratio was determined to be above the lower threshold and below the upper threshold, are considered to be mode waveforms that represent a structure that is close to the actual situation. Therefore, the maximum deformation angle estimator 9 calculates displacement waveforms for all stories based on the results of the final eigenvalue analysis executed by the analytical model updater 8 and the acceleration waveform data for each of the multiple observation stories OF. More specifically, the maximum deformation angle estimator 9 calculates a time history relative displacement response waveform to ground motion as the displacement waveform.

[0036] Fig. 6 is a diagram showing a schematic configuration of a building used as an example to explain the processing in the maximum deformation angle estimation unit. Fig. 7 shows eigenmodes as a result of eigenvalue analysis of the analytical model for the building shown in Fig. 6 after updating in the analytical model update unit. The processing in the maximum deformation angle estimation unit 9 will be described using as an example a building 20A having 39 floors shown in Fig. 6. The observation floors OF are the 1st, 9th, 20th, 30th, and 39th floors. Of these, the 1st floor is the ground motion observation floor BF, and the 9th, 20th, 30th, and 39th floors are the response observation floors RF. Consider a case where the analytical model M, which is generated by the analytical model generation unit 7 and updated by the analytical model update unit 8 as described above based on the acceleration waveform data acquired for this building 20A, is subjected to the last eigenvalue analysis, and a mode waveform as shown in Fig. 7 is obtained. Since the response observation hierarchy RF has four hierarchy levels, mode waveforms up to the fourth order are obtained.

[0037] Specifically, the maximum deformation angle estimation unit 9 first constructs an eigenmode matrix Φ, which is a matrix with N rows and S columns, where N is the number of floors of the building and S is the number of response observation floors RF. For example, in the case of a building 20A shown in Fig. 6, the time history response displacement X of all floors is expressed by the following equation (3) using the eigenmode matrix Φ and the modal response waveform U.

number

[0038] Next, the maximum deflection angle estimation unit 9 calculates the observation layer eigenmode matrix Φ, which is an S-row, S-column matrix for the response observation layer RF. obs The time history response displacement of the observed floor X obs For example, in the case of the building 20A shown in FIG. 6, the observation floor eigenmode matrix Φ obs Using the modal response waveform U, it is expressed by the following equation (4).

number

[0039] By transforming the above equation (4), the following equation (5) is obtained. U=Φ obs -1 X obs ···(5) By substituting the above equation (5) into equation (3), the following equation (6) is obtained. X=ΦΦ obs -1 X obs =TX obs ···(6) In the above equation (6), T is ΦΦ obs -1 is the coefficient matrix, expressed as: The above equation (6) is the time history response displacement X obsis obtained, then the eigenmode matrix Φ and the observation layer eigenmode matrix Φ obs This means that by multiplying the matrix product of this and the matrix product, it is possible to estimate the time history relative displacement response waveform for the ground motion observed story BF at all stories, including the unobserved story NF. Based on this idea, the maximum deformation angle estimator 9 calculates the eigenmode matrix Φ and the observation layer eigenmode matrix Φ obs After calculating, the eigenmode matrix Φ is added to the observation layer eigenmode matrix Φ obs The coefficient matrix T is calculated by multiplying the inverse matrix of The maximum deformation angle estimation unit 9 calculates the time history response displacement X of the observation floor as shown in equation (6). obs Calculate the time history response displacement X of all stories from The maximum deformation angle estimation unit 9 calculates the maximum deformation angle of each story based on the time history response displacement X of all stories calculated in this way.

[0040] The structural performance estimation unit 10 compares the maximum deformation angle calculated by the maximum deformation angle estimation unit 9 with a deformation angle determination threshold for each of the multiple stories, and evaluates the soundness of each story. The structural performance estimation unit 10 outputs the evaluation results to a display device or the like.

[0041] Next, a building health monitoring method using the above-described building health monitoring system 1 will be described with reference to Figures 1 to 7 and 8. Figure 8 is a flowchart of the building health monitoring method. When an earthquake occurs, the earthquake information acquisition unit 5 receives and acquires acceleration waveform data detected by the acceleration sensors 2 at each of the plurality of observation levels OF via an external network or the like (step S1). Based on the acquired acceleration waveform data, the section stiffness calculation unit 6 calculates the first natural frequency of the building 20 and the section stiffness, which is the stiffness in each section R between the plurality of observation stories OF (step S2).

[0042] The analytical model generation unit 7 generates an analytical model for each of the multiple floors by calculating and setting the floor stiffness, which is the stiffness of the floor, based on the section stiffness calculated for the section R that includes that floor and the weight of each of the multiple floors (step S3). The analytical model update unit 8 performs eigenvalue analysis on the analytical model M generated by the analytical model generation unit 7 (step S4). The analytical model update unit 8 calculates a frequency ratio, which is the ratio between the first-order natural frequency as the correct value calculated by the section stiffness calculation unit 6 and the first-order natural frequency calculated by performing eigenvalue analysis on the analytical model M. The analytical model update unit 8 compares the frequency ratio with a lower limit threshold and an upper limit threshold, and determines whether the frequency ratio is equal to or greater than the lower limit threshold and equal to or less than the upper limit threshold (step S5). If the frequency ratio is not equal to or greater than the lower threshold and not equal to or less than the upper threshold, but is smaller than the lower threshold or larger than the upper threshold (No in step S5), the analytical model update unit 8 adjusts each of the hierarchical stiffnesses set in the analytical model to update the analytical model M (step S6). In this case, the process proceeds to step S4, and eigenvalue analysis is performed again on the updated analytical model M. In this way, the analytical model update unit 8 adjusts the hierarchical stiffness and updates the analytical model M until the frequency ratio is greater than or equal to the lower threshold and less than or equal to the upper threshold, and repeats the process of performing eigenvalue analysis on the updated analytical model M.

[0043] If the analytical model update unit 8 determines that the frequency ratio is greater than or equal to the lower threshold and less than or equal to the upper threshold (Yes in step S5), the maximum deformation angle estimation unit 9 calculates displacement waveforms for all floors based on the results of the last eigenvalue analysis performed by the analytical model update unit 8 and the acceleration waveform data for each of the multiple observed floors OF, and calculates the maximum deformation angle for each floor (step S7). The structural performance estimation unit 10 compares the maximum deformation angle calculated by the maximum deformation angle estimation unit 9 with a deformation angle determination threshold for each of the multiple stories, and evaluates the soundness of each story (step S8).

[0044] The building health monitoring system 1 as described above evaluates the health of a building 20 having multiple floors after the occurrence of an earthquake, and includes an earthquake information acquisition unit 5 that acquires acceleration waveform data at each of multiple observation floors OF from acceleration sensors 2 provided at multiple observation floors OF among the multiple floors; a section stiffness calculation unit 6 that calculates the first natural frequency of the building 20 and section stiffness, which is the stiffness at each section R between the multiple observation floors OF, based on the acceleration waveform data; and an analysis model generation unit 7 that generates an analysis model M by calculating and setting the story stiffness, which is the stiffness of each of the multiple floors, based on the section stiffness calculated for the section R including that floor and the weight Wi of each of the multiple floors. The system is equipped with an analytical model updating unit 8 that adjusts the story stiffness to update the analytical model M and repeatedly performs eigenvalue analysis on the updated analytical model M until the frequency ratio, which is the ratio between the primary natural frequency calculated in the section stiffness calculation unit 6 and the primary natural frequency calculated by performing eigenvalue analysis on the analytical model M, becomes equal to or greater than a lower threshold and equal to or less than an upper threshold; a maximum deformation angle estimation unit 9 that calculates displacement waveforms and calculates maximum deformation angles for all stories based on the results of the final eigenvalue analysis performed in the analytical model updating unit 8 and the acceleration waveform data for each of the multiple observed stories OF; and a structural performance estimation unit 10 that compares the maximum deformation angle with a deformation angle judgment threshold for each of the multiple stories to evaluate the soundness. According to the above configuration, acceleration waveform data for each of the observation stories OF is acquired from acceleration sensors 2 installed on a limited number of observation stories OF among the multiple stories. Based on this, the first natural frequency of the building 20 and the section stiffness, which is the stiffness in each section R between the multiple observation stories OF, are calculated. Then, for each of the multiple stories, the story stiffness, which is the stiffness of that story, is calculated and set based on the section stiffness calculated for the section R including that story and the weight Wi of each of the multiple stories, to generate an analytical model M. Eigenvalue analysis is performed on the analytical model M generated in this way. If the frequency ratio, which is the ratio between the first natural frequency calculated by the section stiffness calculation unit 6 and the first natural frequency calculated by performing eigenvalue analysis on the analytical model M, is not equal to or greater than the lower threshold and equal to or less than the upper threshold, the story stiffness is adjusted to update the analytical model M, and eigenvalue analysis is repeatedly performed on the updated analytical model M until the frequency ratio is equal to or greater than the lower threshold and equal to or less than the upper threshold. In this way, because the acceleration sensors 2 are installed only on the observation floors OF, the stiffness value would normally be calculated only as the section stiffness for the section R between the observation floors OF. However, for each of the multiple floors, including the non-observation floors NF where the acceleration sensors 2 are not installed, the story stiffness of that floor is calculated based on the section stiffness to generate an analytical model M. Then, the story stiffness is adjusted based on the results of the eigenvalue analysis performed on the analytical model M, and the analytical model M is updated repeatedly. In this way, the story stiffness of each floor in the last updated analytical model M is adjusted so that the frequency ratio is equal to or greater than the lower threshold and equal to or less than the upper threshold. As a result, it is considered that the story stiffness of each floor has a value close to the stiffness of the actual building 20. Furthermore, the results of the eigenvalue analysis performed on the last updated analytical model M are, in other words, the results of the eigenvalue analysis performed on the analytical model M after the story stiffness adjustment has been completed. Therefore, it is highly likely that the results adequately reflect the characteristics of the actual building 20 on all floors.Based on the results of such eigenvalue analysis, the displacement waveforms and maximum deformation angles are calculated for all floors to evaluate their integrity, so that the integrity of all floors, including those floors where acceleration sensors 2 are not installed, can be evaluated with high accuracy. In this way, a building health monitoring system 1 can be provided that can evaluate with high accuracy the health of each floor, including floors where no acceleration sensors 2 are installed, from acceleration waveform data from acceleration sensors 2 installed on limited floors of a building 20.

[0045] In addition, the analytical model generation unit 7 calculates, for each of the multiple floors, the earthquake story shear force distribution coefficient Ai for that floor based on the weight Wi of each of the multiple floors, to calculate the story shear force Qi for that floor, divides the story shear force Qi for that floor by the average of the story shear forces Qi of the floors in the section R that includes that floor to calculate the story shear force sharing ratio SQi for that floor, divides the story height Hi of that floor by the average of the story heights Hi in the section R that includes that floor to calculate the story height sharing ratio SHi for that floor, and multiplies the section stiffness calculated for the section R that includes that floor by the story shear force sharing ratio SQi for that floor and the story height sharing ratio SHi for that floor, to calculate the story stiffness of that floor. With the above configuration, the analysis model generation unit 7 first calculates the earthquake story shear force distribution coefficient Ai for each of the multiple stories based on the weight Wi of each of the multiple stories, calculates the story shear force Qi at that story, and divides the story shear force Qi at that story by the average story shear force Qi of the stories in the section R that includes that story. In this way, for each of the multiple stories, the extent to which that story bears the story shear force Qi acting on the section R that includes that story among the multiple observation stories OF is calculated as the story shear force sharing ratio SQi at that story. Next, the analytical model generation unit 7 divides, for each of the multiple floors, the floor height Hi of that floor by the average of the floor heights Hi in the section R that includes that floor. As a result, for each of the multiple floors, the extent to which the floor height Hi of that floor contributes to the height of the section R that includes that floor among the multiple observation floors OF is calculated as the floor height contribution ratio SHi of that floor. The value calculated in this way by multiplying the story shear force sharing rate SQi and the story height sharing rate SHi for each story can be considered to be the stiffness sharing rate SRi, which indicates the extent to which the story in question shares stiffness in the section R that includes that story among the multiple observation stories OF. Based on this idea, the analysis model generation unit 7 can provisionally determine (the initial value of) the story stiffness of each story by multiplying the section stiffness calculated for the section R that includes that story by the stiffness sharing rate SRi, which is the value obtained by multiplying the story shear force sharing rate SQi and the story height sharing rate SHi for that story.

[0046] In addition, the frequency ratio is a value obtained by dividing the primary natural frequency calculated in the section stiffness calculation unit 6 by the primary natural frequency calculated by performing eigenvalue analysis on the analysis model M, and if the frequency ratio is smaller than the lower threshold or larger than the upper threshold, the analysis model update unit 8 adjusts the layer stiffness by multiplying each of the layer stiffnesses set in the analysis model M by the frequency ratio. The hierarchical stiffness calculated by the analytical model generating unit 7 is merely a provisional initial value, and may not be an accurate value. Therefore, as already explained, the frequency ratio, which is the ratio between the primary natural frequency calculated by the section stiffness calculating unit 6 and the primary natural frequency calculated by performing eigenvalue analysis on the analytical model M, is calculated, and if the frequency ratio is not equal to or greater than the lower threshold and equal to or less than the upper threshold, the hierarchical stiffness is adjusted to update the analytical model M, and eigenvalue analysis is repeatedly performed on the updated analytical model M until the frequency ratio is equal to or greater than the lower threshold and equal to or less than the upper threshold. According to the above configuration, the analytical model update unit 8 calculates the frequency ratio by dividing the first natural frequency calculated by the section stiffness calculation unit 6 by the first natural frequency calculated by performing eigenvalue analysis on the analytical model M. If this frequency ratio is smaller than the lower threshold or larger than the upper threshold, it is considered that the accuracy of the story stiffness set in the analytical model M is insufficient, and as a result, the first natural frequency calculated by performing eigenvalue analysis on the analytical model M deviates from the first natural frequency calculated by the section stiffness calculation unit 6 as the correct value. Therefore, to improve the accuracy of the story stiffness, the analytical model update unit 8 multiplies each story stiffness set in the analytical model M by the frequency ratio calculated as described above. As a result, if the first natural frequency calculated by performing eigenvalue analysis on the analytical model M is larger than the correct first natural frequency, the story stiffness is adjusted to be reduced. If the first natural frequency calculated by performing eigenvalue analysis on the analytical model M is smaller than the correct first natural frequency, the story stiffness is adjusted to be increased. Furthermore, since this adjustment of the story stiffness is performed by multiplying the story stiffness of each of the multiple stories by the same value, i.e., the frequency ratio, adjusting the story stiffness does not change the story stiffness distribution ratio among the stories. This allows for appropriate adjustment of the story stiffness even when the accuracy of the story stiffness is not considered high.

[0047] (Example of evaluation regarding embodiment) The building health monitoring system 1 of the above embodiment was evaluated. The JMA Kobe waveform was input as the seismic waveform to a model representing a building 20 shown in Figure 2, and system identification was performed using the subspace method based on the acceleration waveform data observed at each observation floor OF. The resulting first-order natural frequency and natural mode were used to estimate the relative displacement waveform of each floor with respect to the ground surface. Fig. 9 is a graph showing relative displacement waveforms obtained from the sixth to tenth floors when the above embodiment is applied to the building shown in Fig. 2. Fig. 10 is a graph showing relative displacement waveforms obtained from the first to fifth floors when the above embodiment is applied to the building shown in Fig. 2. In all layers, the estimated waveforms depicted with the legend "dis_pred" are almost overlapped with the waveforms as the correct values ​​depicted with the legend "dis_true." This shows that the configuration of the above embodiment described as the building health monitoring system 1 has accurately estimated the health of the building 20.

[0048] (Modification of the embodiment: Embodiment using the second hierarchical stiffness calculation method) The building health monitoring system of the present invention is not limited to the above-described embodiment explained with reference to the drawings, and various other modifications are conceivable within the technical scope thereof. For example, a method for calculating the floor stiffness from the section stiffness may be different from that used in the above embodiment. FIG. 11 is a table showing the flow of calculating the story stiffness of each story for the building shown in FIG. 2 in the building health monitoring system according to this modification. In this modification, the analytical model generating unit 7 calculates the story stiffness of each of the multiple stories using the second story stiffness calculation method. In this modification, the analytical model generator 7 calculates the story shear force Qi at each story, as in the above embodiment. Therefore, in Fig. 11, the normalized weight αi, earthquake story shear force distribution coefficient Ai, and story shear force Qi have the same values ​​as in Fig. 3. In Fig. 11, the section stiffness Kb obtained by system identification is shown next to the story shear force Qi.

[0049] In this modification, the analytical model generating unit 7 calculates, for each section R, the sum Bb of the earthquake story shear force distribution coefficients Ai of each story included in the section R. The analytical model generation unit 7 calculates a coefficient βi for each story by multiplying the story shear force Qi at that story by the sum Bb of the earthquake story shear force distribution coefficients Ai in the section R that includes that story. Then, for each story, the analytical model generation unit 7 multiplies the coefficient βi for that story by the section stiffness Kb for the section R that includes that story, to calculate the story stiffness as shown in the rightmost column of Figure 11. The analytical model generating unit 7 generates the analytical model M using the thus calculated hierarchical stiffness as an initial value. The subsequent processing is the same as in the above embodiment.

[0050] In this manner, in this modified example, the analytical model generation unit 7 calculates the earthquake layer shear force distribution coefficient Ai for each of the multiple floors based on the weight Wi of each of the multiple floors, calculates the layer shear force Qi for that floor, multiplies the layer shear force Qi for that floor by the sum Bb of the earthquake layer shear force distribution coefficients Ai for the section R that includes that floor to calculate a coefficient βi, and calculates the floor stiffness of that floor by multiplying the section stiffness Kb calculated for the section R that includes that floor by the coefficient βi for that floor. With the above-described configuration, the analytical model generation unit 7 first calculates the earthquake story shear force distribution coefficient Ai for each of the multiple stories based on the weight Wi of each of the multiple stories, calculates the story shear force Qi for that story, and calculates the coefficient βi by multiplying the story shear force Qi for that story by the sum Bb of the earthquake story shear force distribution coefficients Ai for the section R including that story. The coefficient βi for each story calculated in this way reflects the extent to which that story contributes to the stiffness of the section R including that story among the multiple observation stories OF. Based on this concept, the analytical model generation unit 7 can provisionally determine the story stiffness (initial value) of each of the multiple stories by multiplying the section stiffness Kb calculated for the section R including that story by the coefficient βi for that story.

[0051] Fig. 12 is a graph showing the tier stiffness calculated in this modification. In Fig. 12, line L1 shows the correct tier stiffness value for each tier, and line L2 shows the tier stiffness value calculated as described above. As shown in FIG. 12, the layer stiffness calculated in this modified example is close to the correct value.

[0052] (Other Modifications of the Embodiment) In the above embodiment, when the frequency ratio is smaller than the lower threshold or larger than the upper threshold, the analytical model update unit 8 adjusts the floor stiffness by multiplying the floor stiffness by the same frequency ratio for all floors, but this is not limited to this. When adjusting the floor stiffness, the ratio by which the floor stiffness is multiplied does not have to be the same for all floors, and may be a different value for each section R, for example. In addition to this, it is possible to select and discard the configurations given in the above embodiment and each modified example, or to change them to other configurations as appropriate. [Explanation of symbols]

[0053] 1 Building Health Monitoring System BF Ground Motion Observation Floor 2 Acceleration sensor RF response observation layer 5 Earthquake information acquisition unit NF Non-observation layer 6 Section stiffness calculation section M Analysis model 7 Analysis model generation section R 8 Analysis model update part Wi Weight 9 Maximum deformation angle estimation section Ai Seismic layer shear force distribution coefficient 10 Structural performance estimation section SQi Story shear force distribution ratio 20 Building Hi Floor Height OF Observation Hierarchy

Claims

1. A building health monitoring system that evaluates the health of a multi-story building after an earthquake occurs, comprising: an earthquake information acquisition unit that acquires acceleration waveform data at each of the plurality of observation floors from acceleration sensors provided at the plurality of observation floors among the plurality of floors; a section stiffness calculation unit that calculates a first-order natural frequency of the building and a section stiffness that is stiffness in each section between the plurality of observation stories based on the acceleration waveform data; an analytical model generation unit that generates an analytical model by calculating and setting, for each of the plurality of stories, a story stiffness that is the stiffness of the story based on the section stiffness calculated for the section including that story and the weight of each of the plurality of stories; an analytical model updating unit that adjusts the hierarchical stiffness to update the analytical model and repeatedly performs the eigenvalue analysis on the updated analytical model until a frequency ratio, which is a ratio between the primary natural frequency calculated by the section stiffness calculation unit and the primary natural frequency calculated by performing eigenvalue analysis on the analytical model, becomes equal to or greater than a lower threshold and equal to or less than an upper threshold; a maximum deformation angle estimating unit that calculates displacement waveforms and calculates maximum deformation angles for all of the floors based on the result of the last eigenvalue analysis executed by the analytical model updating unit and acceleration waveform data for each of the plurality of observation floors; a structural performance estimation unit that compares the maximum deformation angle with a deformation angle determination threshold value for each of the plurality of stories to evaluate the soundness; A building health monitoring system comprising:

2. The analysis model generation unit generates, for each of the plurality of layers, calculating a story shear force distribution coefficient for the story based on the weight of each of the plurality of stories, calculating the story shear force at the story, and dividing the story shear force at the story by the average story shear force of the stories in the section to which the story belongs, to calculate the story shear force share at the story; calculating a share ratio of the floor height of the floor by dividing the floor height of the floor by the average floor height of the section including the floor; The section stiffness calculated for the section including the story is multiplied by the story shear force sharing ratio of the story and the story height sharing ratio of the story. The story stiffness of the story is calculated by 2. The building health monitoring system according to claim 1 .

3. The analysis model generation unit generates, for each of the plurality of layers, calculating an earthquake story shear force distribution coefficient for each story based on the weight of each of the plurality of stories, and calculating the story shear force for that story; multiplying the story shear force at the story by the sum of the earthquake story shear force distribution coefficients in the section including the story to calculate a coefficient; The story stiffness of the story is calculated by multiplying the section stiffness calculated for the section including the story by the coefficient of the story.

2. The building health monitoring system according to claim 1 .

4. the frequency ratio is a value obtained by dividing the first natural frequency calculated by the section stiffness calculation unit by the first natural frequency calculated by performing the eigenvalue analysis on the analysis model, When the frequency ratio is smaller than the lower threshold or larger than the upper threshold, the analysis model update unit adjusts the layer stiffness by multiplying each of the layer stiffnesses set in the analysis model by the frequency ratio.

4. A building health monitoring system according to claim 1, wherein:

Images