Temperature measurement result prediction method and temperature measurement result prediction system

Abstract

A temperature measurement result prediction method and a temperature result prediction system are provided that can easily and accurately predict the temperature measurement results of an inspection target structure. [Solution] The method includes a database preparation step S11 for preparing a database 12 in which model structure-related information, which is numerical information, and model structure temperature information, which indicates the temperature distribution at each time of each part calculated by a heat conduction analysis calculation using the model structure-related information, are linked and registered. The method also includes a part temperature estimation step S12 for estimating the temperature distribution at each time of each part of the inspection target structure by comparing the inspection target structure-related information, which is numerical information assumed for the inspection target structure, with the information stored in the database 12. The method also includes a temperature prediction information output step S14 for creating and outputting temperature prediction information for the inspection target structure based on the estimation results in the part temperature estimation step S12.

Description

[Technical Field]

[0001] The present invention relates to a temperature measurement result prediction method and a temperature measurement result prediction system for predicting temperature measurement results before capturing a thermal image on-site when capturing a thermal image showing the temperature distribution on the surface of a structure to be inspected to inspect the condition of the structure. [Background technology]

[0002] Infrastructure structures that are primarily made of concrete, such as bridges and viaducts, may have defects due to aging or poor construction, and in order to prevent serious accidents, it is necessary to detect and address these defects early.

[0003] In recent years, in order to efficiently inspect the numerous infrastructure structures that exist throughout the country, methods have been considered for inspecting defects in structures by capturing images of the structure's surface with a thermal camera mounted on an unmanned aircraft (drone, etc.) and analyzing the captured thermal images. Generally, the surface temperature of a structure has the property that a temperature difference occurs between areas where internal defects exist inside (defective areas) and areas where no internal defects exist (healthy areas). Therefore, by analyzing thermal images, it is possible to estimate the presence or absence of internal defects in the structure, as well as their location and size.

[0004] In order to analyze the condition of a structure with high accuracy, it is preferable that the captured thermal image shows a large temperature difference between the defective area and the healthy area. However, since the surface temperature distribution of a structure changes depending on the season, time, temperature, weather, and solar radiation, even for the same structure, it is important to determine the appropriate time to capture the image.

[0005] For example, in the method for detecting deformation in the surface layer of concrete described in Patent Document 1, an inspector goes to the site and measures the temperature of the surface layer of concrete to be inspected, calculates the rate of change of temperature per unit time from the measurement results, and determines the timing for capturing a thermal image based on this. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Patent No. 4448553 Summary of the Invention [Problem to be solved by the invention]

[0007] Infrastructure structures that require inspection are often scattered across a wide area, so it takes a lot of time and effort for an inspector to simply travel from an office to the site. Therefore, it is preferable that the inspector be able to decide the best time to capture a thermal image while remaining in the office.

[0008] However, methods such as the method for detecting deformation in the surface layer of concrete disclosed in Patent Document 1 require the inspector to go to the site and measure the surface temperature of the surface layer of concrete before capturing a thermal image, which places a heavy burden on the inspector.

[0009] The inventors of the present application thought that if it were possible to predict in advance what kind of temperature measurement results would be obtained for infrastructure structures, etc. to be inspected, depending on the type of defect expected, the season and time when thermal images are to be taken, and the temperature, weather, and solar radiation conditions, the prediction results could be used in a variety of ways. This would make it easy to determine the best time to take images of infrastructure structures, etc., while sitting in an office, etc.

[0010] The present invention has been made in consideration of the above-mentioned background art, and aims to provide a temperature measurement result prediction method and a temperature result prediction system that can easily and accurately predict the temperature measurement results of an inspected structure. [Means for solving the problem]

[0011] The present invention is a method executed by a computer system, in which, when inspecting the state of an inspection target structure based on a thermal image of the surface of the inspection target structure, a temperature measurement result prediction method is used to predict a temperature measurement result before the thermal image is taken on-site, the method comprising: Model structure-related information, which is numerical information about a virtual model structure that simulates the inspection target structure, and information calculated by a thermal conduction analysis using the model structure-related information, It is made up of multiple parts It includes a defect area inside which an internal defect exists and a healthy area other than the defect area. Multiple parts The temperature distribution of the model structure is shown at each time point. change a database preparation step of preparing a database in which information is registered in a mutually linked manner; The inspection target structure related information, which is assumed information about the inspection target structure and is numerical information corresponding to the model structure related information, is compared with the information stored in the database, thereby determining the inspection target structure including the defective area and the sound area. Multiple parts a temperature estimation step of estimating the temperature distribution at each time point; The temperature measurement result prediction method includes a temperature prediction information output step of creating and outputting temperature prediction information for the inspection target structure based on the estimation results in the each part temperature estimation step.

[0012] The model structure-related information preferably includes at least structural information indicating the structural characteristics of the model structure, environmental temperature information indicating the characteristics of the environmental temperature at the installation location of the model structure, site information indicating the characteristics related to the amount of solar radiation on the top surface of the model structure, and defect information indicating the presence or absence of internal defects in the model structure and the characteristics of the internal defects. Furthermore, the each part temperature estimating step may be configured to obtain inspection target structure summary information, which is text information about the inspection target structure, and to derive the inspection target structure-related information based on the inspection target structure summary information.

[0013] In the each-part temperature estimation step, a each-part temperature estimation model created by machine learning using the model structure-related information and the model structure temperature change information registered in the database as training data is prepared, and the inspection target structure-related information is input to the each-part temperature estimation model, thereby estimating the temperature of the inspection target structure including the defective area and the sound area. Multiple parts In this case, the temperature estimation model for each part uses the model structure related information registered in the database as explanatory variables, and estimates the temperature distribution for each time of the defective area and the sound area of the inspection target structure. Multiple parts It is preferable that the model formula is created using a multiple regression analysis technique, with the temperature distribution at each time as the response variable.

[0014] The method may further include a time derivation step of deriving a time suitable for capturing the thermal image of the inspection target structure based on the temperature difference between the defective area and the healthy area calculated from the estimation result in the each-part temperature estimation step, and the temperature prediction information output step may create recommended time information for the inspection target structure based on the derivation result in the time derivation step and output the recommended time information together with the temperature prediction information. In this case, it is preferable that the time derivation step derives a time suitable for capturing the thermal image of the inspection target structure based on the temperature difference between the defective area and the healthy area calculated from the estimation result in the each-part temperature estimation step and the temperature gradient at the boundary between the defective area and the healthy area.

[0015] The present invention also provides a temperature measurement result prediction system, which is a computer system that predicts temperature measurement results before capturing a thermal image on-site when inspecting the state of a structure to be inspected based on a thermal image of the surface of the structure to be inspected, comprising: Model structure-related information, which is numerical information about a virtual model structure that simulates the inspection target structure, and information calculated by a thermal conduction analysis using the model structure-related information, It is made up of multiple partsIt includes a defect area inside which an internal defect exists and a healthy area other than the defect area. Multiple parts The temperature distribution of the model structure is shown at each time point. change A database in which information is linked and registered, The inspection target structure related information, which is assumed information about the inspection target structure and is numerical information corresponding to the model structure related information, is compared with the information stored in the database, thereby determining the inspection target structure including the defective area and the sound area. Multiple parts a temperature estimation unit for estimating the temperature distribution at each time; The temperature measurement result prediction system includes a temperature prediction information output unit that creates and outputs temperature prediction information for the inspection target structure based on the estimation results from the temperature estimation units for each part.

[0016] The model structure-related information preferably includes at least structural information indicating the structural characteristics of the model structure, environmental temperature information indicating the characteristics of the environmental temperature at the installation location of the model structure, site information indicating the characteristics related to the amount of solar radiation on the top surface of the model structure, and defect information indicating the presence or absence of internal defects in the model structure and the characteristics of the internal defects. Furthermore, the each-part temperature estimating unit may be configured to acquire inspection target structure summary information, which is text information about the inspection target structure, and to derive the inspection target structure-related information based on the inspection target structure summary information.

[0017] The each-part temperature estimation unit has a each-part temperature estimation model created by machine learning using the model structure-related information and the model structure temperature change information registered in the database as training data, and by inputting the inspection target structure-related information into the each-part temperature estimation model, it is possible to estimate the defect area and the sound area of the inspection target structure. Multiple parts The temperature distribution at each time can be estimated.

[0018] Furthermore, the device may be configured to include a time derivation unit that derives a time appropriate for capturing the thermal image of the inspection target structure based on the temperature difference between the defective area and the healthy area calculated from the estimation results of the temperature estimation unit for each part, and the temperature prediction information output unit may be configured to create recommended time information for the inspection target structure based on the derivation results of the time derivation unit and output it together with the temperature prediction information. [Effects of the Invention]

[0019] According to the temperature measurement result prediction method and temperature measurement result prediction system of the present invention, users such as inspectors can easily obtain predicted values (temperature prediction information) of the temperature distribution of the structure being inspected while sitting in their offices. Moreover, since the prediction is based on the calculation results of heat conduction analysis, theoretical and highly reliable predicted values can be obtained. Furthermore, by configuring the system to output information on the best time to capture a thermal image (recommended time information), the burden on inspectors can be significantly reduced compared to using the technology of Patent Document 1. [Brief explanation of the drawings]

[0020] [Figure 1] 1A is an overall flowchart showing one embodiment of a temperature measurement result prediction method of the present invention, and FIG. 1B is a block diagram showing the flow of information when a database is created in the database preparation step. [Figure 2] FIG. 1A is a block diagram showing the flow of information in the temperature estimation step for each part, and FIG. 1B is a block diagram showing a method for creating a temperature estimation model for each part. [Figure 3] 10 is a diagram illustrating an example of model structure related information. [Figure 4] 10 is a table showing an example of model structure temperature change information. [Figure 5] 10 is a diagram showing an example of inspection target structure summary information. [Figure 6] 10 is a table showing an example of inspection target structure related information; [Figure 7] 10A is a diagram showing an example of temperature prediction information output in a temperature prediction information output step, and FIG. 10B is a diagram showing an example of recommended time information. [Figure 8] 10 is a chart showing another example of recommended time information output in the temperature prediction information output step. [Figure 9] 1 is a system configuration diagram showing an embodiment of a temperature measurement result prediction system of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0021] <<Temperature measurement prediction method / temperature measurement result prediction system 10 according to an embodiment>> An embodiment of a temperature measurement result prediction method and a temperature measurement result prediction system of the present invention will be described below with reference to the drawings. The temperature measurement prediction method of this embodiment is a method for predicting temperature measurement results before capturing a thermal image on-site when inspecting the condition of a target structure based on a thermal image of the surface of the target structure, and as shown in Figure 1(a), it is composed of a database preparation step S11, a temperature estimation step S12 for each part, a time derivation step S13, and a temperature prediction information output step S14.

[0022] A temperature measurement result prediction system 10, which is one embodiment of the temperature measurement result prediction system of the present invention, is a computer system that executes the temperature measurement prediction method shown in Figure 1(a). As shown in Figure 9, the temperature measurement result prediction system 10 includes a database 12 and a part temperature estimation unit 14, which is a functional block that executes part temperature estimation step S12. It also includes a time derivation unit 16, which is a functional block that executes time derivation step S13, and a temperature prediction information output unit 18, which is a functional block that executes temperature prediction information output step S14. In addition, it includes a display device 20 such as a display, and an information input device 22, such as a keyboard or mouse, that is operated by the user.

[0023] The temperature measurement result prediction method and temperature measurement result prediction system of this embodiment will be explained below in accordance with the overall flowchart of Figure 1(a), assuming that the structure to be inspected is a bridge or viaduct where asphalt pavement is laid on the top surface of a concrete deck to form a walkway.

[0024] In this specification, a thermal image refers to an image composed of a large number of aligned pixels, each of which is assigned temperature data and coordinate data. When analyzing a thermal image, it is treated as image data (a collection of temperature data and coordinate data for each pixel) rather than as an image that provides visual information. <Database preparation step S11 / Database 12> In the database preparation step S11, a database 12 is prepared in which model structure related information and model structure temperature change information are linked to each other and registered, as shown in Fig. 1(b). The model structure is a virtual structure that simulates the structure to be inspected, and is a simulation model for heat conduction analysis.

[0025] First, the model structure-related information will be described. The model structure-related information is numerical information for performing a heat conduction analysis simulation of the model structure, and includes at least structural information indicating the structural characteristics of the model structure, environmental temperature information indicating the characteristics of the environmental temperature at the installation location of the model structure, site information indicating the characteristics of the amount of solar radiation on the top surface of the model structure, and defect information indicating the presence or absence of internal defects in the model structure and the characteristics of the internal defects.

[0026] In the example shown in Figure 3, the thickness of asphalt and concrete, the thermal conductivity of asphalt and concrete, etc. are set as structural information items, and the outdoor temperature data for each month and time are set as environmental temperature information items. For outdoor temperature data, it is advisable to use data for an appropriate location published by the Japan Meteorological Agency, for example.

[0027] Additionally, the site information items include the asphalt's solar absorptivity value and data on global solar radiation by month and time. For global solar radiation data, it is recommended to use data for an appropriate location published by the Japan Meteorological Agency, for example. The asphalt's solar absorptivity and global solar radiation values are used to calculate the equivalent outside temperature required for heat conduction analysis. Furthermore, if the presence of internal defects such as voids is assumed, values such as the depth position, thickness, length, and thermal conductivity of the air layer are set as defect information items.

[0028] Next, the model structure temperature change information will be described. The model structure temperature change information is calculated by performing a heat conduction analysis simulation based on the model structure related information. The model structure is divided into a plurality of parts by predetermined regions. As shown in Figure 4, the temperature distribution data of the model structure Multiple parts of The temperature was calculated for each month and time.

[0029] The calculated temperature distribution data includes at least the temperature distribution on the surface of the model structure. It is made up of multiple parts , the defect area where an internal defect exists inside and the healthy area which is the area other than the defect area. Multiple parts Temperature data is included. If there is an internal defect such as a void inside the model structure, the temperature difference between the defective area and the sound area on the surface of the structure will be large, and the isotherms at the boundary between the two areas will tend to be dense, so the degree of density of the isotherms becomes important information.

[0030] In this way, the model structure-related information shown in Figure 3 and the model structure temperature change information shown in Figure 4 are registered in the database 12 in a linked manner, and this information serves as training data when creating a temperature estimation model 14a (machine learning model) for each part, which will be described later. Here, "preparing the database 12" means creating the database 12 and storing it in the system, making the database 12 already stored in the system usable, making the information in the database 12 stored in another device or storage medium available for retrieval, etc.

[0031] <Each Part Temperature Estimation Step S12 / Each Part Temperature Estimation Unit 14> In the temperature estimation step S12, data on the temperature distribution at each time point in each part of the inspection target structure is estimated by comparing the inspection target structure related information, which is numerical information assumed for the inspection target structure, with the information stored in the database 12. Fig. 6 shows an example of the inspection target structure related information, and each item of the inspection target structure related information corresponds to each item of the model structure related information shown in Fig. 3.

[0032] Although the information related to the inspection target structure [= numerical information] is basically provided by the user, some items have values that are difficult to identify unless the user has specialized knowledge. Therefore, in order to reduce the burden on the user, in this embodiment, the user provides the inspection target structure summary information, which is text information [= character information or numerical information], as shown in Figure 2(a), and a process is performed to derive the inspection target structure related information from the provided inspection target structure summary information.

[0033] To explain this point, let us compare the inspection target structure summary information shown in Figure 5 with the inspection target structure related information shown in Figure 6. For example, the value of the "Asphalt Thickness" item in the inspection target structure related information can be relatively easily identified even by a user without specialized knowledge, so the inspection target structure summary information [= numerical information] of "25" is obtained from the user and used as is as the inspection target structure related information [= numerical information].

[0034] On the other hand, for example, it is difficult for a user to identify the value of the "thermal conductivity of asphalt" item in the information related to the structure to be inspected unless they have specialized knowledge. Therefore, the system obtains from the user the general information (=text information) of the structure to be inspected, such as "standard (density of asphalt)" and "low moisture content (moisture content of asphalt)," and based on this, automatically derives the information (=numerical information) related to the structure to be inspected, such as "1.45 (thermal conductivity of asphalt)." The derivation method is not particularly limited, but one possible method is to prepare a conversion table in advance that converts specific text information into specific numerical information.

[0035] For example, the "outside temperature by time" item in the information related to the structure to be inspected can be determined by the user by examining past data from the Japan Meteorological Agency, but examining each and every one is tedious. Therefore, the system obtains from the user general information (=text information) about the structure to be inspected, such as "April (scheduled inspection month)" and "XX district of Toyama City (installation location)," and based on this, automatically derives the information related to the structure to be inspected, which is the outside temperature data by time (=numeric information). While the derivation method is not particularly limited, one possible method is to store temperature data from all over the country published by the Japan Meteorological Agency in a database in advance, automatically search the database based on the text information, and extract the relevant numerical information.

[0036] In addition, to reduce the burden on the user, items such as "solar radiation absorption rate of asphalt," "total solar radiation by time," and "depth position of air layer" among the information related to the structure under inspection are also included. Text information about the structure under inspection is obtained from the user, and based on this, information related to the structure under inspection [= numerical information] is automatically derived.

[0037] If the user can specify the numerical values for all items, the user may provide information [=numerical information] related to the inspection target structure for all items.

[0038] Next, we will explain the process of comparing the above-mentioned information related to the inspection target structure with the model structure related information and model structure temperature change information stored in database 12, and estimating the temperature distribution at each time point in each part of the inspection target structure.

[0039] In this embodiment, as shown in FIG. 2(a), a temperature estimation model 14a for each part is prepared, and information related to the inspection target structure is input to the temperature estimation model 14a, thereby estimating the temperature of the inspection target structure. Multiple partsThe temperature distribution data for each time is estimated. As shown in FIG. 2(b), the temperature estimation model 14a for each part is a machine learning model created by machine learning using the model structure related information and the model structure temperature change information stored in the database 12 as teacher data. The machine learning algorithm is not particularly limited, but for example, the model structure related information is used as an explanatory variable, and the temperature distribution data for each part is created by machine learning using the model structure related information and the model structure temperature change information as teacher data. Areas are multiple parts into which the model structure is divided. A model formula created using multiple regression analysis, with the temperature distribution at each time as the objective variable, can be used. Other methods such as decision trees, random forests, and neural networks can also be used.

[0040] The "temperature distribution at each time point for each part of the inspected structure" output by the each part temperature estimation model 14a is actually a collection of a huge number of numerical data, but in the subsequent temperature prediction information output step S14, it is organized into a format that is easy for the user to understand, as shown in Figure 7(a), and output as temperature prediction information.

[0041] Incidentally, "preparing each part temperature estimation model 14a" means creating each part temperature estimation model 14a and storing it in the system, making each part temperature estimation model 14a already stored in the system usable, making each part temperature estimation model 14a stored in another device or storage medium usable, etc.

[0042] <Time derivation step S13 / time derivation unit 16> In the time derivation step S13, a time suitable for capturing a thermal image of the inspection target structure is derived based on the results of estimation in the temperature estimation step S12. Specifically, first, the temperature difference between the defective area and the healthy area is derived from the results of estimation in the temperature estimation step S12. For example, as shown in the two graphs in Figure 7(b), the average temperature of the defective area and the average temperature of the healthy area are calculated for each time to determine the temperature difference.

[0043] As mentioned above, in order to analyze the condition of a structure (presence or absence of internal defects, size of internal defects, etc.) with high accuracy, it is preferable that the captured thermal image shows a large temperature difference between the defective area and the healthy area. Therefore, it can be said that the best time to capture a thermal image is the time when the temperature difference between the defective area and the healthy area is large. Therefore, in Figure 7(b), the time between 1:00 PM and 5:00 PM, when the temperature difference exceeds a predetermined reference value Tth, is determined to be the best time to capture a thermal image.

[0044] Furthermore, when deriving the appropriate time to capture a thermal image, it is preferable to also consider the temperature gradient at the boundary between the defective area and the healthy area. This is because even if the temperature difference is relatively large, if the temperature gradient is small, the characteristics of the boundary will be blurred. Therefore, in order to analyze the condition of a structure (the presence or absence of internal defects, the size of internal defects, etc.) with high accuracy, it is preferable that the captured thermal image is one in which a large temperature gradient is apparent.

[0045] There are several methods for quantifying the temperature gradient at the boundary between the defective area and the healthy area, but here, as shown in the third and fourth graphs from the top of Figure 8, the temperature gradient is quantified as the temperature change rate Rp=Ta / Tb. Ta is the difference between the "main temperature of the defective area" and the "temperature at the boundary between the defective area and the healthy area," and Tb is the difference between the "main temperature of the defective area" and the "main temperature of the healthy area." The temperature change rate Rp increases when the temperature gradient is large and decreases when the temperature gradient is small.

[0046] The best time to take a thermal image is when the temperature change rate Rp at the boundary between the defective area and the healthy area is large. Therefore, in the middle part of Figure 8 (fourth graph from the top), the time between 10:00 and 15:00, when the change rate Rp exceeds the specified reference value Rth, is determined to be the best time to take a thermal image.

[0047] In Figure 8, it is concluded that the time period between 13:00 and 17:00, extracted based on the temperature difference, and the time period between 10:00 and 15:00, extracted based on the temperature gradient (rate of change Rp), is the most suitable time to capture thermal images.

[0048] In this way, in the time derivation step S13, the appropriate time for capturing a thermal image is derived by analyzing the temperature difference between the defective area and the healthy area (Figure 7(b)), or by analyzing the temperature gradient at the boundary in addition to the temperature difference (Figure 8).

[0049] It should be noted that the time (numerical data) is derived in the time derivation step S13, but in the subsequent temperature prediction information output step S14, it is organized into a format that is easy for the user to understand, as shown in Figures 7(b) and 8, and output as recommended time information.

[0050] <Temperature prediction information output step S14 / Temperature prediction information output unit 18> In temperature prediction information output step S14, temperature prediction information as shown in Fig. 7(a) is created and output based on the estimation results in each part temperature estimation step S12. Also, recommended time information as shown in Fig. 7(b) or Fig. 8 is created based on the derivation results in time derivation step S13, and output together with the temperature prediction information. Incidentally, "output" means storing in a storage device, displaying on a display device 20 such as a display, printing on paper using a printer or the like, etc.

[0051] <Summary of this embodiment> According to the temperature measurement result prediction method and temperature measurement result prediction system 10 of this embodiment, a user such as an inspector can easily obtain a predicted value (temperature prediction information) of the temperature distribution of a structure to be inspected while sitting in an office or the like. Moreover, since the prediction is based on the calculation results of a heat conduction analysis, a theoretical and highly reliable predicted value can be obtained. In addition, information on the best time to capture a thermal image (recommended time information) is also output, which significantly reduces the burden on the inspector compared to using the technology of Patent Document 1.

[0052] <<Other embodiments and modifications>> The temperature measurement prediction method and temperature measurement result prediction system of the present invention are not limited to the above-described embodiment. For example, the above-described each-part temperature estimation step S12 and each-part temperature estimation unit 14 are configured to estimate each part using a machine learning model (each-part temperature estimation model 14a), but they may be configured to estimate using a method other than machine learning. Furthermore, the temperature measurement prediction method and temperature measurement result prediction system 10 of the above-described embodiment are configured to output information on a time suitable for capturing a thermal image (recommended time information), but if the recommended time information is not required, the time derivation step S13 and the time derivation unit 16 can be omitted.

[0053] The items and specific values of the model structure related information, inspection target structure summary information, and inspection target structure related information shown in Figures 3, 5, and 6 are merely examples and may be changed as appropriate to suit the method of heat conduction analysis simulation and the characteristics of the structure.

[0054] In addition, the present invention can be applied to structures to be inspected as long as they are primarily made of concrete, and is not limited to bridges and viaducts in which asphalt pavement is laid on the top surface of a concrete slab to form a walkway, as in the above embodiment. For example, the present invention can be applied to inspect a variety of structures, such as concrete slabs not covered with asphalt, wall parapets, and buildings with exterior wall tiles attached to a concrete body. [Explanation of symbols]

[0055] 10 Temperature measurement result prediction system 12 Databases 14 Temperature estimation section 16 Time derivation part 18 Temperature prediction information output section S11 Database Preparation Steps S12 Temperature estimation step for each part S13 Time derivation step S14 Temperature prediction information output step

Claims

1. A method executed by a computer system for predicting a temperature measurement result when inspecting a condition of a structure to be inspected based on a thermal image of the surface of the structure to be inspected, the method predicting a temperature measurement result before the thermal image is captured on-site, comprising: a database preparation step of preparing a database in which model structure-related information, which is numerical information about a virtual model structure simulating the inspection target structure, and model structure temperature change information, which is information calculated by a heat conduction analysis calculation using the model structure-related information and indicates the temperature distribution over time in multiple parts of the surface of the model structure, including defective areas where internal defects exist inside and healthy areas other than the defective areas; and a temperature estimation step for estimating temperature distributions at each time in a plurality of parts of the inspection target structure, including the defective area and the sound area, by comparing inspection target structure-related information, which is assumed information about the inspection target structure and is numerical information corresponding to the model structure-related information, with information stored in the database; A temperature measurement result prediction method characterized by comprising a temperature prediction information output step of creating and outputting temperature prediction information for the inspection target structure based on the estimation results in the each part temperature estimation step.

2. A temperature measurement result prediction method as described in claim 1, wherein the model structure-related information includes at least structural information indicating the structural characteristics of the model structure, environmental temperature information indicating the characteristics of the environmental temperature at the installation location of the model structure, site information indicating the characteristics related to the amount of solar radiation on the top surface of the model structure, and defect information indicating the presence or absence of internal defects in the model structure and the characteristics of the internal defects.

3. A temperature measurement result prediction method as described in claim 1, wherein in the temperature estimation step for each part, inspection target structure summary information, which is text information about the inspection target structure, is obtained, and the inspection target structure related information is derived based on the inspection target structure summary information.

4. The method for predicting temperature measurements according to claim 1, wherein the temperature estimation step includes preparing a temperature estimation model for each part created by machine learning using the model structure-related information and the model structure temperature change information registered in the database as training data, and inputting the information related to the inspected structure into the temperature estimation model for each part to estimate the temperature distribution at each time in multiple parts of the inspected structure, including the defective area and the healthy area.

5. The temperature measurement result prediction method described in claim 4, wherein the temperature estimation model for each part is a model formula created using a multiple regression analysis technique in which the model structure-related information registered in the database is used as an explanatory variable and the temperature distribution over time of multiple parts of the inspected structure, including the defective area and the healthy area, is used as a target variable.

6. a time derivation step of deriving a time suitable for capturing the thermal image of the inspection target structure based on a temperature difference between the defective area and the sound area calculated from the estimation result in the each-part temperature estimation step; A temperature measurement result prediction method described in any one of claims 1 to 5, wherein in the temperature prediction information output step, recommended time information for the inspection target structure is created based on the derived result in the time derivation step, and output together with the temperature prediction information.

7. A temperature measurement result prediction method as described in claim 6, wherein the time derivation step derives a time suitable for capturing the thermal image of the inspected structure based on the temperature difference between the defective area and the healthy area calculated from the estimation results in the each part temperature estimation step, and the temperature gradient at the boundary between the defective area and the healthy area.

8. A temperature measurement result prediction system comprising a computer system for predicting temperature measurement results before capturing a thermal image on-site when inspecting the state of a structure to be inspected based on a thermal image of the surface of the structure to be inspected, comprising: a database in which model structure-related information, which is numerical information about a virtual model structure that simulates the inspection target structure, and model structure temperature change information, which is information calculated by a heat conduction analysis calculation using the model structure-related information and indicates the temperature distribution over time in multiple parts of the surface of the model structure, including defective areas where internal defects exist inside and healthy areas other than the defective areas, are linked and registered; a temperature estimation unit for each part that estimates the temperature distribution at each time of a plurality of parts of the inspection target structure, including the defective area and the sound area, by comparing inspection target structure-related information, which is assumed information about the inspection target structure and is numerical information corresponding to the model structure-related information, with information stored in the database; A temperature measurement result prediction system characterized by comprising a temperature prediction information output unit that creates and outputs temperature prediction information for the inspection target structure based on the estimation results from the each part temperature estimation unit.

9. The temperature measurement result prediction system described in claim 8, wherein the model structure related information includes at least structural information indicating the structural characteristics of the model structure, environmental temperature information indicating the characteristics of the environmental temperature at the installation location of the model structure, site information indicating the characteristics related to the amount of solar radiation on the top surface of the model structure, and defect information indicating the presence or absence of internal defects in the model structure and the characteristics of the internal defects.

10. The temperature measurement result prediction system of claim 8, wherein the temperature estimation unit for each part acquires inspection target structure summary information, which is text information about the inspection target structure, and derives the inspection target structure related information based on the inspection target structure summary information.

11. The temperature measurement result prediction system of claim 8, wherein the temperature estimation unit for each part has a temperature estimation model for each part created by machine learning using the model structure-related information and the model structure temperature change information registered in the database as training data, and by inputting the information related to the inspection target structure into the temperature estimation model for each part, estimates the temperature distribution at each time of multiple parts of the inspection target structure, including the defective area and the healthy area.

12. a time derivation unit that derives a time appropriate for capturing the thermal image of the inspection target structure based on a temperature difference between the defective area and the sound area calculated from the estimation result of the each-part temperature estimation unit; A temperature measurement result prediction system as described in any one of claims 8 to 11, wherein the temperature prediction information output unit creates recommended time information for the inspection target structure based on the derivation result in the time derivation unit, and outputs it together with the temperature prediction information.

Images