Coating deterioration estimation device, coating deterioration estimation method, and program

Abstract

The objective is to provide a technology for estimating the deterioration of paint applied to structures using machine learning models. [Solution] One aspect of the present disclosure relates to a paint film deterioration estimation device, comprising: an acquisition unit that acquires painting data of a structure, paint film measurement data for the structure, and weather data for the area where the structure is installed; and a processing unit that uses a paint film deterioration estimation model to acquire a paint film deterioration estimation result for the structure from the painting data, the paint film measurement data, and the weather data, wherein the paint film deterioration estimation model is learned from training data consisting of painting data of the structure, paint film measurement data, weather data, and paint film deterioration data for the structure.

Description

Technical Field

[0001] The present disclosure relates to a coating film deterioration estimation device, a coating film deterioration estimation method, and a program.

Background Art

[0002] Social infrastructure such as roads, rivers, and dams plays an important role in supporting social and economic activities, and it is required to maintain, manage, and renew various facilities. Various infrastructure facilities are corrosion-protected on steel materials according to the usage environment. For example, painting prevents deterioration by physically blocking the external environment and the steel material. However, since the corrosion protection performance of painting decreases due to aging deterioration, repainting is required according to the deterioration state of the coating film. For example, in a steelworks built on a vast site, there are not only factories and furnaces for manufacturing iron products, but also various infrastructure facilities such as roads for vehicles transporting raw materials such as iron ore and coal, intermediate products, and manufactured iron products, bridges, etc. The management of these infrastructure facilities is an essential task for proper operation in the steelworks.

[0003] For example, in the painting of steel structures such as steel bridges, the coating film can deteriorate due to aging deterioration. If this is left unattended, the strength of the steel structure decreases due to rusting, and the proper use of the infrastructure facility is hindered.

[0004] An approach using digital transformation (DX) technology for the maintenance and management of such infrastructure facilities has been proposed.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Non-Patent Documents

[0007] In recent years, research and development of artificial intelligence (AI) technology has progressed, and attempts are being made to solve various problems using machine learning models.

[0008] The objective of this disclosure is to provide a technology for estimating the deterioration of paint applied to structures using a machine learning model. [Means for solving the problem]

[0009] One aspect of the present disclosure relates to a paint film deterioration estimation device, comprising: an acquisition unit that acquires painting data of a structure, paint film measurement data for the structure, and weather data for the area where the structure is installed; and a processing unit that uses a paint film deterioration estimation model to acquire a paint film deterioration estimation result for the structure from the painting data, the paint film measurement data, and the weather data, wherein the paint film deterioration estimation model is learned from training data consisting of painting data of the structure, paint film measurement data, weather data, and paint film deterioration data for the structure. [Effects of the Invention]

[0010] According to this disclosure, a technology can be provided for estimating the deterioration of paint applied to structures using a machine learning model. [Brief explanation of the drawing]

[0011] [Figure 1] Figure 1 is a schematic diagram showing a coating deterioration estimation process according to one embodiment of the present disclosure. [Figure 2] Figure 2 is a block diagram showing the hardware configuration of a coating deterioration estimation device according to one embodiment of the present disclosure. [Figure 3] Figure 3 is a block diagram showing the functional configuration of a coating deterioration estimation device according to one embodiment of the present disclosure. [Figure 4] Figure 4 is a schematic diagram showing the training process of a coating degradation estimation model according to one embodiment of the present disclosure. [Figure 5] Figure 5 is a schematic diagram showing a coating deterioration prediction process according to one embodiment of the present disclosure. [Figure 6] Figure 6 is a flowchart showing the coating deterioration estimation process according to one embodiment of the present disclosure. [Modes for carrying out the invention]

[0012] Embodiments of this disclosure will be described below with reference to the drawings.

[0013] In the following embodiment, a coating deterioration estimation device using a machine learning model is disclosed.

[0014] [Summary of this disclosure] In the coating deterioration estimation process according to the embodiments of the present disclosure described later, as shown in Figure 1, when the coating data relating to the coating applied to the structure to be inspected, such as a steel bridge, coating measurement data such as impedance resistance values ​​measured by electrochemical measurements on the structure, and past weather data in the area where the structure is installed are received as input data, the coating deterioration estimation device 100 uses a trained coating deterioration estimation model 50 to estimate the current coating deterioration state of the structure to be inspected from these input data.

[0015] This makes it possible to estimate the current state of paint deterioration of a structure based on the installation environment of the structure, along with painting data and paint film measurement data. In the embodiment shown in Figure 1, the paint film deterioration estimation model 50 is provided within the paint film deterioration estimation device 100, but this disclosure is not limited thereto. For example, the paint film deterioration estimation model 50 may be stored and operated in a server (not shown) that can communicate with the paint film deterioration estimation device 100. In this case, the paint film deterioration estimation device 100 may provide the server with painting data, paint film measurement data, and weather data of the structure to be inspected as input data, and receive the estimation result of the current paint film deterioration from the server.

[0016] Here, the coating film deterioration estimation device 100 may be realized by a computing device such as a server or a personal computer (PC). For example, it may have a hardware configuration as shown in FIG. 2. That is, the coating film deterioration estimation device 100 has a storage device 101, a processor 102, an interface device 103, and a communication device 104 that are interconnected via a bus B.

[0017] A program or instruction for realizing various functions and processes described later in the coating film deterioration estimation device 100 may be downloaded from any external device via a network or the like, or may be provided from a removable storage medium such as a CD-ROM (Compact Disk-Read Only Memory) or a flash memory.

[0018] The storage device 101 is realized by a random access memory, a flash memory, a hard disk drive, or the like, and stores files, data, etc. used for executing a program or instruction together with the installed program or instruction. The storage device 101 may include a non-transitory storage medium.

[0019] The processor 102 may be realized by one or more CPUs (Central Processing Units), GPUs (Graphics Processing Units), processing circuits, etc. that may be composed of one or more processor cores, and according to the program, instruction, data, etc. such as parameters necessary for executing the program or instruction stored in the storage device 101, executes various functions and processes of the coating film deterioration estimation device 100 described later.

[0020] The interface device 103 provides an interface between the coating deterioration estimation device 100 and the user. For example, the user operates a GUI (Graphical User Interface) displayed on a display or touch panel using a keyboard, mouse, etc., and sends and receives various information, data, instructions, etc., to and from the coating deterioration estimation device 100 via the interface device 103.

[0021] The communication device 104 is implemented by various communication circuits that perform communication processing with external devices, the Internet, LAN (Local Area Network), and other communication networks.

[0022] However, the hardware configuration described above is merely an example, and the coating degradation estimation device 100 according to this disclosure may be implemented using any other suitable hardware configuration.

[0023] [Coating film deterioration estimation device] Next, a coating deterioration estimation device 100 according to one embodiment of the present disclosure will be described. Figure 3 is a block diagram showing the functional configuration of the coating deterioration estimation device 100 according to one embodiment of the present disclosure.

[0024] As shown in Figure 3, the coating deterioration estimation device 100 has an acquisition unit 110 and a processing unit 120. The functions of the acquisition unit 110 and the processing unit 120 may be realized by a computer program stored in the storage device 101 of the coating deterioration estimation device 100 being executed by a processor 102.

[0025] The acquisition unit 110 acquires painting data of the structure, paint film measurement data on the structure, and weather data of the area where the structure is installed. The structure here is a painted steel structure installed outdoors, and may be, for example, a road bridge or a steel bridge installed in a steel mill. The acquisition unit 110 provides the acquired painting data, paint film measurement data, and weather data to the processing unit 120.

[0026] Here, the painting data may include, for example, one or more of the following: paint data used to paint the structure under inspection, the painting method applied, the timing of the painting, pretreatment data in the painting process, and the thickness of the paint film. The paint data may also include the name of the paint used in each painting process of the structure under inspection, the amount of each paint used, etc. The painting method may also include, for example, if the structure under inspection is a steel structure, open blasting, vacuum blasting, water jet blasting, two-component air spray painting, electric air spray painting, air wrap electrostatic airless painting, paint film removal method using a paint stripper, etc. The pretreatment data may include surface treatment and whether or not blasting was performed. Such painting data may be registered in a database or the like when the structure under inspection is painted.

[0027] Furthermore, the coating measurement data may include, for example, the electrochemical measurement results of the coating on the structure being inspected. Specifically, the electrochemical measurement results may include impedance resistance values ​​obtained by applying AC voltages of various frequencies to the structure and acquiring the current response. The acquired impedance resistance values ​​can be used to estimate the deterioration state of the coating on the structure. Specifically, a decrease in impedance resistance values ​​indicates that the coating is deteriorating.

[0028] Furthermore, the meteorological data may include, for example, one or more data points such as precipitation data, temperature data, humidity data, and sunshine duration data for the installation area from the time of installation of the structure under inspection to the estimated time. For example, the meteorological data may include past precipitation, temperature, humidity, and sunshine duration for the installation area of ​​the structure under inspection.

[0029] The processing unit 120 uses the paint film deterioration estimation model 50 to obtain the estimated paint film deterioration of a structure from painting data, paint film measurement data, and weather data. Specifically, the processing unit 120 inputs the painting data, paint film measurement data, and weather data of the structure to be inspected, acquired from the acquisition unit 110, into the paint film deterioration estimation model 50, and obtains the estimated result of the current paint film deterioration of the structure from the paint film deterioration estimation model 50.

[0030] Here, the paint film deterioration estimation model 50 is trained using training data consisting of structure painting data, paint film measurement data, and weather data, as well as structure paint film deterioration data. Here, the paint film deterioration data is ground truth data that shows the deterioration state of the structure's paint film (e.g., rusting state, corrosion state, peeling state, etc.) corresponding to the structure's painting data, paint film measurement data, and weather data. For example, it may show experimental results or simulation results of the structure's paint film deterioration state obtained through experiments or simulations under the structure's painting data, paint film measurement data, and weather data. Such training data is stored in the training data database (DB) 30.

[0031] For example, the paint film degradation estimation model 50 may be implemented by a deep learning model such as a neural network. Specifically, a training device (not shown) extracts training data from the training data DB 30 and inputs the painting data, paint film measurement data, and weather data of each training data into the paint film degradation estimation model 50 to be trained. The training device then runs the paint film degradation estimation model 50 on the input data and obtains the processing results from the paint film degradation estimation model 50 to be trained. The training device adjusts the parameters of the paint film degradation estimation model 50 to be trained according to the error between the obtained processing results and the ground truth data. For example, if the paint film degradation estimation model 50 is implemented by a deep learning model, the parameters of the paint film degradation estimation model 50 to be trained are adjusted according to the error between the processing results and the ground truth data using backpropagation. For example, once the parameter adjustment described above is completed for all the training data in the training dataset extracted from the training data DB 30, the training device terminates the training process, and the finally obtained paint film degradation estimation model 50 is made available to the paint film degradation estimation device 100 as a trained paint film degradation estimation model 50.

[0032] In this way, the processing unit 120 can use the trained paint film deterioration estimation model 50 to estimate the current paint film deterioration state of the structure under inspection, i.e., the current rusting state, corrosion state, peeling state, etc., from the paint data, paint film measurement data, and weather data of the structure under inspection.

[0033] In one embodiment, the acquisition unit 110 may further acquire a climate forecast scenario, and the processing unit 120 may use a paint film deterioration prediction model 60 to obtain a paint film deterioration prediction result for the structure from the painting data, paint film measurement data, weather data, and climate forecast scenario. Such a paint film deterioration prediction model 60 may be trained using training data consisting of the painting data, paint film measurement data, weather data, and climate forecast scenario of the structure, and the paint film deterioration prediction data of the structure. That is, in addition to the painting data, paint film measurement data, and weather data input to the paint film deterioration estimation model 50 described above, the processing unit 120 further inputs a climate forecast scenario to the paint film deterioration prediction model 60 and obtains a prediction result of future paint film deterioration of the structure under inspection from the paint film deterioration prediction model 60.

[0034] Here, the climate prediction scenario may be based, for example, on future climate change predictions for the area where the structure under inspection is located. Specifically, such climate change predictions may be created by government agencies such as the Japan Meteorological Agency, local governments, universities, research institutions, etc., and may typically be provided as multiple climate prediction scenarios. For example, a scenario in which the average annual temperature is the same as the current one, a scenario in which the average annual temperature is 2 degrees higher than the current one, and a scenario in which the average annual temperature is 4 degrees higher than the current one may be created. In this case, the climate prediction scenario input into the coating deterioration prediction model 60 may be one of these multiple climate prediction scenarios selected by the user.

[0035] Furthermore, the paint film deterioration prediction data used as training data is ground truth data that shows the deterioration state of the paint film of a structure (e.g., rusting state, corrosion state, peeling state, etc.) corresponding to the structure's painting data, paint film measurement data, meteorological data, and climate prediction scenario. For example, it may show experimental or simulation results of the paint film deterioration state of a structure obtained through experiments or simulations under the structure's painting data, paint film measurement data, meteorological data, and climate prediction scenario. Such training data is stored in the training data database (DB) 30.

[0036] Such a paint film degradation prediction model 60 may be implemented by a deep learning model such as a neural network. Specifically, a training device (not shown) extracts training data from the training data DB 30 and inputs the painting data, paint film measurement data, weather data, and climate prediction scenario of each training data into the paint film degradation prediction model 60 to be trained. The training device then runs the paint film degradation prediction model 60 on the input data and obtains the processing results from the paint film degradation prediction model 60 to be trained. The training device adjusts the parameters of the paint film degradation prediction model 60 to be trained according to the error between the obtained processing results and the ground truth data. For example, if the paint film degradation prediction model 60 is implemented by a deep learning model, the parameters of the paint film degradation prediction model 60 to be trained are adjusted according to the error between the processing results and the ground truth data using backpropagation. For example, once the parameter adjustments described above are completed for all the training data in the training dataset extracted from the training data DB30, the training device terminates the training process, and the finally acquired coating degradation prediction model 60 is made available to the coating degradation estimation device 100 as the trained coating degradation prediction model 60.

[0037] In this way, the processing unit 120 can use the trained paint film deterioration prediction model 60 to estimate the future paint film deterioration state of the structure under inspection, i.e., the future rusting state, corrosion state, peeling state, etc., from the paint data, paint film measurement data, weather data, and climate prediction scenarios of the structure under inspection.

[0038] According to the paint film deterioration estimation device 100 described above, it is possible to estimate and predict the current and future paint film deterioration state of a structure based on the installation environment of the structure to be inspected.

[0039] [Paint film deterioration estimation process] Next, a coating deterioration estimation process according to one embodiment of the present disclosure will be described. This coating deterioration estimation process is performed by the coating deterioration estimation device 100 described above, and more specifically, it may be realized by one or more processors 102 of the coating deterioration estimation device 100 executing one or more programs or instructions stored in one or more storage devices 101. Figure 6 is a flowchart showing a coating deterioration estimation process according to one embodiment of the present disclosure.

[0040] As shown in Figure 6, in step S101, the paint film deterioration estimation device 100 acquires painting data of the structure, paint film measurement data for the structure, and meteorological data of the area where the structure is installed. Here, the painting data may include data on the paint used to paint the structure, the painting method applied, the timing of the painting, pretreatment data for painting, and / or the film thickness of the paint. The paint film measurement data may include electrochemical measurement results of the paint film of the structure (e.g., impedance resistance value). The meteorological data may include precipitation data, temperature data, humidity data, and / or sunshine duration data for the installation area from the time of installation to the estimation time.

[0041] In step S102, the paint film deterioration estimation device 100 inputs painting data, paint film measurement data, and meteorological data into the paint film deterioration estimation model 50. Such a paint film deterioration estimation model 50 may be trained using training data consisting of painting data, paint film measurement data, and meteorological data of the structure, and paint film deterioration data of the structure. The paint film deterioration data in the training data is ground truth data that indicates the deterioration state of the paint film of the structure (e.g., rusting state, corrosion state, peeling state, etc.) corresponding to the painting data, paint film measurement data, and meteorological data of the structure. Such paint film deterioration data may, for example, represent experimental results or simulation results of the paint film deterioration state of the structure obtained by experimentation or simulation under the painting data, paint film measurement data, and meteorological data of the structure.

[0042] In step S103, the coating deterioration estimation device 100 obtains the coating deterioration estimation result from the coating deterioration estimation model 50. That is, the coating deterioration estimation device 100 estimates the current coating deterioration state of the structure under inspection (e.g., rusting state, corrosion state, peeling state, etc.).

[0043] According to the paint film deterioration estimation process described above, it is possible to estimate and predict the current and future paint film deterioration state of a structure based on the installation environment of the structure being inspected.

[0044] Although embodiments of this disclosure have been described in detail above, this disclosure is not limited to the specific embodiments described above, and various modifications and changes are possible within the scope of the gist of this disclosure as described in the claims. [Explanation of Symbols]

[0045] 30. Training Data Database 50 Paint film deterioration estimation model 60. Paint film degradation prediction model 100 Coating film deterioration estimation device 110 Acquisition Department 120 Processing Unit

Claims

1. An acquisition unit that acquires painting data of a structure, coating measurement data for the structure, and weather data of the area where the structure is installed, A processing unit that uses a paint film deterioration estimation model to obtain the paint film deterioration estimation result of the structure from the painting data, the paint film measurement data and the weather data, It has, The coating deterioration estimation device is trained using training data consisting of painting data of the structure, coating measurement data and weather data, and coating deterioration data of the structure.

2. The coating data includes one or more of the following: paint data used for coating the structure, the coating method applied to the coating, the timing of the coating, pretreatment data for the coating, and the thickness of the coating, as described in claim 1, for the coating deterioration estimation device.

3. The coating film deterioration estimation device according to claim 1, wherein the coating film measurement data shows the electrochemical measurement results of the coating film of the structure.

4. The coating film deterioration estimation apparatus according to claim 3, wherein the electrochemical measurement results include an impedance resistance value to the coating film.

5. The coating deterioration estimation device according to claim 1, wherein the weather data includes one or more precipitation data, temperature data, humidity data, and sunshine duration data for the installation area from the time of installation of the structure to the estimation time.

6. The coating deterioration estimation device according to claim 1, wherein the coating deterioration estimation result relates to one or more of the rusting state, corrosion state, and peeling state.

7. The acquisition unit further acquires climate prediction scenarios, The processing unit uses a coating deterioration prediction model to obtain a coating deterioration prediction result for the structure from the painting data, the coating measurement data, the weather data, and the climate prediction scenario. The coating deterioration prediction device according to claim 1, wherein the coating deterioration prediction model is trained using training data consisting of painting data of the structure, coating measurement data, weather data and climate prediction scenarios, and coating deterioration prediction data of the structure.

8. The coating deterioration estimation device according to claim 7, wherein the climate prediction scenario is based on future climate change predictions in the installation area.

9. To acquire painting data of the structure, coating measurement data for the structure, and weather data of the area where the structure is installed, Using a paint film deterioration estimation model, the paint film deterioration estimation results of the structure are obtained from the painting data, the paint film measurement data, and the weather data. It has, The aforementioned paint film deterioration estimation model is a computer-based method for estimating paint film deterioration, which is trained using training data consisting of painting data of structures, paint film measurement data, and weather data, as well as paint film deterioration data of the structures.

10. To acquire painting data of the structure, coating measurement data for the structure, and weather data of the area where the structure is installed, Using a paint film deterioration estimation model, the paint film deterioration estimation results of the structure are obtained from the painting data, the paint film measurement data, and the weather data. Have the computer run it, The aforementioned paint film deterioration estimation model is a program that is trained using training data consisting of painting data of structures, paint film measurement data, and weather data, as well as paint film deterioration data of the structures.

Images