Pipeline damage prediction system, pipeline damage prediction method, and pipeline damage prediction program

Abstract

[Problem] To provide a pipeline damage prediction system capable of obtaining appropriate damage prediction information for each pipeline by taking into account irregularly shaped pipe portions and integrated portions. [Solution] This pipeline damage prediction system, which uses a computer to predict damage to pipelines caused by natural disasters, comprises: a basic data storage unit for storing pipeline map data indicating the arrangement of pipelines using line segments and connection points of the line segments, pipeline attribute data indicating attributes of the pipelines, ground characteristic data indicating characteristics of the ground in which the pipelines are buried, and damage data showing the degree of damage caused by natural disasters that have occurred in the past; a pipeline configuration identifying unit that identifies pipeline configurations in which irregularly shaped pipes are arranged at the connection points, on the basis of the characteristics of the connection point of each pipeline indicated in the pipeline map data; and a damage prediction calculating unit that generates damage prediction information predicting damage that has occurred in the pipeline on the basis of the pipeline attribute data, the ground characteristic data, the damage data, and the pipeline configuration stored in the basic data storage unit.

Description

Pipeline Damage Prediction System, Pipeline Damage Prediction Method, and Pipeline Damage Prediction Program

[0001] The present invention relates to a pipeline damage prediction system, a pipeline damage prediction method, and a pipeline damage prediction program that use a computer to predict damage to pipelines caused by natural disasters.

[0002] It is desired to formulate efficient renewal plans in response to the aging of pipelines such as water supply pipe networks. In this process, by predicting pipeline damage caused by natural disasters such as earthquakes, ground collapse, volcanic eruptions, and floods, and preferentially renewing pipelines predicted to suffer severe damage, it becomes possible to reconstruct a pipe network that is resilient to disasters.

[0003] For example, as a method for predicting damage caused by earthquakes, the damage estimation formula Rm(v) shown in FIG. 8A proposed by Japan Water Works Association (JWWA), and the damage estimation formula Rm(v) shown in FIG. 8B proposed by Japan Water Research Center (JWRC) are used.

[0004] The former is a standard damage rate R(v) based on the maximum ground surface velocity v (=3.11×10 ―３ ×(v−15) １．３０ ), which is obtained by multiplying by a pipe type correction coefficient Cp, a diameter correction coefficient Cd, a topography correction coefficient Cg, and a liquefaction correction coefficient Cl. The latter is classified according to the presence or absence of liquefaction. When there is no liquefaction, it is obtained by multiplying the standard damage rate R(v) based on the maximum ground surface velocity v (=9.92×10 ―３ ×(v−15) １．１４ ) by the pipe type correction coefficient Cp, the diameter correction coefficient Cd, and the topography correction coefficient Cg. When there is liquefaction, it is obtained by multiplying the liquefaction standard damage rate R Ｌ by the pipe type correction coefficient Cp and the diameter correction coefficient Cd.

[0005] In either case, each correction coefficient is a value obtained using a method such as multiple regression analysis based on the energy of earthquakes that have occurred in the past and the damage status (number of damage cases per pipeline length (damage rate)).

[0006] Japanese Patent Application Laid-Open No. 2001-329574, Japanese Patent Application Laid-Open No. 2004-310307

[0007] Incidentally, if general fittings such as K-type, expansion joints such as GX-type and NS-type, or flexible expansion pipes are located near irregularly shaped pipes such as curved pipes and T-joints, the irregularly shaped pipe section, including the straight pipes connected before and after it, may move due to the unbalanced force caused by water pressure, potentially affecting nearby underground structures. Unbalanced force refers to the force that acts on the bends, branches, and end plugs or valves of a pipeline, attempting to move the pipe due to water pressure. Note that the names K-type, GX-type, and NS-type are names that represent types of ductile iron pipe fittings as defined by the Japan Ductile Iron Pipe Association.

[0008] To stabilize pipelines underground, chain-structure pipelines integrate the pipes connected before and after the irregularly shaped pipes with anti-detachment joints. This method uses friction between the pipe and the soil, ground reaction force behind the pipe, or the bending rigidity of the anti-detachment joint to maintain unbalanced forces.

[0009] Such irregularly shaped pipe sections and integrated sections are considered more susceptible to damage because their movement is restricted by ground movement caused by earthquakes. However, conventional damage prediction methods based on damage estimation formulas have not been able to take these irregularly shaped pipe sections and integrated sections into their damage predictions, and further improvements have been desired to achieve highly accurate pipeline damage prediction.

[0010] In this regard, currently, damage predictions for pipelines are made using the aforementioned damage estimation formula based on pipeline maps obtained from geographic information systems (GIS). However, these pipeline maps are defined as a collection of line segments (pipelines) connecting numerous connection points and two adjacent connection points, and the attribute data for each pipeline does not include data indicating pipeline morphology such as irregularly shaped pipe sections or integrated sections. Furthermore, the entities managing the pipeline network have not adequately prepared such data. As a result, it has been difficult to predict damage that takes into account irregularly shaped pipe sections and integrated sections.

[0011] The object of the present invention is to provide a pipeline damage prediction system, a pipeline damage prediction method, and a pipeline damage prediction program that can obtain appropriate damage prediction information for each pipeline by generating data on irregularly shaped pipe sections and integrated sections from existing pipeline data.

[0012] To achieve the above objectives, the first characteristic configuration of the pipeline damage prediction system according to the present invention is a pipeline damage prediction system that uses a computer to predict damage to pipelines caused by natural disasters, comprising: a basic data storage unit associated with a pipeline diagram showing the arrangement of pipelines at line segments and connection points of line segments, which stores pipeline attribute data indicating the attributes of the pipelines, ground characteristic data indicating the characteristics of the ground in which the pipelines are buried, and damage data indicating the degree of damage caused by natural disasters that have occurred in the past; a pipeline configuration identification unit that identifies a pipeline configuration in which irregularly shaped pipes are arranged at connection points based on the characteristics of the connection points of each pipeline shown in the pipeline diagram; and a damage prediction calculation unit that generates damage prediction information predicting damage to the pipelines based on the pipeline attribute data, ground characteristic data, damage data, and pipeline configurations stored in the basic data storage unit.

[0013] For a pipeline diagram showing the arrangement of pipelines at line segments and connection points, the pipeline configuration identification unit identifies the pipeline configuration based on the characteristics of the connection points, determining that irregularly shaped pipes are placed at the connection points. The damage prediction calculation unit generates damage prediction information predicting the damage to the pipelines based on pipeline attribute data showing the attributes of each pipeline, ground characteristic data, pipeline configuration, and damage data showing the degree of damage caused by natural disasters such as earthquakes that have occurred in the past. Damage prediction information that takes into account susceptibility to the effects of earthquakes and other disasters can be obtained, making it possible to formulate an appropriate pipeline renewal plan.

[0014] The second characteristic configuration is that, in addition to the first characteristic configuration described above, the pipeline shape identification unit includes an angle calculation unit that calculates the angle formed by each pipeline at the connection point, and an irregular pipe identification unit that identifies the shape of the pipeline located within a predetermined range from the connection point as an irregular pipe when the angle calculated by the angle calculation unit is inclined by a predetermined angle or more with respect to 180 degrees.

[0015] If the angle of the pipelines connected at the connection point calculated by the angle calculation unit is inclined at a predetermined angle or more relative to 180 degrees, the irregular pipe identification unit identifies that an irregular pipe is located at the connection point, and identifies the shape of the pipeline located within a predetermined range from that connection point as an irregular pipe. The pipeline shape identification unit can appropriately identify the presence of irregular pipes from the pipeline diagram data.

[0016] The third characteristic configuration is that, in addition to the first characteristic configuration described above, the pipeline shape identification unit includes a diameter change detection unit that detects changes in the diameter of each pipeline at the connection point, and an irregular pipe identification unit that, when the diameter change detection unit detects a change in diameter, identifies the shape of the pipeline located within a predetermined range from the connection point as an irregular pipe.

[0017] The diameter of each pipe constituting the connection point shown in the pipeline diagram can be determined from the diameters included in the attributes of both pipes in the pipeline attribute data. If the diameters of the two pipes are different, the irregular pipe identification unit identifies the shape of the pipe located within a predetermined range from the connection point as an irregular pipe.

[0018] The fourth characteristic configuration is that, in addition to the second or third characteristic configuration described above, the pipeline shape identification unit includes an integrated unit identification unit that, when the irregular pipe identification unit identifies the shape of the pipeline at the connection point as an irregular pipe, identifies the pipeline located within a predetermined range from the irregular pipe as an integrated unit.

[0019] When the pipeline configuration identification unit identifies that an irregularly shaped pipe is located at a connection point, the integration unit identification unit determines that an integration unit is formed in the irregularly shaped pipe section, and the pipeline located within a predetermined range from the irregularly shaped pipe is identified as the integration unit.

[0020] The fifth characteristic configuration is that, in addition to the fourth characteristic configuration described above, the pipeline shape identification unit includes a straight pipe identification unit that identifies the connection point as a straight pipe connection point when the shape of the pipeline at the connection point cannot be identified as an irregular pipe by the irregular pipe identification unit.

[0021] If the shape of the pipeline at the connection point cannot be identified as an irregularly shaped pipe, it will be identified as a straight pipe connection point by the straight pipe identification section.

[0022] The sixth characteristic configuration is that, in addition to the first characteristic configuration described above, the damage prediction calculation unit includes a machine learning unit that generates a decision tree model by ensemble learning of multiple datasets consisting of pipeline attribute data, ground characteristic data, pipeline morphology, and damage data associated with each pipeline, and generates the damage prediction information based on the decision tree model generated by the machine learning unit.

[0023] Preferably, the damage prediction calculation unit includes a machine learning unit that generates the damage prediction information. The machine learning unit prepares a dataset for each of the multiple pipelines, including the pipeline attribute data, ground characteristics data, pipeline morphology, and past damage data, and generates a decision tree model by repeatedly performing ensemble learning. By inputting the pipeline attribute data and ground characteristics data for each pipeline to be evaluated into the decision tree model generated by the machine learning unit, appropriate damage prediction information can be obtained.

[0024] The first characteristic configuration of the pipeline damage prediction method according to the present invention is a pipeline damage prediction method that uses a computer to predict damage to a pipeline caused by a natural disaster, and is characterized by the following steps: a basic data storage step in which pipeline attribute data indicating the attributes of the pipeline, ground characteristic data indicating the characteristics of the ground in which the pipeline is buried, and damage data indicating the degree of damage caused by natural disasters that have occurred in the past are stored in a basic data storage unit, associated with a pipeline diagram showing the arrangement of the pipeline with line segments and connection points; a pipeline configuration identification step in which a pipeline configuration in which irregularly shaped pipes are arranged at connection points is identified based on the characteristics of the connection points of each pipeline shown in the pipeline diagram; and a damage prediction calculation step in which damage prediction information predicting the damage to the pipeline is generated based on the pipeline attribute data, ground characteristic data, damage data, and pipeline configuration stored in the basic data storage unit.

[0025] The second characteristic configuration is that, in addition to the first characteristic configuration described above, the pipeline shape identification step includes an angle calculation step for calculating the angle formed by each pipeline at the connection point, and an irregular pipe identification step for identifying the shape of the pipeline located within a predetermined range from the connection point as an irregular pipe if the angle calculated in the angle calculation step is inclined by a predetermined angle or more with respect to 180 degrees.

[0026] The third characteristic configuration is that, in addition to the first characteristic configuration described above, the pipeline shape identification step includes a diameter change detection step that detects a change in the diameter of each pipeline at the connection point, and an irregular pipe identification step that, when a change in diameter is detected by the diameter change detection step, identifies the shape of the pipeline located within a predetermined range from the connection point as an irregular pipe.

[0027] The fourth characteristic configuration is that, in addition to the second or third characteristic configuration described above, the pipeline shape identification step, after identifying the shape of the pipeline at the connection point as the irregular pipe in the irregular pipe identification step, performs an integrated section identification step which identifies the pipeline located within a predetermined range from the irregular pipe as an integrated section.

[0028] The fifth characteristic configuration is that, in addition to the fourth characteristic configuration described above, the pipeline shape identification step performs a straight pipe identification step to identify the connection point as a straight pipe connection point when the shape of the pipeline at the connection point cannot be identified as a irregular pipe by the irregular pipe identification step.

[0029] The sixth feature configuration is that, in addition to the first feature configuration described above, the damage prediction calculation step performs a machine learning step that generates a decision tree model by ensemble learning of multiple datasets consisting of pipeline attribute data, ground characteristic data, pipeline morphology, and damage data associated with each pipeline, and generates the damage prediction information based on the decision tree model generated by the machine learning step.

[0030] The first characteristic configuration of the pipeline damage prediction program according to the present invention is that the computer functions as a basic data storage processing unit associated with a pipeline diagram showing the arrangement of pipelines with line segments and connection points, and storing pipeline attribute data indicating the attributes of the pipelines, ground characteristic data indicating the characteristics of the ground in which the pipelines are buried, and damage data indicating the degree of damage caused by natural disasters that have occurred in the past; a pipeline configuration identification unit that identifies a pipeline configuration in which irregularly shaped pipes are arranged at connection points based on the characteristics of each connection point of the pipelines shown in the pipeline diagram; and a damage prediction calculation unit that generates damage prediction information predicting damage to the pipelines based on the pipeline attribute data, ground characteristic data, damage data, and pipeline configurations stored in the basic data storage unit.

[0031] The second characteristic configuration is that, in addition to the first characteristic configuration described above, the pipeline shape identification unit includes an angle calculation unit that calculates the angle formed by each pipeline at the connection point, and an irregular pipe identification unit that identifies the shape of the pipeline located within a predetermined range from the connection point as an irregular pipe when the angle calculated by the angle calculation unit is inclined by a predetermined angle or more with respect to 180 degrees.

[0032] The third characteristic configuration is that, in addition to the first characteristic configuration described above, the pipeline shape identification unit includes a diameter change detection unit that detects changes in the diameter of each pipeline at the connection point, and an irregular pipe identification unit that, when the diameter change detection unit detects a change in diameter, identifies the shape of the pipeline located within a predetermined range from the connection point as an irregular pipe.

[0033] The fourth characteristic configuration is that, in addition to the second or third characteristic configuration described above, the pipeline shape identification unit includes an integrated section identification unit that, when the irregular pipe identification unit identifies the shape of the pipeline at the connection point as an irregular pipe, identifies the pipeline located within a predetermined range from the irregular pipe as an integrated section.

[0034] The fifth characteristic configuration is that, in addition to the first characteristic configuration described above, the pipeline shape identification unit includes a straight pipe identification unit that identifies the connection point as a straight pipe connection point when the shape of the pipeline at the connection point cannot be identified as an irregular pipe by the irregular pipe identification unit.

[0035] The sixth characteristic configuration is that, in addition to the first characteristic configuration described above, the damage prediction calculation unit includes a machine learning unit that generates a decision tree model by ensemble learning of multiple datasets consisting of pipeline attribute data, ground characteristic data, pipeline morphology, and damage data associated with each pipeline, and generates the damage prediction information based on the decision tree model generated by the machine learning unit.

[0036] As described above, the present invention provides a pipeline damage prediction system, a pipeline damage prediction method, and a pipeline damage prediction program that can obtain appropriate damage prediction information for each pipeline by taking into account irregularly shaped pipe sections and integrated sections.

[0037] Figure 1 is an explanatory diagram of the computer configuration on which the pipeline damage prediction system is built. Figure 2 is an explanatory diagram of the functional blocks of the pipeline damage prediction system. Figure 3A is an explanatory diagram of the configuration of the dataset for learning or evaluation. Figure 3B is an explanatory diagram of the pipeline length when the pipeline has irregularly shaped pipe sections and integrated sections. Figure 3C is an explanatory diagram of the pipeline length when the pipeline does not have irregularly shaped pipe sections and integrated sections. Figure 4A is an explanatory diagram for formulating ground boundary data. Figure 4B is an explanatory diagram for formulating ground boundary data. Figure 5A is an explanatory diagram of the pipeline diagram obtained from GIS. Figure 5B is an explanatory diagram of the formula for calculating the angle of two pipelines connected at a connection point. Figure 5C is an explanatory diagram of irregularly shaped pipe sections and integrated sections identified by the pipeline shape identification unit from the pipeline diagram in Figure 5A. Figure 6A is an explanatory diagram of the table showing the integrated section length as defined by the design standards. Figure 6B is an explanatory diagram of the table showing the integrated section length as defined by the design standards. Figure 6C is an explanatory diagram of the table showing the integrated section length as defined by the design standards. Figure 7A is an explanatory diagram of a composite piping system in which different types of irregularly shaped pipes are connected. Figure 7B is an explanatory diagram of a composite piping system in which different types of irregularly shaped pipes are connected. Figure 7C is an explanatory diagram of a composite piping system in which different types of irregularly shaped pipes are connected. Figure 8 is a flowchart showing the procedure for the pipeline damage prediction method. Figure 9 is an explanatory diagram of the prediction results by the pipeline damage prediction system. Figure 10A is an explanatory diagram of the damage estimation formula by JWWA. Figure 10B is an explanatory diagram of the damage estimation formula by JWRC.

[0038] The pipeline damage prediction system, pipeline damage prediction method, and pipeline damage prediction program according to the present invention will be described below with reference to the drawings. [Pipeline Damage Prediction System] As shown in Figure 1, the pipeline damage prediction system 10 consists of a computer C, which is connected by a communication bus to a motherboard equipped with an integrated circuit such as a CPU and chipset, a memory board equipped with semiconductor memory such as ROM and RAM, and an I / O board equipped with an input / output interface.

[0039] Computer C is connected to storage devices such as hard disk drives and flash memory devices, portable memory devices such as USB memory, input devices such as keyboards and mice, and output devices such as LCD displays and printers, and is configured to connect to a cloud computer via a communication interface such as the Wi-Fi standard.

[0040] A pipeline damage prediction program, provided via a recording medium such as a cloud computer, storage device, or portable memory device, is installed in the memory on the memory board, and the pipeline damage prediction program is executed by the CPU, thereby realizing each functional block of the pipeline damage prediction system 10.

[0041] Figure 2 shows the functional block configuration of the pipeline damage prediction system 10. The pipeline damage prediction system 10 comprises a basic data storage unit 11, a damage prediction calculation unit 12, and a pipeline morphology identification unit 14. The basic data storage unit 11 is composed of the aforementioned storage devices and cloud computers, and stores pipeline maps acquired from a geographic information system (GIS), pipeline attribute data showing the attributes of each pipeline shown in the pipeline map, ground characteristic data showing the characteristics of the ground in which the pipeline is buried, which includes at least one of the following in GIS units: liquefaction data, ground boundary data, and topographic modification data, and damage data showing the degree of damage caused by natural disasters that have occurred in the past.

[0042] The pipeline shape identifying unit 14 includes an angle calculating unit 15 that calculates the angle formed by each pipeline at a connection point, a diameter change detecting unit 16, a deformed pipe identifying unit 17, an integrated part identifying unit 18, and a straight pipe identifying unit 19.

[0043] The damage prediction calculation unit 12 is a calculation unit that generates damage prediction information predicting damage occurring to pipelines based on pipeline attribute data, ground characteristic data, and damage data stored in the basic data storage unit 11.

[0044] The damage prediction calculation unit 12 includes a machine learning unit 13 that generates a decision tree model by performing ensemble learning on a plurality of data sets configured of pipeline attribute data, ground characteristic data, and damage data associated with each pipeline, and generates the damage prediction information based on the decision tree model generated by the machine learning unit 13.

[0045] A decision tree is a knowledge representation that describes feature separation procedures through a branched tree structure, and determines the property and classification category of a pipeline that is a damage estimation target from a plurality of attributes characterizing the pipeline. The attributes refer to the above-described pipeline attribute data, ground characteristic data, and each classification item configuring the damage data. In the decision tree, branching proceeds according to answers to questions about the attributes, and finally damage estimation information of the pipeline is obtained.

[0046] FIG. 3A illustrates an example of an attribute data set set for each pipeline Pi (i=1, 2, ...) configuring a pipeline network. The pipeline Pi includes, as the pipeline attribute data, pipe type that is data enabling identification of material and earthquake resistance, diameter that is data indicating the size of a pipe, installation year, pipeline extension that is data indicating the length of the pipe, presence / absence of deformed pipe parts / integrated parts included in the pipeline Pi, deformed pipe part extension D1, integrated part extension D2, straight pipe part extension D3, and the like. The presence / absence of deformed pipe parts / integrated parts, deformed pipe part extension D1, integrated part extension D2, and straight pipe part extension D3 included in the pipeline attribute data are identified by the pipeline shape identifying unit 14 described later.

[0047] Furthermore, the ground characteristics data includes values ​​that are basically in polygon units, such as the microtopographic classification where the pipeline is laid, liquefaction data indicating liquefaction characteristics, topographic modification data such as former river channels, former water areas, and residential land development ground data, ground boundary data indicating whether or not the pipeline exists at the ground boundary, maximum ground acceleration, and maximum ground velocity. In addition, damage data includes the number of damages per unit pipeline length (km) from past earthquakes. The microtopographic classification is set with values ​​that reflect the degree of earthquake impact, such as velocity amplification factor, according to the microtopography. The "microtopographic classification" includes data indicating the category of "topography" and numerical data of "topography" indicating the degree of impact, such as velocity amplification factor. This takes into account the degree of earthquake impact due to microtopography.

[0048] The liquefaction data indicates whether or not liquefaction occurred due to past earthquakes as a 1 / 0, the topographic modification data, such as former river channels, former water bodies, and residential land development ground data, indicates whether or not they apply as a 1 / 0, and the ground boundary data indicates whether or not a pipeline exists at the ground boundary as a 1 / 0.

[0049] If the liquefaction data is "1", it can be predicted that liquefaction will likely occur in the event of a future earthquake, increasing the likelihood of damage. The topographic modification data can be predicted that changes in topography will increase the likelihood of damage in the event of a future earthquake. This takes into account areas where ground deformation is likely to occur, which cannot be evaluated using the current topographic classification. Former river channels and former waterways have soft ground, and residential development areas have soft ground due to embankment, so it can be predicted that ground deformation will increase in the event of a future earthquake.

[0050] If a pipeline is located at a ground boundary, that is, at the boundary of adjacent polygons, then in the event of a future earthquake, the ground will not change uniformly, resulting in different ground deformations affecting the pipeline, and consequently, pipeline damage can be predicted.

[0051] Figure 4A shows three adjacent polygons represented by solid lines. The data source is a 1:200,000 seamless geological map, and assuming a boundary line width of 0.05 mm (roughly the width of a single hair), the ground boundary will have a width of approximately 10 m (= 0.05 × 200,000 × 1 / 1,000).

[0052] As shown in Figure 4B, when ground with different characteristics is adjacent, the movement of both during an earthquake will differ, affecting the pipeline at the boundary, and this effect is thought to extend to the adjacent joint. Since the length of a single pipe is 4m to 6m, the range of influence of ground changes on the pipeline is ±6m of the ground boundary, considering the conservative side. If we consider the effect of positional accuracy to be ±5m and the influence of ground changes on the pipeline to be 6m on the conservative side, the range of the ground boundary can be defined as ±11m. Specifically, ground boundary data can be defined as the overlapping area when adjacent boundaries are expanded by a predetermined width based on the pipeline length, using polygons as the unit. This makes it possible to appropriately evaluate the degree of influence of the ground on the pipeline included in each ground boundary data. In this embodiment, we have illustrated the case where the width of the boundary line in the geological map can be estimated to be 10m, but it goes without saying that this value is not limited to this value and depends on the data source.

[0053] Figure 5A shows an example of a pipeline diagram obtained from a Geographic Information System (GIS). A pipeline diagram is defined as a set of line segments (straight lines) connecting numerous connection points and two adjacent connection points, and is diagrammatic information in which the arrangement of pipelines is shown by line segments and connection points.

[0054] Figure 5A shows six pipelines, labeled A through F, and multiple connection points, indicated by black circles. The line types and thicknesses of the lines are varied for convenience to facilitate identification of each pipeline. This pipeline diagram shows five types of piping patterns, numbered 1 through 5 and enclosed in circles. These are: piping pattern 1, where a single pipeline branches; piping pattern 2, where a single pipeline has a bend; piping pattern 3, where multiple pipelines branch; piping pattern 4, where multiple pipelines have bends; and piping pattern 5, where the diameter changes within a straight pipeline.

[0055] The pipeline configuration identification unit 14 determines whether or not a pipeline configuration in which a modified pipe is placed at a connection point is one in which a modified pipe is placed, based on the characteristics of the connection points of each pipeline shown in the pipeline diagram (plan view) described above. If it determines that a pipeline configuration in which a modified pipe is placed is one in which a modified pipe is placed, it registers attribute information related to the pipeline configuration in the attribute information of the corresponding pipe.

[0056] As shown in Figure 5B, the angle calculation unit 15 calculates the angle θ between the two line segments P1(x1, y1) and P2(x2, y2) that enclose the connection point P2(x2, y2) and P2(x2, y2) and P3(x3, y3).

[0057] The irregular pipe identification unit 17 identifies a pipe shape located within a predetermined range from the connection point as an irregular pipe when the angle calculated by the angle calculation unit 15 is inclined at a predetermined angle relative to 180 degrees. The predetermined range can be, for example, within a radius of 1 m centered on the connection point. A radius of 1 m is an example and is not limited to this value. Irregular pipes include curved pipes with bending angles of 90 degrees, 45 degrees, 22.5 degrees, and 11.25 degrees, while the allowable bending angle for straight pipe joints is specified as 4 degrees. Therefore, in this embodiment, an irregular pipe is identified when the angle is inclined at 10 degrees or more relative to 180 degrees. This corresponds to piping patterns 2 and 4 in Figure 5A.

[0058] Similarly, in the case of branch pipes such as T-joints and cross-shaped pipes, the angle calculated by the angle calculation unit 15 is inclined at a predetermined angle relative to 180 degrees, and the number of branches at the connection point is three or more, so it can be characterized as a different shape of pipe than a curved pipe. This corresponds to piping pattern 1 and piping pattern 3 in Figure 5A.

[0059] The diameter change detection unit 16 detects the diameter from the pipe attribute data of both pipes on either side of the connection point and determines whether or not there is a change in diameter. The irregular pipe identification unit 17 identifies the shape of the pipe located within a predetermined range from the connection point as an irregular pipe when the diameter change detection unit 16 detects a change in diameter. This corresponds to the piping pattern 5 in Figure 5A.

[0060] As shown in Figure 5C, the integrated section identification section 18 identifies the shape of the pipeline at the connection point as an irregular pipe using the irregular pipe identification section 17, and then identifies the pipeline located within a predetermined range from the irregular pipe as an integrated section. An integrated section is a part that stabilizes the pipeline underground by integrating the pipes connected before and after the irregular pipe with a detachment prevention joint, and maintaining unbalanced forces through frictional force between the pipe and the soil, ground reaction force on the back of the pipe, or bending rigidity of the detachment prevention joint. In Figure 5C, this is the part shown by the dashed line.

[0061] The range from the irregularly shaped pipe identified as the integrated section is defined by the standard values ​​for integrated design of the irregularly shaped pipe. The lengths L1, L2, and L3 of the integrated section shown in Figure 5C are determined based on the integrated section lengths specified in the design standards, as illustrated in Figure 6.

[0062] The straight pipe identification section 19 identifies a connection point as a straight pipe when the irregular pipe identification section 17 cannot identify the pipe configuration at the connection point as an irregular pipe. It goes without saying that for pipe configurations where the depth direction changes, such as inverted siphons, similar identification is possible if a pipe diagram or data showing the depth direction is available.

[0063] Figure 7A shows an example of a composite piping system consisting of branch pipes and curved pipes. The pipe with the largest angle between the two connecting pipes is defined as the main pipe, and the remaining pipes are defined as branch pipes. The integrated section is defined as the curved pipe, and the longer integrated length of the branch pipe is applied.

[0064] Figure 7B shows an example of a composite piping system consisting of a curved pipe and a drop-off pipe, where the longer of the two integrated lengths is applied to the curved pipe and the drop-off pipe.

[0065] Figure 7C shows an example of a composite piping system consisting of branch pipes, curved pipes, and drop-off pipes. The pipe with the largest angle between the two connecting pipes is defined as the main pipe, and the remaining pipes are defined as branch pipes. The integration section is determined by the longest integration length among the curved pipes, branch pipes, and drop-off pipes.

[0066] In Figure 3A, the attributes identified by the pipeline morphology identification unit 14 are shown in hatched areas as a dataset of attributes set for each pipeline Pi (i = 1, 2, ...) that constitutes the pipeline network.

[0067] The presence or absence of irregularly shaped pipe sections and integrated sections, along with the extensions D1 (m) of the irregularly shaped pipe section, D2 (m) of the integrated section, and D3 (m) of the straight pipe section, are added. As shown in Figure 3B, if irregularly shaped pipe sections and integrated sections are present, the pipeline length D (m) = D1 + D2 + D3. As shown in Figure 3C, if irregularly shaped pipe sections and integrated sections are absent, D1 = D2 = 0, and the pipeline length D (m) becomes the pipeline length of the straight pipe section.

[0068] A decision tree is generated by using the following basic learning algorithm to train the machine learning unit 13 with the dataset shown in Figure 3A as training data. The training dataset is classified into subsets using appropriate attributes. If there are no attributes that can be used for classification, the algorithm terminates without completing the classification knowledge. For each subset, this algorithm is applied to find the difference between the obtained damage prediction information and the actual damage information, and the process of creating a new decision tree (classification rule) to reduce the difference is repeated. At this time, instead of creating a single decision tree, multiple decision trees are created and ensemble learning is used, which uses the whole set of decision trees as a single piece of knowledge, thereby generating a more appropriate decision tree.

[0069] Parallel ensemble methods such as XGBoost and Random Forest are suitable for ensemble learning. For example, in the Random Forest method, multiple training datasets are created by randomly extracting data from a dataset as shown in Figure 3. A corresponding decision tree is created using each training dataset, and the overall result, i.e., damage estimation information, can be obtained by using the average of the outputs of the multiple decision trees.

[0070] Figure 8 shows the series of steps described above. In other words, the pipeline damage prediction method, which uses a computer to predict damage to pipelines caused by natural disasters including earthquakes, comprises a basic data acquisition step (SA1) which is associated with a pipeline diagram that shows the arrangement of pipelines using line segments and connection points, and acquires pipeline attribute data that shows the attributes of the pipelines, ground characteristic data that shows the characteristics of the ground in which the pipelines are buried, and damage data that shows the degree of damage caused by natural disasters that have occurred in the past from a GIS; a pipeline configuration identification step (SA2) which determines whether or not a pipeline configuration in which a non-standard pipe is placed at a connection point is one in which a non-standard pipe is placed, and if so, identifies the integrated length; and a basic data storage step (SA3) which stores pipeline attribute data, ground characteristic data, damage data, and pipeline configuration in a basic data storage unit.

[0071] Furthermore, the system is configured to perform damage prediction calculation steps (SA4 to SA9) that generate damage prediction information predicting damage to the pipeline based on pipeline attribute data including pipeline configuration stored in the basic data storage unit, ground characteristics data, and damage data, and to execute a pipeline renewal plan (SA10) based on the results.

[0072] The damage prediction calculation steps (SA4 to SA9) include a step (SA9) in which a portion (80% in this embodiment) of pipeline attribute data, ground characteristic data, and damage data stored in the basic data storage unit and associated with each pipeline is divided into multiple datasets (SA4), a machine learning step (SA5 to SA7) is performed to generate a decision tree model by ensemble learning, and the remaining pipeline attribute data, ground characteristic data, and damage data (20% in this embodiment) stored in the basic data storage unit as evaluation data is applied to the decision tree model generated by the machine learning step (SA8) to obtain damage prediction information.

[0073] Figure 9 shows the accuracy of recall for actual earthquake damage caused by past earthquakes in a certain region, when five new features are added to the ground characteristics data: presence or absence of liquefaction, presence or absence of residential land development ground, presence or absence of ground boundaries, presence or absence of old river channels / water bodies, and numerical data of microtopography, and when any of the features related to pipeline morphology are removed or combined, as well as the accuracy of recall for a model (M0) that does not take into account any of the features related to pipeline morphology, and the accuracy of recall for existing formulas.

[0074] The features used in the existing formula are "pipe type," "diameter," "micro-topographic classification," "presence or absence of liquefaction," and "PGA or PGV." The recall rate is the percentage of actual damaged pipelines for which the predicted damage was accurate, and is calculated by considering the top 1.8% of pipelines with a probability of damage (five times the number of actual damaged pipelines) as having damage. A higher number indicates better prediction accuracy. It has been observed that the recall rate has been significantly improved by incorporating data on the presence or absence of irregularly shaped pipe sections.

[0075] The pipeline damage prediction system 10 described above executes the pipeline damage prediction method according to the present invention. Specifically, the pipeline damage prediction method includes a pipeline diagram showing the arrangement of pipelines with line segments and connection points, and includes a basic data storage step in which pipeline attribute data indicating the attributes of the pipelines, ground characteristic data indicating the characteristics of the ground in which the pipelines are buried, and damage data indicating the degree of damage caused by natural disasters that have occurred in the past are stored in a basic data storage unit; a pipeline configuration identification step in which, based on the characteristics of the connection points of each pipeline shown in the pipeline diagram, a pipeline configuration in which irregularly shaped pipes are arranged at the connection points; and a damage prediction calculation step in which damage prediction information predicting damage to the pipelines is generated based on the pipeline attribute data, ground characteristic data, damage data, and pipeline configurations stored in the basic data storage unit.

[0076] The pipeline morphology identification step includes an angle calculation step that calculates the angle formed by each pipeline at the connection point, and an irregular pipe identification step that identifies irregular pipes as irregular pipes if the angle calculated in the angle calculation step is inclined by a predetermined angle or more relative to 180 degrees, and the pipeline morphology located within a predetermined range from the connection point is identified as an irregular pipe.

[0077] The pipeline morphology identification step includes a diameter change detection step that detects a change in the diameter of each pipeline at the connection point, and an irregular pipe identification step that, if a change in diameter is detected by the diameter change detection step, identifies the morphology of the pipeline located within a predetermined range from the connection point as an irregular pipe.

[0078] The pipeline morphology identification step involves, after identifying the morphology of the pipeline at the connection point as an irregular pipe in the irregular pipe identification step, performing an integrated section identification step which identifies the pipeline located within a predetermined range from the irregular pipe as an integrated section.

[0079] The pipeline configuration identification step involves executing a straight pipe identification step to identify the connection point as a straight pipe connection point if the configuration of the pipeline at the connection point cannot be identified as a irregularly shaped pipe by the irregularly shaped pipe identification step.

[0080] The damage prediction calculation step involves executing a machine learning step to generate a decision tree model by ensemble learning multiple datasets consisting of pipeline attribute data, ground characteristic data, pipeline morphology, and damage data associated with each pipeline. Damage prediction information is then generated based on the decision tree model generated by the machine learning step.

[0081] The application program that embodies the pipeline damage prediction system 10 described above becomes the pipeline damage prediction program of the present invention, and is installed on a computer via the various recording media and communication interfaces described above.

[0082] In other words, the pipeline damage prediction program is a pipeline damage prediction program that enables a computer to function as follows: a basic data storage processing unit that associates a pipeline diagram showing the arrangement of pipelines with line segments and connection points, and stores pipeline attribute data indicating the attributes of the pipelines, ground characteristic data indicating the characteristics of the ground in which the pipelines are buried, and damage data indicating the degree of damage caused by natural disasters that have occurred in the past; a pipeline configuration identification unit that identifies a pipeline configuration in which irregularly shaped pipes are placed at connection points based on the characteristics of the connection points of each pipeline shown in the pipeline diagram; and a damage prediction calculation unit that generates damage prediction information predicting damage to the pipelines based on the pipeline attribute data, ground characteristic data, damage data, and pipeline configurations stored in the basic data storage unit.

[0083] Furthermore, the pipeline shape identification unit includes an angle calculation unit that calculates the angle formed by each pipeline at the connection point, and an irregular pipe identification unit that identifies the shape of a pipeline located within a predetermined range from the connection point as an irregular pipe when the angle calculated by the angle calculation unit is inclined by a predetermined angle or more relative to 180 degrees.

[0084] The pipeline shape identification unit includes a diameter change detection unit that detects changes in the diameter of each pipeline at the connection point, and an irregular pipe identification unit that, when the diameter change detection unit detects a change in diameter, identifies the shape of a pipeline located within a predetermined range from the connection point as an irregular pipe.

[0085] The pipeline shape identification unit includes an integrated section identification unit that, when the irregular pipe identification unit identifies the shape of the pipeline at the connection point as an irregular pipe, identifies the pipeline located within a predetermined range from the irregular pipe as an integrated section.

[0086] The pipeline configuration identification unit includes a straight pipe identification unit that identifies the connection point as a straight pipe when the configuration of the pipeline at the connection point cannot be identified as a irregularly shaped pipe by the irregularly shaped pipe identification unit.

[0087] The damage prediction calculation unit includes a machine learning unit that generates a decision tree model by ensemble learning of multiple datasets consisting of pipeline attribute data, ground characteristic data, pipeline morphology, and damage data associated with each pipeline, and generates damage prediction information based on the decision tree model generated by the machine learning unit.

[0088] In the embodiment described above, we explained a case where the pipeline attribute data obtained from the GIS does not include irregularly shaped pipe sections or integrated sections. However, if the pipeline information obtained from the GIS contains data that allows for the identification of the piping structure, such as completion drawings, then the replacement of the piping pattern described above is not necessary, and the pipeline attribute data can be extracted from the completion drawings, etc. Design conditions that can be read from the completion drawings include valve locations, concrete protection, pipe ends, bridge pier sections, sections adjacent to other companies' pipelines (such as gas pipes), and sections connecting to structures.

[0089] In the embodiments described above, an example was explained in which the irregularly shaped pipe sections and integrated sections identified by the pipe morphology identification unit are incorporated into the attribute data of the pipes defined by each pipe number shown in the pipe diagram. However, the irregularly shaped pipe sections and integrated sections may be defined as pipes with different pipe numbers connected to each pipe, distinct from the pipes shown in the pipe diagram, and damage prediction may be performed by the damage prediction calculation unit.

[0090] In the embodiments described above, a pipeline damage prediction system due to earthquakes was explained, but the pipeline damage prediction system according to the present invention can also be applied to predicting pipeline damage due to natural disasters such as ground collapse, volcanic eruptions, and floods.

[0091] In the explanations of Figures 7A, 7B, and 7C, an example was described in which the longest value specified by the design standards, as shown in Figures 6A, 6B, and 6C, is applied as the length of the integrated section for composite piping. However, the length of the integrated section may be set to an optimal value depending on the design conditions (unequal forces).

[0092] The embodiments described above represent one aspect of the present invention, and the technical scope of the present invention is not limited based on this description. It goes without saying that the design can be modified as appropriate within the scope in which the effects of the present invention are achieved.

[0093] 10: Pipeline damage prediction system 11: Basic data storage unit 12: Damage prediction calculation unit

Claims

1. A pipeline damage prediction system that uses a computer to predict damage to pipelines caused by natural disasters, comprising: a basic data storage unit associated with a pipeline diagram showing the arrangement of pipelines at line segments and connection points of line segments, which stores pipeline attribute data indicating the attributes of the pipelines, ground characteristic data indicating the characteristics of the ground in which the pipelines are buried, and damage data indicating the degree of damage caused by natural disasters that have occurred in the past; a pipeline configuration identification unit that identifies a pipeline configuration in which irregularly shaped pipes are arranged at connection points based on the characteristics of the connection points of each pipeline shown in the pipeline diagram; and a damage prediction calculation unit that generates damage prediction information predicting damage to the pipelines based on the pipeline attribute data, ground characteristic data, damage data, and pipeline configurations stored in the basic data storage unit.

2. The pipeline damage prediction system according to claim 1, further comprising: an angle calculation unit for calculating the angle formed by each pipeline at the connection point; and an irregular pipe identification unit for identifying the shape of the pipeline located within a predetermined range from the connection point as an irregular pipe when the angle calculated by the angle calculation unit is inclined at or above a predetermined angle with respect to 180 degrees.

3. The pipeline damage prediction system according to claim 1, wherein the pipeline shape identification unit comprises a diameter change detection unit that detects a change in the diameter of each pipeline at the connection point, and a shape-deformed pipe identification unit that, when the diameter change detection unit detects a change in diameter, identifies the shape of the pipeline located within a predetermined range from the connection point as a shape-deformed pipe.

4. The pipeline damage prediction system according to claim 2 or 3, further comprising: a pipeline shape identification unit which, when the shape of the pipeline at the connection point is identified as an irregular pipe by the irregular pipe identification unit, identifies a pipeline located within a predetermined range from the irregular pipe as an integrated unit.

5. The pipeline damage prediction system according to claim 4, wherein the pipeline shape identification unit is further equipped with a straight pipe identification unit that identifies the connection point as a straight pipe connection point when the shape of the pipeline at the connection point cannot be identified as an irregular pipe by the irregular pipe identification unit.

6. The pipeline damage prediction system according to claim 1, wherein the damage prediction calculation unit comprises a machine learning unit that generates a decision tree model by ensemble learning of a plurality of datasets consisting of pipeline attribute data, ground characteristic data, pipeline morphology, and damage data associated with each pipeline, and generates the damage prediction information based on the decision tree model generated by the machine learning unit.

7. A pipeline damage prediction method that uses a computer to predict damage to pipelines caused by natural disasters, comprising: a basic data storage step of storing pipeline attribute data indicating the attributes of the pipeline, ground characteristic data indicating the characteristics of the ground in which the pipeline is buried, and damage data indicating the degree of damage caused by natural disasters that have occurred in the past, associated with a pipeline diagram showing the arrangement of the pipeline with line segments and connection points, in a basic data storage unit; a pipeline configuration identification step of identifying a pipeline configuration in which irregularly shaped pipes are arranged at connection points based on the characteristics of each connection point of the pipeline shown in the pipeline diagram; and a damage prediction calculation step of generating damage prediction information predicting damage to the pipeline based on the pipeline attribute data, ground characteristic data, damage data, and pipeline configuration stored in the basic data storage unit.

8. The pipeline damage prediction method according to claim 7, wherein the pipeline shape identification step comprises: an angle calculation step of calculating the angle formed by each pipeline at the connection point; and an irregular pipe identification step of identifying the shape of the pipeline located within a predetermined range from the connection point as an irregular pipe when the angle calculated in the angle calculation step is inclined by a predetermined angle or more with respect to 180 degrees.

9. The pipeline damage prediction method according to claim 7, wherein the pipeline morphology identification step includes: a diameter change detection step for detecting a change in the diameter of each pipeline at the connection point; and, when a change in diameter is detected by the diameter change detection step, an irregular pipe identification step for identifying the morphology of the pipeline located within a predetermined range from the connection point as an irregular pipe.

10. The pipeline damage prediction method according to claim 8 or 9, wherein the pipeline shape identification step includes, if the shape of the pipeline at the connection point is identified as an irregular pipe by the irregular pipe identification step, an integrated section identification step which includes identifying the pipeline located within a predetermined range from the irregular pipe as an integrated section.

11. The pipeline damage prediction method according to claim 10, wherein the pipeline shape identification step is performed by a straight pipe identification step which identifies the pipeline shape at the connection point as a straight pipe connection point if the shape of the pipeline at the connection point cannot be identified as a straight pipe by the irregular pipe identification step.

12. The pipeline damage prediction method according to claim 7, wherein the damage prediction calculation step involves executing a machine learning step to generate a decision tree model by ensemble learning of a plurality of datasets consisting of pipeline attribute data, ground characteristic data, pipeline morphology, and damage data associated with each pipeline, and generating the damage prediction information based on the decision tree model generated by the machine learning step.

13. A pipeline damage prediction program for a computer to function as follows: a basic data storage processing unit associated with a pipeline diagram showing the arrangement of pipelines with line segments and connection points, which stores pipeline attribute data indicating the attributes of the pipelines, ground characteristic data indicating the characteristics of the ground in which the pipelines are buried, and damage data indicating the degree of damage caused by natural disasters that have occurred in the past; a pipeline configuration identification unit which identifies a pipeline configuration in which irregularly shaped pipes are arranged at connection points based on the characteristics of the connection points of each pipeline shown in the pipeline diagram; and a damage prediction calculation unit which generates damage prediction information predicting damage to the pipelines based on the pipeline attribute data, ground characteristic data, damage data, and pipeline configurations stored in the basic data storage unit.

14. The pipeline damage prediction program according to claim 13, wherein the pipeline shape identification unit comprises an angle calculation unit that calculates the angle formed by each pipeline at the connection point, and an irregular pipe identification unit that identifies the shape of the pipeline located within a predetermined range from the connection point as an irregular pipe when the angle calculated by the angle calculation unit is inclined by a predetermined angle or more with respect to 180 degrees.

15. The pipeline damage prediction program according to claim 13, wherein the pipeline shape identification unit comprises a diameter change detection unit that detects a change in the diameter of each pipeline at the connection point, and a shape-deformed pipe identification unit that, when the diameter change detection unit detects a change in diameter, identifies the shape of the pipeline located within a predetermined range from the connection point as a shape-deformed pipe.

16. The pipeline damage prediction program according to claim 14 or 15, wherein the pipeline shape identification unit identifies the shape of the pipeline at the connection point as an irregular pipe using the irregular pipe identification unit, and then identifies the pipeline located within a predetermined range from the irregular pipe as an integrated unit.

17. The pipeline damage prediction program according to claim 16, wherein the pipeline shape identification unit includes a straight pipe identification unit that identifies the connection point as a straight pipe connection point when the shape of the pipeline at the connection point cannot be identified as a straight pipe by the irregular pipe identification unit.

18. The pipeline damage prediction program according to claim 13, wherein the damage prediction calculation unit comprises a machine learning unit that generates a decision tree model by ensemble learning of a plurality of datasets consisting of pipeline attribute data, ground characteristic data, pipeline morphology, and damage data associated with each pipeline, and generates the damage prediction information based on the decision tree model generated by the machine learning unit.

Images