Buried pipe leakage accident rate prediction device, buried pipe leakage accident rate prediction method, and program

Abstract

To provide a buried pipe leakage accident rate predication device capable of accurately predicting a leakage accident rate of a buried pipe for an arbitrary burial period.SOLUTION: A buried pipe leakage accident rate prediction device 3 comprises a pipe thickness excess probability calculation unit 52 and a leakage accident rate calculation unit 54. The pipe thickness excess probability calculation unit 52 calculates a pipe thickness excess probability of a buried pipe by inputting a burial environment, a burial period and a pipe thickness of the buried pipe to a pipe thickness excess probability prediction model 21. The water leakage accident rate calculation unit 54 uses the pipe thickness excess probability and a conversion factor 23 to calculate a water leakage accident rate of the buried pipe.SELECTED DRAWING: Figure 34

Description

Technical Field

[0001] The present disclosure relates to a buried pipe leakage accident rate prediction device, a buried pipe leakage accident rate prediction method, and a program.

Background Art

[0002] A pipe such as a water pipe is buried in the ground. The pipe is, for example, a cast iron pipe or a ductile pipe. While the pipe is used for a long period of time, the pipe corrodes. Japanese Unexamined Patent Application Publication No. 2007-107882 (Patent Document 1) discloses a pipe corrosion prediction method.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] An object of the present disclosure is to provide a buried pipe leakage accident rate prediction device, a buried pipe leakage accident rate prediction method, and a program that can more accurately predict the leakage accident rate of a buried pipe for any buried period.

Means for Solving the Problems

[0005] The buried pipe leakage accident rate prediction device of the present disclosure includes a pipe thickness exceeding probability calculation unit and a leakage accident rate calculation unit. The pipe thickness exceeding probability calculation unit calculates the pipe thickness exceeding probability of the buried pipe by inputting the buried environment, buried period, and pipe thickness of the buried pipe into the pipe thickness exceeding probability prediction model. The leakage accident rate calculation unit calculates the leakage accident rate of the buried pipe using the pipe thickness exceeding probability of the buried pipe and a conversion coefficient. The pipe thickness exceeding probability prediction model is generated according to the buried environment of the pipe and the pipe thickness of the pipe, and gives the pipe thickness exceeding probability of the pipe that continuously changes with respect to the continuous change of the buried period of the pipe. The conversion coefficient is a coefficient that converts the pipe thickness exceeding probability of the pipe or a first index calculable from the pipe thickness exceeding probability of the pipe into a second index calculable from the leakage accident rate of the pipe or the leakage accident rate of the pipe.

[0006] The buried pipe leakage accident rate prediction method of the present disclosure includes a step of calculating the pipe thickness exceeding probability of the buried pipe by inputting the buried environment, buried period, and pipe thickness of the buried pipe into the pipe thickness exceeding probability prediction model, and a step of calculating the leakage accident rate of the buried pipe using the pipe thickness exceeding probability of the buried pipe and a conversion coefficient. The pipe thickness exceeding probability prediction model is generated according to the buried environment of the pipe and the pipe thickness of the pipe, and gives the pipe thickness exceeding probability of the pipe that continuously changes with respect to the continuous change of the buried period of the pipe. The conversion coefficient is a coefficient that converts the pipe thickness exceeding probability of the pipe or a first index calculable from the pipe thickness exceeding probability of the pipe into a second index calculable from the leakage accident rate of the pipe or the leakage accident rate of the pipe.

[0007] The program of the present disclosure causes a processor to execute each step of the buried pipe leakage accident rate prediction method of the present disclosure.

Advantages of the Invention

[0008] According to the buried pipe leakage accident rate prediction device, buried pipe leakage accident rate prediction method, and program of the present disclosure, the leakage accident rate of the buried pipe can be predicted more accurately.

Brief Description of the Drawings

[0009]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Figure 22

Figure 23

Figure 24

Figure 25

Figure 26

Figure 27

Figure 28

Figure 29

Figure 30

Figure 31

Figure 32

Figure 33

Figure 34

Figure 35

Figure 36

Figure 37

Figure 38

Figure 39

Figure 40

Figure 41

Figure 42

Figure 43

Figure 44

Figure 45

Figure 46

Figure 47

Figure 48

Figure 49

Figure 50

Figure 51

Figure 52

Embodiments for Carrying Out the Invention

[0010] (Embodiment 1) <Buried Pipe Leakage Accident Rate Prediction System 1> Referring to FIG. 1, the buried pipe leakage accident rate prediction system 1 of this embodiment will be described. The buried pipe leakage accident rate prediction system 1 includes a buried pipe leakage accident rate prediction model generation device 2 and a buried pipe leakage accident rate prediction device 3.

[0011] <Buried Pipe Leakage Accident Rate Prediction Model Generation Device 2> Referring to FIGS. 1 to 3, the buried pipe leakage accident rate prediction model generation device 2 generates a buried pipe leakage accident rate prediction model 6 from reference pipe data 20 (refer to FIG. 4). The buried pipe leakage accident rate prediction model generation device 2 transmits it to the buried pipe leakage accident rate prediction device 3. The buried pipe leakage accident rate prediction model 6 includes, for example, a pipe thickness exceeding probability prediction model 21 (refer to FIGS. 5 and 6) and a conversion coefficient 23 (refer to FIG. 7).

[0012] <Hardware Configuration> Referring to FIG. 2, the hardware configuration of the buried pipe leakage accident rate prediction model generation device 2 will be described. The buried pipe leakage accident rate prediction model generation device 2 includes an input device 201, a processor 202, a memory 203, a display 204, a network controller 206, a storage medium drive 207, and a storage 210.

[0013] The input device 201 receives various input operations. The input device 201 is, for example, a keyboard, a mouse, or a touch panel.

[0014] The display 204 displays information necessary for the processing in the buried pipe leakage accident rate prediction model generation device 2, etc. The display 204 is, for example, an LCD (Liquid Crystal Display) or an organic EL (Electroluminescence) display.

[0015] The processor 202 executes the processes necessary for realizing the functions of the buried pipe leakage accident rate prediction model generation device 2 by executing the programs described later. The processor 202 is composed of, for example, a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit), etc.

[0016] The memory 203 provides a storage area for temporarily storing program codes or work memories, etc., when the processor 202 executes the programs described later. The memory 203 is, for example, a volatile memory device such as a DRAM (Dynamic Random Access Memory) or an SRAM (Static Random Access Memory).

[0017] The network controller 206 transmits and receives programs or data to and from any device including the buried pipe leakage accident rate prediction device 3 via the communication network 4 (see FIG. 1). For example, the network controller 206 transmits the buried pipe leakage accident rate prediction model 6 to the buried pipe leakage accident rate prediction device 3 via the communication network 4. The network controller 206 corresponds to any communication method such as Ethernet (registered trademark), wireless LAN (Local Area Network), or Bluetooth (registered trademark), etc.

[0018] The memory medium drive 207 is a device that reads programs or data stored in the memory medium 208. The memory medium drive 207 may further be a device that writes programs or data to the memory medium 208. The memory medium 208 is a non-transitory memory medium and stores programs or data in a non-volatile manner. The memory medium 208 is, for example, an optical memory medium such as an optical disk (e.g., CD-ROM or DVD-ROM), a semiconductor memory medium such as a flash memory or a USB (Universal Serial Bus) memory, a hard disk, a magnetic memory medium such as an FD (Flexible Disk) or a storage tape, or a magneto-optical memory medium such as an MO (Magneto-Optical) disk.

[0019] The storage 210 stores the reference pipe data 20 (see FIG. 4), the buried pipe leakage accident rate prediction model 6 (see FIG. 3), and programs executed in the processor 202, etc. This program includes the buried pipe leakage accident rate prediction model generation program 26 (see FIG. 3). The buried pipe leakage accident rate prediction model generation program 26 is a program for generating the buried pipe leakage accident rate prediction model 6 from the reference pipe data 20. The storage 210 is, for example, a non-volatile memory device such as a hard disk or an SSD.

[0020] A program (including the buried pipe leakage accident rate prediction model generation program 26 (see FIG. 3)) for realizing the functions of the buried pipe leakage accident rate prediction model generation device 2 may be stored in and distributed via the non-transitory memory medium 208 and installed in the storage 210. The program for realizing the functions of the buried pipe leakage accident rate prediction model generation device 2 may be downloaded to the buried pipe leakage accident rate prediction model generation device 2 via the Internet or an intranet.

[0021] In this embodiment, an example is shown in which a general-purpose computer (processor 202) realizes the functions of the buried pipe leakage accident rate prediction model generation device 2 by executing a program. However, the present invention is not limited to this, and all or part of the functions of the buried pipe leakage accident rate prediction model generation device 2 may be realized using an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field-Programmable Gate Array).

[0022] <Functional configuration> Referring to FIG. 3, the functional configuration of the buried pipe leakage accident rate prediction model generation device 2 will be described. The buried pipe leakage accident rate prediction model generation device 2 includes a storage unit 10 and a buried pipe leakage accident rate prediction model generation unit 16.

[0023] <Storage unit 10> The storage unit 10 is realized by, for example, at least one of a storage 210 (see FIG. 2) or a storage medium 208 (see FIG. 2). Referring to FIG. 3, the storage unit 10 includes a reference pipe data storage unit 11, a buried pipe leakage accident rate prediction model storage unit 12, a corrosion lag time storage unit 13, and a program storage unit 14.

[0024] Referring to FIG. 3, reference pipe data 20 (see FIG. 4) is stored in the reference pipe data storage unit 11. The reference pipe data 20 is survey data of a plurality of reference pipes obtained by excavating a plurality of reference pipes at a plurality of locations. The reference pipes are, for example, water pipes. The reference pipes are buried in the ground. The reference pipe data 20 includes, for example, survey data of reference pipes at about 5000 locations throughout Japan. The reference pipe data 20 may be provided by a customer or obtained by the user of the buried pipe leakage accident rate prediction system 1 or the buried pipe leakage accident rate prediction model generation device 2 surveying an area designated by the customer.

[0025] As shown in FIG. 4, the reference pipe data 20 includes the pipeline ID of the reference pipe, the buried environment, and the buried period T refIncluding the corrosion depth, material, nominal pipe thickness, pipeline length, and the number of water leakage accidents per unit time (e.g., 5 years). Pipeline ID, embedding environment, embedding period T ref The reference corrosion depth, nominal pipe thickness, pipeline length, and the number of water leakage accidents per unit time are associated with each other. Embedding period T ref T is the period during which the reference pipe is embedded. The reference corrosion depth is the corrosion depth of the reference pipe. Examples of the material of the pipe can be ductile or cast iron. The nominal pipe thickness of the pipe is the standard pipe thickness of the pipe. The pipeline length is the length of the pipe specified by the pipeline ID. The number of water leakage accidents per unit time is the number of water leakage accidents occurring in the pipe per unit time.

[0026] The embedding environment is the soil quality in which the pipe is embedded. The embedding environment is classified into four embedding environments A, B, C, and D according to the type of soil and soil resistivity. Embedding environment A is a soil quality with a soil resistivity of less than 1500 Ω·cm. Embedding environment B is a clay-based soil with a soil resistivity of 1500 Ω·cm or more. Embedding environment C is a silt-based soil with a soil resistivity of 1500 Ω·cm or more. Embedding environment D is a sand-based soil with a soil resistivity of 1500 Ω·cm or more.

[0027] For the following two reasons, the inventor classified the embedding environment of the pipe into four embedding environments A, B, C, and D. The first reason is that the inventor found that there is a statistically significant difference in the relationship between the corrosion depth of the pipe and the four embedding environments A - D. The second reason is that the number of reference pipe data 20 with the four embedding environments A - D accounts for the majority (80% or more) of the total number of collected reference pipe data 20, and it is found that it is useful to generate a buried pipe water leakage accident rate prediction model 6 for each of the four embedding environments A - D.

[0028] Referring to FIG. 3, the buried pipe water leakage accident rate prediction model 6 is stored in the buried pipe water leakage accident rate prediction model storage unit 12. The buried pipe water leakage accident rate prediction model 6 includes a pipe thickness excess probability prediction model 21 (FIGS. 5 and 6) and a conversion coefficient 23 (FIG. 7).

[0029] The pipe thickness exceeding probability prediction model 21 is generated according to the pipe embedding environment and the nominal pipe thickness. The pipe thickness exceeding probability is the probability that the corrosion depth of the pipe exceeds the nominal pipe thickness of the pipe. The pipe thickness exceeding probability prediction model 21 gives the pipe thickness exceeding probability of the pipe that continuously changes with the continuous change of the pipe embedding period. The pipe thickness exceeding probability prediction model 21 is a model that predicts the pipe thickness exceeding probability at the embedding period of any pipe within a predetermined range (for example, a range from 0 year to 70 years).

[0030] Referring to FIG. 5, an example of the data structure of the pipe thickness exceeding probability prediction model 21 will be described. The data of the pipe thickness exceeding probability prediction model 21 includes the embedding environment, the nominal pipe thickness, and the pipe thickness exceeding probability prediction model 21. The embedding environment, the nominal pipe thickness, and the pipe thickness exceeding probability prediction model 21 are associated with each other.

[0031] In FIG. 6, as an example of the pipe thickness exceeding probability prediction model 21, four pipe thickness exceeding probability prediction models are shown. Specifically, FIG. 6 shows a pipe thickness exceeding probability prediction model for a pipe with a nominal pipe thickness of 7.5 mm and embedded in the embedding environment A, a pipe thickness exceeding probability prediction model for a pipe with a nominal pipe thickness of 7.5 mm and embedded in the embedding environment B, a pipe thickness exceeding probability prediction model for a pipe with a nominal pipe thickness of 7.5 mm and embedded in the embedding environment C, and a pipe thickness exceeding probability prediction model for a pipe with a nominal pipe thickness of 7.5 mm and embedded in the embedding environment D.

[0032] Referring to FIG. 7, the conversion coefficient 23 of the present embodiment is a coefficient that converts the pipe thickness exceeding probability into a leakage accident rate. According to the material of the pipe, a plurality of conversion coefficients 23 may be stored in the buried pipe leakage accident rate prediction model storage unit 12. The plurality of conversion coefficients 23 includes, for example, a conversion coefficient p for ductile iron pipes and a conversion coefficient q for cast iron pipes.

[0033] Referring to FIG. 3, in the corrosion lag time storage unit 13, the probability Q(t L )(see FIG. 13) of the corrosion lag time and the cumulative relative frequency of the corrosion lag time t L (see FIG. 14) and the corrosion lag time t LThe data (see Fig. 16) is stored.

[0034] Referring to Fig. 3, the program storage unit 14 stores a program (for example, the buried pipe leakage accident rate prediction model generation program 26) for realizing the functions of the buried pipe leakage accident rate prediction model generation device 2.

[0035] <Buried pipe leakage accident rate prediction model generation unit 16> Referring to Fig. 3, the buried pipe leakage accident rate prediction model generation unit 16 is realized by executing the buried pipe leakage accident rate prediction model generation program 26 stored in the program storage unit 14 by the processor 202 (see Fig. 2). The buried pipe leakage accident rate prediction model generation unit 16 generates a buried pipe leakage accident rate prediction model 6 (see Fig. 3) from the reference pipe data 20 (see Fig. 4). The buried pipe leakage accident rate prediction model generation unit 16 includes a pipe thickness excess probability prediction model generation unit 17 and a conversion coefficient calculation unit 18.

[0036] (Pipe thickness excess probability prediction model generation unit 17) The pipe thickness excess probability prediction model generation unit 17 shown in Fig. 3 generates a pipe thickness excess probability prediction model 21 (see Figs. 5 and 6) from the reference pipe data 20 (see Fig. 4).

[0037] Referring to Figs. 8 to 22, a method for the pipe thickness excess probability prediction model generation unit 17 to generate the pipe thickness excess probability prediction model 21 from the reference pipe data 20 will be described.

[0038] Fig. 8 shows the reference pipe data 20 of the buried environment D, which is a part of the reference pipe data 20 (see Fig. 4). The reference pipe data 20 includes a large number of data with a reference corrosion depth of 0 mm, and it can be seen that the data has a large variation. Therefore, it is difficult to construct a buried pipe leakage accident rate prediction model 6 that can accurately predict the leakage accident rate of the buried pipe using the reference pipe data 20 as it is.

[0039] The inventor considered that the reference tube data 20 included a large number of data with a reference corrosion depth of 0 mm, and the reason for the large variation in the data was the coating film formed on the outer surface of the tube. That is, corrosion of the tube begins only after the coating film has been penetrated. The fact that the reference tube data 20 includes a large number of data with a reference corrosion depth of 0 mm is because there is a time (hereinafter referred to as "corrosion lag time t L ") from when the tube is buried until corrosion of the tube begins.

[0040] The corrosion lag time t L mainly depends on the buried environment of the tube, as well as the thickness and material of the coating film. The corrosion lag time t L that depends on the buried environment of the tube, as well as the thickness and material of the coating film, is considered to be statistically related to the buried environment, buried period, and corrosion depth of the tube. Therefore, the inventor used the corrosion lag time t L (see FIG. 16) estimated by a statistical method to correct the reference tube data 20 to generate corrected reference tube data (see FIG. 17), and decided to generate a tube thickness excess probability prediction model 21 (see FIGS. 5 and 6) based on the corrected reference tube data.

[0041] Referring to FIG. 9, the tube thickness excess probability prediction model generation unit 17 calculates the corrosion lag time t L corresponding to the buried environment, buried period, and reference corrosion depth from the reference tube data 20 (see FIG. 4) (step S1). Referring to FIG. 10, an example of step S1 will be described.

[0042] Referring to FIG. 10, the tube thickness excess probability prediction model generation unit 17 extracts the reference tube data 20 buried in a predetermined buried environment (for example, buried environment D) selected from the buried environments A, B, C, D from the reference tube data 20 stored in the reference tube data storage unit 11 (see FIG. 3) (step S11).

[0043] Referring to FIG. 10, for the reference tube data 20 extracted in step S11, the tube thickness excess probability prediction model generation unit 17 determines the buried period Tref Each time, calculate the ratio 1 - P(T of the number of data with a reference corrosion depth greater than 0 mm ref )(see FIG. 12), and perform non-linear regression on 1 - P(T ref )(step S12).

[0044] As shown in FIG. 11, P(T ref ) is the ratio of the number of data with a reference corrosion depth of 0 mm for each of the embedding periods T within a predetermined range (for example, one year). Specifically, the pipe wall thickness exceeding probability prediction model generation unit 17 counts the total number of data n ref for each of the embedding periods T within a predetermined range ref (T all ) and the number of data n0(T ref ) with a reference corrosion depth of 0 mm for each of the embedding periods T within a predetermined range ref (T ref ). The pipe wall thickness exceeding probability prediction model generation unit 17 calculates the ratio P(T all (T ref ) of n0(T ref ) to n ref (T ref ). P(T ref ) is given by n0(T all (T ref ) / n ref (T

[0045] As shown in FIG. 12, 1 - P(T ref ) asymptotically approaches from zero to 1 non-linearly as the embedding period T ref increases. Therefore, 1 - P(T ref ) is regressed using a non-linear regression model. Examples of such non-linear regression models include, for example, an exponential distribution model or a fractional function model. As an example of an exponential distribution model, there is an exponential distribution model represented by formula (1). α is a coefficient determined by regression. As an example of a fractional function model, there is a first-order fractional function model represented by formula (2). β and γ are coefficients determined by regression. In the present embodiment, the pipe wall thickness exceeding probability prediction model generation unit 17 uses 1 - P(T ref) is non-linearly regressed with an exponential distribution model represented by the following formula (1) using the least squares method or the like to calculate the coefficient α.

[0046] 1 - P(T ref ) = 1 - 1×exp(-αT ref ) (1) 1 - P(T ref ) = βT ref / (1 + γT ref ) (2) 1 - P(T ref )'s rate of change, that is, the differential value of 1 - P(T ref ) is the ratio of the reference pipes embedded in the embedding environment selected in step S11 that start to corrode during the embedding period T ref . The ratio of the reference pipes that start to corrode during the embedding period T ref is the ratio of the reference pipes embedded in the embedding environment selected in step S11 for which the corrosion lag time t L is the embedding period T ref . In other words, the differential value of 1 - P(T ref ) is the probability Q(T ref ) of the corrosion lag time in the embedding environment selected in step S11. The probability Q(T ref ) of the corrosion lag time in the embedding environment selected in step S11 is the probability that among the corrosion lag times t L of all the reference pipes embedded in the embedding environment selected in step S11, the corrosion lag time t L is the embedding period T ref .

[0047] Referring to FIG. 10, the pipe thickness exceedance probability prediction model generation unit 17 calculates the probability Q(t L ) of the corrosion lag time (see FIG. 13) (step S13). Specifically, the pipe thickness exceedance probability prediction model generation unit 17 calculates 1 - P(T refCalculate the differential curve of the regression curve (see Fig. 12). The pipe thickness exceeding probability prediction model generation unit 17 normalizes the differential curve so that the area of the region sandwiched between the differential curve and the straight line of Q = 0 becomes 1. Thus, the pipe thickness exceeding probability prediction model generation unit 17 calculates the probability Q(t L )(see Fig. 13) of the corrosion lag time. The pipe thickness exceeding probability prediction model generation unit 17 outputs the probability Q(t L ) of the corrosion lag time to the corrosion lag time storage unit 13 (see Fig. 3). The probability Q(t L ) of the corrosion lag time is stored in the corrosion lag time storage unit 13.

[0048] Referring to Fig. 10, the pipe thickness exceeding probability prediction model generation unit 17 calculates the cumulative relative frequency (see Fig. 14) of the corrosion lag time t L ) from the probability Q(t L ) of the corrosion lag time (step S14). For example, the cumulative relative frequency of the corrosion lag time t L for n years is the cumulative sum of the relative frequencies of the corrosion lag time t L that is n years or more, and is given by the sum of the probabilities Q(t L ) of the corrosion lag time t L that is n years or more. The pipe thickness exceeding probability prediction model generation unit 17 calculates the sum of the probabilities Q(t L ) of the corrosion lag time t L that is n years or more as the cumulative relative frequency of the corrosion lag time t L for n years. The pipe thickness exceeding probability prediction model generation unit 17 outputs the cumulative relative frequency of the corrosion lag time t L to the corrosion lag time storage unit 13. The cumulative relative frequency of the corrosion lag time t L is stored in the corrosion lag time storage unit 13. L The cumulative relative frequency of the corrosion lag time t

[0049] The pipe thickness exceeding probability prediction model generation unit 17 obtains the buried period T within a predetermined range from the reference pipe data 20 regarding the buried environment selected in step S11 refAnd for each reference corrosion depth within a predetermined range, the cumulative relative frequency of the reference corrosion depth (see FIG. 14) is calculated (step S15). The cumulative relative frequency of the reference corrosion depth within a predetermined range is the cumulative sum of the relative frequencies of the reference corrosion depths below the said predetermined range.

[0050] Specifically, as shown in FIG. 15, for each predetermined range of the embedding period T, the pipe thickness exceeding probability prediction model generation unit 17 sorts the reference pipe data 20 extracted in step S11 to obtain a plurality of data groups (step S15a). For example, the pipe thickness exceeding probability prediction model generation unit 17 sorts the reference pipe data 20 related to the embedding environment selected in step S11 for each 5-year embedding period T ref to obtain a plurality of data groups. The plurality of data groups include, for example, a data group having an embedding period T ref of more than 15 years and less than 20 years (see FIG. 14). ref (see FIG. 14).

[0051] For one of the plurality of data groups, the pipe thickness exceeding probability prediction model generation unit 17 calculates the cumulative relative frequency of the reference corrosion depth for each reference corrosion depth within a predetermined range (step S15b). The cumulative relative frequency of the reference corrosion depth within a predetermined range is the ratio of the number of data having a reference corrosion depth below the predetermined range among one of the plurality of data groups. The pipe thickness exceeding probability prediction model generation unit 17 calculates the ratio of the number of data having a reference corrosion depth below the predetermined range among one of the plurality of data groups as the cumulative relative frequency of the reference corrosion depth within the predetermined range.

[0052] For example, as shown in FIG. 14, for the data group having an embedding period T ref of more than 15 years and less than 20 years, the pipe thickness exceeding probability prediction model generation unit 17 calculates the cumulative relative frequency of the reference corrosion depth for each 0.5 mm of the reference corrosion depth. For example, in the data group having an embedding period T ref of more than 15 years and less than 20 years, the cumulative relative frequency of the reference corrosion depth of 1.0 mm or more and less than 1.5 mm is the ratio of the number of data having a reference corrosion depth of 0 mm or more and less than 1.5 mm among the said data group. That is, for the embedding period T refIn the data group having [the following], the cumulative relative frequency of the reference corrosion depth of 1.0 mm or more and less than 1.5 mm is calculated as the ratio of the sum of the number of data having a reference corrosion depth of 0 mm or more and less than 0.5 mm, the number of data having a reference corrosion depth of 0.5 mm or more and less than 1.0 mm, and the number of data having a reference corrosion depth of 1.0 mm or more and less than 1.5 mm in the data group to the total number of data in the data group.

[0053] The pipe thickness exceeding probability prediction model generation unit 17 performs step S15b for all of the plurality of data groups (step S15c). Thus, for the embedding environment selected in step S11, the embedding period T within a predetermined range (for example, 5 years) ref and the reference corrosion depth within a predetermined range (for example, 0.5 mm), the cumulative relative frequency of the reference corrosion depth is calculated.

[0054] Generally, the larger the corrosion depth of the pipe, the earlier the corrosion starts after the pipe is embedded, and the corrosion lag time t L is considered to be shorter. Therefore, it is considered that there is a correlation between the cumulative relative frequency of the corrosion lag time t L and the cumulative relative frequency of the reference corrosion depth. Referring to FIG. 10, the pipe thickness exceeding probability prediction model generation unit 17 calculates the corrosion lag time t L from the cumulative relative frequency of the corrosion lag time t calculated in step S14 and the cumulative relative frequency of the reference corrosion depth calculated in step S15 for the predetermined embedding environment selected in step S11, for the embedding period T within a predetermined range ref and the corrosion lag time t corresponding to the reference corrosion depth within a predetermined range. L is calculated (step S16).

[0055] Specifically, as shown by the dotted arrow in FIG. 14, the pipe thickness exceeding probability prediction model generation unit 17 determines that the cumulative relative frequency of the corrosion lag time t L is equal to the cumulative relative frequency of the reference corrosion depth within a predetermined range (for example, 0.0 mm or more and less than 0.5 mm) of the embedding period T within a predetermined range ref (for example, 15 years or more and less than 20 years), and the corrosion lag time t L(For example, 17.5 years) as the embedding period T within a predetermined range ref and the corrosion lag time t corresponding to a reference corrosion depth within a predetermined range (for example, a corrosion depth of 0.0 mm or more and less than 0.5 mm) L (For example, 17.5 years) is calculated.

[0056] Referring to FIG. 10, the pipe wall thickness exceeding probability prediction model generation unit 17 performs steps S11 - S16 for all the embedding environments A - D (step S17). Thus, the pipe wall thickness exceeding probability prediction model generation unit 17 obtains, from the reference pipe data 20 (refer to FIG. 4), the embedding environment, the embedding period T ref and the corrosion lag time t corresponding to the reference corrosion depth L is calculated.

[0057] Referring to FIG. 16, the pipe wall thickness exceeding probability prediction model generation unit 17 generates corrosion lag time data 27 including the embedding environment, the embedding period T ref , the reference corrosion depth, and the corrosion lag time t L . In the corrosion lag time data 27, the embedding environment, the embedding period T ref , the reference corrosion depth, and the corrosion lag time t L are associated with each other. The pipe wall thickness exceeding probability prediction model generation unit 17 outputs the corrosion lag time data 27 to the corrosion lag time storage unit 13 (refer to FIG. 3). The corrosion lag time data 27 is stored in the corrosion lag time storage unit 13.

[0058] Referring to FIG. 9, the pipe wall thickness exceeding probability prediction model generation unit 17 obtains corrected reference pipe data corresponding to the embedding environment (refer to FIG. 17) (step S2). The corrected reference pipe data includes the embedding environment, the corrected embedding period, and the reference corrosion depth. Specifically, the pipe wall thickness exceeding probability prediction model generation unit 17 reads out the embedding environment, the embedding period T ref , the reference corrosion depth, and the corrosion lag time t L from the corrosion lag time data 27 (refer to FIG. 16) stored in the corrosion lag time storage unit 13 (refer to FIG. 3). The pipe wall thickness exceeding probability prediction model generation unit 17 determines the embedding period T of the reference pipe data 20 (refer to FIGS. 4 and 8) refFrom the embedding environment and the embedding period T ref and the corrosion lag time t corresponding to the reference corrosion depth L (see Fig. 16), the corrected embedding period is calculated. In this way, corrected reference pipe data is obtained.

[0059] The corrected reference pipe data (see Fig. 17) is obtained by moving each point of the reference pipe data 20 (see Fig. 4) to the left by the corrosion lag time t L corresponding to each point. It can be seen that the corrected reference pipe data has less data variation than the reference pipe data 20 (see Fig. 8). Note that in the corrected reference pipe data shown in Fig. 17, the data with a reference corrosion depth of 0 mm has been deleted. The pipe thickness exceeding probability prediction model generation unit 17 outputs the corrected reference pipe data to the reference pipe data storage unit 11 (see Fig. 3). The corrected reference pipe data is stored in the reference pipe data storage unit 11.

[0060] Referring to Fig. 9, the pipe thickness exceeding probability prediction model generation unit 17 generates a corrosion depth exceeding probability prediction model (see Fig. 18) according to the embedding environment and the embedding period within a predetermined range (step S3). The corrosion depth exceeding probability is the probability that the corrosion depth of the pipe exceeds a predetermined depth. The corrosion depth exceeding probability prediction model is a model that predicts the corrosion depth exceeding probability in a predetermined embedding environment and within a predetermined range of the embedding period. Referring to Fig. 19, an example of step S3 for generating the corrosion depth exceeding probability prediction model will be described.

[0061] Referring to Fig. 19, the pipe thickness exceeding probability prediction model generation unit 17 extracts the reference pipe data 20 embedded in a predetermined embedding environment (for example, embedding environment D) selected from the embedding environments A, B, C, and D from the corrected reference pipe data (see Fig. 17) stored in the reference pipe data storage unit 11 (see Fig. 3) (step S31).

[0062] Referring to Figs. 19 and 20, the pipe thickness exceeding probability prediction model generation unit 17 obtains a basic regression line for regressing the corrected reference pipe data (step S32). As an example, the corrected reference pipe data is regressed by an exponential model represented by Equation (3) to obtain a basic regression line.

[0063] y = jt k (3) y represents the reference corrosion depth of the reference pipe, j and k represent coefficients, and t represents the modified burial period of the reference pipe.

[0064] Specifically, the logarithm of the modified burial period and the logarithm of the reference corrosion depth are calculated to obtain the double logarithm data of the modified reference pipe data (see Figure 20). Also, as the double logarithm of Equation (3), Equation (4) is obtained.

[0065] log y = log j + k × log t (4) The double logarithm data of the modified reference pipe data is regressed by Equation (4) using the least squares method to obtain the basic regression line (see Figure 20).

[0066] Referring to Figure 20, there is modified reference pipe data that deviates from the basic regression line. The reasons are (i) the variation in the corrosion rate of the reference pipe under a predetermined burial environment, and (ii) the distribution of the corrosion lag time t L for this.

[0067] Referring to Figure 19, based on the basic regression line, the variation in the corrosion rate, and the distribution of the corrosion lag time t L the corrosion depth exceedance prediction model generation unit 17 creates a corrosion depth exceedance prediction model according to the burial environment and the burial period within a predetermined range (step S33). For example, considering that the basic regression line has a certain distribution for each of the corrosion rate and the corrosion lag time t L (for example, the basic regression line is distributed with a certain probability density for each of the corrosion rate and the corrosion lag time t L ), a corrosion depth exceedance prediction model is created.

[0068] The variation in the corrosion rate is mainly reflected in the distribution of the basic regression line due to the variation in the reference corrosion depth in the modified reference tube data. The distribution of the basic regression line is reflected in the distribution of the coefficients j and k in Equation (3) or Equation (4). The distribution of the basic regression line due to the variation in the reference corrosion depth in the modified reference tube data can be expressed, for example, by the probability density function of the distribution of the basic regression line due to the variation in the reference corrosion depth in the modified reference tube data.

[0069] For example, referring to FIGS. 21 and 22, assume that the probability density function of the distribution of the basic regression line is a normal distribution. For each of the double-logarithmic data of the modified reference tube data, the tube thickness exceeding probability prediction model generation unit 17 calculates the deviation amount of the double-logarithmic data of the modified reference tube data from the basic regression line, and calculates the standard deviation σ of this deviation amount. The standard deviation of the probability density function is regarded as equal to the calculated standard deviation σ. Since the integral value of the probability density function over all the deviation amounts from the basic regression line is equal to 1 and the probability density function has a normal distribution with the standard deviation σ, the tube thickness exceeding probability prediction model generation unit 17 can calculate the probability density function of the distribution of the basic regression line.

[0070] The tube thickness exceeding probability prediction model generation unit 17 calculates the cumulative distribution function of the distribution of the basic regression line at the predetermined deviation amount by integrating the probability density function of the distribution of the basic regression line from a negative infinite deviation amount to the predetermined deviation amount.

[0071] The basic regression line is a regression line where the value of the cumulative distribution function (cumulative probability density) of the distribution of the basic regression line is at the 50th percentile, and it is called the 50th percentile regression line. A regression line where the value of the cumulative distribution function (cumulative probability density) of the distribution of the basic regression line is at the pth percentile is called the pth percentile regression line. The pipe wall thickness exceedance probability prediction model generation unit 17 obtains a plurality of percentile regression lines from the basic regression line and the cumulative distribution function of the distribution of the basic regression line. The plurality of percentile regression lines include, for example, the 5th percentile regression line, 6.25th percentile regression line, 12.5th percentile regression line, 18.75th percentile regression line, 25th percentile regression line, 31.25th percentile regression line, 37.5th percentile regression line, 43.75th percentile regression line, 50th percentile regression line, 56.25th percentile regression line, 62.5th percentile regression line, 68.75th percentile regression line, 75th percentile regression line, 81.25th percentile regression line, 87.5th percentile regression line, 93.75th percentile regression line, and 95th percentile regression line.

[0072] The pipe wall thickness exceedance probability prediction model generation unit 17 calculates, from each percentile regression line and the probability density function of the distribution of the basic regression line, the coefficients j and k (refer to items (B) and (C) in FIG. 23) for each of the percentile regression lines and the value of the probability density function (probability density) of each of the percentile regression lines (refer to item (D) in FIG. 23).

[0073] The pipe wall thickness exceedance probability prediction model generation unit 17 L converts the modified embedding period for each of the percentile regression lines into the pipe embedding period based on the distribution of the corrosion lag time t.

[0074] Specifically, the pipe thickness exceeding probability prediction model generation unit 17 sets a predetermined depth (for example, 0 mm or 1.0 mm) (see item (E) in FIG. 23). For each of the percentile regression lines, the pipe thickness exceeding probability prediction model generation unit 17 calculates the corrected burial period (see item (F) in FIG. 23) when the corrosion depth of the pipe is the predetermined depth. The pipe thickness exceeding probability prediction model generation unit 17 substitutes the predetermined depth into the reference corrosion depth y in Equation (3) or Equation (4) that defines each of the percentile regression lines to calculate the corrected burial period corresponding to the predetermined depth.

[0075] Corrosion lag time t L Due to the distribution of, the corrected burial period (see item (F) in FIG. 23) of each of the percentile regression lines corresponds to the burial periods of various pipes. Corrosion lag time t L The distribution can be expressed, for example, as the probability that the corrected burial period corresponds to the burial period of the pipe. Therefore, the pipe thickness exceeding probability prediction model generation unit 17 calculates the probability that the corrected burial period corresponds to the burial period of the pipe.

[0076] The pipe thickness exceeding probability prediction model generation unit 17 sets, for example, the corrosion lag time t L for a predetermined period (for example, 1 year, 2 years) (see item (G) in FIG. 23). The pipe thickness exceeding probability prediction model generation unit 17 calculates the sum of the corrected burial period (see item (F) in FIG. 23) and the corrosion lag time t L as the burial period of the pipe (see item (H) in FIG. 23). The probability that the corrected burial period corresponds to the burial period of the pipe is the probability Q(t L of the corrosion lag time for the said predetermined period L )(see FIG. 13). The pipe thickness exceeding probability prediction model generation unit 17 sets the probability Q(t L of the corrosion lag time stored in the corrosion lag time storage unit 13 (see FIG. 3) (see FIG. 13) as the probability that the corrected burial period corresponds to the burial period of the pipe (see item (I) in FIG. 23).

[0077] The pipe thickness exceeding probability prediction model generation unit 17 calculates the product of the probability density of the percentile regression line (see item (D) in FIG. 23) and the probability that the corrected embedding period corresponds to the pipe embedding period, that is, the probability Q(t L )(see item (I) in FIG. 23) as a probability index (see item (J) in FIG. 23). This product is proportional to the existence probability of the percentile regression line having the corrosion lag time t L and can be regarded as the probability index of the percentile regression line having the corrosion lag time t L .

[0078] The pipe thickness exceeding probability prediction model generation unit 17 calculates the sum of the probability indices (see item (J) in FIG. 23) for each corrosion depth (see item (E) in FIG. 23) and for each embedding period within a predetermined range (see item (H) in FIG. 23) as the corrosion depth reaching probability index. The corrosion depth reaching probability index is an index of the probability of reaching a predetermined corrosion depth within a predetermined range of the embedding period. Then, the pipe thickness exceeding probability prediction model generation unit 17 normalizes the corrosion depth reaching probability index so that the sum of the corrosion depth reaching probability indices over all corrosion depths within a predetermined range of the embedding period becomes 1, and calculates the corrosion depth reaching probability. The corrosion depth reaching probability is the probability of reaching a predetermined corrosion depth within a predetermined range of the embedding period. Thus, as shown in FIG. 24, the pipe thickness exceeding probability prediction model generation unit 17 generates data (see FIG. 24) showing the relationship between the embedding period within a predetermined range, the corrosion depth, and the corrosion depth reaching probability for the embedding environment selected in step S31.

[0079] The corrosion depth reaching probability can be regarded as the corrosion depth exceeding probability which is the probability that the corrosion depth exceeds a predetermined depth (for example, 1.0 mm) within a predetermined range of the embedding period. Thus, as shown in FIG. 18, the pipe thickness exceeding probability prediction model generation unit 17 generates, based on the data shown in FIG. 24, the relationship between the corrosion depth of the pipe, the embedding period of the pipe within a predetermined range, and the corrosion depth exceeding probability (see the corrosion depth exceeding probability data in FIG. 18) for the embedding environment selected in step S31.

[0080] The pipe thickness exceeding probability prediction model generation unit 17 regresses the relationship between the corrosion depth of the pipe, the embedding period of the pipe within a predetermined range, and the corrosion depth exceeding probability (refer to the data of the corrosion depth exceeding probability in FIG. 18), and generates a corrosion depth exceeding probability prediction model according to the embedding environment selected in step S31 and the embedding period within a predetermined range. The corrosion depth exceeding probability is the probability of occurrence of the first event that the corrosion depth exceeds a predetermined depth within a predetermined range of the embedding period and the second event that the corrosion depth does not exceed the predetermined depth within a predetermined range of the embedding period. Therefore, the corrosion depth exceeding probability may be non-linearly regressed by the cumulative distribution function of the binomial distribution. As an example, the pipe thickness exceeding probability prediction model generation unit 17 non-linearly regresses the relationship between the corrosion depth, the embedding period within a predetermined range, and the corrosion depth exceeding probability by the following formula (5), which is one of the cumulative distribution functions of the binomial distribution, to calculate the coefficients a and b.

[0081] R = 100×exp(a + bx) / (1 + exp(a + bx)) (5) R represents the corrosion depth exceeding probability (%), a and b represent the coefficients determined by regression, and x represents the corrosion depth.

[0082] Referring to FIG. 19, the pipe thickness exceeding probability prediction model generation unit 17 performs steps S31 - S33 for all the embedding environments A - D (step S34). In this way, the pipe thickness exceeding probability prediction model generation unit 17 generates a corrosion depth exceeding probability prediction model (refer to FIG. 18) according to the embedding environment and the embedding period within a predetermined range from the reference pipe data 20. The corrosion depth exceeding probability prediction model generated in step S3 (refer to FIG. 9) includes, for example, formula (5). In step S3, the coefficients a and b that define the corrosion depth exceeding probability prediction model are determined with respect to the embedding environment and the embedding period within a predetermined range (refer to the data of coefficient a in FIG. 25 and the data of coefficient b in FIG. 26).

[0083] Referring to FIG. 9, the pipe thickness exceedance probability prediction model generation unit 17 generates a corrosion depth exceedance probability prediction model according to the embedding environment and an arbitrary embedding period (step S4). Specifically, as shown in FIGS. 25 and 26, in step S4, the pipe thickness exceedance probability prediction model generation unit 17 regresses the coefficients a and b for an arbitrary embedding period of the pipe to determine the coefficients a and b for the embedding environment of the pipe and an arbitrary embedding period of the pipe. In this way, a corrosion depth exceedance probability prediction model according to the embedding environment and an arbitrary embedding period is generated.

[0084] The coefficient a may be regressed, for example, by a linear model represented by the following equation (6) as shown in FIG. 25.

[0085] a = c × t + d (6) t represents the embedding period, and c and d represent coefficients determined by regression.

[0086] The coefficient b may be regressed, for example, by an asymptotic model represented by the following equation (7) as shown in FIG. 26.

[0087] b = e - f × exp(-g × t) (7) t represents the embedding period, and e, f, and g represent coefficients determined by regression.

[0088] The corrosion depth exceedance probability prediction model generated in step S4 includes, for example, equations (5), (6), and (7). In step S4, the coefficients a and b defining the corrosion depth exceedance probability prediction model are determined for the embedding environment and an arbitrary embedding period (see the regression line of coefficient a in FIG. 25 and the regression curve of coefficient b in FIG. 26).

[0089] Referring to FIG. 9, the pipe thickness exceeding probability prediction model generation unit 17 generates a pipe thickness exceeding probability prediction model 21 (see FIG. 6) according to the embedding environment and the nominal pipe thickness (step S5). When the corrosion depth of the pipe reaches the nominal pipe thickness of the pipe, a water leakage accident occurs in the pipe. Therefore, the pipe thickness exceeding probability prediction model generation unit 17 inputs the embedding environment, the embedding period, and the nominal pipe thickness into the corrosion depth exceeding probability prediction model generated in step S4, and generates a pipe thickness exceeding probability prediction model 21 according to the embedding environment and the nominal pipe thickness.

[0090] For example, the pipe thickness exceeding probability prediction model generation unit 17 inputs the embedding environment and the embedding period into equations (6) and (7) of the corrosion depth exceeding probability prediction model obtained in step S4 to obtain coefficients a and b for the embedding environment and the embedding period. The pipe thickness exceeding probability prediction model generation unit 17 inputs the coefficients a and b and the nominal pipe thickness as the corrosion depth into equation (5) of the corrosion depth exceeding probability prediction model obtained in step S4, and calculates the corrosion depth exceeding probability of equation (5) as the pipe thickness exceeding probability. In this way, the pipe thickness exceeding probability prediction model generation unit 17 generates a pipe thickness exceeding probability prediction model 21 (see FIG. 6) according to the embedding environment and the nominal pipe thickness.

[0091] The pipe thickness exceeding probability prediction model generation unit 17 outputs the pipe thickness exceeding probability prediction model 21 (see FIG. 6) to the buried pipe water leakage accident rate prediction model storage unit 12 (see FIG. 3). The pipe thickness exceeding probability prediction model 21 is stored in the buried pipe water leakage accident rate prediction model storage unit 12.

[0092] (Conversion coefficient calculation unit 18) The conversion coefficient calculation unit 18 (see FIG. 3) calculates a conversion coefficient 23 (see FIG. 7) based on the reference pipe data 20 (see FIG. 4) and the pipe thickness exceeding probability obtained from the pipe thickness exceeding probability prediction model 21 (see FIGS. 5 and 6). In the present embodiment, the conversion coefficient 23 is a coefficient for converting the pipe thickness exceeding probability of the pipe into the water leakage accident rate of the pipe.

[0093] Referring to FIG. 27, a method for the conversion coefficient calculation unit 18 to calculate the conversion coefficient 23 will be described.

[0094] Referring to FIG. 27, the conversion coefficient calculation unit 18 selects the reference pipe data 20 of a specific material from the reference pipe data 20 stored in the reference pipe data storage unit 11 (see FIG. 3) (step S61). For example, the conversion coefficient calculation unit 18 selects the reference pipe data 20 of the ductile pipe from the reference pipe data 20.

[0095] Referring to FIG. 27, for the reference pipe data 20 selected in step S61, the conversion coefficient calculation unit 18 calculates the probability of pipe thickness exceeding for each pipeline ID (step S62). Specifically, the conversion coefficient calculation unit 18 obtains, from the reference pipe data 20 stored in the reference pipe data storage unit 11 (see FIG. 3), the buried environment, the buried period T ref and the nominal pipe thickness of the reference pipe having the specific material selected in S61. The conversion coefficient calculation unit 18 selects a pipe thickness exceeding probability model (see FIGS. 5 and 6) corresponding to the buried environment and the nominal pipe thickness of the reference pipe. The conversion coefficient calculation unit 18 inputs the buried period T ref of the reference pipe into the selected pipe thickness exceeding probability model to calculate the probability of pipe thickness exceeding for each pipeline ID.

[0096] Referring to FIG. 27, the conversion coefficient calculation unit 18 calculates the leakage accident rate for each probability of pipe thickness exceeding within a predetermined range (step S63).

[0097] Specifically, the conversion coefficient calculation unit 18 generates conversion coefficient calculation reference pipe data 28 (see FIG. 28) for the reference pipe having the specific material selected in step S61 from the pipeline ID, the buried period T ref , the pipeline length, and the number of leakage accident cases per unit time (for example, 5 years) of the reference pipe data 20 stored in the reference pipe data storage unit 11 (see FIG. 3) and the probability of pipe thickness exceeding calculated for each pipeline ID in S62. The conversion coefficient calculation reference pipe data 28 includes the pipeline ID, the probability of pipe thickness exceeding, the pipeline length, and the number of leakage accident cases per unit time (for example, 1 year). In the conversion coefficient calculation reference pipe data 28, the pipeline ID, the probability of pipe thickness exceeding, the buried period T ref , the pipeline length, and the number of leakage accident cases per unit time are associated with each other.

[0098] When the unit time (for example, 5 years) in the reference pipe data 20 (see FIG. 4) is different from the unit time (for example, 1 year) in the reference pipe data 28 for conversion coefficient calculation (see FIG. 28), the conversion coefficient calculation unit 18 reads out the number of water leakage accident cases per unit time of the reference pipe data 20 from the reference pipe data storage unit 11 (see FIG. 3), and converts the number of water leakage accident cases per unit time of the reference pipe data 20 into the number of water leakage accident cases per unit time of the reference pipe data 28 for conversion coefficient calculation. When the unit time in the reference pipe data 20 is the same as the unit time in the reference pipe data 28 for conversion coefficient calculation, the conversion coefficient calculation unit 18 reads out the number of water leakage accident cases per unit time of the reference pipe data 20 from the reference pipe data storage unit 11, and sets the number of water leakage accident cases per unit time of the reference pipe data 20 as the number of water leakage accident cases per unit time of the reference pipe data 28 for conversion coefficient calculation. The reference pipe data 28 for conversion coefficient calculation may further include the specific material selected in step S61.

[0099] The conversion coefficient calculation unit 18 groups the reference pipe data 28 for conversion coefficient calculation (see FIG. 28) for each range of a predetermined pipe thickness exceeding probability. The range of the predetermined pipe thickness exceeding probability is, for example, 1.5% or more of the pipe thickness exceeding probability range and 4.0% or less of the pipe thickness exceeding probability range. Therefore, each group has a sufficient number of data, and the number of data on the relationship between the pipe thickness exceeding probability and the water leakage accident rate required for regression in step S64 described later can be provided. It becomes possible to calculate the conversion coefficient 23 with higher accuracy.

[0100] For example, in the case where the specific material selected in step S61 is ductile, the conversion coefficient calculation unit 18 may group the reference pipe data 28 for conversion coefficient calculation for each range of 2.5% of the pipe thickness exceeding probability (see FIGS. 29 and 30). In the case where the specific material selected in step S61 is cast iron, the conversion coefficient calculation unit 18 may group the reference pipe data 28 for conversion coefficient calculation for each range of 2.5% of the pipe thickness exceeding probability (see FIGS. 31 and 32).

[0101] The conversion coefficient calculation unit 18 calculates the leakage accident rate for each group. Specifically, the conversion coefficient calculation unit 18 calculates the sum of the number of leakage accidents per unit time of the pipelines included in each group as the number of leakage accidents per unit time of each group. The conversion coefficient calculation unit 18 calculates the sum of the pipeline lengths for each group as the pipeline length of each group. The conversion coefficient calculation unit 18 divides the number of leakage accidents per unit time of each group by the pipeline length of each group to calculate the leakage accident rate of each group. In this way, the conversion coefficient calculation unit 18 obtains data on the relationship between the pipe thickness exceeding probability and the leakage accident rate (see the black dots in FIGS. 29 and 31).

[0102] The conversion coefficient calculation unit 18 calculates the conversion coefficient 23 for the specific material selected in S61 from the data on the relationship between the pipe thickness exceeding probability and the leakage accident rate (step S64).

[0103] The unit of the pipe thickness exceeding probability is a percentage and is a value independent of the pipeline length. Since the leakage accident rate is also normalized by the pipeline length, it is a value independent of the pipeline length. Therefore, the pipe thickness exceeding probability is considered to be proportional to the leakage accident rate. Thus, as shown in FIGS. 29 and 31, the conversion coefficient calculation unit 18 calculates, for example, the proportional coefficient obtained by linearly regressing the data on the relationship between the pipe thickness exceeding probability and the leakage accident rate as the conversion coefficient 23. The slope p of the regression line in FIG. 29 represents the conversion coefficient 23 for ductile iron pipes. The slope q of the regression line in FIG. 31 represents the conversion coefficient 23 for cast iron pipes.

[0104] As shown in FIGS. 30 and 32, the pipeline length of each group decreases as the probability of pipe thickness exceeding increases. The reason is that as the probability of pipe thickness exceeding increases, the number of reference pipe data 20 included in each group decreases. Therefore, as the probability of pipe thickness exceeding increases, the reliability of the data on the relationship between the probability of pipe thickness exceeding and the leakage accident rate (see the black dots in FIGS. 29 and 31) decreases. Thus, for the probability of pipe thickness exceeding below a predetermined value, the conversion coefficient 23 can be calculated with higher accuracy by linearly regressing the data on the relationship between the probability of pipe thickness exceeding and the leakage accident rate. The range of the probability of pipe thickness exceeding for performing this linear regression may be, for example, 20.0% or less, or 17.5% or less, or 15.0% or less (see FIGS. 29 and 31), or 12.5% or less, or 10.0% or less.

[0105] Referring to FIG. 27, the conversion coefficient calculation unit 18 determines whether all the conversion coefficients 23 for all materials included in the reference pipe data 20 (see FIG. 4) stored in the reference pipe data storage unit 11 (see FIG. 3) have been calculated (step S69). If not all the conversion coefficients 23 for all materials have been calculated, it returns to S61, selects the reference pipe data 20 of a specific material for which the conversion coefficient 23 has not yet been calculated, and performs S62 to S64 again. When all the conversion coefficients 23 for all materials are calculated, the calculation of the conversion coefficient 23 ends. The conversion coefficient calculation unit 18 outputs the calculated conversion coefficient 23 to the buried pipe leakage accident rate prediction model storage unit 12 (see FIG. 3). The conversion coefficient 23 is stored in the buried pipe leakage accident rate prediction model storage unit 12.

[0106] Referring to FIG. 1, the buried pipe leakage accident rate prediction model generation device 2 transmits the buried pipe leakage accident rate prediction model 6 (pipe thickness exceeding probability prediction model 21 and conversion coefficient 23) to the buried pipe leakage accident rate prediction device 3.

[0107] <Buried pipe leakage accident rate prediction device 3> Referring to FIGS. 1, 33, and 34, the buried pipe leakage accident rate prediction device 3 receives the buried pipe leakage accident rate prediction model 6 from the buried pipe leakage accident rate prediction model generation device 2. The buried pipe leakage accident rate prediction device 3 calculates the leakage accident rate of the buried pipe (refer to FIGS. 41 and 42) using the buried pipe leakage accident rate prediction model 6.

[0108] <Hardware Configuration> Referring to FIG. 33, the hardware configuration of the buried pipe leakage accident rate prediction device 3 will be described. The buried pipe leakage accident rate prediction device 3 includes an input device 301, a processor 302, a memory 303, a display 304, a network controller 306, a storage medium drive 307, and a storage 310.

[0109] The input device 301 accepts various input operations. The input device 301 is, for example, a keyboard, a mouse, or a touch panel.

[0110] The display 304 displays information necessary for the processing in the buried pipe leakage accident rate prediction device 3. The display 304 displays, for example, the leakage accident rate prediction result 60 (refer to FIGS. 41 or 42) described later. The display 304 is, for example, an LCD (Liquid Crystal Display) or an organic EL (Electroluminescence) display.

[0111] The processor 302 executes the processes necessary for realizing the functions of the buried pipe leakage accident rate prediction device 3 by executing the program described later. The processor 302 is composed of, for example, a CPU or a GPU.

[0112] The memory 303 provides a storage area for temporarily storing program codes or work memories when the processor 302 executes a program. The memory 303 is, for example, a volatile memory device such as a DRAM or an SRAM.

[0113] The network controller 306 transmits and receives programs or data to and from any device including the buried pipe leakage accident rate prediction model generation device 2 via the communication network 4 (see FIG. 1). For example, the network controller 306 receives the buried pipe leakage accident rate prediction model 6 from the buried pipe leakage accident rate prediction model generation device 2 via the communication network 4. The network controller 306 supports any communication method such as, for example, Ethernet (registered trademark), wireless LAN, or Bluetooth (registered trademark).

[0114] The memory media drive 307 is a device that reads programs or data stored in the memory media 308. The memory media drive 307 may further be a device that writes programs or data to the memory media 308. The memory media 308 is a non-transitory memory media and stores programs or data non-volatilely. The memory media 308 is, for example, an optical memory media such as an optical disk (e.g., CD-ROM or DVD-ROM), a semiconductor memory media such as a flash memory or a USB memory, a magnetic memory media such as a hard disk, an FD, or a storage tape, or a magneto-optical memory media such as an MO disk.

[0115] The storage 310 stores the buried pipe data 40 (see FIGS. 34 to 36), the preprocessed buried pipe data 46 (see FIG. 40), the buried pipe leakage accident rate prediction model 6 (see FIG. 34), and programs executed in the processor 302, etc. This program includes the buried pipe leakage accident rate prediction program 48 (see FIG. 34). The buried pipe leakage accident rate prediction program 48 is a program for calculating the leakage accident rate of the buried pipe from the buried pipe data 40. The storage 310 is, for example, a non-volatile memory device such as a hard disk or an SSD.

[0116] The program for realizing the functions of the buried pipe leakage accident rate prediction device 3 may be stored in a non-volatile storage medium 308, distributed, and installed in the storage 310. The program for realizing the functions of the buried pipe leakage accident rate prediction device 3 may be downloaded to the buried pipe leakage accident rate prediction device 3 via the Internet or an intranet.

[0117] In this embodiment, an example is shown in which a general-purpose computer (processor 302) realizes the functions of the buried pipe leakage accident rate prediction device 3 by executing a program. However, the present invention is not limited to this, and all or part of the functions of the buried pipe leakage accident rate prediction device 3 may be realized using an integrated circuit such as an ASIC or an FPGA.

[0118] <Functional configuration> With reference to FIG. 34, an example of the functional configuration of the buried pipe leakage accident rate prediction device 3 will be described. The buried pipe leakage accident rate prediction device 3 includes a storage unit 30, a buried pipe data reception unit 50, a buried pipe data preprocessing unit 51, a pipe thickness excess probability calculation unit 52, a leakage accident rate calculation unit 54, and a leakage accident rate prediction result output unit 57.

[0119] <Storage unit 30> The storage unit 30 is realized by at least one of a storage 310 (see FIG. 33) or a storage medium 308 (see FIG. 33). As shown in FIG. 34, the storage unit 30 includes a buried pipe data storage unit 31, a nominal pipe thickness database unit 32, a buried environment map storage unit 33, a preprocessed buried pipe data storage unit 34, a buried pipe leakage accident rate prediction model storage unit 36, a leakage accident rate storage unit 37, and a program storage unit 38.

[0120] With reference to FIG. 34, the buried pipe data storage unit 31 stores data of buried pipes received from customers (hereinafter referred to as "buried pipe data 40"). The buried pipe is, for example, a water pipe. The buried pipe is buried in the soil. The buried pipe data 40 includes, for example, a pipeline map 41 (see FIG. 35) and buried pipe attribute data 42 (see FIG. 36).

[0121] As shown in FIG. 35, in the pipeline map 41, the positions of the buried pipes are displayed on the map for each pipeline ID of the buried pipes.

[0122] Referring to FIG. 36, the buried pipe attribute data 42 includes the pipeline ID of the buried pipe, the laying (burial) year, the nominal diameter, the joint form, the type of pipe thickness, the pipeline length, and the material. In the buried pipe attribute data 42, the pipeline ID, the laying (burial) year, the nominal diameter, the joint form, the type of pipe thickness, the pipeline length, and the material are associated with each other. The laying (burial) year of the buried pipe is the year when the buried pipe was laid (buried). Examples of the joint form can include A-shaped, K-shaped, T-shaped, or NS-shaped, etc. Examples of the type of pipe thickness can include type 1, type 2, or type 3, etc. The pipeline length is the length of the pipe specified by the pipeline ID. Examples of the material of the buried pipe can include ductile or cast iron.

[0123] Referring to FIG. 34, the nominal pipe thickness database section 32 stores nominal pipe thickness data 43 (refer to FIG. 37) including the laying year of the pipe, the nominal diameter, the joint form, the type of pipe thickness, and the nominal pipe thickness. In the nominal pipe thickness data 43, the laying year of the pipe, the nominal diameter, the joint form, the type of pipe thickness, and the nominal pipe thickness are associated with each other. The nominal pipe thickness of the pipe is the standard pipe thickness of the pipe.

[0124] Referring to FIG. 34, the buried environment map storage section 33 stores a buried environment map 44 (refer to FIG. 38). The buried environment map 44 is a map showing the positions of the buried environments A, B, C, and D described above.

[0125] The buried environment map 44 can be created, for example, from a generally available ground information map (not shown) and ground-buried environment correspondence data 45 (see Fig. 39). The ground information map is, for example, a land classification survey provided by a public agency such as the Ministry of Land, Infrastructure, Transport and Tourism. Ground information such as the geology and topography of the surface layer of the ground (see Fig. 39) is shown on the ground information map. The inventor has discovered that there is a statistical correlation between the buried environments A, B, C, D of the buried pipe and the ground information with respect to the corrosion of the buried pipe. Therefore, the inventor has created the ground-buried environment correspondence data 45 showing the correspondence between the buried environments A, B, C, D and the ground information. The buried environment map 44 is created by applying the ground-buried environment correspondence data 45 to the ground information map.

[0126] Referring to Fig. 34, the preprocessed buried pipe data storage unit 34 stores preprocessed buried pipe data 46 (see Fig. 40) to be described later.

[0127] Referring to Fig. 34, the buried pipe leakage accident rate prediction model storage unit 36 stores a buried pipe leakage accident rate prediction model 6. The buried pipe leakage accident rate prediction device 3 receives the buried pipe leakage accident rate prediction model 6 from the buried pipe leakage accident rate prediction model generation device 2. The buried pipe leakage accident rate prediction model 6 includes, for example, a pipe thickness excess probability prediction model 21 (see Figs. 5 and 6) and a conversion coefficient 23 (see Fig. 7).

[0128] Referring to Fig. 34, the leakage accident rate storage unit 37 stores the leakage accident rate of the buried pipe calculated by the leakage accident rate calculation unit 54 (see Fig. 34). As shown in Fig. 41, the leakage accident rate of the buried pipe is associated with the pipeline ID of the buried pipe and stored in the leakage accident rate storage unit 37.

[0129] Referring to Fig. 34, the program storage unit 38 stores a program (for example, a buried pipe leakage accident rate prediction program 48) for realizing the functions of the buried pipe leakage accident rate prediction device 3.

[0130] <Buried pipe data reception unit 50> Referring to FIG. 34, the buried pipe data reception unit 50 receives buried pipe data 40 (refer to FIGS. 34 to 36) from a customer. The buried pipe data 40 is stored in the buried pipe data storage unit 31. The buried pipe data 40 may be stored in a storage medium 308 (refer to FIG. 33) provided by the customer. The buried pipe data 40 may be stored in a storage 310 (refer to FIG. 33) in advance.

[0131] <Buried pipe data preprocessing unit 51> Referring to FIG. 34, the buried pipe data preprocessing unit 51 creates preprocessed buried pipe data 46 (refer to FIG. 40) from the buried pipe data 40 (refer to FIGS. 34 to 36). The preprocessed buried pipe data 46 includes, for example, the pipeline ID of the buried pipe, the buried environment, the buried period T, the nominal pipe thickness, and the material. The buried environment is the soil type in which the buried pipe is buried. As already described, the buried environment is classified into four buried environments A, B, C, and D according to the type of soil and the soil resistivity. The buried period T is the period during which the buried pipe has been buried. For example, when obtaining the prediction result of the leakage accident rate of the buried pipe in the current year (the year in which the prediction of the leakage accident rate of the buried pipe is executed), the buried period T is the difference between the current year (the year in which the prediction of the leakage accident rate of the buried pipe is executed) and the laying year of the buried pipe (refer to FIG. 36) stored in the buried pipe data storage unit 31 (refer to FIG. 34). When obtaining the prediction result of the leakage accident rate of the buried pipe in a future year, the buried period T is the difference between the future year and the laying year of the buried pipe (refer to FIG. 36) stored in the buried pipe data storage unit 31 (refer to FIG. 34).

[0132] <Pipe thickness excess probability calculation unit 52> Referring to FIG. 34, the pipe thickness excess probability calculation unit 52 calculates the pipe thickness excess probability of the buried pipe for each pipeline ID. Specifically, the pipe thickness excess probability calculation unit 52 inputs the buried environment, the buried period, and the nominal pipe thickness of the buried pipe into the pipe thickness excess probability prediction model 21 stored in the buried pipe leakage accident rate prediction model storage unit 36 (refer to FIG. 34) to calculate the pipe thickness excess probability of the buried pipe.

[0133] <Leakage accident rate calculation unit 54> Referring to FIG. 34, the leakage accident rate calculation unit 54 calculates the leakage accident rate of the buried pipe for each pipeline ID from the preprocessed buried pipe data 46. The leakage accident rate of the buried pipe is the number of leakage accident cases of the buried pipe per unit time (for example, 1 year) and per unit distance (for example, 1 km). The leakage accident rate calculation unit 54 includes a conversion coefficient selection unit 55.

[0134] <conversion coefficient selection unit 55> Referring to FIG. 34, the conversion coefficient selection unit 55 selects the conversion coefficient 23 (refer to FIG. 7) corresponding to the material of the buried pipe from the conversion coefficients 23 stored in the buried pipe leakage accident rate prediction model storage unit 36 (refer to FIG. 34).

[0135] <leakage accident rate prediction result output unit 57> The leakage accident rate prediction result output unit 57 outputs the leakage accident rate prediction result 60 of the buried pipe to at least one of the display 304, the storage medium 308, or the storage 310 shown in FIG. 33. The leakage accident rate prediction result 60 may be, for example, a leakage accident rate prediction table 61 (refer to FIG. 41) or a leakage accident rate prediction map 62 (refer to FIG. 42). In the leakage accident rate prediction table 61, the pipeline ID of the buried pipe and the leakage accident rate are associated with each other. In the leakage accident rate prediction map 62, the position of the buried pipe and the leakage accident probability are displayed on the map. The leakage accident rate prediction map 62 is created by the leakage accident rate prediction result output unit 57 from the leakage accident rate of the buried pipe (FIG. 41) associated with the pipeline ID and the pipeline map 41 (refer to FIG. 35) including the pipeline ID and position of the buried pipe.

[0136] <method for predicting the leakage accident rate of buried pipes> Referring to FIGS. 43 and 44, the method for predicting the leakage accident rate of the buried pipe according to the present embodiment will be described.

[0137] Referring to FIG. 43, the buried pipe data preprocessing unit 51 creates preprocessed buried pipe data 46 (refer to FIG. 40) from the buried pipe data 40 (refer to FIGS. 34 to 36) stored in the buried pipe data storage unit 31 (refer to FIG. 34) (step S81).

[0138] Specifically, the buried pipe data preprocessing unit 51 calculates the buried period T of the buried pipe for each pipeline ID. For example, when obtaining the prediction result of the leakage accident rate of the buried pipe in the current year (the year when the prediction of the leakage accident rate of the buried pipe is executed), the buried pipe data preprocessing unit 51 calculates the difference between the current year stored in the storage unit 30 and the laying year of the buried pipe (refer to FIG. 36) stored in the buried pipe data storage unit 31 (refer to FIG. 34) as the buried period T of the buried pipe (refer to FIG. 40). When obtaining the prediction result of the leakage accident rate of the buried pipe in a future year, the buried pipe data preprocessing unit 51 calculates the difference between the future year received by the input device 301 (refer to FIG. 33) and stored in the storage unit 30 and the laying year of the buried pipe (refer to FIG. 36) stored in the buried pipe data storage unit 31 (refer to FIG. 34) as the buried period T of the buried pipe (refer to FIG. 40).

[0139] The buried pipe data preprocessing unit 51 obtains the nominal pipe thickness of the buried pipe for each pipeline ID from the laying (buried) year, nominal diameter, joint form, and type of pipe thickness (refer to FIG. 36) stored in the buried pipe data storage unit 31 (refer to FIG. 34) and the nominal pipe thickness data 43 (refer to FIG. 37) stored in the nominal pipe thickness database unit 32 (refer to FIG. 34).

[0140] The buried pipe data preprocessing unit 51 obtains the buried environment of the buried pipe for each pipeline ID from the pipeline map 41 (refer to FIG. 35) stored in the buried pipe data storage unit 31 (refer to FIG. 34) and the buried environment map 44 (refer to FIG. 38) stored in the buried environment map storage unit 33. The buried pipe data preprocessing unit 51 acquires the material from the buried pipe data 40 shown in FIG. 34 (refer to FIGS. 34 to 36). In this way, the buried pipe data preprocessing unit 51 creates the preprocessed buried pipe data 46 (refer to FIG. 40) from the buried pipe data 40. The buried pipe data preprocessing unit 51 outputs the preprocessed buried pipe data 46 to the preprocessed buried pipe data storage unit 34 (refer to FIG. 34). The preprocessed buried pipe data 46 is stored in the preprocessed buried pipe data storage unit 34.

[0141] Referring to FIG. 43, the pipe thickness exceeding probability calculation unit 52 calculates the pipe thickness exceeding probability of the buried pipe for each pipeline ID (step S82). Specifically, the pipe thickness exceeding probability calculation unit 52 reads out the pipeline ID, the buried environment, the buried period T, and the nominal pipe thickness from the pre-processed buried pipe data 46 (FIG. 40) stored in the pre-processed buried pipe data storage unit 34. Among the pipe thickness exceeding probability prediction models 21 stored in the buried pipe leakage accident rate prediction model storage unit 36 (refer to FIG. 34), the pipe thickness exceeding probability prediction model 21 corresponding to the read buried environment and nominal pipe thickness is selected by the pipe thickness exceeding probability calculation unit 52. The pipe thickness exceeding probability calculation unit 52 inputs the buried period T of the buried pipe into the selected pipe thickness exceeding probability prediction model 21, and calculates the pipe thickness exceeding probability of the buried pipe for each pipeline ID.

[0142] Referring to FIG. 43, the leakage accident rate calculation unit 54 calculates the leakage accident rate of the buried pipe for each pipeline ID (step S83). Specifically, referring to FIG. 44, the leakage accident rate calculation unit 54 reads out the pipeline ID and the material of the buried pipe from the pre-processed buried pipe data 46 (FIG. 40) stored in the pre-processed buried pipe data storage unit 34 (refer to FIG. 34). The conversion coefficient selection unit 55 selects the conversion coefficient 23 corresponding to the material of the buried pipe from the conversion coefficients 23 (refer to FIG. 7) stored in the buried pipe leakage accident rate prediction model storage unit 36 (refer to FIG. 34) (step S83a). The leakage accident rate calculation unit 54 multiplies the pipe thickness exceeding probability of the buried pipe calculated in step S82 by the conversion coefficient 23 selected in step S83a for each pipeline ID, and calculates the leakage accident rate of the buried pipe (step S83b).

[0143] The leakage accident rate calculation unit 54 outputs the leakage accident rate of the buried pipe calculated in step S83b to the leakage accident rate storage unit 37 (refer to FIG. 34). As shown in FIG. 41, the leakage accident rate of the buried pipe is stored in the leakage accident rate storage unit 37 in association with the pipeline ID.

[0144] Referring to FIG. 43, the leakage accident rate prediction result output unit 57 outputs the leakage accident rate prediction result 60 to at least one of the display 304, the storage medium 308, or the storage 310 shown in FIG. 33 (step S84). The leakage accident rate prediction result 60 may be, for example, a leakage accident rate prediction table 61 (see FIG. 41) or a leakage accident rate prediction map 62 (see FIG. 42).

[0145] The buried pipe leakage accident rate prediction program 48 (see FIG. 34) causes the processor 302 (see FIG. 33) to execute the buried pipe leakage accident rate prediction method of the present embodiment. The computer-readable recording medium (non-transitory computer-readable recording medium, for example, the storage medium 308) of the present embodiment stores a program for causing the processor 302 to execute the buried pipe leakage accident rate prediction method of the present embodiment.

[0146] The effects of the buried pipe leakage accident rate prediction device 3, the buried pipe leakage accident rate prediction method, and the program of the present embodiment will be described.

[0147] The buried pipe leakage accident rate prediction device 3 of this embodiment includes a pipe thickness exceeding probability calculation unit 52 and a leakage accident rate calculation unit 54. The pipe thickness exceeding probability calculation unit 52 calculates the pipe thickness exceeding probability of the buried pipe by inputting the buried environment, buried period, and pipe thickness (for example, nominal pipe thickness) of the buried pipe into the pipe thickness exceeding probability prediction model 21. The leakage accident rate calculation unit 54 calculates the leakage accident rate of the buried pipe using the pipe thickness exceeding probability of the buried pipe and the conversion coefficient 23. The pipe thickness exceeding probability prediction model 21 is generated according to the buried environment of the pipe and the pipe thickness (for example, nominal pipe thickness) of the pipe, and gives the pipe thickness exceeding probability of the pipe that continuously changes with respect to the continuous change of the buried period of the pipe. The pipe thickness exceeding probability of the pipe is the probability that the corrosion depth of the pipe exceeds the pipe thickness (for example, nominal pipe thickness) of the pipe. The pipe thickness exceeding probability of the buried pipe is the probability that the corrosion depth of the buried pipe exceeds the pipe thickness (for example, nominal pipe thickness) of the buried pipe. The conversion coefficient 23 is a coefficient that converts the pipe thickness exceeding probability of the pipe or a first index calculable from the pipe thickness exceeding probability of the pipe into a second index calculable from the leakage accident rate of the pipe or the leakage accident rate of the pipe. The leakage accident rate of the pipe is the number of pipe leakage accidents per unit time and per unit distance. The leakage accident rate of the buried pipe is the number of buried pipe leakage accidents per unit time and per unit distance.

[0148] According to the buried pipe leakage accident rate prediction device 3 of this embodiment, the leakage accident rate of the buried pipe can be predicted more accurately for any buried period.

[0149] In the buried pipe leakage accident rate prediction device 3 of this embodiment, the leakage accident rate calculation unit 54 includes a conversion coefficient selection unit 55 that selects the conversion coefficient 23 corresponding to the material of the buried pipe.

[0150] According to the buried pipe leakage accident rate prediction device 3 of this embodiment, the leakage accident rate of the buried pipe can be predicted more accurately for any buried period according to the material of the buried pipe.

[0151] In the buried pipe leakage accident rate prediction device 3 of this embodiment, the conversion coefficient 23 is a coefficient that converts the pipe thickness exceeding probability of the pipe into the leakage accident rate of the pipe. The leakage accident rate calculation unit 54 calculates the leakage accident rate of the buried pipe by multiplying the pipe thickness exceeding probability of the buried pipe and the conversion coefficient 23.

[0152] Therefore, the leakage accident rate of the buried pipe can be calculated more simply. The calculation process of the leakage accident rate of the buried pipe is clear, and the reliability of customers for the leakage accident rate of the buried pipe calculated by the buried pipe leakage accident rate prediction device 3 of the present embodiment can be improved.

[0153] The buried pipe leakage accident rate prediction method of the present embodiment includes a step (step S82) of calculating the probability of pipe thickness exceeding of the buried pipe by inputting the buried environment, the buried period, and the pipe thickness (for example, the nominal pipe thickness) of the buried pipe into the probability prediction model 21 of pipe thickness exceeding, and a step (step S83) of calculating the leakage accident rate of the buried pipe by using the probability of pipe thickness exceeding of the buried pipe and the conversion coefficient 23. The probability prediction model 21 of pipe thickness exceeding is generated according to the buried environment of the pipe and the pipe thickness (for example, the nominal pipe thickness) of the pipe, and gives the probability of pipe thickness exceeding of the pipe that continuously changes with respect to the continuous change of the buried period of the pipe. The probability of pipe thickness exceeding of the pipe is the probability that the corrosion depth of the pipe exceeds the pipe thickness (for example, the nominal pipe thickness) of the pipe. The probability of pipe thickness exceeding of the buried pipe is the probability that the corrosion depth of the buried pipe exceeds the pipe thickness (for example, the nominal pipe thickness) of the buried pipe. The conversion coefficient 23 is a coefficient that converts the probability of pipe thickness exceeding of the pipe into the leakage accident rate of the pipe. The leakage accident rate of the pipe is the number of leakage accidents of the pipe per unit time and per unit distance. The leakage accident rate of the buried pipe is the number of leakage accidents of the buried pipe per unit time and per unit distance.

[0154] According to the buried pipe leakage accident rate prediction method of the present embodiment, the leakage accident rate of the buried pipe can be predicted more accurately for any buried period.

[0155] In the buried pipe leakage accident rate prediction method of the present embodiment, the step (step S83) of calculating the leakage accident rate of the buried pipe includes a step (step S83a) of selecting the conversion coefficient 23 corresponding to the material of the buried pipe.

[0156] According to the buried pipe leakage accident rate prediction method of the present embodiment, the leakage accident rate of the buried pipe can be predicted more accurately for any buried period according to the material of the buried pipe.

[0157] In the method for predicting the buried pipe leakage accident rate according to this embodiment, the conversion coefficient 23 is a coefficient for converting the probability of the pipe thickness exceeding the standard of the pipe into the leakage accident rate of the pipe. The step of calculating the leakage accident rate of the buried pipe (step S83) is a step of calculating the leakage accident rate of the buried pipe by multiplying the probability of the pipe thickness exceeding the standard of the buried pipe by the conversion coefficient 23.

[0158] Therefore, the leakage accident rate of the buried pipe can be calculated more simply from the probability of the pipe thickness exceeding the standard of the buried pipe. The calculation process of the leakage accident rate of the buried pipe is clear, and the reliability of customers with respect to the leakage accident rate of the buried pipe calculated by the buried pipe leakage accident rate prediction device 3 according to this embodiment can be enhanced.

[0159] The program according to this embodiment (the buried pipe leakage accident rate prediction program 48) causes the processor 302 to execute each step of the method for predicting the buried pipe leakage accident rate according to this embodiment.

[0160] According to the program according to this embodiment (the buried pipe leakage accident rate prediction program 48), the leakage accident rate of the buried pipe can be predicted more accurately for any buried period.

[0161] (Embodiment 2) With reference to FIGS. 1 to 7, FIGS. 33 to 42, and FIGS. 45 to 51, the buried pipe leakage accident rate prediction system 1 and the buried pipe leakage accident rate prediction method according to Embodiment 2 will be described.

[0162] The buried pipe leakage accident rate prediction system 1 of this embodiment is mainly different from the buried pipe leakage accident rate prediction system 1 of Embodiment 1 in the following two points. First, this embodiment is different from Embodiment 1 in the calculation method (see FIG. 45) of the conversion coefficient 23 in the conversion coefficient calculation unit 18 (see FIG. 3) of the buried pipe leakage accident rate prediction model generation device 2. Second, this embodiment is different from Embodiment 1 in the pre-processed buried pipe data 46 (see FIG. 51) created by the buried pipe data pre-processing unit 51 (see FIG. 34) of the buried pipe leakage accident rate prediction device 3. The buried pipe leakage accident rate prediction method of this embodiment is different from the buried pipe leakage accident rate prediction method of Embodiment 1 in the calculation method (see FIG. 52) of the leakage accident rate in the leakage accident rate calculation unit 54 (see FIG. 34).

[0163] <Conversion Coefficient Calculation Unit 18 and Conversion Coefficient Calculation Method> In this embodiment, the conversion coefficient calculation unit 18 (see FIG. 3) calculates a conversion coefficient 23 (see FIG. 7) that converts a first index that can be calculated from the probability of pipe wall thickness exceeding of the pipe into a second index that can calculate the leakage accident rate of the pipe or the leakage accident rate of the pipe.

[0164] The first index is not particularly limited. For example, it is an evaluation index for the number of buried pipe leakage accidents. The evaluation index for the number of buried pipe leakage accidents is given by the product of the probability of pipe wall thickness exceeding of the pipe and the pipeline length of the pipe. The second index is not particularly limited. For example, it is the number of buried pipe leakage accidents per unit time (for example, 1 year). The leakage accident rate of the pipe can be calculated by dividing the number of buried pipe leakage accidents per unit time by the pipeline length of the pipe. The conversion coefficient 23 of this embodiment is, for example, a coefficient that converts the evaluation index for the number of buried pipe leakage accidents into the number of buried pipe leakage accidents per unit time. The second index may be an index proportional to the first index.

[0165] Referring to FIG. 45, the method by which the conversion coefficient calculation unit 18 of this embodiment calculates the conversion coefficient 23 will be described.

[0166] Referring to Fig. 45, steps S61 and S62 of the present embodiment are the same as steps S61 and S62 (refer to Fig. 27) of Embodiment 1.

[0167] Referring to Fig. 45, the conversion coefficient calculation unit 18 calculates a leakage accident count evaluation index for each pipeline ID (step S66). Specifically, the conversion coefficient calculation unit 18 reads out the pipeline ID and the pipeline length from the reference pipeline data 20 (refer to Fig. 4) stored in the reference pipeline data storage unit 11 (refer to Fig. 3). The conversion coefficient calculation unit 18 calculates the leakage accident count evaluation index by multiplying the pipe thickness excess probability calculated in step S62 by the pipeline length for each pipeline ID.

[0168] Referring to Fig. 45, the conversion coefficient calculation unit 18 calculates a leakage accident count evaluation index and the number of leakage accidents per unit time for each buried environment (step S67).

[0169] Specifically, the conversion coefficient calculation unit 18 generates reference pipeline data 66 for conversion coefficient calculation (refer to Fig. 46) for the reference pipelines having the specific material selected in step S61 from the pipeline ID, the buried environment, and the number of leakage accidents per unit time (for example, 5 years) of the reference pipeline data 20 (refer to Fig. 4) stored in the reference pipeline data storage unit 11 (refer to Fig. 3) and the leakage accident count evaluation index calculated in S66. The reference pipeline data 66 for conversion coefficient calculation includes the pipeline ID, the buried environment, the leakage accident count evaluation index, and the number of leakage accidents per unit time (for example, 1 year). In the reference pipeline data 66 for conversion coefficient calculation, the pipeline ID, the buried environment, the leakage accident count evaluation index, and the number of leakage accidents per unit time are associated with each other.

[0170] When the unit time (e.g., 5 years) in the reference pipe data 20 (see FIG. 4) is different from the unit time (e.g., 1 year) in the reference pipe data 66 for conversion coefficient calculation (see FIG. 46), the conversion coefficient calculation unit 18 reads out the number of water leakage accident cases per unit time of the reference pipe data 20 from the reference pipe data storage unit 11 (see FIG. 3), and converts the number of water leakage accident cases per unit time of the reference pipe data 20 into the number of water leakage accident cases per unit time of the reference pipe data 66 for conversion coefficient calculation. When the unit time in the reference pipe data 20 is the same as the unit time in the reference pipe data 66 for conversion coefficient calculation, the conversion coefficient calculation unit 18 reads out the number of water leakage accident cases per unit time of the reference pipe data 20 from the reference pipe data storage unit 11, and sets the number of water leakage accident cases per unit time of the reference pipe data 20 as the number of water leakage accident cases per unit time of the reference pipe data 66 for conversion coefficient calculation. The reference pipe data 66 for conversion coefficient calculation may further include the specific material selected in step S61, the probability of pipe thickness exceeding calculated in step S62, and the pipeline length of the reference pipe data 20 (see FIG. 4).

[0171] The conversion coefficient calculation unit 18 groups the reference pipe data 66 for conversion coefficient calculation (see FIG. 46) according to the embedding environment. The conversion coefficient calculation unit 18 divides the reference pipe data 66 for conversion coefficient calculation into a group of embedding environment A, a group of embedding environment B, a group of embedding environment C, and a group of embedding environment D.

[0172] The conversion coefficient calculation unit 18 calculates the water leakage accident case number evaluation index and the number of water leakage accident cases per unit time for each group. The conversion coefficient calculation unit 18 calculates the sum of the water leakage accident case number evaluation indexes of the pipelines included in each group as the water leakage accident case number evaluation index of each group. The conversion coefficient calculation unit 18 calculates the sum of the number of water leakage accident cases per unit time of the pipelines included in each group as the number of water leakage accident cases per unit time of each group. In this way, the conversion coefficient calculation unit 18 obtains the data on the relationship between the water leakage accident case number evaluation index and the number of water leakage accident cases per unit time (see the points in FIGS. 47 and 49).

[0173] Referring to FIG. 45, the conversion coefficient calculation unit 18 calculates a conversion coefficient 23 for a specific material selected in S61 from the data on the relationship between the number of water leakage accident evaluation indicators and the number of water leakage accidents per unit time (see the points in FIGS. 47 and 49) (step S68).

[0174] The number of water leakage accident evaluation indicators of the pipe is proportional to the pipeline length of the pipe. The number of water leakage accidents of the pipe per unit time is also proportional to the pipeline length of the pipe. Therefore, it is considered that the number of water leakage accident evaluation indicators is an indicator proportional to the number of water leakage accidents per unit time. Therefore, the conversion coefficient calculation unit 18 calculates, for example, a proportionality coefficient obtained by linearly regressing the data on the relationship between the number of water leakage accident evaluation indicators and the number of water leakage accidents per unit time (see the points in FIGS. 47 and 49) as the conversion coefficient 23. The slope p of the regression line in FIG. 47 represents the conversion coefficient 23 for ductile iron pipes. The slope q of the regression line in FIG. 49 represents the conversion coefficient 23 for cast iron pipes.

[0175] As shown in FIGS. 48 and 50, the pipeline length of each group is sufficiently long. The reason is that the number of reference pipe data 20 included in each group is sufficiently large. Therefore, the reliability of the data on the relationship between the number of water leakage accident evaluation indicators and the number of water leakage accidents per unit time (see the points in FIGS. 47 and 49) is high, and the conversion coefficient 23 can be calculated with higher accuracy.

[0176] Referring to FIG. 45, step S69 of the present embodiment is the same as step S69 of Embodiment 1. The conversion coefficient calculation unit 18 outputs the calculated conversion coefficient 23 to the buried pipe water leakage accident rate prediction model storage unit 36 (see FIG. 3). The conversion coefficient 23 is stored in the buried pipe water leakage accident rate prediction model storage unit 36.

[0177] The buried pipe leakage accident rate prediction model generation device 2 transmits a buried pipe leakage accident rate prediction model 6 (see Fig. 3), which includes a pipe thickness excess probability prediction model 21 (see Figs. 5 and 6) and a conversion coefficient 23 (see Fig. 7), to a buried pipe leakage accident rate prediction device 3 (see Fig. 1). The buried pipe leakage accident rate prediction device 3 receives the buried pipe leakage accident rate prediction model 6 (see Fig. 34) from the buried pipe leakage accident rate prediction model generation device 2. The buried pipe leakage accident rate prediction model 6 is stored in the buried pipe leakage accident rate prediction model storage unit 36 (see Fig. 34) of the buried pipe leakage accident rate prediction device 3.

[0178] <Buried pipe data preprocessing unit 51 and preprocessed buried pipe data 46> The preprocessed buried pipe data 46 (see Fig. 51) created by the buried pipe data preprocessing unit 51 of this embodiment includes, in addition to the pipeline ID, buried environment, burial period T, nominal pipe thickness, and material of the buried pipe included in the preprocessed buried pipe data 46 (see Fig. 40) created by the buried pipe data preprocessing unit 51 of Embodiment 1, the pipeline length of the buried pipe. The buried pipe data preprocessing unit 51 reads out the pipeline length (see Fig. 36) stored in the buried pipe data storage unit 31 (see Fig. 34) and includes the pipeline length in the preprocessed buried pipe data 46.

[0179] <Buried pipe leakage accident rate prediction method> Referring to Figs. 43 and 52, the buried pipe leakage accident rate prediction method of this embodiment will be described. The buried pipe leakage accident rate prediction method of this embodiment is the same as the buried pipe leakage accident rate prediction method of Embodiment 1, but is different from the buried pipe leakage accident rate prediction method of Embodiment 1 in the following points.

[0180] Referring to FIG. 43, in step S81 of the present embodiment, the buried pipe data preprocessing unit 51 creates preprocessed buried pipe data 46 (refer to FIG. 51) from the buried pipe data 40 (refer to FIGS. 34 to 36) stored in the buried pipe data storage unit 31 (refer to FIG. 34). Step S81 of the present embodiment is the same as step S81 of the first embodiment. However, in the present embodiment, the buried pipe data preprocessing unit 51 further obtains the pipeline length of the buried pipe (refer to FIG. 36) from the buried pipe data 40 stored in the buried pipe data storage unit 31 (refer to FIG. 34).

[0181] Referring to FIG. 43, step S82 of the present embodiment is the same as step S82 of the first embodiment. However, step S82 of the present embodiment is different from step S82 of the first embodiment in that the leakage accident rate calculation unit 54 reads out the pipeline ID, the buried environment, the buried period T, and the nominal pipe thickness of the buried pipe from the preprocessed buried pipe data 46 (refer to FIG. 51) stored in the preprocessed buried pipe data storage unit 34.

[0182] Referring to FIG. 43, the leakage accident rate calculation unit 54 calculates the leakage accident rate of the buried pipe for each pipeline ID (step S83).

[0183] Specifically, referring to FIG. 52, the leakage accident rate calculation unit 54 calculates the evaluation index of the number of leakage accident cases of the buried pipe for each pipeline ID (step S83c). The leakage accident rate calculation unit 54 reads out the pipeline ID, the material, and the pipeline length of the buried pipe from the preprocessed buried pipe data 46 (refer to FIG. 51) stored in the preprocessed buried pipe data storage unit 34 (refer to FIG. 3). The leakage accident rate calculation unit 54 multiplies the pipe thickness excess probability of the buried pipe calculated in step S82 by the pipeline length of the buried pipe for each pipeline ID to calculate the evaluation index of the number of leakage accident cases of the buried pipe.

[0184] The conversion coefficient selection unit 55 selects, for each pipeline ID, the conversion coefficient 23 corresponding to the material of the buried pipe from the conversion coefficients 23 stored in the buried pipe leakage accident rate prediction model storage unit 36 (see FIG. 34) (step S83d). The leakage accident rate calculation unit 54 multiplies, for each pipeline ID, the buried pipe leakage accident count evaluation index calculated in step S83d by the conversion coefficient 23 selected in step S83d to calculate the number of buried pipe leakage accidents per unit time (for example, 1 year) (step S83e). The leakage accident rate calculation unit 54 divides, for each pipeline ID, the number of buried pipe leakage accidents per unit time by the pipeline length of the buried pipe to calculate the leakage accident rate of the buried pipe (step S83f).

[0185] The leakage accident rate calculation unit 54 outputs the leakage accident rate of the buried pipe calculated in step S83f to the leakage accident rate storage unit 37 (see FIG. 34). As shown in FIG. 41, the leakage accident rate of the buried pipe is stored in the leakage accident rate storage unit 37 in association with the pipeline ID.

[0186] The program of the present embodiment causes the processor 302 (see FIG. 33) to execute the buried pipe leakage accident rate prediction method of the present embodiment. The computer-readable recording medium (non-transitory computer-readable recording medium, for example, the storage medium 308) of the present embodiment stores a program for causing the processor 302 to execute the buried pipe leakage accident rate prediction method of the present embodiment.

[0187] The effects of the buried pipe leakage accident rate prediction device 3, the buried pipe leakage accident rate prediction method, and the program of the present embodiment will be described. The buried pipe leakage accident rate prediction device 3, the buried pipe leakage accident rate prediction method, and the program of the present embodiment have the following effects similar to those of the buried pipe leakage accident rate prediction device 3, the buried pipe leakage accident rate prediction method, and the program of Embodiment 1.

[0188] The buried pipe leakage accident rate prediction device 3 of this embodiment includes a pipe thickness exceeding probability calculation unit 52 and a leakage accident rate calculation unit 54. The pipe thickness exceeding probability calculation unit 52 calculates the pipe thickness exceeding probability of the buried pipe by inputting the buried environment, buried period, and pipe thickness (for example, nominal pipe thickness) of the buried pipe into the pipe thickness exceeding probability prediction model 21. The leakage accident rate calculation unit 54 calculates the leakage accident rate of the buried pipe using the pipe thickness exceeding probability of the buried pipe and the conversion coefficient 23. The pipe thickness exceeding probability prediction model 21 is generated according to the buried environment of the pipe and the pipe thickness (for example, nominal pipe thickness) of the pipe, and gives the pipe thickness exceeding probability of the pipe that continuously changes with respect to the continuous change of the buried period of the pipe. The pipe thickness exceeding probability of the pipe is the probability that the corrosion depth of the pipe exceeds the pipe thickness (for example, nominal pipe thickness) of the pipe. The pipe thickness exceeding probability of the buried pipe is the probability that the corrosion depth of the buried pipe exceeds the pipe thickness (for example, nominal pipe thickness) of the buried pipe. The conversion coefficient 23 is a coefficient that converts a first index that can be calculated from the pipe thickness exceeding probability of the pipe into a second index that can calculate the leakage accident rate of the pipe. The leakage accident rate of the pipe is the number of pipe leakage accidents per unit time and per unit distance. The leakage accident rate of the buried pipe is the number of buried pipe leakage accidents per unit time and per unit distance.

[0189] According to the buried pipe leakage accident rate prediction device 3 of this embodiment, the leakage accident rate of the buried pipe can be predicted more accurately for any buried period.

[0190] In the buried pipe leakage accident rate prediction device 3 of this embodiment, the leakage accident rate calculation unit 54 includes a conversion coefficient selection unit 55 that selects the conversion coefficient 23 corresponding to the material of the buried pipe.

[0191] According to the buried pipe leakage accident rate prediction device 3 of this embodiment, the leakage accident rate of the buried pipe can be predicted more accurately for any buried period according to the material of the buried pipe.

[0192] In the buried pipe leakage accident rate prediction device 3 of the present embodiment, the conversion coefficient is a coefficient for converting the first index to the second index. The first index is an evaluation index of the number of pipe leakage accidents, and the second index is the number of pipe leakage accidents per unit time. The leakage accident rate calculation unit 54 calculates an evaluation index of the number of buried pipe leakage accidents by multiplying the probability of the buried pipe wall thickness exceeding the standard and the pipeline length of the buried pipe, calculates the number of buried pipe leakage accidents per unit time by multiplying the evaluation index of the number of buried pipe leakage accidents and the conversion coefficient 23, and calculates the leakage accident rate of the buried pipe by dividing the number of buried pipe leakage accidents per unit time by the pipeline length of the buried pipe.

[0193] Since the number of pipe leakage accidents per unit time is used as the second index, the conversion coefficient 23 can be calculated more simply from the reference pipe data 20. Generation of the buried pipe leakage accident rate prediction model 6 (see FIGS. 3, 5 to 7, and 34) including the conversion coefficient 23 becomes easier.

[0194] The buried pipe leakage accident rate prediction method of the present embodiment includes a step of calculating the probability of the buried pipe wall thickness exceeding the standard (step S82) by inputting the buried environment, the buried period, and the wall thickness (for example, the nominal wall thickness) of the buried pipe into the wall thickness exceeding probability prediction model 21 of the pipe, and a step of calculating the leakage accident rate of the buried pipe (step S83) using the probability of the buried pipe wall thickness exceeding the standard and the conversion coefficient 23. The wall thickness exceeding probability prediction model 21 of the pipe is generated according to the buried environment of the pipe and the wall thickness (for example, the nominal wall thickness) of the pipe, and gives the probability of the wall thickness exceeding the standard of the pipe that continuously changes in response to the continuous change of the buried period of the pipe. The probability of the wall thickness exceeding the standard of the pipe is the probability that the corrosion depth of the pipe exceeds the wall thickness (for example, the nominal wall thickness) of the pipe. The probability of the buried pipe wall thickness exceeding the standard is the probability that the corrosion depth of the buried pipe exceeds the wall thickness (for example, the nominal wall thickness) of the buried pipe. The conversion coefficient 23 is a coefficient for converting a first index that can be calculated from the probability of the wall thickness exceeding the standard of the pipe to a second index that can be used to calculate the leakage accident rate of the pipe. The leakage accident rate of the pipe is the number of pipe leakage accidents per unit time and per unit distance. The leakage accident rate of the buried pipe is the number of buried pipe leakage accidents per unit time and per unit distance.

[0195] According to the buried pipe leakage accident rate prediction method of the present embodiment, the leakage accident rate of the buried pipe can be predicted more accurately for any buried period.

[0196] In the buried pipe leakage accident rate prediction method of the present embodiment, the step of calculating the leakage accident rate of the buried pipe (step S83) includes a step of selecting a conversion coefficient 23 corresponding to the material of the buried pipe (step S83d).

[0197] According to the buried pipe leakage accident rate prediction method of the present embodiment, the leakage accident rate of the buried pipe can be predicted more accurately for any buried period according to the material of the buried pipe.

[0198] In the buried pipe leakage accident rate prediction method of the present embodiment, the conversion coefficient is a coefficient for converting the first index to the second index. The first index is an evaluation index of the number of leakage accidents of the pipe, and the second index is the number of leakage accidents of the pipe per unit time. The step of calculating the leakage accident rate of the buried pipe (step S83) includes a step of calculating an evaluation index of the number of leakage accidents of the buried pipe by multiplying the probability of the pipe thickness exceeding the buried pipe and the pipeline length of the buried pipe (step S83c), and an evaluation index of the number of leakage accidents of the buried pipe and the conversion coefficient 23. A step of calculating the number of leakage accidents of the buried pipe per unit time by multiplication (step S83e), and a step of calculating the leakage accident rate of the buried pipe by dividing the number of leakage accidents of the buried pipe per unit time by the pipeline length of the buried pipe (step S83f).

[0199] Since the number of leakage accidents of the pipe per unit time is used as the second index, the conversion coefficient 23 can be calculated more simply from the reference pipe data 20. The generation of the buried pipe leakage accident rate prediction model 6 (see FIGS. 3, 5 to 7, and 34) including the conversion coefficient 23 becomes easy.

[0200] The program of the present embodiment (the buried pipe leakage accident rate prediction program 48 (see FIG. 34)) causes the processor 302 to execute each step of the buried pipe leakage accident rate prediction method of the present embodiment.

[0201] According to the program of this embodiment (the buried pipe leakage accident rate prediction program 48), the leakage accident rate of the buried pipe can be predicted more accurately for any buried period.

[0202] (Modification example) A function of generating the buried pipe leakage accident rate prediction model 6 from the reference pipe data 20 may be added to the buried pipe leakage accident rate prediction device 3, and the buried pipe leakage accident rate prediction model generation device 2 may be omitted from the buried pipe leakage accident rate prediction system 1. The conversion coefficient 23 may be a coefficient that converts the probability of pipe wall thickness exceeding the standard into a second index from which the leakage accident rate of the pipe can be calculated. The conversion coefficient 23 may be a coefficient that converts a first index calculable from the probability of pipe wall thickness exceeding the standard into the leakage accident rate of the pipe.

[0203] The pipe wall thickness exceeding probability prediction model 21 may be generated by another method. For example, the pipe wall thickness exceeding probability prediction model 21 may be generated without considering the distribution of the basic regression line.

[0204] It should be considered that the disclosed Embodiment 1 and Embodiment 2 and their modification examples are illustrative in all respects and not restrictive. The scope of the present disclosure is shown by the claims rather than the above description, and is intended to include all modifications within the meaning and scope equivalent to the claims.

Description of reference numerals

[0205] 1 Buried pipe leakage accident rate prediction system, 2 Buried pipe leakage accident rate prediction model generation device, 3 Buried pipe leakage accident rate prediction device, 4 Communication network, 6 Buried pipe leakage accident rate prediction model, 10 Memory unit, 11 Reference pipe data memory unit, 12 Buried pipe leakage accident rate prediction model memory unit, 13 Corrosion lag time memory unit, 14 Program memory unit, 16 Buried pipe leakage accident rate prediction model generation unit, 17 Pipe thickness exceeding probability prediction model generation unit, 18 Conversion coefficient calculation unit, 20 Reference pipe data, 21 Pipe thickness exceeding probability prediction model, 23 Conversion coefficient, 26 Buried pipe leakage accident rate prediction model generation program, 27 Corrosion lag time data, 28,66 Conversion coefficient calculation reference pipe data, 30 Memory unit, 31 Buried pipe data memory unit, 32 Nominal pipe thickness database unit, 33 Buried environment map memory unit, 34 Pretreated buried pipe data memory unit, 36 Buried pipe leakage accident rate prediction model memory unit, 37 Leakage accident rate memory unit, 38 Program memory unit, 40 Buried pipe data, 41 Pipeline map, 42 Buried pipe attribute data, 43 Nominal pipe thickness data, 44 Buried environment map, 45 Ground-buried environment correspondence data, 46 Pretreated buried pipe data, 48 Buried pipe leakage accident rate prediction program, 50 Buried pipe data reception unit, 51 Pretreatment unit, 52 Pipe thickness exceeding probability calculation unit, 54 Leakage accident rate calculation unit, 55 Conversion coefficient selection unit, 57 Leakage accident rate prediction result output unit, 60 Leakage accident rate prediction result, 61 Leakage accident rate prediction table, 62 Leakage accident rate prediction map, 201,301 Input device, 202,302 Processor, 203,303 Memory, 204,304 Display, 206,306 Network controller, 207,307 Memory media drive, 208,308 Memory media, 210,310 Storage.

Claims

1. A pipe wall thickness exceeding probability calculation unit that calculates the probability of the embedded pipe's wall thickness exceeding by inputting the embedding environment, embedding period, and wall thickness of the embedded pipe into a wall thickness exceeding probability prediction model; A leakage accident rate calculation unit that calculates the leakage accident rate of the embedded pipe using the probability of the embedded pipe's wall thickness exceeding and a conversion coefficient; The wall thickness exceeding probability prediction model is generated according to the pipe's embedding environment and the pipe's wall thickness, and gives the probability of the pipe's wall thickness exceeding that continuously changes with respect to the continuous change of the pipe's embedding period; The conversion coefficient is a coefficient that converts the probability of the pipe's wall thickness exceeding into the leakage accident rate of the pipe; The conversion coefficient is calculated from the relationship between the probability of the wall thickness exceeding and the leakage accident rate among a plurality of groups of reference pipes obtained by grouping the reference pipe data for calculating the conversion coefficient for each range of the probability of the wall thickness exceeding; The reference pipe data for calculating the conversion coefficient includes the probability of the wall thickness exceeding, the pipeline length, and the number of leakage accident cases per unit time of the plurality of reference pipes; The probability of the wall thickness exceeding of the plurality of reference pipes is calculated by inputting the embedding environment, embedding period, and wall thickness of the plurality of reference pipes into the wall thickness exceeding probability prediction model; The leakage accident rate of each of the plurality of groups is calculated by dividing the sum of the number of leakage accident cases per unit time of the reference pipes included in each of the plurality of groups by the sum of the pipeline lengths of the reference pipes included in each of the plurality of groups. An embedded pipe leakage accident rate prediction device.

2. The embedded pipe leakage accident rate prediction device according to Claim 1, wherein the leakage accident rate calculation unit includes a conversion coefficient selection unit that selects the conversion coefficient corresponding to the material of the embedded pipe.

3. The conversion coefficient is a coefficient that converts the probability of the pipe's wall thickness exceeding into the leakage accident rate of the pipe; The embedded pipe leakage accident rate prediction device according to Claim 1 or Claim 2, wherein the leakage accident rate calculation unit calculates the leakage accident rate of the embedded pipe by multiplying the probability of the embedded pipe's wall thickness exceeding and the conversion coefficient.

4. A step of calculating the probability of the embedded pipe's wall thickness exceeding by inputting the embedding environment, embedding period, and wall thickness of the embedded pipe into a wall thickness exceeding probability prediction model; A step of calculating the leakage accident rate of the embedded pipe using the probability of the embedded pipe's wall thickness exceeding and a conversion coefficient. The pipe wall thickness exceeding probability prediction model is generated according to the pipe embedding environment and the wall thickness of the pipe, and gives the probability of the pipe wall thickness exceeding continuously changing with respect to the continuous change of the pipe embedding period. The conversion coefficient is a coefficient for converting the probability of the pipe wall thickness exceeding of the pipe into the leakage accident rate of the pipe. The conversion coefficient is calculated from the relationship between the probability of the pipe wall thickness exceeding and the leakage accident rate among a plurality of groups of a plurality of reference pipes obtained by grouping the reference pipe data for calculating the conversion coefficient for each range of a predetermined probability of the pipe wall thickness exceeding. The reference pipe data for calculating the conversion coefficient includes the probability of the pipe wall thickness exceeding, the pipeline length, and the number of leakage accident cases per unit time of the plurality of reference pipes. The probability of the pipe wall thickness exceeding of the plurality of reference pipes is calculated by inputting the embedding environment, the embedding period, and the wall thickness of the plurality of reference pipes into the pipe wall thickness exceeding probability prediction model. The leakage accident rate of each of the plurality of groups is calculated by dividing the sum of the number of leakage accident cases per unit time of the reference pipes included in each of the plurality of groups by the sum of the pipeline lengths of the reference pipes included in each of the plurality of groups. This is a method for predicting the leakage accident rate of buried pipes.

5. The step of calculating the leakage accident rate of the buried pipe includes the step of selecting the conversion coefficient corresponding to the material of the buried pipe. The method for predicting the leakage accident rate of a buried pipe according to claim 4.

6. The conversion coefficient is a coefficient for converting the probability of the pipe wall thickness exceeding of the pipe into the leakage accident rate of the pipe. The step of calculating the leakage accident rate of the buried pipe is a step of calculating the leakage accident rate of the buried pipe by multiplying the probability of the pipe wall thickness exceeding of the buried pipe and the conversion coefficient. The method for predicting the leakage accident rate of a buried pipe according to claim 4 or claim 5.

7. A program for causing a processor to execute each step of the method for predicting the leakage accident rate of a buried pipe according to any one of claims 4 to 6.

Images