JP2019079320A - End pressure control support device, end pressure control support method and computer program - Google Patents

Abstract

To provide an end pressure control support device, an end pressure control support method and a computer program capable of achieving end pressure control of a water distribution system at low cost.SOLUTION: An end pressure control support device of an embodiment comprises an end pressure learning unit and an end pressure estimation unit. The end pressure learning unit learns relationship between end pressure of a water distribution network and quantity related to water distribution through the water distribution network using machine learning method. The end pressure estimation unit estimates end pressure of the water distribution network based on the relationship learned by the end pressure learning unit and information indicating the quantity.SELECTED DRAWING: Figure 2

Description

  Embodiments of the present invention relate to an end pressure control support device, an end pressure control support method, and a computer program.

  As a control method of a water distribution system for supplying water to a consumer, end pressure control effective for reduction of extra power consumption of a water distribution pump, reduction of water leakage amount, etc. is known. However, in order to perform end pressure control, it is necessary to install an end pressure measuring device in the water distribution pipe network, and a simulation model (hereinafter referred to as “pipe network model”) for performing pipe network analysis is constructed. Need and cost a lot. Therefore, a technology that enables the end pressure control of the water distribution system to be realized at lower cost is desired.

JP 2001-188614 A Patent No. 3851783 gazette Patent No. 5010504

  The problem to be solved by the present invention is to provide an end pressure control support device, an end pressure control support method, and a computer program that can realize end pressure control of a water distribution system at lower cost.

  The distal pressure control support device of the embodiment has a distal pressure learning unit and a distal pressure estimation unit. The end pressure learning unit learns the relationship between the end pressure of the water distribution network and the quantities related to water distribution through the water distribution network using a machine learning method. The end pressure estimation unit estimates the end pressure of the water distribution network based on the relationship learned by the end pressure learning unit and the information indicating the quantities.

The figure which shows the specific example of the system configuration | structure of the water distribution system 100 in 1st Embodiment. FIG. 2 is a block diagram showing a specific example of a functional configuration of the distal end pressure control support device 1 of the first embodiment. 6 is a flowchart showing an operation example of the distal end pressure control support device 1 of the first embodiment. The figure which shows the specific example of the system configuration | structure of the water distribution system 100a in 2nd Embodiment. The block diagram which shows the specific example of a function structure of the terminal pressure control assistance apparatus 1a of 2nd Embodiment. The flowchart which shows the operation example of the terminal pressure control assistance apparatus 1a of 2nd Embodiment. The block diagram which shows the specific example of a function structure of the terminal pressure control assistance apparatus 1b of 3rd Embodiment. The flowchart which shows the operation example of the terminal pressure control assistance apparatus 1b of 3rd Embodiment. The figure which shows the specific example of the system configuration | structure of the water distribution system 100c in 4th Embodiment. The block diagram which shows the specific example of a function structure of the terminal pressure control assistance apparatus 1c of 4th Embodiment. The flowchart which shows the operation example of the terminal pressure control assistance apparatus 1c of 4th Embodiment. The block diagram which shows the specific example of a function structure of the terminal pressure control assistance apparatus 1d of 5th Embodiment. The flowchart which shows the operation example of the terminal pressure control assistance apparatus 1d of 5th Embodiment.

  Hereinafter, an end pressure control support device, an end pressure control support method, and a computer program according to an embodiment will be described with reference to the drawings.

First Embodiment
FIG. 1 is a diagram showing a specific example of the system configuration of the water distribution system 100 in the first embodiment. The water distribution system 100 includes an end pressure control support device 1, a monitoring control system 2, one or more water distribution stations 3, one or more end pressure measurement devices 4, and one or more consumers 6 via water distribution network 5. Supply system. FIG. 1 shows water distribution areas 3-1 and 3-2 as an example of one or more water distribution sites, and shows end pressure measuring apparatuses 4-1 and 4-2 as an example of one or more end pressure measuring apparatuses 4, 1 As an example of the customer 6 above, customers 6-1 to 6-3 are shown. The number of water distribution stations 3, terminal pressure measuring devices 4 and customers 6 may be different from the example of FIG. Each of the end pressure control support device 1, the monitoring control system 2, each water distribution station 3, and each end pressure measurement device 4 is configured to be communicable with each other.

  The end pressure control support device 1 provides the monitoring control system 2 that monitors or controls the water distribution site 3 with information (hereinafter referred to as “support information”) that supports the control or the monitoring. Specifically, the end pressure control support device 1 acquires end pressure information indicating the measured value of the end pressure of the water distribution network 5 from each end pressure measurement device 4 and distributes water through the water distribution pipe network 5 The information indicating the various quantities is acquired, and the support information is generated based on the acquired terminal pressure information and the information indicating the above-mentioned various quantities. More specifically, the terminal pressure control support device 1 acquires distribution area information indicating various quantities (hereinafter referred to as "distribution quantities") from each distribution site 3, and the acquired terminal pressure information Support information is generated based on the and water distribution site information. For example, the support information includes an estimated value of the end pressure at each end of the water distribution network 5. Further, the support information may include information generated based on the estimated value of the end pressure, a determination result based on the estimated value of the end pressure, and the like.

  The monitoring control system 2 is a system that monitors or controls the water distribution station 3 using support information provided from the end pressure control support device 1. For example, the monitoring control system 2 adjusts the number of operating water distribution pumps in each water distribution station 3 and the output intensity, and monitors the water level of the water distribution reservoir which is a water supply source.

  The distribution station 3 has a distribution reservoir for storing water to be supplied to the customer 6 and various facilities (not shown) such as a distribution pump for delivering the water of the distribution reservoir to the distribution network 5, and distributes the water of the distribution reservoir. It is a facility that supplies each customer 6 through the water network 5. The various facilities which the water distribution station 3 has are monitored or controlled by the monitoring control system 2. In addition, the water distribution station 3 transmits, to the terminal pressure control support device 1, water distribution site information indicating various amounts (distribution lot amounts) which can be observed in the own facility.

  Generally, it is considered that the terminal pressure of the water distribution network 5 has a high correlation with the discharge flow rate or discharge pressure of the pump at the water distribution site 3. Therefore, the water distribution site information includes at least information indicating the discharge flow rate or discharge pressure of the pump. Moreover, since it can be considered that the end pressure of the distribution pipe network 5 is also correlated with the water level of the distribution reservoir in the distribution site 3, the distribution site information may include information indicating the water level of the distribution reservoir.

  The end pressure measurement device 4 is a device that measures the end pressure of the water distribution network 5. The end pressure measurement device 4 transmits end pressure information indicating the measured end pressure to the end pressure control support device 1.

  FIG. 2 is a block diagram showing a specific example of a functional configuration of the distal end pressure control support device 1 of the first embodiment. The end pressure control support device 1 includes a CPU (Central Processing Unit), a memory, an auxiliary storage device, and the like connected by a bus, and executes a program. The end pressure control support device 1 includes a water distribution place information storage unit 101, an end pressure information storage unit 102, an information acquisition unit 103, an end pressure learning unit 104, an end pressure estimation unit 105, and an output unit 106 by executing a program. Function. Note that all or part of the functions of the end pressure control support device 1 are realized using hardware such as an application specific integrated circuit (ASIC), a programmable logic device (PLD), or a field programmable gate array (FPGA). It is also good. The program may be recorded on a computer readable recording medium. The computer readable recording medium is, for example, a portable medium such as a flexible disk, a magneto-optical disk, a ROM, a CD-ROM, or a storage device such as a hard disk built in a computer system. The program may be transmitted via a telecommunication link.

  The water distribution site information storage unit 101 and the end pressure information storage unit 102 are configured using storage devices such as a magnetic hard disk drive and a semiconductor storage device. The water distribution site information storage unit 101 stores water distribution site information, and the end pressure information storage unit 102 stores end pressure information.

  The information acquisition unit 103 includes a communication interface, and communicates with each water distribution station 3 and each terminal pressure measurement device 4 via the communication interface. The information acquisition unit 103 acquires distribution facility information from each distribution facility 3, and stores the acquired distribution facility information in the distribution facility information storage unit 101. Further, the information acquisition unit 103 acquires end pressure information from each end pressure measurement device 4 and stores the acquired end pressure information in the end pressure information storage unit 102. The information acquisition unit 103 repeatedly stores the distribution site information and the end pressure information in the distribution site information storage unit 101 and the end pressure information storage unit 102 by repeatedly executing the acquisition of the distribution site information and the end pressure information at a predetermined timing. .

  The end pressure learning unit 104 performs machine learning processing using the water distribution center information stored in the water distribution center information storage unit 101 and the end pressure information stored in the end pressure information storage unit 102 as teacher data. The terminal pressure learning unit 104 applies a machine learning technique for generating or updating a classifier by supervised learning. The end pressure learning unit 104 learns the relationship between the distribution field quantity and the end pressure of the water distribution network 5 by performing a machine learning process based on the teacher data. More specifically, the end pressure learning unit 104 generates or updates a discriminator indicating the relationship between the distribution field quantity and the end pressure of the water distribution network 5. The end pressure learning unit 104 outputs data of parameters or the like representing the generated or updated classifier to the end pressure estimation unit 105 as a learning result. Hereinafter, the learning of the relationship is referred to as “learning of end pressure” as necessary.

  The end pressure estimation unit 105 estimates the end pressure corresponding to the given distribution field amount based on the relationship between the distribution field amounts learned by the end pressure learning unit 104 and the end pressure of the distribution pipe network 5 Do. The end pressure estimation unit 105 outputs estimated end pressure information indicating the end pressure obtained by estimation to the output unit 106. The end pressure estimation unit 105 can estimate the end pressure for the end at which the end pressure learning unit 104 learns the end pressure.

  The output unit 106 includes a communication interface, and communicates with the monitoring control system 2 via the communication interface. The output unit 106 generates support information based on the estimated end pressure information output from the end pressure estimation unit 105, and transmits the generated support information to the monitoring control system 2.

  FIG. 3 is a flowchart showing an operation example of the distal end pressure control support device 1 of the first embodiment. First, the information acquisition unit 103 acquires water distribution place information from each water distribution place 3, and acquires end pressure information from each end pressure measurement device 4 (step S101). The information acquiring unit 103 stores the acquired water distribution site information in the water distribution site information storage unit 101, and stores the acquired terminal pressure information in the terminal pressure information storage unit 102. By repeatedly executing acquisition of such information, the end pressure control supporting device 1 accumulates water distribution place information and end pressure information (step S102).

  Subsequently, based on the water distribution site information stored in the water distribution site information storage unit 101 and the terminal pressure information stored in the terminal pressure information storage unit 102, the terminal pressure learning unit 104 distributes various distribution channel quantities and water distribution pipes. The relationship with the end pressure of the net 5 is learned. That is, the end pressure learning unit 104 learns the end pressure of a certain end based on the past end pressure information of the end and the past water distribution place information (step S103). The end pressure learning unit 104 outputs the learning result to the end pressure estimation unit 105.

  Subsequently, the information acquisition unit 103 acquires current distribution site information from each distribution site 3 (step S104). The terminal pressure estimation unit 105 estimates the current terminal pressure of the water distribution network 5 based on the learning result by the terminal pressure learning unit 104 and the acquired current distribution site information (step S105). The end pressure learning unit 104 outputs the estimated end pressure information obtained by the estimation to the output unit 106. The output unit 106 generates support information based on the estimated terminal pressure information, and transmits the generated support information to the monitoring control system 2 (step S106).

  The terminal pressure control support device 1 of the first embodiment configured as described above is applied to, for example, a water distribution system in which the terminal pressure measurement device is not permanently installed. Specifically, in the water distribution system 100 shown in FIG. 1, the terminal pressure measurement device 4 is temporarily present. In this case, the end pressure control support device 1 learns the relationship between the distribution site quantities and the end pressure based on the end pressure information and the distribution site information acquired in the period in which the end pressure measurement device 4 exists, After the end pressure measuring device 4 is removed, the end pressure is estimated based on the learning result and the water distribution site information. That is, the end pressure control support device 1 learns the relationship between the distribution site quantities and the end pressure based on the past end pressure information and the past water distribution site information, and the learning result and the current distribution site information Estimate the current end pressure based on it.

  According to such an end pressure control support device 1, the water distribution system 100 can perform end pressure control without requiring the end pressure measurement device 4 to be permanently installed. For example, the terminal pressure measuring device 4 is temporarily installed only at the design stage or adjustment period of the water distribution system 100 to learn in advance the relationship between the distribution ground quantity and the terminal pressure, and in actual operation, the terminal pressure It becomes possible to estimate the end pressure without arranging the measuring device 4. Therefore, the end pressure control of the water distribution system 100 can be realized at lower cost by applying the end pressure control support device 1 of the first embodiment.

  Specifically, it is possible to estimate the end pressure of the water distribution network 5 without the need to construct a network network model. Moreover, it becomes possible to estimate the end pressure of the distribution pipe network 5 without always providing the end pressure measurement device 4 in the water distribution system 100. In addition, the end pressure control support device 1 can learn and estimate the end pressure for any end of the water distribution network 5 as long as it is possible to obtain past end pressure information. Therefore, it becomes possible to control the end pressure of distribution network 5 more precisely. This makes it possible to reduce the excess end pressure in the water distribution network 5. In addition, it is possible to reduce the amount of water leakage in the water distribution pipe network 5 by reducing the excess end pressure.

  Note that any machine learning method may be used to learn the end pressure in the present embodiment. For example, methods such as SVM (Support Vector Machine) and DL (Deep Learning) may be used. However, experiments have shown that machine learning using a random forest is effective (that is, the estimation accuracy is high) for the estimation of the terminal pressure of the water distribution network. Therefore, the terminal pressure control support devices of the present embodiment and the following embodiments assume a case where terminal pressure is estimated by machine learning based on random forest.

Second Embodiment
FIG. 4 is a view showing a specific example of the system configuration of the water distribution system 100a in the second embodiment. The water distribution system 100a is different from the water distribution system 100 in the first embodiment in that the end pressure control support device 1a is replaced with the end pressure control support device 1a. The other configuration of the water distribution system 100a is the same as the water distribution system 100 in the first embodiment. Therefore, about the other composition, explanation is omitted by attaching the same numerals as FIG.

  FIG. 5 is a block diagram showing a specific example of a functional configuration of the distal end pressure control support device 1a of the second embodiment. The end pressure control support device 1a further includes a permanent end pressure measurement device 107 (an example of a measurement unit), a point including an end pressure learning unit 104a instead of the end pressure learning unit 104, and an end pressure estimation unit 105 It differs from the end pressure control support device 1 of the first embodiment in that the end pressure estimation unit 105a is provided. The other configuration of the end pressure control support device 1a is the same as that of the end pressure control support device 1 of the first embodiment. Therefore, about the other composition, explanation is omitted by attaching the same numerals as FIG.

  The permanent end pressure measuring device 107 is an end pressure measuring device that is permanently installed at a specific end (hereinafter referred to as a “permanent point”) of the water distribution pipe network 5. In the first embodiment, the end pressure measuring device 4 is installed to obtain end pressure information used to learn the relationship between water distribution field quantities and end pressure, whereas the permanent end pressure measuring device 107 Are installed to obtain end pressure information used for both learning and estimation of the end pressure described above. The permanent end pressure measuring device 107 measures the end pressure at the permanent setting point, and outputs, to the information acquisition unit 103, terminal pressure information (hereinafter referred to as “the permanent end pressure information”) indicating the measured terminal pressure at the permanent setting point. The permanent end pressure information is accumulated by the information acquisition unit 103 in the end pressure information storage unit 102 together with the end pressure information acquired by the other end pressure measurement device 4.

  The end pressure learning unit 104 a distributes water, as in the first embodiment, based on the past end pressure information acquired by the end pressure measurement device 4 and the permanent end pressure measurement device 107 and the past water distribution place information. The relationship between field quantity and terminal pressure of water distribution network 5 is learned. More specifically, the end pressure learning unit 104a learns the relationship between the distribution field quantity and the end pressure at the permanent setting point, and the end pressure at any end other than the permanent setting point at which past terminal pressure information can be obtained. Do. In this respect, the end pressure learning unit 104a is different from the end pressure learning unit 104 in the first embodiment. The end pressure information of the end is used to learn the end pressure of the end other than the permanent set point to be estimated, as in the first embodiment.

  The end pressure estimation unit 105a estimates the end pressure corresponding to the given distribution field amount based on the relationship between the distribution field amounts learned by the end pressure learning unit 104a and the end pressure of the distribution pipe network 5 Do. Here, while the terminal pressure estimating unit 105 in the first embodiment estimates the current terminal pressure using the present water distribution site information as an input, the terminal pressure estimating unit 105 a estimates the current water distribution site information and the current permanent installation. The present embodiment differs from the end pressure estimation unit 105 in the first embodiment in that the end pressure information is used as an input to estimate the current end pressure at any end other than the constant setting point.

  FIG. 6 is a flowchart showing an operation example of the distal end pressure control support device 1a according to the second embodiment. First, as in the first embodiment, the information acquiring unit 103 acquires water distribution site information and terminal pressure information (step S201), and each acquired information is distributed to the water distribution site information storage unit 101 and the terminal pressure information storage unit 102. (Step S202). The terminal pressure information acquired here includes the terminal pressure information acquired by the terminal pressure measuring device 4 and the terminal pressure information (permanent terminal pressure information) acquired by the permanent terminal pressure measuring device 107.

  Subsequently, the terminal pressure learning unit 104a learns the terminal pressure based on the past water distribution site information and the past terminal pressure information accumulated in the water distribution site information storage unit 101 and the terminal pressure information storage unit 102 (step S203). ). Specifically, the end pressure learning unit 104 a learns the relationship between the distribution field quantity and the end pressure at the permanent installation point, and the end pressure at any end of the water distribution network 5. That is, according to this learning, a discriminator that outputs the end pressure at any end of the water distribution network 5 with the water distribution site information and the permanent end pressure information as inputs is generated. The end pressure learning unit 104a outputs the learning result to the end pressure estimation unit 105a.

  Subsequently, the information acquisition unit 103 acquires current permanent terminal pressure information and current water distribution site information (step S204). The end pressure estimation unit 105a estimates the current end pressure at any end of the water distribution network 5 based on the learning result by the end pressure learning unit 104a, the current distribution site information, and the current permanent end pressure information. (Step S205). The end pressure learning unit 104 a outputs the estimated end pressure information obtained by the estimation to the output unit 106. The output unit 106 generates support information based on the estimated terminal pressure information, and transmits the generated support information to the monitoring control system 2 (step S206).

  The terminal pressure control supporting device 1a according to the second embodiment configured as described above uses various terminal water pressure information in the past and the terminal pressure information in the past acquired for the other terminals. The relationship between the pressure and the terminal pressure is learned, and the current terminal pressure at any terminal of the water distribution network 5 is estimated based on the current distribution site information and the current permanent terminal pressure information. In this way, learning of the relationship between the distribution field quantity and the end pressure, and estimation of the present end pressure based on the learning result are performed based on the measured value of the end pressure at a specific end (that is, a permanent set point) Thus, the end pressure control supporting device 1a can estimate the end pressure of the water distribution network 5 more accurately.

Third Embodiment
A water distribution system 100b (not shown) in the third embodiment has the same system configuration as the water distribution system 100 in the first embodiment, and the end pressure control support device 1b is used in place of the end pressure control support device 1. It differs from the water distribution system 100 in the first embodiment in that it comprises.

  FIG. 7 is a block diagram showing a specific example of a functional configuration of the distal end pressure control support device 1b of the third embodiment. The terminal pressure control support device 1b further includes a permanent candidate point determination unit 108, a terminal pressure learning unit 104b in place of the terminal pressure learning unit 104, and a terminal pressure estimation unit 105b in place of the terminal pressure estimation unit 105. It differs from the end pressure control support device 1 of the first embodiment in that it comprises. The other configuration of the end pressure control support device 1a is the same as that of the end pressure control support device 1 of the first embodiment. Therefore, about the other composition, explanation is omitted by attaching the same numerals as FIG.

  The permanent candidate point determination unit 108 is a terminal considered to have a high effect of installing the terminal pressure measurement device based on the learning result of the terminal pressure by the terminal pressure learning unit 104b and the estimation result of the terminal pressure by the terminal pressure estimation unit 105b. Is estimated, and the estimated end is determined as a candidate for a permanent set point (hereinafter referred to as a “permanent candidate point”).

  FIG. 8 is a flowchart showing an operation example of the distal end pressure control support device 1b according to the third embodiment. The flowchart of FIG. 8 shows the flow of processing in which the terminal pressure control support device 1b determines a permanent installation candidate point. First, the permanent set candidate point determination unit 108 acquires the terminal pressure information and the distribution site information accumulated in the water distribution site information storage unit 101 and the terminal pressure information storage unit 102 (step S301). The permanent candidate point determination unit 108 generates learning data and evaluation data based on the acquired terminal pressure information and water distribution site information (step S302). The learning data is data used to learn the end pressure, and the evaluation data is data used to estimate a permanent point candidate point based on the learning result.

  For example, the permanent candidate point determination unit 108 classifies the end pressure information into learning and evaluation, and sets a set of end pressure information classified as learning and the corresponding water distribution place information as learning data. . Similarly, the candidate permanent point determination unit 108 sets a set of end pressure information classified for evaluation and water distribution site information corresponding thereto as evaluation data. In addition, classification of water distribution place information and terminal pressure information may be performed at random, and may be performed based on arbitrary statistical methods. Note that the learning data and the evaluation data are generated based on any calculation formula that outputs the learning data or the evaluation data using the water distribution place information and the end pressure information as input regardless of the classification as described above. It is also good. Furthermore, in view of the reliability of the permanent candidate point to be estimated, it is desirable that the learning data and the evaluation data be generated so as not to include mutually overlapping data.

  On the other hand, the permanent candidate point determination unit 108 generates a plurality of combinations (hereinafter, referred to as “permanent candidate patterns”) consisting of one or more end that can be a permanent candidate point among the ends of the water distribution network 5 (hereinafter referred to as “constant candidate pattern”). S303). The permanent candidate pattern may be a combination of randomly extracted ends, or may be a combination of ends extracted on any basis. In addition, permanent candidate patterns corresponding to all combinations of all the ends may be generated, or permanent candidate patterns corresponding to some combinations may be generated. Since the number of permanent candidate patterns affects the time required to determine a permanent candidate point, it is adjusted according to the required reliability, the number of locatable end pressure measuring devices, the allowable processing time, etc. Good.

  Subsequently, the end pressure learning unit 104b learns the end pressure for each permanent candidate pattern generated by the permanent candidate point determination unit 108 based on the learning data (step S304). Specifically, assuming that the permanent pressure candidate point is a permanent setting point, the terminal pressure learning unit 104b determines the relationship between the distribution field quantity and the terminal pressure at the permanent setting point, and the terminal pressure at any terminal other than the permanent setting point. learn. This learning is similar to the learning of the end pressure in the second embodiment. In addition, water distribution place information and end pressure information included in data for learning are used for this learning.

  Then, based on the learning result in step S304, the water distribution place information included in the evaluation data, and the terminal pressure information of the permanent installation candidate point included in the evaluation data, the end pressure estimation unit 105b determines any other than the permanent installation point. The end pressure at the end is estimated (step S305). This estimation is similar to the estimation of the end pressure in the second embodiment.

  The permanent candidate point determination unit 108 determines, based on the difference between the estimated value of the terminal pressure estimated in step S305 and the measured value of the terminal pressure included in the evaluation data, for each permanent candidate pattern, a permanent candidate point is a permanently set point. It is determined whether or not the criteria (hereinafter referred to as "constant setup criteria") are satisfied (step S306). For example, if the sum of the difference between the estimated value of the terminal pressure and the actual measurement value in a permanent design candidate pattern is equal to or less than a predetermined threshold, the permanent design candidate point determination unit 108 determines that the permanent design candidate point satisfies the permanent standard. Do. In addition, if the determination method of a permanent installation standard is a permanent installation candidate point as a permanent installation point, it can be determined by any method as long as it can be determined whether the estimation accuracy of the end pressure is equal to or higher than a predetermined accuracy. It may be.

  Subsequently, the permanent installation candidate point determination unit 108 determines whether or not it is determined whether or not the permanent installation standard is satisfied for all permanent installation candidate patterns (step S307). If there is an undecided permanent installation candidate pattern (NO in step S307), the permanent installation candidate point determination unit 108 selects a permanent installation candidate pattern to be determined next, and the process returns to step S304. On the other hand, when there is no undecided permanent installation candidate pattern (step S307-YES), the permanent installation candidate point determination unit 108 outputs information indicating a permanent installation candidate pattern determined to satisfy the permanent installation criteria to the output unit 106 as a processing result. (Step S308).

  The permanent candidate point determining unit 108 does not necessarily have to make the above determination for all permanent candidate patterns. For example, the constant installation candidate point determination unit 108 may be configured to output the estimation result when at least one constant installation candidate pattern that satisfies the standard installation condition is obtained. In addition, the constant installation candidate point determination unit 108 does not necessarily have to output information indicating all the permanent installation candidate patterns that satisfy the regular installation standard. For example, the permanent setting candidate point determination unit 108 may be configured to select and output a permanent setting candidate pattern that is determined to satisfy the constant setting criteria and having a score of a predetermined level or more. In this case, for example, the permanent installation candidate point determination unit 108 gives a higher score to a permanent installation candidate pattern in which the estimation accuracy of the terminal pressure in step S305 is high and the number of permanent installation candidate points is smaller. The higher the number of permanent pattern candidates, the lower the score. That is, when there are a plurality of combinations of the ends determined as constant candidate points (constant candidate patterns), the constant candidate point determination unit 108 determines an end represented by a combination with a smaller number of ends as a constant candidate point. It is also good.

  According to such an end pressure control support device 1b of the third embodiment, in the water distribution system 100b, it is possible to select an appropriate permanent installation point in terms of the end pressure estimation accuracy and the installation cost. Therefore, the end pressure control of the water distribution system 100b can be realized at lower cost by applying the end pressure control support device 1b of the third embodiment.

Fourth Embodiment
FIG. 9 is a diagram showing a specific example of the system configuration of the water distribution system 100c in the fourth embodiment. The water distribution system 100c differs from the first water distribution system 100 in that the end pressure control support device 1c is replaced with the end pressure control support device 1c. Further, in the generation of support information to be provided to the monitoring control system 2, the terminal pressure control support device 1c uses water demand information indicating the demand for water of each customer 6 in addition to the distribution site information and the terminal pressure information. It differs from the end pressure control support device 1 of the first embodiment in the point. In this case, the end pressure control support device 1c is configured to be communicable with a measuring device (not shown) such as a smart meter that measures the amount of water used by each customer 6, and each terminal device communicates with these measuring devices. Acquire water demand information of customer 6

  FIG. 10 is a block diagram showing a specific example of a functional configuration of an end pressure control assisting device 1c according to the fourth embodiment. The terminal pressure control support device 1c according to the fourth embodiment further includes a water demand information storage unit 109, a point including an information acquisition unit 103c instead of the information acquisition unit 103, and an end pressure instead of the terminal pressure learning unit 104. It differs from the end pressure control support device 1 of the first embodiment in that the end pressure estimation unit 105c is replaced with a learning unit 104c and the end pressure estimation unit 105. The other configuration of the end pressure control support device 1c is the same as that of the end pressure control support device 1 of the first embodiment. Therefore, about the other composition, explanation is omitted by attaching the same numerals as FIG.

  The water demand information storage unit 109 is configured using a storage device such as a magnetic hard disk drive or a semiconductor storage device. The water demand information storage unit 109 stores water demand information.

  The information acquisition unit 103c differs from the information acquisition unit 103 in the first embodiment in that it acquires water demand information from each customer 6 in addition to water distribution site information and terminal pressure information. The information acquisition unit 103c stores the acquired water demand information in the water demand information storage unit 109. The information acquisition unit 103c accumulates the water demand information in the water demand information storage unit 109 by repeatedly executing acquisition of the water demand information at a predetermined timing.

  The terminal pressure learning unit 104 c is configured to distribute water distribution site information stored in the water distribution site information storage unit 101, terminal pressure information stored in the terminal pressure information storage unit 102, and water demand information stored in the water demand information storage unit 109. And learn the terminal pressure of the water distribution network 5 based on. Specifically, the end pressure learning unit 104 c learns the relationship between the distribution facility quantity and the water demand of each customer 6 and the end pressure at an arbitrary end of the distribution pipe network 5. In this respect, the end pressure learning unit 104c is different from the end pressure learning unit 104 in the first embodiment. In addition, about the arbitrary terminal used as estimation object, the point which uses the past terminal pressure information of the terminal for learning of the terminal pressure is the same as that of 1st Embodiment.

  The end pressure estimation unit 105c estimates an end pressure corresponding to the amount of water distribution places and the water demand of each customer 6 based on the relationship learned by the end pressure learning unit 104c. That is, the end pressure estimation unit 105c is different from the end pressure estimation unit 105 in the first embodiment in that the current pressure distribution site information and the current water demand information are input to estimate the current end pressure.

  FIG. 11 is a flowchart showing an operation example of the distal end pressure control support device 1c of the fourth embodiment. First, the information acquisition unit 103c acquires distribution site information, terminal pressure information, and water demand information (step S401), and each acquired information is allocated to the distribution site information storage unit 101, terminal pressure information storage unit 102, and water demand information storage. It accumulate | stores in the part 109 (step S402).

  Subsequently, the terminal pressure learning unit 104 c stores the past water distribution site information stored in the water distribution site information storage unit 101, the past terminal pressure information stored in the terminal pressure information storage unit 102, and the water demand information storage unit 109. The terminal pressure is learned based on the past water demand information accumulated in (step S403). Specifically, the end pressure learning unit 104c learns the relationship between the distribution facility quantity and the water demand of each customer 6 and the end pressure at an arbitrary end of the distribution pipe network 5. That is, according to this learning, a discriminator that outputs the end pressure at an arbitrary end of the water distribution pipe network 5 with the water distribution site information and the water demand information as inputs is generated. The end pressure learning unit 104c outputs the learning result to the end pressure estimation unit 105c.

  Subsequently, the information acquisition unit 103c acquires current water distribution place information and current water demand information (step S404). The end pressure estimation unit 105c estimates the current end pressure at an arbitrary end of the water distribution network 5 based on the learning result by the end pressure learning unit 104c and the current water distribution place information and the current water demand information ((4) Step S405). The end pressure learning unit 104 c outputs estimated end pressure information indicating the estimated end pressure to the output unit 106. The output unit 106 generates support information based on the estimated terminal pressure information, and transmits the generated support information to the monitoring control system 2 (step S406).

  The terminal pressure control supporting device 1c according to the fourth embodiment configured as described above uses the past water distribution site information, the past terminal pressure information, and the past water demand information to use the terminal pressure of the water distribution pipe network 5 And estimate the current terminal pressure at any end of the water distribution network 5 based on the current water distribution site information and the current water demand information. As described above, the terminal pressure control supporting device 1 c performs the learning and estimation of the terminal pressure by using the water demand information of each customer 6 in addition to the distribution site information and the terminal pressure information. It is possible to estimate the current end pressure more accurately.

Fifth Embodiment
A water distribution system 100d (not shown) in the fifth embodiment has the same system configuration as the water distribution system 100a in the second embodiment, and includes an end pressure control assistance device 1d in place of the end pressure control assistance device 1a. The point is different from the water distribution system 100a in the second embodiment.

  FIG. 12 is a block diagram showing a specific example of a functional configuration of an end pressure control assisting device 1d of the fifth embodiment. The end pressure control support device 1d is different from the end pressure control support device 1a of the second embodiment in that the end pressure control support device 1d further includes an estimation result information storage unit 110 and a water leakage diagnosis unit 111. The other configuration of the end pressure control support device 1d is the same as that of the end pressure control support device 1a of the second embodiment. Therefore, about the other composition, explanation is omitted by attaching the same numerals as FIG.

  The estimation result information storage unit 110 is configured using a storage device such as a magnetic hard disk drive or a semiconductor storage device. The estimation result information storage unit 110 stores estimated end pressure information indicating the end pressure estimated by the end pressure estimation unit 105. As a result, estimated end pressure information in the past is accumulated in the estimation result information storage unit 110.

  The water leakage diagnosis unit 111 detects the water leakage in the water distribution network 5 based on the past terminal pressure information accumulated in the terminal pressure information storage unit 102 and the past estimated terminal pressure information accumulated in the estimation result information storage unit 110. To diagnose the presence or absence of

  FIG. 13 is a flow chart showing an operation example of the distal end pressure control support device 1d of the fifth embodiment. The flowchart of FIG. 13 shows the flow of processing in which the terminal pressure control support device 1d diagnoses the presence or absence of water leakage in the water distribution network 5. Here, the same processes as those of the operation example in the second embodiment are denoted by the same reference numerals as those in FIG. That is, each process of step S201 to S205 is the same as that of the operation example in the second embodiment, and learning and estimation of the end pressure are performed in these steps.

  Subsequently, the end pressure estimation unit 105a stores the estimated end pressure information in the estimation result information storage unit 110. The end pressure estimation unit 105a accumulates estimated end pressure information in the estimation result information storage unit 110 by repeatedly performing estimation of the end pressure (step S501).

  The water leakage diagnosis unit 111 detects the water leakage in the water distribution network 5 based on the past terminal pressure information accumulated in the terminal pressure information storage unit 102 and the past estimated terminal pressure information accumulated in the estimation result information storage unit 110. It diagnoses the presence or absence of (step S502). Specifically, the water leakage diagnosis unit 111 estimates that water leakage has occurred in the water distribution network 5 when the state where the measured value of the end pressure is larger than the estimated value of the end pressure continues for a predetermined time. When it is determined that water leakage is occurring at a certain end, the water leakage diagnosis unit 111 notifies the output unit 106 to that effect (step S503). Then, the output unit 106 outputs the information indicating the occurrence of water leakage (hereinafter referred to as “water leakage information”) in a predetermined mode in response to the notification of the water leakage diagnosis unit 111, thereby generating the water leakage in the water distribution pipe network 5. To inform For example, the leaked water information may be displayed on the display device or may be output as an audio through a speaker. In addition, the water leakage information may be included in the support information and transmitted to the monitoring control system 2.

  The terminal pressure control supporting device 1d of the fifth embodiment configured as described above can diagnose the presence or absence of water leakage in the water distribution system 100d and notify the diagnosis result. With such notification, the water distribution system 100d can perform appropriate processing at an appropriate timing according to the occurrence of water leakage. For example, when the diagnosis result is notified to the monitoring control system 2, the monitoring control system 2 can reduce the amount of water leakage by lowering the discharge pressure of the pump of the water distribution station 3. Therefore, according to the end pressure control assisting device 1d of the fifth embodiment, the end pressure of the water distribution network 5 can be more appropriately controlled.

(Modification)
In the fifth embodiment, although the leak detection unit 111 determines the presence or absence of leak based on the measured value and the measured value of the end pressure at each end, the leak detection unit 111 determines the presence or absence of water leak at each end The presence or absence of water leakage may be determined based on the result of water leakage diagnosis at the end of the periphery. For example, in this case, pipe network information indicating the distance or pipe length between the permanently installed points is stored in advance in the terminal pressure control support device 1d. If the water leakage diagnosis unit 111 determines that leakage is occurring at the end of a certain diagnosis target, the other end located within a predetermined range from the end of the diagnosis target based on the pipe network information (hereinafter “peripheral end” Identification). Then, when the water leakage diagnosis unit 111 determines that the water leakage is generated also at the peripheral end, the water leakage diagnosis unit 111 may finally determine that the water leakage is generated at the end of the diagnosis target. As described above, by determining the presence or absence of water leakage at a certain end based on the diagnosis results of the plurality of ends, it becomes possible to diagnose the presence or absence of water leakage more accurately.

  The terminal pressure control support device 1c of the fourth embodiment may include the permanent terminal pressure measurement device 107 of the second embodiment. In this case, the end pressure control support device 1c is an arbitrary end other than the permanent installation point based on the past water distribution site information, the past permanent end pressure information, the past end pressure information, and the past water demand information. You may learn the end pressure in and estimate the current end pressure at any end other than the permanent set point based on the current distribution site information, the current permanent end pressure information, and the current water demand information .

  According to at least one embodiment described above, an end pressure learning unit that learns by machine learning the relationship between water distribution site quantities and end pressure based on past end pressure information and past water distribution site information The end pressure of the water distribution system by having the end pressure estimation unit that estimates the current end pressure in the water distribution network based on the relationship learned by the end pressure learning unit and the current distribution site information. Control can be realized at lower cost.

  While certain embodiments of the present invention have been described, these embodiments have been presented by way of example only, and are not intended to limit the scope of the invention. These embodiments can be implemented in other various forms, and various omissions, replacements, and modifications can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the invention described in the claims and the equivalents thereof as well as included in the scope and the gist of the invention.

1, 1a, 1b, 1c, 1d ... end pressure control support device, 2 ... monitoring control system, 3, 3-1, 3- 2 ... water distribution station, 4, 4 -1, 4-2 ... end pressure measurement device, 5 ... water distribution pipe network, 6,6-1 to 6-3 ... consumer, 100, 100a, 100b, 100c, 100d ... water distribution system, 101 ... water distribution site information storage unit, 102 ... end pressure information storage unit, 103, 103c ... information acquisition unit, 104, 104a, 104b, 104c ... terminal pressure learning unit, 105, 105a, 105b, 105c ... terminal pressure estimation unit 106 ... output unit 107: permanent terminal pressure measuring device 108: permanent candidate point Determination unit, 109 ... water demand information storage unit, 110 ... estimation result information storage unit, 111 ... water leakage diagnosis unit

Claims (11)

An end pressure learning unit that learns the relationship between an end pressure of a water distribution network and various quantities related to water distribution through the distribution network using a machine learning method;
An end pressure estimation unit configured to estimate an end pressure of the water distribution network based on the relationship learned by the end pressure learning unit and information indicating the quantities;
Terminal pressure control support device comprising: An end pressure information storage unit that accumulates end pressure information indicating an end pressure of the water distribution network;
A distribution station information storage unit that stores distribution station information indicating various quantities relating to a distribution station distributed to the distribution pipe network;
And further
The terminal pressure learning unit learns the relationship between the quantities and the terminal pressure based on the past terminal pressure information and the past water distribution site information,
The terminal pressure estimation unit estimates a current terminal pressure in the water distribution network based on the relationship learned by the terminal pressure learning unit and current water distribution site information.
The end pressure control support device according to claim 1. It further comprises a measurement unit that measures an end pressure at a specific end of the water distribution network,
The end pressure learning unit is configured to calculate the quantities and the end pressure at the specific end based on past end pressure information including the end pressure information acquired at the specific end and past distribution site information. Learn the relationship with the end pressure at any end other than the specific end in the distribution network,
The terminal pressure estimation unit is configured to calculate the distribution network based on the relationship learned by the terminal pressure learning unit, current distribution site information, and current terminal pressure information acquired at the specific terminal. Estimate the current end pressure at said any end of
The terminal pressure control support device according to claim 2. And a determination unit configured to determine a specific end at which the end pressure is measured in the water distribution network based on the end pressure estimation result by the end pressure estimation unit,
The end pressure learning unit assumes a predetermined end of the end whose end pressure is indicated by the end pressure information as the specific end based on the distribution field information and the end pressure information given for learning. Learn the relationship between quantities and end pressure at the specified end and end pressure at any end other than the specified end in the water distribution network,
The terminal pressure estimation unit is configured to calculate the relationship learned by the terminal pressure learning unit, the distribution field information given for evaluation, and the terminal pressure of the specific terminal among the terminal pressure information given for evaluation. End pressure at the arbitrary end of the water distribution network based on the end pressure information indicating
The determination unit may set the predetermined end as the specific end based on the difference between the end pressure indicated by the end pressure information given for evaluation and the end pressure estimated by the end pressure estimation unit Determine whether or not
The terminal pressure control assistance device according to claim 2 or 3. The determining unit determines, as the specific end, an end represented by a combination having a smaller number of ends, when there is a plurality of combinations of the ends determined as the specific ends.
The terminal pressure control support device according to claim 4. The system further comprises a water demand information storage unit that stores water demand information indicating water demand of each customer,
The end pressure learning unit learns the relationship between the various amounts, the end pressure, and the water demand based on the past water distribution site information, the past end pressure information, and the past water demand information,
The terminal pressure estimation unit is configured to calculate a current terminal at an arbitrary terminal of the water distribution network based on the relationship learned by the terminal pressure learning unit, current water distribution site information, and current water demand information. Estimate the pressure,
The terminal pressure control support device according to claim 2. It further comprises a water leakage diagnosis unit for diagnosing the presence or absence of water leakage in the water distribution network,
The terminal pressure estimation unit estimates the current terminal pressure at the specific terminal based on the relationship learned by the terminal pressure learning unit and current distribution site information.
The water leakage diagnosis unit is configured such that the end pressure at the specific end measured by the measurement unit continues to be higher than the end pressure at the specific end estimated by the end pressure estimation unit for a predetermined time period. Determining that water leakage has occurred in the water distribution network,
The terminal pressure control support device according to claim 3. The water leakage diagnosis unit is configured to, when a state where the measured value of the end pressure is larger than the estimated value of the end pressure continues for a predetermined time at a specific end and an end around the specific end. It is determined that water leakage has occurred at the end,
The end pressure control support device according to claim 7. The terminal pressure learning unit learns the relationship by a machine learning method based on a random forest.
The end pressure control support device according to any one of claims 1 to 8. A learning step of learning a relationship between an end pressure of a water distribution network and various quantities related to water distribution through the distribution network using a machine learning method;
Estimating the terminal pressure of the water distribution network based on the relationship learned in the learning step and the information indicating the quantities;
Terminal pressure control support method having: A learning step of learning a relationship between an end pressure of a water distribution network and various quantities related to water distribution through the distribution network using a machine learning method;
Estimating the terminal pressure of the water distribution network based on the relationship learned in the learning step and the information indicating the quantities;
A computer program that causes a computer to run.

Images