KR20240062396A - System for managing water distribution supply network using sentimental analysis and big data machine learning, and method for the same - Google Patents

Abstract

By applying web crawling-based consumer sentiment analysis and big data spatial information machine learning model in an environment where urban buildings and underground water pipe networks are implemented in a 3D cyber physical system, the water quality of the urban water pipe network is improved. Accident tracking management and safety evaluation management can be performed. In addition, the water supply network can be spatially visualized by implementing it in a 3D cyber-physical system environment to enable proactive response to water quality accidents in the water pipe network as well as tracking management in the event of an accident. In addition, in the event of a water quality accident or disaster, nearby residents can quickly evacuate the extent of damage, and the competent administrative agency can quickly predict the extent of damage and establish a response system using constant water analysis and big data machine learning. A pipe network management system and method are provided.

Description

Water pipe network management system and method using sentiment analysis and big data machine learning {SYSTEM FOR MANAGING WATER DISTRIBUTION SUPPLY NETWORK USING SENTIMENTAL ANALYSIS AND BIG DATA MACHINE LEARNING, AND METHOD FOR THE SAME}

The present invention relates to a water distribution network management system, and more specifically, to web crawling in an environment where city buildings and underground water distribution networks are implemented as a 3D cyber physical system. ) - A water pipe network management system using emotional analysis and big data machine learning that can perform water quality accident tracking management and safety evaluation management of the urban water pipe network by applying consumer emotional analysis and big data spatial information machine learning model. and methods thereof.

Recently, water quality accidents such as red water and larvae have occurred frequently in the process of supplying tap water. The main causes of these water quality accidents are pointed out to be poor operation by unskilled drivers, aging water pipes (water pipes), and old indoor water supply pipes.

In the case of drinking tap water, safety is ensured through the installation of water purifiers in each household, but in the case of domestic water, there are insufficient measures to respond when contaminated tap water is supplied. In addition, some households are responding by installing and using individual water purification filters at the water supply end of faucets, showers, etc., but there are limits to regular response due to reliability issues and frequent replacement cycles of water purification filters. there is.

In addition, in the case of the water quality of tap water, each local government's water supply business periodically analyzes the water quality and discloses it on its website, etc., but the only way for the general public to actually check the safety of the water quality of the water supplied to homes is through turbidity (foreign substances). It is difficult for the general public to check the reliability of the water supply because the only way is to check the chromaticity (red number) with the naked eye or check whether the water has an odor.

However, it is known that the taste and smell that sometimes occurs in tap water are aesthetic items and do not pose a threat to public health to a certain level.

However, because water quality accidents in water pipe networks occur due to various causes, the variables that provide various causes are systematized and stratified.

In addition, when a water quality accident occurs, tracking and management must be possible based on the water pipe network diagram, and it is also necessary to establish a safety evaluation system for water quality in the water pipe network based on a prediction model for various factors.

Meanwhile, as prior art, Republic of Korea Patent No. 10-1875885 discloses an invention titled “Waterworks Pipe Network Operation and Management System through Pipe Network Analysis,” which is a pipe network analysis operation simulation system to predict and minimize the extent of accidents when an accident occurs. This will be described with reference to FIGS. 1A and 1B.

Figure 1a is a diagram showing a water pipe network operation and management system through pipe network analysis according to conventional technology, and Figure 1b is a diagram showing the process of the water pipe network operation and management system through pipe network analysis shown in Figure 1a.

Referring to Figure 1a, in the case of a water pipe network operation management system through pipe network analysis according to conventional technology, it is a system for predicting and minimizing the extent of accidents in an emergency through the operation management server 10, and the operation management server 10 is a data It includes a collection module, model construction module, pipe network analysis module, model correction module, and control module.

At this time, the control module is a device for supplying water to consumers and responding to accidents based on pipe network analysis and simulation results.

This operation management server 10 controls the operations of the data collection module, model construction module, pipe network analysis module, and model correction module through the control module, simulating water supply conditions to consumers, simulating water quality conditions, simulating stagnant water, and high pressure/discharge. It can perform the respective functions of operating software that performs functions such as water simulation and customer meter reading collection, and operating simulation operating system (OS) that performs basic file management and various simulations.

The operation simulation (pipe network analysis) DB (20) not only manages linkage between other systems (30) such as the monitoring and control system, customer management system, demand forecasting system, production planning system, and geographic information system (GIS),

It can be equipped with basic file management functions such as unit management, analysis setting management, coefficient management, curve management, pattern management, basic file creation, and basic file editing.

Referring to Figure 1b, when analyzing the pipe network in the water pipe network operation and management system through pipe network analysis according to conventional technology, the data collection module sets the block of the pipe based on the water pipe network diagram provided by the GIS server, and compares the water pipe network diagram with the water pipe network diagram. A basic file for pipe network analysis is created based on the water rate, flow rate, water level information of the drainage basin, and water quality information provided by the linked server.

Next, the model building module is based on the basic file of the pipe network analysis, and includes operation status information including water supply to customers, stagnant water, high pressure and discharged water, and water quality changes; And it generates accident situation information for each situation, including water leakage situation, pipe cleaning, water outage, and water supply to consumers.

Next, the pipe network analysis module generates pipe network analysis information including flow rate, pressure, and water quality information of the block, scenarios and results for each accident situation, according to the pipe network analysis algorithm based on operation status information and situation-specific accident situation information.

Next, the model correction module applies the pipe network analysis information to the model correction algorithm to correct the existing pipe friction coefficient or residual chlorine reduction coefficient.

In other words, the water pipe network operation and management system through pipe network analysis according to conventional technology is a water pipe network operation and management system through pipe network analysis tailored to the residual chlorine water quality simulation considering the pipe friction coefficient and residual chlorine reduction coefficient on the water pipe network.

However, in the case of a water pipe network operation and management system through pipe network analysis according to conventional technology, in the actual environment, due to lack of operation by skilled operators, aging of water pipes, and old indoor water pipes, red water, rust, black water, foreign substances and larvae, etc. There are limitations in that prediction is impossible with a pipe network analysis model and tracking and management is difficult in case of water quality abnormalities.

Meanwhile, as another prior art, Republic of Korea Patent No. 10-1585552 discloses an invention titled “Big data-based integrated water supply operation management system and method,” which is explained with reference to FIGS. 2A and 2B.

Figure 2a is a schematic network configuration diagram of an integrated water supply operation and management system according to conventional technology, and Figure 2b is a specific configuration diagram of the water supply integrated operation and management system.

Referring to FIGS. 2A and 2B, in the case of a water supply integrated operation management system according to the prior art, the measuring device 51 may include a flow meter 51a, a water quality meter 51b, a water pressure meter 51c, etc. There is,

It is installed within each block of the water supply network, which is divided into multiple blocks for each region, and measures various measurement data related to water supply use for each block. The flow meter 51a measures the flow rate, flow rate, and water level of the water for each block, the water quality meter 51b measures the water quality of the water, and the water pressure meter 51c measures the water pressure of the water.

The water supply network monitoring device 54 sets up a block of the water supply network into multiple blocks for each region based on the water supply network map, and monitors the status of the water supply network using real-time measurement data including the flow rate and water pressure of each pipe. At this time, the water supply network map can be created based on GIS data provided by the GIS server 52.

The water supply integrated operation management system 53 blocks the water supply pipe network into a number of blocks for each region, and for each blocked region, human and social information including information on the number of residents, age, gender, and number of businesses, and water supply use related information. Based on the information, we evaluate the water supply usage and analyze the characteristics of the region to analyze the correlation between the water usage by region, the deterioration of the pipe network, and the distribution of the resident population.

Specifically, the water supply integrated operation management system 53 includes a communication department 53a, a pipe network information collection department 53f, a water usage evaluation department 53g, a regional characteristics analysis department 43h, an applied statistics department 53d, and policy setting. It may include a portion 53c, etc.

The communication unit 53a interfaces for transmission and reception of pipe network information data between the waterworks integrated operation management system 53 and the waterworks pipe network monitoring device 54, and transmits and receives GIS data between the waterworks integrated operation management system 53 and the GIS server 52. interface.

The pipe network information collection unit 53f collects GIS data of the water supply pipe network to be managed and regional water leakage, water pressure, and water flow rate information of the water supply pipe network from the communication network GIS server 52.

The water usage evaluation unit 53g evaluates the water usage by region using information collected from the pipe network information collection unit 53f.

At this time, in order to evaluate the water usage amount, the water usage evaluation unit 53ㅎ, for example, analyzes factors causing deviations in water usage per person per day, and analyzes the water usage by population by age, day and night residence, business, and household/house type in detail by water supply unit. Factor analysis, etc. can be performed.

In other words, the water usage evaluation unit 53g not only monitors the water usage by region, but also evaluates the water usage through various analyzes linking the water usage with human and social information by region, thereby providing information on water supply network management, maintenance, and repair. It can provide more useful information to efficiently establish policies during work.

Referring again to FIG. 2A, the user terminal 55 is a terminal device capable of accessing the waterworks integrated operation and management system 53 through a communication network such as the Internet, and may be a terminal device such as a personal PC, laptop, tablet, or smartphone. It can be.

The manager or operator uses the user terminal 55 to access the water supply integrated operation and management system 53 through a communication network such as the Internet, thereby providing water supply pipe network monitoring information or water supply information provided by the water supply integrated operation management system 53. You can easily check pipe network analysis information, etc.

In other words, in the case of a big data-based integrated water supply operation management system according to conventional technology, the water pressure and water quality management unit that performs flow data for water leakage management in the blocked water pipe network and water pressure and water quality data of each block is GIS-based. Generates an alarm indicating the location on the water supply network map,

Using evaluation information on the appropriateness of water usage and applied statistical information, the water supply network can be managed by presenting various statistical information by displaying the aging of facilities, facility density distribution, civil complaint density, or water leakage distribution in the water supply network at the relevant location.

However, the big data-based integrated water supply operation and management system according to conventional technology has limitations in that water quality items are not specified in relation to the water quality of the water pipe network, and in particular, tracking and management is difficult when water quality abnormalities occur.

Meanwhile, Republic of Korea Patent No. 10-2085122 discloses an invention titled “Tap water real-time smart management system using big data-based artificial intelligence,” which will be described with reference to FIGS. 3A to 3C.

Figure 3a is a schematic diagram of a water supply facility to which a real-time smart tap water management system using big data-based artificial intelligence according to conventional technology is applied;

Figure 3b is a detailed configuration diagram of a collection device of a real-time smart management system for tap water using big data-based artificial intelligence, and Figure 3c is a detailed configuration diagram of an analysis device for a real-time smart management system for tap water using artificial intelligence based on big data.

As shown in Figure 3a, the real-time smart tap water management system using big data-based artificial intelligence according to conventional technology includes a sensor device 70, a collection device 81, an analysis device 82, and a display device 83. It is composed including.

Here, a general water treatment facility for water supply includes a water intake station (61) that intakes water from a water source through a water intake pipe (61a); A water purification plant (62) that receives water taken from the water intake plant (61) through a water pipe (62a) and purifies it; A drainage basin (63) that receives purified water from the water purification plant (62) through the water supply pipe (63a) and temporarily stores it; and a customer 64 that receives water stored in the drainage basin 63 through a water supply pipe 64a.

The water purification plant 62 includes a silt basin, water intake pumping station, submerged well, pre-ozone treatment plant, mixed coagulation sedimentation pond, sand filter, post-ozone treatment plant, activated carbon filter, chlorine input room, purification pond, and pressurization plant.

The sensor device 70 is installed in each pipe, such as a water intake pipe, water pipe, water supply pipe, and water supply pipe, and detects the water quality of the water transported through the pipe.

Specifically, as shown in FIG. 3B, the collection device 81 includes a sensing value receiving unit 81a, a sensing value storage unit 81b, and a sensing value transmitting unit 81c.

The sensing value receiver 81a receives the sensing value detected from the sensor device 70 in real time, and the sensing value includes identification information. In addition, the sensing value receiver 81a includes a Real Time Clock (RTC) and thus receives real-time time information and day and date information based on the RTC.

The sensing value storage unit 81b functions to store and manage the sensing values received from the sensing value receiving unit 81a by classifying them according to identification information. For example, the sensing value storage unit 81b stores and manages the received sensing value based on identification information including location information and pipeline information.

The sensing value transmitting unit 81c transmits the sensing value stored in the sensing value storage unit 81a to the analysis device 82.

The analysis device 82 analyzes and predicts risk factors based on the sensing values stored in the collection device 81, classifies the collected sensing values by identification information, and collects contextual data about the sensing values classified by identification information. It analyzes and predicts risk factors for the sensing value through artificial intelligence analysis using the constructed situational data and provides visual information.

Specifically, referring to FIG. 3C, the analysis device 82 includes a data separation unit 82a, a data construction unit 82b, a prediction determination unit 82c, and a self-learning update unit 82d.

The data separation unit 82a stores and manages the collected sensing values by dividing them by identification information.

That is, the data separation unit 82a stores and manages the sensed values by dividing them by identification information included in the collected sensed values.

The data construction unit 82b constructs context-specific data by patterning the sensing values classified by identification information in the data separation unit 82a.

The prediction determination unit 82c analyzes sensing values corresponding to abnormal signs based on related sensing values among the situational data constructed in the data construction unit 82b and performs the function of determining the prediction direction.

Specifically, the prediction determination unit 82c includes an abnormality symptom confirmation module 82c-1, a related sensing value derivation module 82c-2, a warning limit value setting module 82c-3, and a prediction direction determination module 82c-4. It is composed including.

The abnormality symptom confirmation module 82c-1 detects and confirms the sensing value corresponding to the abnormality symptom among the sensing values. That is, the abnormality symptom confirmation module 82c-1 performs the function of checking whether the detected sensing value is in the normal range, and checks whether the detected sensing value is in the normal range calculated by the data construction unit 82b. , If it does not fall within the normal range, it is judged to be an abnormality.

The related sensing value derivation module 82c-2 determines the factors affecting the sensing value confirmed in the abnormality symptom confirmation module 82c-1 as the related sensing value, and derives the determined related sensing value. At this time, the sensing value determined to be an abnormality may be related to the aging of water purification chemicals and pipes used in water purification plants, etc.

Therefore, the related sensing value derivation module 82c-2 derives the related sensing values of various factors (iron ion concentration, copper ion concentration, suspended matter, electrical conductivity, turbidity, residual chlorine, etc.) linked to the sensing value determined to be an abnormality. do.

The self-learning update unit 82d performs a function of updating situation-specific data constructed in the data construction unit. The data accumulated through these updates can be used for various learning of patterned situation-specific data and can also be used as data for algorithms for predictive judgment.

According to the real-time smart management system for tap water using artificial intelligence based on big data according to conventional technology, the water quality status of all pipes from the process of collecting raw water to sending purified tap water to consumers is collected, and based on this, whether the water quality is abnormal or not. Since it can be predicted, the reliability of tap water can be improved.

Additionally, since turbidity, pH, water temperature, redox potential, and conductivity can be detected through a single sensor device, the sensor device can be easily installed in the pipeline, and transmission and reception of the detected sensing values is relatively easy.

However, even in the case of a real-time smart tap water management system using artificial intelligence based on big data according to conventional technology, there is a limitation in that it is difficult to track and manage water quality abnormalities.

Meanwhile, considering that consumer complaints occur through various channels before large-scale water quality accidents occur, web crawling-based consumer sentimental analysis is an important input variable when building a water pipe network water quality safety evaluation model. You can use it.

Specifically, sentiment analysis, also called opinion mining, can be defined as a natural language processing technology that analyzes subjective data such as people's attitudes, opinions, and tendencies expressed in text.

This sentiment analysis is a technology that analyzes public opinion and opinions expressed on websites and social media and reprocesses them into useful information. By analyzing text, netizens' emotions and opinions can be converted into objective information by statistics and numbers.

Meanwhile, the typical water pipe network safety evaluation management system aims to minimize the affected area in emergencies such as water quality accidents and to ensure efficient operation and management in normal times by linking network operation data analysis and pipe network interpretation for each unit area of the water pipe network.

This water pipe network safety evaluation management system can determine the grade of the water pipe network through not only the general status of the water pipe network but also various technical diagnosis data.

However, since most water pipe networks are buried underground and are affected by topographical characteristics such as buried soil and water pressure, it is very difficult to identify and predict the exact degree of corrosion and deterioration.

Accordingly, an evaluation method that can indirectly diagnose the deterioration of the water pipe network is needed rather than an evaluation through actual measurements. In this case, since it must be interpreted in combination with directly observed image data, expertise in the relevant field is required.

In addition, some organizations use a score evaluation method that assigns points based on the status of each factor related to deterioration and sums the scores of each factor item to determine which pipes need priority maintenance.

Alternatively, a Probabilistic Neural Network algorithm is introduced to determine the degree of deterioration of buried pipes using a model that divides the degree of deterioration of buried pipes into 5 levels and classifies them according to the magnitude of probability, thereby determining the soundness of the buried pipes. is evaluating.

However, since the deterioration evaluation method varies depending on the material and information of the buried pipe, field applicability decreases as the model becomes more complex, and it is difficult to say that the deterioration of the buried pipe is the direct cause of the water quality accident issue.

In the end, the method of evaluating the deterioration of the water pipe network requires various information such as the deterioration of the buildings in the area where the buried pipes are located, household density, building use, road environment, soil environment, and corrosiveness information of the buried pipes, and is suitable for the site. Model construction can be said to be very important, and a water pipe network management system that takes this into account is needed.

Republic of Korea Patent No. 10-2085122 (registration date: February 28, 2020), title of invention: “Real-time smart management system for tap water using big data-based artificial intelligence” Republic of Korea Patent No. 10-1875885 (registration date: July 2, 2018), title of invention: “Water pipe network operation and management system through pipe network analysis” Republic of Korea Patent No. 10-1585552 (registration date: January 8, 2016), title of invention: “Big data-based integrated water supply operation management system and method” Republic of Korea Patent No. 10-1736666 (registration date: May 10, 2017), title of invention: “Water pipe network water operation method” Republic of Korea Patent No. 10-1687636 (registration date: December 13, 2016), title of invention: “Waterwork network management method, server and system” Republic of Korea Patent No. 10-1205103 (registration date: November 20, 2012), title of invention: “Water pipe network operation management system” Republic of Korea Patent No. 10-1183656 (registration date: September 11, 2012), title of invention: “Waterwork operation management system and control method thereof”

The technical task to be achieved by the present invention to solve the above-mentioned problems is to conduct web crawling-based consumer sentiment analysis and big data analysis in an environment where city buildings and underground water pipe networks are implemented as a 3D cyber physical system (3D Cyber Physical System). The purpose is to provide a water pipe network management system and method using emotional analysis and big data machine learning that can perform water quality accident tracking management and safety evaluation management of the urban water pipe network by applying data spatial information machine learning model. .

Another technical task that the present invention aims to achieve is to implement emotional analysis and big data that can be spatially visualized by implementing the water supply network in a 3D cyber-physical system environment to enable proactive response to water quality accidents in the water pipe network as well as tracking and management in the event of an accident. The purpose is to provide a water pipe network management system and method using machine learning.

Another technical task that the present invention aims to achieve is to provide a sensitive system that allows nearby residents to quickly evacuate the extent of damage in the event of a water quality accident or disaster, and allows the competent administrative agency to quickly predict the extent of damage and establish a response system. The purpose is to provide a water pipe network management system and method using analysis and big data machine learning.

As a means of achieving the above-described technical task, the water pipe network management system using emotional analysis and big data machine learning according to the present invention includes water pipe network asset information, water quality complaint-related data, administrative data, and water quality-related data. A water quality-related information provision unit that preprocesses data at each point to enable hierarchical spatial analysis of water quality-related information and provides it as big data; A water quality-related consumer sentiment analysis server that obtains complaints related to tap water quality through web crawling and provides consumer sentiment analysis information related to tap water quality; A weather information provider that provides real-time weather information in the area where the water pipe network is located; Geographic information provider that provides geographic information about the area where the water pipe network is located; And based on the water quality-related consumer emotional analysis information, internal and external data provided by the water quality-related information provision department, the weather information, and the geographic information are collected to derive priority areas for water quality management in the water pipe network, and when a water quality accident occurs, It supports tracking management by considering the connection nodes of the water pipe network system, calculates the damaged area and number of households, and includes a water pipe network management server that supports water quality accident safety evaluation in the water pipe network.

Meanwhile, as another means of achieving the above-described technical problem, the water pipe management method using emotional analysis and big data machine learning according to the present invention includes the steps of a) presetting input information for water pipe network management; b) a step where the water pipe network management server collects real-time weather information from a weather information provider; c) the water pipe network management server collecting water quality data measured in real time by sensors installed in the water pipe network; d) the water pipe network management server collecting and monitoring internal data/external data from the water quality-related information provider; e) the water pipe network management server processing 3D meteorological physical system-based spatial information according to geographic information provided by a geographic information provider using preset input information and real-time measured measurement information; f) a water quality-related consumer emotional analysis server generating consumer emotional analysis information according to emotional trends collected through web crawling; g) the water pipe network management server analyzing water pipe network-related data according to emotional analysis and big data machine learning; and h) a step of the water pipe network management server visualizing and transmitting the results of water quality accident tracking management and safety evaluation management, wherein the water pipe network management server is configured to visually display and deliver the results of the water quality accident tracking and safety evaluation management, wherein the water pipe network management server is configured to visually display and transmit the results of the water quality accident tracking and safety assessment management, wherein the water pipe network management server is configured to visually display and deliver the results of the water quality accident tracking and safety assessment management, wherein the water pipe network management server is configured to visually display and transmit the results of the water quality accident tracking and safety evaluation management. By collecting internal and external data provided by the information provider, the weather information, and the geographic information, priority areas for water quality management in the water pipe network are derived, and when a water quality accident occurs, tracking management is supported by considering the connection nodes of the water pipe network system. It is characterized by calculating the damaged area and number of households, and supporting safety evaluation of water quality accidents in the water pipe network.

According to the present invention, by applying web crawling-based consumer sentiment analysis and big data spatial information machine learning model in an environment where urban buildings and underground water pipe networks are implemented as a 3D cyber physical system, urban You can perform water quality accident tracking and safety evaluation management of your water pipe network.

According to the present invention, the water supply network can be spatially visualized by implementing it in a three-dimensional cyber-physical system environment to enable proactive response to water quality accidents in the water pipe network as well as tracking and management in the event of an accident.

According to the present invention, in the event of a water accident or disaster, nearby residents can quickly evacuate the extent of damage, and the competent administrative agency can quickly predict the extent of damage and establish a response system.

Figure 1a is a diagram showing a water pipe network operation and management system through pipe network analysis according to conventional technology, and Figure 1b is a diagram showing the process of the water pipe network operation and management system through pipe network analysis shown in Figure 1a.
Figure 2a is a schematic network configuration diagram of an integrated water supply operation and management system according to conventional technology, and Figure 2b is a specific configuration diagram of the water supply integrated operation and management system.
Figure 3a is a schematic diagram of a water supply facility to which a real-time smart tap water management system using big data-based artificial intelligence according to conventional technology is applied, and Figure 3b is a diagram of the collection device of the real-time smart tap water management system using big data-based artificial intelligence. This is a specific configuration diagram, and Figure 3c is a specific configuration diagram of the analysis device of the tap water real-time smart management system using big data-based artificial intelligence.
Figure 4 is a schematic configuration diagram of a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention.
Figure 5 is a detailed configuration diagram of a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention.
Figure 6 is a diagram summarizing the data purification process used to evaluate water quality safety in the water pipe network and derive water quality management areas in the water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention.
Figure 7 is a diagram illustrating the classification of civil complaints that occurred on a specific day, the water pipe network, and the expression of old pipes in the water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention.
Figure 8 is a diagram illustrating an urban water pipe network/water supply pipe in a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention.
Figures 9A to 9C are diagrams for explaining consumer emotional analysis related to water quality in a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention.
Figure 10a is a diagram illustrating a neural network algorithm for classification in a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention, and Figure 10b is an operation flow chart showing the data processing sequence according to the neural network algorithm for classification. am.
Figure 11 is a detailed configuration diagram of the data analysis unit in the water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention.
Figure 12a is a detailed configuration diagram of the water supply status and civil complaint analysis unit in the data analysis unit shown in Figure 11, Figure 12B is a specific configuration diagram of the water supply deterioration analysis unit, and Figure 12C is a specific configuration diagram of the water supply water quality management area analysis unit.
Figure 13 is a diagram illustrating a tracking management function that enables the scale of damage and direction of spread when a water quality abnormality or water quality accident occurs for a target block in a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention. .
Figure 14 is an operation flowchart showing a water pipe management method using emotional analysis and big data machine learning according to an embodiment of the present invention.

Below, with reference to the attached drawings, embodiments of the present invention will be described in detail so that those skilled in the art can easily implement the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. In order to clearly explain the present invention in the drawings, parts unrelated to the description are omitted, and similar parts are given similar reference numerals throughout the specification.

Throughout the specification, when a part is said to “include” a certain element, this means that it may further include other elements rather than excluding other elements, unless specifically stated to the contrary. Additionally, terms such as “… unit” used in the specification refer to a unit that processes at least one function or operation, and may be implemented as hardware, software, or a combination of hardware and software.

First, various water quality abnormalities or water quality accidents related to red water, green matter, foreign substances, larvae, and other small organisms that occur in distribution water systems such as urban water pipe networks are social issues, and the resulting social costs are increasing. In addition, since the deterioration of underground pipes is not the direct cause of water quality abnormalities or accidents, improvements are induced by methods such as pipe cleaning based on the safety evaluation results for each block calculated by the water pipe network management system according to the present invention. Social costs such as replacement can be reduced.

Accordingly, the water pipe network management system supports tracking management by considering the connection nodes of the water pipe network system when a water quality accident occurs, calculates the damaged area and number of households, and is a system that supports safety evaluation of water quality accidents in the water pipe network before the accident occurs. Alternatively, the information must be able to be quickly delivered to the competent administrative agency or nearby residents.

In the case of a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention, the water pipe network management system not only responds proactively to water quality accidents in the water pipe network that are difficult to respond to with existing water pipe network management methods, but also enables tracking management in the event of an accident. The network is implemented in a 3D virtual physical system environment to visualize it spatially. In addition, in the event of a water quality accident or disaster, nearby residents can quickly evacuate the extent of damage, and the competent administrative agency quickly predicts the extent of damage and establishes a response system. can be established.

Hereinafter, with reference to FIGS. 4 to 14, a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention will be described. Also, with reference to FIG. 15, a water pipe network management system according to an embodiment of the present invention will be described. We explain how to manage water pipe networks using emotional analysis and big data machine learning.

[Water pipe network management system using emotional analysis and big data machine learning]

Figure 4 is a schematic configuration diagram of a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention.

Referring to FIG. 4, the water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention is largely divided into a water pipe network management server 100, a water quality-related information provider 200, and a water quality-related consumer. It is comprised of a sentiment analysis server 300, a weather information provider 400, a geographic information provider 500, and a user terminal 600.

The water quality-related information provision unit 200 preprocesses data at each point to enable hierarchical spatial analysis of water quality-related information in the water pipe network, including water pipe network asset information, water quality complaint-related data, administrative data, and water quality-related data, and creates big data. Provided as (Big Data).

Here, big data is a technology that extracts value from data and analyzes the results, including large amounts of structured or even unstructured data sets that exceed the capabilities of existing database management tools. In other words, it refers to a huge amount of data that is difficult to process with existing databases. Therefore, in the case of a decision-making system that determines the priority of water quality key management points in the water pipe network and the optimal location of water quality sensors according to an embodiment of the present invention, big data in the water supply field is analyzed and utilized.

Specifically, the water quality-related information provision unit 200 includes a water pipe network asset information DB 210, an administrative data DB 220, a water quality data DB 230, and a water quality complaint DB 240. Specifically, the water quality information DB 200 includes The pipe network asset information DB 210 stores water pipe property information including pipe material, pipe use, pipe diameter, pipe extension length, and number of years of use, but is not limited to this.

The administrative data DB 220 stores administrative data related to population information, road information, and building information (including building durability) for each water supply block, but is not limited to this. Here, the water usage for each water supply block is calculated in the event of a water quality accident according to the population information for each water supply block stored in the administrative data DB 113, and thereafter, the water usage for each water supply block is mapped in the mapped service area. You can.

The water quality data DB 230 stores water quality-related data including points where water quality standards have been violated in the past and water quality measurement information, but is not limited to this.

The water quality complaint DB (240) stores information related to water quality complaints, including the water quality complaint reception date, type of water quality complaint, and location of water quality complaint, but is not limited to this.

Additionally, the water quality-related consumer emotional analysis server 300 provides consumer emotional analysis information related to tap water quality.

The weather information provider 400 provides real-time weather information of the area where the water pipe network is located.

The geographic information provider 500 provides geographic information about the area where the water pipe network is located.

The water pipe network management server 100 collects internal and external data, the weather information, and the geographical information provided from the water quality-related information provider 200 based on the water quality-related consumer emotional analysis information, and collects the water pipe network water quality. We derive management priority areas, support tracking management by considering connection nodes in the water pipe network system when a water quality accident occurs, calculate the damaged area and number of households, and support safety evaluation of water quality accidents in the water pipe network.

The user terminal 600 receives the analysis result of the water pipe network management server 100. In other words, the water pipe network management server 100 can quickly deliver information to the user terminal 600, for example, the competent administrative agency or nearby residents, before and after an accident occurs.

In the end, in the case of a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention, water quality civil complaint data regarding red water, green matter, foreign substances, small organisms such as larvae, etc. occurring in urban water pipe networks, etc. Social and social information including completion date, number of years of use, building use, and household information constructed from the water supply network's population, pipe deterioration, accident history such as water leakage and pipe damage, buried soil environment and road information, and building history information. Consumer emotional analysis and big data spatial information machine learning models are applied in an environment where environmental big data is implemented as a 3D virtual physical system.

Afterwards, the water pipe network management server 100 derives priority areas for water quality management in the water pipe network, supports tracking management by considering the connection nodes of the water pipe network system when a water quality accident occurs, calculates the damaged area and number of households, and calculates the water quality of the water pipe network. Support accident safety assessment. Accordingly, information can be quickly delivered to the user terminal 600, for example, to the competent administrative agency or nearby residents, before and after an accident occurs.

Meanwhile, Figure 5 is a detailed configuration diagram of a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention.

Referring to FIG. 5, in the water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention, the water pipe network management server 100 includes an input information setting unit 110 and a data collection unit 120. ), real-time weather information collection unit 130, spatial data processing unit 140, cyber-physics-based spatial information processing unit 150, data analysis unit 160, water supply management DB (170), water quality accident tracking management unit (180) and a safety evaluation management department (190).

The input information setting unit 110 presets input information including geographic coordinates according to the location of each measurement sensor installed in the water pipe network.

The data collection unit 120 collects water quality-related information from the water quality-related information providing unit 200 through a network, and collects consumer emotional analysis information from the water quality-related consumer emotional analysis server 300.

The real-time weather information collection unit 130 collects real-time weather information of the area where the water pipe network is located from the weather information provider 400. For example, the real-time weather information collection unit 130 collects real-time weather information by extracting weather information from a nearby point in the target facility or installing a weather measuring device.

The cyber-physical-based spatial information processing unit 150 matches the input information set by the input information setting unit 110 and the geographic information provided by the geographic information provider 500 and visualizes it.

The spatial data processing unit 140 collects the data and geographic information displayed based on cyber-physics in the cyber-physical spatial information processing unit 150 according to the consumer emotional analysis information collected from the water quality-related consumer emotional analysis server 300. Water quality-related information collected in unit 120 is matched to form spatial data.

Specifically, in order to provide spatial information, methods of visualizing and processing data must be considered. Geographic information used in this field of spatial information can be largely divided into vector data and raster data.

At this time, the vector data is provided in the form of a topographic map and can express points of interest (POI: points of interest), buildings, roads, etc. as data consisting of points, lines, and surfaces. When provided to the general public, the vector data is CAD (CAD). It is provided in DXF (Drawing exchange Format), which is a CAD) format data file structure, or as a shape file specialized for spatial information vector type file structure.

In addition, the raster data can be viewed as image data, and examples include aerial maps or satellite map data that are widely used in general map services. Such raster data is provided by pre-processing several sheets of data into a tile format. The vector data and raster data provided in this way are usually in a form that cannot be visualized directly on the web, so visualization processing is required to visualize them. Accordingly, the cyber-physical-based spatial information processing unit 150 matches the input information set by the input information setting unit 110 and the geographic information provided by the geographic information provider 500 and visualizes it.

The data analysis unit 160 analyzes the water supply status and civil complaints, water supply deterioration, and water quality management area according to the spatial data processed by the spatial data processing unit 140.

The water supply management DB 170 stores data processed by the spatial data processing unit 140 and data analyzed by the data analysis unit 160.

The water quality accident tracking and management unit 180 derives priority areas for water quality management in the water pipe network according to the analysis results of the data analysis unit 160, and tracks and manages water quality accidents by considering the connection nodes of the water pipe network system when a water quality accident occurs.

The safety evaluation management unit 190 manages the safety evaluation of the water pipe network according to the analysis results of the data analysis unit 160 and the water quality accident tracking of the water quality accident tracking management unit 180.

In the case of a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention, social network services (SNS) or administrative services are provided in an environment where city buildings and underground water pipe networks are implemented as a three-dimensional virtual physical system. Consumer emotional analysis information collected by acquiring and analyzing water pipe network complaint information submitted to local cafes or administrative autonomous organizations in autonomous districts is collected.

In addition, in addition to city building history information data, road environment and soil information data, underground water pipe network asset information data, and water pipe network corrosivity index, a database organized based on geographical coordinates such as population and number of households receiving water in the water supply system is used as a neural network for classification. By applying the model and boosting model, we classify each block into “Concern,” “Caution,” “Warning,” and “Danger” levels to derive priority areas for water quality management in the water pipe network.

In addition, when an actual water quality accident occurs, based on the connection nodes and operation and model data of the water pipe network, it visually shows which valves and pipes it starts from and spreads to, supports tracking and management of areas with abnormal water quality, and calculates the damaged area and number of households, etc. It can support the safety assessment of water quality accidents in water pipe networks. Afterwards, in the event of a water quality abnormality or water quality accident, information can be quickly delivered to the competent administrative agency or nearby residents.

In the case of a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention, when water of abnormal quality flows from the water intake source in the supply system line supplied from the water intake facility to the water purification facility in the drinking water supply system, the water is drained from the water purification facility. It is possible to identify the cause point in the system that extends to facilities and water supply facilities, and if the expected damage time and scale are collected in real time when the damage spreads, it is combined with public data. Spatial visualization can be implemented in real time on an open geographic information system by the water pipe network management system according to an embodiment of the present invention.

Meanwhile, various water quality abnormalities or water quality accidents related to red water, green matter, foreign substances, larvae, and other small organisms that occur in distribution water systems such as urban water pipe networks are not legal management items, and water quality abnormalities including emotional opinions by consumers are not subject to legal management. It takes a lot of time to confirm the actual facts, and there are limitations in responding to existing water quality accident analysis methods.

To explain the above-mentioned water quality-related data in more detail, water quality accidents in the water pipe network are caused not only by the inflow of external pollutants and foreign substances, but also by factors such as aging pipes, sediment at the bottom, and operational accidents.

Here, red water is more of an aesthetic problem than a health problem, but it is one of the most common causes of water accidents. Corrosion of metal pipes and plumbing fixtures increases the concentration of metal compounds in the water. Specifically, each metal undergoes a different corrosion process, but in general, low pH, high dissolved oxygen, high temperature, and high dissolved solids concentration can increase the corrosion rate.

For example, heavy metals such as lead and cadmium can also enter the water from plumbing materials. Metals can leach into water from pipes such as copper (household water pipes), iron (drainage pipes) and galvanized steel pipes, causing taste, odor and color problems in addition to minor health risks.

Additionally, rust water coming out of the faucet is one of the very sensitive issues that causes discomfort to tap water consumers. Recently, expectations for tap water quality have increased, and citizens tend to react more sensitively to the occurrence of rust than in the past.

Figure 6 is a diagram summarizing the data purification process used to evaluate water quality safety in the water pipe network and derive water quality management areas in the water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention.

As shown in Figure 6, the water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention largely creates individual layers through extraction or calculates specific attribute values using extracted data. Data is refined by extracting or merging to create individual layers.

Specifically, in the case of a water pipe network DB, asset groups and types can be classified by item and extracted and created by classifying each individual layer from a single layer.

At this time, rather than merging the entire attribute information of the water supply population layer and the water quality complaint layer of the supply system, only the necessary attribute values (e.g., population, number of complaints) are extracted, and an analysis tool is used to create a new layer with added layer attribute information. You can create individual feature layers.

Additionally, if necessary, new individual layers are additionally created by merging new attribute information into the resulting layer. In particular, because it is difficult to obtain water quality complaint data collected in autonomous districts or because the data is limited in number, consumer sensitivity exposed on social networks such as SNS, cafes, blogs, etc. is collected through web crawling, and the collected consumer sensitivity is collected. Use the diagram as analysis data.

In addition, data on pipe deterioration analyzed through various administrative district data, consumer sensitivity data, old building data according to the building register, road and soil environment data, and water pipe network data were synthesized to create a classification neural network model and self-organization map (Self-organization map). -Organizing Map) is applied to evaluate the water quality safety of the water pipe network and derive water quality management areas.

At this time, you can use multiple layers by overlapping them, but when you create a single layer with multiple attribute information, you can write the necessary queries or apply filters to visualize with a single layer, and you can use a variety of features rather than using multiple feature layers. Analysis and visualization tasks can be performed more easily.

In addition, because water quality abnormalities or water quality accidents that may occur in a water pipe network buried underground are not necessarily directly affected by the degree of deterioration of the buried pipes, there is no need to prioritize replacing buried pipes, which requires high costs.

Accordingly, the water pipe network management system, which is equipped with tracking management and safety evaluation management functions to respond to water quality abnormalities or water quality accidents that occur in the drainage system and water supply system, is a three-dimensional system with spatial coordinates for urban buildings and underground water pipe networks, etc. It runs in a spatial environment implemented with a virtual physical system.

In addition, various water quality abnormalities or water quality accidents related to small organisms such as red water, green matter, foreign substances, larvae, etc. are reported by consumers through social network services, through subjective opinions, or through local administrative autonomous district cafes or administrative agencies. It is based on civil complaint data.

In addition, in addition to the internal factors of various water quality abnormalities or water quality accidents and underground buried pipes, there are also human and social factors such as the age of the building, the external environment due to the road and soil environment, and the characteristics of the number of households, so these can be classified into past water quality occurrence areas or water quality abnormalities. When implementing a model for an area with a high concentration of consumer sensitivities, the evaluation factors can be simplified by reducing the explanatory variables by implementing a classification model and neural network model by dividing the buried pipe material characteristics.

In addition, the areas classified by the neural network model for classification are largely classified into “Concern,” “Caution,” “Warning,” and “Danger” levels for each block, thereby deriving priority areas for water quality management in the water pipe network. In addition, when an actual water quality accident occurs in an environment implemented with a virtual physical system, it visually shows which valves and pipes it starts from and spreads based on the connection nodes of the water pipe network system and operation and model data, providing a tracking management function for areas with abnormal water quality. It can be granted.

Meanwhile, Figure 7 is a diagram illustrating the classification of civil complaints that occurred on a specific day, the display of water pipe networks and old pipes in the water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention.

As shown in Figure 7, the water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention can be expressed as a classification of civil complaints that occurred on a specific day, a water pipe network, and an old pipe, and also spatial coordinates Buildings, underground water pipe networks, water quality complaints, and consumer sensitivities are implemented as a database, and this is a platform that displays them in a 3D virtual physical environment to predict the extent of damage and visualize the degree of spread in the event of water quality abnormalities or water quality accidents in the water pipe network. You can.

Meanwhile, Figure 8 is a diagram illustrating an urban water pipe network/water supply pipe in a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention.

In the case of a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention,

As shown in Figure 8, the water pipe network (water pipe/water supply pipe) within the city can be displayed,

For example, the water pipe network buried underground is broadly classified into metal and non-metallic pipes, including pipe type, pipe diameter, burial year, burial depth, corrosiveness index, road environment, soil environment, exterior coating, interior coating, and number of civil complaints per year. The evaluation is performed by applying weights to 10 variables according to the importance of the evaluation items.

Afterwards, this can be expanded into a deep learning model for classification to reduce the evaluation parameters for each major type.

In particular, the number of annual complaints can be calculated using data used to analyze consumer sensitivity in addition to water quality complaints filed with administrative autonomous districts.

The neural network algorithm for classification according to an embodiment of the present invention converts multidimensional data into two dimensions by applying Self-Organizing Map (SOM), a technique that simultaneously performs dimensionality reduction and clustering. By mapping, it is easy to visualize and easily understand the characteristics of complex water pipe-related data.

In addition, for information on building deterioration, road environment, soil environment, etc., a virtual physical environment is implemented using public open data, and spatial analysis is performed based on the geographical coordinates of the underground water pipe network to utilize model variables.

In addition, the tracking and management function for water quality accidents or water quality abnormalities configures a topology that physically connects all elements of the underground water pipe network (e.g. valves, pipes, various water quality/flow/water pressure sensors, etc.) Based on the operation data and model data of the pipe network, it provides a function to visually show which valve and pipe it starts from and spreads out.

By combining the implemented models as described above, we classify each block in the target area into “Concern,” “Caution,” “Warning,” and “Danger” levels to derive priority areas for water quality management in the water supply network.

Meanwhile, FIGS. 9A to 9C are diagrams for explaining consumer emotional analysis related to water quality in a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention, and FIG. 9A is a tap water emotional big data analysis service. shows a web screen implemented to provide, FIG. 9b is a configuration diagram of a water quality-related consumer emotional analysis server, and FIG. 9c is a diagram schematically illustrating the emotional analysis results.

In the case of a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention, the data collected by crawling social networks such as Mom Cafe and blogs operated in apartment complexes in the target administrative district is distributed to the relevant area. As shown in Figure 9a, a web service that provides a tap water emotional big data analysis service can be implemented so that it can be utilized as consumer sensitivity.

At this time, consumer sentiment analysis on water quality problems can be largely divided into machine learning approaches and dictionary-based approaches based on emotional word dictionaries. Machine learning methods generally involve large amounts of documents. Used for analysis, the data is pre-processed through tokenization, etc., and through learning, it is a method of performing emotional classification (class) such as positive and negative. Naive Bayes method and support vector machine are used as machine learning classifiers. It can be mainly used.

Specifically, the dictionary-based water quality-related consumer emotional analysis technique is based on a dictionary of emotional words, in which the researcher builds a dictionary of emotional words and finally classifies the sentiment through matching with the dictionary of emotional words for the object to be analyzed. It's a method. For example, this emotional analysis is largely performed through five steps, and the emotions of consumers, that is, citizens, regarding abnormal tap water quality are recognized, interpreted, and processing logic are also performed through the same steps.

Referring to FIG. 9b, the water quality-related consumer emotion analysis server 300 includes a data collection unit 310, a tokenization performance unit 320, an emotion analysis unit 330, an emotional index analysis unit 340, and consumer sensitivity analysis. It may include a unit 350 and a consumer emotional analysis DB 360.

The first step is the data collection step, where the data collection unit 310 collects and stores unstructured data related to tap water through the Internet, such as news, Internet cafes, blogs, and SNS.

The second step is the tokenization step, in which the tokenization performing unit 320 separates the collected data into morphemes, which are the minimum units with meaning, and tags them with parts of speech to define them. At this time, meaningless words such as articles and prepositions and symbols such as special characters are filtered out and excluded.

Through this, only necessary information is classified into morphemes and stored, and this series of processes is called tokenization.

The third step is a sentiment analysis step, in which the sentiment analysis unit 330 measures the positive and negative meanings of vocabulary using frequencies. For this purpose, a dictionary of emotional words is used.

The fourth step is the emotional index analysis step, in which the emotional index analysis unit 340 performs detailed emotional index analysis, such as positive/negative, through a predefined emotional word dictionary.

The fifth and final step is the consumer sensitivity analysis step, where the consumer sensitivity analysis unit 350 analyzes consumer sensitivity regarding abnormal tap water quality using the analyzed emotional index.

At this time, the consumer sentiment analysis DB (360) is equipped with a dictionary of emotional words to distinguish emotional words related to tap water, and collects water pipe network complaint information received from social network services, local cafes in administrative autonomous districts, or administrative governments. Acquire and store consumer emotional analysis results.

For example, as shown in Figure 9c, as a tool to analyze consumer sensitivity to tap water quality based on web crawling, as a result of analyzing the sentiment expressed on various social networks for water accidents that occurred in 2019, tap water-related As a result of the sentiment analysis, a total of 1,922 cases were mentioned, with a daily maximum of 52 and a daily minimum of 13, which showed a large daily variation and did not show a consistent pattern.

In addition, on daily average, 33.14 emotional words related to "tap water" were searched, and when analyzing the respective ratios including all positive, negative, and neutral language, they are analyzed as 48%, 29%, and 23%, with positive and negative words Considering only the respective proportions, it was analyzed that they accounted for 62% and 38%.

Meanwhile, Figure 10a is a diagram illustrating a neural network algorithm for classification in a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention, and Figure 10b shows the data processing sequence according to the neural network algorithm for classification. This is an operation flow chart.

In the case of a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention, as shown in Figure 10a, a classification neural network algorithm is used, and the data processing order according to the classification neural network algorithm is, As shown in Figure 10b, in order to run or terminate an algorithm in a self-organizing map, two values, the current input vector and the current number of iterations, must be maintained.

Specifically, the value of each element of the weight matrix is initialized to a random value greater than 0 and less than 1, and for the input vector and j nodes existing in the competitive layer, the distance between the input vector and the nodes ( ) to calculate the current input vector and After selecting the node in the competitive layer with the smallest value, the weight is modified. Afterwards, if the current input vector is the last input vector, move to the next process, and if the current input vector is not the last input vector, go back to the front of the flowchart and calculate the distance between the input vector and the competitive layer node. If the maximum number of repetitions is as defined in advance, application of the algorithm is terminated. Otherwise, the current input vector is set as the first input vector and the learning rate is modified.

For example, the distance between the input vector and the node ( ) is calculated by Equation 1.

[Equation 1]

here, represents the size of the input vector, represents the value of row i and column j in the weight table. The operation of selecting a node in the competitive layer and then modifying the weight of that node follows [Equation 2] below.

[Equation 2]

here, represents the new weight, represents the previous weight, is the learning rate, is a value determined when designing an algorithm, and the value gets smaller as the algorithm is repeated.

Meanwhile, Figure 11 is a detailed configuration diagram of the data analysis unit in the water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention, and Figure 12a shows the water supply status and civil complaints within the data analysis unit shown in Figure 11. This is a detailed configuration diagram of the analysis unit, Figure 12b is a specific configuration diagram of the water supply deterioration analysis unit, and Figure 12C is a specific configuration diagram of the water supply water quality management area analysis unit.

In the water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention, the data analysis unit 160, as shown in FIG. 11, includes the water supply status and civil complaints analysis unit 161, and the water supply deterioration degree. It includes an analysis unit 162 and a water supply water quality management area analysis unit 163.

As shown in FIG. 12A, the water supply status and civil complaints analysis unit 161, the data analysis unit 160,

The water supply status statistical analysis unit (161a) analyzes the distribution of water quality complaints according to internal and external data collected from the water quality information provision unit 200, analyzes the status of the water pipe network, and analyzes water usage and patterns by block. ); and a target area water supply status display unit 161b that displays the target area water supply status according to search conditions, and the internal data collected from the water quality-related information provision unit 200 is water usage status data. It includes water quality complaints and leakage complaint data, water pipe network attribute data, and water operation data, and external data collected from the water quality-related information provision unit 200 may be road name address data, soil drainage grade data, and population data by block. .

In addition, as shown in FIG. 12b, the water supply deterioration analysis unit 162 analyzes the distribution of water quality complaints according to internal data and external data collected from the water quality-related information providing unit 200, and determines the status of the water pipe network. a water supply status statistical analysis unit (162a) that analyzes water usage and patterns for each block; And a target area water supply deterioration mapping unit 162b that maps the target area water supply deterioration according to the analysis results of the water supply status statistical analysis unit 162a,

The internal data collected from the water quality information provision unit 200 includes water pipe network (water pipe/water supply pipe) specifics, ID, pipe material, pipe diameter, pipe extension length, burial depth, pipe use, consumer water quality complaints, and number of water leaks. , and water pipe network diagnosis results, and the external data collected from the water quality-related information providing unit 200 may be road name address data and soil drainage grade data scored by road type and section.

In addition, as shown in FIG. 12c, the water supply water quality management area analysis unit 163 analyzes and visualizes the management area according to internal data and external data collected from the water quality-related information providing unit 200. and visualization unit 163a; And a water quality intensive management area mapping unit (163b) that maps the water quality intensive management area according to the analysis results of the management area analysis and visualization unit (163a),

The internal data collected from the water quality information provision unit 200 includes specific matters regarding water usage, specific matters at the point where water quality complaints occur, specific matters at the point where water leakage complaints occur, and special matters regarding the water pipe network based on the results of the analysis of the degree of deterioration, and the water quality External data collected from the related information provision unit 200 may include road name address data, soil drainage grade data, and population data for each block.

Meanwhile, Figure 13 illustrates a tracking management function that enables the scale of damage and direction of spread when a water quality abnormality or water quality accident occurs for a target block in a water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention. It is a drawing.

The water pipe network management system using emotional analysis and big data machine learning according to an embodiment of the present invention determines the priority of the water quality management area based on the results graded according to the water quality safety evaluation of the water pipe network for the target block, As shown in Figure 13, when a water quality abnormality or water quality accident occurs for a target block, the extent of damage and direction of spread can be tracked and managed.

Ultimately, according to an embodiment of the present invention, web crawling-based consumer sentiment analysis and big data spatial information machine learning model are used in an environment where city buildings and underground water pipe networks are implemented as a 3D cyber physical system. By applying, it is possible to perform water quality accident tracking management and safety evaluation management of the urban water pipe network.

In addition, the water supply network can be spatially visualized by implementing it in a 3D cyber-physical system environment to enable proactive response to water quality accidents in the water pipe network as well as tracking and management in the event of an accident. Additionally, in the event of a water quality accident or disaster, nearby residents can monitor the extent of damage. can be evacuated quickly, and the competent administrative agency can quickly estimate the extent of damage and establish a response system.

[Water supply network management method using emotional analysis and big data machine learning]

Figure 14 is an operation flowchart showing a water pipe management method using emotional analysis and big data machine learning according to an embodiment of the present invention.

Referring to FIG. 14, the water pipe management method using emotional analysis and big data machine learning according to an embodiment of the present invention first presets input information for water pipe network management (S110).

Next, the water pipe network management server 100 collects real-time weather information from the weather information provider 400 (S120).

Next, the water pipe network management server 100 collects water quality data measured in real time by sensors installed in the water pipe network (S130).

Next, the water pipe network management server 100 collects and monitors internal data/external data from the water quality-related information provider 200 (S140).

Next, the water pipe network management server 100 processes 3D meteorological physical system-based spatial information according to the geographic information provided by the geographic information provider 500 using preset input information and real-time measured measurement information (S150) ).

Next, the water quality-related consumer emotional analysis server 300 generates consumer emotional analysis information according to emotional trends collected through web crawling (S160).

Next, the water pipe network management server 100 analyzes water pipe network-related data according to emotional analysis and big data machine learning (S170).

Next, the water pipe network management server 100 visually displays and delivers the results of water quality accident tracking management and safety evaluation management (S180).

In addition, when the water pipe network management server 100 generates an alarm according to water quality accident tracking management and safety evaluation management, it can deliver an alarm such as a text message to the user terminal 600 owned by the administrative government or nearby residents. .

In the end, according to the water pipe management method using emotional analysis and big data machine learning according to the embodiment of the present invention, consumers are concerned about red water, rust, foreign substances, small organisms such as larvae, etc. generated in the water pipe network supply system buried underground in the city. Completion date, number of years of use, building use, household information, etc. constructed from sensitivity and civil complaint data, water supply population of the water pipe network, pipe deterioration, accident history such as leakage and pipe damage, buried soil environment and road information, and building history information, etc. In an environment where social and environmental big data, etc., are implemented as a cyber-physical system, consumer emotional analysis and big data spatial information machine learning models are applied to derive priority areas for water quality management in the water pipe network.

Afterwards, in the event of a water quality accident, tracking and management is conducted by considering the connection nodes of the water pipe network system, so that information can be quickly delivered to the competent administrative agency or nearby residents in the event of a consumer sensibility outlier or water quality abnormality or water quality accident.

The description of the present invention described above is for illustrative purposes, and those skilled in the art will understand that the present invention can be easily modified into other specific forms without changing the technical idea or essential features of the present invention. will be. Therefore, the embodiments described above should be understood in all respects as illustrative and not restrictive. For example, each component described as single may be implemented in a distributed manner, and similarly, components described as distributed may also be implemented in a combined form.

The scope of the present invention is indicated by the claims described below rather than the detailed description above, and all changes or modified forms derived from the meaning and scope of the claims and their equivalent concepts should be construed as being included in the scope of the present invention. do.

100: Water pipe network management server
110: input information setting unit 120: data collection unit
130: Real-time weather information collection unit 140: Spatial data processing unit
150: Cyberphysics-based spatial information processing unit
160: Data analysis department
161: Water supply status and civil complaints analysis department 162: Water supply deterioration analysis department
163: Water supply water quality management area analysis department 161a: Water supply status statistical analysis department
161b: Water supply status display unit for target area 162a: Water supply status statistical analysis unit
162b: Target area water supply deterioration mapping unit 163a: Management area analysis and visualization unit
163b: Water quality intensive management area mapping department
170: Water supply management DB
180: Water Accident Tracking and Management Department
190: Safety Evaluation Management Department
200: Water quality information provision department
210: Water pipe network asset information DB 220: Administrative data DB
230: Water quality data DB 240: Water quality complaint DB
300: Water quality-related consumer emotional analysis server
310: Data collection unit 320: Tokenization performance unit
330: Emotional analysis unit 340: Emotional index analysis unit
350: Consumer sensitivity analysis unit 360: Consumer sentiment analysis DB
400: Weather information provider
500: Geographic information provider 600: User terminal

Claims (18)

Water quality-related information provided as big data by pre-processing data at each point to enable hierarchical spatial analysis of water quality-related information in the water pipe network, including water pipe network asset information, water quality complaint-related data, administrative data, and water quality-related data. Information provision department (200);
A water quality-related consumer sentiment analysis server 300 that obtains complaints related to tap water quality through web crawling and provides consumer sentiment analysis information related to tap water quality;
A weather information provider (400) that provides real-time weather information of the area where the water pipe network is located;
Geographic information provider 500 that provides geographic information about the area where the water pipe network is located; and
Based on the water quality-related consumer emotional analysis information, internal data and external data provided from the water quality-related information provision unit 200, the weather information, and the geographical information are collected to derive water quality management priority areas for the water pipe network, and to derive water quality management priority areas. When an accident occurs, it supports tracking management considering the connection nodes of the water pipe network system to calculate the damaged area and number of households, and emotional analysis and big data machine learning including the water pipe network management server (100) that supports water quality accident safety evaluation in the water pipe network. A water pipe network management system using . The method of claim 1, wherein the water quality-related information providing unit 200,
A water pipe network asset information DB (210) that stores water pipe attribute information including pipe material, pipe use, pipe diameter, pipe extension length, and number of years of use;
Administrative data DB (220), which stores administrative data related to population information, road information, and building information for each water supply block;
Water quality data DB (230), which stores water quality-related data including points where water quality standards have been violated in the past and water quality measurement information; and
A water pipe network management system using emotional analysis and big data machine learning, including a water quality complaint DB (240) that stores water quality complaint information, including water quality complaint reception date, water quality complaint type, and water quality complaint occurrence point. The method of claim 1, wherein the water pipe network management server 100,
An input information setting unit 110 that pre-sets input information including geographic coordinates according to the locations of each measurement sensor installed in the water pipe network;
a data collection unit 120 that collects water quality-related information from the water quality-related information providing unit 200 through a network and collects consumer emotional analysis information from the water quality-related consumer emotional analysis server 300;
A real-time weather information collection unit 130 that collects real-time weather information of the area where the water pipe network is located from the weather information provider 400;
a cyber-physics-based spatial information processing unit 150 that matches and visualizes the input information set by the input information setting unit 110 displayed in a cyber-physics-based manner and the geographic information provided by the geographic information provider 500;
According to the consumer emotional analysis information collected from the water quality-related consumer emotional analysis server 300, the geographic information displayed as cyberphysical-based in the cyberphysical-based spatial information processing unit 150 and the data collected by the data collection unit 120 are displayed. A spatial data processing unit 140 that forms spatial data by matching water quality-related information;
a data analysis unit 160 that analyzes water supply status and civil complaints, water supply deterioration, and water quality management areas according to the spatial data processed by the spatial data processing unit 140;
a water quality accident tracking unit 180 that tracks and manages water quality accidents according to the analysis results of the data analysis unit 160; and
Sentiment analysis and big data machine learning, including a safety evaluation management unit 190 that manages the safety evaluation of the water pipe network according to the analysis results of the data analysis unit 160 and the water quality accident tracking of the water quality accident tracking management unit 180. Water pipe network management system used. According to paragraph 3,
The water quality accident tracking and management unit 180 classifies the areas classified by the classification neural network model into “concern,” “caution,” “warning,” and “danger” levels for each block, and derives priority areas for water quality management in the water pipe network. In an environment implemented with a cyber-physical system, when an actual water quality accident occurs, it is possible to track and manage areas with abnormal water quality by visually showing which valves and pipes it starts from and spreads based on the connection nodes and operation and model data of the water pipe network system. A water pipe network management system using emotional analysis and big data machine learning. The method of claim 3, wherein the data analysis unit 160,
a water supply status and civil complaints analysis unit (161) that analyzes the water supply status and civil complaints according to internal and external data collected from the water quality-related information provision unit (200);
A water supply deterioration analysis unit 162 that analyzes the deterioration of water supply according to internal and external data collected from the water quality information provision unit 200; and
Water supply network management using emotional analysis and big data machine learning, including a water supply water quality management area analysis unit 163 that analyzes the water supply water quality management area according to internal and external data collected from the water quality information provision unit 200. system. According to claim 5, the water supply status and civil complaint analysis unit 161,
The water supply status statistical analysis unit (161a) analyzes the distribution of water quality complaints according to internal and external data collected from the water quality information provision unit 200, analyzes the status of the water pipe network, and analyzes water usage and patterns by block. ); and
Includes a target area water supply status display unit (161b) that displays the target area water supply status by search condition,
The internal data collected from the water quality information provision unit 200 is water usage status data. It includes water quality complaints and leakage complaint data, water pipe network attribute data, and water operation data, and external data collected from the water quality-related information provision unit 200 includes road name address data, soil drainage grade data, and population data by block. A water pipe network management system using emotional analysis and big data machine learning. The method of claim 5, wherein the water supply deterioration analysis unit 162,
The water supply status statistical analysis unit (162a) analyzes the distribution of water quality complaints according to internal and external data collected from the water quality information provision unit 200, analyzes the status of the water pipe network, and analyzes water usage and patterns by block. ); and
It includes a target area water supply deterioration mapping unit 162b that maps the target area water supply deterioration according to the analysis results of the water supply status statistical analysis unit 162a,
The internal data collected from the water quality information provision unit 200 includes water pipe network (water pipe/water supply pipe) specifics, ID, pipe material, pipe diameter, pipe extension length, burial depth, pipe use, consumer water quality complaints, and number of water leaks. , and water pipe network diagnosis results, and the external data collected from the water quality-related information provision unit 200 is emotional analysis and big data, characterized in that road name address data and soil drainage grade data scored by road type and section. Water pipe network management system using machine learning. The method of claim 5, wherein the water supply water quality management area analysis unit 163,
a management area analysis and visualization unit (163a) that analyzes and visualizes the management area according to internal and external data collected from the water quality-related information provision unit (200); and
It includes a water quality intensive management area mapping unit (163b) that maps the water quality intensive management area according to the analysis results of the management area analysis and visualization unit (163a),
The internal data collected from the water quality information provision unit 200 includes specific matters regarding water usage, specific matters at the point where water quality complaints occur, specific matters at the point where water leakage complaints occur, and special matters regarding the water pipe network based on the results of the analysis of the degree of deterioration, and the water quality External data collected from the related information provision unit 200 is a water pipe network management system using emotional analysis and big data machine learning, characterized in that it includes road name address data, soil drainage grade data, and population data for each block. The method of claim 1, wherein the water quality-related consumer emotional analysis server 300,
A data collection unit 310 that collects and stores unstructured data related to tap water through the Internet, such as news, Internet cafes, blogs, and SNS;
The collected data is separated into morphemes, which are the minimum units with meaning, and defined by tagging them with parts of speech. Meaningless vocabulary and symbols such as special characters are filtered out and tokenized to classify and store only necessary information into morphemes. execution department 320;
A sentiment analysis unit 330 that measures the positive and negative meanings of vocabulary using a dictionary of emotional words using frequencies;
An emotional index analysis unit 340 that performs detailed emotional index analysis such as positive/negative through a predefined emotional word dictionary; and
A water pipe network management system using emotional analysis and big data machine learning, including a consumer sensitivity analysis unit 350 that analyzes consumer sensitivity regarding abnormal tap water quality using the analyzed emotional index. According to clause 9,
It is equipped with a dictionary of emotional words to distinguish emotional words related to tap water, and stores the results of consumer emotional analysis by acquiring information on water pipe network complaints received from social network services, local cafes in administrative autonomous districts, or administrative governments. A water pipe network management system using emotional analysis and big data machine learning that additionally includes an analysis DB (360). a) Pre-setting input information for water pipe network management;
b) the water pipe network management server 100 collecting real-time weather information from the weather information provider 400;
c) the water pipe network management server 100 collecting water quality data measured in real time by a sensor installed in the water pipe network;
d) the water pipe network management server 100 collecting and monitoring internal data/external data from the water quality-related information providing unit 200;
e) the water pipe network management server 100 processing 3D meteorological physical system-based spatial information based on geographic information provided by a geographic information provider 500 using preset input information and real-time measured measurement information;
f) the water quality-related consumer emotional analysis server 300 generating consumer emotional analysis information according to emotional trends collected through web crawling;
g) the water pipe network management server 100 analyzing water pipe network-related data according to emotional analysis and big data machine learning; and
h) The water pipe network management server includes a step of visualizing and delivering the results of water quality accident tracking management and safety evaluation management,
The water pipe network management server 100 collects internal and external data, the weather information, and the geographic information provided from the water quality-related information provider 200 based on the water quality-related consumer emotional analysis information, and collects the water pipe network Sentiment analysis and big data are used to derive water quality management priority areas, support tracking management by considering connection nodes in the water pipe network system when a water quality accident occurs, calculate the damaged area and number of households, and support water quality accident safety evaluation in the water pipe network. Water pipe network management method using data machine learning. According to clause 11,
In step h) above, the areas classified by the neural network model for classification are classified into "Concern", "Caution", "Warning" and "Danger" levels for each block, and priority areas for water quality management in the water pipe network are derived, and the cyber-physical system Sentiment analysis, which is characterized by tracking and managing areas with abnormal water quality by visually showing which valves and pipes it starts from and spreads based on the connection nodes of the water pipe network system and operation and model data when an actual water quality accident occurs in an environment implemented by and water pipe network management method using big data machine learning. According to clause 11,
The water quality-related information provision unit 200 in step d) above provides data for each point to enable hierarchical spatial analysis of the water quality-related information of the water pipe network, including water pipe network asset information, water quality complaint-related data, administrative data, and water quality-related data. A water pipe network management method using emotional analysis and big data machine learning, characterized by pre-processing and providing it as big data. The method of claim 13, wherein the water quality-related information providing unit 200,
A water pipe network asset information DB (210) that stores water pipe attribute information including pipe material, pipe use, pipe diameter, pipe extension length, and number of years of use;
Administrative data DB (220), which stores administrative data related to population information, road information, and building information for each water supply block;
Water quality data DB (230), which stores water quality-related data including points where water quality standards have been violated in the past and water quality measurement information; and
A water pipe network management method using emotional analysis and big data machine learning, including a water quality complaint DB (240) that stores water quality complaint information, including water quality complaint reception date, water quality complaint type, and water quality complaint occurrence point. The method of claim 11, wherein the water pipe network management server 100,
An input information setting unit 110 that pre-sets input information including geographic coordinates according to the locations of each measurement sensor installed in the water pipe network;
a data collection unit 120 that collects water quality-related information from the water quality-related information providing unit 200 through a network and collects consumer emotional analysis information from the water quality-related consumer emotional analysis server 300;
A real-time weather information collection unit 130 that collects real-time weather information of the area where the water pipe network is located from the weather information provider 400;
a cyber-physics-based spatial information processing unit 150 that matches and visualizes the input information set by the input information setting unit 110 displayed in a cyber-physics-based manner and the geographic information provided by the geographic information provider 500;
According to the consumer emotional analysis information collected from the water quality-related consumer emotional analysis server 300, the geographic information displayed as cyberphysical-based in the cyberphysical-based spatial information processing unit 150 and the data collected by the data collection unit 120 are displayed. A spatial data processing unit 140 that forms spatial data by matching water quality-related information;
a data analysis unit 160 that analyzes water supply status and civil complaints, water supply deterioration, and water quality management areas according to the spatial data processed by the spatial data processing unit 140;
a water quality accident tracking management unit 180 that tracks and manages water quality accidents according to the analysis results of the data analysis unit 160; and
Sentiment analysis and big data machine learning including a safety evaluation management unit 190 that manages the safety evaluation of the water pipe network according to the analysis results of the data analysis unit 160 and the water quality accident tracking of the water quality accident tracking management unit 180. Water pipe network management method used. The method of claim 15, wherein the data analysis unit 160,
a water supply status and civil complaints analysis unit (161) that analyzes the water supply status and civil complaints according to internal and external data collected from the water quality-related information provision unit (200);
A water supply deterioration analysis unit 162 that analyzes the deterioration of water supply according to internal and external data collected from the water quality information provision unit 200; and
Water supply network management using emotional analysis and big data machine learning, including a water supply water quality management area analysis unit 163 that analyzes the water supply water quality management area according to internal and external data collected from the water quality information provision unit 200. method. The method of claim 11, wherein the water quality-related consumer emotional analysis server 300,
A data collection unit 310 that collects and stores unstructured data related to tap water through the Internet, such as news, Internet cafes, blogs, and SNS;
The collected data is separated into morphemes, which are the minimum units with meaning, and defined by tagging them with parts of speech. Meaningless vocabulary and symbols such as special characters are filtered out and tokenized to classify and store only necessary information into morphemes. execution department 320;
A sentiment analysis unit 330 that measures the positive and negative meanings of vocabulary using a dictionary of emotional words using frequencies;
An emotional index analysis unit 340 that performs detailed emotional index analysis such as positive/negative through a predefined emotional word dictionary; and
A water pipe network management method using emotional analysis and big data machine learning, including a consumer sensitivity analysis unit 350 that analyzes consumer sensitivity regarding abnormal tap water quality using the analyzed emotional index. According to clause 17,
It is equipped with a dictionary of emotional words to distinguish emotional words related to tap water, and stores the results of consumer emotional analysis by acquiring information on water pipe network complaints received from social network services, local cafes in administrative autonomous districts, or administrative governments. Water pipe network management method using emotional analysis including analysis DB (360) and big data machine learning.