ARTIFICIAL INTELLIGENCE AND IoT-BASED EARLY WATER INFRASTRUCTURE LEAK DETECTION AND INTERVENTION SYSTEM AND METHOD

TR202615551A2Pending Publication Date: 2026-09-21T C USKUDAR UNIVERSITESI
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
TR202615551
Authority / Receiving Office
TR · TR
Patent Type
Applications
Current Assignee / Owner
Filing Date
2026-09-10
Publication Date
2026-09-21

Smart Images

  • Figure 00000033_0000
    Figure 00000033_0000
Patent Text Reader

Abstract

The invention relates to an AI and IoT-based water infrastructure leak detection and response system (1) and method for early detection of leaks and pipeline anomalies occurring in water distribution networks, determining the location of the leak in the pipeline and automatically isolating the relevant pipeline in critical situations; the system includes IoT sensor nodes (1.1) that obtain vibration and / or pressure, flow and temperature measurement data related to the pipeline, a central processing unit (1.2) and a valve control unit (1.3); the central processing unit (1.2) has a live hydraulic digital twin unit (1.2.2) that generates dynamic hydraulic reference data instead of fixed thresholds using the network model and current measurement data, and the features extracted from the measurement data are combined with the said hydraulic reference data by the AI-based anomaly analysis unit (1.2.3) By combining and classifying the leaks, the location of the leak is determined by the leak positioning unit (1.2.4) by evaluating the difference in arrival time of the sensor signals and the damping characteristics together, and in case of critical pipe burst and / or high volume leak detection, the relevant pipeline is automatically isolated via the valve control device (1.3) with the valve control command generated by the intervention decision and control unit (1.2.5).
Need to check novelty before this filing date? Find Prior Art

Description

1 TARIFF ARTIFICIAL INTELLIGENCE AND IoT-BASED EARLY WATER INFRASTRUCTURE LEAK DETECTION AND INTERVENTION SYSTEM AND METHOD Technical Area 5 The invention relates to the technical field of monitoring and management of water distribution networks, early detection of leaks and pipeline anomalies, especially in water networks detection, determination of the leak location, and adjustment of water flow according to the detected anomaly. Sensors for control, hydraulic modeling, artificial intelligence and automatic valve 10 water infrastructure leak detection and intervention system and working method based on control. It is related. State of the Art Detecting and locating leaks in water distribution networks. The known techniques for this include various sensor-based monitoring, hydraulic modeling, and anomaly detection. There are systems in place. In these systems, acoustic data obtained from the pipeline, Evaluation of vibration, pressure and / or flow data with hydraulic model results. Comparison of actual field measurements and leakage or 20 based on the data obtained Different technical approaches are used, such as identifying network anomalies. Patent document number US2024035630A1, included in the known art, describes a liquid pipeline. Obtaining acoustic and / or vibration measurement data from multiple sensors placed in the network to convert the data in question from the time domain to the frequency domain and obtain 25 By evaluating the obtained frequency values, the leakage situation and the leakage It describes a system and method for determining the location of the source. The subject of the document is the comparison of measurement values ​​from different sensors, leakage. by identifying the sensors closest to the source and the relative geographic locations of the sensors. It is also explained how to estimate the location of the leak using this data. Furthermore, measurement 30 inaccuracies in the data caused by traffic, rain, and similar environmental noises. Data processing and threshold calibration aimed at reducing positive results. is being carried out. 2 However, the document in question includes pressure, flow, and temperature data along with vibration data. Integrated assessment of the current hydraulic behavior of the water distribution network. Generating dynamic reference data from a real-time hydraulic digital twin representing and the reference data in question, along with sensor data, are analyzed using AI-based anomalies. Its use in classification is not explained. Furthermore, the document in question mentions leak 5. its position is determined by the time-to-arrival difference (TDOA) of signals received from neighboring sensor nodes. Determining this by evaluating the signal attenuation characteristics together is a critical step. When an anomaly is detected, the relevant valve is automatically controlled, and the pipeline is monitored. There is no closed-loop technical process involving physical isolation. The patent document numbered US11494853B2, which is included in the known technology, is for a water distribution system. a method for determining the amount of water leakage from different leakage areas in the system The method is explained. In this method, hydraulic modeling and real field measurements are used. They are evaluated together and the spatial distribution of water leakage amounts is optimized. Leakage amounts are determined for different areas. In this approach, hydraulic 15 The model is used in conjunction with the difference between the monitored and simulated pressure values, and Leakage amounts in different virtual spaces through optimization based on genetic algorithms The aim is to determine this. However, the document states that the hydraulic model is based on data from multiple physical sensors, 20 Dynamic reference to currently synchronized and AI-based anomaly analysis. It is not explained whether it is used as a live hydraulic digital twin providing the data. Furthermore... The document in question contains data obtained from vibration, pressure, flow, and temperature. Attributes are combined with hydraulic digital twin reference data to create Random Forest and Classification using a batch learning architecture that utilizes XGBoost algorithms together; 25 point by point based on the time-to-arrival difference of the leak location and signal attenuation characteristics. automatic physical detection via valve actuator after identification and critical leak detection. There is no technical process in place for carrying out an intervention. Patent document number WO2015087378A1, included in the known art, describes 30 in a water treatment plant network. using pressure measurement data obtained from numerous pressure gauges a device and method for detecting network anomalies The document explains the pressure difference between different pairs of pressure gauges. Creating a water consumption rate index between sensors from the ratio of their differences, 3 anomalies can be identified by evaluating the change in the index over time. The process involves identifying the anomaly and estimating its probable location. In addition, some applications use GIS to integrate business, asset, customer, and external data with historical data. inclusion of series in the evaluation and use of machine learning techniques It is anticipated. 5 However, the document in question mentions different physical factors including vibration, pressure, flow, and temperature. obtaining the parameters together via IoT sensor nodes and realizing this data combined with dynamic reference data generated by a real-time hydraulic digital twin. It is not explained. Also, the Random Forest and XGBoost algorithms together have 10 The ensemble learning architecture used combines TDOA and signal fading features. Valve control is performed as a result of point leak location and critical leak detection. The pipeline is physically isolated by automatically activating the device. The document in question describes a closed-loop intervention structure that enables this. It is not explained. 15 Therefore, detecting physical water losses in traditional water distribution networks is largely... It relies on manual methods of intervention after the event has occurred. Known These approaches in technology bring with them the following technical problems and shortcomings: It brings: 20 - Delayed and Inadequate Early Detection: Existing acoustic listening or ground radar (GPR) The methods involve stopping the leak only after saturating the soil around the pipe and bringing it to the surface. after it comes out or causes large-scale collapses It can be used for microleaks (physical cracks) in their initial stages. It cannot be detected by these methods. 25 - High False Alarm (False Positive) Rate: Only detects pressure or flow rate changes. Simple threshold-based SCADA systems monitor normal consumption peaks on the network, Pump activation / deactivation moments or sudden increases in consumption can be interpreted as leaks. In terms of interpretation, this leads to false alarms in 45-60% of cases. This situation causes field teams to waste time and incur unnecessary excavation costs. 30 is happening. - Lack of Positioning and Accuracy: Existing macro-scale regional measurement areas (DMA - District Metered Area) methods indicate that a leak is located in a specific area. 4 They can say, but not exactly which of the kilometers-long pipelines It cannot indicate a fault at that point (in meters). - Environmental Noise Sensitivity: In acoustic-based systems, traffic in urban areas, construction, industrial vibrations or pipe flow noises are highly prevalent. The signal is distorted, and incorrect analysis results are produced. 5 - Lack of Automated and Real-Time Response: Known solutions are often ineffective. It focuses solely on the "detection and reporting" function. In the event of a leak or pipe burst, A real-time, automated shutdown to prevent water waste and infrastructure damage. They do not work in an integrated manner with the mechanism. In summary, when the solutions in question within the known technology are examined, sensor-based leak detection is identified. detection and positioning, comparison of hydraulic model with field measurements and sensor Techniques such as anomaly detection through data appear to be known individually. However, in the known technique, vibration, pressure, flow, and temperature measurement data are combined. By obtaining and pre-processing the measurement data, the said measurement data is used in the water distribution network. 15 generated by a live hydraulic digital twin representing current hydraulic behavior Combining dynamic reference data, combining the combined data into a Random Forest and Analysis using a batch learning architecture where XGBoost algorithms are run in parallel. By determining the type of anomaly, the difference in arrival time of the detected leak and the signal By evaluating the damping characteristics together, 20 on the pipeline locating and detecting a critical level of leak or pipe burst In this case, the relevant valve is automatically controlled and the pipeline is physically isolated. a closed-loop technical structure in which the processes are carried out in a connected manner It is not available. Brief Description of the Invention The purpose of the invention is to relate to the technical field of monitoring and management of water distribution networks. and, in particular, early detection of leaks and pipeline anomalies occurring in water networks. detection, determination of the leak location, and adjustment of water flow according to the detected anomaly. Sensors, hydraulic modeling, artificial intelligence, and automatic valves for control purposes. 30 Control-based water infrastructure leak detection and response system and operating method. to accomplish. Explanation of the Figures Figure 1. General block of the water infrastructure leak detection and intervention system that is the subject of the invention. It is a diagram. Explanation of References in Figures To better understand the invention, the numbers in the figures correspond to the following: 5 1. Water Infrastructure Leak Detection and Response System 1.1. IoT sensor node 1.1.1. Mechanical parameter sensor 1.1.2. Flow sensor 10 1.1.3. Temperature sensor 1.1.4. Microcontroller 1.1.5. Signal preprocessing unit 1.1.6. Wireless communication unit 1.2. Central Processing Unit 1.2.1. Data receiving unit 1.2.2. Live hydraulic digital twin unit 1.2.2.1. Network modeling unit 1.2.2.2. Hydraulic simulation unit 20 1.2.2.3. Reference data production unit 1.2.3. Artificial intelligence-based anomaly analysis unit 1.2.3.1. Feature extraction unit 1.2.3.2. Data Fusion Unit 1.2.3.3. Collective learning classification unit 25 1.2.3.4. Anomaly scoring unit 1.2.4. Leak location unit 1.2.4.1. Arrival time difference determination unit 1.2.4.2. Signal attenuation analysis unit 1.2.4.3. Location determination unit 30 1.2.5. Intervention decision and control unit 1.2.6. Data communication unit 6 1.3. Valve control device 1.3.1. Communication unit 1.3.2. Valve actuator 1.3.3. Valve status determination unit 1.3.4. Actuator control unit 5 1.4. User device 1.4.1. Communication unit 1.4.2. User interface 1.4.3. Geographic visualization unit 10 1.4.4. Alarm notification unit 1.4.5. Work order and navigation unit 1.5. Data storage unit 1.5.1. Sensor database 15 1.5.2. Network database 1.5.3. Anomaly registration unit Detailed Description of the Invention Leaks and pipeline anomalies occurring in water distribution networks, vibrations, Representing the current hydraulic behavior of the water distribution network using pressure, flow, and temperature data. identification by evaluating the digital twin reference data together, Determining the location of the detected leak in the pipeline and, according to the identified anomaly The invention concerns water infrastructure 25, which enables the automatic control of the relevant valve. Leak detection and intervention system (1); measurement from a pipeline of a water distribution network At least one IoT sensor node (1.1) that enables the acquisition of data, the IoT in question By processing the measurement data obtained from the sensor node (1.1), the water distribution network the hydraulic status is determined by the measurement data and hydraulic reference data. By analyzing them together, they detect anomalies in the network and locate the detected leak. 30 at least one central processing device (1.2) that specifies the central processing device (1.2) a valve in the water distribution network according to the control command generated by at least one valve control device (1.3) that changes its status and the central processing device in question 7 (1.2) provides the user with leak, anomaly and network status information generated by. It includes at least one user device (1.4) that enables its presentation. System (1), the IoT sensor node (1.1), central processing device (1.2), valve control 5 Transmission of data and / or control commands between the device (1.3) and the user device (1.4) It includes at least one communication network that provides The system (1) uses the measurement data obtained from the said IoT sensor node (1.1) to process the water data relating to the distribution network and data relating to the detected anomalies It contains at least one data storage unit (1.5) that enables its storage. 10 The IoT sensor node in question (1.1) detects vibrations in the pipeline of the water distribution network. and / or at least one that enables the sensing of mechanical parameters representing pressure. mechanical parameter sensor (1.1.1), flow velocity and / or flow rate in the pipeline in question at least one flow sensor (1.1.2) that enables the detection of the pipeline and / or pipe 15 at least one temperature sensor that enables the detection of the temperature of the fluid passing through the pipeline (1.1.3), receiving and processing the measurement data obtained from the said sensors at least one microcontroller (1.1.4) that provides the noise in the measurement data in question at least one signal that enables the reduction and preparation of measurement data for analysis. preprocessing unit (1.1.5) and central processing of the preprocessed measurement data 20 at least one wireless communication unit (1.1.6) that enables transmission to the device (1.2) It includes. The mechanical parameter sensor (1.1.1) detects vibrations and pressures occurring in the pipeline. to enable the detection of changes, at least one accelerometer and at least one piezoelectric 25 It includes a pressure sensor. This accelerometer detects structural events occurring in the pipeline. It generates vibration data by detecting vibrations, and this is the piezoelectric pressure in question. The sensor detects static and dynamic pressure changes in the pipeline and provides pressure data. It constitutes. In one application of the invention, the IoT sensor node (1.1) belongs to the water distribution network. externally and / or internally to the pipeline wall and / or valve chamber It can be installed. The said IoT sensor node (1.1) is resistant to water and moisture ingress. 8 in an enclosure with IP68 protection rating to provide protection It can be positioned. In one application of the invention, the flow sensor (1.1.2) measures the fluid passing through the pipeline. It is an ultrasonic flow sensor that determines the flow rate and / or flow rate. 5 The microcontroller mentioned (1.1.4); the mechanical parameter sensor (1.1.1), flow at least one sensor (1.1.2) and temperature sensor (1.1.3) that receives the measurement data from the data input unit, at least one processor that processes the measurement data in question, and the processed 10 It includes at least one memory unit that stores measurement data. The signal preprocessing unit in question (1.1.5) processes the measurement data obtained from the sensors. In order to ensure the reduction of environmental noise, high levels of these measurement data are used. It applies pass-through and / or low-pass filtering. 15 The wireless communication unit in question (1.1.6) processes the pre-processed measurement data. transmitting wirelessly to the central processing device (1.2) via the communication network This wireless communication unit (1.1.6) provides the central measurement data. LoRaWAN and / or NB-IoT communication 20 to enable transmission to the processing device (1.2). It uses the protocol. The wireless communication unit in question (1.1.6) is pre-processed. Measurement data is collected at specified time intervals and / or based on anomaly detection. It transmits the trigger to the central processing unit (1.2). The specific invention In the application, the measurement data in question is related to the instantaneous data transmission based on anomaly detection. If not, 25 to the central processing unit (1.2) at 30-second intervals. is transmitted. In one application of the invention, the IoT sensor node (1.1) and the central processing device Measurement data between (1.2) are protected against unauthorized access and / or data modification. For protection purposes, the measurement data in question is transmitted in encrypted form. The data in question... TLS 1.3 communication security protocol and / or AES-256 encryption algorithm in transmission It is available. 30 The central processing unit (1.2) processes the measurement data and water transmitted from the IoT sensor node (1.1). At least one data acquisition unit (1.2.1) that receives data relating to the distribution network, the measurement in question data and network data, the expected hydraulic behavior of the water distribution network. 9 at least one live hydraulic digital twin unit (1.2.2) that determines the measurement data in question. Reference data generated by the live hydraulic digital twin unit (1.2.2) are analyzed together. at least one AI-based anomaly analysis unit that identifies anomalies in the network (1.2.3), at least one leak that identifies the location of the detected leak in the pipeline The positioning unit (1.2.4) controls valve 5 according to the type and / or level of the identified anomaly. At least one intervention decision and which constitutes the control command to be sent to the control device (1.3) control unit (1.2.5) and the other electronic components of the central processing unit (1.2) of the system at least one data communication unit that enables the exchange of data and control commands with its components It includes (1.2.6). The data receiving unit in question (1.2.1) receives the measurement data transmitted from the IoT sensor nodes (1.1). with water distribution network pipe diameter, pipe material, pipe coordinates, network age and It enables the acquisition of topographic elevation data into the central processing unit (1.2). Live hydraulic digital twin unit (1.2.2); 15 representing the physical structure of the water distribution network at least one network modeling unit (1.2.2.1) that enables the creation of a network model, Using the network model in question and current measurement data, the water distribution network at least one hydraulic simulation unit (1.2.2.2) that calculates the hydraulic behavior and the calculated a reference point to be compared with the hydraulic parameters measured from the hydraulic behavior in question. It contains at least one reference data generation unit (1.2.2.3) that makes up the data. 20 The network modeling unit in question (1.2.2.1) measures the pipe diameter and pipe size of the water distribution network. using material, pipe coordinate, network age and topographic elevation data, the statement the subject is the digital grid model representing the physical structure of the water distribution network It constitutes. 25 The hydraulic simulation unit (1.2.2.2) and the network modeling unit (1.2.2.1) are mentioned. with the aforementioned digital grid model created by IoT sensor nodes (1.1) By processing the latest measurement data obtained, the expected pressure in the water distribution network is determined. It calculates the flow distribution. In one application of the invention, the hydraulic system in question 30 Simulation unit (1.2.2.2), modeling of the hydraulic behavior of the water distribution network and an EPANET-based hydraulic simulation algorithm for the purpose of simulation It uses the hydraulic simulation algorithm in question, developed by the network modeling unit. Pipe diameter, pipe material, pipe coordinate, network age and as determined by (1.2.2.1) Current measurements obtained from IoT sensor nodes (1.1) with topographic elevation data It determines the expected pressure and flow distribution of the network using the data. The reference data production unit (1.2.2.3) and the hydraulic simulation unit (1.2.2.2) are mentioned. From the expected pressure and flow distribution calculated by IoT sensor 5 hydraulic reference to be compared with the measurement data obtained from nodes (1.1) It generates its data. In this way, the aforementioned reference data production unit (1.2.2.3), a predetermined fixed lower and / or upper limit to be used in anomaly assessment Instead of a threshold value, it is dynamically adjusted according to the current hydraulic condition of the water distribution network. It establishes the defined hydraulic reference data. The dynamic hydraulic reference in question is 10. data includes network aging, seasonal consumption variations, and pump operating status. and / or leaks due to hydraulic changes resulting from normal consumption variations. It is used to distinguish changes resulting from pipeline anomalies. AI-based anomaly analysis unit (1.2.3); 15 obtained from IoT sensor nodes (1.1) at least one feature extraction unit that identifies features for analysis from measurement data. (1.2.3.1), with the aforementioned attributes, by the live hydraulic digital twin unit (1.2.2) at least one data fusion unit (1.2.3.2) combining the generated hydraulic reference data, by processing the combined data with multiple machine learning models at least one ensemble learning classification unit (1.2.3.3) that classifies the state of the network and 20 at least one that creates an anomaly score for the classified network status in question It includes anomaly scoring units (1.2.3.4). The feature extraction unit in question (1.2.3.1) applies a fast Fourier transform to the measurement data. (FFT) is applied and the signal energy and crest factor (crest 25) are obtained from the measurement data. It obtains the attributes of factor), dominant frequency, and pressure drop rate. The feature extraction unit (1.2.3.1) derives its characteristics from the normal operating behavior in the measurement data. Z-score and / or median absolute deviation are used to determine deviations. It performs statistical analysis based on Deviation-MAD (Material Deviation Dependent Reduction). The aforementioned statistical analysis... The deviation information obtained as a result of the analysis is 30, which is obtained as a result of the fast Fourier transform. AI-based anomaly analysis with frequency domain features It is used. 11 The data fusion unit (1.2.3.2) in question is obtained by the feature extraction unit (1.2.3.1). Hydraulics generated by the live hydraulic digital twin unit (1.2.2) with the acquired attributes. The reference data will be combined and processed in the ensemble learning classification unit (1.2.3.3). It forms a combined dataset. The aforementioned collective learning classification unit (1.2.3.3) and feature extraction unit (1.2.3.1) The features extracted by the Random Forest and XGBoost algorithms are processed in parallel. by processing the status of the water distribution network using a batch learning architecture in which it is run It classifies. The collective learning classification unit in question (1.2.3.3), water distribution network condition normal consumption variation, microleakage, catastrophic pipe burst 10 or it is classified as pressure fluctuation originating from the pump and / or valve. The anomaly scoring unit in question (1.2.3.4) and the collective learning classification unit (1.2.3.3) Probability of the anomaly detected according to the classification output generated by It calculates the score. 15 Leakage positioning unit (1.2.4); detected by different IoT sensor nodes (1.1). Identify at least one time-of-arrival difference that determines the difference in arrival times between signals. unit (1.2.4.1), damping depending on the propagation of the signals in the pipeline. at least one signal attenuation analysis unit (1.2.4.2) that determines the characteristic and the specified word 20 The subject is to use the time-of-arrival difference and signal attenuation data to detect leaks in the pipe. It includes at least one positioning unit (1.2.4.3) that determines its position on the line. The time-of-arrival difference determination unit (1.2.4.1) determines the first IoT signal originating from the leak. The arrival time to sensor node (1.1) and the arrival time to the second IoT sensor node (1.1) is 25 It calculates the difference between them. The signal attenuation analysis unit (1.2.4.2) in question is used for different IoT sensor nodes (1.1) By comparing the amplitudes of the signals detected by the pipe, it is possible to determine the nature of those signals. It determines the damping characteristic depending on its propagation along the line. 30 The location determination unit (1.2.4.3) and the time difference determination unit (1.2.4.1) Signal attenuation analysis unit (1.2.4.2) with arrival time difference determined by By using the damping property determined by it, the leak on the pipeline 12 It determines the location. The location determination unit in question (1.2.4.3) determines the location of the leak. By relating its location to the pipe coordinates of the water distribution network, the leak point can be identified. It constitutes the geographical location information of the relevant parties. In a specific application of the invention, the said Leak location, time-of-arrival difference, and signal attenuation analysis together. It is determined with a positional accuracy of 1-2 meters through evaluation. 5 The intervention decision and control unit (1.2.5) uses AI-based anomaly analysis. Critical level pipe burst or high volume leak detected by unit (1.2.3) If this happens, it automatically triggers the emergency scenario and the relevant pipeline. The valve closing command to be transmitted to the valve control device (1.3) for the purpose of isolation is 10 It constitutes. In one application of the invention, the valve closing command in question is controlled by the valve control. to the device (1.3) in the form of an encrypted data packet via the wireless communication network. is transmitted. In a specific application of the invention, a critical level of pipe burst and / or Intervention decision and control unit upon detection of high volume leakage (1.2.5) Automatic valve closing command is given to the relevant valve control device (1.3) by 15 The information is transmitted and the leak is detected by operating the valve control device (1.3). The pipeline can be insulated in less than 15 minutes. The data communication unit (1.2.6) in question is the central processing unit (1.2) and the IoT sensor node. (1.1), valve control device (1.3), user device (1.4) and data storage unit (1.5) 20 It enables the transmission of electronic data and / or control commands. The valve control device (1.3) receives the valve control command sent from the central processing device (1.2). The area has at least one communication unit (1.3.1) that controls the valve according to the valve control command. at least one valve actuator (1.3.2) that mechanically moves the valve in question, 25 at least one valve status determination unit (1.3.3) that determines the status of the valve in question at least one actuator control unit that operates its actuator (1.3.2) according to the valve control command (1.3.4) includes. In one application of the invention, the field intervention regarding the leakage. work order closing information showing that it has been completed from the user device (1.4) central processing Upon transmission to the device (1.2), the intervention decision and control unit (1.2.5) controls the relevant valve 30 It generates a valve opening command to the device (1.3). The said actuator control unit (1.3.4) controls the valve actuator (1.3.2) according to the valve opening command. the valve being opened gradually and the isolated pipeline being put back into operation It provides. 13 The communication unit in question (1.3.1) receives the valve sent from the central processing unit (1.2). receiving the control command and the valve status determination unit (1.3.3) from the valve status determination unit It transmits the status information to the central processing device (1.2). The valve actuator (1.3.2) receives control from the actuator control unit (1.3.4). It converts the signal into mechanical motion, moving the valve stem and operating the valve. It allows for opening, closing, and / or changing the level of openness. The valve status determination unit (1.3.3) determines whether the valve is open, closed and / or has an intermediate opening of 10. by determining its status, the status data representing the status of that valve. It constitutes. The actuator control unit (1.3.4) receives the valve signal from the communication unit (1.3.1). It controls the movement of the valve actuator (1.3.2) by processing the control command. 15 The user device (1.4) enables electronic data exchange with the central processing device (1.2). a small communication unit (1.4.1), leakage, anomaly and from the central processing unit (1.2) at least one user interface that provides network status information to the user. (1.4.2), at least one 20 that enables the location of the identified leak to be shown on the map. The geographic visualization unit (1.4.3) alerts the user about the detected anomaly. at least one alarm notification unit (1.4.4) that ensures transmission and intervention to field personnel. at least one work order that provides information on accessing the leak location and Includes navigation unit (1.4.5). The user device in question (1.4) is a smartphone, tablet and in the specific applications of the invention. It can be a computer. The communication unit in question (1.4.1) received the leakage from the central processing unit (1.2), It receives anomaly and network status data and user-generated data. It transmits the subject to the central processing unit (1.2). The user interface in question (1.4.2) is leaking anomalies from the central processing unit (1.2). and visually presents network status data to the user, and the information entered by the user. 14 It enables the reception of commands and / or data. In an application of the invention, this is the case. The user interface (1.4.2) displays sensor measurement data and signals related to the detected anomaly. graphs, hydraulic fluctuation data, leak location information, anomaly score, a document containing information about the valve intervention performed and the time frame related to the intervention. It generates an anomaly report. This anomaly report is in electronic document format 5 It can be presented to the user and / or stored in the data storage unit (1.5). In one application of the invention, the user interface (1.4.2) provides an estimate of the detected leaks. information about water loss and water savings achieved as a result of the system's operation is provided to the user. It offers. The aforementioned geographic visualization unit (1.4.3) and leak location unit (1.2.4) A geographical map representing the water distribution network shows the location of the leak as determined by the authorities. It shows on it. The geographical visualization unit in question (1.4.3) shows water distribution. the hydraulic status of the network and the identified anomaly areas on a geographical map. It visualizes the hydraulic condition and / or 15 in question in an application of the invention. Anomaly density is shown as a heat map on a geographical map. The alarm notification unit (1.4.4) and the AI-based anomaly analysis unit (1.2.3) The system provides the user with visual and / or auditory alarm information regarding the anomaly detected. It is conveyed as follows: 20 The work order and navigation unit (1.4.5) in question pertain to the field response to the identified leak. to create the work order and assign field personnel to the leak location unit (1.2.4) It directs to the designated leak location. This refers to the work order and navigation unit. (1.4.5), the geographical location information of the leak determined within the scope of the field intervention work order, 25 anomaly priority information and the pipe belonging to the relevant pipeline registered in the network database (1.5.2) It provides information about the material and / or pipe depth to field personnel. Data storage unit (1.5); measurement data obtained from IoT sensor nodes (1.1) at least one sensor database storing (1.5.1), physical and geographical 30 of the water distribution network at least one network database (1.5.2) that stores network data representing its characteristics and anomalies identified by the artificial intelligence-based anomaly analysis unit (1.2.3) It contains at least one anomaly register unit (1.5.3) that stores the data. The sensor database in question (1.5.1) contains vibration data obtained from IoT sensor nodes (1.1), It stores pressure, flow, and temperature measurement data by correlating them with time information. The network database in question (1.5.2) contains information about the pipe diameter and pipe material of the water distribution network. It stores data such as pipe coordinates, pipe depth, network age, and topographic elevation. 5 The anomaly recording unit in question (1.5.3) includes the type of anomaly detected, the anomaly score, data regarding the identified leak location and the valve intervention performed It is hiding. Therefore, leaks and pipeline anomalies occurring in water distribution networks detection, determination of the location of the leak in the pipeline and the detected Water infrastructure leak detection and monitoring to ensure water flow control according to anomalies. intervention system (1); - enabling the acquisition of measurement data from the pipeline of the water distribution network. at least one IoT sensor node (1.1), - by processing the measurement data obtained from the said IoT sensor node (1.1) water detecting anomalies in the distribution network and identifying the leak in the pipe at least one central processing device (1.2) that determines its position in the line and - according to the control command generated by the central processing device (1.2) in question 20 at least one valve that changes the state of another valve in the water distribution network control device (1.3); - vibration and / or pressure of the said IoT sensor node (1.1) on the pipeline at least one mechanical parameter that enables the perception of the mechanical parameters representing The parameter sensor (1.1.1) measures the flow velocity and / or flow rate in the pipeline in question. at least one flow sensor (1.1.2) that enables its detection and of the pipeline in question and / or at least one that enables the detection of the temperature of the fluid passing through the pipeline. It should include a temperature sensor (1.1.3); - the central processing device in question (1.2), - a network model representing the physical structure of the water distribution network in question, 30 Using the current measurement data obtained from the IoT sensor node (1.1), water determining the expected hydraulic behavior of the distribution network and the expected from hydraulic behavior, instead of a predetermined fixed lower and / or upper threshold value 16 Dynamically determined according to the current hydraulic condition of the water distribution network. at least one live hydraulic digital twin unit (1.2.2) that generates hydraulic reference data, - analysis from measurement data obtained from the said IoT sensor node (1.1) defining the relevant attributes, and the live hydraulic digital twin with those attributes. combining the hydraulic reference data generated by unit (1.2.2) and 5 by processing the combined data with multiple machine learning models at least one AI-based anomaly that classifies the condition of the water distribution network unit of analysis (1.2.3), - arrival between signals detected by different IoT sensor nodes (1.1) Attenuation due to time difference and the propagation of these signals in the pipeline 10 By using these features together, the location of the leak in the pipeline is detected. at least one leak location unit (1.2.4) that determines and - critical by the AI-based anomaly analysis unit (1.2.3) in case a pipe burst and / or high-volume leak is detected at this level In order to isolate the relevant pipeline, the valve control device (1.3) 15 the valve closing command to be transmitted and the said valve closing command According to the valve control device (1.3), the pipeline is physically closed by closing the relevant valve. at least one intervention decision and control unit that ensures its isolation (1.2.5) It includes. The live hydraulic digital twin unit in question (1.2.2) measures the physical structure of the water distribution network. at least one network modeling tool that enables the creation of a network model representing the network. The unit (1.2.2.1) distributes water using the network model and current measurement data. at least one hydraulic simulation unit (1.2.2.2) that calculates the hydraulic behavior of the network and The calculated hydraulic behavior will be compared with the measured hydraulic parameters. 25 It contains at least one reference data generation unit (1.2.2.3) that generates the reference data. The reference data production unit in question (1.2.2.3); hydraulic simulation unit (1.2.2.2) The expected pressure and flow distribution is obtained from IoT sensor nodes (1.1). 30 Representing the current hydraulic status of the water distribution network by correlating it with measurement data. and independently of a predetermined fixed lower and / or upper threshold value to provide dynamically updated hydraulic reference data It is structured. 17 The AI-based anomaly analysis unit (1.2.3) is derived from IoT sensor nodes (1.1) at least one attribute that identifies the features for analysis from the obtained measurement data inference unit (1.2.3.1), live hydraulic digital twin unit (1.2.2) with the said attributes at least one data fusion unit that combines hydraulic reference data generated by (1.2.3.2) processes the combined data with multiple machine learning models 5 at least one ensemble learning classification unit (1.2.3.3) that classifies the state of the network and at least one that creates an anomaly score for the classified network status in question It includes anomaly scoring units (1.2.3.4). Feature extraction unit (1.2.3.1); fast Fourier transform (FFT) of measurement data. By applying this method, signal energy, peak factor, dominant frequency, and 10 can be obtained from the measurement data. It will obtain the pressure drop rate attributes and normal operation in the measurement data. Z-score and / or median absolute deviation to identify deviations from behavior It is configured to perform statistical analysis based on (MAD). The aforementioned collective learning classification unit (1.2.3.3) and feature extraction unit (1.2.3.1) The features extracted by the Random Forest and XGBoost algorithms are processed in parallel for 15 seconds. by processing the status of the water distribution network using a batch learning architecture in which it is run Normal consumption variation, microleak, catastrophic pipe burst, or pump and / or valve failure. It is structured to be classified as a pressure fluctuation caused by a specific source. The leaked positioning unit (1.2.4) is located at different IoT sensor nodes (1.1) 20 At least one configured time-of-arrival difference determination unit (1.2.4.1), of the signals in question The most structured to determine the damping characteristic depending on its propagation in the pipeline a small signal attenuation analysis unit (1.2.4.2) and the signal with the said arrival time difference By using both damping and other properties together, it will determine the location of the leak in the pipeline. It includes at least one location determination unit (1.2.4.3) configured in this way. 25 The location determination unit in question (1.2.4.3); the time difference determination unit (1.2.4.1) Signal attenuation analysis unit (1.2.4.2) with arrival time difference determined by By using the damping property determined by it, the leak on the pipeline will determine the location and the identified leak location will be transmitted to the water distribution network. By associating the leak point's geographical location information with the corresponding pipe coordinates, 30 It is structured in a way that will create it. 18 The valve control device in question (1.3) receives the valve sent from the central processing device (1.2). At least one communication unit configured to receive the control command (1.3.1), word The subject is to mechanically move the valve according to the valve control command. At least one valve actuator configured (1.3.2) indicates the physical condition of the valve in question. at least one valve status determination unit (1.3.3) configured to determine and word 5 The subject is to operate the valve actuator (1.3.2) according to the valve control command. It includes at least one configured actuator control unit (1.3.4). The intervention decision and control unit (1.2.5) is responsible for the field response to the leak. Upon receiving the work order closing information indicating completion, the relevant valve is reopened. It is configured to generate a valve opening command to open it; word 10 The subject actuator control unit (1.3.4) controls the valve according to the valve opening command. by controlling the actuator (1.3.2) the valve is opened gradually and the insulated pipe It is configured to enable the line to be reactivated. The user device (1.4) exchanges electronic data with the central processing device (1.2). at least one communication unit configured to provide (1.4.1), central processing 15 the leak, anomaly and network status information received from the device (1.2) is presented to the user. At least one user interface configured to provide (1.4.2), identified leakage at least one geographical configuration that allows its location to be shown on a map The visualization unit (1.4.3) ensures that the warning regarding the detected anomaly is conveyed to the user. at least one alarm notification unit configured to provide (1.4.4) and 20 field personnel This will enable the provision of intervention information and access information to the leak location. It includes at least one configured work order and navigation unit (1.4.5). The data storage unit in question (1.5) stores the measurements obtained from IoT sensor nodes (1.1). At least one sensor database configured to store data (1.5.1), water distribution 25 will store network data representing the physical and geographical characteristics of the network at least one network database structured in this way (1.5.2) and AI-based anomaly detection to store data on anomalies identified by the analysis unit (1.2.3) It contains at least one configured anomaly register unit (1.5.3). the IoT sensor node in question (1.1); from the mechanical parameter sensor (1.1.1), flow Obtaining measurement data from the sensor (1.1.2) and the temperature sensor (1.1.3) 30 and at least one microcontroller configured to enable its processing (1.1.4), 19 The goal is to reduce noise in the measurement data and make the measurement data suitable for analysis. at least one signal preprocessing unit (1.1.5) configured to enable its delivery and the transmission of the pre-processed measurement data to the central processing unit (1.2) It includes at least one wireless communication unit (1.1.6) configured to provide. The central processing unit (1.2) in question receives the measurement from the IoT sensor node (1.1). at least one data set configured to receive data on the water distribution network and related data. the receiving unit (1.2.1) and the central processing unit (1.2) with the other electronic components of the system at least one data transmitter configured to enable the exchange of data and control commands It includes unit (1.2.6). Detecting leaks and pipeline anomalies in water distribution networks, 10 Determining the location of the leak in the pipeline and adjusting the water supply according to the detected anomaly. The invention concerns the control of water flow, water infrastructure leak detection and intervention. method; - Measurement data from the water distribution network pipeline from at least one IoT sensor Obtaining via node (1.1), 15 - at least one of the measurement data obtained from the said IoT sensor node (1.1) transmitting to the central processing unit (1.2), - processing of the said measurement data by the central processing unit (1.2) and - at least one valve of the control command generated by the central processing device (1.2) transmitting to the control device (1.3); 20 - IoT of vibration and / or pressure, flow and temperature measurement data related to the pipeline. obtaining via sensor node (1.1), - IoT sensor with network model representing the physical structure of the water distribution network Live hydraulic digital twin of current measurement data obtained from node (1.1) The expected hydraulic 25 of the water distribution network is used by unit (1.2.2). determining its behavior and anticipating the expected hydraulic behavior in advance. Instead of a defined fixed lower and / or upper threshold value, the current water distribution network dynamically determined hydraulic reference data according to hydraulic status creation, - 30 measurement data obtained from the IoT sensor node (1.1) for analysis. Attribute definition, matching those attributes with the live hydraulic digital twin unit. Combining the hydraulic reference data generated by (1.2.2) and the combined data is processed by multiple machine learning models Artificial intelligence-based anomaly analysis unit for the status of the water distribution network (1.2.3) classification by means of, - arrival between signals detected by different IoT sensor nodes (1.1) Attenuation due to time difference and the propagation of these signals in the pipeline 5 The location of the leak in the pipeline is determined by using these features together. Determination via the leak location unit (1.2.4), - Critical level pipe by artificial intelligence based anomaly analysis unit (1.2.3) in case of a burst and / or high-volume leak, the relevant pipe Valve closing 10 to be transmitted to valve control device (1.3) in order to isolate the line. the command is generated by the intervention decision and control unit (1.2.5) and the word The issue is that the valve is closed according to the valve closing command, thereby isolating the pipeline. It includes the steps to be taken. Network modeling is a network model that represents the physical structure of the water distribution network. The creation of the unit (1.2.2.1) and the current measurement data with the said network model 15 hydraulic simulation of the expected pressure and flow distribution of the water distribution network using Determination of the unit (1.2.2.2) and the determined expected pressure and flow hydraulic reference data to be compared with hydraulic parameters measured from the distribution The reference data is generated via the data production unit (1.2.2.3). The expected pressure and flow distribution by the hydraulic simulation unit (1.2.2.2) of IoT 20 correlation with the measurement data obtained from the sensor nodes (1.1) and water distribution hydraulic reference data representing the current hydraulic status of the network in advance dynamically, independently of a defined fixed lower and / or upper threshold value It is being updated. In the method, 25 measurement data obtained from IoT sensor nodes (1.1) are used for analysis. Determining the attributes via the attribute extraction unit (1.2.3.1), Hydraulic reference generated by live hydraulic digital twin unit (1.2.2) with attributes Combining the data through the data fusion unit (1.2.3.2), the combined data By processing the data with multiple machine learning models, a cumulative assessment of the network's state is obtained. Classification of learning through the classification unit (1.2.3.3) and the classified word 30 the anomaly scoring unit of the network status anomaly score (1.2.3.4) This is ensured by means of 21 A fast Fourier transform (FFT) is applied to the measurement data, and the measurement in question... signal energy, peak factor, dominant frequency, and pressure drop rate features from the data obtaining and identifying deviations from normal operating behavior in the measurement data. For the purpose of statistical analysis based on Z-score and / or median absolute deviation (MAD). is being carried out. 5 The features extracted by the feature extraction unit (1.2.3.1) are generated in a Random Forest and With a batch learning architecture where XGBoost algorithms are run in parallel being processed and the condition of the water distribution network, normal consumption variation, microleakage, as a catastrophic pipe burst or pressure surge caused by a pump and / or valve It is classified. 10 Arrival time between signals detected by different IoT sensor nodes (1.1) the difference is determined by means of the arrival time difference determination unit (1.2.4.1), Signal attenuation analysis of attenuation characteristics depending on signal propagation in the pipeline. determination via unit (1.2.4.2) and the signal with the difference in arrival time in question. By using the damping feature in combination, the location of the leak in the pipeline can be determined at position 15. Determination is ensured through the determination unit (1.2.4.3). with the arrival time difference determined by the arrival time difference determination unit (1.2.4.1) the damping characteristic determined by the signal damping analysis unit (1.2.4.2) By using them together, the location of the leak on the pipeline can be determined and identified. The location of the leak is 20, corresponding to the pipe coordinates of the water distribution network. By correlating these factors, geographical location information for the leak point is generated. Communication unit of valve control command sent from central processing device (1.2) (1.3.1) is received via the valve actuator according to the valve control command in question. (1.3.2) operation of the valve via the actuator control unit (1.3.4) mechanically moving and the physical condition of the valve in question is determined by valve status unit 25 Determination is ensured through (1.3.3). Work order closing information indicating that the field response to the leak has been completed. Upon receipt, the valve opening command is issued to reopen the relevant valve. the creation of the decision and control unit (1.2.5) to the valve opening command in question According to the valve actuator (1.3.2), it is controlled via the actuator control unit (1.3.4) 30 22 by gradually opening the valve and reactivating the isolated pipeline is provided. Electronic data exchange between the central processing device (1.2) and the user device (1.4) This is done via the communication unit (1.4.1) from the central processing device (1.2) The leak, anomaly and network status information received is transmitted via the user interface (1.4.2) 5 the presentation to the user of the geographic visualization unit of the determined leak location (1.4.3) By displaying the detected anomaly on a map, the warning about the anomaly is sent as an alarm. notification unit (1.4.4) to the user and intervention information to field personnel with access information to the leak location via the work order and navigation unit (1.4.5) is provided. 10 Measurement data obtained from IoT sensor nodes (1.1) are stored in the sensor database (1.5.1) storage, a network representing the physical and geographical characteristics of the water distribution network. storage of data in network database (1.5.2) and AI-based anomaly analysis Data regarding anomalies determined by unit (1.2.3) are recorded in the anomaly registration unit. (1.5.3) storage is provided. 15 from the mechanical parameter sensor (1.1.1), the flow sensor (1.1.2) and the temperature sensor (1.1.3) receiving the obtained measurement data via the microcontroller (1.1.4) and processing, reducing noise in the measurement data and the measurement data in question The measurement data is processed by a signal preprocessing unit to make it suitable for analysis. Pre-processing via (1.1.5) and the said measurement that has been pre-processed 20 data transmitted to the central processing device (1.2) via the wireless communication unit (1.1.6) The transmission is ensured. Measurement data transmitted from the IoT sensor node (1.1) and water distribution network data is received via the data receiving unit (1.2.1) and the central processing device (1.2). Data exchange and control command exchange between other electronic components of the system. This is ensured through the communication unit (1.2.6). Industrial Applicability of the Invention The invention solves leaks and pipeline anomalies occurring in water distribution networks. Detection, location, and control of water flow according to the anomalies in question. 30 23 It relates to a system and method for carrying out; municipal water distribution networks, applicable in industrial water infrastructures and similar pressurized water distribution systems. It is of a certain quality. The invention is not limited to the above descriptions; a person skilled in the field can easily identify 5 aspects of the invention. It can present different applications. These are the claims of the invention and the protection sought. should be evaluated within this context. 15 25

Claims

24 REQUESTS 1. Detection of leaks and pipeline anomalies in water distribution networks. the process involves determining the location of the leak in the pipeline and analyzing the identified anomaly. to ensure that water flow is controlled accordingly; - enables the acquisition of measurement data from the pipeline of the water distribution network. at least one IoT sensor node (1.1), - by processing the measurement data obtained from the said IoT sensor node (1.1) water detecting anomalies in the distribution network and identifying the leak in the pipe at least one central processing device (1.2) that determines its position in the line and - according to the control command generated by the central processing device (1.2) in question 10 at least one valve that changes the state of another valve in the water distribution network containing control device (1.3); - vibration and / or pressure of the said IoT sensor node (1.1) on the pipeline at least one mechanical parameter that enables the perception of the mechanical parameters representing The parameter sensor (1.1.1) measures the flow velocity and / or flow rate in the pipeline in question. at least one flow sensor (1.1.2) that enables its detection and of the pipeline in question and / or at least one that enables the detection of the temperature of the fluid passing through the pipeline. It should include a temperature sensor (1.1.3); - the central processing device in question (1.2), - a network model representing the physical structure of the water distribution network in question, 20 Using the current measurement data obtained from the IoT sensor node (1.1), water determining the expected hydraulic behavior of the distribution network and the expected from hydraulic behavior, instead of a predetermined fixed lower and / or upper threshold value Dynamically determined according to the current hydraulic condition of the water distribution network. At least one live hydraulic digital twin unit (1.2.2) that generates hydraulic reference data, 25 - analysis from measurement data obtained from the said IoT sensor node (1.1) defining the relevant attributes, and the live hydraulic digital twin with those attributes. combining the hydraulic reference data generated by unit (1.2.2) and by processing the combined data with multiple machine learning models at least one AI-based anomaly classifying the status of the water distribution network 30 unit of analysis (1.2.3), - arrival between signals detected by different IoT sensor nodes (1.1) attenuation depending on the time difference and the propagation of these signals in the pipeline. By using these features together, the location of the leak in the pipeline is detected. at least one leak location unit (1.2.4) that determines and - critical by the AI-based anomaly analysis unit (1.2.3) in case a pipe burst and / or high-volume leak is detected at this level In order to isolate the relevant pipeline, the valve control device (1.3) 5 the valve closing command to be transmitted and the said valve closing command According to the valve control device (1.3), the pipeline is physically closed by closing the relevant valve. It must include at least one intervention decision and control unit (1.2.5) that enables its isolation. Water infrastructure leak detection and intervention system characterized by (1).

2. The live hydraulic digital twin unit in question (1.2.2); the physical 10 of the water distribution network at least one network that enables the creation of a network model representing its structure The modeling unit (1.2.2.1) uses the network model and current measurement data in question. using at least one hydraulic system that calculates the hydraulic behavior of the water distribution network. simulation unit (1.2.2.2) and measured from the calculated hydraulic behavior in question At least one reference data point that will be compared with hydraulic parameters. Water infrastructure as in claim 1, characterized by including a production unit (1.2.2.3). Leak detection and intervention system (1).

3. Pressure and flow distribution expected by the hydraulic simulation unit (1.2.2.2) for IoT water distribution by correlating it with the measurement data obtained from the sensor nodes (1.1) a predetermined fixed sub-20 representing the current hydraulic status of the network and / or dynamically updated hydraulics, regardless of the upper threshold value. the aforementioned reference data generation is structured to provide reference data. Water infrastructure leak detection as in claim 2, characterized by including unit (1.2.2.3). and intervention system (1).

4. The aforementioned AI-based anomaly analysis unit (1.2.3); IoT sensor 25 Features for analysis from measurement data obtained from nodes (1.1) at least one feature extraction unit (1.2.3.1) that determines the features in question and the living organism hydraulic reference data generated by the hydraulic digital twin unit (1.2.2) at least one data fusion unit (1.2.3.2) combining the data in question, one of which is the combined data. at least one 30-bit machine learning model that classifies the state of the network by processing it. collective learning classification unit (1.2.3.3) and the classified network in question at least one anomaly scoring unit (1.2.3.4) that constitutes an anomaly score relating to the condition a water infrastructure as in any of claims 1-3 characterized by including Leak detection and intervention system (1). 26 5. By applying the fast Fourier transform (FFT) to the measurement data, the measurement in question signal energy, peak factor, dominant frequency, and pressure drop rate from the data will obtain the features and from the normal operating behavior in the measurement data Z-score and / or median absolute deviation (MAD) for determining deviations The aforementioned feature is structured to perform statistical analysis based on attribute 5. a water like the one in claim 4, characterized by containing the inference unit (1.2.3.1). Infrastructure leak detection and response system (1).

6. Features extracted by the feature extraction unit (1.2.3.1) are placed in a Random Forest and With a batch learning architecture where XGBoost algorithms are run in parallel by processing the condition of the water distribution network, normal consumption variation, microleakage, 10 catastrophic pipe burst or pressure surge caused by pump and / or valve The aforementioned collective learning classification is structured in a way that will categorize them as such. a water infrastructure as in claim 4 or 5 characterized by including unit (1.2.3.3). Leak detection and intervention system (1).

7. Arrival time between signals detected by different IoT sensor nodes (1.1) is 15 at least one arrival time difference determination unit configured to identify the difference (1.2.4.1) refers to the attenuation characteristic of the signals in question depending on their propagation in the pipeline. at least one signal attenuation analysis unit (1.2.4.2) configured to determine and by using the aforementioned arrival time difference together with the signal attenuation feature at least one location configured to determine the location of the leak in the pipeline 20 the leak location unit (1.2.4) which includes the detection unit (1.2.4.3) a water infrastructure as in any of claims 1-6 characterized by including Leak detection and intervention system (1).

8. Arrival time difference determined by the arrival time difference determination unit (1.2.4.1) The attenuation characteristic determined by the signal attenuation analysis unit (1.2.4.2) is 25 Using this, it will determine the location of the leak on the pipeline and the identified word the issue is relating the leak location to the pipe coordinates of the water distribution network. The statement is structured to generate geographical location information for the leak point. Claim 7, characterized by its inclusion of the subject matter location determination unit (1.2.4.3). such as a water infrastructure leak detection and intervention system (1). 30 9. To receive the valve control command sent from the central processing unit (1.2). At least one communication unit configured (1.3.1) to respond to the valve control command in question at least one valve configured to be mechanically operated according to the system. actuator (1.3.2) to determine the physical condition of the valve in question 27 at least one valve status determination unit (1.3.3) is configured and the valve in question at least the actuator (1.3.2) is configured to operate according to the valve control command. including the valve control device (1.3) which includes an actuator control unit (1.3.4) Water infrastructure leak detection as in any of claims 1-8 characterized by and intervention system (1). 5 10. Work order closing information indicating that the field response to the leak has been completed. upon receipt, the valve opening command is given to reopen the relevant valve. the intervention decision and control unit (1.2.5) which is structured to form and by controlling the valve actuator (1.3.2) according to the valve opening command in question the valve should be opened gradually and the isolated pipeline should be put back into operation 10 the said actuator control unit (1.3.4) which is configured to enable the taking of. a water infrastructure leak detection and leak detection system as described in claim 9, characterized by its inclusion. intervention system (1).

11. To enable electronic data exchange with the central processing unit (1.2). at least one configured communication unit (1.4.1), 15 received from the central processing device (1.2) This will enable the user to be provided with information on leaks, anomalies, and network status. At least one configured user interface (1.4.2), map of the identified leak location. at least one geographic visualization structured to allow it to be displayed on the map. Unit (1.4.3) will ensure that the user is notified of the detected anomaly. at least one alarm notification unit configured as (1.4.4) and 20 field personnel This will enable the provision of intervention information and access information to the leak location. at least one work order and navigation unit (1.4.5) configured in this way any of the claims 1-10 characterized by including the user device (1.4). such as a water infrastructure leak detection and intervention system (1).

12. To store the measurement data obtained from IoT sensor nodes (1.1) 25 at least one configured sensor database (1.5.1), physical and of the water distribution network the most structured to store network data representing its geographical features a small network database (1.5.2) and an AI-based anomaly analysis unit (1.2.3) the most structured to store data on anomalies identified by 30 It must contain at least one data storage unit (1.5) containing at least one anomaly record unit (1.5.3). Water infrastructure leak detection as in any of claims 1-11 characterized by and intervention system (1).

13. The IoT sensor node in question (1.1) receives flow from the mechanical parameter sensor (1.1.1). measurement data obtained from the sensor (1.1.2) and the temperature sensor (1.1.3) 28 at least one microcontroller configured to enable the reception and processing of data (1.1.4) aims to reduce noise in the measurement data and to improve the measurement data. at least one signal pre-configured in a way that will make it suitable for analysis the processing unit (1.1.5) and the central processing unit of the said pre-processed measurement data at least one wireless 5 configured to transmit to the processing device (1.2) Any of the requests 1-12 characterized by containing a communication unit (1.1.6). a water infrastructure leak detection and intervention system like one (1).

14. The central processing device (1.2) receives the measurement data transmitted from the IoT sensor node (1.1). at least one configured to receive data and data relating to the water distribution network Data receiving unit (1.2.1) and central processing unit (1.2) other electronic 10 of the system at least a component configured to enable the exchange of data and control commands. Any of the requests 1-13 characterized by containing a data communication unit (1.2.6). a water infrastructure leak detection and intervention system like one (1).

15. Detection of leaks and pipeline anomalies occurring in water distribution networks. the process involves determining the location of the leak in the pipeline and addressing the identified anomaly. according to; - Measurement data from the water distribution network pipeline from at least one IoT sensor obtaining via node (1.1), - at least one of the measurement data obtained from the said IoT sensor node (1.1) Transmission to the central processing unit (1.2), 20 - processing of the said measurement data by the central processing unit (1.2) and - at least one valve of the control command generated by the central processing device (1.2) including transmission to the control device (1.3); - IoT of vibration and / or pressure, flow and temperature measurement data related to the pipeline. Obtained via sensor node (1.1), 25 - IoT sensor with network model representing the physical structure of the water distribution network Live hydraulic digital twin of current measurement data obtained from node (1.1) The expected hydraulic water distribution network is used by unit (1.2.2). determining its behavior and anticipating the expected hydraulic behavior in advance. Instead of a defined fixed lower and / or upper threshold value, the current 30 of the water distribution network dynamically determined hydraulic reference data according to hydraulic status creation, 29 - Analysis of measurement data obtained from the IoT sensor node (1.1) Attribute definition, matching those attributes with the live hydraulic digital twin unit. Combining the hydraulic reference data generated by (1.2.2) and the combined data is processed by multiple machine learning models Artificial intelligence-based anomaly analysis unit for the status of the water distribution network (1.2.3) 5 classification by means of, - arrival between signals detected by different IoT sensor nodes (1.1) attenuation depending on the time difference and the propagation of these signals in the pipeline. The location of the leak in the pipeline is determined by using these features together. Determination via the leak positioning unit (1.2.4) and 10 - Critical level pipe by artificial intelligence based anomaly analysis unit (1.2.3) in case of a burst and / or high-volume leak, the relevant pipe Valve closing signal to be transmitted to valve control device (1.3) for the purpose of isolating the line the command is generated by the intervention decision and control unit (1.2.5) and the word The issue is that the valve is closed according to the valve closing command, thus isolating the pipeline. 15 A water infrastructure leak detection and response method characterized by its use.

16. Network modeling of the network model representing the physical structure of the water distribution network. the creation of the unit (1.2.2.1) and current measurement with the said network model The expected pressure and flow distribution of the water distribution network can be determined using hydraulic data. Determination via the simulation unit (1.2.2.2) and the aforementioned expected 20 hydraulic parameters to be compared with hydraulic parameters measured from pressure and flow distribution by generating reference data through the reference data generation unit (1.2.2.3) A water infrastructure leak detection and response method characterized as in claim 15.

17. Expected pressure and flow distribution by the hydraulic simulation unit (1.2.2.2) for IoT Correlation with the measurement data obtained from the sensor nodes (1.1) and water distribution 25 hydraulic reference data representing the current hydraulic status of the network in advance dynamically, independently of a defined fixed lower and / or upper threshold value Water infrastructure leak detection and update, as in request 16, characterized by its functionality. intervention method.

18. 30 for analysis of measurement data obtained from IoT sensor nodes (1.1). Determining the attributes via the attribute extraction unit (1.2.3.1), Hydraulics created by live hydraulic digital twin unit (1.2.2) with attributes Combining reference data via the data fusion unit (1.2.3.2), combined The data in question is processed using multiple machine learning models to analyze the network. classification of the situation through the collective learning classification unit (1.2.3.3) and anomaly scoring of the anomaly score related to the classified network status any of the requests 15-17 characterized by being created via unit (1.2.3.4) A water infrastructure leak detection and repair method, like the one in 5.

19. Applying the fast Fourier transform (FFT) to the measurement data, the measurement in question signal energy, peak factor, dominant frequency, and pressure drop rate from the data obtaining features and normal operating behavior in measurement data Z-score and / or median absolute deviation (MAD) for determining deviations a 10 like the one in request 18 characterized by performing a statistically based analysis. Water infrastructure leak detection and repair method.

20. Features extracted by the feature extraction unit (1.2.3.1) in Random Forest and With a batch learning architecture where XGBoost algorithms are run in parallel processing and condition of the water distribution network, normal consumption variation, microleakage, Catastrophic pipe burst or pressure surge caused by pump and / or valve 15 a water infrastructure like the one in claim 18 or 19, characterized by its classification as such. Leak detection and response method.

21. Arrival time between signals detected by different IoT sensor nodes (1.1) The difference is determined by means of the arrival time difference determination unit (1.2.4.1), signal attenuation 20 depends on the attenuation characteristic of the signals propagation in the pipeline. determination via the analysis unit (1.2.4.2) and the signal with the said arrival time difference By using damping properties in combination, the location of the leak in the pipeline is determined. The request is characterized by its determination via the location determination unit (1.2.4.3). A water infrastructure leak detection and repair method, like any of the 15-20 methods.

22. Arrival time difference determined by the arrival time difference determination unit (1.2.4.1) and 25 the damping characteristic determined by the signal damping analysis unit (1.2.4.2) By using them together, the location of the leak on the pipeline can be determined and The identified leak location is determined by the pipe coordinates of the water distribution network. by associating and generating geographical location information for the leak point A water infrastructure leak detection and response method characterized as in claim 21. 30 23. Communication unit of valve control command sent from central processing device (1.2) (1.3.1) is received via the valve actuator according to the valve control command in question. (1.3.2) operation of the valve via the actuator control unit (1.3.4) mechanically moving and determining the physical condition of the valve in question. 31 any of the claims 15-23 characterized by being determined by means of unit (1.3.3). A method for detecting and responding to water infrastructure leaks, similar to one in another.

24. Work order closing information indicating that the field response to the leak has been completed. upon receipt of the valve opening command, the valve is reopened. The opening of the valve in question is to be established by the intervention decision and control unit (1.2.5). According to the command, the valve actuator (1.3.2) is controlled by the actuator control unit (1.3.4). by checking and gradually opening the valve and reconnecting the isolated pipeline. Water infrastructure leak detection, as in system 23, characterized by its commissioning. and intervention method.

25. Electronic data exchange between the central processing device (1.2) and the user device (1.4) 10 This is done via the communication unit (1.4.1) from the central processing device (1.2) The leak, anomaly and network status information received is transmitted via the user interface (1.4.2). the presentation to the user of the geographic visualization unit of the determined leak location (1.4.3) By displaying the detected anomaly on a map, the warning about the anomaly is sent as an alarm. Notification unit (1.4.4) to be communicated to the user and intervention to field personnel 15 work order and navigation unit (1.4.5) with information on access to leak location. a request characterized by being provided through a means such as in any of 15-24 Water infrastructure leak detection and repair method.

26. Measurement data obtained from IoT sensor nodes (1.1) in the sensor database (1.5.1) storage, representing the physical and geographical characteristics of the water distribution network 20 Storage of network data in the network database (1.5.2) and artificial intelligence based Data relating to anomalies determined by the anomaly analysis unit (1.2.3) any of the requests 15-25 characterized by being stored in the registration unit (1.5.3). A method for detecting and responding to water infrastructure leaks, similar to one in another.

27. Mechanical parameter sensor (1.1.1), flow sensor (1.1.2) and temperature 25 Measurement data obtained from the sensor (1.1.3) is transmitted via the microcontroller (1.1.4) acquisition and processing of the measurement data, reduction of noise in the measurement data and measurement In order to make the measurement data suitable for analysis, the signal of the said measurement data... preprocessing via the preprocessing unit (1.1.5) and the preprocessed word The subject is the measurement data transmitted to the central processing unit via the wireless communication unit (1.1.6) 30 The request is characterized by being transmitted to the device (1.2) as in any of 15-26. A method for detecting and responding to water infrastructure leaks.

28. Measurement data transmitted from the IoT sensor node (1.1) and related to the water distribution network data is received via the data receiving unit (1.2.1) and the central processing device (1.2). 32 data exchange of data and control commands between other electronic components of the system Request 15- characterized by being carried out via communication unit (1.2.6). A water infrastructure leak detection and response method, like any of the 27. 10 20