An intelligent water management system based on big data technology and internet of things

Abstract

This invention relates to the field of intelligent management, specifically disclosing an intelligent water management system based on big data technology and the Internet of Things (IoT). The system collects real-time water flow data from sensors deployed at key nodes of a pipeline network. It then utilizes artificial intelligence data analysis and processing algorithms based on big data technology and IoT to analyze and correlate the time-series patterns of water flow at these key nodes. Based on a comparison of the similarity between the time-series patterns of the analyzed key node and the sets of key nodes with a predetermined threshold, the system automatically determines whether the analyzed key node is a leak node. In this way, the system can automatically identify potential leak nodes, improving the accuracy and response speed of leak detection, while reducing the need for manual inspections, lowering operating costs, and thus improving the reliability and stability of the entire water system.

Description

Technical Field

[0001] This invention relates to the field of intelligent management, and more specifically, to an intelligent water management system based on big data technology and the Internet of Things. Background Technology

[0002] Water management refers to a series of activities encompassing the comprehensive planning, development, allocation, utilization, protection, and restoration of water resources. It involves all aspects of water resources, including surface water and groundwater monitoring, water quality management, construction and maintenance of water conservancy projects, flood and drought prevention, management of irrigation and water supply systems, and wastewater treatment and reuse. The purpose of water management is to ensure the sustainable use of water resources and meet the needs of socio-economic development and ecological environmental protection.

[0003] Pipeline leak detection is a crucial aspect of water management. It involves using specific technologies and methods to identify and locate leaks in water supply networks—specifically, the locations where leaks occur within the pipes. This work is essential for maintaining the safe operation of water supply systems, reducing water waste, minimizing economic losses, and improving water supply efficiency.

[0004] However, traditional pipeline leakage detection methods, such as acoustic and sonic methods, largely rely on the professional knowledge and experience of the inspectors. This can lead to highly subjective and inconsistent results. Furthermore, manual inspection is time-consuming and inefficient, making it difficult to adapt to large-scale urban water supply systems with complex pipe networks. In addition, traditional methods have a limited scope and low accuracy in locating leaks, especially in noisy environments or where pipelines are deeply buried, resulting in poor detection performance.

[0005] Therefore, a smart water management system based on big data technology and the Internet of Things is desired. Summary of the Invention

[0006] To address the aforementioned technical problems, this application is proposed. Embodiments of this application provide an intelligent water management system based on big data technology and the Internet of Things (IoT). This system collects real-time water flow data from water flow sensors deployed at various key nodes of the pipeline network system. It then utilizes artificial intelligence data analysis and processing algorithms based on big data technology and IoT to analyze and correlate the time-series patterns of the water flow at these key nodes. Based on a comparison of the similarity between the water flow time-series patterns of the analyzed key node and the sets of key nodes with a predetermined threshold, the system automatically determines whether the analyzed key node is a leaking node. In this way, the system can automatically identify potential leaking nodes, improving the accuracy and response speed of leak detection, while reducing the need for manual inspections, lowering operating costs, and thus improving the reliability and stability of the entire water management system.

[0007] According to one aspect of this application, a smart water management system based on big data technology and the Internet of Things is provided, comprising:

[0008] The water flow acquisition module is used to acquire a collection of time series data of water flow collected by water flow sensors deployed at various key nodes of the pipeline network system.

[0009] The water flow time series arrangement module is used to arrange the time series of each water flow in the set of water flow time series according to the time dimension to form a water flow time series input vector to obtain a set of water flow time series input vectors;

[0010] The water flow time-series pattern feature extraction module is used to extract water flow time-series pattern features from each water flow time-series input vector in the set of water flow time-series input vectors to obtain a set of water flow time-series associated feature vectors;

[0011] The water flow clustering module is used to obtain water flow clustering pattern feature vectors by passing the set of water flow time-series associated feature vectors through a joint clustering analysis network.

[0012] The key node water flow extraction module is used to extract the water flow time-series correlation feature vector corresponding to the key node to be analyzed as the query feature vector from the set of water flow time-series correlation feature vectors.

[0013] The similarity calculation module is used to calculate the hash similarity between the query feature vector and the water flow clustering pattern feature vector;

[0014] The leakage node determination module is used to determine whether the key node to be analyzed is a leakage node based on the comparison between the hash similarity and a predetermined threshold.

[0015] Compared to existing technologies, this application provides an intelligent water management system based on big data technology and the Internet of Things (IoT). This system collects real-time water flow data from water flow sensors deployed at key nodes of the pipeline network. It then utilizes AI data analysis and processing algorithms based on big data technology and IoT to analyze and correlate the time-series patterns of the water flow at these key nodes. Based on a comparison of the similarity between the water flow time-series patterns of the analyzed key node and the sets of key nodes with a predetermined threshold, the system automatically determines whether the analyzed key node is a leak node. In this way, the system can automatically identify potential leak nodes, improving the accuracy and response speed of leak detection, while reducing the need for manual inspections, lowering operating costs, and thus improving the reliability and stability of the entire water management system. Attached Figure Description

[0016] The embodiments of this application will be described in more detail with reference to the accompanying drawings, outlining a series of activities described above and thereafter. These activities relate to various aspects of water resources, including the monitoring of surface water and groundwater, water quality management, construction and maintenance of water conservancy projects, flood and drought prevention, management of irrigation and water supply systems, and wastewater treatment and reuse. The purpose of water management is to ensure the sustainable use of water resources and meet the needs of socio-economic development and ecological environmental protection.

[0017] Pipeline leak detection is a crucial aspect of water management. It involves using specific technologies and methods to identify and locate leaks in water supply networks—specifically, the locations where leaks occur within the pipes. This work is essential for maintaining the safe operation of water supply systems, reducing water waste, minimizing economic losses, and improving water supply efficiency.

[0018] However, traditional pipeline leakage detection methods, such as acoustic and sonic methods, largely rely on the professional knowledge and experience of the inspectors. This can lead to highly subjective and inconsistent results. Furthermore, manual inspection is time-consuming and inefficient, making it difficult to adapt to large-scale urban water supply systems with complex pipe networks. In addition, traditional methods have a limited scope and low accuracy in locating leaks, especially in noisy environments or where pipelines are deeply buried, resulting in poor detection performance.

[0019] It's understandable that big data technology involves a range of tools, techniques, and methods for collecting, processing, and analyzing data from large, complex datasets. Big data is enormous, ranging from terabytes (TB) to petabytes (PB), and its generation and processing speed are extremely fast. The Internet of Things (IoT) refers to connecting various objects such as embedded systems, sensors, machines, and home appliances through the internet to form a vast network, enabling intelligent identification, location, tracking, monitoring, and management. Big data technology and IoT are often used in combination. IoT devices, as data sources, can generate massive amounts of real-time data, while big data technology can process and analyze this real-time data, providing immediate insights and decision support. Therefore, combining the analytical capabilities of big data with data collected through IoT devices can predict equipment failures and maintenance needs, thereby improving efficiency and reducing costs. This combination allows the two technologies to reinforce each other, jointly driving the development of intelligent systems and automated solutions. In particular, the application of big data technology and IoT in water management is also very widespread, jointly promoting the intelligentization and automation of water management. For example, by installing various sensors (such as water level, flow rate, and water quality sensors), comprehensive perception and real-time data collection of the water environment can be achieved; through data analysis and technical means, leakage rates in the water supply system can be effectively controlled, reducing operating costs; and through a zoned metering management system, leakage control and optimization of the water supply network can be achieved. These applications demonstrate the deep integration of big data technology and the Internet of Things in water management. They not only improve the efficiency and accuracy of water management but also provide strong technical support for the sustainable use of water resources.

[0020] Based on this, this application proposes an intelligent water management system based on big data technology and the Internet of Things (IoT). It collects real-time water flow data from water flow sensors deployed at various key nodes of the pipeline network system. Then, it utilizes AI data analysis and processing algorithms based on big data technology and IoT to analyze and correlate the time-series patterns of water flow at these key nodes. By comparing the similarity of the water flow time-series patterns between the analyzed key node and the sets of key nodes with a predetermined threshold, the system automatically determines whether the analyzed key node is a leak node. In this way, the system can automatically identify potential leak nodes, improving the accuracy and response speed of leak detection, while reducing the need for manual inspections, lowering operating costs, and thus improving the reliability and stability of the entire water management system.

[0021] Figure 1 This is a block diagram of an intelligent water management system based on big data technology and the Internet of Things according to an embodiment of this application. Figure 2This is a schematic diagram of the architecture of a smart water management system based on big data technology and the Internet of Things, according to an embodiment of this application. Figure 1 and Figure 2 As shown, the intelligent water management system 100 based on big data technology and the Internet of Things according to an embodiment of this application includes: a water flow acquisition module 110, used to acquire a set of time series of water flow collected by water flow sensors deployed at various key nodes of the pipeline network system; a water flow time series arrangement module 120, used to arrange the time series of each water flow in the set of water flow time series according to the time dimension to obtain a set of water flow time series input vectors; and a water flow time series pattern feature extraction module 130, used to extract water flow time series pattern features from each water flow time series input vector in the set of water flow time series input vectors to obtain the water flow. The system includes: a set of time-series associated feature vectors; a water flow clustering module 140, used to pass the set of water flow time-series associated feature vectors through a joint clustering analysis network to obtain water flow clustering pattern feature vectors; a key node water flow extraction module 150, used to extract the water flow time-series associated feature vectors corresponding to the key nodes to be analyzed from the set of water flow time-series associated feature vectors as query feature vectors; a similarity calculation module 160, used to calculate the hash similarity between the query feature vectors and the water flow clustering pattern feature vectors; and a leak node judgment module 170, used to determine whether the key node to be analyzed is a leak node based on the comparison between the hash similarity and a predetermined threshold.

[0022] In this embodiment, the water flow acquisition module 110 is used to acquire a set of time series data of water flow collected by water flow sensors deployed at various key nodes of the pipeline network system. It should be understood that water flow data from key nodes is crucial for intelligent water management. Specifically, the time series data of water flow collected by the water flow sensors at each key node reflects the changes in water flow in the pipeline network at different times, including the flow differences between peak and off-peak periods. Therefore, in order to achieve comprehensive monitoring and refined management of the water supply network, the technical solution of this application acquires a set of time series data of water flow collected by water flow sensors deployed at various key nodes of the pipeline network system, and performs time-series analysis and processing on the set of water flow time series data. This allows for the timely detection and location of leaks; that is, an abnormal decrease in flow may indicate a leak in the pipeline network. This reduces water waste and economic losses, and prevents water supply problems from affecting residents' lives and production activities.

[0023] In this embodiment, the water flow time series arrangement module 120 is used to arrange the time series of each water flow in the set of water flow time series according to the time dimension to obtain a set of water flow time series input vectors. Accordingly, considering that each water flow in the set of water flow has time-dimensional characteristic information, meaning that subsequent water flow data points are often related to previous water flow data points, in order to retain the characteristic information of water flow changes over time at each key node and ensure that the time-series change trend and pattern of water flow at each key node can be subsequently mined, in the technical solution of this application, the time series of each water flow in the set of water flow time series is arranged according to the time dimension to obtain a set of water flow time series input vectors. This allows for a more accurate analysis of the trend and structural pattern of water flow changes at each key node over time, thereby effectively improving the reliability of water system management.

[0024] In this embodiment, the water flow time-series pattern feature extraction module 130 is used to extract water flow time-series pattern features from each water flow time-series input vector in the set of water flow time-series input vectors to obtain a set of water flow time-series associated feature vectors. Specifically, in this embodiment, the water flow time-series pattern feature extraction module is used to: pass each water flow time-series input vector in the set of water flow time-series input vectors through a water flow time-series pattern feature extractor based on a deep neural layer to obtain the set of water flow time-series associated feature vectors, wherein the water flow time-series pattern feature extractor based on a deep neural layer is a water flow time-series pattern feature extractor based on a one-dimensional convolutional layer. It should be understood that each water flow time-series input vector in the set of water flow time-series input vectors has time-series pattern features and patterns, such as peak values, valley values, or specific fluctuation patterns of water flow at each key node. A one-dimensional convolutional layer can identify local time-series patterns and features in time-series data and can effectively capture and perceive them from time-series data. Therefore, in the technical solution of this application, each water flow time-series input vector in the set of water flow time-series input vectors is respectively passed through a water flow time-series pattern feature extractor based on a one-dimensional convolutional layer to capture and mine the local time-series feature information about water flow in each water flow time-series input vector, thereby obtaining a set of water flow time-series associated feature vectors containing richer water flow time-series features.

[0025] In this embodiment, the water flow clustering module 140 is used to obtain a water flow clustering pattern feature vector by passing the set of water flow time-series associated feature vectors through a joint clustering analysis network. Accordingly, considering the potential correlations and hidden pattern information among the various water flow time-series associated feature vectors in the set—that is, the various water flow time-series associated feature vectors exhibit similar water flow time-series characteristics and patterns—this application integrates the water flow time-series features of each key node and mines the overall pattern features in the set of water flow time-series associated feature vectors. Therefore, in order to integrate the water flow time-series features of each key node and mine the overall pattern features in the set of water flow time-series associated feature vectors, the technical solution of this application uses a joint clustering analysis network to obtain a water flow clustering pattern feature vector. It should be understood that by clustering the set of water flow time-series correlation feature vectors through the joint clustering analysis network, it is possible to capture the similar water flow time-series correlation features of the set of water flow time-series correlation feature vectors, and aggregate these similar feature information to reveal the commonalities and variation patterns of water flow in the time-series feature information of the set of water flow time-series correlation feature vectors, forming a water flow clustering pattern feature vector with the most commonality and overall representativeness, which provides help and support for subsequent analysis and processing.

[0026] Figure 3 This is a block diagram of a water flow clustering module in a smart water management system based on big data technology and the Internet of Things, according to an embodiment of this application. Specifically, in an embodiment of this application, as... Figure 3 As shown, the water flow clustering module 140 includes: an adjacency matrix and degree matrix construction unit 141, used to construct an adjacency matrix and a degree matrix of the set of water flow time-series associated feature vectors; a Laplacian matrix calculation unit 142, used to calculate a Laplacian matrix based on the adjacency matrix and the degree matrix; a normalization processing unit 143, used to normalize the Laplacian matrix to obtain a normalized Laplacian matrix; an eigenvalue calculation unit 144, used to arrange the eigenvalues ​​of the normalized Laplacian matrix from largest to smallest, and extract the first K eigenvalues ​​to calculate the eigenvectors of the first K eigenvalues; and a feature vector normalization unit 145, used to normalize the eigenvectors of the first K eigenvalues ​​and combine the normalized eigenvectors of the first K eigenvalues ​​into a fused feature vector to obtain the water flow clustering pattern feature vector.

[0027] Specifically, in this embodiment, the adjacency matrix degree matrix construction unit is used to: calculate the association weight values ​​between each water flow time-series associated feature vector in the set of water flow time-series associated feature vectors using the following weight formula to obtain the adjacency matrix; wherein, the weight formula is:

[0028]

[0029] Among them, V i and V j Let be the i-th and j-th water flow time-series correlated feature vectors in the set of water flow time-series correlated feature vectors, respectively; σ be the variance between the i-th and j-th water flow time-series correlated feature vectors; exp(·) denotes an exponential function with the natural constant e as the base; W i,j Let be the eigenvalue at position (i,j) in the adjacency matrix, ||·||2 2 This represents the square of the L2 norm of a vector.

[0030] More specifically, in this embodiment, the Laplace matrix calculation unit is used to: calculate the Laplace matrix based on the adjacency matrix and the degree matrix using the following Laplace formula; wherein the Laplace formula is:

[0031] L=DW

[0032] Where W is the adjacency matrix, D is the degree matrix, and L is the Laplace matrix.

[0033] More specifically, in this embodiment, the standardization processing unit is configured to: standardize the Laplace matrix using the following standardization formula to obtain the standardized Laplace matrix; wherein the standardization formula is:

[0034]

[0035] Where W is the adjacency matrix, D is the degree matrix, I is the identity matrix, and L is the degree matrix. norm Let be the standardized Laplace matrix.

[0036] In this embodiment, the key node water flow extraction module 150 is used to extract the water flow time-series correlation feature vector corresponding to the key node to be analyzed from the set of water flow time-series correlation feature vectors as query feature vectors. It should be understood that this is to further perform centralized analysis and judgment on each key node among all key nodes, so that when a problem occurs, the problem node can be accurately located, allowing for rapid identification of such anomalies and timely implementation of measures. In the technical solution of this application, the water flow time-series correlation feature vector corresponding to the key node to be analyzed is extracted from the set of water flow time-series correlation feature vectors as query feature vectors. It should be understood that by extracting the water flow time-series correlation feature vector corresponding to the key node to be analyzed, the water flow pattern of that node can be analyzed in a targeted manner to understand its water flow changes over different time periods. Specifically, if there is a leak in the key node to be analyzed, its flow pattern may exhibit anomalies. This allows for rapid identification of the key node, thereby enabling advance scheduling of maintenance work, reducing unexpected downtime, and thus allowing for more refined management of the water supply network, improving system operating efficiency and reliability.

[0037] In this embodiment of the application, the similarity calculation module 160 is used to calculate the hash similarity between the query feature vector and the water flow clustering pattern feature vector. Figure 4 This is a block diagram of a similarity calculation module in a smart water management system based on big data technology and the Internet of Things, according to an embodiment of this application. Specifically, in an embodiment of this application, as... Figure 4 As shown, the similarity calculation module 160 includes: a feature optimization unit 161, used to optimize the query feature vector and the water flow clustering pattern feature vector to obtain an optimized query feature vector and an optimized water flow clustering pattern feature vector; and a hash similarity calculation unit 162, used to calculate the hash similarity between the optimized query feature vector and the optimized water flow clustering pattern feature vector.

[0038] Specifically, the feature optimization unit 161 is used to optimize the query feature vector and the water flow clustering pattern feature vector to obtain optimized query feature vector and optimized water flow clustering pattern feature vector. It should be understood that in the above-described technical solution, each water flow time-series associated feature vector in the set of water flow time-series associated feature vectors expresses the time-series associated semantic features of the water flow of the corresponding key node. Therefore, after passing the set of water flow time-series associated feature vectors through a joint clustering analysis network, the water flow clustering pattern feature vector can obtain the set water flow clustering time-series semantic feature representation of the key node set based on the joint clustering distribution of the time-series associated semantic features of the water flow of each key node.

[0039] Thus, when calculating the hash similarity between the query feature vector and the water flow clustering pattern feature vector, and mapping the query feature vector and the water flow clustering pattern feature vector to the hash similarity probability space respectively, it is desirable to improve the deterministic probability mapping of the distinguishable feature distributions of the query feature vector and the water flow clustering pattern feature vector, thereby improving the accuracy of hash similarity calculation. Based on this, in the technical solution of this application, feature optimization is performed on the query feature vector and the water flow clustering pattern feature vector to obtain optimized query feature vectors and optimized water flow clustering pattern feature vectors.

[0040] Specifically, in this embodiment, the feature optimization unit includes: a weighting coefficient calculation subunit, used to calculate the weighting coefficients of the query feature vector and the water flow clustering pattern feature vector to obtain query weighting coefficients and water flow weighting coefficients; and a weighting optimization subunit, used to perform weighted optimization on the query feature vector and the water flow clustering pattern feature vector using the query weighting coefficients and the water flow weighting coefficients respectively to obtain the optimized query feature vector and the optimized water flow clustering pattern feature vector.

[0041] More specifically, in this embodiment of the application, the weighting coefficient calculation subunit is used to: calculate the weighting coefficients of the query feature vector and the water flow clustering pattern feature vector using the following weighting formula to obtain the query weighting coefficient and the water flow weighting coefficient; wherein, the weighting formula is:

[0042]

[0043] Among them, v 1-max and v 2-max These are the maximum eigenvalues ​​of the query feature vector V1 and the water flow clustering pattern feature vector V2, respectively. 1-j and v 2-j is the j-th feature value of the query feature vector V1 and the water flow clustering pattern feature vector V2, L is the length of the feature vector, log represents the logarithm to the base 2, and α is the weight hyperparameter, ω1 and ω2 represent the query weighting coefficient and the water flow weighting coefficient, respectively, and exp(·) represents the exponential function to the base e.

[0044] Specifically, by using the majority voting mechanism in the crowdsourcing form of water flow temporal semantic association features of the query feature vector and the water flow clustering pattern feature vector respectively, the individual information expectation of each feature vector is maximized under the distribution subordinate probability model of the query feature vector V1 and the water flow clustering pattern feature vector V2 relative to the hash similarity probability space. Then, by weighting the query feature vector V1 and the water flow clustering pattern feature vector V2 with coefficients ω1 and ω2 to optimize the query feature vector V1 and the water flow clustering pattern feature vector V2, the distribution of the query feature vector V1 and the water flow clustering pattern feature vector V2 can be jointly estimated relative to the hash similarity probability space based on the average information confidence of the feature values ​​of the query feature vector V1 and the water flow clustering pattern feature vector V2. This improves the determinism of their probability mapping to the hash similarity probability space while maintaining the distinguishability of the feature distribution of the query feature vector V1 and the water flow clustering pattern feature vector V2, thereby improving the accuracy of hash similarity calculation. In this way, the system can automatically identify potential leak points, improving the accuracy and response speed of leak detection, while reducing the need for manual inspections, lowering operating costs, and thus improving the reliability and stability of the entire water system.

[0045] Specifically, the hash similarity calculation unit 162 is used to calculate the hash similarity between the optimized query feature vector and the optimized water flow clustering pattern feature vector. Correspondingly, considering that the optimized query feature vector represents the time-series correlation feature information of the water flow of the node to be analyzed, the optimized water flow clustering pattern feature vector represents the time-series feature module and / or hardware module of water flow in normal mode, which is integrated into the wireless terminal. For example, the intelligent water management system 100 based on big data technology and the Internet of Things can be a software module in the operating system of the wireless terminal, or it can be an application developed for the wireless terminal; of course, the intelligent water management system 100 based on big data technology and the Internet of Things can also be one of many hardware modules of the wireless terminal.

[0046] Alternatively, in another example, the smart water management system 100 based on big data technology and the Internet of Things can also be a separate device from the wireless terminal, and the smart water management system 100 based on big data technology and the Internet of Things can be connected to the wireless terminal via wired and / or wireless networks, and transmit interactive information in accordance with an agreed data format.

[0047] Various implementations of this disclosure have been described above, and the foregoing description is exemplary and not exhaustive. Furthermore, it is not limited to the disclosed implementations, and many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described implementations. The terminology used herein is chosen to best explain the principles, practical applications, or improvements to technology in the market, or to enable others skilled in the art to understand the various practical methods disclosed herein.

Claims

1. A smart water management system based on big data technology and the Internet of Things, characterized in that, include: The water flow acquisition module is used to acquire a collection of time series data of water flow collected by water flow sensors deployed at various key nodes of the pipeline network system. The water flow time series arrangement module is used to arrange the time series of each water flow in the set of water flow time series according to the time dimension to form a water flow time series input vector to obtain a set of water flow time series input vectors; The water flow time-series pattern feature extraction module is used to extract water flow time-series pattern features from each water flow time-series input vector in the set of water flow time-series input vectors to obtain a set of water flow time-series associated feature vectors; The water flow clustering module is used to obtain water flow clustering pattern feature vectors by passing the set of water flow time-series associated feature vectors through a joint clustering analysis network. The key node water flow extraction module is used to extract the water flow time-series correlation feature vector corresponding to the key node to be analyzed as the query feature vector from the set of water flow time-series correlation feature vectors. The similarity calculation module is used to calculate the hash similarity between the query feature vector and the water flow clustering pattern feature vector; The leakage node determination module is used to determine whether the key node to be analyzed is a leakage node based on the comparison between the hash similarity and a predetermined threshold.

2. The intelligent water management system based on big data technology and the Internet of Things as described in claim 1, characterized in that, The water flow time-series pattern feature extraction module is used to: pass each water flow time-series input vector in the set of water flow time-series input vectors through a water flow time-series pattern feature extractor based on a deep neural layer to obtain the set of water flow time-series associated feature vectors.

3. The intelligent water management system based on big data technology and the Internet of Things as described in claim 2, characterized in that, The water flow temporal pattern feature extractor based on deep neural layers is a water flow temporal pattern feature extractor based on one-dimensional convolutional layers.

4. The intelligent water management system based on big data technology and the Internet of Things as described in claim 3, characterized in that, The water flow clustering module includes: The adjacency matrix and degree matrix construction unit is used to construct the adjacency matrix and degree matrix of the set of time-series correlated feature vectors of water flow. A Laplacian matrix calculation unit is used to calculate the Laplacian matrix based on the adjacency matrix and the degree matrix; A standardization processing unit is used to standardize the Laplacian matrix to obtain a standardized Laplacian matrix. The eigenvalue calculation unit is used to arrange the eigenvalues ​​of the standardized Laplacian matrix from largest to smallest, and extract the first K eigenvalues ​​to calculate the eigenvector of the first K eigenvalues. The feature vector standardization unit is used to standardize the feature vectors of the first K feature values ​​and combine the standardized feature vectors of the first K feature values ​​into a fused feature vector to obtain the water flow clustering pattern feature vector.

5. The intelligent water management system based on big data technology and the Internet of Things according to claim 4, characterized in that, The adjacency matrix degree matrix construction unit is used to: calculate the association weight value between each water flow time-series associated feature vector in the set of water flow time-series associated feature vectors using the following weight formula to obtain the adjacency matrix; The weighting formula is as follows: Among them, V i and V j Let be the i-th and j-th water flow time-series correlated feature vectors in the set of water flow time-series correlated feature vectors, respectively; σ be the variance between the i-th and j-th water flow time-series correlated feature vectors; exp(·) denotes an exponential function with the natural constant e as the base; W i,j Let be the eigenvalue at position (i,j) in the adjacency matrix, ||·||2 2 This represents the square of the L2 norm of a vector.

6. The intelligent water management system based on big data technology and the Internet of Things according to claim 5, characterized in that, The Laplace matrix calculation unit is used to: calculate the Laplace matrix based on the adjacency matrix and the degree matrix using the following Laplace formula; The Laplace formula is as follows: L=DW Where W is the adjacency matrix, D is the degree matrix, and L is the Laplace matrix.

7. The intelligent water management system based on big data technology and the Internet of Things as described in claim 6, characterized in that, The standardization processing unit is used to: standardize the Laplacian matrix using the following standardization formula to obtain the standardized Laplacian matrix; The standardized formula is as follows: Where W is the adjacency matrix, D is the degree matrix, I is the identity matrix, and L is the degree matrix. norm Let be the standardized Laplace matrix.

8. The intelligent water management system based on big data technology and the Internet of Things according to claim 7, characterized in that, The similarity calculation module includes: The feature optimization unit is used to optimize the query feature vector and the water flow clustering pattern feature vector to obtain the optimized query feature vector and the optimized water flow clustering pattern feature vector. The hash similarity calculation unit is used to calculate the hash similarity between the optimized query feature vector and the optimized water flow clustering pattern feature vector.

9. The intelligent water management system based on big data technology and the Internet of Things as described in claim 8, characterized in that, The feature optimization unit includes: The weighting coefficient calculation subunit is used to calculate the weighting coefficients of the query feature vector and the water flow clustering pattern feature vector to obtain the query weighting coefficient and the water flow weighting coefficient; The weighted optimization subunit is used to perform weighted optimization on the query feature vector and the water flow clustering pattern feature vector using the query weighting coefficient and the water flow weighting coefficient, respectively, to obtain the optimized query feature vector and the optimized water flow clustering pattern feature vector.

Images