A smart water management method and system based on the Internet of Things and artificial intelligence

Abstract

This invention discloses a smart water management method and system based on the Internet of Things (IoT) and artificial intelligence (AI), belonging to the field of water management. The method includes: determining the water management area, collecting water supply-related data, calculating a similarity coefficient, and constructing IoT units to manage water supply nodes; constructing an edge computing module for each IoT unit to fuse water supply-related data into fused data; setting a water supply regulation cycle, dividing data segments, and constructing a water delivery prediction model; assessing whether water supply nodes are malfunctioning, and using the IoT units to upload the malfunction assessment results and required water delivery volume to a smart water management platform. The system includes IoT units and a smart water management platform, comprising an edge computing module, a sensor array installed on the water supply nodes, an IoT communication module, and a processor. This invention achieves closed-loop management of "perception-fusion-prediction-decision-scheduling," effectively reducing the complexity and difficulty of water management.

Description

Technical Field

[0001] This invention relates to the field of smart water management, specifically to a smart water management method and system based on the Internet of Things and artificial intelligence. Background Technology

[0002] Smart water management is characterized by three key features: water management awareness, water management operations, and the intelligence of water management personnel. It holds promise as a crucial approach to solving urban water resource problems and has become an essential path for the development of traditional water utilities. As part of smart cities, smart water management provides public services such as water supply and drainage, flood control and drainage, water pollution control and environmental protection, and disaster prevention and mitigation for urban development. Simultaneously, it enhances the efficiency and quality of water utilities' work, manages and promptly handles various emergency water incidents, improves service levels and customer satisfaction, and provides service guarantees.

[0003] Existing smart water management systems suffer from drawbacks in managing urban water supply systems. They rely on a single data format and cannot integrate and manage multiple data types, leading to inaccurate water supply control and an inability to respond to fluctuating regional water demands. Furthermore, the various stages of water supply management and control are slow to react, often resulting in delayed and inaccurate fault assessments when dealing with malfunctions at water supply nodes. Summary of the Invention

[0004] In view of the above-mentioned shortcomings in the prior art, the present invention provides a smart water management method and system based on the Internet of Things and artificial intelligence.

[0005] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows:

[0006] A smart water management method based on the Internet of Things and artificial intelligence is provided, which includes:

[0007] Step S1: Determine the area for water management, collect time-series data on water supply flow, water pressure, water quality, and meteorological data for each water supply node in the water supply network within the area, calculate the similarity coefficient of the water supply process between water supply nodes, and construct an Internet of Things (IoT) unit for managing water supply nodes based on the location data between water supply nodes.

[0008] Step S2: Construct an edge computing module for each IoT unit, collect water flow, water pressure, water quality data and meteorological data of each water supply node in the IoT unit at different times, and fuse the water flow, water pressure, water quality data and meteorological data to obtain time-series based fused data of the water supply node.

[0009] Step S3: Set the water supply regulation cycle, divide the time series-based fusion data into different data segments according to the water supply regulation cycle, collect the cumulative water transmission volume of each data segment's water supply node, construct a water transmission volume prediction model, and fit the water transmission volume prediction model using the fusion data and water transmission volume within the data segment to obtain the fitted water transmission volume prediction model.

[0010] Step S4: Using the fitted water delivery prediction model, output the predicted water delivery and demand water delivery of the water supply node in different water supply regulation cycles, assess whether the water supply node has failed, and use the Internet of Things unit to upload the failure assessment results and demand water delivery to the smart water management platform.

[0011] Further, step S1 includes:

[0012] Step S11: Determine the area for water management and collect water flow, water pressure, water quality data and meteorological data of each water supply node in the water supply network within the area; install GPS positioning sensors on each water supply node and collect positioning data of each water supply node;

[0013] Step S12: Construct time-series-based water supply flow data Water supply pressure data Water quality data and meteorological data Where i is the water supply node number, t n Where n is the data acquisition time, and n is the data number. They are time t respectively n The collected data include water supply flow rate, water supply pressure, water quality, and meteorological data.

[0014] Step S13: Calculate the similarity coefficient of the water supply process between water supply nodes based on the water supply flow, water supply pressure, water quality data and meteorological data collected from different water supply nodes at the same time.

[0015] ;

[0016] Where u is the data number. Let i be the water supply node at time t. u The collected water supply flow rate, water supply pressure, water quality data, and meteorological data, These represent the water supply nodes i-1 at time t. u The collected water supply flow rate, water supply pressure, water quality data, and meteorological data, These are the threshold values ​​for differences in water supply flow rate, water supply pressure, water quality data, and meteorological data, respectively. The influence weights of water supply flow, water supply pressure, water quality data, and meteorological data are respectively. Let be the similarity coefficient of the water supply process between water supply node i and water supply node i-1;

[0017] Step S14: Set reference values ​​for similarity assessment ;

[0018] like If so, it is determined that the water supply process between water supply node i and water supply node i-1 is similar;

[0019] like If so, it is determined that the water supply processes between water supply node i and water supply node i-1 are not similar;

[0020] Step S15: Obtain the location data corresponding to two water supply nodes i and i-1 with similar water supply processes. Calculate the distance between two water supply nodes i and i-1 with similar water supply processes. ;

[0021] ;

[0022] in, This is a distance calculation function based on location data;

[0023] Step S16: Set the distance threshold between water supply nodes ;

[0024] like Then, water supply node i and water supply node i-1 are classified as the same Internet of Things unit;

[0025] like Then, water supply node i and water supply node i-1 are divided into different IoT units;

[0026] Get all water supply nodes managed in the same IoT unit.

[0027] Further, step S2 includes:

[0028] Step S21: Construct an edge computing module for each IoT unit to collect the water supply flow rate of each water supply node in the IoT unit at different times k. Water supply pressure Water quality data and meteorological data ;

[0029] Step S22: Construct a state prediction equation that integrates water supply flow, water supply pressure, water quality data, and meteorological data;

[0030] ;

[0031] in, Let k be the state vector at time k. , Let A be the state vector at time k-1, and A be the state transition matrix. The water supply nodes within the IoT unit maintain stable water supply over a short period. , Let k be the control vector at time k-1. For the control matrix, The process noise at time k-1 is... Follows a normal distribution , Let be the covariance matrix of the process noise;

[0032] Step S23: Based on the collected water supply flow rate Water supply pressure Water quality data and meteorological data Construct observation vectors And construct observation equations based on the observation vectors;

[0033] ;

[0034] in, For the observation matrix, The observation noise for data collection at the water supply node. Follows a normal distribution R is the observation noise covariance matrix of the data collected at the water supply node;

[0035] Step S24: Using the observation matrix Calculate the Kalman gain;

[0036] ;

[0037] in, Let Kalman gain be at time k. Let k be the state prediction error covariance matrix at time k;

[0038] Step S25: Utilize Kalman gain and state prediction error covariance matrix The state prediction equation is modified to output fused data. ;

[0039] ;

[0040] in, It is the identity matrix. Let k be the state prediction matrix based on time k-1. Let k be the state update error covariance matrix at time k;

[0041] Step S26: Obtain time-series-based fused data for each water supply node in the IoT unit. , For water supply node i at time... The fused data.

[0042] Further, step S3 includes:

[0043] Step S31: Set the water supply regulation cycle According to the water supply adjustment cycle Data fusion Divide the data into N consecutive segments of equal size. And satisfy , For the m-th fused data segment, To integrate data At the corresponding time, e is the data segment number;

[0044] Step S32: Collect the time period corresponding to each data segment. The accumulated water transfer volume W at internal water supply node i is used to obtain N consecutive water transfer volume data. , For the Nth water transfer volume, construct a water transfer volume prediction model for the water supply node;

[0045] ;

[0046] in, Let u be the coefficient of the fused data in the data segment. This is the bias of the water transfer prediction model;

[0047] Step S33: Divide the different data segments and the corresponding water transfer volume As fitting data points N fitted data points are obtained, and these N fitted data points are input into the water flow prediction model to construct a fitted data matrix. and output vector , This is the Nth data segment;

[0048] Step S34: Using the fitted data matrix and output vector Calculate the coefficient vector of the water transfer prediction model ;

[0049] ;

[0050] Step S35: Convert the coefficient vector The bias is calculated by substituting the data into the water transfer prediction model and using the fitted data points. Thus, a water transfer prediction model with a complete fit was obtained.

[0051] Further, step S4 includes:

[0052] Step S41: Obtain the water supply flow, water supply pressure, water quality data and meteorological data collected at the water supply nodes in the next water supply regulation cycle, input them into the fitted water transmission prediction model, and output the predicted water transmission volume of the water supply nodes in the next water supply regulation cycle.

[0053] Collect the cumulative actual water delivery volume of the water supply node within the next water supply regulation cycle, calculate the water delivery volume error between the predicted water delivery volume and the actual water delivery volume. If the water delivery volume error is greater than the error threshold, it is determined that the water supply node has malfunctioned; otherwise, it is determined that the water supply node is operating normally.

[0054] The IoT unit uploads the faulty water supply node information to the smart water management platform.

[0055] Step S42: The smart water management platform sets the water supply flow, water pressure, water quality data and forecast meteorological data for each water supply node in different future water supply regulation cycles, sends them to the Internet of Things unit, uses the water transmission volume prediction model to predict the required water transmission volume of the water supply node in different future water supply regulation cycles, and feeds it back to the smart water management platform.

[0056] A smart water management system based on the Internet of Things and artificial intelligence is provided, comprising:

[0057] The Internet of Things (IoT) unit includes an edge computing module, a sensor array installed on the water supply node, an IoT communication module, and a processor. The sensor array includes a flow sensor, a water pressure sensor, a water quality sensor, and a meteorological sensor. The sensor array and the processor are connected through the IoT communication module.

[0058] The edge computing module is built inside the processor. The edge computing module stores a computer-readable storage medium, which stores computer program instructions. When the computer program instructions are executed, they run the above-mentioned smart water management method based on the Internet of Things and artificial intelligence.

[0059] The smart water management platform communicates wirelessly with the IoT unit via a wireless communication module.

[0060] The beneficial effects of this invention are as follows: This invention utilizes water supply-related data collected by multiple types of sensors to filter water supply nodes within a small area, constructing an IoT unit for small-area management. By leveraging edge computing and artificial intelligence prediction, it assesses the water supply process of each node within the small area, including fault node assessment and water supply volume prediction. This effectively reduces the latency of water supply management at each node, ensures timely response to water supply faults at each node, and increases the accuracy of water supply volume prediction and management. This invention achieves closed-loop management of "perception-fusion-prediction-decision-scheduling," effectively reducing the complexity and difficulty of water management. Attached Figure Description

[0061] Figure 1 This is a flowchart of a smart water management method based on the Internet of Things and artificial intelligence. Detailed Implementation

[0062] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0063] like Figure 1 As shown, a smart water management method based on the Internet of Things and artificial intelligence includes:

[0064] Step S1: Determine the area for water management, collect time-series data on water supply flow, water pressure, water quality, and meteorological data for each water supply node in the water supply network within the area, calculate the similarity coefficient of the water supply process between water supply nodes, and construct an Internet of Things (IoT) unit for managing water supply nodes based on the location data between water supply nodes.

[0065] Step S1 specifically includes the following steps:

[0066] Step S11: Determine the area for water management and collect water flow, water pressure, water quality data and meteorological data of each water supply node in the water supply network within the area; install GPS positioning sensors on each water supply node and collect positioning data of each water supply node;

[0067] Step S12: Construct time-series-based water supply flow data Water supply pressure data Water quality data and meteorological data Where i is the water supply node number, t n Where n is the data acquisition time, and n is the data number. They are time t respectivelyn The collected data include water supply flow rate, water supply pressure, water quality, and meteorological data.

[0068] Step S13: Calculate the similarity coefficient of the water supply process between water supply nodes based on the water supply flow, water supply pressure, water quality data and meteorological data collected from different water supply nodes at the same time.

[0069] ;

[0070] Where u is the data number. Let i be the water supply node at time t. u The collected water supply flow rate, water supply pressure, water quality data, and meteorological data, These represent the water supply nodes i-1 at time t. u The collected water supply flow rate, water supply pressure, water quality data, and meteorological data, These are the threshold values ​​for differences in water supply flow rate, water supply pressure, water quality data, and meteorological data, respectively. The influence weights of water supply flow, water supply pressure, water quality data, and meteorological data are respectively. Let be the similarity coefficient of the water supply process between water supply node i and water supply node i-1;

[0071] The difference threshold provides a reference for comparing the similarity of each data point. By reasonably setting the difference threshold for each data point, the accuracy of similarity calculation between water supply nodes can be effectively improved. During the water supply process, the closer the values ​​of each data point are, the higher the similarity of the water supply process; the smaller the similarity coefficient, the higher the similarity, and vice versa. The influence weight represents the weight of different data points in the similarity assessment. When determining the specific value, the fluctuation range of each data point should be considered; the larger the fluctuation range, the larger the influence weight, and it should also satisfy the following condition: .

[0072] For two water supply nodes with a smaller similarity coefficient, their water supply processes are similar, and they are generally located on the same water supply pipeline or within a small water supply area. Within this area, the water supply-related data between the nodes changes relatively little. By constructing Internet of Things (IoT) units in this small area and equipping them with edge computing modules, the onboard artificial intelligence algorithms can predict whether faults will occur in the water supply process, thereby enabling water supply management in this small area. Water supply nodes are typically identified by the locations of control valves, pipe connection points, and pipe convergence points.

[0073] Step S14: Set reference values ​​for similarity assessment ;

[0074] like If so, it is determined that the water supply process between water supply node i and water supply node i-1 is similar;

[0075] like If so, it is determined that the water supply processes between water supply node i and water supply node i-1 are not similar;

[0076] Step S15: Obtain the location data corresponding to two water supply nodes i and i-1 with similar water supply processes. Calculate the distance between two water supply nodes i and i-1 with similar water supply processes. ;

[0077] ;

[0078] in, This is a distance calculation function based on location data;

[0079] Step S16: Set the distance threshold between water supply nodes ;

[0080] like Then, water supply node i and water supply node i-1 are classified as the same Internet of Things unit;

[0081] like Then, water supply node i and water supply node i-1 are divided into different IoT units;

[0082] Get all water supply nodes managed in the same IoT unit.

[0083] By utilizing the distance difference between two water supply nodes, the accuracy of IoT unit construction can be further improved, avoiding the accidental inclusion of two water supply nodes that are far apart into the same IoT unit, thereby avoiding increasing the difficulty of IoT unit management of regional water supply.

[0084] Step S2: Construct an edge computing module for each IoT unit, collect water flow, water pressure, water quality data and meteorological data of each water supply node in the IoT unit at different times, and fuse the water flow, water pressure, water quality data and meteorological data to obtain time-series fused data of the water supply node.

[0085] Step S2 specifically includes the following steps:

[0086] Step S21: Construct an edge computing module for each IoT unit to collect the water supply flow rate of each water supply node in the IoT unit at different times k. Water supply pressure Water quality data and meteorological data ;

[0087] Step S22: Construct a state prediction equation that integrates water supply flow, water supply pressure, water quality data, and meteorological data;

[0088] ;

[0089] in, Let k be the state vector at time k. , Let A be the state vector at time k-1, and A be the state transition matrix. The water supply nodes within the IoT unit maintain stable water supply over a short period. , Let k be the control vector at time k-1. For the control matrix, The process noise at time k-1 is... Follows a normal distribution , Let be the covariance matrix of the process noise;

[0090] The control vector represents the changes in water flow rate, water pressure, water quality data, and current meteorological data when control commands are sent from the IoT unit to different water supply nodes. Process noise represents the error between the IoT unit and the sensor array collecting data at the water supply nodes, including signal transmission and processing errors. The sensor array includes flow sensors, water pressure sensors, water quality sensors, and meteorological sensors that collect water flow rate, water pressure, water quality data, and meteorological data.

[0091] Step S23: Based on the collected water supply flow rate Water supply pressure Water quality data and meteorological data Construct observation vectors And construct observation equations based on the observation vectors;

[0092] ;

[0093] in, For the observation matrix, The observation noise for data collection at the water supply node. Follows a normal distribution R is the observation noise covariance matrix of the data collected at the water supply node. The observation noise is determined by the accuracy of the sensors that collect water supply flow, water supply pressure, water quality data and meteorological data. The observation noise is the data acquisition error of the sensors.

[0094] Step S24: Using the observation matrix Calculate the Kalman gain;

[0095] ;

[0096] in, Let Kalman gain be at time k. Let k be the state prediction error covariance matrix at time k;

[0097] Step S25: Utilize Kalman gain and state prediction error covariance matrix The state prediction equation is modified to output fused data. ;

[0098] ;

[0099] in, It is the identity matrix. Let k be the state prediction matrix based on time k-1. Let k be the state update error covariance matrix at time k;

[0100] Step S26: Obtain time-series-based fused data for each water supply node in the IoT unit. , For water supply node i at time... The fused data.

[0101] Step S3: Set the water supply regulation cycle, divide the time series-based fused data into different data segments according to the water supply regulation cycle, collect the cumulative water transmission volume of each data segment's water supply node, construct a water transmission volume prediction model, and fit the water transmission volume prediction model using the fused data and water transmission volume within the data segment to obtain the fitted water transmission volume prediction model.

[0102] Step S3 specifically includes the following steps:

[0103] Step S31: Set the water supply regulation cycle According to the water supply adjustment cycle Data fusion Divide the data into N consecutive segments of equal size. And satisfy , For the m-th fused data segment, To integrate data At the corresponding time, e is the data segment number;

[0104] Step S32: Collect the time period corresponding to each data segment. The accumulated water transfer volume W at internal water supply node i is used to obtain N consecutive water transfer volume data. , For the Nth water transfer volume, construct a water transfer volume prediction model for the water supply node;

[0105] ;

[0106] in, Let u be the coefficient of the fused data in the data segment. This is the bias of the water transfer prediction model;

[0107] Step S33: Divide the different data segments and the corresponding water transfer volume As fitting data points N fitted data points are obtained, and these N fitted data points are input into the water flow prediction model to construct a fitted data matrix. and output vector , This is the Nth data segment;

[0108] Step S34: Using the fitted data matrix and output vector Calculate the coefficient vector of the water transfer prediction model ;

[0109] ;

[0110] Step S35: Convert the coefficient vector The bias is calculated by substituting the data into the water transfer prediction model and using the fitted data points. Thus, a water transfer prediction model with a complete fit was obtained.

[0111] coefficient vector It includes the coefficients corresponding to each fused data point within the data segment in the water transfer prediction model. The water transfer prediction model, completed through fitting, can predict the water transfer volume in different future time periods, providing an assessment basis for the health of the water supply status in a small area. Simultaneously, by setting water flow rate, water pressure, water quality data, and forecasted meteorological data for future time periods, it accurately estimates the water demand in the region, providing a reference for water transfer within the region.

[0112] Step S4: Using the fitted water delivery prediction model, output the predicted water delivery and demand water delivery of the water supply node in different water supply regulation cycles, assess whether the water supply node has failed, and use the Internet of Things unit to upload the failure assessment results and demand water delivery to the smart water management platform.

[0113] Step S4 specifically includes the following steps:

[0114] Step S41: Obtain the water supply flow, water supply pressure, water quality data and meteorological data collected at the water supply nodes in the next water supply regulation cycle, input them into the fitted water transmission prediction model, and output the predicted water transmission volume of the water supply nodes in the next water supply regulation cycle.

[0115] Collect the cumulative actual water delivery volume of the water supply node within the next water supply regulation cycle, calculate the water delivery volume error between the predicted water delivery volume and the actual water delivery volume. If the water delivery volume error is greater than the error threshold, it is determined that the water supply node has malfunctioned; otherwise, it is determined that the water supply node is operating normally.

[0116] The IoT unit uploads the faulty water supply node information to the smart water management platform.

[0117] Step S42: The smart water management platform sets the water supply flow, water pressure, water quality data and forecast meteorological data for each water supply node in different future water supply regulation cycles, sends them to the Internet of Things unit, uses the water transmission volume prediction model to predict the required water transmission volume of the water supply node in different future water supply regulation cycles, and feeds it back to the smart water management platform.

[0118] A smart water management system based on the Internet of Things and artificial intelligence includes:

[0119] The Internet of Things (IoT) unit includes an edge computing module, a sensor array installed on the water supply node, an IoT communication module, and a processor. The sensor array includes a flow sensor, a water pressure sensor, a water quality sensor, and a meteorological sensor. The sensor array and the processor are connected through the IoT communication module.

[0120] The edge computing module is built inside the processor. The edge computing module stores computer-readable storage media, which stores computer program instructions. When the computer program instructions are executed, they run the aforementioned smart water management method based on the Internet of Things and artificial intelligence.

[0121] The smart water management platform communicates wirelessly with the IoT unit via a wireless communication module, which can use 4G wireless communication.

Claims

1. A smart water management method based on the Internet of Things and artificial intelligence, characterized in that, include: Step S1: Determine the area for water management, collect time-series data on water supply flow, water pressure, water quality, and meteorological data for each water supply node in the water supply network within the area, calculate the similarity coefficient of the water supply process between water supply nodes, and construct an Internet of Things (IoT) unit for managing water supply nodes based on the location data between water supply nodes. Step S2: Construct an edge computing module for each IoT unit, collect water flow, water pressure, water quality data and meteorological data of each water supply node in the IoT unit at different times, and fuse the water flow, water pressure, water quality data and meteorological data to obtain time-series fused data of the water supply node. Step S3: Set the water supply regulation cycle, divide the time series-based fusion data into different data segments according to the water supply regulation cycle, collect the cumulative water transmission volume of each data segment water supply node, construct a water transmission volume prediction model, and use the fusion data and water transmission volume in the data segment to fit the water transmission volume prediction model to obtain the fitted water transmission volume prediction model. Step S4: Using the fitted water delivery prediction model, output the predicted water delivery and demand water delivery of the water supply node in different water supply regulation cycles, assess whether the water supply node has failed, and use the Internet of Things unit to upload the failure assessment results and demand water delivery to the smart water management platform. Step S1 includes: Step S11: Determine the area for water management and collect water flow, water pressure, water quality data and meteorological data of each water supply node in the water supply network within the area; install GPS positioning sensors on each water supply node and collect positioning data of each water supply node; Step S12: Construct time-series-based water supply flow data Water supply pressure data Water quality data and meteorological data Where i is the water supply node number, t n Where n is the data acquisition time, and n is the data number. They are time t respectively n The collected data include water supply flow rate, water supply pressure, water quality, and meteorological data. Step S13: Calculate the similarity coefficient of the water supply process between water supply nodes based on the water supply flow, water supply pressure, water quality data and meteorological data collected from different water supply nodes at the same time. ； Where u is the data number. Let i be the water supply node at time t. u The collected water supply flow rate, water supply pressure, water quality data, and meteorological data, These represent the water supply nodes i-1 at time t. u The collected water supply flow rate, water supply pressure, water quality data, and meteorological data, These are the threshold values ​​for differences in water supply flow rate, water supply pressure, water quality data, and meteorological data, respectively. The influence weights of water supply flow, water supply pressure, water quality data, and meteorological data are respectively. Let be the similarity coefficient of the water supply process between water supply node i and water supply node i-1; Step S14: Set reference values ​​for similarity assessment ; like If so, it is determined that the water supply process between water supply node i and water supply node i-1 is similar; like If so, it is determined that the water supply processes between water supply node i and water supply node i-1 are not similar; Step S15: Obtain the locations of two water supply nodes i and i-1 with similar water supply processes. data Calculate the distance between two water supply nodes i and i-1 with similar water supply processes. ; ； in, This is a distance calculation function based on location data; Step S16: Set the distance threshold between water supply nodes ; like Then, water supply node i and water supply node i-1 are classified as the same Internet of Things unit; like Then, water supply node i and water supply node i-1 are divided into different IoT units; Get all water supply nodes managed in the same IoT unit.

2. The smart water management method based on the Internet of Things and artificial intelligence according to claim 1, characterized in that, Step S2 includes: Step S21: Construct an edge computing module for each IoT unit to collect the water supply flow rate of each water supply node in the IoT unit at different times k. Water supply pressure Water quality data and meteorological data ; Step S22: Construct a state prediction equation that integrates water supply flow, water supply pressure, water quality data, and meteorological data; ； in, Let k be the state vector at time k. , Let A be the state vector at time k-1, and A be the state transition matrix. The water supply nodes within the IoT unit maintain stable water supply over a short period. , Let k be the control vector at time k-1. For the control matrix, For the process noise at time k-1, the process noise is... Follows a normal distribution , Let be the covariance matrix of the process noise; Step S23: Based on the collected water supply flow rate Water supply pressure Water quality data and meteorological data Construct observation vectors And construct observation equations based on the observation vectors; ； in, For the observation matrix, The observation noise for data collection at the water supply node. Follows a normal distribution R is the observation noise covariance matrix of the data collected at the water supply node; Step S24: Using the observation matrix Calculate the Kalman gain; ； in, Let Kalman gain be at time k. Let k be the state prediction error covariance matrix at time k; Step S25: Utilize Kalman gain and state prediction error covariance matrix The state prediction equation is modified to output fused data. ; ； in, It is the identity matrix. Let k be the state prediction matrix based on time k-1. Let k be the state update error covariance matrix at time k; Step S26: Obtain time-series-based fused data for each water supply node in the IoT unit. , For water supply node i at time... The fused data.

3. The smart water management method based on the Internet of Things and artificial intelligence according to claim 2, characterized in that, Step S3 includes: Step S31: Set the water supply regulation cycle According to the water supply adjustment cycle Data fusion Divide the data into N consecutive segments of equal size. And satisfy , For the m-th fused data segment, To integrate data At the corresponding time, e is the data segment number; Step S32: Collect the time period corresponding to each data segment. The accumulated water transfer volume W at internal water supply node i is used to obtain N consecutive water transfer volume data. , For the Nth water transfer volume, construct a water transfer volume prediction model for the water supply node; ； in, Let u be the coefficient of the fused data in the data segment. For the bias of the water transfer prediction model Place; Step S33: Divide the different data segments and the corresponding water transfer volume As fitting data points , N fitted data points are obtained, and these N fitted data points are input into the water flow prediction model to construct a fitted data matrix. and output vector , This is the Nth data segment; Step S34: Using the fitted data matrix and output vector Calculate the coefficient vector of the water transfer prediction model ; ； Step S35: Convert the coefficient vector The bias is calculated by substituting the data into the water transfer prediction model and using the fitted data points. Thus, a water transfer prediction model with a complete fit was obtained.

4. The smart water management method based on the Internet of Things and artificial intelligence according to claim 3, characterized in that, Step S4 includes: Step S41: Obtain the water supply flow, water supply pressure, water quality data and meteorological data collected at the water supply nodes in the next water supply regulation cycle, input them into the fitted water transmission prediction model, and output the predicted water transmission volume of the water supply nodes in the next water supply regulation cycle. Collect the cumulative actual water delivery volume of the water supply node within the next water supply regulation cycle, calculate the water delivery volume error between the predicted water delivery volume and the actual water delivery volume. If the water delivery volume error is greater than the error threshold, it is determined that the water supply node has malfunctioned; otherwise, it is determined that the water supply node is operating normally. The IoT unit uploads the faulty water supply node information to the smart water management platform; Step S42: The smart water management platform sets the water supply flow, water pressure, water quality data and forecast meteorological data for each water supply node in different future water supply regulation cycles, sends them to the Internet of Things unit, uses the water transmission volume prediction model to predict the required water transmission volume of the water supply node in different future water supply regulation cycles, and feeds it back to the smart water management platform.

5. A smart water management system based on the Internet of Things and artificial intelligence, characterized in that, include: The Internet of Things (IoT) unit includes an edge computing module, a sensor array installed on a water supply node, an IoT communication module, and a processor. The sensor array includes a flow sensor, a water pressure sensor, a water quality sensor, and a meteorological sensor. The sensor array and the processor are connected via the IoT communication module. The edge computing module is built within the processor, and the edge computing module stores a computer-readable storage medium. The computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed, they run the smart water management method based on the Internet of Things and artificial intelligence as described in any one of claims 1-4. The smart water management platform communicates wirelessly with the IoT unit via a wireless communication module.

Images