A rapid analysis method for urban water supply network damage accidents

Abstract

The present invention discloses a method for rapid analysis of urban water supply pipe network damage accidents, including statistical analysis of pipe network water pressure monitoring data change characteristics, full-condition pipe network damage simulation, generation of pressure sample set, establishment of monitoring point perception node set based on pipe network hydraulic analysis, damage accident identification and damage accident location. By establishing a perception node set, the model drive and data drive are combined, and the hydraulic characteristics of the pipe network and the fluctuation law of the user's water use are considered at the same time, so that the prediction result is more accurate. At the same time, the pipe burst data samples generated by simulation are more and richer than previous samples, so that the prediction accuracy is higher. Moreover, the calculation rules adopted are relatively simple, the analysis and prediction speed is faster, and the efficiency is improved.

Description

Technical Field

[0001] The invention relates to an analysis method capable of quickly identifying and locating damage accidents of urban water supply pipe networks. Background Art

[0002] The water supply network is an important part of the city's lifeline project. The normal operation of the water supply network system is a prerequisite for ensuring the quality of life of urban residents and the production and operation of enterprises. At present, the construction of urban water supply networks is developing rapidly, and the scale and complexity of water supply networks are increasing, which poses new challenges to the safety monitoring and operation management of the networks. Water supply network damage accidents are one of the most common types of accidents in the water supply system. When the water supply network is damaged, the water pressure is released instantly, causing a large amount of water to gush out, resulting in a large amount of water resources being wasted and causing water pressure fluctuations in the water supply area; a large amount of water may flow to the streets and roads, causing traffic congestion and even causing personal safety problems; closing valves and valve cutting operations during emergency repairs will also seriously affect the normal function of the network system and affect residents' lives and corporate production. Therefore, quickly and efficiently identifying and locating network damage accidents can effectively control the adverse effects of network damage accidents, which is of great significance for improving the monitoring and management level of water supply networks.

[0003] The traditional pipeline damage analysis method is mainly based on the analysis of the fluctuation degree of pipeline pressure monitoring data. When the pressure monitoring values ​​of some monitoring points in the pipeline network fluctuate abnormally (the water pressure fluctuation exceeds the set threshold range), a preliminary analysis of the pipeline damage area is carried out according to the location of different monitoring points and the degree of water pressure fluctuation. Combined with manual inspection and other methods, the specific location of the pipeline damage is searched. The accuracy of pipeline damage identification is limited, and the efficiency of pipeline damage positioning is low.

[0004] Therefore, a rapid analysis method for urban water supply network damage accidents is needed. Combining the time difference of abnormal pressure fluctuations at different monitoring points and the difference in water pressure fluctuation values, combined with the analysis of the hydraulic characteristics of network damage in the network hydraulic model, the occurrence of network damage accidents can be quickly identified, and the location of the network damage can be effectively located. The establishment of the above method plays an important technical support role in strengthening the maintenance and management of water supply networks, establishing and improving emergency repair mechanisms, and improving the efficiency of safe operation and maintenance of water supply networks. Summary of the invention

[0005] In order to solve the problems in the prior art, the present invention provides a method for rapid analysis of urban water supply pipe network damage accidents, which solves the problems in the prior art that the traditional damage analysis method has a low accuracy rate in identifying pipe network damage and takes a long time to determine the damage location.

[0006] The present invention is achieved through the following technical solutions:

[0007] A method for quickly analyzing damage of a city water supply network comprises the following steps:

[0008] (1) Statistical analysis of the changing characteristics of water pressure monitoring data in the pipeline network:

[0009] Under normal operation of the water supply network, the water pressure state of the network has certain characteristic laws, including the fluctuation of the water head of the network itself at the monitoring point, and the drop of the water head from the water source point (factory pressure) to the monitoring point. The above characteristics are statistically analyzed based on the water pressure monitoring data of the network.

[0010] Statistical characteristics of the water head fluctuation at the monitoring point itself. Based on the water pressure monitoring data accumulated by the monitoring system, the data of each monitoring point is statistically analyzed on a daily basis, and the standard deviation (σ mt ), corresponding to the average water head of monitoring point m at time t On this basis, the water head fluctuation threshold of the monitoring point itself at time t is calculated:

[0011] The statistical characteristics of the head drop at the monitoring point relative to the overall network. During the water supply process of the water supply network, the head from the water source to each node of the network gradually decreases, and the head drop from the factory head to each monitoring point has a certain statistical regularity. Define the network benchmark head that represents the overall water supply of the network. In day D, when calculating time t, the network benchmark head H st Water pressure fluctuation ΔH at each monitoring point m (s-m)t , calculate its average value and its random fluctuation standard deviation (σ (s-m)t ). On this basis, the drop threshold of the water pressure at the monitoring point relative to the reference water head of the pipe network at time t is calculated:

[0012] (2) Perform full-condition pipeline damage simulation to generate a pressure sample set:

[0013] Based on the pipeline network model with low, medium and high nuclear cycles, pipeline network damage events with D days, T moments and P damage degree gradients are simulated, and each node is traversed to generate a simulated pressure sample set.

[0014] (3) Based on the hydraulic analysis of the pipe network, establish a set of monitoring point sensing nodes:

[0015] The shortest path matrix of the monitoring points is calculated based on the Dijkstra algorithm to obtain the water flow path length from each monitoring point to each node of the pipe network, and the shortest path matrix L is formed by combining them. jm .

[0016] Based on the benchmark model of the pipe network hydraulic model, a damage accident simulation of the same pipe network damage degree at each node under the average daily average water volume is carried out. The monitoring point set M and the node set J are established, and the water pressure fluctuation ΔH of the monitoring point m (m = 1...M) caused by the damage of node j (j = 1...J) is calculated. jm . Traverse each node.

[0017] If the water pressure fluctuation ΔH jm More than 2 times That is, if monitoring point m can effectively sense the abnormal water pressure fluctuation caused by the damage accident of node j, then node j belongs to the node sensing set of monitoring point m. For another monitoring point (m+1), ΔH j(m+1) Fluctuation greater than 2 times Then compare the water flow path length l jm and l j(m+1) , j is classified into the set of monitoring point sensing nodes with short water flow path length.

[0018] Traverse each monitoring point to obtain the set of sensing nodes at each monitoring point.

[0019] (4) Damage accident identification

[0020] Based on the actual monitoring data of the monitoring point, the pressure threshold at the current time t is calculated. If the actual monitoring pressure value ΔH 测m Less than the pressure threshold Then enter the damage accident identification alternative state 1.

[0021] If multiple monitoring points simultaneously experience damage accident identification alternative state 1, the monitoring point sensing node sets corresponding to the monitoring points are sorted from large to small according to the water pressure fluctuation values ​​to form an abnormal monitoring point sequence 1;

[0022] At the same time, calculate the drop value ΔH of the monitoring point m relative to the hydraulic reference pressure of the pipe network at the current time t (s-m)t , if the value exceeds the pressure drop threshold ΔH at point m at time T (s-m)t下限 , indicating that the pressure reduction at the monitoring point m exceeds the normal hydraulic gradient range of the entire pipe network, and the damage accident identification alternative state 2 is entered.

[0023] If multiple monitoring points simultaneously experience damage accident identification alternative state 2, the monitoring point sensing node sets corresponding to the monitoring points are sorted from large to small according to the water pressure fluctuation values ​​to form an abnormal monitoring point sequence 2;

[0024] When damage accident identification alternative states 1 and 2 appear at the same time, it indicates that the pressure at the monitoring point has experienced abnormal fluctuations relative to itself and the overall situation of the pipeline network. At this time, the pipeline network is identified to have a damage accident state.

[0025] (5) Damage accident location

[0026] After identifying the damage accident state of the pipeline network, the sensing node sets of each monitoring point corresponding to the abnormal monitoring point sequence 1 and the abnormal monitoring point sequence 2 at the current time t are extracted. According to the sorting order, the data in the simulated water pressure fluctuation sample set is extracted, and the error value is calculated. The errors are sorted from small to large, and the nodes closest to the water pressure fluctuation characteristics of the current pipeline damage accident and the corresponding node sorting are used as the location of the damage accident and the priority order for inspection.

[0027] Beneficial effects: 1. The present invention clearly characterizes the pressure fluctuation characteristics of the monitoring point itself and the pressure drop relative to the overall pipe network during the normal operation of the pipe network by calculating the standard deviation of the water pressure fluctuation of the monitoring point itself and the pressure drop of the monitoring point relative to the overall pipe network;

[0028] 2. Based on the above characteristics, a comprehensive identification and analysis of pipeline damage accidents is performed. Damage accident identification alternative state 1 indicates that the pressure at the monitoring point has abnormally fluctuated relative to its own changes, and damage accident identification alternative state 2 indicates that the pressure at the monitoring point has dropped beyond the normal range relative to the overall situation of the pipeline network. Comprehensive identification avoids the one-sidedness of a single judgment method, effectively improves the accuracy of identification, and reduces the false alarm rate.

[0029] 3. The present invention establishes a set of sensing nodes at monitoring points. First, the length of the water flow path from each monitoring point to each node of the pipe network is calculated to form the shortest path matrix L jm , and then judge the damage condition perception ability of each monitoring point to each node. Combined with the comparison of the water flow path length, each node can be clearly divided into the perception range of the monitoring point with the shortest water flow path, that is, the set of monitoring point perception nodes is obtained, which represents the monitoring range of the monitoring point with the best perception ability.

[0030] 4. In the identification of damage accidents, the water pressure fluctuation value ΔH of the monitoring point itself is 测m , and the drop value ΔH of the monitoring point m relative to the hydraulic reference pressure of the pipe network (s-m)t The order of monitoring points with abnormal pressure is obtained. On this basis, in the damage accident location analysis, the samples in the sensing node set corresponding to these monitoring points can be preferentially extracted and calculated, avoiding the global search of all node samples in the pipeline network and improving the search efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 A general framework diagram of a rapid analysis method for urban water supply network damage accidents;

[0032] Figure 2 Schematic diagram of the simulated sample set of pipeline damage accidents;

[0033] Figure 3The shortest water flow path matrix L of the monitoring point jm Schematic diagram. DETAILED DESCRIPTION

[0034] The following is a further description of the method for rapid analysis of urban water supply network damage accidents according to the present invention in conjunction with the accompanying drawings and through specific implementation methods:

[0035] Step 1: Based on the characteristic law of water pressure state of the pipe network, construct Figure 1 The schematic diagram shows a rapid analysis method for the hydraulic status of a water supply network. The method is described by two characteristic indicators: the statistical characteristics of the head fluctuation of the monitoring point itself and the statistical characteristics of the monitoring point relative to the overall network.

[0036] In this method, the pressure monitoring data collected by the monitoring point is considered to contain four dimensions of information, namely, the monitoring point number m at which the data is recorded, the date d at which the data is recorded, the time t at which the data is recorded (generally 1 minute is one time, 24*60=1440 times a day), the size H of the data dt . The total number of monitoring points is recorded as M, the total number of recorded data dates is recorded as D, and the total number of recorded data times in a day is recorded as T. Under normal operating conditions of the water supply network, the water pressure state of the network has certain characteristic laws, including the fluctuation of the water head of the network itself at the monitoring point, and the drop of the water head from the water source point (factory pressure) to the monitoring point. The above characteristics are statistically analyzed from a statistical perspective based on the water pressure monitoring data of the network.

[0037] (1) Calculation of statistical characteristic indicators of the water head fluctuation at the monitoring point itself.

[0038] The statistical characteristic index of the water head fluctuation of the monitoring point itself is calculated by calculating the water head H at each time t of the monitoring point m within D days. mdt The standard deviation σ mt To quantify, the specific calculation formula is as follows:

[0039]

[0040] Where: mt ——the standard deviation of the water head monitoring value at monitoring point m at time t; mdt ——The water head monitoring value at monitoring point m at time t on date d; ——The average value of the head monitoring value of monitoring point m at time t on all dates; D——The total number of days of the date.

[0041] First, by calculating the standard deviation of the pressure data of each monitoring point at different times of the day, the characteristic value σ of the water consumption fluctuation of each monitoring point at each time within a certain period is obtained. mt , and then calculate the water head fluctuation threshold of the monitoring point itself at time t: The statistical characteristic indicators of the water head fluctuation at the monitoring point itself are obtained.

[0042] (2) Calculation of statistical characteristic indicators of the water head drop at the monitoring point relative to the overall pipe network.

[0043] During the water supply process of the water supply network, the water head from the water source to each node of the network gradually decreases, and the head reduction from the factory head to each monitoring point has a certain statistical law. To facilitate the description of this law, first define the network base water head as the outlet water head of the water plant water supply pump station at time t, and use H st Calculate the reference water head H at all times t in D days st The pressure drop ΔH to each monitoring point m (s-m)t The average and standard deviation σ (s-m)t To describe the statistical characteristics of the head drop of the monitoring point relative to the overall pipe network, the specific calculation formula is as follows:

[0044] ΔH (s-m)t =H st -H mt

[0045]

[0046] Where: ΔH (s-m)t ——pressure drop from the reference water head of the pipe network to each monitoring point m at time t; H st ——Base water head of the pipe network, i.e., the outlet water head of the water supply pump station of the water plant at time t; H mt ——the water pressure recorded at monitoring point m at time t; σ (s-m)t ——The standard deviation of the pressure drop from the network base water head to each monitoring point at time t; ——The average value of the pressure drop from the reference water head to each monitoring point at all times t in D days; D——The total number of days of data recording date;

[0047] According to statistical laws, it is believed that when the reference water head H st The pressure drop to the monitoring point m is greater than the head drop threshold of the monitoring point relative to the entire pipe network. When the water pressure is abnormal, the monitoring point detects abnormal water pressure.

[0048] Step 2: Perform full-condition pipeline damage simulation to generate Figure 2 The schematic diagram shows a sample set of simulated pipeline damage accidents. Based on the pipeline hydraulic model d calibrated based on the low-day, average-day, and high-day monitoring point data, the EPANET software toolkit is used to traverse and simulate the damage events of all nodes in the pipeline network by programming. Considering that the degree of the outbreak of damage events is different and the time of occurrence is also different, for each day's model, the damage event of each node specifically includes P damage degree gradients and a total of P*T small events at T times.

[0049] Record the pressure value H of each monitoring point when each damage event occurs djpmt And the pressure value H of the monitoring point under normal working conditions of the model at this moment dmt , subtract the latter from the former to get the water pressure fluctuation ΔH at the monitoring point after the damage event occurs djpmt .

[0050] The water pressure fluctuation data of all monitoring points are summarized according to the pipeline hydraulic model d, node number j, damage degree p, monitoring point m and accident time t to form a sample set of water pressure fluctuations for pipeline damage simulation under all working conditions.

[0051] Step 3: Based on the hydraulic analysis of the pipe network, establish a set of monitoring point sensing nodes. In this step, the hydraulic fluctuation sensing index and the water flow path sensing index are used to divide the monitoring point sensing node set.

[0052] (1) Hydraulic fluctuation perception index

[0053] Based on the hydraulic model of the pipeline network, a simulation of damage accidents with the same degree of damage at each node under the average daily average time conditions of the pipeline network is carried out. The pressure data at the monitoring point m in the simulated damage accident of each node is subtracted from the pressure data under normal conditions to obtain the water pressure fluctuation data ΔH of the monitoring point m. jm , and record it as an indicator of hydraulic fluctuation perception.

[0054] (2) Water flow path perception indicators

[0055] When the pipe network topology data is obtained, the shortest water flow path of the monitoring point is calculated based on the Dijkstra algorithm, and the following is obtained: Figure 3 The shortest water flow path matrix L shown in the schematic diagram jm The calculation process of the Dijkstra algorithm is as follows: First, according to the topology of the pipeline network, an undirected graph adjacency matrix G is constructed to describe the node connection relationship in the pipeline network node set J. Then, based on the adjacency matrix G, the Dijkstra algorithm is used to calculate the shortest paths from all monitoring points in the monitoring point set M to all nodes in the pipeline network node set J, which are used as the water flow path perception indicators of the pipeline network monitoring points, and the shortest water flow path matrix L is formed. jm .

[0056] Among them, the adjacency matrix G is a matrix representing the adjacent relationship between the nodes of the pipeline network. The matrix G uses the length of the pipe section between two connected nodes as the weight value, and the weight value of the unconnected node is infinite. At the same time, to avoid repeated calculations, the adjacency matrix only retains the weight value of the upper triangle.

[0057] The Dijkstra algorithm is a method for calculating the shortest path between two nodes based on the adjacency matrix. An auxiliary array V is required in the calculation process. m , each of its elements V m [n] indicates the current found starting point j m (i.e. monitoring point m) to each node j n The length of is retrieved from the adjacency matrix G of the pipe network. The calculation process of Dijkstra algorithm is as follows:

[0058] 1) To facilitate the description of the algorithm's calculation process, an auxiliary array V is introduced m , auxiliary array V m The initial state of the elements in are all infinite, and the node j is updated m The length of the water flow path to itself is V m [m] is 0, update node j m To the adjacent node j ni The water flow path length is V m [ni]; 2) Find all nodes that are close to the starting point j m The node value with the minimum water flow path length min{V m} to mark and further update the starting point j m to min{V m} value corresponding to the marked node min{j ni} Adjacent node j nii The total water flow path length is retained as V m 3) Repeat step 2) until all node values ​​are marked as the shortest water flow path length, and output the auxiliary array V m As the shortest water flow path array from monitoring point m to all nodes; 4) Repeat steps 1) to 3) for each monitoring point to form M arrays V m , and finally combine them to generate the shortest water flow path matrix L jm .

[0059] (3) Determination rules of monitoring point perception node set

[0060] For the nodes in the pipe network, the hydraulic fluctuation perception index is first used for judgment. According to the judgment criteria of the hydraulic fluctuation perception index in the third step, when a damage accident occurs at the j node under the average daily average time condition, the water pressure fluctuation ΔH at the monitoring point m is jm More than 2 times That is, if it is assumed that monitoring point m can effectively sense the abnormal water pressure fluctuation caused by the damage accident of node j, then node j belongs to the node perception set of monitoring point m.

[0061] Secondly, when the fluctuation perception index cannot determine the unique affiliation of the monitoring point, the hydraulic distance perception index is further used for judgment. For example, when the damage event of the above node j occurs, for another monitoring point (m+1), its water pressure fluctuation ΔH j(m+1) Also greater than 2 times At this time, the shortest path matrix L jm The shortest water flow path information from the two monitoring points to the node is compared with the shortest water flow path length l jm and l j(m+1) , node j is included in the set of monitoring point sensing nodes with short water flow path length.

[0062] Traverse each monitoring point to obtain the set of sensing nodes at each monitoring point.

[0063] Step 4: Identification of pipeline damage accidents.

[0064] (1) Damage accident identification alternative state 1 judgment rule

[0065] Based on the actual monitoring data of the monitoring point, the pressure threshold at the current time t is calculated. If the actual monitoring pressure value ΔH 测m Less than the pressure threshold Then it enters the damage accident identification alternative state 1. If multiple monitoring points appear in the damage accident identification alternative state 1 at the same time, the monitoring point sensing node sets corresponding to the monitoring points are sorted from large to small according to the water pressure fluctuation value to obtain the abnormal detection point sequence 1.

[0066] (2) Damage accident identification alternative state 2 judgment rule

[0067] Calculate the hydraulic reference head H of the monitoring point m relative to the pipe network at the current time t st The head drop value ΔH (s-m)t , if the value exceeds the pressure drop threshold at point m at time t This indicates that the pressure drop at the monitoring point m exceeds the normal hydraulic gradient fluctuation range of the entire pipe network, and the damage accident identification alternative state 2 is entered.

[0068] If multiple monitoring points simultaneously experience damage accidents, alternative state identification 2 is performed, and the monitoring point sensing node sets corresponding to the monitoring points are sorted from large to small according to the water pressure fluctuation values ​​to obtain abnormal monitoring point sequence 2.

[0069] When damage accident identification alternative states 1 and 2 appear at the same time, it indicates that the pressure at the monitoring point has experienced abnormal fluctuations relative to itself and the overall situation of the pipeline network. At this time, the pipeline network is identified to have a damage accident state.

[0070] Step 5: Search and analysis of damaged areas of the pipeline network based on hydraulic model simulation

[0071] (1) Water pressure fluctuation value sequence at pipe network monitoring points

[0072] Compared with the identification of pipeline damage events, the search and analysis of pipeline damage areas requires more accurate positioning results. Therefore, it is necessary to use the water pressure fluctuation value sequence formed by data from multiple monitoring points to comprehensively locate the damaged area.

[0073] This section uses a pipeline damaged area search and analysis method based on hydraulic model simulation to search and analyze damaged areas. The search and analysis of damaged areas is achieved by comparing the differences in the simulated data set and the water pressure fluctuation sequences of the monitoring points during the actual measurement process.

[0074] For the simulation data, since the hydraulic model is not limited by time, it can simulate two parallel working conditions, damage accident and normal working condition, at the same time. Therefore, for the current time t, the water pressure fluctuation value ΔH of each simulation sequence is 模tm It is equal to the pressure difference between the simulated damage accident condition at time t and the simulated normal condition at time t. In actual monitoring, only one condition can occur at a time. Therefore, for the measured data, the water pressure fluctuation value in each set of water pressure fluctuation value sequence is calculated and generated using the time window method. The water pressure fluctuation value ΔH at each time t is 测tm It is equal to the pressure difference between the current time t and the previous time t'.

[0075] (2) Damaged area search and analysis

[0076] After the damage accident of the pipeline network is identified, the water pressure fluctuation values ​​of all monitoring points at the current time t are extracted in order from small to large according to the number of the monitoring points to form a water pressure fluctuation value sequence S 测t , indicating the overall water pressure fluctuation characteristics of the pipeline network under the current pipeline damage accident.

[0077] The damage events of all nodes in the sensing node set of each monitoring point corresponding to the abnormal monitoring point sequence 1 and the abnormal monitoring point sequence 2 are extracted, and the water pressure fluctuation sequence S of the damage event is extracted after classification according to different nodes j and different degrees p. 测jpt The measured water pressure fluctuation sequence S 测t With the simulated water pressure fluctuation sequence S 测jpt By comparison, the mean square error σ of the two sets of sequence deviations is calculated. jpt , and sort them from small to large. The sorting result corresponds to the similarity sorting result of the water pressure fluctuation characteristics of each simulated damage accident and the actual damage accident of the current pipeline network. According to the number of each simulated damage accident, the node corresponding to the damage accident is searched to finally locate the location of the actual damage event, and the priority of manual inspection is determined according to the sorting order of the corresponding node damage accidents.

[0078] (3) Visualization

[0079] The top 50 nodes in the sorting results in step (2) are selected as exploration nodes, and their geographic location data are imported into the waterdesk software for visualization, forming an intuitive distribution map of pipeline damage event prediction to assist inspection staff in conducting inspections along the line.

[0080] The above description is only a preferred embodiment of the present invention or invention and is not intended to limit the present invention or invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention or invention should be included in the protection scope of the present invention or invention.

Claims

1. A method for rapid analysis of damage to a city water supply network, characterized in that: The specific steps include: (1) Statistical analysis of the changing characteristics of water pressure monitoring data in the pipeline network: Statistical characteristics of the water head fluctuations at the monitoring point itself; Based on the water pressure monitoring data accumulated by the monitoring system, the data of each monitoring point are statistically analyzed on a daily basis to calculate the standard deviation (σ) of the water head fluctuation at each monitoring point at the same time every day. mt ), corresponding to the average water head at monitoring point m at time t On this basis, the water head fluctuation threshold of the monitoring point itself at time t is calculated: Statistical characteristics of the head drop at the monitoring point relative to the overall network; During the water supply process of the water supply network, the head from the water source to each node of the network gradually decreases, and the head drop from the factory head to each monitoring point has a statistical law; Define the network benchmark head representing the overall water supply of the network. In D days, calculate the network benchmark head H at time t. st Water pressure fluctuation ΔH at each monitoring point m (s-m)t , calculate its average value and its random fluctuation standard deviation (σ (s-m)t ); on this basis, the threshold of the decrease of the water pressure in the pipe network at the monitoring point relative to the reference water head of the pipe network at time t is calculated: (2) Perform full-condition pipeline damage simulation to generate a pressure sample set: Based on the pipeline network model with low, medium and high nuclear cycles, simulate the pipeline network damage events with D days, T moments and P damage degree gradients, traverse each node and generate a simulated pressure sample set; (3) Based on the hydraulic analysis of the pipe network, establish a set of monitoring point sensing nodes: The shortest path matrix of the monitoring points is calculated based on the Dijkstra algorithm to obtain the water flow path length from each monitoring point to each node of the pipe network, and the shortest path matrix L is formed by combining them. jm ; Based on the benchmark model of the pipeline network hydraulic model, a damage accident simulation with the same degree of pipeline damage at each node under the average daily average time condition is carried out; a monitoring point set M and a node set J are established, and the water pressure fluctuation ΔH of the monitoring point m (m = 1 ... M) caused by the damage of node j (j = 1 ... J) is calculated. jm ; Traverse each node; If the water pressure fluctuation ΔH jm More than 2 times That is, monitoring point m can effectively sense the abnormal water pressure fluctuation caused by the damage accident of node j, then node j belongs to the node perception set of monitoring point m; for another monitoring point (m+1), ΔH j(m+1) Fluctuation greater than 2 times Then compare the water flow path length l jm and l j(m+1) , j is classified into the set of monitoring point sensing nodes with small water flow path length; Traverse each monitoring point to obtain the sensing node set of each monitoring point; (4) Damage accident identification Based on the actual monitoring data of the monitoring point, the pressure threshold at the current time t is calculated. If the actual monitoring pressure value ΔH 测m Less than the pressure threshold Then enter the damage accident identification alternative state 1; If multiple monitoring points simultaneously experience damage accident identification alternative state 1, the monitoring point sensing node sets corresponding to the monitoring points are sorted from large to small according to the water pressure fluctuation values ​​to form an abnormal monitoring point sequence 1; At the same time, calculate the drop value ΔH of the monitoring point m relative to the hydraulic reference pressure of the pipe network at the current time t (s-m)t , if the value exceeds the pressure drop threshold ΔH at point m at time T (s-m)t下限 , indicating that the pressure reduction at the monitoring point m exceeds the normal hydraulic gradient range of the entire pipe network, and the damage accident identification alternative state 2 is entered; If multiple monitoring points simultaneously experience damage accident identification alternative state 2, the monitoring point sensing node sets corresponding to the monitoring points are sorted from large to small according to the water pressure fluctuation values ​​to form an abnormal monitoring point sequence 2; When damage accident identification alternative states 1 and 2 appear at the same time, it indicates that the pressure at the monitoring point has experienced abnormal fluctuations relative to itself and the overall situation of the pipeline network. At this time, the pipeline network is identified to have a damage accident state; (5) Damage accident location After identifying the damage accident state of the pipeline network, the sensing node sets of each monitoring point corresponding to the abnormal monitoring point sequence 1 and the abnormal monitoring point sequence 2 at the current time t are extracted. According to the sorting order, the data in the simulated water pressure fluctuation sample set is extracted, and the error value is calculated. The errors are sorted from small to large, and the nodes closest to the water pressure fluctuation characteristics of the current pipeline damage accident and the corresponding node sorting are used as the location of the damage accident and the priority order for inspection.

2. The rapid analysis method for damage of urban water supply pipe network according to claim 1 is characterized in that: The statistical characteristic index of the water head fluctuation of the monitoring point itself in step (1) is calculated by calculating the water head H of the monitoring point m at each time t within D days mdt The standard deviation σ mt To quantify, the specific calculation formula is as follows: Where: mt ——the standard deviation of the water head monitoring value at monitoring point m at time t; mdt ——The water head monitoring value at monitoring point m at time t on date d; ——The average value of the head monitoring value of monitoring point m at time t on all dates; D——The total number of days of the date.

3. The rapid analysis method for damage of urban water supply pipe network according to claim 1 is characterized in that: The statistical characteristics of the head drop of the monitoring point relative to the entire pipe network in step (1) are calculated by calculating the pipe network reference head H at time t. st Average water pressure fluctuation to each monitoring point m and its random fluctuation standard deviation σ (s-m)t The specific calculation formula is as follows: ΔH (s-m)t =H st -H mt Where: ΔH (s-m)t ——pressure drop from the reference water head of the pipe network to each monitoring point m at time t; H st ——Base water head of the pipe network, i.e., the outlet water head of the water supply pump station of the water plant at time t; H mt ——the water pressure recorded at monitoring point m at time t; σ (s-m)t ——The standard deviation of the pressure drop from the network base water head to each monitoring point at time t; ——The average value of the pressure drop from the reference water head to each monitoring point at all times t in D days; D——The total number of days of data recording date.

Images