A water supply management method and system

Abstract

The present invention provides a water supply management method and system, the method comprising: obtaining an initial water supply data set, filling missing values ​​and replacing abnormal values ​​in the initial water supply data set to obtain a water supply data set to be used; segmenting the water supply data set to be used by a time window to generate a plurality of data segments, constructing and training an initial water supply prediction network model to obtain a final prediction network model; obtaining a predicted water supply through the final prediction network model, determining a predicted pipeline stress threshold of a pipeline of a water supply system based on the predicted water supply; obtaining the real-time stress of the pipeline of the water supply system, and comparing the real-time stress with the predicted pipeline stress threshold to judge the health status of the pipeline. The stress corresponding to the pipeline itself is used as the judgment standard of the health status, replacing the water pressure with a high degree of variation, avoiding the judgment error of the health status of the pipeline caused by the uncertainty of the water pressure, and improving the accuracy of monitoring the health status of the pipeline.

Description

Technical Field

[0001] The present invention relates to the technical field of data prediction, and in particular to a water supply management method and system. Background Art

[0002] With the rapid expansion of urbanization, municipal infrastructure construction has developed rapidly. As the "blood" buried underground, the urban water supply system is one of the most important municipal infrastructures. Its orderly operation is the guarantee for the normal operation of the city. The urban water supply system is a general term for facilities that provide water for people's lives, production, municipal administration and fire fighting. It supplies water quantities and water quality that meet different needs to various categories of users.

[0003] The health status of the pipeline corresponding to the water supply system is currently monitored by monitoring the water pressure in the pipeline to determine whether the pipeline is damaged. However, during the water supply process, transient flows generated by valve opening and closing or water pump power failure will cause changes in water pressure. That is, changes in water pressure are not only related to pipeline damage. Judging the health status of the pipeline by monitoring water pressure is often not accurate enough. Summary of the invention

[0004] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a water supply management method and system, aiming to solve the technical problem in the prior art of determining the health status of the pipeline by monitoring the water pressure in the pipeline of the water supply system, and the monitoring results are not accurate enough.

[0005] In order to achieve the above objectives, in a first aspect, an embodiment of the present application provides a water supply management method, comprising the following steps:

[0006] Acquire an initial water supply data set, wherein the initial water supply data set includes a plurality of daily water supply data sets, wherein the daily water supply data sets include water supply at a plurality of time points, wherein the time points in different daily water supply data sets are the same, and perform missing value filling and outlier replacement on the initial water supply data set to obtain a water supply data set to be used;

[0007] The water supply data set to be used is segmented by time windows to generate a plurality of data segments, and an initial water supply prediction network model is constructed. In two adjacent data segments, the data segment of the previous time window is used as an input value, and the water supply at the first time point in the data segment of the next time window is used as an output value, and the initial water supply prediction network model is trained to obtain a final prediction network model;

[0008] Obtaining a predicted water supply through the final prediction network model, obtaining a pipeline flow corresponding to a pipeline of a water supply system through the predicted water supply, obtaining a pipeline flow rate according to the pipeline flow, obtaining a first head loss according to the pipeline flow, and obtaining a second head loss according to the pipeline flow rate, obtaining a predicted water pressure in the pipe based on the first head loss and the second head loss, and determining a predicted pipeline stress threshold according to the predicted water pressure in the pipe;

[0009] The real-time stress of the pipeline of the water supply system is obtained, and the real-time stress is compared with the predicted pipeline stress threshold to determine the health status of the pipeline.

[0010] Furthermore, the step of filling missing values ​​and replacing outliers on the initial water supply data set includes:

[0011] Detecting whether there are missing points in the initial water supply data set;

[0012] If the missing point exists, two adjacent points adjacent to the missing point are detected, and a filling water supply is generated based on the water supply at the time points corresponding to the two adjacent points, and the filling water supply is added to the missing point;

[0013] Determine the daily average water supply in the daily water supply data set, and determine the abnormal threshold value through the daily average water supply and the water supply at the time point;

[0014] Determining whether there is abnormal data in the daily water supply data set based on the daily average water supply and the abnormal threshold;

[0015] If the abnormal data exists, an updated water supply amount is obtained based on the adjacent daily water supply data set, and the abnormal data is replaced by the updated water supply amount.

[0016] Furthermore, the step of detecting whether there are missing points in the initial water supply data set includes:

[0017] Constructing a coordinate system with the time point as the horizontal coordinate and the water supply at the time point as the vertical coordinate, so as to map the initial water supply data set into a plurality of coordinate points;

[0018] Obtaining the Euclidean distance between two adjacent coordinate points, and determining whether the Euclidean distance is greater than a distance threshold;

[0019] If the Euclidean distance is greater than the distance threshold, it is determined that there is a missing point between the two coordinate points.

[0020] Furthermore, the calculation formula for the average daily water supply is:

[0021] ,

[0022] in, represents the daily average water supply of the i-th daily water supply data set, represents the water supply at the jth time point in the i-th day water supply dataset, represents the total number of time points in the water supply dataset for the ith day;

[0023] The calculation formula of the abnormal threshold is:

[0024] ,

[0025] in, represents the abnormal threshold of the water supply dataset on the ith day;

[0026] The formula for obtaining the updated water supply is:

[0027] ,

[0028] in, represents the updated water supply at the jth time point in the i-th day water supply dataset, represents the water supply at the jth time point in the i+1th daily water supply dataset, represents the time point of the jth time point in the i-1th daily water supply dataset, , are weight coefficients, and .

[0029] Furthermore, the initial water supply prediction network model includes an LSTM layer, an attention layer and a fully connected layer, the data segment of the previous time window is used as an input value, and the water supply at the first time point in the data segment of the next time window is used as an output value, and the steps of training the initial water supply prediction network model include:

[0030] Input the water supply at several time points in the data segment of the previous time window into the LSTM layer to obtain several first feature vectors, and input the several first feature vectors into the attention layer;

[0031] Input the water supply at the first time point in the data segment of the next time window into the attention layer to obtain a prompt vector, determine a weight vector of the first feature vector based on the prompt vector, and input a plurality of the first feature vectors and the weight vectors corresponding thereto into the fully connected layer;

[0032] The water supply at the first time point in the data segment of the next time window is used as the output value of the fully connected layer, and the predicted value is output through the fully connected layer, a loss function is constructed based on the predicted value and the output value, and the initial water supply prediction network model is trained through the loss function.

[0033] Furthermore, the formula for obtaining the weight vector is:

[0034] ,

[0035] in, represents the weight vector corresponding to the mth first eigenvector, represents the mth first eigenvector, Represents the transposition character, represents the prompt vector, represents the total number of first eigenvectors in the data segment of the previous time window, Represents an exponential function.

[0036] Furthermore, the step of determining the predicted pipeline stress threshold according to the predicted water pressure in the pipe includes:

[0037] The predicted annular stress and the predicted axial stress of the pipeline of the water supply system are respectively obtained by the predicted water pressure in the pipeline;

[0038] The predicted pipeline stress threshold is obtained through the predicted annular stress and the predicted axial stress.

[0039] Furthermore, the calculation formula for predicting the hoop stress is:

[0040] ,

[0041] in, represents the predicted hoop stress, represents the predicted water pressure in the pipe, Indicates the inner diameter of the pipe. Indicates the wall thickness of the pipe;

[0042] The calculation formula for the predicted axial stress is:

[0043] ,

[0044] in, represents the predicted axial stress, represents the interface internal force of the pipeline, represents the cross-sectional area of ​​the pipeline, Indicates the outer diameter of the pipe. represents the bending moment acting on the pipeline, Indicates the section coefficient of the pipeline;

[0045] The calculation formula for predicting the pipeline stress threshold is:

[0046] ,

[0047] in, Indicates the predicted pipeline stress threshold.

[0048] In a second aspect, an embodiment of the present application provides a water supply management system, which is applied to the water supply management method as described in the first aspect above, and the system includes:

[0049] A collection module is used to obtain an initial water supply data set, wherein the initial water supply data set includes a plurality of daily water supply data sets, wherein the daily water supply data sets include water supply at a plurality of time points, wherein the time points in different daily water supply data sets are the same, and the initial water supply data set is filled with missing values ​​and replaced with abnormal values ​​to obtain a water supply data set to be used;

[0050] A processing module, used to segment the water supply data set to be used by time windows to generate a plurality of data segments, construct an initial water supply prediction network model, and train the initial water supply prediction network model by taking the data segment of the previous time window as an input value and the water supply at the first time point in the data segment of the next time window as an output value in two adjacent data segments, so as to obtain a final prediction network model;

[0051] a calculation module, configured to obtain a predicted water supply through the final prediction network model, obtain a pipeline flow corresponding to a pipeline of a water supply system through the predicted water supply, obtain a pipeline flow rate according to the pipeline flow, obtain a first head loss according to the pipeline flow, obtain a second head loss according to the pipeline flow rate, obtain a predicted in-pipe water pressure based on the first head loss and the second head loss, and determine a predicted pipeline stress threshold according to the predicted in-pipe water pressure;

[0052] The analysis module is used to obtain the real-time stress of the pipeline of the water supply system, and compare the real-time stress with the predicted pipeline stress threshold to determine the health status of the pipeline.

[0053] In a third aspect, an embodiment of the present application provides a computer, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the water supply management method as described in the first aspect above is implemented.

[0054] In a fourth aspect, an embodiment of the present application provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, the water supply management method as described in the first aspect above is implemented.

[0055] Compared with the prior art, the beneficial effects of the present invention are as follows: by filling missing values ​​and replacing abnormal values ​​in the initial water supply data set, data deviation caused by abnormalities or damage of the collection equipment during data collection can be avoided, thereby improving the accuracy of subsequent predictions of the health status of the pipeline; by constructing the final prediction network model, based on the time series correspondence of the water supply at different time points in the stand-by water supply data set, the correlation between the water supply at multiple time points in the data segment of the previous time window and the water supply at the first time point in the data segment of the next time window is analyzed, and the predicted water supply is obtained based on multiple data, thereby improving the accuracy of data. Accuracy and reliability; by predicting the predicted water pressure in the pipe and obtaining the predicted pipeline stress threshold based on the predicted water pressure in the pipe, the stress corresponding to the pipeline itself is used as the judgment standard for the health status, which replaces the water pressure with a higher degree of variation, avoids the judgment error of the health status of the pipeline caused by the uncertainty of the water pressure, and improves the accuracy of monitoring the health status of the pipeline; at the same time, by combining the predicted water supply to update the predicted pipeline stress threshold in real time, a dynamic evaluation standard is realized, which avoids the misjudgment of the health status of the pipeline caused by the mismatch between a single threshold and the changing water supply, and further improves the accuracy of monitoring the health status of the pipeline. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a flow chart of the water supply management method in the first embodiment of the present invention;

[0057] Figure 2 It is a structural block diagram of a water supply management system in a second embodiment of the present invention;

[0058] The following specific implementation manner will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION

[0059] In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. Several embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

[0060] It should be noted that when an element is referred to as being "fixed to" another element, it may be directly on the other element or there may be a central element. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may be a central element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.

[0061] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0062] See also Figure 1 The water supply management method provided by the first embodiment of the present invention comprises the following steps:

[0063] S10: Obtain an initial water supply data set, wherein the initial water supply data set includes a plurality of daily water supply data sets, wherein the daily water supply data sets include water supply at a plurality of time points, wherein the time points in different daily water supply data sets are the same, and perform missing value filling and outlier replacement on the initial water supply data set to obtain a water supply data set to be used;

[0064] Different daily water supply data sets represent water supply data collected on different days. Assuming that seven days of water supply data are extracted, there are seven daily water supply data sets. Taking a single daily water supply data set as an example, if the water supply data is obtained at a frequency of 1 hour / time, the daily water supply data set has water supply at 24 time points. Different daily water supply data sets include water supply at 24 time points, and the collection time points are the same.

[0065] The step S10 comprises:

[0066] S110: Detecting whether there are missing points in the initial water supply data set;

[0067] Specifically, a coordinate system is constructed with the time point as the horizontal coordinate and the water supply at the time point as the vertical coordinate to map the initial water supply data set into a number of coordinate points; the Euclidean distance between two adjacent coordinate points is obtained, and it is determined whether the Euclidean distance is greater than a distance threshold; if the Euclidean distance is greater than the distance threshold, it is determined that there is a missing point between the two coordinate points.

[0068] S120: If the missing point exists, two adjacent points adjacent to the missing point are detected, and a filling water supply is generated based on the water supply at the time points corresponding to the two adjacent points, and the filling water supply is added to the missing point;

[0069] In this embodiment, the replenishment water supply is generated by the average of the two adjacent points.

[0070] S130: Determine the daily average water supply in the daily water supply data set, and determine the abnormal threshold value according to the daily average water supply and the water supply at the time point;

[0071] The calculation formula for the average daily water supply is:

[0072] ,

[0073] in, represents the daily average water supply of the i-th daily water supply data set, represents the water supply at the jth time point in the i-th day water supply dataset, Represents the total number of time points in the water supply dataset for the ith day.

[0074] The calculation formula of the abnormal threshold is:

[0075] ,

[0076] in, Represents the abnormal threshold of the water supply dataset on the ith day.

[0077] S140: Determine whether there is abnormal data in the daily water supply data set based on the daily average water supply and the abnormal threshold;

[0078] After obtaining the abnormal threshold, determine whether the abnormal data exists by the following formula:

[0079] ,

[0080] When the difference between the water supply at the jth time point in the i-th daily water supply data set and the daily average water supply of the i-th daily water supply data set is greater than 4.5 times the abnormal threshold, the water supply at that time point is determined to be abnormal data.

[0081] S150: If the abnormal data exists, obtaining an updated water supply amount based on the adjacent daily water supply data set, and replacing the abnormal data with the updated water supply amount.

[0082] The formula for obtaining the updated water supply is:

[0083] ,

[0084] in, represents the updated water supply at the jth time point in the i-th day water supply dataset, represents the water supply at the jth time point in the i+1th daily water supply dataset, represents the time point of the jth time point in the i-1th daily water supply dataset, , are weight coefficients, and .

[0085] S20: dividing the water supply data set to be used by time windows to generate a plurality of data segments, constructing an initial water supply prediction network model, and in two adjacent data segments, using the data segment of the previous time window as an input value and the water supply at the first time point in the data segment of the next time window as an output value, training the initial water supply prediction network model to obtain a final prediction network model;

[0086] In this embodiment, the number of time points in the daily water supply data set is used as the time window to complete the segmentation of the water supply data set to be used. The initial water supply prediction network model includes an LSTM layer, an attention layer and a fully connected layer, the LSTM layer is connected to the attention layer, and the attention layer is connected to the fully connected layer.

[0087] The step S20 comprises:

[0088] S210: Inputting the water supply at several time points in the data segment of the previous time window into the LSTM layer to obtain several first feature vectors, and inputting the several first feature vectors into the attention layer;

[0089] S220: Inputting the water supply at the first time point in the data segment of the next time window into the attention layer to obtain a prompt vector, determining a weight vector of the first feature vector based on the prompt vector, and inputting a plurality of the first feature vectors and the weight vectors corresponding thereto into the fully connected layer;

[0090] The formula for obtaining the weight vector is:

[0091] ,

[0092] in, represents the weight vector corresponding to the mth first eigenvector, represents the mth first eigenvector, Represents the transposition character, represents the prompt vector, represents the total number of first eigenvectors in the data segment of the previous time window, Represents an exponential function.

[0093] S230: taking the water supply at the first time point in the data segment of the next time window as the output value of the fully connected layer, and outputting a predicted value through the fully connected layer, constructing a loss function based on the predicted value and the output value, and training the initial water supply prediction network model through the loss function;

[0094] It can be understood that the fully connected layer combines the first feature vector and the weight vector corresponding thereto into a second feature vector, and generates the predicted value according to a plurality of the second feature vectors. The purpose of the loss function is to shorten the distance between the predicted value and the output value, and based on the difference between the predicted value and the output value, reversely adjust the setting parameters of the initial water supply prediction network model, thereby obtaining the final prediction network model.

[0095] S30: obtaining a predicted water supply through the final prediction network model, obtaining a pipeline flow corresponding to a pipeline of the water supply system through the predicted water supply, obtaining a pipeline flow rate according to the pipeline flow, obtaining a first head loss according to the pipeline flow, and obtaining a second head loss according to the pipeline flow rate, obtaining a predicted water pressure in the pipe based on the first head loss and the second head loss, and determining a predicted pipeline stress threshold according to the predicted water pressure in the pipe;

[0096] It can be understood that the predicted water supply can be obtained by obtaining the real-time water supply and inputting the real-time water supply into the final prediction network model.

[0097] The formula for obtaining the pipeline flow rate is:

[0098] ,

[0099] in, Indicates the pipeline flow rate, Indicates the pipeline flow rate, Indicates the inner diameter of the pipe.

[0100] The calculation formula of the first head loss is:

[0101] ,

[0102] in, represents the first head loss, Represents the characteristic coefficient of the water meter in the water supply system.

[0103] The calculation formula of the second head loss is:

[0104] ,

[0105] in, represents the second head loss, It represents the resistance coefficient of the pipeline. Represents the acceleration due to gravity.

[0106] It should be noted that, depending on the locations of different pipelines, the number of other pipelines between the starting water supply point and the current pipeline is different. For the predicted water pressure in the current pipeline, the water pressure of the starting water supply needs to be superimposed on the losses of multiple pipelines to form the predicted water pressure in the pipe.

[0107] S40: Acquire the real-time stress of the pipeline of the water supply system, and compare the real-time stress with the predicted pipeline stress threshold to determine the health status of the pipeline.

[0108] The step S40 comprises:

[0109] S410: Obtaining predicted annular stress and predicted axial stress of the pipeline of the water supply system respectively through the predicted water pressure in the pipeline;

[0110] The calculation formula for the predicted hoop stress is:

[0111] ,

[0112] in, represents the predicted hoop stress, represents the predicted water pressure in the pipe, Indicates the inner diameter of the pipe. Indicates the wall thickness of the pipe;

[0113] The calculation formula for the predicted axial stress is:

[0114] ,

[0115] in, represents the predicted axial stress, represents the interface internal force of the pipeline, represents the cross-sectional area of ​​the pipeline, Indicates the outer diameter of the pipe. represents the bending moment acting on the pipeline, Indicates the section coefficient of the pipeline.

[0116] S420: Obtaining a predicted pipeline stress threshold value through the predicted annular stress and the predicted axial stress;

[0117] The calculation formula for predicting the pipeline stress threshold is:

[0118] ,

[0119] in, Indicates the predicted pipeline stress threshold.

[0120] It should be noted that in the water supply system, stress sensors are installed in different pipelines to detect the real-time stress in the pipelines. After obtaining the predicted pipeline stress threshold, if the real-time stress is less than the predicted pipeline stress threshold, the pipeline is judged to be in a healthy state; if the real-time stress is greater than the predicted pipeline stress threshold, the pipeline is judged to be in an unhealthy state and needs to be repaired.

[0121] By filling missing values ​​and replacing abnormal values ​​in the initial water supply data set, data deviation caused by abnormalities or damage of the collection equipment during the data collection process can be avoided, thereby improving the accuracy of subsequent predictions of the health status of the pipeline; by constructing the final prediction network model, based on the time series correspondence of the water supply at different time points in the stand-by water supply data set, the correlation between the water supply at multiple time points in the data segment of the previous time window and the water supply at the first time point in the data segment of the next time window is analyzed, and the predicted water supply is obtained based on multiple data, thereby improving the accuracy and reliability of the data; The predicted water pressure in the pipe is predicted, and a predicted pipeline stress threshold is obtained based on the predicted water pressure in the pipe, and the stress corresponding to the pipeline itself is used as the judgment standard for the health status, thereby replacing the water pressure with a higher degree of variation, avoiding the judgment error of the health status of the pipeline caused by the uncertainty of the water pressure, and improving the accuracy of monitoring the health status of the pipeline; at the same time, the predicted pipeline stress threshold is updated in real time by combining with the predicted water supply, thereby realizing a dynamic evaluation standard, avoiding the misjudgment of the health status of the pipeline caused by the mismatch between a single threshold and the changing water supply, and further improving the accuracy of monitoring the health status of the pipeline.

[0122] See also Figure 2 The second embodiment of the present invention provides a water supply management system, which is applied to the water supply management method described in the above embodiment, and will not be repeated here. As used below, the terms "module", "unit", "subunit", etc. can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0123] The system comprises:

[0124] The acquisition module 10 is used to obtain an initial water supply data set, wherein the initial water supply data set includes a plurality of daily water supply data sets, wherein the daily water supply data sets include water supply at a plurality of time points, wherein the time points in different daily water supply data sets are the same, and the initial water supply data set is filled with missing values ​​and replaced with abnormal values ​​to obtain a water supply data set to be used;

[0125] The acquisition module 10 includes:

[0126] The first unit is used to detect whether there are missing points in the initial water supply data set;

[0127] The first unit is specifically used to construct a coordinate system with the time point as the horizontal coordinate and the water supply at the time point as the vertical coordinate, so as to map the initial water supply data set into a plurality of coordinate points; obtain the Euclidean distance between two adjacent coordinate points, and determine whether the Euclidean distance is greater than a distance threshold; if the Euclidean distance is greater than the distance threshold, it is determined that there is a missing point between the two coordinate points;

[0128] The second unit is used for detecting two adjacent points adjacent to the missing point if the missing point exists, generating a filling water supply based on the water supply at the time points corresponding to the two adjacent points, and filling the missing point with the filling water supply;

[0129] The third unit is used to determine the daily average water supply in the daily water supply data set, and determine the abnormal threshold value according to the daily average water supply and the water supply at the time point;

[0130] A fourth unit is used to determine whether there is abnormal data in the daily water supply data set based on the daily average water supply and the abnormal threshold;

[0131] A fifth unit is configured to obtain an updated water supply amount based on the adjacent daily water supply data set if the abnormal data exists, and replace the abnormal data with the updated water supply amount;

[0132] The processing module 20 is used to divide the water supply data set to be used by time windows to generate a plurality of data segments, construct an initial water supply prediction network model, and train the initial water supply prediction network model by taking the data segment of the previous time window as the input value and the water supply at the first time point in the data segment of the next time window as the output value in two adjacent data segments to obtain a final prediction network model.

[0133] The processing module 20 includes:

[0134] A sixth unit is used to input the water supply at several time points in the data segment of the previous time window into the LSTM layer to obtain several first feature vectors, and input the several first feature vectors into the attention layer;

[0135] A seventh unit is used to input the water supply at the first time point in the data segment of the next time window into the attention layer to obtain a prompt vector, determine a weight vector of the first feature vector based on the prompt vector, and input a plurality of the first feature vectors and the weight vectors corresponding thereto into the fully connected layer;

[0136] An eighth unit is used for taking the water supply at the first time point in the data segment of the next time window as the output value of the fully connected layer, outputting the predicted value through the fully connected layer, constructing a loss function based on the predicted value and the output value, and training the initial water supply prediction network model through the loss function;

[0137] A calculation module 30, configured to obtain a predicted water supply through the final prediction network model, obtain a pipeline flow corresponding to a pipeline of a water supply system through the predicted water supply, obtain a pipeline flow rate according to the pipeline flow, obtain a first head loss according to the pipeline flow, obtain a second head loss according to the pipeline flow rate, obtain a predicted in-pipe water pressure based on the first head loss and the second head loss, and determine a predicted pipeline stress threshold according to the predicted in-pipe water pressure;

[0138] The analysis module 40 is used to obtain the real-time stress of the pipeline of the water supply system, and compare the real-time stress with the predicted pipeline stress threshold to determine the health status of the pipeline;

[0139] The analysis module 40 includes:

[0140] A ninth unit, for respectively obtaining the predicted hoop stress and the predicted axial stress of the pipeline of the water supply system through the predicted water pressure in the pipeline;

[0141] The tenth unit is used to obtain a predicted pipeline stress threshold value through the predicted annular stress and the predicted axial stress.

[0142] The present invention also provides a computer, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the water supply management method as described in the above technical solution when executing the computer program.

[0143] The present invention also provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, the water supply management method as described in the above technical solution is implemented.

[0144] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0145] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A water supply management method, characterized in that: The following steps are involved: Acquire an initial water supply data set, wherein the initial water supply data set includes a plurality of daily water supply data sets, wherein the daily water supply data sets include water supply at a plurality of time points, wherein the time points in different daily water supply data sets are the same, and perform missing value filling and outlier replacement on the initial water supply data set to obtain a water supply data set to be used; The water supply data set to be used is segmented by time windows to generate a number of data segments, and an initial water supply prediction network model is constructed. The initial water supply prediction network model includes an LSTM layer, an attention layer, and a fully connected layer. In two adjacent data segments, the data segment of the previous time window is used as an input value, and the water supply at the first time point in the data segment of the next time window is used as an output value. The initial water supply prediction network model is trained to obtain a final prediction network model; The step of training the initial water supply prediction network model by taking the data segment of the previous time window as input value and the water supply at the first time point in the data segment of the next time window as output value comprises: Input the water supply at several time points in the data segment of the previous time window into the LSTM layer to obtain several first feature vectors, and input the several first feature vectors into the attention layer; Input the water supply at the first time point in the data segment of the next time window into the attention layer to obtain a prompt vector, determine a weight vector of the first feature vector based on the prompt vector, and input a plurality of the first feature vectors and the weight vectors corresponding thereto into the fully connected layer; Using the water supply at the first time point in the data segment of the next time window as the output value of the fully connected layer, and outputting the predicted value through the fully connected layer, constructing a loss function based on the predicted value and the output value, and training the initial water supply prediction network model through the loss function; Obtaining a predicted water supply through the final prediction network model, obtaining a pipeline flow corresponding to a pipeline of a water supply system through the predicted water supply, obtaining a pipeline flow rate according to the pipeline flow, obtaining a first head loss according to the pipeline flow, and obtaining a second head loss according to the pipeline flow rate, obtaining a predicted water pressure in the pipe based on the first head loss and the second head loss, and determining a predicted pipeline stress threshold according to the predicted water pressure in the pipe; The formula for obtaining the pipeline flow rate is: ， in, Indicates the pipeline flow rate, Indicates the pipeline flow rate, Indicates the inner diameter of the pipeline; The calculation formula of the first head loss is: ， in, represents the first head loss, Indicates the characteristic coefficient of the water meter in the water supply system; The calculation formula of the second head loss is: ， in, represents the second head loss, It represents the resistance coefficient of the pipeline. represents the acceleration due to gravity; The step of obtaining the predicted water pressure in the pipe based on the first head loss and the second head loss comprises: Obtain the number of pipelines between the starting water supply point and the current pipeline and the water pressure of the starting water supply point; The water pressure at the starting water supply point is superimposed on the first head loss and the second head loss of the number of pipelines to obtain a predicted water pressure in the pipe; The real-time stress of the pipeline of the water supply system is obtained, and the real-time stress is compared with the predicted pipeline stress threshold to determine the health status of the pipeline.

2. The water supply management method according to claim 1, characterized in that: The step of filling missing values ​​and replacing outliers on the initial water supply data set comprises: Detecting whether there are missing points in the initial water supply data set; If the missing point exists, two adjacent points adjacent to the missing point are detected, and a filling water supply is generated based on the water supply at the time points corresponding to the two adjacent points, and the filling water supply is added to the missing point; Determine the daily average water supply in the daily water supply data set, and determine the abnormal threshold value through the daily average water supply and the water supply at the time point; Determining whether there is abnormal data in the daily water supply data set based on the daily average water supply and the abnormal threshold; If the abnormal data exists, an updated water supply amount is obtained based on the adjacent daily water supply data set, and the abnormal data is replaced by the updated water supply amount.

3. The water supply management method according to claim 2, characterized in that: The step of detecting whether there are missing points in the initial water supply data set comprises: Constructing a coordinate system with the time point as the horizontal coordinate and the water supply at the time point as the vertical coordinate, so as to map the initial water supply data set into a plurality of coordinate points; Obtaining the Euclidean distance between two adjacent coordinate points, and determining whether the Euclidean distance is greater than a distance threshold; If the Euclidean distance is greater than the distance threshold, it is determined that there is a missing point between the two coordinate points.

4. The water supply management method according to claim 2, characterized in that: The calculation formula for the average daily water supply is: ， in, represents the daily average water supply of the i-th daily water supply data set, represents the water supply at the jth time point in the i-th day water supply dataset, represents the total number of time points in the water supply dataset of the ith day; The calculation formula of the abnormal threshold is: ， in, represents the abnormal threshold of the water supply dataset on the ith day; The formula for obtaining the updated water supply is: ， in, represents the updated water supply at the jth time point in the i-th day water supply dataset, represents the water supply at the jth time point in the i+1th daily water supply dataset, represents the time point of the jth time point in the i-1th daily water supply dataset, , are weight coefficients, and .

5. The water supply management method according to claim 1, characterized in that: The formula for obtaining the weight vector is: ， in, represents the weight vector corresponding to the mth first eigenvector, represents the mth first eigenvector, Represents the transposition character, represents the prompt vector, represents the total number of first eigenvectors in the data segment of the previous time window, Represents an exponential function.

6. The water supply management method according to claim 1, characterized in that: The step of determining the predicted pipeline stress threshold according to the predicted water pressure in the pipe comprises: The predicted annular stress and the predicted axial stress of the pipeline of the water supply system are respectively obtained by the predicted water pressure in the pipeline; The predicted pipeline stress threshold is obtained through the predicted annular stress and the predicted axial stress.

7. The water supply management method according to claim 6, characterized in that: The calculation formula for the predicted hoop stress is: ， in, represents the predicted hoop stress, represents the predicted water pressure in the pipe, Indicates the inner diameter of the pipe. Indicates the wall thickness of the pipe; The calculation formula for the predicted axial stress is: ， in, represents the predicted axial stress, represents the interface internal force of the pipeline, represents the cross-sectional area of ​​the pipeline, Indicates the outer diameter of the pipe. represents the bending moment acting on the pipeline, Indicates the section coefficient of the pipeline; The calculation formula for predicting the pipeline stress threshold is: ， in, Indicates the predicted pipeline stress threshold.

8. A water supply management system, applied to the water supply management method according to any one of claims 1 to 7, characterized in that: The system comprises: A collection module is used to obtain an initial water supply data set, wherein the initial water supply data set includes a plurality of daily water supply data sets, wherein the daily water supply data sets include water supply at a plurality of time points, wherein the time points in different daily water supply data sets are the same, and the initial water supply data set is filled with missing values ​​and replaced with abnormal values ​​to obtain a water supply data set to be used; A processing module, used to perform data segmentation on the water supply data set to be used by time windows to generate a plurality of data segments, and construct an initial water supply prediction network model, wherein the initial water supply prediction network model comprises an LSTM layer, an attention layer and a fully connected layer, and in two adjacent data segments, the data segment of the previous time window is used as an input value, and the water supply at the first time point in the data segment of the next time window is used as an output value, and the initial water supply prediction network model is trained to obtain a final prediction network model; The processing module comprises: A sixth unit is used to input the water supply at several time points in the data segment of the previous time window into the LSTM layer to obtain several first feature vectors, and input the several first feature vectors into the attention layer; A seventh unit is used to input the water supply at the first time point in the data segment of the next time window into the attention layer to obtain a prompt vector, determine a weight vector of the first feature vector based on the prompt vector, and input a plurality of the first feature vectors and the weight vectors corresponding thereto into the fully connected layer; An eighth unit is used for taking the water supply at the first time point in the data segment of the next time window as the output value of the fully connected layer, outputting a predicted value through the fully connected layer, constructing a loss function based on the predicted value and the output value, and training the initial water supply prediction network model through the loss function; The processing module is also used to obtain the number of pipelines between the starting water supply point and the current pipeline and the water pressure of the starting water supply point; superimpose the water pressure of the starting water supply point on the first head loss and the second head loss of the number of pipelines to obtain the predicted water pressure in the pipe; a calculation module, configured to obtain a predicted water supply through the final prediction network model, obtain a pipeline flow corresponding to a pipeline of a water supply system through the predicted water supply, obtain a pipeline flow rate according to the pipeline flow, obtain a first head loss according to the pipeline flow, obtain a second head loss according to the pipeline flow rate, obtain a predicted in-pipe water pressure based on the first head loss and the second head loss, and determine a predicted pipeline stress threshold according to the predicted in-pipe water pressure; The formula for obtaining the pipeline flow rate is: ， in, Indicates the pipeline flow rate, Indicates the pipeline flow rate, Indicates the inner diameter of the pipeline; The calculation formula of the first head loss is: ， in, represents the first head loss, Indicates the characteristic coefficient of the water meter in the water supply system; The calculation formula of the second head loss is: ， in, represents the second head loss, It represents the resistance coefficient of the pipeline. represents the acceleration due to gravity; The analysis module is used to obtain the real-time stress of the pipeline of the water supply system, and compare the real-time stress with the predicted pipeline stress threshold to determine the health status of the pipeline.

Images