Online distribution method for water demand of nodes of water supply network

Abstract

The embodiment of the invention provides a water supply network node water demand on-line distribution method, which relates to the field of water supply networks, and comprises the following steps: calculating flow data in unit time according to working condition information of adjacent moments, and then calculating water demand of different network nodes according to the flow data in unit time and the like. And then a pressure simulation value is obtained through calculation of the water supply network hydraulic model, the pressure simulation value is compared with a pressure monitoring value actually monitored by the monitoring equipment, the water demand corresponding to the pressure simulation value with the error meeting the preset condition is determined as the water demand of the corresponding node, and therefore the water real-time distribution model is trained. Under the condition that the water supply system operates in real time, the working condition information at the current moment is input into the trained water volume real-time distribution model, the water demand of the pipe network node can be rapidly and accurately determined, the simulation precision of the water supply pipe network hydraulic model is effectively improved, and therefore the problems that in the prior art, errors are large, cost is high and implementation is difficult are solved.

Description

Technical Field

[0001] The present application relates to the field of water supply networks, and in particular to a method for online distribution of water demand at nodes in a water supply network. Background Art

[0002] The urban water supply network is a vital part of urban infrastructure. It undertakes the important task of providing drinking water to residents, industrial and commercial areas. The pipe network system meets the daily life and production needs of urban residents and various facilities by transporting and distributing water resources, thus ensuring the normal operation and development of the city.

[0003] As an important tool for pipe network design and operation, the pipe network hydraulic model simulates the characteristics and behavior of water flow in the pipe network through mathematical and physical principles. It can help engineers predict the flow rate, flow rate, pressure and head loss of water in the pipe, so as to effectively optimize the design of the pipe network, improve water supply efficiency, reduce energy consumption, and ensure the safe and stable operation of the pipe network system. An important parameter in the pipe network hydraulic model is the node water demand, which describes the water demand of each node (such as community, residential building, building, factory, etc.). Reasonable and accurate determination of node water demand helps the pipe network system to reasonably allocate water resources, ensure the normal water use of each node, and minimize waste and loss. Therefore, accurate determination of node water demand is of great significance to the planning, design and operation management of the pipe network system.

[0004] At present, the main methods for determining node water demand in hydraulic models are average distribution method, flow ratio method and distribution based on data analysis results of meter reading system. Among them, the first two methods ignore the differences in the number of people and water consumption along the water supply line, and the water demand obtained is quite different from the actual situation. The third method requires a large amount of meter reading data, which is costly and cumbersome to operate, and is difficult to implement in practice. Summary of the invention

[0005] The embodiment of the present application provides an online allocation method for water demand of water supply network nodes, which can quickly obtain a distribution plan for node water demand based on the current operating condition information of the water supply network, so as to solve the problems of large error and high cost in the current method for determining node water demand.

[0006] The present application provides a method for online distribution of water demand at a water supply network node, comprising: Establishing a hydraulic model of the water supply network according to basic information of the water supply network, wherein the basic information includes water supply pipeline information and network node information; Acquire the working condition information of the water supply network at different times, wherein the working condition information includes the operating status of valves and water pumps and the location information and monitoring data of monitoring equipment, wherein the monitoring data includes pressure monitoring values; Determine the closest pipe network node in the water supply pipe network hydraulic model according to the location information, and associate the pipe network node with the corresponding monitoring equipment; Calculate the flow data per unit time based on the working condition information at adjacent moments; Calculate the water demand of different pipe network nodes according to the basic information, flow data per unit time and working condition information at the corresponding time; Input the water demand of different pipe network nodes into the water supply pipe network hydraulic model to obtain the pressure simulation values ​​of different pipe network nodes, compare the pressure simulation values ​​of different pipe network nodes with the corresponding pressure monitoring values, and store the water demand and flow data corresponding to the pressure simulation values ​​whose errors meet the preset conditions into the data set; Establishing a real-time water allocation model, training the real-time water allocation model according to the data set, and saving the real-time water allocation model after the training; The current working condition information is input into the trained real-time water distribution model to obtain the water demand of the pipe network node.

[0007] In a feasible implementation, after the flow data per unit time is calculated according to the operating condition information at adjacent moments, the flow data per unit time and the operating status of the valve and the water pump at the corresponding moment are associated and stored.

[0008] In a feasible implementation, the basic information also includes water pump information, valve information and reservoir information.

[0009] In a feasible implementation, the water supply pipeline information includes pipeline code, pipe diameter, pipe length, Hazen-Williams coefficient, start node code and end node code; The pipe network node information includes node code, node coordinates and node elevation; The water pump information includes water pump code, starting node code, ending node code, water pump type and head-flow curve; The valve information includes valve code, start node code, end node code and valve type; The reservoir information includes reservoir code, reservoir coordinates and water level elevation.

[0010] In a feasible implementation, the monitoring device includes a flow meter and a pressure meter, and the monitoring data also includes instantaneous flow and accumulated flow; The calculating the flow data per unit time according to the working condition information at adjacent moments includes: calculating the flow data per unit time according to the accumulated flow at adjacent moments.

[0011] In a feasible implementation, the step of inputting the water demand of different pipe network nodes into the water supply pipe network hydraulic model to obtain the pressure simulation values ​​of different pipe network nodes includes: Calculating the head loss according to the water demand of different pipe network nodes; The pressure simulation value of the pipe network node is determined according to the head loss.

[0012] In a feasible implementation, the calculation formula of the head loss is:

[0013] In the formula, is the head loss, is the HW coefficient, is the pipe diameter, For the tube leader, is the instantaneous flow rate in the pipe.

[0014] In a feasible implementation, the data set is divided into a training set, a validation set and a test set, and the ratio of the training set, the validation set and the test set is 7:2:1.

[0015] In a feasible implementation, the real-time water distribution model is established based on the LightGBM algorithm.

[0016] In a feasible implementation, the preset condition is that the difference between the pressure simulation value and the pressure monitoring value is less than 5% of the pressure monitoring value.

[0017] The embodiment of the present application provides an online water demand allocation method for water supply network nodes, which calculates the flow data per unit time based on the working condition information at adjacent moments, and then calculates the water demand of different network nodes based on the flow data per unit time, thereby obtaining a pressure simulation value, which is then compared with the pressure monitoring value actually monitored by the monitoring equipment, and the water demand corresponding to the pressure simulation value whose error meets the preset conditions is determined as the water demand of the corresponding node, so as to train the real-time water distribution model. When the water supply system is running in real time, the working condition information at the current moment is input into the trained real-time water distribution model, which can quickly and accurately determine the water demand of the network nodes, effectively improve the simulation accuracy of the hydraulic model of the water supply network, and thus solve the problems of large errors, high costs and difficulty in implementation in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of an online water demand allocation method for a water supply network node provided in an embodiment of the present application; Figure 2 It is a schematic diagram of a flow chart of calculating pressure simulation values ​​of different pipe network nodes provided by an embodiment of the present application; Figure 3 It is a flow chart of establishing a real-time water distribution model provided by an embodiment of the present application; Figure 4 It is a schematic diagram of a hydraulic model of a water supply network with a simple topology provided in an embodiment of the present application. DETAILED DESCRIPTION

[0019] In order to enable those skilled in the art to better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work should fall within the scope of protection of the present application.

[0020] At present, the methods for determining the node water demand in the hydraulic model mainly include the following: 1. Average distribution method: Count the total number of nodes in the area between the two flow meters and evenly distribute the total water demand to each node.

[0021] Among them, for two flow meters connected in series in the water supply network, the monitoring value of the upstream flow meter minus the monitoring value of the downstream flow meter is the total water demand of the area between the two flow meters.

[0022] 2. Specific flow method: This method is used to calculate the flow along the line, assuming that the water volume is evenly distributed on the main pipeline, so as to calculate the flow rate per unit length of the pipeline. That is, the ratio of the water demand at each node to the total water demand is determined by the pipeline length distance between the node and the upstream flow meter.

[0023] 3. Allocate water according to the data analysis results of the meter reading system; associate each node with the water user, obtain the meter reading data of all water users in the area through the meter reading system for analysis, determine the water consumption ratio of each node at different times, and then allocate the total water demand to each node according to the ratio.

[0024] Among the above methods, methods 1 and 2 are relatively simple to implement in practice, but they have certain defects, because this method ignores the differences in the number of people and water consumption along the water supply line, so the node water demand obtained by this method is quite different from the actual situation; method 3 takes into account the relationship between the number of people and water consumption, but because the meter reading cycle is generally once a day (remote water meter) or even once a month (mechanical water meter), it is impossible to obtain water volume information on a finer time scale. At the same time, associating water users with hydraulic model nodes is a very labor-intensive task, because the location of water users, water consumption information and information of hydraulic model nodes are stored in different systems and cannot be matched one by one. Therefore, this method is difficult to implement in practice.

[0025] Based on the above problems, an embodiment of the present application provides an online allocation method for water demand at water supply network nodes, which can quickly and accurately determine the node water demand. It aims to solve the problems of large node water demand error and low hydraulic simulation accuracy when the online hydraulic model of the water supply network is simulated in real time. It can improve the simulation accuracy of the online real-time simulation of the hydraulic model, provide data support for intelligent business, and thus improve the efficiency and reliability of the entire water supply system.

[0026] Figure 1 It is a flow chart of an online water demand allocation method for a water supply network node provided in an embodiment of the present application; Figure 2 It is a schematic diagram of a flow chart of calculating pressure simulation values ​​of different pipe network nodes provided by an embodiment of the present application; Figure 3 1 is a flow chart of establishing a real-time water distribution model provided by an embodiment of the present application. Figure 1-Figure 3 As shown, an embodiment of the present application provides a method for online distribution of water demand at a water supply network node, comprising: S100: Establishing a hydraulic model of a water supply network according to basic information of the water supply network, wherein the basic information includes water supply pipeline information and network node information.

[0027] In some examples, the basic information also includes water pump information, valve information, and reservoir information.

[0028] The water supply pipeline information includes pipeline code, pipe diameter, pipe length, Hazen-Williams coefficient, start node code and end node code; The pipe network node information includes node code, node coordinates and node elevation; The water pump information includes water pump code, starting node code, ending node code, water pump type and head-flow curve; The valve information includes valve code, start node code, end node code and valve type; The reservoir information includes reservoir code, reservoir coordinates and water level elevation.

[0029] It should be noted that after the hydraulic model of the water supply network is established based on the network adjustment theory, the network topology needs to be verified to ensure the correctness of the network topology.

[0030] In some examples, the hydraulic model of the water supply network may be established using open source Epanet software, commercial WaterGems software, or the like.

[0031] Model validation can be done in two ways: 1) Conduct field investigation to determine whether the data on the actual pipeline network, such as its depth and direction, are consistent with the basic information; 2) Hydraulic solution operation: If there are problems with the topology (such as isolated points, unconnected pipes, etc.), the hydraulic model of the water supply network will report an error during the solution, and then a field investigation is required to ensure that the data is true and correct.

[0032] It is easy to understand that other methods can also be used to verify the model, and this application does not limit this.

[0033] S200: Acquire the working condition information of the water supply network at different times, wherein the working condition information includes the operating status of valves and water pumps and the location information and monitoring data of monitoring equipment, and the monitoring data includes pressure monitoring values.

[0034] In some examples, the monitoring device includes a flow meter and a pressure meter, and the monitoring data also includes instantaneous flow and accumulated flow.

[0035] It should be noted that in the hydraulic model of the water supply network, a pressure gauge is installed at each network node.

[0036] In some examples, a monitoring data time series can be established, including a cumulative flow time series and a pressure time series. For example, the cumulative flow at time t is f and the pressure monitoring value is p, the cumulative flow at time t1 is f1 and the pressure monitoring value is p1, and so on.

[0037] S300: Determine the closest pipe network node in the water supply pipe network hydraulic model according to the location information, and associate the pipe network node with the corresponding monitoring equipment.

[0038] It is easy to understand that associating the pipe network node with the corresponding monitoring equipment facilitates the subsequent calculation of water demand using the corresponding monitoring data of the pipe network node, and after obtaining the pressure simulation value of the pipe network node, it is compared with the corresponding monitoring data.

[0039] S400: Calculate the flow data per unit time according to the working condition information at adjacent moments.

[0040] The calculating of the flow data per unit time according to the working condition information at adjacent moments includes: calculating the flow data per unit time according to the accumulated flow at adjacent moments.

[0041] According to the established monitoring data time series, the difference between the cumulative flow at time t1 and the cumulative flow at time t is calculated, which is the flow data per unit time. It is easy to understand that t1=t+1, and the time difference between adjacent moments is 1h.

[0042] In some examples, after the flow data per unit time is calculated based on the operating condition information at adjacent moments, the flow data per unit time and the operating status of the valve and the water pump at the corresponding moment are associated and stored.

[0043] S500: Calculate the water demand of different pipe network nodes according to the basic information, the flow data per unit time and the working condition information at the corresponding time.

[0044] In some examples, an evolutionary genetic algorithm may be used to optimize and solve the water demand of the pipe network nodes. In other examples, a particle swarm algorithm, an annealing algorithm, a genetic algorithm, etc. may also be used, and this application does not limit this.

[0045] S600: Input the water demand of the different pipe network nodes into the hydraulic model of the water supply pipe network to obtain the pressure simulation values ​​of the different pipe network nodes, compare the pressure simulation values ​​of the different pipe network nodes with the corresponding pressure monitoring values, and store the water demand and flow data corresponding to the pressure simulation values ​​whose errors meet the preset conditions into the data set.

[0046] It should be noted that the preset condition may be in the form of a percentage, for example, the difference between the pressure monitoring value and the pressure simulation value is within 3% of the pressure monitoring value, or within 5%, and needs to be set according to actual conditions.

[0047] In some examples, inputting the water demand of the different pipe network nodes into the water supply pipe network hydraulic model to obtain the pressure simulation values ​​of the different pipe network nodes includes: Calculating the head loss according to the water demand of different pipe network nodes; The pressure simulation value of the pipe network node is determined according to the head loss.

[0048] Wherein, the calculation formula of the head loss is:

[0049] In the formula, is the head loss, is the HW coefficient, is the pipe diameter, For the tube leader, is the instantaneous flow rate in the pipe.

[0050] S700: Establish a real-time water allocation model, train the real-time water allocation model according to the data set, and save the trained real-time water allocation model.

[0051] In some examples, the real-time water distribution model is established based on the LightGBM algorithm. In other examples, other machine learning and deep learning models, such as convolutional neural networks, long short-term memory, etc., can also be used.

[0052] In some examples, the data set is divided into a training set, a validation set, and a test set, and the ratio of the training set, the validation set, and the test set is 7:2:1.

[0053] S800: Inputting the current working condition information into the trained real-time water distribution model to obtain the water demand of the pipe network node.

[0054] Figure 4 This is a schematic diagram of a hydraulic model of a water supply network with a simple topology provided in an embodiment of the present application. The technical solution of the present application is further described below in conjunction with a specific example.

[0055] The basic information of the water supply network is shown in Table 1-3:

[0056] Table 1 Pipeline network node information

[0057] Table 2 Water supply pipeline information

[0058] Table 3 Reservoir information like Figure 4 As shown in the figure, in the above water supply network hydraulic model, there are 1 reservoir, 3 water supply pipelines, 3 network nodes and 4 flow meters. The water supply network hydraulic model takes water from reservoir R1 and flows into the water supply pipeline by gravity. Among them, network nodes J1, J2 and J3 can be understood as water intake points. Water users such as residential areas and factories around each network node take water from the corresponding network node.

[0059] Assume that the monitoring value of flow meter 1# is 3600 under the current working conditions , the monitoring value of flow meter 2# is 2400 , the monitoring value of flow meter 3# is 1000 .

[0060] The monitoring value of the flow meter can be understood as the total water consumption downstream of the flow meter. Therefore, it can be concluded that the water consumption of node 1 is 3600-2400=1200 , the water consumption of node 2 is 2400-1000=1400 , the water consumption of node 3 is 1000 .

[0061] It can be seen from the calculation formula of head loss that head loss is related to HW coefficient, pipe diameter, pipe length and instantaneous flow rate. The HW coefficient, pipe diameter and pipe length can be obtained through the basic information of the water supply network, while the instantaneous flow rate changes with the working conditions. The instantaneous flow rate can be measured by a flow meter, but due to the high cost of flow meters, it is impossible to arrange a flow meter at each node in the network. Therefore, the instantaneous flow rate of the network node where the flow meter is not installed needs to be calculated by the water demand of the network node. According to the balance requirements of the flow continuity equation in the hydraulic model solution, the water demand of the network node = the inflow pipe flow - the outflow pipe flow. The instantaneous flow rate of the network node to be calculated is equal to the sum of the water demands of all network nodes downstream of the network node. In order to determine the head loss calculation formula for each pipeline , you need to know the water demand of each pipe network node J1, J2, and J3.

[0062] Assuming that the values ​​of flow meters 2# and 3# are unknown, or that flow meters are not installed at the network nodes J1 and J2, water demands are assigned to the three network nodes artificially, and then input into the hydraulic model of the water supply network to calculate the head loss, obtain the pressure simulation values ​​of the network nodes J1 and J2, and then compare them with the pressure monitoring values ​​at the network nodes J1 and J2. The water demand of each network node corresponding to the pressure simulation value when the error is minimum is used as the water allocation plan.

[0063] It should be noted that when water demand is assigned artificially, the water demand must be a positive number.

[0064] It can be seen from the above embodiments that the embodiments of the present application provide a method for online allocation of water demand at water supply network nodes, which simulates the water demand of each network node through past data, and trains the real-time water allocation model according to the simulated water demand. When the water supply system is running in real time, the current operating condition information is input into the trained real-time water allocation model, which can quickly and accurately determine the water demand of the network nodes, and effectively improve the simulation accuracy of the hydraulic model of the water supply network, thereby solving the problems of large errors, high costs and difficulty in implementation in the prior art.

[0065] It is easy to understand that those skilled in the art can combine, split, reorganize, etc. the embodiments of the present application to obtain other embodiments based on the several embodiments provided in the present application, and these embodiments do not exceed the protection scope of the present application.

[0066] The above specific implementation methods further explain in detail the purpose, technical solutions and beneficial effects of the embodiments of the present application. It should be understood that the above are only specific implementation methods of the embodiments of the present application and are not used to limit the protection scope of the embodiments of the present application. Any modifications, equivalent substitutions, improvements, etc. made on the basis of the technical solutions of the embodiments of the present application should be included in the protection scope of the embodiments of the present application.

Claims

1. A method for online distribution of water demand at a water supply network node, characterized in that: include: Establishing a hydraulic model of the water supply network according to basic information of the water supply network, wherein the basic information includes water supply pipeline information and network node information; Acquire the working condition information of the water supply network at different times, wherein the working condition information includes the operating status of valves and water pumps and the location information and monitoring data of monitoring equipment, wherein the monitoring data includes pressure monitoring values; Determine the closest pipe network node in the water supply pipe network hydraulic model according to the location information, and associate the pipe network node with the corresponding monitoring equipment; Calculate the flow data per unit time based on the working condition information at adjacent moments; Calculate the water demand of different pipe network nodes according to the basic information, flow data per unit time and working condition information at the corresponding time; Input the water demand of different pipe network nodes into the water supply pipe network hydraulic model to obtain the pressure simulation values ​​of different pipe network nodes, compare the pressure simulation values ​​of different pipe network nodes with the corresponding pressure monitoring values, and store the water demand and flow data corresponding to the pressure simulation values ​​whose errors meet the preset conditions into the data set; Establishing a real-time water allocation model, training the real-time water allocation model according to the data set, and saving the real-time water allocation model after the training; The current working condition information is input into the trained real-time water distribution model to obtain the water demand of the pipe network node.

2. The method for online water demand allocation of water supply network nodes according to claim 1, characterized in that: After the flow data per unit time is calculated according to the working condition information at adjacent moments, the flow data per unit time and the operating states of the valve and the water pump at the corresponding moments are stored in association.

3. The method for online water demand allocation of water supply network nodes according to claim 1, characterized in that: The basic information also includes water pump information, valve information and reservoir information.

4. The method for online water demand allocation of water supply network nodes according to claim 3 is characterized in that: The water supply pipeline information includes pipeline code, pipe diameter, pipe length, Hazen-Williams coefficient, start node code and end node code; The pipe network node information includes node code, node coordinates and node elevation; The water pump information includes water pump code, starting node code, ending node code, water pump type and head-flow curve; The valve information includes valve code, start node code, end node code and valve type; The reservoir information includes reservoir code, reservoir coordinates and water level elevation.

5. The method for online distribution of water demand at a water supply network node according to any one of claims 1 to 4, characterized in that: The monitoring equipment includes a flow meter and a pressure gauge, and the monitoring data also includes instantaneous flow and accumulated flow; The calculating the flow data per unit time according to the working condition information at adjacent moments includes: calculating the flow data per unit time according to the accumulated flow at adjacent moments.

6. The method for online distribution of water demand at a water supply network node according to any one of claims 1 to 4, characterized in that: The step of inputting the water demand of different pipe network nodes into the water supply pipe network hydraulic model to obtain pressure simulation values ​​of different pipe network nodes includes: Calculating the head loss according to the water demand of different pipe network nodes; The pressure simulation value of the pipe network node is determined according to the head loss.

7. The method for online water demand allocation of water supply network nodes according to claim 6, characterized in that: The calculation formula of the head loss is: ； In the formula, is the head loss, is the HW coefficient, is the pipe diameter, For the tube leader, is the instantaneous flow rate in the pipe.

8. The method for online water demand allocation of water supply network nodes according to any one of claims 1 to 4, characterized in that: The data set is divided into a training set, a validation set and a test set, and the ratio of the training set, the validation set and the test set is 7:2:

1.

9. The method for online distribution of water demand at a water supply network node according to any one of claims 1 to 4, characterized in that: The real-time water allocation model is established based on the LightGBM algorithm.

10. The method for online distribution of water demand at a water supply network node according to any one of claims 1 to 4, characterized in that: The preset condition is that the difference between the pressure simulation value and the pressure monitoring value is less than 5% of the pressure monitoring value.

Images