A method for generating simulated flow and water quality data at drainage pipe network locations
By generating node-level flow and water quality data for drainage pipe networks through Fourier fitting and phase shifting, the problems of insufficient total quantity conservation and randomness in existing technologies are solved, realizing efficient and low-cost dynamic simulation data generation, which is applicable to drainage pipe networks of different scales.
Patent Information
- Application Number
- CN202611081754.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-21
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2046-07-21
AI Technical Summary
Existing technologies struggle to generate dynamic water quality and quantity data in drainage networks, especially node-level data. They fail to meet the total conservation of flow and water quality, and the high monitoring costs and complex maintenance lead to inaccurate simulation results.
The flow rate and water quality change functions are generated by using a Fourier fitting function. Combined with the pipeline network model, the flow rate and water quality changes at each sewage inflow node are calculated to ensure total conservation. Phase offset and noise fluctuation are introduced to simulate randomness and generate dynamic data for each point.
Based on the monitoring data of the total regional outlet, node-level flow and water quality data that conform to the total conservation and have random fluctuations can be generated, which reduces monitoring costs and improves the accuracy and reliability of simulation.
Smart Images

Figure CN122595929B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of municipal drainage network monitoring and numerical simulation technology, and more specifically, to a method for generating simulated flow and water quality data at drainage network points. Background Technology
[0002] Quantitatively assessing external water intrusion (groundwater / river backflow, etc.) using regional water quality-flow balance relationships is a common method in drainage network inspection and operation and maintenance engineering. This method uses the flow rate and characteristic pollutant concentration at the regional total outlet as a benchmark, and infers the external water volume based on the mass conservation principle. When applied in engineering implementation, it generally relies on two types of data: first, water quality and quantity monitoring data at the regional total outlet; and second, baseline water quality and quantity data at each pollutant-generating point (manhole / node) within the network. To determine the feasibility and accuracy of this method under fluctuating and uncertain conditions, numerical models of the drainage network, such as the SWMM, are constructed. Virtual experiments are conducted by setting specific external water inflow characteristics and wastewater background values. The SWMM model discretizes the drainage network into a network composed of nodes, pipe segments, pumping stations, and storage facilities. It solves unsteady flow based on the Saint-Venant equations and can couple one-dimensional convection-diffusion equations to simulate pollutant migration, making it a key tool for urban drainage system planning, operation, maintenance, and optimization.
[0003] However, in practical applications, node-level dynamic input data still has shortcomings: limited by the installation cost, maintenance complexity, and spatial distribution characteristics of monitoring equipment, actual projects can usually only obtain summary data of the total outlet of the area, making it difficult to achieve synchronous monitoring of numerous nodes (especially small nodes) within the pipeline network, and unable to obtain flow and water quality fluctuations at each point; during numerical simulation, constant or simple regular inflow conditions are usually used, or only a single total curve is relied upon for average distribution, without considering the flow fluctuations, water quality concentration changes, and randomness of discharge time patterns caused by differences in living habits in different areas, so the generated data deviates from the actual working conditions; and directly randomly generated node data cannot guarantee that the sum of the flow at each point equals the total flow in the area, and it is difficult to guarantee the balance between the water quality at each point and the total water quality in the area, resulting in simulation results that do not conform to the law of conservation of mass, making it difficult to support the verification of water quality and quantity balance methods; field experiments also make it difficult to accurately obtain the water quality and quantity fluctuations of the background sewage in the pipeline network, make it impossible to achieve comprehensive point sampling, and make it difficult to accurately grasp the information of external water intrusion, making it difficult to achieve verification based on real data.
[0004] Therefore, there is an urgent need for a method to generate dynamic water quality and quantity data, which can generate dynamic water quality and quantity data at each point in the pipeline network based on regional total monitoring data, and make the dynamic water quality and quantity data both random and volatile, and satisfy the total amount conservation constraint, so as to solve the problem of missing real inflow data at nodes in numerical simulation studies. Summary of the Invention
[0005] To achieve the above objectives, this application provides a method for generating simulated flow and water quality data at drainage network locations, comprising the following steps: A pipe network model is established based on the actual drainage pipe network topology of the study area; pollution-generating units and total emission monitoring points are determined in the pipe network model; the pollution-generating units are the monitoring points. Flow rate and characteristic pollutant concentrations were collected at the study area corresponding to the total flow monitoring points to obtain total flow rate and total water quality sequences; these sequences were then fitted using Fourier functions. The total flow rate fitting function F is obtained by fitting the total flow rate sequence. The total water quality fitting function G is obtained by fitting the total water quality sequence. ; Generate the flow rate change function for each wastewater inflow node. ; Generate water quality change function for each wastewater inflow node ; Based on flow change function and water quality change function The flow time series and water quality time series of each point are generated by combining the pipeline network model.
[0006] Among them, the Fourier fitting function Represented as: ,in, The Fourier fitting function is... The DC component, , Here, represents the Fourier coefficients of the nth harmonic, and T is the function period. These are harmonic orders.
[0007] Among them, the flow rate change function for each wastewater inflow node is generated. Includes the following steps: Calculate the flow contribution rate at each location. Traffic contribution rate satisfy N = number of points; Calculate the phase offset at each point Phase shift This indicates that the peak traffic flow at a given location has a random shift over time; Combining traffic contribution rate With phase shift The flow rate change function at point i is generated by the total flow rate fitting function F(t). , represented as: ,in, Let be the flow rate at the i-th point at time t.
[0008] Calculate the flow contribution rate at each location. Includes the following steps: The contribution rate of each point i to the total flow is generated based on the normal distribution. , represented as: ,in, Let μ be the flow contribution rate of the i-th point, where μ is the mean and μ = 1 / N, and N is the number of points; σ 2 For variance; Contribution to traffic Applying constraints is represented as: ,in, This is a lower limit for the value to avoid invalid points with zero or minimal contribution; This is the upper limit of the value, used to prevent an imbalance in traffic distribution caused by a single point having an excessively high proportion; Traffic contribution rate After constraint truncation and normalization, .
[0009] Among them, the phase offset of each point The values follow a uniform distribution, represented as: ,in, To ensure uniform distribution, Let be the phase offset of the i-th point relative to the total flow. This represents the duration of the maximum phase shift.
[0010] Among them, the water quality change function for each wastewater inflow node is generated. Includes the following steps: The flow weight and time average of each flow point are calculated based on the flow rate at each point, and are expressed as follows: , ,in, For the i-th point at time... Traffic weight, where M is the total number of time nodes; Let be the average weight of the i-th point across all M time points; Generate basic volatility coefficient For the basic volatility coefficient The modified base volatility coefficient is generated by performing a correction. ; Generate noise fluctuations and determine corrected noise. ; The adjusted base volatility coefficient for each point at each time node With corrected noise Superimposed, a discrete sequence of water quality data at each location is obtained; the discrete sequence of water quality data at each location is then fitted using a Fourier function. Perform Fourier fitting and output the water quality change function at each point, expressed as: .
[0011] Among them, the basic volatility coefficient Generated through a uniform distribution, represented as: ,in, Let be the initial basic volatility coefficient for the i-th point. For fluctuation amplitude parameters, ; Using the average weight of each point For the basic volatility coefficient The modified base volatility coefficient is generated by performing a correction. , represented as: ,in, For the corrected basic volatility coefficient, The average weight of the i-th point This represents the weighted average concentration across all locations.
[0012] The generated noise fluctuations include: The initial noise for each point and time node is generated based on a normal distribution, and is expressed as follows: , ,in, Initial noise, The standard deviation of noise. This is the noise proportionality factor, and ; The weighted error of the initial noise is calculated and evenly distributed to each point, as shown below: ,in, This is the weighted error; The corrected noise is calculated and expressed as: ,in, For weighted error, To correct noise; Verify the weighted average error caused by the correction noise. If the set threshold is met, then the noise is corrected; otherwise, the noise is regenerated until the constraint is met.
[0013] Furthermore, the weighted average error The calculation method is as follows: .
[0014] Furthermore, the generation of flow time series and water quality time series at each location by combining the pipeline network model is achieved through a program, including the following steps: Determine the corresponding location of each pollution-generating unit in the pipeline network model within the study area, determine the start time, end time, and total duration of the simulation, and determine the sampling data within the study area; Based on the flow change function and water quality change function The flow rate and water quality at each time point are calculated node by node, forming a flow rate time series object and a water quality time series object; Read the target INP file of the pipeline network model, locate or create the time series module area, and write the flow time series objects and water quality time series objects of each node into the target INP file according to the SWMM specification format. Generate compliant external inflow configuration statements based on the target INP file; Run the SWMM model to complete the dynamic hydraulic and water quality simulation of the pipeline network, and generate the flow time series and water quality time series of each point.
[0015] This invention enables a modeling approach that generates component data in reverse from total quantity constraints. It addresses the challenges of numerical simulation studies, such as the lack of real inflow data at nodes and the difficulty in generating simulation data that combines randomness and conservation. Furthermore, the data acquisition cost is low, requiring only total quantity monitoring data from the regional outlet to reconstruct the dynamic water quality and quantity data of each node within the pipe network. This overcomes the limitations of high cost, difficult maintenance, and lack of synchronous monitoring for node-level monitoring equipment. The generated point data follows a probability distribution (normal / uniform distribution) while simultaneously satisfying Fourier periodicity and dual total quantity conservation constraints for flow and water quality, more closely resembling the actual operating state of the pipe network. This invention provides a near-realistic virtual simulation scenario for areas lacking node monitoring, offering data support for quantitative identification of external water intrusion, drainage pattern research, pipe network hydraulic characteristic analysis, and optimized scheduling. Moreover, the implementation method is highly versatile: applicable to drainage pipe networks of different scales and with varying numbers of nodes, with parameters flexibly set according to the actual region. Attached Figure Description
[0016] Figure 1 This is a flowchart illustrating the steps of a method for generating simulated flow and water quality data at drainage network locations according to an embodiment of the present invention. Figure 2 This is a schematic diagram of a pipeline network model provided according to an embodiment of the present invention; Figure 3 This is a schematic diagram comparing the total flow fitting curve with the original data according to an embodiment of the present invention; Figure 4 This is a schematic diagram comparing the flow distribution and total flow at each point, generated by the method for generating simulated flow and water quality data at each point of the drainage network according to an embodiment of the present invention. Figure 5 This is a schematic diagram comparing the Fourier fitting curves of water quality at various points generated by the method for generating simulated flow water quality data at drainage pipe network points according to an embodiment of the present invention. Detailed Implementation
[0017] The specific implementation of the present invention will now be described in detail with reference to the accompanying drawings.
[0018] The method for generating simulated flow and water quality data at drainage pipe network points provided by this invention is as follows: Figure 1 As shown, it includes the following steps: Step S100: Establish a pipe network model based on the actual drainage pipe network topology of the study area; determine the pollution-generating units and total emission monitoring points in the pipe network model; In this step, the study area is divided into several catchment sub-regions, and the following nodes are determined in the pipeline network model accordingly: 1) There are N pollution-generating units, each corresponding to a sewage inflow node (i.e., inspection well) in the actual drainage network, representing the discharge point where sewage from a certain residential area is collected; the sewage inflow node in the actual drainage network and the pollution-generating units in the network model are both points; 2) One total monitoring point, namely the total outlet of the area in the actual drainage network.
[0019] This invention provides specific implementation examples: such as Figure 2 As shown, a SWMM model was constructed, containing 20 sewage inflow nodes (points 1-20, representing inspection wells collecting sewage from different residential areas) and 1 total discharge monitoring point (point 21, used to collect and calculate the total discharge of upstream sewage). The unified parameters of the pipeline network in the model are: pipe segment length 200m, maximum confluence length 1600m, pipe diameter 0.8m, slope 0.001, and other parameters adopt the default values of SWMM.
[0020] Step S110: Collect flow rate and characteristic pollutant concentration at the study area location corresponding to the total flow monitoring point to obtain the total flow rate sequence and total water quality sequence; generate the continuous function F(t) of the total flow rate sequence and the continuous function G(t) of the total water quality sequence; When collecting flow rate and characteristic pollutant concentration data, the flow rate and characteristic pollutant concentration were monitored at the actual regional outlet in the study area. The data sampling frequency was less than 4 hours to ensure that the daily variation patterns of water quality and quantity could be captured (e.g., using a 2-hour interval, covering 12 time points from 0:00 to 22:00). Among these, characteristic pollutants with low degradation coefficients (such as COD) were preferentially selected for the characteristic pollutant concentration to reduce errors caused by attenuation during transport.
[0021] Furthermore, a total flow sequence is generated based on the actual sampled flow data, and a total water quality sequence is generated based on the actual sampled characteristic pollutant concentration data.
[0022] Because the flow rate and water quality data fluctuate greatly and exhibit periodic characteristics, this step uses Fourier series to fit the discrete interval sampling data into a continuous function so that values at any time can be generated subsequently and a phase shift can be introduced.
[0023] For a function f(x) with period T, the Fourier series expansion is: ,in, The Fourier fitting function is... The DC component, , Here, T represents the Fourier coefficient of the nth harmonic, and T is the function period (e.g., for daily variation, T = 24h). These are harmonic orders.
[0024] The DC component and each harmonic component (i.e.) The fitted function value is obtained by superimposing the results.
[0025] In this step, the Fourier fitting function is used. The total flow rate fitting function F is obtained by fitting the total flow rate sequence. The total water quality fitting function G is obtained by fitting the total water quality sequence. This serves as the benchmark for generating subsequent location data.
[0026] In the case provided by this invention, at the emission outlet (e.g. Figure 2 At point 21, flow rate and COD concentration were monitored synchronously every 2 hours, covering 12 time points from 0:00 to 22:00, to obtain the total flow rate sequence and total water quality sequence.
[0027] Furthermore, Fourier series was used to fit the total flow series, and the comparison results between the fitted curve and the original data are as follows: Figure 3 As shown, the fitted curve accurately captures the fluctuation characteristics of the original monitoring data: the peak water usage is concentrated between 6:00 and 22:00, and the flow rate drops rapidly to less than 50% of the peak value after 22:00. The highest point of the day occurs around 16:00.
[0028] Step S120: Generate the flow rate change function for each wastewater inflow node. ; In actual drainage networks, the flow rate at the regional total outlet is the sum of the flow rates at the sewage inflow nodes. Influenced by the unified work schedule within the region, the water usage patterns at each point (i.e., the locations corresponding to the sewage inflow nodes) are generally similar, but their contributions to the total flow rate differ. This invention generates a flow rate variation function for each point based on the principles of flow rate contribution rate allocation and phase offset, ensuring that the sum of the flow rates at each point equals the total flow rate. The specific implementation includes the following steps: 1) Calculate the flow contribution rate at each location; Given N points (i.e., N pollution-generating units), what is the flow contribution rate of each point i to the total flow? It follows a normal distribution, represented as: ,in, σ represents the flow contribution rate of the i-th location; μ is the mean, taken as μ=1 / N, where N=number of locations;2 The variance reflects the degree of difference in contribution among locations (e.g., σ). 2 =0.05); To ensure that each location has a reasonable water output and that the proportions are not excessively disparate, a constraint is imposed on the flow contribution rate, expressed as: ,in, This is the lower limit of the value to avoid invalid points with zero or minimal contribution; for example, a value of 0.02 is used. This is the upper limit of the range, used to prevent an excessively high proportion of a single location from causing an imbalance in traffic distribution; for example, a value of 0.2. This range can be adjusted and varied depending on the total number of locations N.
[0029] Traffic contribution rate at each location After constraint truncation and normalization, This ensures the conservation of the total flow rate.
[0030] 2) Calculate the phase offset; Although the flow rate changes at each location are synchronized with the overall flow rate, the peak flow rates at each location exhibit random shifts over time. To reflect this temporal randomness, a phase parameter is introduced into the Fourier model. Its values follow a uniform distribution, expressed as: ,in, To ensure uniform distribution, Let be the phase offset of the i-th point relative to the total flow. This represents the maximum phase offset duration, with a value ranging from 0 < Δt ≤ 3h, determined based on the area size and the duration of the pipeline confluence. For example, a value of Δt = 2h means that the phase difference between each point and the total flow rate fluctuates within ±2h.
[0031] 3) Combining traffic contribution rate With phase shift The flow rate change function at point i is generated by the total flow rate fitting function F(t). , represented as: ,in, Let i be the flow rate at the i-th point at time t; because The sum of the flow rates at each point satisfies That is, to achieve total conservation, where, This refers to the total outflow rate of the region; simultaneously, it is determined by... The numerical range at each point is determined by... Determine the time phase while maintaining random differences.
[0032] In this embodiment, N=20, the mean flow contribution rate μ=1 / 20=0.05, and the variance σ 2=0.05, constraint range [0.005, 0.5]; phase offset Δt = 2h, flow distribution at each point and comparison with total flow are as follows Figure 4 As shown: the flow rate at each point maintains good randomness in both magnitude and temporal distribution. The point with the largest flow rate accounts for approximately 0.1 times the total flow rate, while the point with the smallest flow rate is consistently less than 0.02 times the maximum total flow rate. The flow rate trends at each point are highly consistent with the total flow rate, with only a difference in temporal phase.
[0033] Step S130: Generate the water quality change function for each wastewater inflow node. ; In practical applications, the total effluent quality is typically assumed to be influenced by the water quality and flow rate weights at each point. Therefore, the water quality at each point... Satisfying the flow-weighted constraint is represented as: ,in, Let i be the flow weight at time t. This is the fitting function for the total effluent water quality.
[0034] In practice, water quality data at each point is generated by superimposing basic fluctuations, noise fluctuations, and phase fluctuations. Constraint control and verification are used to ensure that the weighted average error is less than a set threshold. The calculation process includes the following steps: 1) Calculate the flow weight and time average of each flow point based on the flow rate at each point, expressed as: , ,in, For the i-th point at time... Traffic weight, where M is the total number of time nodes; Let be the average weight of the i-th point across all M time points.
[0035] 2) Calculate the basic volatility coefficient For the basic volatility coefficient The modified base volatility coefficient is generated by performing a correction. The basic fluctuation coefficient can reflect the inherent differences in water quality background between different monitoring points; First, the basic volatility coefficient is generated through a uniform distribution, expressed as: ,in, Let be the initial basic volatility coefficient for the i-th point; The fluctuation amplitude parameter and It is used to control the dispersion of water quality between monitoring points, and multiple points can be selected. Numerical experiments were conducted on the values.
[0036] Then, the average weight of each point is used. The basic volatility coefficient is corrected and expressed as follows: ,in, For the corrected basic volatility coefficient, This represents the weighted average concentration across all locations.
[0037] Corrected basic volatility coefficient The weighted average is zero, thus ensuring that the weighted average concentration remains unchanged after the basic fluctuations are superimposed, so that the basic fluctuations do not affect the total amount constraint.
[0038] 3) Generate noise fluctuations to simulate random fluctuations in water quality monitoring; First, initial noise for each location and time node is generated based on a normal distribution, expressed as: , ,in, Initial noise, The standard deviation of noise. This is the noise proportionality factor, and (like );in, It is proportional to the water quality concentration at that time point, ensuring that the noise intensity adapts to the baseline value.
[0039] Furthermore, the weighted error of the initial noise is calculated and evenly distributed to each point to eliminate the influence of the initial noise on the weighted average and determine the noise correction; specifically expressed as: , ,in, For weighted error, To correct for noise.
[0040] Based on this, the weighted average error caused by the correction noise is verified. If the set threshold (e.g., 10%) is met, then the noise is corrected; otherwise, the noise is regenerated until the constraint is met.
[0041] Among them, the weighted average error Based on total water quality concentration The weighted sum of the deviations from each point is calculated as follows: .
[0042] 4) Calculate the basic fluctuation coefficient of each point at each time node. With corrected noise Superimposed, a discrete sequence of water quality data at each location is obtained; the discrete sequence of water quality data at each location is then fitted using a Fourier function. Perform Fourier fitting and output the water quality change function at each point, expressed as: .
[0043] In this step, by superimposing basic fluctuations, noise fluctuations, and phase fluctuations, the water quality at each point follows the rules of the overall regional water quality fluctuations, while achieving randomness in concentration and time, and the weighted average concentration matches the overall regional water quality (with a set error of <10%). Furthermore, by adjusting the amplitude of the basic fluctuations... It can flexibly control the difference in the direct average value of water quality at each point while ensuring that the weighted average value remains consistent, thereby regulating the degree of data dispersion.
[0044] In this embodiment, COD curves for each location are generated, with a weighted average error threshold of 10%. The generated Fourier fit function curves for the water quality at 20 locations are shown below. Figure 5 As shown in Table 1, the water quality data retains a strong degree of randomness. The weighted average water quality calculated using the flow rate of each of the 20 points was compared with the original total water quality data. The results show a high degree of agreement between the weighted average water quality and the original water quality data.
[0045] Table 1. Comparison of weighted average values of each sampling point with total effluent quality (unit: mg / L)
[0046] For different fluctuation amplitudes The arithmetic mean and weighted mean of the generated data (with values of 0, 0.2, 0.4, and 0.8) are compared, and the results are shown in Table 2. It can be seen that as... As the concentration increases, the time-averaged concentration difference between the two average values monotonically rises from 0 to 26.44 mg / L, while the weighted average remains consistent. This demonstrates that this method can achieve consistency in weighted averages by adjusting... Flexible control over the degree of data dispersion.
[0047] Table 2 Differences Comparison of direct average and weighted average of data generated under the given values
[0048] Step S140: Based on the flow rate change function and the water quality change function, generate the flow rate time series and water quality time series of each point in combination with the pipeline network model.
[0049] The specific implementation can be achieved through a program (such as a Python script), and includes the following steps: 1) Determine the corresponding location of each pollution-generating unit in the study area in the pipeline network model, determine the start time, end time and total duration of the simulation, and accordingly determine the simulation duration and sampling step size (sampling step size is less than or equal to the data sampling frequency) for the flow time series and the point water quality time series; determine the sampling data in the study area, including the total flow, DC component and harmonic components at each time point; the sampling data can be used as the basis for subsequent calculation of the dynamic data points of each node.
[0050] 2) Based on the simulation duration and sampling step size, combined with the flow rate change function and water quality change function The flow and water quality values are calculated node by node at each time point, incorporating random fluctuations to form a flow time series object and a water quality time series object containing sequence name, complete data sequence and start and end time information; 3) Read the target INP file of the pipeline network model, locate or create the time series module area, and write the flow time series objects and water quality time series objects of each node into the corresponding positions of the file according to the SWMM specification format; 4) Generate compliant external inflow configuration statements based on the target INP file, including: locating or creating the [INFLOWS] module in the INP file, configuring the time series names associated with flow and water quality for each node, and generating compliant external inflow configuration statements; 5) Run the SWMM model to complete the dynamic hydraulic and water quality simulation of the pipeline network and generate the flow time series and water quality time series of each point.
[0051] In the embodiments provided by the present invention, the flow time series objects and water quality time series objects generated by 20 nodes are written into the INP file of SWMM ([TIMESERIES] and [INFLOWS] modules) using a Python script, and the model is run to obtain the hydraulic and water quality simulation results of the pipeline network driven by the dynamic inflow of each node.
[0052] In summary, the drainage network point flow and water quality simulation data generation method provided by this invention can generate dynamic flow and water quality fluctuation curves with spatiotemporal differences for each point (node) in the network, even with only monitoring data of water quality and quantity at the total regional outlet. Furthermore, the generated point flow data satisfies the total quantity conservation constraint that "the sum of the flow at each point equals the total regional flow," while preserving the random differences in numerical amplitude and temporal phase at each point. The generated point water quality data satisfies the constraint that "the weighted average concentration of each point, weighted by flow rate, equals the total regional effluent concentration" (with the weighted average error controlled within a set threshold), while preserving the spatial differences and temporal randomness of water quality. Finally, the generated point time series data can be efficiently and systematically written into the network model (such as the INP file of SWMM) to drive the model to complete dynamic simulation.
[0053] This invention achieves a point-based flow generation mechanism that combines proportional allocation with time phase adjustment: the flow contribution rate of each point is determined by a normal distribution, while the phase parameter follows a uniform distribution to reflect the random shift in peak time at each point. This coupling ensures that the flow curve at each point reflects both the overall trend and unique temporal fluctuation characteristics, while flow contribution rate normalization guarantees the conservation of total flow. Furthermore, while characterizing the dynamic fluctuations in water quality at each point, a weighted average constraint based on flow rate and error equalization correction are introduced to ensure that the weighted average concentration at each point is consistent with the overall water quality of the region (error <10%). This reflects both the spatial differences in pollutant concentration and satisfies the conservation of total flow, solving the problem of randomly generated data not meeting quality balance. When generating water quality data, the basic fluctuation amplitude parameter is adjusted... While maintaining a consistent weighted average, the method can flexibly control the differences (dispersion) of the direct (arithmetic) average of water quality at each point, thereby adapting to the spatial differentiation characteristics of water quality in different regions and enhancing the applicability and controllability of the method.
[0054] The above-disclosed embodiments are merely a few specific examples of the present invention. However, the present invention is not limited thereto, and any variations that can be conceived by those skilled in the art should fall within the protection scope of the present invention.
Claims
1. A method for generating sewer network point flow water quality simulation data, characterized in that, Includes the following steps: A pipe network model is established based on the actual drainage pipe network topology of the study area; pollution-generating units and total emission monitoring points are determined in the pipe network model; the pollution-generating units are the monitoring points. Flow rate and characteristic pollutant concentrations were collected at the study area corresponding to the total flow monitoring points to obtain total flow rate and total water quality sequences; these sequences were then fitted using Fourier functions. The total flow rate fitting function is obtained by fitting the total flow rate sequence. The total water quality fitting function is obtained by fitting the total water quality sequence. Wherein, the Fourier fitting function Represented as: ,in, The Fourier fitting function is... The DC component, , For the first Fourier coefficients of subharmonics For the period of the function, Harmonic order; Generate the flow rate change function for each wastewater inflow node. This includes: calculating the flow contribution rate at each location. The traffic contribution rate satisfy N = number of points; calculate the phase offset for each point. The phase shift This indicates that the peak flow rate at each location has a random shift over time; combined with the aforementioned flow contribution rate. With phase shift Fitting function based on total flow Point generation Flow change function , represented as: ,in, For the first Each point at time The traffic flow; where the traffic contribution rate is calculated. Includes: generating each point based on a normal distribution. Contribution rate to total traffic , represented as: ,in, For the first Traffic contribution rate at each location The mean and , σ represents the number of points. 2 Variance; contribution rate to flow rate Applying constraints is represented as: ,in, The lower limit of the value, This represents the upper limit of the possible values; it also represents the traffic contribution rate. After constraint truncation and normalization, ; Generate water quality change function for each wastewater inflow node This includes: calculating the flow weight of each point and the time average of the flow weight based on the flow at each point, expressed as: , ,in, For the first Each point at time Traffic weight, The total number of time points. For the first Each point in the whole Average weights at each time point; generating the basic volatility coefficient. For the aforementioned basic volatility coefficient The modified base volatility coefficient is generated by performing a correction. ; Generate noise fluctuations and determine corrected noise The adjusted base volatility coefficient for each point at each time point. With corrected noise Superimposed, a discrete sequence of water quality data at each location is obtained; the discrete sequence of water quality data at each location is then fitted using a Fourier function. Perform Fourier fitting and output the water quality change function at each point, expressed as: ; Based on flow change function and water quality change function The flow time series and water quality time series of each point are generated by combining the pipeline network model.
2. The method for generating simulated flow and water quality data at drainage network points according to claim 1, characterized in that, The phase offset of each point The values follow a uniform distribution, represented as: ,in, To ensure uniform distribution, For the first Phase offset of each point relative to the total flow. This represents the maximum phase shift duration.
3. The method for generating simulated flow and water quality data at drainage network points according to claim 1, characterized in that, The basic volatility coefficient Generated through a uniform distribution, represented as: ,in, For the first The initial basic volatility coefficient at each point, The fluctuation amplitude parameter and ; Using the average weight of each point For the basic volatility coefficient The modified base volatility coefficient is generated by performing a correction. , represented as: ,in, For the corrected basic volatility coefficient, For the first Average weight of each point This is the weighted average concentration across all locations.
4. The method for generating simulated flow and water quality data at drainage network points according to claim 1, characterized in that, The generated noise fluctuations include: The initial noise for each point and time node is generated based on a normal distribution, and is expressed as follows: , ,in, Initial noise, The standard deviation of noise. This is the noise proportionality factor, and ; The weighted error of the initial noise is calculated and evenly distributed to each point, as shown below: ,in, This is the weighted error; The corrected noise is calculated and expressed as: ,in, For weighted error, To correct noise; Verify the weighted average error caused by the correction noise. If the set threshold is met, then the noise is corrected; otherwise, the noise is regenerated until the constraint is met.
5. The method for generating simulated flow and water quality data at drainage network points according to claim 4, characterized in that, The weighted average error The calculation method is as follows: .
6. The method for generating simulated flow and water quality data at drainage network points according to claim 1, characterized in that, The generation of flow time series and water quality time series at each location by combining the pipeline network model is implemented through a program, including the following steps: Determine the corresponding location of each pollution-generating unit in the pipeline network model within the study area, determine the start time, end time, and total duration of the simulation, and determine the sampling data within the study area; Based on the flow change function and water quality change function The flow rate and water quality at each time point are calculated node by node, forming a flow rate time series object and a water quality time series object; Read the target INP file of the pipeline network model, locate or create the time series module area, and write the flow time series objects and water quality time series objects of each node into the target INP file according to the SWMM specification format. Generate compliant external inflow configuration statements based on the target INP file; Run the SWMM model to complete the dynamic hydraulic and water quality simulation of the pipeline network, and generate the flow time series and water quality time series of each point.
Citation Information
Patent Citations
Drainage pipe network pollution path monitoring and identifying method, device and electronic equipment
CN110196083A
Simulation prediction method and system for dynamic process of sewage treatment based on digital twinning
CN120197403A
