A sewage plant water collection pipe network infiltration water quality and quantity dynamic estimation method and system

Abstract

The application discloses a sewage plant water collection pipe network infiltration water quality and quantity dynamic estimation method and system, including the following steps, standardization obtains the daily scale flow sequence and daily scale concentration monitoring sequence of the sewage plant in the research area within the specified observation period N days, and the two are identified, removed and missing value filled in abnormal high value; the background water quality concentration of each water quality component is constructed; the trend item after the daily scale flow sequence is decomposed is used as a trend flow item; the initial estimated groundwater infiltration flow, additional wastewater flow and infiltration concentration estimation value of groundwater infiltration under each candidate dry day base flow quantile are calculated; a target function is constructed, the target quantile and pipe network groundwater infiltration water quality and quantity results are determined based on the optimization result; and the pipe network groundwater infiltration water quality and quantity results of the sewage plant are output. The application can dynamically identify the composition of dry day inflow flow and accurately quantify the groundwater infiltration quantity and water quality, and the dynamic identification cost is low.

Description

Technical Field

[0001] This invention belongs to the field of urban drainage network modeling, simulation and analysis technology, and in particular relates to a method and system for dynamic estimation of infiltration water quality and quantity in sewage treatment plant water intake networks. Background Technology

[0002] Groundwater infiltration is an objective phenomenon in the operation of urban drainage systems. This unintended groundwater enters the sewage network through manholes, pipe defects, and improper connections, eventually flowing into sewage treatment plants and significantly altering the hydraulic load within the network. While drainage networks are designed to transport and treat domestic, commercial, and industrial wastewater (i.e., sanitary flow, constituting the baseline flow in dry weather), allowing for a certain percentage of groundwater infiltration, in actual operation, groundwater continuously infiltrates the network, causing the total amount of wastewater collected by sewage treatment plants to deviate from design expectations. Furthermore, the scale of infiltration often far exceeds initial design assumptions due to factors such as pipe aging and geological conditions.

[0003] Due to groundwater infiltration, urban drainage and sewage treatment systems face a series of challenges, including facility safety, increased maintenance costs, and impacts on water quality. More critically, the construction of hydraulic and water quality simulation models for drainage networks requires accurate inflow boundary conditions. Dry-day baseline flow, groundwater infiltration, and groundwater concentration data are the core input parameters for these models. However, due to a lack of monitored groundwater infiltration data and accurate decomposition data, the flow input for these models often relies on empirical assumptions, leading to significant discrepancies between simulation results and actual operating conditions. This hinders their ability to effectively support applications such as network optimization and scheduling, severely limiting their application value. Furthermore, the lack of unified, low-cost, and scalable methods for decomposing dry-day baseline flow and quantifying groundwater infiltration makes cross-regional data comparisons difficult, hindering the development of systematic network management solutions. In practical applications, this directly impacts the scientific validity of network repair priority determination and investment decisions.

[0004] In summary, groundwater infiltration volume and quality are among the core boundary conditions of drainage network models. Their accuracy directly affects the model's simulation precision of hydraulic conditions and pollutant loads in dry weather. Inaccurate data will lead to biases in decisions regarding network repair and wastewater treatment plant scheduling. Existing technologies such as the minimum nighttime flow method, the plugging measurement method, and the characteristic factor method, when used to estimate groundwater infiltration, suffer from problems such as overly simplified assumptions, high implementation costs, and limited applicability to space and scenarios. These methods cannot meet the needs of complex urban drainage systems for accurate decomposition of dry-weather inflow components and dynamic identification of groundwater infiltration volume. Summary of the Invention

[0005] Purpose of the invention: The first purpose of this invention is to provide a method for dynamically estimating the infiltration water quality and quantity in sewage treatment plant receiving networks that can meet the needs of complex urban drainage systems for accurate decomposition of dry-day inflow components, dynamic identification of groundwater infiltration volume and its water quality.

[0006] The second objective of this invention is to provide a dynamic estimation system for the quality and quantity of infiltrated water in the sewage treatment plant's water intake network.

[0007] Technical solution: This invention discloses a method for dynamically estimating the quality and quantity of infiltrated water in a wastewater treatment plant's intake network, comprising the following steps:

[0008] S1: Obtain high-frequency influent flow data for wastewater treatment plants within the study area over a specified observation period of N days. Monitoring data of the concentration of all water components entering the plant. Where i represents the type of water quality component; for high-frequency influent flow data and influent water quality concentration monitoring data After standardization, daily flow sequences were obtained. and daily-scale concentration monitoring sequences ;

[0009] S2: Daily-scale flow sequence and daily-scale concentration monitoring sequences Perform identification, removal, and missing value imputation for abnormally high values;

[0010] S3: Obtain rainfall meteorological data within the study area and historical water quality concentration monitoring data of all water quality components from the outlet pipes of all residential areas within the study area. Based on rainfall meteorological data, periods without rainfall were identified, and then historical water quality concentration monitoring data of all water quality components during these periods were selected from the outlet pipes. Based on historical water quality concentration monitoring data from the outlet pipe and influent water quality concentration monitoring data Constructing the dry-day background water quality concentration for each water quality component ;

[0011] S4: The daily flow series after identifying, removing, and imputing outliers using the Time Series Smoothing (STL) algorithm is processed. Decompose the daily-scale flow series The decomposed trend term is used as the trend flow term. ;

[0012] S5: Set multiple candidate dry-day baseflow quantiles 'a' and combine them with the trend flow term. Generate corresponding candidate drought baseline flow Combined with trend traffic items Baseline flow rates for each candidate drought day Daily-scale concentration monitoring sequences of each water quality component i Background water quality concentration during drought The groundwater infiltration flow rate under each candidate dry-day baseflow quantile 'a' was calculated. Additional wastewater flow Estimated infiltration concentration of groundwater ;

[0013] S6: Construct the objective function and use the objective function to calculate the groundwater infiltration flow for each candidate dry-day baseflow quantile a. Additional wastewater flow and estimated infiltration concentration Perform joint optimization and determine the target quantile based on the optimization results. , and the target quantile The corresponding data set represents the quality and quantity of groundwater infiltration into the sewage treatment plant's pipe network.

[0014] S7: Output the quality and quantity of groundwater infiltration in the sewage treatment plant's pipeline network.

[0015] Furthermore, the high-frequency inlet flow data described in step S1 and influent water quality concentration monitoring data The standardization process is based on high-frequency inbound flow data. Calculate the average daily inflow rate over the observation period N days. Average daily inflow Daily-scale flow series of high-frequency inflow data obtained by sorting by time ,in This represents the average daily total inflow to the plant on day N;

[0016] Influent water concentration monitoring data The daily average concentration of influent water was obtained by taking the daily average value. Daily average concentration of water entering the plant The daily-scale concentration monitoring sequence was obtained by sorting by time. ,in This represents the average concentration of water component i on day N.

[0017] Furthermore, in step S2, the daily-scale flow sequence... and daily-scale concentration monitoring sequences The method for identifying and removing outlier values ​​is as follows: The local outlier factor algorithm is used to analyze the daily-scale flow series. Outlier analysis was performed, and daily flow sequences were selected based on Euclidean distance. The set of points in the k-neighborhood , And calculate the reachable distance within the k-neighborhood. Locally achievable density and the corresponding local outlier factors ;

[0018] For data points of daily scale flow series achievable distance The calculation method is as follows:

[0019]

[0020] in , Data points representing daily-scale flow series To its first The distance between neighbors Data points representing daily-scale flow series and The Euclidean distance between them;

[0021] Where the locally reachable density is The calculation method is as follows:

[0022]

[0023] Data points for each daily-scale flow sequence Local outlier The calculation method is as follows:

[0024] ;

[0025] when And the corresponding daily-scale flow sequence data points Exceeding daily scale flow sequence When the preset quantile is used, the data points of the daily scale flow series It was identified as a locally abnormal high value point;

[0026] For date m where an abnormally high value is identified, remove the daily-scale flow series data points within that date. Simultaneously remove daily-scale concentration monitoring sequences. Daily-scale concentration monitoring for that day ;

[0027] Step S2 involves analyzing the daily-scale flow sequence. and daily-scale concentration monitoring sequences The method for handling missing values ​​is as follows: High-frequency inbound flow data... and influent water quality concentration monitoring data The original missing data points and daily scale flow series and daily-scale concentration monitoring sequences After identifying and removing abnormally high values, the resulting vacancies are marked as NaN, and linear interpolation is used to fill in the NaNs.

[0028] Furthermore, in step S3, the arid-day background water quality concentration of each water quality component is constructed. The steps are as follows:

[0029] S31: Calculate historical water quality concentration monitoring data for outdoor pipes during periods without rainfall. The mean value is used as the background concentration of water quality component i in the outlet pipe during dry weather. ;

[0030] S32: Data from influent water quality concentration monitoring Select the maximum daily concentration value Constitutes the daily maximum concentration sequence And remove the daily maximum concentration sequence. Outliers in the sequence will be removed from the daily maximum concentration series. As a measure of background concentration during drought ;

[0031] S33: Background concentration during dry weather will be measured at the outdoor pipe. and detection of background concentration during drought days The larger values ​​are considered as water quality components. Background water quality concentration during drought ;

[0032] S34: Repeat steps S31-S33 to calculate the dry-day background water quality concentration corresponding to each water quality component i. .

[0033] Furthermore, the groundwater infiltration flow rate under each candidate dry-weather baseflow quantile a, as described in step S5, is calculated. Additional wastewater flow Estimated infiltration concentration of groundwater The steps are as follows:

[0034] S51: Set multiple candidate dry-weather base current quantiles a, Based on candidate dry-day baseflow quantile 'a' and trend flow term Set the candidate dry-day baseline flow corresponding to each candidate dry-day baseflow quantile 'a'. , ;

[0035] S52: For each candidate dry-day baseline flow Based on the trend flow term during the observation period N and candidate drought baseline flow Calculate the initial estimated daily groundwater infiltration rate. ,and ;

[0036] S53: Divide the observation period N days into multiple non-overlapping time periods of equal length in chronological order. Where K∈[1,N / L], and calculate in Time period Initial daily estimated groundwater infiltration Daily-scale concentration monitoring sequences of each water quality component i Correlation coefficient between Based on the initial estimate of groundwater infiltration volume and daily-scale concentration monitoring sequences During the corresponding time period Internal trends and correlation coefficients , each time period Initial estimated groundwater infiltration rate Reclassification as groundwater infiltration flow Or additional wastewater flow ;

[0037] S54: Based on candidate dry weather baseline flow Groundwater infiltration flow Background water quality concentration during drought Calculate the water quality components under different candidate dry-day baseflow quantiles a. Estimates of infiltration concentration .

[0038] Furthermore, in step S53, each time period... Initial estimated groundwater infiltration rate Reclassification as groundwater infiltration flow Or additional wastewater flow The method is as follows:

[0039] Calculate time period Initial estimate of groundwater infiltration time slope and daily-scale concentration monitoring sequences of ;

[0040] like , And the correlation coefficient Greater than the preset threshold Then this time period Initial estimated groundwater infiltration volume Reclassified as additional wastewater flow ;

[0041] like , or correlation coefficient Less than or equal to the preset threshold Then this time period Initial estimated groundwater infiltration volume Reclassification as groundwater infiltration flow ;

[0042] like During this time period Initial estimated groundwater infiltration volume Reclassification as groundwater infiltration flow .

[0043] Furthermore, in step S54, the water quality components under different candidate dry-day base current quantiles a are calculated. Estimates of infiltration concentration The method is as follows:

[0044] Based on candidate drought baseline flow and groundwater infiltration flow Calculate the initial estimate of the daily infiltration concentration. The calculation formula is:

[0045] ;

[0046] The initial estimates of infiltration concentration for all daily times within the observation period N were calculated. According to groundwater infiltration flow rate The weighted time average yielded the candidate dry-day base current quantile 'a' and water quality components. Estimates of infiltration concentration The calculation formula is as follows:

[0047] .

[0048] Furthermore, step S6, obtaining the results of the quality and quantity of groundwater infiltration in the sewage treatment plant's pipeline network, includes the following steps:

[0049] S61: Set daily-scale flow mixing weights Used to dynamically regulate groundwater infiltration flow. and additional wastewater flow The distribution ratio between them;

[0050] S62: Based on daily-scale flow mixed weighting Reclassified groundwater infiltration flow and additional wastewater flow The corrected groundwater infiltration flow rate is obtained by making adjustments. and correction of additional wastewater flow ;

[0051] S63: For each water quality component Estimates of infiltration concentration After perturbation and expansion, the estimated infiltration concentration is: Based on the modified groundwater infiltration flow rate Correcting additional wastewater flow rate The estimated infiltration concentration after perturbation expansion is: Predicted concentrations of each water quality component i ;

[0052] S64: Minimizing diurnal concentration monitoring sequences Compared with predicted concentration Sum of squared residuals between Jointly optimize daily-scale flow mixed weights Estimated infiltration concentration after perturbation expansion The simulated annealing algorithm is used to solve the problem. and Corresponding daily-scale flow mixed weight Estimated infiltration concentration after perturbation expansion ;

[0053] S65: Compare the joint optimized values ​​for each candidate dryland baseflow quantile a. Take multiple joint optimization results The candidate dry-weather base current quantile 'a' corresponding to the smallest median value is used as the target quantile. ; and target quantile The corresponding data sets include , , and .

[0054] Furthermore, in step S62, the reclassified groundwater infiltration flow rate is... and additional wastewater flow The correction method is as follows:

[0055]

[0056] ;

[0057] Predicting concentration in step S63 The calculation formula is as follows:

[0058] ;

[0059] The expression for the objective function in step S64 is as follows:

[0060]

[0061] in Indicates the quantity of water quality components. ; The time index is the number of daily-scale samples corresponding to the observation duration. .

[0062] Based on the same inventive concept, this invention also discloses a dynamic estimation system for the quality and quantity of infiltrated water in a wastewater treatment plant's intake network, comprising:

[0063] The data standardization module is used to obtain high-frequency influent flow data of wastewater treatment plants within the study area over a specified observation period of N days. Monitoring data of the concentration of all water components entering the plant. Where i represents the type of water quality component; and high-frequency influent flow data and influent water quality concentration monitoring data After standardization, daily flow sequences were obtained. and daily-scale concentration monitoring sequences ;

[0064] The data preprocessing module is used for daily-scale flow series. and daily-scale concentration monitoring sequences Perform identification, removal, and missing value imputation for abnormally high values;

[0065] The drought background water quality concentration module acquires rainfall meteorological data for the study area and historical water quality concentration monitoring data of all water components in the residential areas within the study area from the outlet pipes. Based on rainfall meteorological data, periods without rainfall were identified, and then historical water quality concentration monitoring data of all water quality components during these periods were selected from the outlet pipes. Based on historical water quality concentration monitoring data from the outlet pipe and influent water quality concentration monitoring data Constructing the dry-day background water quality concentration for each water quality component ;

[0066] The trend flow item module uses the Time Series Smoothing (STL) algorithm to identify, remove, and impute missing values ​​in the daily-scale flow series. Decompose the daily-scale flow series The decomposed trend term is used as the trend flow term. ;

[0067] Infiltration initial estimation module; sets multiple candidate dry-day baseflow quantiles 'a' and combines them with the trend flow term. Generate corresponding candidate drought baseline flow Combined with trend traffic items Baseline flow rates for each candidate drought day Daily-scale concentration monitoring sequences of each water quality component i Background water quality concentration during drought The groundwater infiltration flow rate under each candidate dry-day baseflow quantile 'a' was calculated. Additional wastewater flow Estimated infiltration concentration of groundwater ;

[0068] Objective function optimization module: Constructs the objective function and uses it to calculate the groundwater infiltration flow rate for each candidate dry-weather baseflow quantile a. Additional wastewater flow and estimated infiltration concentration Perform joint optimization and determine the target quantile based on the optimization results. , and the target quantile The corresponding data set represents the quality and quantity of groundwater infiltration into the sewage treatment plant's pipe network.

[0069] The results output module outputs the results of the groundwater infiltration quality and quantity in the sewage treatment plant's pipeline network.

[0070] Beneficial Effects: Compared with existing technologies, this invention has the following significant advantages: Through multi-level data preprocessing, trend decomposition, background concentration determination, and correlation analysis, this invention can dynamically identify the composition of dry-day inflow and accurately quantify groundwater infiltration, with low dynamic identification costs. Simultaneously, by combining multi-source data such as flow rate and water quality with optimization algorithms, this invention can further improve the accuracy of identifying the composition of dry-day inflow and the accuracy of groundwater infiltration and its quality, which is beneficial for providing accurate inflow boundary conditions for sewage pipe network models. By capturing the time-varying trend of groundwater infiltration, this invention achieves accurate assessment of the pipe network status, which is beneficial for providing key technical support for the optimization of urban sewage system operation and the refined management of pipe networks. Attached Figure Description

[0071] Figure 1 This is a flowchart of the present invention;

[0072] Figure 2 This is a schematic diagram of the system structure of the present invention;

[0073] Figure 3 Predicted concentration for embodiments of the present invention Daily-scale concentration monitoring sequence Comparison chart;

[0074] Figure 4 As an embodiment of the present invention, baseline flow rate during candidate dry weather Daily scale flow series Correcting groundwater infiltration flow and correction of additional wastewater flow The overlay image. Detailed Implementation

[0075] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0076] Example 1

[0077] This invention discloses a method for dynamically estimating the quality and quantity of infiltrated water in a wastewater treatment plant's intake network, such as... Figure 1 As shown, it includes the following steps:

[0078] S1: Obtain high-frequency influent flow data for wastewater treatment plants within the study area over a specified observation period of N days. Monitoring data of the concentration of all water components entering the plant. Where i represents the type of water quality component; for high-frequency influent flow data and influent water quality concentration monitoring data Standardization processing was performed to obtain the daily-scale flow sequence of high-frequency inflow data. Daily-scale concentration monitoring sequence of incoming water quality .

[0079] In practical applications, the types of water components to be acquired are selected based on actual needs. Among these, high-frequency influent flow data... The term "high frequency" refers to the continuous collection of inflow data at intervals of minutes or hours, such as collecting inflow data every 5 minutes or every hour.

[0080] High-frequency inlet flow data and influent water quality concentration monitoring data The standardization process is as follows:

[0081] High-frequency inflow flow data and influent water quality concentration monitoring data The timestamp is converted to a daily-scale index t, with the format YYYY-MM-DD. For example, 2025-05-05 represents the influent flow rate or influent water quality concentration monitoring on May 5, 2025. Preferably, high-frequency influent flow rate data is used. The unit standardization is The monitoring data of water quality concentration entering the plant The unit standardization is .

[0082] Based on high-frequency inflow data Calculate the average daily inflow rate over the observation period N days. ,and ,in For the first Number of monitoring data entries within the day Indicates the first Day 1 The original high-frequency flow monitoring value of the strip, To determine the frequency of daily monitoring of inflow. Average daily inflow. Daily-scale flow series of high-frequency inflow data obtained by sorting by time ,in This represents the average daily total inflow to the plant on day N.

[0083] Influent water concentration monitoring data The daily average concentration of influent water was obtained by taking the daily average value. Daily average concentration of influent water monitored The daily-scale concentration monitoring sequence was obtained by sorting by time. ,in This represents the average concentration of water component i on day N.

[0084] S2: Daily-scale flow sequence and daily-scale concentration monitoring sequences Perform identification, removal, and missing value completion for abnormally high values.

[0085] Daily scale flow series and daily-scale concentration monitoring sequences The methods for identifying and removing abnormally high values ​​are as follows:

[0086] The Local Outlier Factor (LOF) algorithm was used to analyze daily flow sequences. Outlier analysis was performed. Daily flow sequences were selected based on Euclidean distance. The set of points in the k-neighborhood , , And calculate the reachable distance within the k-neighborhood. Locally achievable density and the corresponding local outlier factors .

[0087] For data points of daily scale flow series achievable distance The calculation method is as follows:

[0088]

[0089] in , ; Data points representing daily-scale flow series To its first The distance between neighbors Data points representing daily-scale flow series and The Euclidean distance between them.

[0090] Locally achievable density is The calculation method is as follows:

[0091] ;

[0092] Data points for each daily-scale flow sequence Local outlier The calculation method is as follows:

[0093] ;

[0094] when And the corresponding daily-scale flow sequence data points Exceeding daily scale flow sequence When the preset quantile is used, it represents the data points of the daily-scale flow series. The local density is significantly lower than that of its neighboring points, and it is a local anomaly. For the date m where an anomaly is identified, the daily-scale flow series data points within that date are removed. Simultaneously, daily-scale concentration monitoring sequences were removed. Daily-scale concentration monitoring for that day This ensures consistency between the incoming flow rate and water quality concentration monitoring.

[0095] Preferably, the preset quantile is 75%; in practical applications, users can adjust the specific value of the preset quantile according to actual needs. In this embodiment, the preset quantile is set to 75% only to exclude daily-scale flow sequences. The abnormally high values ​​are not excluded, but other types of outliers are not removed. The exclusion of abnormally high values ​​is to eliminate the interference of rainfall events and abnormal single-cell events on concentration monitoring. Single-cell abnormal events are periodic, short-term external disturbances; such single-cell events include, but are not limited to, concentrated sewage discharge during peak tourist seasons, unconventional emissions from industrial enterprises, and sudden pipeline accidents.

[0096] Daily scale flow series and daily-scale concentration monitoring sequences The missing value imputation method is as follows:

[0097] High-frequency inflow flow data and influent water quality concentration monitoring data Original missing data points, daily scale flow sequence and daily-scale concentration monitoring sequences After identifying and removing abnormally high values, all resulting gaps were marked as NaN for daily-scale flow sequences. and daily-scale concentration monitoring sequences Linear interpolation is applied to all NaN points to obtain continuous data on a daily scale. The interpolation method is as follows:

[0098]

[0099] Where x(t) represents the interpolated value of the target variable at interpolation time t, and the target variable can be the daily total flow sequence after removing outliers. or daily-scale concentration monitoring sequence Any data point in the data that needs to be completed; , These are the nearest non-NaN times before and after the interpolation point, respectively. and These are the non-NaN values ​​at these two consecutive known times.

[0100] S3: Obtain rainfall meteorological data within the study area and historical water quality concentration monitoring data of all water quality components from the outlet pipes of all residential areas within the study area. Based on rainfall meteorological data, periods without rainfall were identified, and then historical water quality concentration monitoring data of all water quality components during these periods were selected from the outlet pipes. Based on historical water quality concentration monitoring data from the outlet pipe and influent water quality concentration monitoring data Constructing the dry-day background water quality concentration for each water quality component .

[0101] Preferably, the historical water quality concentration monitoring data of the residential area's door pipes during periods without rainfall are selected. No significant increase, i.e., historical water quality concentration monitoring data from the household pipe. The growth rate should be less than a pre-set threshold. Add the screening criterion "historical water quality concentration monitoring data from the outlet pipe". The reasons for "no significant increase" are as follows: If only "no rainfall" is used as the screening condition, external disturbances such as peak tourism seasons and unconventional industrial emissions may still cause abnormally high pollutant concentrations; if the aforementioned factors are not excluded, such abnormally high values ​​caused by non-rainfall will introduce bias, causing an overestimation of the background concentration in dry weather, which in turn leads to an underestimation of the inflow and infiltration components; adding this constraint can effectively ensure that the extracted monitoring data represents the typical background level in dry weather, ensuring data stability and the accuracy of calculation results.

[0102] S31: Calculate historical water quality concentration monitoring data for outdoor pipes during periods without rainfall. The mean value is used as the background concentration of water quality component i in the outlet pipe during dry weather. ,and Where L represents the number of days without rainfall. This indicates historical water quality concentration monitoring data from the household pipes. The concentration of water quality component i was monitored on day b, and .

[0103] Preferably, if historical water quality concentration monitoring data of the outflow pipes of the residential areas within the study region cannot be obtained during practical applications, literature data can be used as historical water quality concentration monitoring data for the outflow pipes. .

[0104] S32: Data from influent water quality concentration monitoring Select the maximum daily concentration value Constitutes the daily maximum concentration sequence ,and The expression is as follows:

[0105]

[0106] in, Indicates water quality components In the The highest concentration value monitored per day, where N is the total number of days in the observation period.

[0107] Elimination of daily maximum concentration sequence Outliers in the sequence will be removed from the daily maximum concentration series. As a measure of background concentration during drought .

[0108] The method for removing outliers is as follows: remove the daily maximum concentration sequence. Data from the 98th percentile or higher, i.e., retaining the daily maximum concentration sequence. The 98th percentile, and the detection of background concentrations during dry days. The expression is as follows:

[0109] .

[0110] S33: Background concentration during dry weather will be measured at the outdoor pipe. and detection of background concentration during drought days The larger values ​​are considered as water quality components. Background water quality concentration during drought ,Right now The reason for choosing the larger value between the two is as follows: historical water quality monitoring data from the residential area's outlet pipes better reflects the typical pollutant concentrations at the source discharge end, while the influent water quality concentration monitoring data from the sewage treatment plant reflects the overall background level after multiple sources are superimposed and aggregated; if the smaller value is chosen, the actual pollutant concentration under dry weather conditions may be underestimated, leading to systematic biases in subsequent model calculations and pollution load assessments; choosing the larger value can ensure the reasonableness of the estimation results while ensuring stronger robustness and conservatism in determining the background concentration under dry weather conditions, thus providing a reliable benchmark for subsequent anomaly detection and rainy weather load separation.

[0111] Preferably, in practical applications, if historical water quality concentration monitoring data of the outflow pipes of the residential areas within the study area cannot be obtained, and literature data cannot be obtained to replace historical water quality concentration monitoring data of the outflow pipes, then... Then directly detect the background concentration during drought days. As a component of water quality Background water quality concentration during drought .

[0112] S34: Repeat steps S31-S33 to calculate the dry-day background water quality concentration corresponding to each water quality component i. .

[0113] S4: The daily flow series after identifying, removing, and imputing outliers using the Time Series Smoothing (STL) algorithm is processed. Decompose the daily-scale flow series The decomposed trend term is used as the trend flow term. .

[0114] To identify the stable baseline components during groundwater inflow, STL was used to analyze diurnal flow sequences. Trend extraction is performed on daily-scale flow series using the Time Series Smoothing (STL) algorithm. The hierarchical model for decomposition is as follows:

[0115]

[0116] in, The trend term represents the stable growth or decline trend of the drainage network caused by continuous groundwater inflow and basic domestic sewage load. This indicates the seasonal component, used to capture cyclical fluctuations. This represents the residual term, reflecting random factors such as sudden drainage and equipment vibration; and the trend flow term... Equal to trend term .

[0117] The STL algorithm estimates the seasonal term iteratively. With trend items And update the residual term after each iteration. Until convergence, the final extracted trend flow term The daily-scale flow time series data, after preprocessing and smoothing to remove noise, is used to characterize the stable growth or decline trend in the drainage network caused by continuous groundwater inflow and basic domestic sewage load.

[0118] S5: Set multiple candidate dry-day baseflow quantiles 'a' and combine them with the trend flow term. Generate corresponding candidate drought baseline flow Combined with trend traffic items Baseline flow rates for each candidate drought day Daily-scale concentration monitoring sequences of each water quality component i Background water quality concentration during drought The groundwater infiltration flow rate under each candidate dry-day baseflow quantile 'a' was calculated. Additional wastewater flow Estimated infiltration concentration of groundwater .

[0119] S51: Set multiple candidate dry-weather base current quantiles a, Based on candidate dry-day baseflow quantile 'a' and trend flow term Set the candidate dry-day baseline flow corresponding to each candidate dry-day baseflow quantile 'a'. .

[0120] The set of multiple candidate dry-day base current quantiles constitutes the candidate quantile set. Where n is the total number of candidate dry-weather base current quantiles a; preferably, For 1, It is 20, that is In practical applications, the settings can be configured according to actual needs. and , and a takes a random value within the threshold range.

[0121] For each candidate dry-day base current quantile Set a corresponding candidate drought baseline flow. And candidate drought baseline flow The expression is .

[0122] S52: For each candidate dry-day baseline flow Based on the trend flow term during the observation period N and candidate drought baseline flow Calculate the initial estimated daily groundwater infiltration rate. ,and .

[0123] Used to ensure Greater than or equal to 0, to prevent occurrence The case of less than 0 makes It has the correct physical meaning.

[0124] S53: Divide the observation period N days into multiple non-overlapping time periods of equal length in chronological order. Where K∈[1,N / L], and calculate in each time period Initial daily estimated groundwater infiltration Daily-scale concentration monitoring sequences of each water quality component i Correlation coefficient between Based on the initial estimate of groundwater infiltration volume and daily-scale concentration monitoring sequences During the corresponding time period Internal trends and correlation coefficients , each time period Initial estimated groundwater infiltration volume Reclassification as groundwater infiltration flow Or additional wastewater flow .

[0125] The observation period of N days is divided into multiple non-overlapping time periods of equal length in chronological order. The length of each time period is denoted as L days, preferably L=10. All time periods form a set. ,in This represents the Kth time interval, and The covered date is Days; M represents the total number of time periods, and , and K∈[1,N / L]; M time periods cover all observation periods N days, and the dates of the time periods arranged in chronological order are consecutive and do not overlap.

[0126] For each candidate dry day baseline flow To differentiate between groundwater infiltration and additional wastewater affecting high-frequency influent flow data The contribution of the change is calculated separately for each time period. Initial daily estimated groundwater infiltration Daily-scale concentration monitoring sequences of each water quality component i Correlation coefficient between And the correlation coefficient The calculation formula is as follows:

[0127] ,

[0128] in Indicates the time period Initial estimated groundwater infiltration volume The mean, Indicates time period Daily-scale concentration monitoring sequence of water quality component i The mean.

[0129] Comprehensive time period Correlation coefficient within Initial estimate of groundwater infiltration Trends and daily-scale concentration monitoring sequences The trend during this period Initial estimated groundwater infiltration volume Reclassified as additional wastewater flow or groundwater infiltration flow .

[0130] Calculate time period Initial estimate of groundwater infiltration time slope and daily-scale concentration monitoring sequences of In practical applications, the least squares linear regression method can be used to calculate the time slope.

[0131] like , That is, the initial estimate of groundwater infiltration. and daily-scale concentration monitoring sequences In time period The internal values ​​all showed an upward trend over time, and the correlation coefficient was... Greater than the preset threshold This indicates that within the time period The increasing trend in the concentration of this water component was not caused by the infiltration of low-concentration groundwater, but rather by the actual inflow of high-concentration wastewater exceeding the initially estimated groundwater infiltration rate. Caused by, therefore this time period Initial estimated groundwater infiltration volume Reclassified as additional wastewater flow ,Right now Among them, the preset threshold For positive numbers, a preset threshold is used. The preferred value is 0.3, but in practical applications, it can be adjusted according to actual needs. The specific value.

[0132] like , That is, the initial estimate of groundwater infiltration. In time period The concentrations within the range showed an increasing trend over time, according to the daily-scale concentration monitoring sequence. In time period The internal average shows a decreasing trend over time, or the correlation coefficient... Less than or equal to the preset threshold This indicates that the flow exceeds the candidate drought baseline flow. The flow rate most likely originates from low-concentration groundwater infiltration or other low-concentration external water recharge, therefore, during this period... Initial estimated groundwater infiltration volume Reclassification as groundwater infiltration flow ,Right now .

[0133] like That is, the initial estimate of groundwater infiltration. In time period The internal trend did not increase over time, but rather showed a downward or flat trend. Initial estimated groundwater infiltration volume Reclassification as groundwater infiltration flow ,Right now .

[0134] S54: Based on candidate dry weather baseline flow Groundwater infiltration flow Background water quality concentration during drought Calculate the water quality components under different candidate dry-day baseflow quantiles a. Estimates of infiltration concentration .

[0135] Preliminary estimation of daily infiltration concentration based on mass conservation, for each candidate dry-day baseline flow. Under a given candidate dry-day base current quantile 'a', for each water quality component Construct a daily-scale mass balance relationship: the observed total pollutant mass equals the candidate dry-day baseline flow. Mass carried and groundwater infiltration flow The sum of the masses carried is used to calculate the initial estimate of the daily infiltration concentration. :

[0136]

[0137] in, Daily time Daily average concentration monitoring of incoming water quality; Daily scale flow series The decomposed trend term represents the total daily flow after preprocessing and smoothing; Background water quality concentration during dry weather; This refers to the groundwater infiltration flow rate obtained after reclassification in step S53.

[0138] Preferably, when When the infiltration concentration is equal to or close to 0, to avoid numerical instability, the initial estimate of the infiltration concentration at the corresponding time point can be used. If set to missing, use linear interpolation of adjacent valid days as a replacement.

[0139] The initial estimates of infiltration concentration for all daily times within the observation period N were calculated. Then, a time-weighted average based on flow rate was performed to finally obtain the candidate dry-day baseflow quantile 'a' and the correlation between water quality components. Estimates of infiltration concentration :

[0140]

[0141] The weighting process uses groundwater infiltration flow rate. As a weighting factor, it emphasizes the contribution of days with higher groundwater infiltration to the overall estimate; at the same time, to ensure physical rationality, it also considers... Apply boundary constraints to remove extreme out-of-boundary points. The above processing completes the process in... The initial estimation of infiltration concentration driven by the conservation of mass in the dimension provides a physically consistent starting point for the subsequent optimal estimation of the joint inverse dynamic variables of weights and concentrations.

[0142] This invention obtains the baseline flow rate in dry weather based on flow quantile thresholds, mass conservation optimization, and multi-component fitting, and decomposes groundwater infiltration. Through multi-level data preprocessing, trend decomposition, determination of background concentration in dry weather, correlation analysis, and anomaly data processing techniques, it can effectively solve the problems of difficulty in quantifying groundwater infiltration and low accuracy of baseline flow rate decomposition in urban drainage systems. Furthermore, the method of estimating the quality and quantity of infiltrated water in sewage treatment plant networks is low-cost, highly applicable in practical applications, and demonstrates significant beneficial effects, social benefits, and economic value.

[0143] S6: Construct the objective function and use the objective function to calculate the groundwater infiltration flow for each candidate dry-day baseflow quantile a. Additional wastewater flow and estimated infiltration concentration Perform joint optimization and determine the target quantile based on the optimization results. , and the target quantile The corresponding data set represents the quality and quantity of groundwater infiltration into the sewage treatment plant's pipeline network.

[0144] Based on the joint optimization of diurnal scale decomposition and infiltration concentration calibration, after completing the initial mass conservation estimation in step S54, for each candidate dry day baseflow quantile a, the infiltration flow rate and infiltration concentration are further corrected through joint optimization to ensure consistency among different components and overall mass balance.

[0145] S61: Set daily-scale flow mixing weights Used to dynamically regulate groundwater infiltration flow. and additional wastewater flow The allocation ratio between them. Daily-scale flow mixing weight. Range constraints This ensures the physical rationality of the allocation.

[0146] S62: Based on daily-scale flow mixed weighting Reclassified groundwater infiltration flow and additional wastewater flow The corrected groundwater infiltration flow rate is obtained by making adjustments. and correction of additional wastewater flow .

[0147] For each day time Based on mixed weights Regarding the groundwater infiltration flow rate in step S53 and additional wastewater flow The decomposition is corrected to obtain the corrected groundwater infiltration flow rate. and correction of additional wastewater flow And the correction method is as follows:

[0148]

[0149] .

[0150] As can be seen from the above correction methods, the correction process involves adjusting the groundwater infiltration rate after correction. Correcting additional wastewater flow Apply nonnegativity constraints to ensure that the corrected groundwater infiltration flow rate Correcting additional wastewater flow The predicted concentrations of each water quality component are greater than or equal to zero throughout the entire diurnal timescale series. The result is greater than zero throughout the entire diurnal scale sequence, fundamentally eliminating any unreasonable negative values.

[0151] S63: For each water quality component Estimates of infiltration concentration After perturbation and expansion, the estimated infiltration concentration is: Based on the modified groundwater infiltration flow rate Correcting additional wastewater flow rate Estimated infiltration concentration after perturbation expansion Predicted concentrations of each water quality component i .

[0152] Preferably, the estimated infiltration concentration The range of disturbance expansion is ,Right now The reasonable concentration range constraint specifies the concentration of each water quality component in the infiltration water (the estimated infiltration concentration after perturbation expansion is...). The concentration must be within the upper and lower limits allowed by experience to avoid abnormal concentrations that do not conform to reality, that is, to ensure that the adjustment of the infiltration concentration does not deviate from the physically acceptable range.

[0153] Predicted concentration calculations driven by infiltration concentration yield corrected groundwater infiltration flow rates. and correction of additional wastewater flow Then, based on the mass conservation constraint, for each daily time point... Each water quality component Predicted concentration Perform prediction calculations:

[0154] .

[0155] The mass conservation constraint ensures the predicted groundwater infiltration rate (corrected groundwater infiltration flow rate). ) and additional wastewater flow (corrected additional wastewater flow) The total amount and water concentration are balanced to ensure that the total water volume and the concentration of each component after mixing are consistent with the monitoring data.

[0156] S64: Minimizing diurnal concentration monitoring sequences Compared with predicted concentration The sum of squared residuals between them is the objective function. Jointly optimize daily-scale flow mixed weights Estimated infiltration concentration after perturbation expansion The simulated annealing algorithm is used to solve the problem, and the jointly optimized solution is obtained. and Corresponding daily-scale flow mixed weight Estimated infiltration concentration after perturbation expansion .

[0157] By minimizing the objective function Daily-scale flow mixed weight Estimated infiltration concentration after perturbation expansion Joint optimization is performed, which means simultaneously minimizing the predicted concentrations of all water quality components. Compared with the measured daily-scale concentration monitoring sequence To minimize deviations and ensure that all water quality components conform to the daily concentration monitoring sequence to the greatest extent possible across all daily time periods. :

[0158]

[0159] in Indicates the quantity of water quality components. ; The time index is the number of daily-scale samples corresponding to the observation duration. The above objective function During the entire observation period Inside, for all The errors of each water quality component are summed to ensure that the parameter estimation reaches a consistent optimal value in both the time dimension and the water quality component dimension.

[0160] As can be seen from steps S62 to S64, the joint optimization is carried out under multiple constraints, specifically non-negativity constraints, reasonable concentration range constraints, and mass conservation constraints.

[0161] In this embodiment, the simulated annealing algorithm is used to solve the above-mentioned multidimensional, nonlinear objective function that may have multiple local minima. In other words, simulated annealing is used as the global optimization strategy. Simulated annealing is a well-known algorithm in the field. During the search process, it controls the "temperature" to gradually decrease, allowing for the acceptance of poor solutions with a certain probability at high temperatures to escape local minima. It gradually converges as the temperature decreases, thereby finding the global or near-global optimum within a wide parameter space. In this embodiment, the simulated annealing algorithm runs at each candidate dry-day base current quantile 'a', iteratively updating the daily-scale flow mixing weights. Estimated infiltration concentration after perturbation expansion And force the above constraints to be satisfied until convergence.

[0162] S65: Compare the joint optimized values ​​for each candidate dryland baseflow quantile a. Take multiple joint optimization results The candidate dry-weather base current quantile 'a' corresponding to the smallest median value is used as the target quantile. ; and target quantile The corresponding data sets include , , and predicted concentration .

[0163] The optimal choice of candidate dry-day base current quantile 'a' for each candidate dry-day base current quantile All were jointly optimized according to S64 to obtain the corresponding quantile 'a' for each candidate dry-day base current. as well as Corresponding daily-scale flow mixed weight Estimated infiltration concentration after perturbation expansion , thus obtaining the objective function Minimal parameter set , and the corresponding minimization objective ,by As a unified performance metric, in the candidate set Execution meta selection: That is, selecting multiple jointly optimized The candidate dry-weather base current quantile 'a' corresponding to the smallest median value is used as the target quantile. .and The corresponding candidate drought baseline flow in step S4 Correcting groundwater infiltration flow Correcting additional wastewater flow rate and predicted concentration .

[0164] Preferably, when multiple candidate dry-day base current quantiles a appear... When the variances are the same or similar, residual variance should be chosen first. Smaller ones.

[0165] S7: Outputs the quality and quantity of groundwater infiltration into the sewage treatment plant's pipe network. Output and The corresponding candidate drought baseline flow in step S4 Correcting groundwater infiltration flow Correcting additional wastewater flow rate and predicted concentration .

[0166] Preferably, the output also includes the transformed daily-scale flow sequence. Simultaneously, it generates charts for verification and display, including total flow (daily-scale flow sequence). ) and various types of flow rates (corrected groundwater infiltration flow rate) and correction of additional wastewater flow Overlay plot of ) and predicted concentration; Daily-scale concentration monitoring sequence A comparison chart between them.

[0167] Existing technologies for quantifying groundwater infiltration in urban drainage systems mainly include the minimum nighttime flow method, the closure measurement method, and the characteristic factor method, or rely on traditional design standards. However, these methods have significant drawbacks in practical applications, as follows:

[0168] The minimum nighttime flow method assumes that the minimum nighttime flow during the dry season is equivalent to the groundwater infiltration rate, and extrapolates the groundwater infiltration rate using the average nighttime flow over a longer timescale during the dry season. Limitations of this method include: the nighttime flow rate used still includes a small amount of sanitary water flow, meaning the monitored minimum flow rate cannot represent the pure groundwater infiltration rate.

[0169] The core assumption of this method, which relies on the fact that "the nighttime sanitary flow rate is negligible during the dry season," does not take into account the fluctuations in domestic sewage due to seasonal and holiday factors, which leads to an inherent bias in groundwater infiltration. This method is only applicable to small-scale areas with simple pipe network conditions. When the research scope is expanded, the fluctuations in sanitary water use within the pipe network become more complex, and its core assumption is difficult to hold. Therefore, this method has poor generalizability.

[0170] The plugging-and-metering method involves sealing the inlet and outlet of a single pipe segment and measuring the changes in water level or volume within that segment over a period of time to estimate groundwater infiltration. However, this method has several drawbacks: it requires sealing and measuring each pipe segment individually, consuming significant manpower and resources, and can temporarily disrupt the normal operation of the pipe network, interfering with the stability of the urban drainage system. Furthermore, the plugging-and-metering method is difficult to apply on a large scale; for large-scale pipe network systems, the sheer number of pipelines, complex topological relationships, and the extreme difficulty of implementing segment-by-segment sealing make it unsuitable for widespread use.

[0171] The characteristic factor method selects conservative substances with significant concentration differences in wastewater as water quality characteristic factors and solves the water quantity and quality balance equation to separate the inflow and infiltration components. This method has relatively high accuracy, but its drawbacks include: the characteristic factor method relies on high-frequency water quality sampling, which results in high long-term monitoring costs; when applied to large areas, the representativeness of the characteristic factors is easily affected by the wide coverage of the pipe network and the large differences in the spatial and temporal distribution of water quality, and the monitoring cost increases exponentially with the expansion of the area, thus limiting the practicality of the method.

[0172] Traditional design standards stipulate that the infiltration rate in pipeline design should be 10% to 25% of the domestic sewage volume. This method not only seriously underestimates the actual infiltration rate, but also ignores the hydrological differences between regions by relying on the assumption of a fixed proportion. It does not take into account the differences in pipeline aging, groundwater depth, soil moisture, etc. in different regions, and its estimation lacks universality.

[0173] In summary, existing traditional methods for estimating groundwater infiltration volume cannot meet the needs of complex urban drainage systems for accurate decomposition of dry-day inflow components and dynamic identification of groundwater infiltration volume due to problems such as oversimplification of assumptions, high implementation costs, and limited applicability to space and scenarios.

[0174] Compared to existing quantification methods that rely on simplified assumptions (such as the minimum nighttime flow method) or single data dimensions, this invention accurately identifies rainwater infiltration flow using the Local Outlier Factor (LOF) algorithm, completes missing data using linear interpolation, and combines multi-source data fusion analysis of flow and water quality to achieve dynamic decomposition of baseline flow during dry weather and accurate quantification of groundwater infiltration. In particular, the introduction of technical mechanisms such as "dynamic separation of domestic sewage (extra wastewater) and groundwater" and "time series output of water quantity and quality linkage analysis" systematically ensures the accuracy of the decomposition results in key dimensions such as flow composition, concentration characteristics, and seasonal dynamics, truly achieving the quantification goal of "data-driven, dynamically adapted" and breaking through the limitations of traditional methods that rely on fixed-ratio assumptions.

[0175] This invention effectively solves problems such as interference from fluctuations in domestic sewage flow and difficulty in distinguishing complex flow components by constructing a full-process technical framework of "abnormal data cleaning - missing value completion - multi-component fitting". This framework can accurately separate the background domestic sewage and groundwater infiltration contribution from the baseline flow in dry weather. By capturing the seasonal trend and dynamic changes of groundwater infiltration, it can indirectly assess the condition of the pipe network, fundamentally avoiding the underestimation or misjudgment of infiltration caused by the simplification of assumptions in traditional methods, and ensuring the reliability and scientific nature of the results in practical engineering applications such as pipe network repair decisions and sewage treatment plant scheduling.

[0176] This invention eliminates the need for costly methods such as high-frequency water sampling and pipe blockage. It achieves precise quantification using only existing flow and water quality monitoring data from wastewater treatment plants, significantly reducing the manpower and resources required for long-term monitoring and data analysis. Furthermore, its output of continuous and complete daily-scale sequences provides accurate boundary conditions for wastewater network models, reducing model calibration and validation costs. It is particularly suitable for the efficient assessment and refined management of complex urban and regional drainage systems, significantly improving the economic efficiency of prioritizing network repairs and making investment decisions.

[0177] With the digital and intelligent upgrading of urban drainage system operation and maintenance, this invention can support dynamic monitoring of groundwater infiltration and precise control of baseline flow during dry weather, helping sewage treatment plants to operate stably, reducing the risk of pollutant discharge, and improving the level of urban water environment governance. At the same time, its indirect assessment capability of the pipe network status can promote the optimization of drainage facility lifecycle management, providing solid technical support for the construction of smart urban water management and resilient drainage systems.

[0178] Example 2

[0179] This invention discloses a dynamic estimation system for the quality and quantity of infiltrated water in a wastewater treatment plant's intake network, such as... Figure 2 As shown, it includes a data standardization module, a data preprocessing module, a dry weather background water quality concentration module, a trend flow term module, an infiltration initial estimation module, an objective function optimization module, and a result output module.

[0180] The data standardization module is used to obtain high-frequency influent flow data of wastewater treatment plants within the study area over a specified observation period of N days. Monitoring data of the concentration of all water components entering the plant. Where i represents the type of water quality component; and high-frequency influent flow data and influent water quality concentration monitoring data After standardization, daily flow sequences were obtained. and daily-scale concentration monitoring sequences The data conversion module performs step S1 in Example 1.

[0181] The data preprocessing module is used for daily-scale flow series. and daily-scale concentration monitoring sequences The system identifies, removes, and fills in missing values ​​for abnormally high values. The data processing module executes step S2 from Example 1.

[0182] The drought background water quality concentration module acquires rainfall meteorological data for the study area and historical water quality concentration monitoring data of all water components in the residential areas within the study area from the outlet pipes. Based on rainfall meteorological data, periods without rainfall were identified, and then historical water quality concentration monitoring data of all water quality components during these periods were selected from the outlet pipes. Based on historical water quality concentration monitoring data from the outlet pipe and influent water quality concentration monitoring data Constructing the dry-day background water quality concentration for each water quality component The drought background module executes step S3 in Example 1.

[0183] The trend flow item module uses the Time Series Smoothing (STL) algorithm to identify, remove, and impute missing values ​​in the daily-scale flow series. Decompose the daily-scale flow series The decomposed trend term is used as the trend flow term. The trend extraction module executes step S4 in Example 1.

[0184] Infiltration initial estimation module; sets multiple candidate dry-day baseflow quantiles 'a' and combines them with the trend flow term. Generate corresponding candidate drought baseline flow Combined with trend traffic items Baseline flow rates for each candidate drought day Daily-scale concentration monitoring sequences of each water quality component i Background water quality concentration during drought The groundwater infiltration flow rate under each candidate dry-day baseflow quantile 'a' was calculated. Additional wastewater flow Estimated infiltration concentration of groundwater The initial infiltration module performs step S5 in Example 1.

[0185] Objective function optimization module: Constructs the objective function and uses it to calculate the groundwater infiltration flow rate for each candidate dry-weather baseflow quantile a. Additional wastewater flow and estimated infiltration concentration Perform joint optimization and determine the target quantile based on the optimization results. , and the target quantile The corresponding data set represents the quality and quantity of groundwater infiltration into the sewage treatment plant's pipe network. The target optimization module executes step S6 in Example 1.

[0186] The result output module outputs the water quality and quantity results of groundwater infiltration into the sewage treatment plant's pipe network. The result output module executes step S7 in Example 1.

[0187] Example 3

[0188] The present invention discloses a method for dynamically estimating the quality and quantity of infiltrated water in a wastewater treatment plant's influent network. The method is applied to Area A in Shenzhen and includes the following steps:

[0189] Shenzhen Area A was selected as the study area, and step S1 in Example 1 was performed. The observation period N days was 731 days, specifically from January 2023 to January 2025. This example only requires analysis of the chemical oxygen demand (COD) component of the water quality, therefore only the COD concentration monitoring data of the influent water quality was obtained.

[0190] Perform step S2 in Example 1 to process the daily-scale flow sequence. and daily-scale concentration monitoring sequences The process involves identifying and removing abnormally high values, as well as imputing missing values. Furthermore, this embodiment utilizes a daily-scale flow sequence. The data sequence after identifying, removing, and imputing outlying high values ​​is as follows: Figure 4 The black broken line in the figure shows the daily-scale concentration monitoring sequence of this embodiment. The data sequence after identifying, removing, and imputing outlying high values ​​is as follows: Figure 3As shown by the black broken line in the image.

[0191] Perform step S3 in Example 1 to obtain historical water quality concentration monitoring data of all water components in residential areas within Shenzhen Area A during periods of no rainfall. And construct the arid-day background water quality concentration of this water quality component. .

[0192] Perform step S4 in Example 1 to process the daily-scale flow sequence. Decompose to obtain trend flow items .

[0193] Perform step S5 in Example 1, in A total of 20 candidate dry-day base current quantiles 'a' are set, forming a candidate quantile set. The calculated baseline flow rates for 20 candidate drought days .

[0194] Perform step S6 in Example 1 to calculate the target quantile. and the target quantile The corresponding data sets include , , and And the final output candidate drought baseline flow like Figure 4 The red broken line in the figure shows the corrected groundwater infiltration flow rate. like Figure 4 The blue broken line in the image indicates the correction for additional wastewater flow. like Figure 4 The green broken line in the image shows the predicted concentration. like Figure 3 The red broken line in the figure shows that the water quality component calculated in step S5 is at the target quantile. Estimated infiltration concentration like Figure 3 As shown by the blue broken line in the image.

[0195] Performing step S7 in Example 1, the output includes the transformed daily-scale flow sequence. Candidate dry weather baseline flow Correcting groundwater infiltration flow Correcting additional wastewater flow rate and predicted concentration The final output is as follows Figure 3 and Figure 4 As shown.

Claims

1. A method for dynamically estimating the quality and quantity of infiltrated water in a wastewater treatment plant's intake network, characterized in that: Includes the following steps, S1: Obtain high-frequency influent flow data for wastewater treatment plants within the study area over a specified observation period of N days. Monitoring data of the concentration of all water components entering the plant. , where i represents the type of water quality component; High-frequency inlet flow data and influent water quality concentration monitoring data After standardization, daily flow sequences were obtained. and daily-scale concentration monitoring sequences ; S2: Daily-scale flow sequence and daily-scale concentration monitoring sequences Perform identification, removal, and missing value imputation for abnormally high values; S3: Obtain rainfall meteorological data within the study area and historical water quality concentration monitoring data of all water quality components from the outlet pipes of all residential areas within the study area. Based on rainfall meteorological data, periods without rainfall were identified, and then historical water quality concentration monitoring data of all water quality components during these periods were selected from the outlet pipes. Based on historical water quality concentration monitoring data from the outlet pipe and influent water quality concentration monitoring data Constructing the dry-day background water quality concentration for each water quality component ; S4: The daily flow series after identifying, removing, and imputing outliers using the Time Series Smoothing (STL) algorithm is processed. Decompose the daily-scale flow series The decomposed trend term is used as the trend flow term. ; S5: Set multiple candidate dry-day baseflow quantiles 'a' and combine them with the trend flow term. Generate corresponding candidate drought baseline flow Combined with trend traffic items Baseline flow rates for each candidate drought day Daily-scale concentration monitoring sequences of each water quality component i Background water quality concentration during drought The groundwater infiltration flow rate under each candidate dry-day baseflow quantile 'a' was calculated. Additional wastewater flow Estimated infiltration concentration of groundwater ; S6: Construct the objective function and use the objective function to calculate the groundwater infiltration flow for each candidate dry-day baseflow quantile a. Additional wastewater flow and estimated infiltration concentration Perform joint optimization and determine the target quantile based on the optimization results. , and the target quantile The corresponding data set represents the quality and quantity of groundwater infiltration into the sewage treatment plant's pipe network. S7: Output the quality and quantity of groundwater infiltration in the sewage treatment plant's pipeline network.

2. The method for dynamically estimating the quality and quantity of infiltrated water in the wastewater treatment plant's receiving network according to claim 1, characterized in that: The high-frequency inlet flow data mentioned in step S1 and influent water quality concentration monitoring data The standardization process is based on high-frequency inbound flow data. Calculate the average daily inflow rate over the observation period N days. Average daily inflow Daily-scale flow series of high-frequency inflow data obtained by sorting by time ,in This represents the average daily total inflow to the plant on day N; Influent water concentration monitoring data The daily average concentration of influent water was obtained by taking the daily average value. Daily average concentration of water entering the plant The daily-scale concentration monitoring sequence was obtained by sorting by time. ,in This represents the average concentration of water component i on day N.

3. The method for dynamically estimating the quality and quantity of infiltrated water in the wastewater treatment plant's receiving network according to claim 1, characterized in that: Step S2 involves analyzing the daily-scale flow sequence. and daily-scale concentration monitoring sequences The method for identifying and removing outlier values ​​is as follows: The local outlier factor algorithm is used to analyze the daily-scale flow series. Outlier analysis was performed, and daily flow sequences were selected based on Euclidean distance. The set of points in the k-neighborhood , And calculate the reachable distance within the k-neighborhood. Locally achievable density and the corresponding local outlier factors ; For data points of daily scale flow series achievable distance The calculation method is as follows: ， in , Data points representing daily-scale flow series To its first The distance between neighbors Data points representing daily-scale flow series and The Euclidean distance between them; Where the locally reachable density is The calculation method is as follows: ， Data points for each daily-scale flow sequence Local outlier The calculation method is as follows: ； when And the corresponding daily-scale flow sequence data points Exceeding daily scale flow sequence When the preset quantile is used, the data points of the daily scale flow series It was identified as a locally abnormal high value point; For date m where an abnormally high value is identified, remove the daily-scale flow series data points within that date. Simultaneously remove daily-scale concentration monitoring sequences. Daily-scale concentration monitoring for that day ; Step S2 involves analyzing the daily-scale flow sequence. and daily-scale concentration monitoring sequences The method for handling missing values ​​is as follows: High-frequency inbound flow data... and influent water quality concentration monitoring data The original missing data points and daily scale flow series and daily-scale concentration monitoring sequences After identifying and removing abnormally high values, the resulting vacancies are marked as NaN, and linear interpolation is used to fill in the NaNs.

4. The method for dynamically estimating the quality and quantity of infiltrated water in the wastewater treatment plant's receiving network according to claim 1, characterized in that: In step S3, the dry-day background water quality concentration of each water quality component is constructed. The steps are as follows: S31: Calculate historical water quality concentration monitoring data for outdoor pipes during periods without rainfall. The mean value is used as the background concentration of water quality component i in the outlet pipe during dry weather. ; S32: Data from influent water quality concentration monitoring Select the maximum daily concentration value Constitutes the daily maximum concentration sequence And remove the daily maximum concentration sequence. Outliers in the sequence will be removed from the daily maximum concentration series. As a measure of background concentration during drought ; S33: Background concentration during dry weather will be measured at outdoor pipes. and detection of background concentration during drought days The larger values ​​are considered as water quality components. Background water quality concentration during drought ; S34: Repeat steps S31-S33 to calculate the dry-day background water quality concentration corresponding to each water quality component i. .

5. The method for dynamically estimating the quality and quantity of infiltrated water in the wastewater treatment plant's receiving network according to claim 1, characterized in that: The groundwater infiltration flow rate at each candidate dry-weather baseflow quantile a, as described in step S5, is obtained through calculation. Additional wastewater flow Estimated infiltration concentration of groundwater The steps are as follows: S51: Set multiple candidate dry-weather base current quantiles a, Based on candidate dry-day baseflow quantile 'a' and trend flow term Set the candidate dry-day baseline flow corresponding to each candidate dry-day baseflow quantile 'a'. , ; S52: For each candidate dry-day baseline flow Based on the trend flow term during the observation period N and candidate drought baseline flow Calculate the initial estimated daily groundwater infiltration rate. ,and ; S53: Divide the observation period of N days into multiple non-overlapping time periods of equal length in chronological order. Where K∈[1,N / L], and calculate in Time period Initial daily estimated groundwater infiltration Daily-scale concentration monitoring sequences of each water quality component i Correlation coefficient between Based on the initial estimate of groundwater infiltration volume and daily-scale concentration monitoring sequences During the corresponding time period Internal trends and correlation coefficients , each time period Initial estimated groundwater infiltration volume Reclassification as groundwater infiltration flow Or additional wastewater flow ; S54: Based on candidate dry weather baseline flow Groundwater infiltration flow Background water quality concentration during drought Calculate the water quality components under different candidate dry-day baseflow quantiles a. Estimates of infiltration concentration .

6. The method for dynamically estimating the quality and quantity of infiltrated water in the sewage treatment plant's influent network according to claim 5, characterized in that: In step S53, each time period Initial estimated groundwater infiltration volume Reclassification as groundwater infiltration flow Or additional wastewater flow The method is as follows: Calculate time period Initial estimate of groundwater infiltration time slope and daily-scale concentration monitoring sequences of ; like , And the correlation coefficient Greater than the preset threshold Then this time period Initial estimated groundwater infiltration volume Reclassified as additional wastewater flow ; like , or correlation coefficient Less than or equal to the preset threshold Then this time period Initial estimated groundwater infiltration volume Reclassification as groundwater infiltration flow ; like During this time period Initial estimated groundwater infiltration volume Reclassification as groundwater infiltration flow .

7. The method for dynamically estimating the quality and quantity of infiltrated water in the wastewater treatment plant's receiving network according to claim 5, characterized in that: In step S54, the water quality components under different candidate dry-day baseflow quantiles a are calculated. Estimates of infiltration concentration The method is as follows: Based on candidate drought baseline flow and groundwater infiltration flow Calculate the initial estimate of the daily infiltration concentration. The calculation formula is: ； The initial estimates of infiltration concentration for all daily times within the observation period N were calculated. According to groundwater infiltration flow rate The weighted time average yielded the candidate dry-day base current quantile 'a' and water quality components. Estimates of infiltration concentration The calculation formula is as follows: 。 8. The method for dynamically estimating the quality and quantity of infiltrated water in the sewage treatment plant's receiving network according to claim 5, characterized in that: Step S6, obtaining the results of groundwater infiltration quality and quantity in the sewage treatment plant's pipeline network, includes the following steps: S61: Set daily-scale flow mixing weights Used to dynamically regulate groundwater infiltration flow. and additional wastewater flow The distribution ratio between them; S62: Based on daily-scale flow mixed weighting Reclassified groundwater infiltration flow and additional wastewater flow The corrected groundwater infiltration flow rate is obtained by making adjustments. and correction of additional wastewater flow ; S63: For each water quality component Estimates of infiltration concentration After perturbation and expansion, the estimated infiltration concentration is: Based on the modified groundwater infiltration flow rate Correcting additional wastewater flow rate Estimated infiltration concentration after perturbation expansion Predicted concentrations of each water quality component i ; S64: Minimizing diurnal concentration monitoring sequences Compared with predicted concentration The sum of squared residuals between them is the objective function. Jointly optimize daily-scale flow mixed weights Estimated infiltration concentration after perturbation expansion The simulated annealing algorithm is used to solve the problem, and the jointly optimized solution is obtained. and Corresponding daily-scale flow mixed weight Estimated infiltration concentration after perturbation expansion ; S65: Compare the joint optimized values ​​for each candidate dryland baseflow quantile a. Take multiple joint optimization results The candidate dry-weather base current quantile 'a' corresponding to the smallest median value is used as the target quantile. ; and target quantile The corresponding data sets include , , and .

9. The method for dynamically estimating the quality and quantity of infiltrated water in the sewage treatment plant's influent network according to claim 8, characterized in that: In step S62, the groundwater infiltration flow rate after reclassification is measured. and additional wastewater flow The correction method is as follows: ， ； Predicting concentration in step S63 The calculation formula is as follows: ； The expression for the objective function in step S64 is as follows: ， in Indicates the quantity of water quality components. ; The time index is the number of daily-scale samples corresponding to the observation duration. .

10. A dynamic estimation system for the quality and quantity of infiltrated water in a wastewater treatment plant's influent network according to any one of claims 1 to 9, characterized in that: include, The data standardization module is used to obtain high-frequency influent flow data of wastewater treatment plants within the study area over a specified observation period of N days. Monitoring data of the concentration of all water components entering the plant. , where i represents the type of water quality component; And high-frequency inflow flow data and influent water quality concentration monitoring data After standardization, daily flow sequences were obtained. and daily-scale concentration monitoring sequences ; The data preprocessing module is used for daily-scale flow series. and daily-scale concentration monitoring sequences Perform identification, removal, and missing value imputation for abnormally high values; The drought background water quality concentration module acquires rainfall meteorological data for the study area and historical water quality concentration monitoring data of all water components in the residential areas within the study area from the outlet pipes. Based on rainfall meteorological data, periods without rainfall were identified, and then historical water quality concentration monitoring data of all water quality components during these periods were selected from the outlet pipes. Based on historical water quality concentration monitoring data from the outlet pipe and influent water quality concentration monitoring data Constructing the dry-day background water quality concentration for each water quality component ; The trend flow item module uses the Time Series Smoothing (STL) algorithm to identify, remove, and impute missing values ​​in the daily-scale flow series. Decompose the daily-scale flow series The decomposed trend term is used as the trend flow term. ; Infiltration initial estimation module; sets multiple candidate dry-day baseflow quantiles 'a' and combines them with the trend flow term. Generate corresponding candidate drought baseline flow Combined with trend traffic items Baseline flow rates for each candidate drought day Daily-scale concentration monitoring sequences of each water quality component i Background water quality concentration during drought The groundwater infiltration flow rate under each candidate dry-day baseflow quantile 'a' was calculated. Additional wastewater flow Estimated infiltration concentration of groundwater ; Objective function optimization module; Construct an objective function and use the objective function to calculate the groundwater infiltration flow for each candidate dry-day baseflow quantile a. Additional wastewater flow and estimated infiltration concentration Perform joint optimization and determine the target quantile based on the optimization results. , and the target quantile The corresponding data set represents the quality and quantity of groundwater infiltration into the sewage treatment plant's pipe network. The results output module outputs the results of the groundwater infiltration quality and quantity in the sewage treatment plant's pipeline network.

Images