A sewage COD component dynamic estimation method, system, device and medium based on sludge properties

Abstract

This invention relates to a method, system, equipment, and medium for dynamic estimation of wastewater COD components based on sludge properties. It includes: correlating and fusing organic pollution index data with process operation data to generate a contextual feature vector; pairing and fusing a COD component baseline ratio vector with the contextual feature vector to generate a preliminary COD component time series; obtaining the ratio of biodegradable COD to total COD in the sludge sample based on sludge property analysis data; using this ratio as a sludge property constraint parameter, combined with the analytical relationship constraints of the wastewater under test during the wastewater treatment process, to perform constraint optimization and error redistribution processing on the preliminary COD component time series, thereby obtaining the dynamic estimation result of the COD components of the wastewater under test. This invention achieves high-frequency, dynamic, and physically reasonable estimation of wastewater COD components, avoiding the unidirectional accumulation of errors towards inert components and improving the accuracy of the estimation results.

Description

Technical Field

[0001] This invention relates to the field of wastewater testing technology, and in particular to a method, system, equipment, and medium for dynamic estimation of COD components in wastewater based on sludge properties. Background Technology

[0002] Chemical oxygen demand (COD) is a key indicator for evaluating the degree of organic pollution in wastewater. In activated sludge wastewater treatment systems, total COD alone cannot accurately reflect the biodegradability, form, and degradation kinetics of organic matter. To support process design, model building, and operational optimization, the total influent COD needs to be divided into several components with clear biochemical significance, such as biodegradable dissolved substances (volatile fatty acids SVFA, readily biodegradable dissolved matrix SF), recalcitrant dissolved substances (SU), colloidal substances (CB), slowly biodegradable particulate matter (XB), and recalcitrant particulate matter (XU).

[0003] Currently, COD component analysis mainly relies on laboratory standard fractionation tests. Although this method yields accurate results, it is time-consuming, labor-intensive, and infrequent, typically only applicable under typical operating conditions. This results in static parameters that fail to reflect the dynamic changes in actual water quality, and errors can easily accumulate in inert components, affecting the accuracy of long-term predictions.

[0004] In recent years, rapid analysis methods based on online COD and BOD meters combined with multi-stage filtration have emerged. While these methods improve the efficiency of single-sample analysis, they remain limited to instantaneous sample analysis and fail to organically integrate with long-term operational data such as sludge age and influent load. They also lack the use of activated sludge properties as historical load integral constraints, leading to inconsistencies in component estimation over time, with errors still shifting towards inert components. Other studies have used machine learning methods to directly predict individual components from total COD; however, due to the lack of physical constraints such as mass conservation and material balance, the stability and physical plausibility of the prediction results are insufficient when operating conditions fluctuate.

[0005] Therefore, those skilled in the art urgently need a technical solution that can integrate multi-scale monitoring data, introduce sludge properties as long-term physical constraints, realize dynamic estimation of COD components, and have good engineering applicability, so as to improve the accuracy and reliability of wastewater treatment process simulation and control. Summary of the Invention

[0006] (a) Technical problems to be solved

[0007] This invention addresses the problems of existing COD component estimation methods, such as large experimental workload, static results, errors concentrated on inert components, and lack of long-term physical constraints. It proposes a dynamic estimation method, system, equipment, and medium for influent COD components based on sludge properties and multi-scale data fusion. The aim is to achieve high-frequency, dynamic, and physically reasonable long-term estimation of COD components, providing reliable input for the simulation, optimization, and control of wastewater treatment processes.

[0008] (II) Technical Solution

[0009] To achieve the above objectives, the main technical solutions adopted by the present invention include:

[0010] In a first aspect, embodiments of the present invention provide a method for dynamically estimating the COD components of wastewater based on sludge properties, including:

[0011] The organic pollution index data monitored by graded filtration during the wastewater treatment process are correlated and fused with the process operation data to generate a context feature vector characterizing the dynamic changes of organic pollution indexes in the wastewater to be tested.

[0012] The baseline proportion vector of COD components in wastewater obtained from COD fractionation experimental data is paired and fused with the context feature vector to generate an initial time series of COD components of the wastewater under test changing over time.

[0013] Based on the sludge property analysis data of sludge samples in the aerobic tank, the ratio of biodegradable COD to total COD in the sludge samples was obtained, and this ratio was used as a constraint parameter for the sludge properties of the wastewater to be tested.

[0014] Based on the sludge property constraint parameters and combined with the analytical relationship constraints of the wastewater to be tested in the wastewater treatment process, the initial COD component time series is constrained and optimized and the error is redistributed to obtain the dynamic estimation results of the COD components of the wastewater to be tested.

[0015] Optionally, the organic pollution index data monitored through staged filtration during wastewater treatment are correlated and fused with process operation data to generate a contextual feature vector characterizing the dynamic changes of organic pollution indexes in the wastewater under test, including:

[0016] Acquire the graded filtration monitoring data of the wastewater to be tested in a time series within a set period. The graded filtration monitoring data includes organic pollution index data and corresponding process operation data.

[0017] The organic pollution index data and process operation data are time-stamped and standardized to form a standardized time series data table.

[0018] Based on the time series data table, multi-level feature construction is performed on the data at each time point to generate basic feature variables to characterize the state of organic pollution in wastewater, derived feature variables to characterize the influence of external conditions on wastewater treatment, and time series feature variables to characterize the dynamic change trend of organic pollution indicators.

[0019] The basic feature variables, derived feature variables, and time-series feature variables at each time point are concatenated according to a predetermined feature order to generate a multi-dimensional context feature vector.

[0020] The organic pollution index data include: particulate CODgf and BODgf obtained through the first-stage filtration, and dissolved CODmf and BODmf obtained through the second-stage filtration.

[0021] Optionally, based on the time series data table, multi-level feature construction is performed on the data at each time point to generate basic feature variables characterizing the state of organic pollution in wastewater, derived feature variables characterizing the influence of external conditions on wastewater treatment, and time-series feature variables characterizing the dynamic changing trend of organic pollution indicators, including:

[0022] Numerical extraction of organic pollution index data at each time point in the time series data table yields basic characteristic variables representing the current organic pollution status of wastewater.

[0023] Process operation data are used as environmental and load factors and coupled with basic characteristic variables to generate derived characteristic variables that characterize the impact of external conditions on wastewater treatment.

[0024] A sliding window analysis was performed on the organic pollution index data along the time dimension to extract statistical features including mean, standard deviation, and rate of change, thus representing time-series characteristic variables that characterize the dynamic changing trend of organic pollution indicators.

[0025] Optionally, before pairing and fusing the baseline proportion vector of COD components in wastewater obtained from COD fractionation experimental data with the context feature vector to generate a preliminary time series estimate of COD components in the wastewater under test over time, the method further includes:

[0026] Based on COD fractionation test data of wastewater samples under different operating conditions, a multi-component benchmark database of wastewater COD corresponding to different operating conditions was established.

[0027] Based on the wastewater COD multi-component benchmark database, the detailed chemical components of COD under various operating conditions were determined, including: readily biodegradable volatile fatty acids S. VFA Easily biodegradable sugars S F Slow-degradable organic matter S U Colloidal biodegradable organic matter C B Particulate biodegradable organic matter XB and recalcitrant organic matter X U And constitute the reference proportion vector F of influent COD components under the current operating conditions. base =[S VFA ,S F ,S U C B ,X B ,X U ];

[0028] The identification information of different operating conditions is associated with the corresponding COD component benchmark ratio vector to generate a contextual wastewater fingerprint database, in which the identification information includes influent time period, season, process stage and water purification status.

[0029] Optionally, the baseline proportion vector of COD components in wastewater obtained based on COD fractionation experimental data is paired and fused with the context feature vector to generate a preliminary time series estimate of COD components in the wastewater under test over time, including:

[0030] From the pre-set contextual wastewater fingerprint database, retrieve the historical benchmark proportion vector that matches the working conditions represented by the contextual feature vector. The historical benchmark proportion vector is obtained by performing component analysis and statistics on COD fractionation experimental data of similar wastewater.

[0031] Based on the organic pollution index data of the wastewater to be tested at the current time, the content distribution of various forms of organic matter is analyzed, and the historical baseline proportion vector is dynamically corrected based on the analysis results to obtain the optimal baseline proportion vector that is suitable for the current working conditions.

[0032] The optimal baseline scaling vector and the context feature vector are fused at the feature level to generate the initial estimate of COD components at the current sampling time.

[0033] The context feature vectors and corresponding organic pollution index data at each time point are traversed sequentially. The retrieval, correction and fusion steps are repeated to generate a preliminary time series of COD components of the wastewater to be tested that changes continuously over time.

[0034] Optionally, based on the sludge property analysis data of the sludge samples in the aerobic tank, the ratio of biodegradable COD to total COD in the sludge samples is obtained, and this ratio is used as a sludge property constraint parameter for the wastewater to be tested, including:

[0035] Sludge samples were collected from the aerobic tank at a set sampling period, and the total COD of the sludge samples was measured by a preset COD detection device.

[0036] After performing aeration biodegradation on the sludge sample, the remaining COD in the sludge sample is measured by a COD detection device, and the remaining COD is subtracted from the total COD to obtain the COD of aeration degradation.

[0037] The ratio of COD from aeration degradation to total COD is determined as a sludge property constraint parameter characterizing the biochemical degradation properties of the wastewater under test.

[0038] Optionally, based on sludge property constraint parameters and combined with the analytical relationship constraints of the wastewater under test during the wastewater treatment process, the initial COD component time series is subjected to constraint optimization and error redistribution processing to obtain the dynamic estimation results of the COD components of the wastewater under test, including:

[0039] Within the set sliding time window, the initial estimated time series of COD components is used as the initial solution for optimization iteration, and an optimization model including mass conservation constraints, analytical relation constraints, and sludge property constraints is established.

[0040] With the objective of minimizing the overall estimation bias of the initial COD component time series, the optimization model is solved to obtain the optimal COD component time series.

[0041] During the optimization process, based on the sensitivity of each component to changes in the graded filtration monitoring, the estimation error of the inert component in the initial COD component time series is allocated to a predefined transition component group between the biodegradable component and the inert component.

[0042] The optimal COD component time series obtained by optimization is output as the dynamic estimation result of COD components of the wastewater to be tested.

[0043] Among them, the mass conservation constraint is the sum of the estimated values ​​of all COD components at each time point, which is consistent with the organic pollution index data at that time point; the analytical relationship constraint is based on the organic pollution index data obtained from graded filtration monitoring, which constrains the analytical quantitative relationship between each component, including the estimated value of colloidal COD equals the difference between the BOD after primary filtration and the BOD after secondary filtration, the estimated value of soluble inert COD equals the difference between the COD after secondary filtration and the BOD after secondary filtration, and the estimated value of particulate COD equals the difference between the total COD and the COD after primary filtration.

[0044] Secondly, embodiments of the present invention provide a dynamic estimation system for wastewater COD components based on sludge properties, comprising:

[0045] The online monitoring module is used to correlate and fuse the organic pollution index data monitored by graded filtration during the wastewater treatment process with the process operation data to generate a context feature vector characterizing the dynamic changes of organic pollution indexes in the wastewater under test.

[0046] The component estimation module is used to pair and fuse the baseline proportion vector of COD components in wastewater obtained based on COD fractionation experimental data with the context feature vector to generate a preliminary time series of COD components of the wastewater under test changing over time.

[0047] The sludge property constraint module is used to obtain the ratio of biodegradable COD to total COD in the sludge sample based on the sludge property analysis data of the sludge sample in the aerobic tank, and use this ratio as the sludge property constraint parameter of the wastewater to be tested.

[0048] The constraint optimization and error redistribution module is used to perform constraint optimization and error redistribution processing on the initial COD component time series based on sludge property constraint parameters and combined with the analytical relationship constraints of the wastewater to be tested in the wastewater treatment process, so as to obtain the dynamic estimation results of COD components of the wastewater to be tested.

[0049] Thirdly, embodiments of the present invention provide a dynamic estimation device for wastewater COD components based on sludge properties, comprising:

[0050] At least one processor;

[0051] and memory that is communicatively connected to at least one processor;

[0052] The memory stores instructions that can be executed by at least one processor, which enables the at least one processor to perform the above-described method for dynamic estimation of wastewater COD components based on sludge properties.

[0053] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method for dynamic estimation of wastewater COD components based on sludge properties.

[0054] (III) Beneficial Effects

[0055] The beneficial effects of this invention are as follows: The proposed method for dynamic estimation of wastewater COD components based on sludge properties, by employing online graded filtration monitoring and process operation data fusion to construct dynamic features, introducing historical COD fractionation experimental benchmark ratios as prior knowledge, and innovatively utilizing sludge property analysis data as long-term physical constraints, combined with an optimization model incorporating mass conservation, analytical relationships, and sludge property constraints for error redistribution, achieves high-frequency, dynamic, and physically reasonable estimation of influent COD components compared to existing technologies. This effectively avoids the unidirectional accumulation of errors towards inert components, significantly improving the fractionation accuracy and long-term reliability of the estimation results. Attached Figure Description

[0056] Figure 1 This is a flowchart illustrating a method for dynamically estimating wastewater COD components based on sludge properties, provided in an embodiment of the present invention.

[0057] Figure 2 This is a software architecture diagram of a dynamic estimation method for wastewater COD components based on sludge properties, provided in an embodiment of the present invention.

[0058] Figure 3 This is a schematic diagram comparing the dynamic curve of COD components in one embodiment of the present invention with the traditional static COD component parameters. Detailed Implementation

[0059] To better explain and facilitate understanding of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0060] refer to Figures 1 to 3 As shown in the embodiment of the present invention, a dynamic estimation method for wastewater COD components based on sludge properties is proposed. This method is applicable to activated sludge wastewater treatment systems with biochemical reaction tanks. The method includes: correlating and fusing organic pollution index data monitored through graded filtration during wastewater treatment with process operation data to generate a context feature vector for the dynamic changes of organic pollution indexes in the wastewater to be tested; pairing and fusing a baseline proportion vector of COD components in wastewater obtained based on COD fractionation experimental data with the context feature vector to generate a preliminary time series of COD components of the wastewater to be tested changing over time; obtaining the ratio of biodegradable COD to total COD in the sludge sample based on sludge property analysis data of sludge samples in the aerobic tank, and using this ratio as a sludge property constraint parameter for the wastewater to be tested; and performing constraint optimization and error redistribution processing on the preliminary COD component estimation time series based on the sludge property constraint parameter and the analytical relationship constraint of the wastewater to be tested during the wastewater treatment process to obtain the dynamic estimation result of COD components of the wastewater to be tested.

[0061] This embodiment, by employing online graded filtration monitoring and process operation data fusion to construct dynamic features, introducing historical COD fractionation experimental benchmark ratios as prior knowledge, and innovatively utilizing sludge property analysis data as long-term physical constraints, combined with an optimization model incorporating mass conservation, analytical relationships, and sludge property constraints for error redistribution, can achieve high-frequency, dynamic, and physically reasonable estimation of influent COD components compared to existing technologies. It effectively avoids the unidirectional accumulation of errors towards inert components, significantly improving the temporal consistency and long-term reliability of the estimation results.

[0062] To better understand the above technical solutions, exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that the present invention can be understood more clearly and thoroughly, and that the scope of the present invention can be fully conveyed to those skilled in the art.

[0063] Specifically, refer to Figure 1 and Figure 2 As shown, this embodiment specifically discloses a method for dynamically estimating the COD components of wastewater based on sludge properties. The method includes the following steps S100 to S400:

[0064] S100. The organic pollution index data monitored by graded filtration during the wastewater treatment process are correlated and fused with the process operation data to generate a context feature vector characterizing the dynamic changes of the organic pollution index in the wastewater to be tested.

[0065] In this embodiment, by associating the organic pollution index data of the wastewater to be tested with the process operation data of wastewater treatment, a contextual feature vector that can comprehensively and dynamically characterize the organic pollution status of wastewater is constructed. This contextual feature vector not only includes the original pollution level monitoring values ​​of the pollution indicators, but also incorporates the influence of the wastewater treatment system's operating status and the inherent laws governing the evolution of the indicators over time, thereby providing a high-information-density input for subsequent assessment and prediction of wastewater COD components. Specifically, step S100 may include the following sub-steps S110 to S140:

[0066] S110. Obtain the graded filtration monitoring data of the wastewater to be tested, arranged in time series within a set time period. The graded filtration monitoring data includes organic pollution index data and corresponding process operation data. The organic pollution index data includes: particulate CODgf and BODgf obtained through the first stage of filtration, and dissolved CODmf and BODmf obtained through the second stage of filtration.

[0067] For example, an online automatic monitoring device for the classification of organic matter in wastewater influent can be installed. This device has at least two interconnected filtration channels with filter membranes of 1.2 μm and 0.45 μm, respectively. Before filtration, the total COD of the influent can be measured as CODtot using an online COD meter based on the potassium dichromate method. After filtration through the 1.2 μm filter membrane, the particulate CODgf and BODgf of the first filtrate can be measured using the same online COD meter. The dissolved CODmf and BODmf of the filtrate after filtration through the 0.45 μm filter membrane can then be measured using the same online COD meter based on the same online COD method. The wastewater treatment process operation data includes: influent flow rate, influent temperature, SRT (sludge retention time), precipitation, and primary sedimentation tank operating conditions.

[0068] S120. Perform timestamp alignment and standardization on organic pollution index data and process operation data to form a standardized time series data table.

[0069] Furthermore, timestamp alignment ensures comparability of all variables at the same time point by uniformly interpolating or aggregating data from different sources to the same time granularity and filling in missing points using methods such as linear interpolation or forward imputation. Standardization processes include data cleaning and Z-score standardization. Data cleaning is used to identify and handle outliers and erroneous records to ensure data validity; Z-score standardization processes all numerical variables to make their mean 0 and standard deviation 1, thereby eliminating the negative impact of different physical units on subsequent model calculations and avoiding numerical instability or model bias caused by differences in variable scales.

[0070] S130. Based on the time series data table, construct multi-level features for the data at each time point to generate basic feature variables that characterize the state of organic pollution in wastewater, derived feature variables that characterize the influence of external conditions on wastewater treatment, and time series feature variables that characterize the dynamic change trend of organic pollution indicators.

[0071] Further, step S130 may specifically include the following sub-steps S131 to S133:

[0072] S131. Numericalize the organic pollution index data for each time node in the time series data table to obtain the basic characteristic variables representing the organic pollution status of wastewater at the current moment.

[0073] The basic characteristic variables include: the proportion of colloidal pollutants: R_CODgf=(CODgf-CODmf) / total COD, R_BODgf=(BODgf-BODmf) / total BOD; the proportion of dissolved pollutants: R_CODmf=CODmf / total COD, R_BODmf=BODmf / total BOD; the comprehensive biodegradability characteristics: the ratio of total BOD / COD: F_BOD / COD=total BOD / total COD; and the biodegradability characteristics of the dissolved phase: the ratio of dissolved BOD / COD: F_BOD / CODmf=BODmf / CODmf.

[0074] S132. The process operation data is used as environmental factors and load factors, and coupled with the basic characteristic variables to generate derived characteristic variables that characterize the influence of external conditions on wastewater treatment.

[0075] The derived characteristic variables include: pollutant load characteristics: colloidal chemical oxygen demand load L_CODgf=(CODgf-CODmf)×influent flow rate Q, dissolved chemical oxygen demand load L_CODmf=CODmf×influent flow rate Q; sludge age correlation characteristics: constructing the combination characteristics of sludge retention time and biodegradability, I_BOD=SRT / (F_BOD / COD), where SRT is the sludge age; rainfall impact characteristics: based on precipitation P data, generating Boolean-type rainfall event flags (true when P exceeds a preset threshold), and / or continuous rainfall impact coefficients (C=P / P_avg, where P_avg is the historical average precipitation).

[0076] S133. Perform sliding window analysis on organic pollution index data along the time dimension to extract statistical features including mean, standard deviation, and rate of change, and time-series characteristic variables that characterize the dynamic change trend of organic pollution indicators.

[0077] The time-series characteristic variables include: sliding statistical characteristics: calculating the moving average AV_CODmf, moving standard deviation SD_CODmf, and rate of change ΔCODmf of dissolved chemical oxygen demand (CODmf) within a preset time window; event triggering characteristics: identifying abrupt events in the process operation data. When the instantaneous rate of change of the influent flow rate Q exceeds a preset threshold, a flow impact flag is generated and the corresponding peak values ​​of chemical oxygen demand (CODgf) and CODmf are recorded.

[0078] S140. The basic feature variables, derived feature variables, and time-series feature variables of each time node are concatenated according to a predetermined feature order to generate a multi-dimensional context feature vector.

[0079] S200. The baseline proportion vector of COD components in wastewater obtained based on COD fractionation experimental data is paired and fused with the context feature vector to generate an initial time series of COD components of the wastewater to be tested changing over time.

[0080] In this embodiment, a dynamic, data-driven recursive estimation method is adopted to construct a preliminary time series of COD components in the wastewater under test over time. Instead of directly using a static baseline ratio, it organically combines historical baseline data from fractionation experiments, real-time monitored graded organic pollution indicators, and dynamically changing contextual operating conditions. Through intelligent fusion and online correction, a preliminary COD component estimation series that conforms to general patterns and is adapted to specific scenarios is generated.

[0081] In this embodiment, before step S200, the following steps G100 to G300 are also included:

[0082] G100. Based on COD fractionation test data of wastewater samples under different operating conditions, establish a multi-component benchmark database of wastewater COD corresponding to different operating conditions.

[0083] Furthermore, in multiple typical context scenarios of the wastewater treatment system, standard COD fractionation tests were conducted on representative influent samples to obtain multi-component benchmark values ​​of influent COD under each scenario; the context scenarios include at least one or more of the following dividing dimensions: (1) rainy season / dry season, or before and after a rainfall event; (2) weekday / weekend or holiday / peak and off-peak tourist season; (3) whether the primary sedimentation tank and pretreatment facilities are in operation; (4) the ratio of domestic sewage to industrial wastewater, the service population, and the industrial load level; (5) temperature range, SRT range, etc.

[0084] G200. Based on the multi-component benchmark database of wastewater COD, determine the detailed chemical components of COD under each operating condition, including: readily biodegradable volatile fatty acids S. VFA Easily biodegradable sugars S F Slow-degradable organic matter S U Colloidal biodegradable organic matter C B Particulate biodegradable organic matter X B and recalcitrant organic matter X U And constitute the reference proportion vector F of influent COD components under the current operating conditions. base =[S VFA ,S F ,S U C B ,X B ,X U ].

[0085] G300 associates the identification information of different operating conditions with the corresponding COD component benchmark ratio vector to generate a contextual wastewater fingerprint database, in which the identification information includes influent time period, season, process stage and water purification status.

[0086] In this embodiment, step S200 may include the following sub-steps S210 to S240:

[0087] S210. Retrieve a historical benchmark proportion vector from the preset context wastewater fingerprint database that matches the operating conditions represented by the context feature vector. The historical benchmark proportion vector is obtained by performing component analysis and statistics on COD fractionation experimental data of similar wastewater.

[0088] Furthermore, based on the context feature vector C(t) at the current time t, one or more reference working condition scenarios most similar to the context feature vector C(t) are selected from the wastewater fingerprint database to obtain the corresponding COD component benchmark proportion vector F. base(t), the selection method can be: matching based on manually set rules (e.g., dividing by rainy / sunny days), nearest neighbor matching based on statistical distance or similarity measures, or scene recognition and interpolation based on machine learning classification / regression models.

[0089] S220. Based on the organic pollution index data of the wastewater to be tested at the current time, analyze the content distribution of various forms of organic matter, and dynamically correct the historical benchmark ratio vector with the analysis results to obtain the optimal benchmark ratio vector that is suitable for the current working conditions.

[0090] Furthermore, by utilizing the analytical relationship between online COD and BOD obtained from hierarchical monitoring, the baseline proportion vector F is... base (t) is corrected, and the analytical relations include: C B (t) = BODgf(t) - BODmf(t), S U (t) = CODmf(t) - BODmf(t), X B (t)+X U (t) = CODtot(t) - CODgf(t).

[0091] S230. Perform feature-level fusion of the optimal baseline scale vector and the context feature vector to generate the initial estimate of COD components at the current sampling time.

[0092] S240. Iterate through the context feature vectors and corresponding organic pollution index data at each time point in sequence, repeat the retrieval, correction and fusion steps, and generate the preliminary estimated time series of COD components of the wastewater to be tested that change continuously over time.

[0093] Furthermore, the initial time series of COD components F raw (t)=[S VFAraw (t),S Fraw (t),S Uraw (t), C Braw (t),X Braw (t),X Uraw (t)]. Where X Braw (t) represents the weighted average method used to calculate C. Braw (t) and C given by the fingerprint database B X is obtained by fusing the ratio × CODtot(t). Uraw (t) represents the weighted average for S. U The ratio × CODtot(t) is finely adjusted to obtain the result.

[0094] It is worth mentioning that unsupervised clustering methods based on K-means can also be used to cluster the historical context feature vector C(t) data and CODtot(t) data, automatically forming contextual operating scenarios, and then building a fingerprint database within each cluster scenario; alternatively, supervised learning methods such as random forests or gradient boosting trees can be used, with actual fractionation test results as labels, to learn the mapping relationship between C(t) and COD components, and output F in the online stage. base (t).

[0095] S300. Based on the sludge property analysis data of the sludge sample in the aerobic tank, obtain the ratio of biodegradable COD to total COD in the sludge sample, and use this ratio as the sludge property constraint parameter of the wastewater to be tested.

[0096] In this embodiment, the acquisition of sludge property constraint parameters is based on a core biochemical characteristic of the activated sludge treatment process: the actual degradation capacity of sludge for organic matter. This parameter aims to provide a crucial, dynamic, on-site constraint for accurate estimation of subsequent COD components from the perspective of the working state of sludge microorganisms. By measuring the proportion of degradable COD in sludge samples after aerobic aeration within a specific time period, the actual utilization efficiency of the microbial community in the sludge mixed liquor for biodegradable components in wastewater under the current operating conditions can be directly quantified, thereby linking the static component proportions in the laboratory with the dynamic biological treatment process of the wastewater treatment plant. Specifically, step S300 may include the following sub-steps S310 to S340:

[0097] S310. Collect sludge samples from the aerobic tank at the set sampling period, and measure the total COD of the sludge samples using a preset COD detection device.

[0098] S320. After performing aeration biodegradation on the sludge sample, the remaining COD in the sludge sample is measured by a COD detection device, and the remaining COD is subtracted from the total COD to obtain the COD of the aeration degradation.

[0099] S330. The ratio of COD degraded by aeration to total COD is determined as a sludge property constraint parameter characterizing the biochemical degradation characteristics of the wastewater to be tested.

[0100] For example, on a timescale corresponding to the average sludge settling time (SRT) (e.g., 30 days), mixed liquor samples are collected every 15 days at the end of the aerobic tank. After sedimentation and separation, the sludge portion is used for the following experiment: COD is determined on the sludge sample to obtain the COD. tot_sludge (Or, VSS can be used to characterize the concentration of the organic fraction). Then, a portion of the sludge sample is placed in a sealed aeration bottle and aerated for an extended period (e.g., 7 days) under constant temperature and oxygenation conditions. The COD or OUR curves within the bottle are periodically measured, and the total degradable COD in the sludge sample is calculated.degradable_sludge Finally, based on the initial COD tot_sludge With COD degradable_sludge The difference: X BU_sludge =COD degradable_sludge / (COD degradable_sludge +COD inert_sludge ), X BU_sludge Sludge property constraint parameters that reflect the historical biodegradability / inertness of influent.

[0101] S400. Based on the sludge property constraint parameters and combined with the analytical relationship constraints of the wastewater to be tested in the wastewater treatment process, the initial COD component time series is constrained and optimized and the error is redistributed to obtain the dynamic estimation results of the COD components of the wastewater to be tested.

[0102] In this embodiment, a multi-constraint optimization framework is established to forcibly match and synergistically optimize the initial COD component estimation time series generated in the preceding steps with the sludge biochemical degradation characteristics (sludge property constraints) and the inherent component relationships resolved through staged filtration monitoring during wastewater treatment (relationship constraints). This optimization process not only corrects potential systematic biases in the initial estimation series but also ensures, through a specific error redistribution strategy, that the estimation results conform to the basic principles of mass conservation and component resolution, and accurately reflect the actual degradation capacity of the sludge under current operating conditions, thereby obtaining a more reliable and practically accurate dynamic estimation result of COD components. Specifically, step S400 may include the following sub-steps S410 to S440.

[0103] S410. Within the set sliding time window, the initial estimated time series of COD components is used as the initial solution for optimization iteration, and an optimization model including mass conservation constraints, analytical relation constraints, and sludge property constraints is established.

[0104] Furthermore, the mass conservation constraint is the sum of the estimated values ​​of all COD components at each time point, consistent with the organic pollution index data at that time point; the analytical relationship constraint is based on the organic pollution index data obtained from graded filtration monitoring, constraining the analytical quantitative relationship between each component, including the estimated value of colloidal COD equal to the difference between the BOD after primary filtration and the BOD after secondary filtration, the estimated value of soluble inert COD equal to the difference between the COD after secondary filtration and the BOD after secondary filtration, and the estimated value of particulate COD equal to the difference between the total COD and the COD after primary filtration.

[0105] S420. With the objective of minimizing the overall estimation bias of the initial COD component time series, solve the optimization model to obtain the optimal COD component time series.

[0106] S430. During the optimization process, based on the sensitivity of each component to changes in the graded filtration monitoring, the estimation error of the inert component in the initial COD component time series is allocated to the predefined transition component group between the biodegradable component and the inert component.

[0107] S440. The optimal COD component time series obtained from the optimization solution is output as the dynamic estimation result of COD components of the wastewater to be tested.

[0108] To further explain, taking the time point corresponding to the current sludge property test as the center, a length of T is selected. w Time window [t0, t0+T] w Within this time window, the estimated COD component F(t) for each time step should satisfy the following: an analytical relationship with CODtot(t), CODgf(t), CODmf(t), BODgf(t), and BODmf(t); and each component should be non-negative and not exceed a reasonable upper bound. Throughout the entire time window, based on the activated sludge mass balance theory, the influent COD component F(t), influent flow rate Q(t), and sludge property constraint parameter X can be used to determine the optimal parameters. BU_sludge Integrate constraints. For example, an approximate constraint or objective can be constructed as follows: the cumulative biodegradable COD load per unit time calculated by F(t) simulation versus the actually observed X. BU_sludge Consistent; or the theoretical sludge production derived from F(t) matches the actual sludge production.

[0109] Next, let F raw F(t) represents the initial estimated component sequence, and F(t) represents the comprehensive objective function constructed from the optimized sequence. ,in, This indicates that adjustments will be made without deviating too much from the initial estimate. This indicates the relationship between the sludge properties derived from F(t) and the actual X. BU_sludge The difference between them is measured, and λ is the weighting parameter.

[0110] Then, quadratic programming or other optimization methods are used to solve the above optimization problem to obtain the COD component Fopt(t) at each time t within the time window. In this process, to satisfy the sludge property constraints and analytical relationships, X... B (t) and X U The time series of (t) is coordinated and adjusted to shift the focus from X to X. Uraw The error in (t) is redistributed within a reasonable range to the biodegradable and inert components.

[0111] Finally, the time window is shifted forward, for example, updating every Δt = 1 hour, and the above optimization process is repeated to achieve continuous updating and long-term operation of COD component estimation. The optimized dynamic COD component estimation result F is then displayed.opt (t) is stored in the historical database and converted into the input format required by the ASM series activated sludge model through the model interface module. It is used for operation status playback and diagnosis, simulation of effluent water quality and energy consumption under different control strategies, guidance for optimization of process parameters such as aeration rate, return ratio, and sludge discharge rate, and as a feature input of the intelligent control system to improve the sensitivity and robustness of the control strategy to changes in influent water quality.

[0112] Secondly, this embodiment also proposes a dynamic estimation system for wastewater COD components based on sludge properties, including:

[0113] The online monitoring module is used to correlate and fuse organic pollution index data monitored through graded filtration during the wastewater treatment process with process operation data to generate a context feature vector characterizing the dynamic changes of organic pollution indexes in the wastewater to be tested.

[0114] The component preliminary estimation module is used to pair and fuse the baseline proportion vector of COD components in wastewater obtained based on COD fractionation experimental data with the context feature vector to generate a preliminary time series of COD components of the wastewater under test changing over time.

[0115] The sludge property constraint module is used to obtain the ratio of biodegradable COD to total COD in sludge samples based on sludge property analysis data from sludge samples in the aerobic tank, and to use this ratio as the sludge property constraint parameter for the wastewater to be tested.

[0116] The constraint optimization and error redistribution module is used to perform constraint optimization and error redistribution processing on the initial COD component time series based on sludge property constraint parameters and combined with the analytical relationship constraints of the wastewater to be tested in the wastewater treatment process, so as to obtain the dynamic estimation results of COD components of the wastewater to be tested.

[0117] Thirdly, this embodiment also proposes a wastewater COD component dynamic estimation device based on sludge properties, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute the above-described wastewater COD component dynamic estimation method based on sludge properties.

[0118] Fourthly, this embodiment also proposes a computer-readable storage medium storing computer-executable instructions. When the computer-executable instructions are executed by a processor, they implement the above-described method for dynamic estimation of wastewater COD components based on sludge properties.

[0119] In one specific embodiment, such as Figure 3 The figure shows particulate biodegradable organic matter X in the COD composition.B Data examples of the index in different estimation schemes: CODtot is the total COD measured by the water plant's own online monitoring instrument; XB_stable is the static parameter × CODtot after the laboratory standard fractionation test, so the index data change characteristics are completely consistent with CODtot; XB_base is the value after one correction (i.e., the X value corresponding to the initial estimate of COD components obtained in step S200 of this embodiment). B The indicator data changes in tandem with CODtot, but has been dynamically corrected based on time series and contextual characteristics; XB_opt is the final corrected value (i.e., the X in the final dynamic estimation result of COD components obtained in this embodiment). B The indicator data generally changes with CODtot, but due to the addition of sludge property analysis features and constraints, the problem of fractionation error transferring to inert components is further avoided. It fully considers the biochemical degradation characteristics of wastewater and the characteristics of sludge formed under long-term wastewater input, thus combining X... B The accuracy of fractionation of the indicators (and other components) and their sensitivity to changes in operating conditions, based on more accurate results, will enable the construction of more precise process control inputs in the future, further demonstrating the advantages of the scheme described in this embodiment.

[0120] In summary, this embodiment presents a method, system, equipment, and medium for dynamic estimation of wastewater COD components based on sludge properties. This method utilizes a small number of standard COD fractionation tests to construct a contextual wastewater fingerprint database, combined with online COD and BOD instruments to achieve preliminary estimation of high-frequency COD components. Furthermore, it uses the ratio of biodegradable to inert COD in the sludge as a long-term constraint, implementing optimization and error redistribution under mass conservation conditions within a time window matched to the sludge age. This embodiment upgrades traditional static COD fractionation parameters to a parameter system that can be dynamically adjusted with time and operating conditions without significantly increasing workload. This effectively avoids the problem of long-term error accumulation in inert particulate components in traditional methods, significantly improving the physical rationality, robustness, and responsiveness to changes in operating conditions of the COD component estimation results. It has the following beneficial effects:

[0121] (1) Multi-timescale collaboration, taking into account both representativeness and dynamism: By integrating low-frequency standard COD fractionation test, medium-frequency sludge property test and high-frequency online COD and BOD monitoring, a dynamic estimation framework for COD components that can be operated for a long time was constructed, achieving a balance between timeliness and representativeness of water quality characterization under limited experimental load.

[0122] (2) Contextual wastewater fingerprint database supports scenario-aware initial estimation: Based on the working conditions such as rainy season / dry season, weekday / weekend, and primary sedimentation tank operation status, a wastewater fingerprint database is constructed. During the online stage, it automatically identifies and matches the real-time operating scenario. Through interpolation, the initial estimation of COD components is dynamically adjusted, overcoming the problem that traditional static parameters are difficult to adapt to water quality fluctuations.

[0123] (3) Using the properties of activated sludge as a long-term constraint to enhance the rationality of the results: The ratio of biodegradable COD to inert COD in activated sludge samples is used to reflect the cumulative effect of historical influent load. This ratio is used as one of the constraints for estimating COD components within a time window, so that the estimation results are not only based on instantaneous water quality, but also in line with the long-term operation process, thus improving the physical rationality.

[0124] (4) Time window optimization and error redistribution to alleviate the accumulation of errors in inert components: Within the time window corresponding to the sludge age, the initial COD component estimation sequence is optimized and the error is redistributed by combining the online analytical relationship between COD and BOD, sludge property constraints and mass conservation conditions, so as to avoid unexplained errors being simply attributed to inert particle COD components, thereby improving numerical stability and sensitivity to changes in operating conditions.

[0125] (5) Highly compatible with activated sludge mathematical models: The output dynamic COD component results can be directly used as the input boundary conditions of activated sludge mathematical models such as the ASM series, which significantly improves the accuracy and reliability of the model in engineering applications such as medium and long-term simulation, energy consumption optimization, sludge reduction and emission compliance control.

[0126] Since the systems / devices described in the above embodiments of the present invention are systems / devices used to implement the methods of the above embodiments of the present invention, those skilled in the art can understand the specific structure and modifications of the systems / devices based on the methods described in the above embodiments of the present invention, and therefore will not be repeated here. All systems / devices used in the methods of the above embodiments of the present invention fall within the scope of protection of the present invention.

[0127] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0128] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions.

[0129] It should be noted that in the description of this invention, the word "a" or "an" preceding a component does not exclude the existence of multiple such components. This invention can be implemented by means of hardware comprising several different components and by means of a suitably programmed computer. The use of terms such as first, second, third, etc., is merely for convenience and does not indicate any order. These terms can be understood as part of the component names.

[0130] Furthermore, it should be noted that in the description of this specification, the terms "one embodiment," "some embodiments," "embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Furthermore, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0131] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning of the basic inventive concept, can make other changes and modifications to these embodiments.

[0132] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from the spirit and scope of the invention.

Claims

1. A method for dynamically estimating the COD components of wastewater based on sludge properties, characterized in that, include: The organic pollution index data monitored by graded filtration during the wastewater treatment process are correlated and fused with the process operation data to generate a context feature vector characterizing the dynamic changes of organic pollution indexes in the wastewater to be tested. The baseline proportion vector of COD components in wastewater obtained from COD fractionation experimental data is paired and fused with the context feature vector to generate an initial time series of COD components of the wastewater under test changing over time. Based on the sludge property analysis data of sludge samples in the aerobic tank, the ratio of biodegradable COD to total COD in the sludge samples was obtained, and this ratio was used as a constraint parameter for the sludge properties of the wastewater to be tested. Based on the sludge property constraint parameters and combined with the analytical relationship constraints of the wastewater to be tested in the wastewater treatment process, the initial COD component time series is constrained and optimized and the error is redistributed to obtain the dynamic estimation results of the COD components of the wastewater to be tested.

2. The method as described in claim 1, characterized in that, By correlating and fusing organic pollution index data monitored through graded filtration during wastewater treatment with process operation data, a contextual feature vector characterizing the dynamic changes of organic pollution indexes in the wastewater under test is generated, including: Acquire the graded filtration monitoring data of the wastewater to be tested in a time series within a set period. The graded filtration monitoring data includes organic pollution index data and corresponding process operation data. The organic pollution index data and process operation data are time-stamped and standardized to form a standardized time series data table. Based on the time series data table, multi-level feature construction is performed on the data at each time point to generate basic feature variables to characterize the state of organic pollution in wastewater, derived feature variables to characterize the influence of external conditions on wastewater treatment, and time series feature variables to characterize the dynamic change trend of organic pollution indicators. The basic feature variables, derived feature variables, and time-series feature variables at each time point are concatenated according to a predetermined feature order to generate a multi-dimensional context feature vector. The organic pollution index data include: particulate CODgf and BODgf obtained through the first-stage filtration, and dissolved CODmf and BODmf obtained through the second-stage filtration.

3. The method of claim 2, wherein, Based on time series data tables, multi-level feature construction is performed on the data at each time point to generate basic feature variables characterizing the state of organic pollution in wastewater, derived feature variables characterizing the influence of external conditions on wastewater treatment, and time-series feature variables characterizing the dynamic changing trends of organic pollution indicators, including: Numerical extraction of organic pollution index data at each time point in the time series data table yields basic characteristic variables representing the current organic pollution status of wastewater. Process operation data are used as environmental and load factors and coupled with basic characteristic variables to generate derived characteristic variables that characterize the impact of external conditions on wastewater treatment. A sliding window analysis was performed on the organic pollution index data along the time dimension to extract statistical features including mean, standard deviation, and rate of change, thus representing time-series characteristic variables that characterize the dynamic changing trend of organic pollution indicators.

4. The method of claim 1, wherein, Before pairing and fusing the baseline proportion vector of COD components in wastewater obtained from COD fractionation experimental data with the context feature vector to generate a preliminary time series estimate of COD components in the wastewater under test over time, the following steps are also included: Based on COD fractionation test data of wastewater samples under different operating conditions, a multi-component benchmark database of wastewater COD corresponding to different operating conditions was established. Based on the wastewater COD multi-component benchmark database, the detailed chemical components of COD under various operating conditions were determined, including: readily biodegradable volatile fatty acids S. VFA Easily biodegradable sugars S F Slow-degradable organic matter S U Colloidal biodegradable organic matter C B Particulate biodegradable organic matter X B and recalcitrant organic matter X U And constitute the reference proportion vector F of influent COD components under the current operating conditions. base =[S VFA ,S F ,S U C B ,X B ,X U ]; The identification information of different operating conditions is associated with the corresponding COD component benchmark ratio vector to generate a contextual wastewater fingerprint database, in which the identification information includes influent time period, season, process stage and water purification status.

5. The method of claim 1, wherein, The baseline proportion vector of COD components in wastewater obtained from COD fractionation experimental data is paired and fused with the context feature vector to generate a preliminary time series estimate of COD components in the wastewater over time, including: From the pre-set contextual wastewater fingerprint database, retrieve the historical benchmark proportion vector that matches the working conditions represented by the contextual feature vector. The historical benchmark proportion vector is obtained by performing component analysis and statistics on COD fractionation experimental data of similar wastewater. Based on the organic pollution index data of the wastewater to be tested at the current time, the content distribution of various forms of organic matter is analyzed, and the historical baseline proportion vector is dynamically corrected based on the analysis results to obtain the optimal baseline proportion vector that is suitable for the current working conditions. The optimal baseline scaling vector and the context feature vector are fused at the feature level to generate the initial estimate of COD components at the current sampling time. The context feature vectors and corresponding organic pollution index data at each time point are traversed sequentially. The retrieval, correction and fusion steps are repeated to generate a preliminary time series of COD components of the wastewater to be tested that changes continuously over time.

6. The method of claim 1, wherein, Based on the sludge property analysis data of sludge samples in the aerobic tank, the ratio of biodegradable COD to total COD in the sludge samples was obtained, and this ratio was used as a constraint parameter for the sludge properties of the wastewater to be tested, including: Sludge samples were collected from the aerobic tank at a set sampling period, and the total COD of the sludge samples was measured by a preset COD detection device. After performing aeration biodegradation on the sludge sample, the remaining COD in the sludge sample is measured by a COD detection device, and the remaining COD is subtracted from the total COD to obtain the COD of aeration degradation. The ratio of COD from aeration degradation to total COD is determined as a sludge property constraint parameter characterizing the biochemical degradation properties of the wastewater under test.

7. The method as described in claim 2, characterized in that, Based on sludge property constraint parameters and combined with the analytical relationship constraints of the wastewater in the wastewater treatment process, the initial COD component time series is optimized and the error is redistributed to obtain the dynamic estimation results of the COD components of the wastewater, including: Within the set sliding time window, the initial estimated time series of COD components is used as the initial solution for optimization iteration, and an optimization model including mass conservation constraints, analytical relation constraints, and sludge property constraints is established. With the objective of minimizing the overall estimation bias of the initial COD component time series, the optimization model is solved to obtain the optimal COD component time series. During the optimization process, based on the sensitivity of each component to changes in the graded filtration monitoring, the estimation error of the inert component in the initial COD component time series is allocated to a predefined transition component group between the biodegradable component and the inert component. The optimal COD component time series obtained by optimization is output as the dynamic estimation result of COD components of the wastewater to be tested. Among them, the mass conservation constraint is the sum of the estimated values ​​of all COD components at each time point, which is consistent with the organic pollution index data at that time point; the analytical relationship constraint is based on the organic pollution index data obtained from graded filtration monitoring, which constrains the analytical quantitative relationship between each component, including the estimated value of colloidal COD equals the difference between the BOD after primary filtration and the BOD after secondary filtration, the estimated value of soluble inert COD equals the difference between the COD after secondary filtration and the BOD after secondary filtration, and the estimated value of particulate COD equals the difference between the total COD and the COD after primary filtration.

8. A dynamic estimation system for wastewater COD components based on sludge properties, characterized in that, include: The online monitoring module is used to correlate and fuse the organic pollution index data monitored by graded filtration during the wastewater treatment process with the process operation data to generate a context feature vector characterizing the dynamic changes of organic pollution indexes in the wastewater under test. The component estimation module is used to pair and fuse the baseline proportion vector of COD components in wastewater obtained based on COD fractionation experimental data with the context feature vector to generate a preliminary time series of COD components of the wastewater under test changing over time. The sludge property constraint module is used to obtain the ratio of biodegradable COD to total COD in the sludge sample based on the sludge property analysis data of the sludge sample in the aerobic tank, and use this ratio as the sludge property constraint parameter of the wastewater to be tested. The constraint optimization and error redistribution module is used to perform constraint optimization and error redistribution processing on the initial COD component time series based on sludge property constraint parameters and combined with the analytical relationship constraints of the wastewater to be tested in the wastewater treatment process, so as to obtain the dynamic estimation results of COD components of the wastewater to be tested.

9. A dynamic estimation device for wastewater COD components based on sludge properties, characterized in that, include: At least one processor; and memory that is communicatively connected to at least one processor; The memory stores instructions that can be executed by at least one processor, which enables the at least one processor to perform a dynamic estimation method for wastewater COD components based on sludge properties as described in any one of claims 1-7.

10. A computer-readable storage medium storing computer-executable instructions thereon, characterized in that, When executed by a processor, the computer-executable instructions implement the method for dynamically estimating the COD components of wastewater based on sludge properties as described in any one of claims 1-7.

Images