Intelligent sewage plant whole-process management and control method based on digital twinning

Abstract

The invention relates to the technical field of sewage treatment optimization, in particular to an intelligent sewage plant full-process management and control method based on digital twinning, which comprises the following steps: acquiring concentration data to calculate rate clustering fluctuation, setting a threshold value mark stable section to analyze sedimentation association, constructing dynamic threshold value output control information, and adjusting stirring to generate a sludge discharge instruction set. According to the method, through the time sequence clustering of the concentration change rate, the stable section in the sludge sedimentation process is identified, the concentration fluctuation identification precision is improved, the sedimentation hysteresis is judged based on the time difference between the sedimentation track and the standard curve, the particle layering identification capability is improved, and the dynamic threshold matrix is constructed by using the correlation degree and the particle size data; the concentration range, the settling velocity deviation and the particle size distribution dispersion are fused, the multidimensional property of a regulation and control basis is enhanced, the interface position movement trend is combined, the concentration range is matched, a stirring intervention instruction is output, real-time regulation and accurate control of the layered settling process are achieved, and the treatment stability and the reaction time efficiency are improved.

Description

Technical Field

[0001] This invention relates to the field of wastewater treatment optimization technology, and in particular to a method for full-process control of intelligent wastewater treatment plants based on digital twins. Background Technology

[0002] The field of wastewater treatment optimization technology encompasses multiple stages, including wastewater collection, transmission, treatment, and discharge. It aims to improve wastewater treatment efficiency, reduce energy consumption and operating costs, and ensure stable effluent quality that meets standards. Core components include dynamic monitoring of wastewater quality, water quantity regulation, reaction process control, sludge treatment, and resource utilization. It leverages interdisciplinary approaches such as environmental engineering, automation control, information technology, and process optimization to achieve refined management of each treatment stage. Currently, it is gradually developing towards intelligent and automated systems. By constructing an interactive relationship between physical systems and virtual models, it achieves precise perception, scientific prediction, and efficient scheduling of wastewater treatment systems.

[0003] Among them, the intelligent wastewater treatment plant full-process control method based on digital twins refers to taking the wastewater treatment process as the object, continuously collecting relevant information such as influent water quality, water quantity distribution, pollutant concentration, aeration intensity, sludge content, equipment operating parameters and energy consumption data, and combining it with the physical characteristics and operating rules of each link in the wastewater treatment process to conduct process simulation and state identification, thereby forming a data foundation that can be used for control reference. It mainly obtains real-time data through online monitoring equipment, conducts pattern analysis using historical operating data, identifies the changing trends of key treatment links, and completes the dynamic judgment and control command generation of the wastewater treatment process through mathematical derivation and control strategy combination.

[0004] While existing technologies cover the entire wastewater treatment process and rely on online monitoring, model simulation, and control strategy combinations to achieve comprehensive evaluation of system operation, they suffer from response delays in identifying dynamic characteristics. Due to the nonlinear changes and multi-level distribution characteristics of sludge behavior, relying solely on overall concentration or traditional indicators is insufficient to accurately capture the changing trends at key levels, easily leading to the failure to identify some abnormal states in a timely manner. During sedimentation tank operation, changes in particle settling velocity may cause settling lag due to abrupt changes in particle size distribution. If these changes are not accurately identified and intervened in a timely manner, they can easily cause phenomena such as interface floating and fluctuations in effluent quality, affecting effluent stability. In existing technologies, intervention decisions often rely on human experience or judgment patterns based on single indicators, lacking the ability to integrate multi-dimensional information. This limits the accuracy of response to abnormal states and the ability to adapt and regulate the intensity of intervention. Although existing treatment processes have a certain degree of intelligence, they have not yet formed a truly efficient closed-loop feedback control system. Insufficient linkage between treatment links results in time differences and adjustment lags between information response and operation execution, affecting the overall system's operational stability and energy efficiency. Summary of the Invention

[0005] To address the technical problems existing in the prior art, this invention provides a method for the full-process management and control of a smart wastewater treatment plant based on digital twins. The technical solution is as follows: To achieve the above objectives, the present invention adopts the following technical solution: a smart wastewater treatment plant full-process management and control method based on digital twins, comprising the following steps: S1: Acquire time-series data of multi-layer deep sludge concentration through distributed sensors, calculate the concentration change rate within the time window, use K-means clustering algorithm to distinguish fluctuation characteristics, set a standard deviation threshold based on the variance of the concentration change rate within the sliding window, and mark the continuous time period with a standard deviation lower than the threshold in the clustering results as the stable segment identification result. S2: Based on the stable segment identification results, obtain the settling trajectory of the target sludge particles, use the grey relational analysis method to calculate the time difference of the sludge particles reaching the bottom monitoring point, compare it with the standard settling curve and analyze the correlation, and generate the lag judgment result. S3: Call the particle size data in the hysteresis determination result whose correlation is lower than the benchmark value obtained by Spearman's rank correlation coefficient, combine it with the average concentration in the current stable segment identification result, use the TOPSIS method to construct a dynamic threshold matrix including concentration range, settling velocity deviation and particle size distribution dispersion, and output control threshold information. S4: Based on the control threshold information, extract the hierarchical data where the concentration range exceeds the safety threshold, match the interface position movement vector collected by the online turbidity meter, adjust the stirring intensity level and duration, and output the stirring intervention command.

[0006] As a further aspect of the present invention, the stable segment identification result includes identification time period number, depth level distribution, and fluctuation classification label; the hysteresis determination result includes particle size grouping identifier, response delay value, and offset coefficient level; the control threshold information includes concentration interval level, particle size adjustment weight, and settling velocity deviation factor; and the stirring intervention command includes stirring level number, stirring duration, and target depth identifier.

[0007] As a further aspect of the present invention, step S1 specifically comprises: S101: The time series data of multi-layer deep sludge concentration is acquired through distributed sensors. The continuous concentration value sequence of each layer within a fixed time window is extracted. The concentration change rate is calculated based on the adjacent sampling time points of each concentration sequence. After standardizing the sequence samples of the concentration change rate of the deep layer, the data is input into the K-means clustering algorithm to obtain the category label of the sample point in the cluster and obtain the fluctuation characteristic category distribution value. The K-means clustering algorithm is an unsupervised learning method that divides standardized concentration change rate sequences into multiple categories based on similarity. S102: Based on the fluctuation feature category distribution value, perform sliding window division on the corresponding concentration change rate sequence in each sample group, call the variance of the rate sample in each window to calculate, and obtain the mean of the window variance value as the mean threshold of the concentration change rate variance. S103: Call the distribution value of the fluctuation feature category and the threshold of the mean variance of the concentration change rate, compare the standard deviation of each sample group, filter the sample segments with a standard deviation less than the threshold and stable within three consecutive time windows, and generate stable segment identification results.

[0008] As a further aspect of the present invention, step S2 specifically includes: S201: Based on the stable segment identification results, obtain the concentration sequence and sampling time data of the depth position within the stable time period, identify the change of the settling position of sludge particles in the vertical direction, and record the time node when the particles reach the bottom monitoring point by combining the particle size distribution index and the monitoring point depth value, and generate a time-consuming dataset corresponding to the particle size. S202: Based on the time-consuming dataset corresponding to the particle size, call the theoretical sedimentation time series under the same particle size distribution in the standard sedimentation curve, establish a comparison vector between the actual sedimentation time of each group of particle sizes and the corresponding time nodes of the standard sequence, use the grey relational analysis method to calculate the correlation value between the particle size sedimentation sequence and the standard sequence, and obtain a list of particle size sedimentation correlation. The standard settling curve refers to the particle size-time mapping reference curve formed by experimentally determining the theoretical settling time of particles of different sizes within a unit depth under ideal static water conditions. S203: Based on the particle size sedimentation correlation list, combined with particle size time data and number index, filter the particle size groups with correlation values ​​lower than the lag judgment threshold, record the time difference range and number mapping structure of the corresponding particle size, and establish the lag judgment result. The hysteresis determination threshold is the average grey relational value of particle size samples that have the same trend as the standard sedimentation curve within the current period.

[0009] As a further aspect of the present invention, step S3 specifically comprises: S301: Call the particle size samples with trajectory correlation degree lower than the preset threshold in the hysteresis determination results, calculate the correlation between particle size and sedimentation response time difference based on Spearman's rank correlation coefficient, identify particle size segments with correlation coefficient lower than the statistical threshold, extract the corresponding sedimentation response time difference and particle size distribution index, and combine the number proportion and distribution spacing of the corresponding particle size segments to establish a related particle size distribution dataset. The preset threshold is a judgment standard set based on the similarity level between historical settlement behavior and standard trajectory, combined with empirical data. The statistical threshold is a decision boundary calculated using the distribution characteristics of the Spearman rank correlation coefficient on historical samples. S302: Call the average concentration value in the associated particle size distribution dataset and the stable segment identification result, extract the concentration range value and settling velocity change range value in the corresponding time window, calculate the particle size distribution dispersion in the settling zone, obtain the concentration range, settling velocity deviation, and particle size dispersion data, and generate a dynamic threshold evaluation index set. S303: Call the dynamic threshold evaluation index set, take the proximity and ranking consistency between each index as input variables, use the TOPSIS method to calculate the distance between multiple indices and rank them, calculate the concentration control factor score, extract the variable combination with the best score, and establish control threshold information.

[0010] As a further aspect of the present invention, the dynamic threshold evaluation index set is invoked, and the closeness and ranking consistency among each index are used as input variables. The specific formula for the ranking consistency is as follows: ; Calculate the ranking consistency index value; in, Representing the The first in the group The ranking consistency index value of the items. Representing the The first in the group The indicator in the first Sorting offset adjustment factor under variable Representing the The first in the group The indicator in the first The sorting position value under the variable, Representing the The first in the group The average of the ranking positions of the item indicator under all variables. Representing the The first in the group The absolute value of the variance of the ranking of each indicator under the variable. Representing the The first in the group The indicator is based on the empirical expected offset value of historical ranking stability. Representing the The total number of variables in the group that are involved in the sorting. For index number, For variable index number, This is the group number.

[0011] As a further aspect of the present invention, step S4 specifically comprises: S401: Based on the control threshold information, extract the hierarchical data where the concentration range exceeds the safety threshold, and identify the corresponding spatial number and range value to obtain the concentration threshold abnormal hierarchical set; The safety threshold is set based on the combination of main control factor variables, and is mapped to the concentration range of each level through the TOPSIS sorting results. S402: Call the spatial number data in the concentration threshold anomaly level set, match the interface position movement vector collected in real time by the online turbidity meter, calculate the offset distance and velocity change rate of the interface corresponding to the spatial number in the sampling period through time series comparison of vector position, and generate interface offset trend parameter group. S403: Based on the offset speed and distance values ​​of each item in the interface offset trend parameter group, according to the classification criteria of speed range and displacement amplitude based on the stirring intensity level, obtain the matching stirring level, and output the stirring intervention instruction set in combination with the duration of action.

[0012] As a further aspect of the present invention, the method further includes step S5: S5: Based on the execution feedback data of the stirring intervention command, obtain the changes in the operating frequency of the sludge discharge pump and the sludge return ratio after stirring, use the fuzzy comprehensive evaluation method to evaluate the adjustment range and response effect, construct the operation status level classification result, and output a sludge discharge control command set including multiple adjustment parameters. The sludge discharge control instruction set includes sludge discharge rate levels, control frequency levels, and response adjustment ranges.

[0013] As a further aspect of the present invention, step S5 specifically includes: S501: Based on the execution feedback data of the stirring intervention command, obtain the operating frequency data of the periodic sludge discharge pump after stirring and the monitoring data of the sludge return ratio, calculate the frequency variation and return ratio variation amplitude within the period in chronological order, and establish a set of periodic operation characteristic data. S502: Call the frequency change rate and reflux ratio change rate in the periodic operation feature data set, set the membership function of frequency change and reflux ratio change, divide the fuzzy level interval, use the fuzzy comprehensive evaluation method to evaluate the adjustment amplitude and response effect, calculate the comprehensive membership distribution result, and obtain the fuzzy adjustment level factor. The membership function is used to map the frequency and reflux ratio changes to their degree of belonging to different regulation state levels; The fuzzy level intervals are defined by setting a division standard to classify variable values ​​into multiple fuzzy level ranges, including stable ranges and fluctuating ranges. The fuzzy comprehensive evaluation method synthesizes the membership degree and weight of indicators at different levels to comprehensively evaluate the overall performance of the adjustment state. The comprehensive membership distribution result refers to the set of membership values ​​obtained by fusing the membership of multiple input variables at each fuzzy level; S503: Based on the membership level results in the fuzzy adjustment level factor, and combined with the current cycle sludge discharge control instruction template, the sludge discharge control instruction is matched by filtering the multi-level parameter combinations corresponding to the level, and the sludge discharge control instruction set is obtained.

[0014] As a further aspect of the present invention, the specific formula for the membership function of setting the frequency variation and the return current ratio variation is as follows: ; Calculate the fuzzy response coordination parameters; in, Let i be the fuzzy response coordination parameter. Let i be the rate of change of the reflux ratio. Let i be the rate of change of frequency of the i-th term. This is the frequency-dependent amplification factor for adjusting the return current ratio. Let be the subjective weight value corresponding to the i-th fuzzy level. Let i be the fuzzy attribution trust factor. Let i be the feedback response intensity coefficient. Let j be the adjustment intensity conversion factor. Let j be the system damping response ratio. For fuzzy evaluation sample numbers, For the index number of the adjustment item, This represents the total number of samples for the rate of change of frequency and the rate of change of reflux ratio.

[0015] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: By acquiring time-series data on the concentration variations of multi-layered, deep sludge, and analyzing concentration fluctuation characteristics in real time to identify stable periods, the accuracy of perception of sludge dynamic behavior is effectively enhanced. The variance of the concentration change rate is used as a key criterion for clustering judgment, improving the objectivity and adaptability of stability identification. Based on this, by analyzing the correlation between the settling trajectory of sludge particles and the standard settling curve, the lag in settling efficiency is further identified, significantly improving the ability to identify particle size settling velocity deviations. This type of settling performance analysis introduces a dynamic comparison between particle size data and settling behavior, effectively improving the early warning capability for abnormal states in the reaction process. Furthermore, a dynamic threshold matrix is ​​constructed using correlation and concentration, taking into account multi-dimensional information such as settling rate, concentration changes, and particle size distribution, providing a comprehensive judgment basis for regulation and enhancing the accuracy of intervention decisions. Finally, by matching the concentration range and interface displacement trend, the stirring intensity and duration are dynamically adjusted to achieve precise intervention operations, significantly improving the refinement and response speed of sludge stratification control. The overall processing logic embodies a continuous feedback closed loop from data acquisition, status identification, anomaly judgment, threshold generation to instruction output. It integrates time-series data analysis, multi-index modeling, and action adjustment optimization to achieve real-time control and efficiency improvement of sedimentation behavior during wastewater treatment, thereby enhancing the robustness and operational adaptability of the treatment process. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the workflow of the present invention. Detailed Implementation

[0018] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0019] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0020] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0021] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0022] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0023] Please see Figure 1 This invention provides a technical solution: a method for full-process management and control of a smart wastewater treatment plant based on digital twins, comprising the following steps: S1: Acquire time-series data of multi-layer deep sludge concentration through distributed sensors, calculate the concentration change rate within the time window, use K-means clustering algorithm to distinguish fluctuation characteristics, set a standard deviation threshold based on the variance of the concentration change rate within the sliding window, and mark the continuous time period with a standard deviation lower than the threshold in the clustering results as the stable segment identification result. S2: Based on the stable segment identification results, the settling trajectory of the target sludge particles is obtained. The grey relational analysis method is used to calculate the time difference of the sludge particles reaching the bottom monitoring point. The result is compared with the standard settling curve and the correlation is analyzed to generate the lag judgment result. S3: Call the particle size data in the hysteresis judgment result whose correlation is lower than the benchmark value obtained by Spearman's rank correlation coefficient, combine it with the average concentration in the current stable segment identification result, and use the TOPSIS method to construct a dynamic threshold matrix including concentration range, settling velocity deviation and particle size distribution dispersion, and output control threshold information. S4: Based on the control threshold information, extract the hierarchical data where the concentration range exceeds the safety threshold, match the interface position movement vector collected by the online turbidity meter, adjust the stirring intensity level and duration, and output the stirring intervention command; S5: Based on the execution feedback data of the stirring intervention command, obtain the changes in the operating frequency of the sludge discharge pump and the sludge return ratio after stirring, use the fuzzy comprehensive evaluation method to evaluate the adjustment range and response effect, construct the operation status level classification result, and output a sludge discharge control command set including multiple adjustment parameters.

[0024] The stable segment identification results include the identification time period number, depth level distribution, and fluctuation classification label. The hysteresis determination results include particle size grouping identifier, response delay value, and offset coefficient level. The control threshold information includes concentration range level, particle size adjustment weight, and settling velocity deviation factor. The stirring intervention command includes stirring level number, stirring duration, and target depth identifier. The fuzzy comprehensive evaluation method synthesizes the membership degree and weight of the indicators at different levels to comprehensively evaluate the overall performance of the regulation state.

[0025] Please see Figure 1 The specific steps of S1 are as follows: S101: The time series data of multi-layer deep sludge concentration is acquired through distributed sensors. The continuous concentration value sequence of each layer within a fixed time window is extracted. The concentration change rate is calculated based on the adjacent sampling time points of each concentration sequence. After standardizing the sequence samples of the concentration change rate of the deep layer, the data is input into the K-means clustering algorithm to obtain the category label of the sample point in the cluster and obtain the fluctuation characteristic category distribution value. Multi-layer depth sludge concentration time-series data were acquired using distributed sensors. One sensor node was installed every 0.5 meters in the wastewater treatment tank, with a depth range of 0.5 to 2.0 meters, forming a total of four monitoring nodes. The sampling interval between each node was 2 minutes, and a total of 15 concentration value sequences were collected every 30 minutes. After acquiring the concentration values, a concentration value sequence array was constructed for each depth node. The difference between any two adjacent values ​​in the execution sequence. The calculation operation is performed, and the difference is divided by the corresponding time interval of 2 minutes to obtain the concentration change rate within each interval. For example, if the concentration of a node at a depth of 1.0 meter is 5.0 mg / L at the first time point and 5.3 mg / L at the second time point, then the rate is... mg / L / min; the concentration change rate sequence was obtained sequentially. Calculate the average value of the rate sequence. and standard deviation Then perform Z-score normalization, applying it to each rate. For example, a certain sequence is Its mean is 0.1 and its standard deviation is 0.1581. Therefore, the standardized value of the first term is... The final standardized matrix is ​​input into K-means clustering calculation, and the number of clusters is set. The Euclidean distance between the standardized rate sequence and the three initial cluster centers was calculated, assuming the sample sequence... , with a certain center Distance is By comparing the distances of the sample to the three centers, the category corresponding to the smallest distance is selected, and finally each time window sample obtains a category label, forming a fluctuating feature category distribution; K-means clustering is an unsupervised learning method that divides standardized concentration change rate sequences into multiple categories based on similarity.

[0026] S102: Based on the distribution value of the fluctuation feature category, perform sliding window division on the corresponding concentration change rate sequence in each sample group, call the variance of the rate sample in each window to calculate, and obtain the mean of the window variance value as the threshold of the mean variance of the concentration change rate. Based on the obtained fluctuation characteristic category distribution values, a sliding window operation is performed on the sample rate sequences belonging to the same category. Each window width is set to 5 rate points, meaning each concentration change rate sequence has a length of 14. The sliding window starts from point 1 and slides one position forward each time, forming a total of 10 windows. Five rate sample values ​​are extracted from each window. The variance of the samples in that window is calculated by taking the mean of the five values ​​in that window, squared the differences between each value and the mean, summing these sums, and then dividing by the number of samples. For example, if a window value is... The mean is 0.12 and the variance is... After calculating the variance values ​​for each of the 10 windows, the mean value was taken to obtain the threshold value for the mean variance of the concentration change rate. The variance values ​​are as follows: The threshold is The threshold serves as the basis for standard judgment and participates in the subsequent judgment of the stability of sample segments.

[0027] S103: Call the distribution value of the fluctuation feature category and the threshold of the mean variance of the concentration change rate, compare the standard deviation of each sample group, filter the sample segments with a standard deviation less than the threshold and stable within three consecutive time windows, and generate stable segment identification results. Using the aforementioned fluctuation characteristic category distribution values ​​and the mean threshold of the variance of the concentration change rate, the standard deviation of the rate of change of all samples in each sample group is taken, and the standard deviation is calculated one by one. ,in Taking the actual sample rate sequence as an example The mean is 0.135 and the standard deviation is... The standard deviation is compared with the calculated mean variance threshold of 0.00235. If the standard deviation is less than the threshold, the sequence is judged as a candidate stable segment. Further checks are made to see if the standard deviation remains less than the threshold in three consecutive time windows. That is, three adjacent time windows (such as the 13th window and the 24th window) need to be slid. If the standard deviation is less than the threshold in all three windows, the segment is marked as a stable segment. Finally, a stable segment identification result is generated. For example, if the standard deviation of sample segment [10~24] is [0.0021, 0.0022, 0.0023] in three time windows, which is less than the threshold of 0.00235, the judgment condition is met, and the segment is output as a stable segment.

[0028] Please see Figure 1The specific steps of S2 are as follows: S201: Based on the stable segment identification results, obtain the concentration sequence and sampling time data of the depth position within the stable time period, identify the change of the settling position of sludge particles in the vertical direction, and combine the particle size distribution index and the depth value of the monitoring point to record the time node when the particles reach the bottom monitoring point and generate a time dataset corresponding to the particle size. Based on the stable segment identification results, the collected sedimentation monitoring data was first segmented to identify multiple stable segments in the sedimentation process. Each stable segment needed to meet the conditions that the pressure fluctuation at the monitoring point was less than 0.2 kPa and the concentration change was less than 5 mg / L within a continuous time window. After screening, the corresponding depth location and concentration value sequence were extracted within each stable segment. Data at sampling time points were extracted in 5-second increments, and the concentration sequences were compared longitudinally. The comparison results were used to determine the sedimentation trend by differentiating the data at each depth point. In this process, a particle size distribution index was incorporated, for example, for a particle size of 125 μm. Three groups of particles of different sizes (125μm, 150μm, and 200μm) were compared, and the time when their concentrations appeared at each depth point was compared. The time when the concentration of a certain particle size first increased significantly at the bottom monitoring point (e.g., at a depth of 1.0m) (e.g., the concentration increased by more than 20% of the initial baseline value) was recorded as the time node when the particle of that size reached the bottom. This operation was performed cyclically for each stable segment to obtain a set of "particle size-time" data items. For example, it took 85s for 125μm particles to arrive, 112s for 150μm particles, and 154s for 200μm particles. The obtained data constituted the time dataset corresponding to the particle size.

[0029] S202: Based on the time-consuming dataset corresponding to particle size, call the theoretical sedimentation time series under the same particle size distribution in the standard sedimentation curve, establish a comparison vector between the actual sedimentation time of each group of particle sizes and the corresponding time nodes of the standard sequence, use the grey relational analysis method to calculate the correlation value between the particle size sedimentation sequence and the standard sequence, and obtain the particle size sedimentation correlation list. Based on the time-based dataset corresponding to particle size, the actual recorded time sequences are compared with the theoretical time sequences under the standard settling curve. The standard settling time for each particle size group needs to be set based on the results of laboratory static water column tests. For example, the theoretical settling times are set to 80s for 125μm, 110s for 150μm, and 150s for 200μm. By constructing a comparison vector, the actual and theoretical settling times for each pair of particle sizes are paired according to the particle size number to generate a difference vector. The difference vector is as follows: Δt=[5s,2s,4s] Then, based on each group of differences, a correlation assessment operation is performed. The assessment process involves calculating the absolute difference between the actual sequence and the theoretical sequence, and then normalizing them to a standard score. The normalization method uses an interval mapping of [0,1], and the scoring benchmark is set to a difference interval of ±10s. For example, when the difference is 5s, the normalization value is 0.5. Then, the correlation coefficient between the actual and theoretical sequences is calculated, and the discrimination precision is set to 0.1 units. The normalized values ​​are substituted into the following process one by one: The correlation coefficient calculation logic is to calculate the correlation coefficient for each particle size difference. Perform the following operations: 1. 2. Compare with all other particle size differences to calculate its relative shift in the total sequence, and normalize to For example, if the difference is 5 seconds, then Then put all Combined to form a correlation coefficient vector, 3. Output a list of particle size sedimentation correlation degrees, such as: [0.1667, 0.3333, 0.2], corresponding to 125μm, 150μm, and 200μm respectively; A standard settling curve is a particle size-time mapping baseline curve formed by experimentally determining the theoretical settling time of particles of different sizes within a unit depth under ideal static water conditions.

[0030] S203: Based on the particle size sedimentation correlation list, combined with particle size time data and number index, filter the particle size groups with correlation values ​​lower than the lag judgment threshold, record the time difference range and number mapping structure of the corresponding particle size, and establish the lag judgment result. According to the particle size sedimentation correlation list, the lag judgment threshold is set to 0.25. If the correlation coefficient of a certain particle size is lower than the threshold, it is judged as a sedimentation lag particle group. For example, in the previous results, 125μm (0.1667) and 200μm (0.2) meet the conditions and are judged as lag groups. Then, the time difference interval between these particle sizes and their actual consumption time and theoretical consumption time is extracted, and a mapping structure between the number and the time difference is constructed. For example, number 1 corresponds to 125μm, its actual consumption time is 85s, its theoretical consumption time is 80s, and the difference is 5s. Number 3 corresponds to 200μm, its actual consumption time is 154s, its theoretical consumption time is 150s, and the difference is 4s. It is recorded as the mapping structure {1:[5],3:[4]}, and finally the lag judgment result is formed. The particle size number, corresponding consumption time difference and particle size value are output as combined information. Table 1. Results of particle size sedimentation analysis

[0031] As shown in Table 1, by comparing the sedimentation correlation coefficients calculated for each particle size with the theoretical sequence, the hysteresis-determining particle group was screened out, providing a basis for subsequent sedimentation process control and particle group identification. The hysteresis threshold is the mean grey relational degree of particle size samples that have the same trend as the standard sedimentation curve within the current period.

[0032] Please see Figure 1 The specific steps of S3 are as follows: S301: Call the particle size samples with trajectory correlation degree lower than the preset threshold in the hysteresis judgment results, calculate the correlation between particle size and sedimentation response time difference based on Spearman's rank correlation coefficient, identify particle size segments with correlation coefficient lower than the statistical threshold, extract the corresponding sedimentation response time difference and particle size distribution index, and combine the number proportion and distribution spacing of the corresponding particle size segments to establish a related particle size distribution dataset. To retrieve particle size samples with a trajectory correlation coefficient below a preset threshold from the lag determination results, it is necessary to first read the data of particles already identified as lagging sedimentation particles in the previous processing flow. The trajectory correlation coefficient is defined as the degree of sequence consistency between the actual sedimentation path and the standard path for that particle size. If the correlation coefficient corresponding to the particle size number is below 0.25, its trajectory correlation is considered poor. The sedimentation time and corresponding concentration change trend curves of this type of particle size are then read, and a time difference sequence is constructed for each particle size group. This is the difference between the actual settling time and the theoretical settling time. For example, if the actual settling time for a particle with a diameter of 125μm is 85s and the theoretical settling time is 80s, then... The minimum trajectory correlation coefficient threshold is set to 0.25. This value is derived from the trajectory consistency of more than 70% of samples in historical data under normal settlement conditions. Its minimum distribution boundary is set as this threshold to filter atypical settlement path data. After obtaining particle size samples that meet the conditions, a particle size sequence needs to be constructed. Its corresponding time difference series Based on this, a rank transformation operation is performed, assigning rank values ​​to the data in the particle size sequence and time difference sequence in ascending order. For example, for a particle size sequence of [125, 150, 200] μm, the transformed rank is [1, 2, 3], and for a time difference sequence of [5, 2, 4] s, the transformed rank is [3, 1, 2]. Then, the squared differences between the corresponding rank values ​​in each group are summed to calculate the sum of squared rank differences. Substituting into the Spearman rank correlation coefficient formula: ,in, This represents the rank correlation coefficient between the difference in particle size and time consumption. This represents the number of particle size samples; in this example, it is 3. Correlation coefficient. The correlation is lower than the empirically set statistical threshold of 0.2. This statistical threshold is derived from the 25th percentile value of the rank correlation coefficient distribution calculated from a large number of historical samples. Therefore, a value lower than this is considered to be a weak correlation. After selecting these low-correlation particle size segments, the corresponding time difference data and particle size numbers are extracted. Based on the proportion of each particle size segment in the total sample (assuming 125μm accounts for 15%, 150μm accounts for 25%, and 200μm accounts for 60%) and the particle size segment spacing (the particle size spacing is uniformly 25μm), a correlated particle size distribution dataset is constructed. This dataset is used as one of the variable sets affecting the sedimentation state in the next stage of calculation. Table 2. Rank Value Analysis of Particle Size and Sedimentation Response Time Difference

[0033] As shown in Table 2, the sum of squared rank differences between particle size and response is 6. Substituting this into the formula, the correlation coefficient is -0.5, which is considered to be a significant negative correlation. This result is used to extract the associated particle size distribution dataset. The preset threshold is a judgment standard set based on the similarity level between historical settlement behavior and standard trajectory, combined with empirical data. The statistical threshold is the decision boundary calculated by the distribution characteristics of the Spearman rank correlation coefficient on historical samples.

[0034] S302: Call the average concentration value in the associated particle size distribution dataset and the stable segment identification results, extract the concentration range value and settling velocity change range value in the corresponding time window, calculate the particle size distribution dispersion in the settling zone, obtain the concentration range, settling velocity deviation, and particle size dispersion data, and generate a dynamic threshold evaluation index set. Using the average concentration values ​​extracted from the aforementioned associated particle size distribution dataset and stable segment identification results, it is necessary to identify the difference between the maximum and minimum concentration values ​​within the corresponding time window for each particle size segment, which is defined as the concentration range value. Simultaneously, the settling velocity variation value was extracted. The settling velocity variation was sampled from the velocity differences at different time points throughout the entire settling process for different particle size ranges, and the settling velocity deviation value was calculated. For example, the average concentration variation within a certain particle size range is [130, 155, 145] mg / L, with a concentration range of 25 mg / L; the settling velocity is 1.2 mm / s in the first stage and 0.8 mm / s in the second stage, with a settling velocity deviation of 0.4 mm / s; the particle size distribution dispersion is obtained based on the deviation of the particle size sample distribution within a set range, and is defined as the standard deviation. The sample particle size is [125, 150, 200] μm, the average value is 158.3 μm, and the dispersion is calculated as follows: , Subsequently, the concentration range, settling velocity deviation, and particle size dispersion were combined to form a dynamic threshold evaluation index set, which served as reference input parameters for subsequent system control and monitoring. This index set is summarized below by particle size segment number and numerical dimension: Table 3. Dynamic Threshold Evaluation Index Set

[0035] As shown in Table 3, a dynamic threshold evaluation index set was constructed by numerically aggregating data from three dimensions, providing input for subsequent particle size segment control of the system. The results indicate that the particle size sample exhibits characteristics such as significant concentration fluctuations, unstable sedimentation rates, and a wide distribution range, making it suitable as a representative of unstable sedimentation particle segments.

[0036] S303: Call the dynamic threshold evaluation index set, take the closeness and ranking consistency between each index as input variables, use the TOPSIS method to calculate the distance between multiple indices and rank them, calculate the concentration control factor score, extract the variable combination with the best score, and establish control threshold information. The dynamic threshold evaluation index set is invoked, with the degree of similarity and ranking consistency among each index as input variables. The specific formula for ranking consistency is as follows: ; Calculate the ranking consistency index value; in, Representing the The first in the group The ranking consistency index value of the items. Representing the The first in the group The indicator in the first Sorting offset adjustment factor under variable Representing the The first in the group The indicator in the first The sorting position value under the variable, Representing the The first in the group The average of the ranking positions of the item indicator under all variables. Representing the The first in the group The absolute value of the variance of the ranking of each indicator under the variable. Representing the The first in the group The indicator is based on the empirical expected offset value of historical ranking stability. Representing the The total number of variables in the group that are involved in the sorting. For index number, For variable index number, Group number; formula: ; Detailed explanation of the formula and its calculation derivation: The formula is used to calculate the ranking consistency index value γ of the a-th indicator in group c. a ^(c), this result is used to assess the consistency of the ranking of indicators under different variables; Parameter meanings and settings: The total number of variables involved in the sorting in group c is set to 5; The ranking position of indicator a in group c under variable b is obtained by ranking the performance of the indicators under each variable, and is set as follows: , , , , ; The average of the ranking positions of the a-th indicator in group c under all variables is calculated as follows: ; The absolute value of the variance of the ranking values ​​of each indicator under the b-th variable in group c is set as: , , , , ; The ranking offset adjustment factor for the a-th indicator in group c under variable b is set as follows: , , , , ; The expected offset value of the a-th indicator in group c, based on the historical ranking stability, is set to 0.5. Substitute the parameters into the formula to calculate: Calculate the absolute difference for each item: ; ; ; ; ; Calculate the score for each item: Item 1: ; Item 2: ; Item 3: ; Item 4: ; Item 5: ; Summing and calculating the average: ; Calculate the final Value: ; The result of 0.0409 indicates that the ranking consistency of indicator a in group c is relatively high under different variables; the closer the value is to 0, the stronger the consistency. This result is used to evaluate the stability of indicators in a multivariate environment, providing a basis for subsequent multi-indicator distance calculation and ranking scoring.

[0037] Please see Figure 1 The specific steps of S4 are as follows: S401: Based on the control threshold information, extract the hierarchical data where the concentration range exceeds the safety threshold, and identify the corresponding spatial number and range value to obtain the concentration threshold anomaly hierarchical set; Based on the control threshold information, the concentration range data of all particle size samples are first extracted from the established dynamic threshold evaluation index set. The range value corresponding to each particle size level is read one by one to determine whether it exceeds the set safety threshold. The safety threshold is set based on the comprehensive score ranking result formed by the combination of main control factor variables. The main control factors include the concentration range value, settling velocity deviation value, and particle size dispersion value, with weights of 0.3, 0.4, and 0.3 for each factor. The three factor values ​​for each particle size level are normalized, multiplied by their corresponding weights, and summed to obtain the comprehensive score. For example, if a particle size level is numbered Z3, its normalized concentration range is 0.6, settling velocity deviation is 0.5, and particle size dispersion is 0.8, then its score is... After sorting the scores of all levels, the concentration range of the level with the highest score is selected as the safety threshold. For example, if the highest score is 0.73 and the corresponding concentration range is 30 mg / L, then 30 mg / L is set as the current safety threshold. All particle size level data are traversed, and each range value is compared with 30 mg / L. For example, if the concentration range of Z1 is 28 mg / L, Z2 is 32 mg / L, and Z3 is 35 mg / L, then Z2 and Z3 are identified as abnormal levels because their values ​​exceed 30 mg / L. The corresponding spatial number and concentration range value are recorded, and the output is the concentration threshold abnormal level set. Table 4. Concentration Threshold Abnormal Level Judgment Table

[0038] As shown in Table 4, the concentration range of spatial numbers Z2 and Z3 both exceeded the safety threshold of 30 mg / L, and were marked as abnormal levels. The safety threshold is set based on the combination of main control factor variables, and is mapped to the concentration range of each level through the TOPSIS method ranking results.

[0039] S402: Call the spatial number data in the concentration threshold anomaly level set, match the interface position movement vector collected in real time by the online turbidity meter, calculate the offset distance and velocity change rate of the interface corresponding to the spatial number in the sampling period through time series comparison of vector position, and generate interface offset trend parameter group. The system retrieves the spatial number data already identified in the anomaly hierarchy set, such as Z2 and Z3, and reads the interface position movement vector from the corresponding turbidimeter's real-time monitoring record according to the number. The vector data contains interface height data point pairs at multiple time points. For example, record number Z2 is: Calculate the velocity for every two adjacent time points. ,like , Then calculate the rate of change of velocity. , and thus Meanwhile, the maximum offset distance is calculated as the difference between the initial height and the lowest height. For example, if the height of the Z2 interface drops from 90mm to 72mm, the offset distance is 18mm. The rate of change of speed and the offset distance are combined and organized into the interface offset trend parameter group. Table 5 Interface Offset Trend Parameter Table

[0040] As shown in Table 5, both Z2 and Z3 show a significant downward trend in interface movement. Their offset distance and rate of change of velocity are used to determine the stirring level in the next step.

[0041] S403: Based on the offset speed and distance values ​​of each item in the interface offset trend parameter group, and according to the classification criteria of speed range and displacement amplitude based on the stirring intensity level, obtain the matching stirring level, and output the stirring intervention instruction set in combination with the duration of action. To demonstrate the inherent coupling mechanism between dynamic threshold evaluation indicators and final stirring intervention decisions, a dynamic threshold matrix needs to be constructed before generating stirring intervention commands. This matrix uses particle size dispersion, concentration range, and settling velocity deviation as columns and spatial numbers as rows, forming a two-dimensional index set. For each row vector, the degree of matching with the grading threshold standard is calculated. The specific matching process involves defining interval judgment conditions for each index, such as particle size dispersion. If the dispersion is classified as low (0, 10) μm, medium (10, 25) μm, and high (≥25 μm); if the concentration range is <20 mg / L, it is considered low fluctuation; if it is [20, 30) μm, it is considered medium fluctuation; and if it is ≥30 mg / L, it is considered high fluctuation; if the settling velocity deviation is <0.2 mm / s, it is considered stable settling; if it is [0.2, 0.5) μm, it is considered medium fluctuation; and if it is ≥0.5 mm / s, it is considered violent fluctuation. By judging the range in which each indicator falls, the indicator combination is mapped to a preset stirring level template. For example, if the three indicators of a space number are in the ranges of high dispersion, high concentration range, and medium settling velocity deviation, respectively, it is judged to fall into the predefined Level III response group according to the weighted score. Then, a second verification is performed by combining the actual offset value and velocity change rate in the interface offset trend parameters. When the offset distance or velocity meets the corresponding stirring level conditions, a response is triggered. The record in that row of the dynamic threshold matrix is ​​mapped one by one with the generated stirring level. For example, the index numbered Z3 is concentration range = 35mg / L, settling velocity deviation = 0.3mm / s, and particle size dispersion = 26μm. Its dynamic matrix is ​​[high, medium, high], and the corresponding scoring combination is Level III stirring. After successful matching, the duration of action is output synchronously. The final output stirring intervention command is the response strategy to the dynamic index change behavior, indicating that the stirring behavior is directly controlled by the dynamic threshold three-factor combination characteristics. Table 6. Mapping Table of Dynamic Threshold Matrix and Stirring Level

[0042] As shown in Table 6, Z3 is matched to Level III stirring intervention level by the system according to the set rules because two of its dynamic threshold feature combinations are at the high fluctuation level. Z2 corresponds to Level II stirring because all three indicators are in the median range, forming a complete dynamic evaluation-decision response mapping link. This supplementary content verifies the control coordination and decision-making closed-loop consistency of dynamic parameters in the stirring response process.

[0043] Please see Figure 1 The specific steps of S5 are as follows: S501: Based on the execution feedback data of the stirring intervention command, obtain the operating frequency data of the periodic sludge discharge pump and the monitoring data of the sludge return ratio after stirring, calculate the frequency variation and return ratio variation within the period in chronological order, and establish a set of periodic operating characteristic data. Based on the execution feedback data of the stirring intervention command, the periodic time window after the stirring ends is first extracted. For example, if the sampling period is set to 60 minutes, real-time monitoring data of the sludge pump operating frequency and sludge return ratio are collected within this period. The sludge pump frequency data is recorded in units of sampling frequency per minute. For example, if the sampling frequency in a certain period is... Hz, calculate the maximum and minimum difference for all recorded values, frequency variation. Hz, reflux ratio data set as a sample value sequence Then the range of change All periodic data are arranged in chronological order to form a time series, and the mean change and fluctuation trend of each series are calculated to form a set of periodic operation characteristic data. The start time, end time and corresponding instruction number also need to be recorded to form a complete periodic record item, which is used for subsequent adjustment rule matching and control template mapping. Table 7. Data on Periodic Operation Characteristics

[0044] As shown in Table 7, the sludge pump frequency fluctuates by 7 Hz in period P1, and the sludge return ratio changes by 6%. This feature dataset can be used to match the sludge discharge command parameter combinations corresponding to the fuzzy level.

[0045] S502: Call the frequency change rate and reflux ratio change rate in the periodic operation feature data set, set the membership function of frequency change and reflux ratio change, divide the fuzzy level interval, use the fuzzy comprehensive evaluation method to evaluate the adjustment amplitude and response effect, calculate the comprehensive membership distribution result, and obtain the fuzzy adjustment level factor. The specific formulas for setting the membership functions of frequency variation and reflux ratio variation are as follows: ; Calculate the fuzzy response coordination parameters; in, Let i be the fuzzy response coordination parameter. Let i be the rate of change of the reflux ratio. Let i be the rate of change of frequency of the i-th term. This is the frequency-dependent amplification factor for adjusting the return current ratio. Let be the subjective weight value corresponding to the i-th fuzzy level. Let i be the fuzzy attribution trust factor. Let i be the feedback response intensity coefficient. Let j be the adjustment intensity conversion factor. Let j be the system damping response ratio. For fuzzy evaluation sample numbers, For the index number of the adjustment item, This represents the total number of samples for the rate of change of frequency and the rate of change of reflux ratio. formula: ; Detailed explanation of the formula and its calculation derivation: The formula is used to calculate the fuzzy response coordination parameters, which are used as the coordination judgment values ​​before the adjustment of the membership level in the fuzzy comprehensive evaluation. Parameter meanings and settings: The rate of change of the reflux ratio for the i-th term is set to 0.34; The rate of change of the frequency of the i-th term is set to 0.21; The frequency-dependent amplification factor for adjusting the return current ratio is set to 1.5. The subjective weight value corresponding to the i-th fuzzy level is set to 0.60 in the three-level fuzzy level; The fuzzy attribution trust factor for the i-th item is set to 0.72; The feedback response intensity coefficient for the i-th term is set to 0.35. The j-th adjustment intensity conversion factor is obtained by dividing the actual intervention signal output intensity by the theoretical standard intensity. The values ​​for the three samples in the data acquisition period are set to 1.02, 0.96, and 1.08, respectively. denoted as the j-th system damping response ratio, in the sludge return control system, it represents the reduction ratio of the fluid resistance inside the pipeline to the disturbance response. It is the ratio of the dynamic pressure loss caused by pipe wall friction and flow disturbance at bends during the passage of a unit volume of return sludge through the pipeline section to the system design flow resistance benchmark value. It is calculated by measuring the pressure loss of the return fluid between the inlet and outlet of the pipeline section using a flow velocity sensor and a differential pressure sensor, combined with the change in flow rate per unit time. It indicates the buffering hysteresis capability of the sludge pipeline system to input disturbances. The higher the value, the more passive the system response and the stronger the effect of the control command. The three sets of data are set to 0.18, 0.20, and 0.22, respectively. The total number of samples is set to 3; Substitute the parameters into the formula to calculate: ; ; ; Substitute each term into the complete formula: ; The result of 0.3167 indicates that the fuzzy level parameter corresponding to the change in the current input frequency and return ratio is at the lower level interval boundary in the control rules, which means that the control coordination of the sample point is weak before intervention. This value will be used as the index of the level factor matching in the fuzzy control module and will be entered into the subsequent fuzzy rule evaluation matrix to generate control level instructions.

[0046] The membership function is used to map the frequency and reflux ratio changes to their degree of belonging to different regulation state levels; Fuzzy level intervals classify variable values ​​into multiple fuzzy level ranges by setting division criteria, including stable ranges and fluctuating ranges. Fuzzy comprehensive evaluation method synthesizes the membership degree and weight of indicators at different levels to comprehensively evaluate the overall performance of the adjustment state; The comprehensive membership distribution result refers to the set of membership values ​​obtained by fusing the membership degrees of multiple input variables at each fuzzy level.

[0047] S503: Based on the membership level results in the fuzzy adjustment level factor, combined with the current cycle sludge discharge control instruction template, the sludge discharge control instruction set is obtained by filtering the multi-level parameter combinations corresponding to the level and matching the sludge discharge control instruction. Based on the membership level results of the fuzzy adjustment level factors, the fuzzy membership degree of each operating parameter in the current cycle is extracted. For example, if the frequency variation corresponds to the fuzzy level "medium" and the reflux ratio amplitude corresponds to the fuzzy level "high", then the fuzzy combination is [medium, high]. This combination is used as the input item for fuzzy rules and compared with the fuzzy rules defined in the sludge discharge control instruction template library. Each rule is defined as "If the frequency level = medium and the reflux ratio level = high, then the sludge discharge instruction group is {pump speed = 52Hz, duration = 300s, cycle = 3 times / h}". The system selects all multi-level parameter groups that match this level combination from the template and then performs similarity calculation. If multiple parameter groups meet the condition, the Euclidean distance between them and the current cycle's operating feature data with the smallest distance is compared. For example, if the current cycle is (frequency = 51Hz, reflux ratio = 73%), and the two instruction groups in the template are (52Hz, 75%) and (50Hz, 70%), then the Euclidean distance of the first group is [missing information]. The second group is Therefore, the first group is selected as the final sludge discharge control command, with the output format being: [pump speed, duration, cycle], and the final result being [52Hz, 300s, 3 times / h], which serves as the sludge discharge control command set.

[0048] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0049] It should be understood that the term "and / or" in this article merely describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0050] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0051] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0052] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0053] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0054] In the embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0055] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0056] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0057] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0058] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for full-process control of intelligent wastewater treatment plants based on digital twins, characterized in that: The method includes: S1: Acquire time-series data of multi-layer deep sludge concentration through distributed sensors, calculate the concentration change rate within the time window, use K-means clustering algorithm to distinguish fluctuation characteristics, set a standard deviation threshold based on the variance of the concentration change rate within the sliding window, and mark the continuous time period with a standard deviation lower than the threshold in the clustering results as the stable segment identification result. S2: Based on the stable segment identification results, obtain the settling trajectory of the target sludge particles, use the grey relational analysis method to calculate the time difference of the sludge particles reaching the bottom monitoring point, compare it with the standard settling curve and analyze the correlation, and generate the lag judgment result. S3: Call the particle size data in the hysteresis determination result whose correlation is lower than the benchmark value obtained by Spearman's rank correlation coefficient, combine it with the average concentration in the current stable segment identification result, use the TOPSIS method to construct a dynamic threshold matrix including concentration range, settling velocity deviation and particle size distribution dispersion, and output control threshold information. S4: Based on the control threshold information, extract the hierarchical data where the concentration range exceeds the safety threshold, match the interface position movement vector collected by the online turbidity meter, adjust the stirring intensity level and duration, and output the stirring intervention command.

2. The method for full-process control of intelligent wastewater treatment plants based on digital twins according to claim 1, characterized in that, The stable segment identification results include the identification time period number, depth level distribution, and fluctuation classification label; the hysteresis determination results include particle size grouping identifier, response delay value, and offset coefficient level; the control threshold information includes concentration range level, particle size adjustment weight, and settling velocity deviation factor; and the stirring intervention command includes stirring level number, stirring duration, and target depth identifier.

3. The method for full-process control of intelligent wastewater treatment plants based on digital twins according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: The time series data of multi-layer deep sludge concentration is acquired through distributed sensors. The continuous concentration value sequence of each layer within a fixed time window is extracted. The concentration change rate is calculated based on the adjacent sampling time points of each concentration sequence. After standardizing the sequence samples of the concentration change rate of the deep layer, the data is input into the K-means clustering algorithm to obtain the category label of the sample point in the cluster and obtain the fluctuation characteristic category distribution value. The K-means clustering algorithm is an unsupervised learning method that divides standardized concentration change rate sequences into multiple categories based on similarity. S102: Based on the fluctuation feature category distribution value, perform sliding window division on the corresponding concentration change rate sequence in each sample group, call the variance of the rate sample in each window to calculate, and obtain the mean of the window variance value as the mean threshold of the concentration change rate variance. S103: Call the distribution value of the fluctuation feature category and the threshold of the mean variance of the concentration change rate, compare the standard deviation of each sample group, filter the sample segments with a standard deviation less than the threshold and stable within three consecutive time windows, and generate stable segment identification results.

4. The method for full-process control of intelligent wastewater treatment plants based on digital twins according to claim 3, characterized in that, The specific steps of S2 are as follows: S201: Based on the stable segment identification results, obtain the concentration sequence and sampling time data of the depth position within the stable time period, identify the change of the settling position of sludge particles in the vertical direction, and record the time node when the particles reach the bottom monitoring point by combining the particle size distribution index and the monitoring point depth value, and generate a time-consuming dataset corresponding to the particle size. S202: Based on the time-consuming dataset corresponding to the particle size, call the theoretical sedimentation time series under the same particle size distribution in the standard sedimentation curve, establish a comparison vector between the actual sedimentation time of each group of particle sizes and the corresponding time nodes of the standard sequence, use the grey relational analysis method to calculate the correlation value between the particle size sedimentation sequence and the standard sequence, and obtain a list of particle size sedimentation correlation. The standard settling curve refers to the particle size-time mapping reference curve formed by measuring the theoretical settling time of particles per unit depth under ideal static water conditions. S203: Based on the particle size sedimentation correlation list, combined with particle size time data and number index, filter the particle size groups with correlation values ​​lower than the lag judgment threshold, record the time difference range and number mapping structure of the corresponding particle size, and establish the lag judgment result. The hysteresis determination threshold is the average grey relational value of particle size samples that have the same trend as the standard sedimentation curve within the current period.

5. The method for full-process control of intelligent wastewater treatment plants based on digital twins according to claim 4, characterized in that, The specific steps for S3 are as follows: S301: Call the particle size samples with trajectory correlation degree lower than the preset threshold in the hysteresis determination results, calculate the correlation between particle size and sedimentation response time difference based on Spearman's rank correlation coefficient, identify particle size segments with correlation coefficient lower than the statistical threshold, extract the corresponding sedimentation response time difference and particle size distribution index, and combine the number proportion and distribution spacing of the corresponding particle size segments to establish a related particle size distribution dataset. The preset threshold is a judgment standard set based on the similarity level between historical settlement behavior and standard trajectory, combined with empirical data. The statistical threshold is a decision boundary calculated using the distribution characteristics of the Spearman rank correlation coefficient on historical samples. S302: Call the average concentration value in the associated particle size distribution dataset and the stable segment identification result, extract the concentration range value and settling velocity change range value in the corresponding time window, calculate the particle size distribution dispersion in the settling zone, obtain the concentration range, settling velocity deviation, and particle size dispersion data, and generate a dynamic threshold evaluation index set. S303: Call the dynamic threshold evaluation index set, take the proximity and ranking consistency between each index as input variables, use the TOPSIS method to calculate the distance between multiple indices and rank them, calculate the concentration control factor score, extract the variable combination with the best score, and establish control threshold information.

6. The method for full-process control of intelligent wastewater treatment plants based on digital twins according to claim 5, characterized in that, The dynamic threshold evaluation index set is invoked, with the degree of similarity and ranking consistency among each index as input variables. The specific formula for ranking consistency is as follows: ； Calculate the ranking consistency index value; in, Representing the The first in the group The ranking consistency index value of the items. Representing the The first in the group The indicator in the first Sorting offset adjustment factor under variable Representing the The first in the group The indicator in the first The sorting position value under the variable, Representing the The first in the group The average of the ranking positions of the item indicator under all variables. Representing the The first in the group The absolute value of the variance of the ranking of each indicator under the variable. Representing the The first in the group The indicator is based on the empirical expected offset value of historical ranking stability. Representing the The total number of variables in the group that are involved in the sorting. For index number, For variable index number, This is the group number.

7. The method for full-process control of intelligent wastewater treatment plants based on digital twins according to claim 5, characterized in that, The specific steps for S4 are as follows: S401: Based on the control threshold information, extract the hierarchical data where the concentration range exceeds the safety threshold, and identify the corresponding spatial number and range value to obtain the concentration threshold abnormal hierarchical set; The safety threshold is set based on the combination of main control factor variables, and is mapped to the concentration range of each level through the TOPSIS sorting results. S402: Call the spatial number data in the concentration threshold anomaly level set, match the interface position movement vector collected in real time by the online turbidity meter, calculate the offset distance and velocity change rate of the interface corresponding to the spatial number in the sampling period through time series comparison of vector position, and generate interface offset trend parameter group. S403: Based on the offset speed and distance values ​​of each item in the interface offset trend parameter group, according to the classification criteria of speed range and displacement amplitude based on the stirring intensity level, obtain the matching stirring level, and output the stirring intervention instruction set in combination with the duration of action.

8. The method for full-process control of intelligent wastewater treatment plants based on digital twins according to claim 1, characterized in that, The method further includes step S5: S5: Based on the execution feedback data of the stirring intervention command, obtain the changes in the operating frequency of the sludge discharge pump and the sludge return ratio after stirring, use the fuzzy comprehensive evaluation method to evaluate the adjustment range and response effect, construct the operation status level classification result, and output a sludge discharge control command set including multiple adjustment parameters. The sludge discharge control instruction set includes sludge discharge rate levels, control frequency levels, and response adjustment ranges.

9. The method for full-process control of intelligent wastewater treatment plants based on digital twins according to claim 8, characterized in that, The specific steps of S5 are as follows: S501: Based on the execution feedback data of the stirring intervention command, obtain the operating frequency data of the periodic sludge discharge pump after stirring and the monitoring data of the sludge return ratio, calculate the frequency variation and return ratio variation amplitude within the period in chronological order, and establish a set of periodic operation characteristic data. S502: Call the frequency change rate and reflux ratio change rate in the periodic operation feature data set, set the membership function of frequency change and reflux ratio change, divide the fuzzy level interval, use the fuzzy comprehensive evaluation method to evaluate the adjustment amplitude and response effect, calculate the comprehensive membership distribution result, and obtain the fuzzy adjustment level factor. The membership function is used to map the frequency and reflux ratio changes to their degree of belonging to different regulation state levels; The fuzzy level intervals are defined by setting a division standard to classify variable values ​​into multiple fuzzy level ranges, including stable ranges and fluctuating ranges. The fuzzy comprehensive evaluation method synthesizes the membership degree and weight of indicators at different levels to comprehensively evaluate the overall performance of the adjustment state. The comprehensive membership distribution result refers to the set of membership values ​​obtained by fusing the membership of multiple input variables at each fuzzy level; S503: Based on the membership level results in the fuzzy adjustment level factor, and combined with the current cycle sludge discharge control instruction template, the sludge discharge control instruction is matched by filtering the multi-level parameter combinations corresponding to the level, and the sludge discharge control instruction set is obtained.

10. The method for full-process control of a smart wastewater treatment plant based on digital twins according to claim 9, characterized in that, The specific formula for the membership function of setting the frequency variation and the return current ratio variation is as follows: ； Calculate the fuzzy response coordination parameters; in, Let i be the fuzzy response coordination parameter. Let i be the rate of change of the reflux ratio. Let i be the rate of change of frequency of the i-th term. This is the frequency-dependent amplification factor for adjusting the return current ratio. Let be the subjective weight value corresponding to the i-th fuzzy level. Let i be the fuzzy attribution trust factor. Let i be the feedback response intensity coefficient. Let j be the adjustment intensity conversion factor. Let j be the system damping response ratio. For fuzzy evaluation sample numbers, For the index number of the adjustment item, This represents the total number of samples for the rate of change of frequency and the rate of change of reflux ratio.

Images