Sewage treatment process model parameter correction method and system

Abstract

The invention relates to the technical field of sewage treatment, in particular to a sewage treatment process model parameter correction method and system. The method comprises the following steps: acquiring sensing data of a multi-point facility through a sewage treatment sensing network, extracting a topological structure of a sewage pipeline, projecting the data into the topological structure, marking position information, constructing an initial sewage treatment process model based on the information, extracting a sewage data set of a treatment process, and establishing a sewage treatment process model; the system simulates a sewage treatment environment, generates virtual flow field data, tracks particle motion, draws a sedimentation curve and generates particle sedimentation track offset data, obtains morphological parameters of a sedimentation tank, reconstructs the framework of the sedimentation tank, predicts a sedimentation imbalance state, backtracks and analyzes model parameters through an abnormal state, and corrects the model parameters. And generating an optimized sewage treatment process model. According to the invention, the operation efficiency and the treatment effect of the sewage treatment facility are integrally improved, and the intelligent and refined development of the sewage treatment technology is promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of sewage treatment, and particularly to a method and system for correcting parameters of a sewage treatment process model. Background Art

[0002] Existing sewage treatment technologies often rely on experience and static models, resulting in insufficient response to the dynamic changes in the sewage treatment process. Traditional monitoring methods cannot reflect the treatment status of sewage in real time, causing low treatment efficiency and resource waste. Especially in complex sewage treatment environments, the changes in sewage characteristics cannot be accurately captured, limiting the improvement of treatment effects. Many sewage treatment facilities face various unforeseen situations during operation, such as water quality fluctuations and flow rate changes, which have a direct impact on the treatment effects. Existing technologies are powerless in dealing with these changes, lacking flexibility and adaptability, unable to achieve rapid adjustment, resulting in increased operating costs of treatment facilities and instability of treatment effects, thereby affecting the effects of environmental protection and resource reuse. Traditional sewage treatment facilities face problems such as uneven particle sedimentation and sedimentation imbalance. The identification and feedback of these problems by existing technologies are relatively lagging, and effective dynamic adjustment and optimization have not been achieved. Traditional methods usually rely on regular manual monitoring and adjustment, resulting in slow response speed and unstable treatment effects during the treatment process. Lack of scientific sediment analysis and flow field simulation makes it difficult to provide real-time data support for operation decisions. Summary of the Invention

[0003] Based on this, it is necessary to provide a method and system for correcting parameters of a sewage treatment process model to solve at least one of the above technical problems.

[0004] To achieve the above object, a method for correcting parameters of a sewage treatment process model includes the following steps: Step S1: Collect sensing data of multi-point sewage treatment facilities through a sewage treatment sensing network; extract the sewage pipeline topological structure of the sewage treatment sensing network; project the multi-point sewage facility sensing data into the sewage pipeline topological structure, and mark the corresponding position information; Step S2: Construct an initial sewage treatment process model based on the corresponding position information and multi-point sewage facility sensing data; extract the sewage data set of the treatment process in the initial sewage treatment process model; Step S3: Perform sewage treatment environment simulation according to the sewage data set of the treatment process to generate a simulated sewage treatment process environment; simulate virtual flow field data based on the simulated sewage treatment process environment and the sewage data set of the treatment process; Step S4: Perform particle motion tracking based on the sewage data set of the treatment process and the virtual flow field data, and draw a particle sedimentation curve; perform sedimentation offset mapping fitting based on the preset ideal sedimentation situation and the particle sedimentation curve to generate particle sedimentation trajectory offset data; Step S5: Obtain the morphological parameters of the sedimentation tank; reconstruct the sedimentation tank architecture according to the morphological parameters of the sedimentation tank, and determine the reference inclination angle of the influent diversion plate based on the sedimentation tank architecture; predict the sedimentation imbalance abnormal state based on the reference inclination angle of the influent diversion plate and the particle sedimentation offset data; Step S6: Backtrack and analyze the abnormal over-limit parameters in the initial sewage treatment process model through the sedimentation imbalance abnormal state; perform model parameter correction on the initial sewage treatment process model based on the abnormal over-limit parameters to generate a corrected sewage treatment process model.

[0005] Through the multi-point data collection of the sewage treatment sensing network, the present invention ensures the comprehensive monitoring of the operation status of sewage treatment facilities. The extracted pipeline topological structure provides a clear spatial basis for subsequent data analysis. The position information marked after data projection enhances the traceability and practicability of the data. The construction of the initial sewage treatment process model ensures the accurate description of the sewage treatment process. The extracted sewage data set provides a necessary basis for the optimization of the model. The simulated sewage treatment process environment generated by environmental simulation makes the simulation of the actual treatment situation more realistic. The simulation of virtual flow field data provides important support for particle motion analysis. The drawing of particle motion tracking and sedimentation curves reveals the particle behavior characteristics in the sewage treatment process. The generation of sedimentation offset mapping fitting provides a data basis for the optimization of the sedimentation process. The acquisition of the morphological parameters of the sedimentation tank and the reconstruction of the sedimentation tank architecture ensure the systematic understanding of the sedimentation process. The prediction based on the reference inclination angle of the diversion plate and the sedimentation data can identify the sedimentation imbalance abnormal state in a timely manner. The backtrack analysis of the abnormal over-limit parameters effectively points to the adjustment direction of the model parameters. The sewage treatment process model after model parameter correction improves the accuracy and applicability of the model, provides a reliable theoretical basis for the subsequent optimization of sewage treatment, overall improves the operation efficiency and treatment effect of sewage treatment facilities, and promotes the intelligent and refined development of sewage treatment technology.

[0006] The present invention also provides a sewage treatment process model parameter correction system for executing the sewage treatment process model parameter correction method as described above. The sewage treatment process model parameter correction system includes: A data collection module for collecting multi-point sewage treatment facility sensing data through the sewage treatment sensing network; extracting the sewage pipeline topological structure of the sewage treatment sensing network; projecting the multi-point sewage facility sensing data onto the sewage pipeline topological structure and marking the corresponding position information; A model construction module for constructing an initial sewage treatment process model based on the corresponding position information and the multi-point sewage facility sensing data; extracting the sewage data set during the treatment process in the initial sewage treatment process model; An environmental simulation module, which is used to perform sewage treatment environmental simulation based on the sewage data set of the treatment process, generate an environmental simulation of the sewage treatment process; and simulate virtual flow field data based on the environmental simulation of the sewage treatment process and the sewage data set of the treatment process; A particle tracking module, which is used to track the movement of particles based on the sewage data set of the treatment process and the virtual flow field data, and draw a particle sedimentation curve; perform sedimentation offset mapping fitting based on the preset ideal sedimentation condition and the particle sedimentation curve, and generate particle sedimentation trajectory offset data; A sedimentation analysis module, which is used to obtain the morphological parameters of the sedimentation tank; reconstruct the sedimentation tank structure according to the morphological parameters of the sedimentation tank, and determine the reference inclination angle of the inlet guide plate based on the sedimentation tank structure; predict the sedimentation imbalance abnormal state based on the reference inclination angle of the inlet guide plate and the particle sedimentation offset data; A parameter correction module, which is used to retrospectively analyze the abnormal over-limit parameters in the initial sewage treatment process model through the sedimentation imbalance abnormal state; perform model parameter correction on the initial sewage treatment process model based on the abnormal over-limit parameters, and generate a corrected sewage treatment process model.

[0007] Through the application of the data acquisition module, the present invention realizes the comprehensive monitoring and real-time data acquisition of sewage treatment facilities, ensures the accuracy and timeliness of data, the extracted sewage pipeline topological structure provides a clear spatial framework for subsequent analysis, the data projection and position information marking enhance the traceability of data, effectively support the basis of model construction, the model construction module generates an initial sewage treatment process model based on multi-point sensing data, provides an accurate description of the treatment process, the extracted sewage data set provides a necessary basis for model optimization, the environmental simulation generated by the environmental simulation module ensures the effective simulation of the actual treatment situation, the simulation of virtual flow field data provides important support for particle motion analysis, the particle tracking module reveals the behavioral characteristics of particles during the treatment process through the drawing of the sedimentation curve, the generation of sedimentation offset mapping fitting provides a data basis for the optimization of the sedimentation process, the morphological parameters of the sedimentation tank obtained by the sedimentation analysis module provide a scientific basis for the reconstruction of the sedimentation tank structure, the prediction based on the reference inclination angle of the guide plate can timely identify the sedimentation imbalance abnormal state, the parameter correction module clarifies the direction of model parameter adjustment through the retrospective analysis of abnormal over-limit parameters, the sewage treatment process model after model parameter correction significantly improves the accuracy and applicability of the model, provides reliable theoretical support for the optimized operation of sewage treatment facilities, overall improves the sewage treatment effect and treatment efficiency, and promotes the intelligent and refined development of sewage treatment technology. Description of the Drawings

[0008] Figure 1 It is a schematic diagram of the step flow of a method for correcting the parameters of a sewage treatment process model; Figure 2It is a schematic flowchart of the detailed implementation steps of step S2; The realization, functional characteristics and advantages of the purpose of the present invention will be further described in conjunction with the embodiments and with reference to the accompanying drawings. Specific implementation manners

[0009] The technical method of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those skilled in the art within the scope of the present invention without creative work belong to the scope of protection of the present invention.

[0010] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings represent the same or similar parts, and thus their repeated description will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor methods and / or microcontroller methods.

[0011] It should be understood that although the terms "first", "second", etc. may be used here to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, the first unit can be called the second unit, and similarly the second unit can be called the first unit. The term "and / or" used here includes any and all combinations of one or more of the listed associated items.

[0012] To achieve the above object, please refer to Figures 1 to 2 , a method for correcting parameters of a sewage treatment process model, comprising the following steps: Step S1: Collect multi-point sewage treatment facility sensing data through a sewage treatment sensing network; extract the sewage pipeline topological structure of the sewage treatment sensing network; project the multi-point sewage facility sensing data into the sewage pipeline topological structure, and mark the corresponding position information; Step S2: Construct an initial sewage treatment process model based on the corresponding position information and multi-point sewage facility sensing data; extract the sewage treatment process sewage data set in the initial sewage treatment process model; Step S3: Perform sewage treatment environment simulation according to the sewage treatment process sewage data set to generate a simulated sewage treatment process environment; simulate virtual flow field data based on the simulated sewage treatment process environment and the sewage treatment process sewage data set; Step S4: Perform particle motion tracking based on the processed process sewage dataset and virtual flow field data, and plot the particle settlement curve; perform settlement offset mapping fitting based on the preset ideal deposition situation and the particle settlement curve to generate particle settlement trajectory offset data; Step S5: Obtain the sedimentation tank morphology parameters; reconstruct the sedimentation tank structure according to the sedimentation tank morphology parameters, and determine the reference inclination angle of the inlet guide plate based on the sedimentation tank structure; predict the sedimentation imbalance abnormal state based on the reference inclination angle of the inlet guide plate and the particle settlement offset data; Step S6: Backtrack and analyze the abnormal overlimit parameters in the initial sewage treatment process model through the sedimentation imbalance abnormal state; perform model parameter correction on the initial sewage treatment process model based on the abnormal overlimit parameters to generate a corrected sewage treatment process model.

[0013] Through the multi-point data collection of the sewage treatment sensing network, the present invention ensures the comprehensive monitoring of the operating state of sewage treatment facilities. The extracted pipeline topology structure provides a clear spatial basis for subsequent data analysis. The position information marked after data projection enhances the traceability and practicability of the data. The construction of the initial sewage treatment process model ensures the accurate description of the sewage treatment process. The extracted sewage dataset provides a necessary basis for the optimization of the model. The simulated sewage treatment process environment generated by environmental simulation makes the simulation of the actual treatment situation more realistic. The simulation of virtual flow field data provides important support for particle motion analysis. The particle motion tracking and the plotting of the settlement curve reveal the particle behavior characteristics in the sewage treatment process. The generation of settlement offset mapping fitting provides a data basis for the optimization of the sedimentation process. The acquisition of the sedimentation tank morphology parameters and the reconstruction of the sedimentation tank structure ensure a systematic understanding of the sedimentation process. The prediction based on the reference inclination angle of the guide plate and the settlement data can timely identify the sedimentation imbalance abnormal state. The backtracking analysis of the abnormal overlimit parameters effectively points to the adjustment direction of the model parameters. The sewage treatment process model after model parameter correction improves the accuracy and applicability of the model, provides a reliable theoretical basis for the subsequent sewage treatment optimization, overall improves the operating efficiency and treatment effect of sewage treatment facilities, and promotes the intelligent and refined development of sewage treatment technology.

[0014] In an embodiment of the present invention, the method for correcting the parameters of a sewage treatment process model includes the following steps: Step S1: Collect multi-point sewage treatment facility sensing data through a sewage treatment sensing network; extract the sewage pipeline topology structure of the sewage treatment sensing network; project the multi-point sewage facility sensing data onto the sewage pipeline topology structure and mark the corresponding position information; In this embodiment, a sensor node network is deployed in the sewage treatment plant. A multi-parameter water quality sensor of model YSI EXO2 is selected to collect parameters such as dissolved oxygen (DO), chemical oxygen demand (COD), ammonia nitrogen (NH3-N), and water temperature (T) in the sewage pipe network. The sampling frequency is set to 1 Hz, and the number of deployed nodes is 50. The sensor nodes are distributed to cover the main sewage pipelines, branch pipelines, and key nodes. A unique spatial number is set for each node. The topological structure information of the sewage pipe network is extracted by using the spatial coordinate annotation method based on the geographic information system (GIS). The water quality data collected by each sensor is projected onto the corresponding sewage pipe network topological structure in the form of a space-time triple (node number, sampling time, measured value). The flow direction attribute is determined by using the topological relationship, and the pipe segment number, distance from the inlet, pipe diameter number, and corresponding time stamp corresponding to each group of data are marked, thus completing the spatial mapping process of multi-point sensing data.

[0015] Step S2: Construct an initial sewage treatment process model based on the corresponding location information and multi-point sewage facility sensing data; extract the sewage treatment process sewage data set in the initial sewage treatment process model; In this embodiment, after obtaining the spatially marked sensing data, an initial model of the sewage treatment process is constructed. The sewage treatment unit process based on the oxidation ditch process is selected as the modeling object. The initial model is constructed by combining data-driven modeling and physical rule modeling. The mapped water quality data described above is used as the input variable. The dynamic correlation between different measurement points is analyzed by the grey relational analysis method, and an initial prediction model with the influent water quality parameters as the input and the effluent water quality parameters as the output is established. The initial model includes process variables such as the reaction tank volume V (unit: m³), sludge retention time SRT (unit: d), and reflux ratio R (unit: dimensionless ratio). After extracting the intermediate data of the sewage treatment process stage in the initial model, a sewage treatment process sewage data set is constructed. The data set includes dynamic change data such as DO, COD, NH3-N, water temperature, and flow velocity per hour, and is segmented and coded according to time periods to form a daily-level sample set containing at least 720 records.

[0016] Step S3: Perform sewage treatment environment simulation according to the sewage treatment process sewage data set to generate a simulated sewage treatment process environment; simulate virtual flow field data based on the simulated sewage treatment process environment and the sewage treatment process sewage data set; In this embodiment, based on the above constructed sewage dataset of the treatment process, a three-dimensional treatment unit geometric model is constructed in the ANSYS Fluent simulation platform. The fluid property is set as weakly compressible fluid, and the RNG k-ε turbulence model is used for simulation solution. The fluid boundary conditions are set as the influent flow rate Q = 2000 m³ / d in the actually collected data. The inlet is set as the velocity inlet boundary condition, and the outlet is set as the pressure outlet. Each time step is set as 1 s, and the simulation period is 24 h. A virtual sewage treatment environment such as the internal flow velocity distribution and vorticity structure is generated through simulation. Subsequently, combined with the above sewage dataset of the treatment process with time synchronization, the data of each time step is mapped to the cross-section position of the flow field, and the Euler-Lagrange two-way coupling method is used to simulate the particle behavior in this virtual environment, and the velocity, acceleration and displacement of each particle at each time node are extracted to form virtual flow field particle motion data.

[0017] Step S4: Track the particle motion based on the sewage dataset of the treatment process and the virtual flow field data, and draw the particle settlement curve; perform settlement offset mapping fitting based on the preset ideal deposition situation and the particle settlement curve to generate particle settlement trajectory offset data; In this embodiment, the above obtained virtual flow field data is input into the particle trajectory tracking module, and the initial particle radius r = 0.05 mm and density = 1.3 g / cm³ are set. The particle tracking algorithm is used for hourly evolution calculation to track the complete trajectory of each particle from the inlet to its settlement at the bottom of the tank, and a two-dimensional function relationship curve h(t) of the settlement height h and time t is drawn. The particle swarm optimization (PSO) algorithm is used to fit the difference between this settlement curve and the set ideal settlement model, and the fitting residual is defined as the settlement offset degree , and a settlement trajectory offset dataset is generated through three-dimensional spatial interpolation, including the offset direction, offset velocity and offset duration, and the spatial path deviation value of the particle in the actual structure is marked.

[0018] Step S5: Obtain the morphological parameters of the sedimentation tank; reconstruct the sedimentation tank architecture according to the morphological parameters of the sedimentation tank, and determine the reference inclination angle of the inlet deflector based on the sedimentation tank architecture; predict the sedimentation imbalance abnormal state based on the reference inclination angle of the inlet deflector and the particle settlement offset data; In this embodiment, a three-dimensional laser scanning device (model: Leica BLK360) is used to conduct on-site measurement of the existing sedimentation tank to obtain structural parameters such as the boundary contour, tank depth, tank length, and tank width of the sedimentation tank structure. The tank depth is 4.5 m, the tank length is 20 m, and the tank width is 10 m. These parameters are used to reconstruct the sedimentation tank CAD geometric model. Subsequently, according to the average value V0 = 0.35 m / s of the inlet flow velocity field of the sedimentation tank, the reference inclination angle of the inlet deflector is determined by the streamline projection method , making the initial streamline tangential to the upper surface of the deflector, and the inclination angle The calculation formula is tan⁻¹(h / L), where h is the height difference between the end of the deflector and the bottom of the tank, and L is the projected length of the deflector. It is calculated that it is about 18°. This angle is used as the standard deflector design benchmark, and combined with the aforementioned particle settlement offset data, the abnormal state of sedimentation imbalance is predicted based on the analysis of the difference between the sedimentation trajectory concentration area and the deflector design vector.

[0019] Step S6: Retroactively analyze the abnormal over-limit parameters in the initial sewage treatment process model through the abnormal state of sedimentation imbalance; based on the abnormal over-limit parameters, correct the model parameters of the initial sewage treatment process model to generate a corrected sewage treatment process model.

[0020] In this embodiment, the time period corresponding to the predicted sedimentation imbalance area is traced back to the initial sewage treatment process model, the processing parameters corresponding to the time period are matched, the fluctuation ranges of indexes such as DO, COD, and NH3-N during this time period are calculated, and the error between them and the standard input variables of the model is determined for over-limit. When any parameter deviates from its optimal control range by more than 10%, it is marked as an abnormal over-limit parameter. For example, if DO drops from the normal value of 2 mg / L to 1.2 mg / L during a specific time period, it is determined as over-limit. Subsequently, the corresponding DO model control parameter α in the initial model is corrected, and the mapping relationship between the influent and effluent DO is re-fitted using the least squares method to complete the model parameter correction of the initial sewage treatment process model. Finally, the corrected model structure and parameter values are output and synchronously recorded in the model version control library.

[0021] Preferably, step S1 includes the following steps: Step S11: Conduct a layout design for the key nodes of the sewage treatment facility to obtain the data of the sensing deployment sites; configure multi-modal acquisition units for each sensing deployment site to construct a multi-point composite acquisition unit; Step S12: Construct a sewage treatment sensing network based on the multi-point composite acquisition unit; perform multi-point synchronous sampling based on the sewage treatment sensing network to obtain the sensing data of the multi-point sewage treatment facility. Among them, the sampling frequency is set to 1 Hz to 5 Hz, and the sampling duration is not less than 10 min; Step S13: Perform a structural decoding process on the sensing data of the multi-point sewage treatment facility to obtain the standardized sewage sensing node data; Step S14: Analyze the topological structure data of the sewage pipeline based on the sewage treatment sensing network. Among them, the number of nodes is not less than 20, and the number of connected edges is not less than the number of nodes - 1; Step S15: Project the sensing data of the multi-point sewage treatment facility into the topological structure data of the sewage pipeline to obtain the position information mapping data, and the projection error threshold is limited to ≤0.5 m; Step S16: Perform site marking processing according to the position information mapping data to obtain the corresponding position information of the multi-point sewage sensing projection data.

[0022] In this embodiment, according to the sewage treatment process flow chart and the positional relationship of each treatment unit, key nodes such as the sewage inlet, outlet, primary sedimentation tank, activated sludge tank, secondary sedimentation tank, and sludge return port are extracted. The digital map platform ArcGIS is used in combination with the on-site BIM model for three-dimensional positioning analysis. Initial point layout simulation is carried out based on the geometric distribution of the nodes and the main transmission paths. The flow velocity, temperature, and pH distribution at the key nodes are simulated by the layout simulation software COMSOL Multiphysics. Representative sensing deployment sites are screened. Each site needs to meet at least one structural continuity constraint and two data stability thresholds to obtain the sensing deployment site data. A multimodal acquisition unit composed of a temperature sensor, turbidity sensor, pH sensor, and laser particle monitoring unit is configured at each site. All units are assembled into a composite housing structure, the material of which is polytetrafluoroethylene and embedded with an anti-corrosion platinum coating. A multi-point composite acquisition unit with a power supply module, signal conditioning module, and analog-to-digital conversion module is constructed. In the process of constructing the sewage treatment sensing network, the multi-point composite acquisition unit is connected to the Internet of Things data acquisition platform in the LoRaWAN (Long Range Wide Area Network with Low Power Consumption) mode. Each node ID and its network protocol address are set, and a master-slave communication mechanism is established. The master node receives the data of all slave nodes and uploads it to the edge computing server. The structure of the sensing network is defined as G=(V,E), where V is the set of sensor deployment points and E is the set of edges of the network connectivity relationship between nodes. The system sets the side length threshold that the shortest path between every two points does not exceed 25 meters, forming a sewage treatment sensing network structure with a stable topology. At the same time, multi-point synchronous sampling operations are started at each node. The sampling frequency is set between 1Hz and 5Hz, and the sampling duration is not less than 600 seconds. Each data packet needs to contain a timestamp, site number, four-channel original measurement values, multimodal fusion identifier, and time calibration label. When performing structural decoding processing on the multi-point sewage treatment facility sensing data, first, the HEX format data transmitted back by LoRaWAN is restored to the original data by the decoding module of the edge server. The protocol parsing template is used to separate each field. The parsed fields include channel type, sampled voltage value, temperature correction value, and timestamp. The field-level double-check mechanism is used for CRC (Cyclic Redundancy Check) error identification. Data with a correction rate lower than 95% will be directly excluded. After successful decoding, the values of each channel are respectively converted to the unit and the sensing source channel is marked, and then converted into standardized sewage sensing node data in a unified format and organized into a two-dimensional array according to the site serial number and sampling time. In the process of analyzing the sewage treatment sensing network, based on the aforementioned sensing deployment site data, an undirected graph structure is generated through the NetworkX library in Python. Each site is used as a graph node, and the physical distance between nodes less than 30 meters is set as a connected edge. It is required that the number of graph nodes in the network is not less than 20, and the number of connected edges is not less than the number of nodes minus one, forming a basic topological structure that meets the minimum spanning tree constraint.The shortest path distances between all nodes are solved by the Dijkstra algorithm and written into the topological adjacency matrix for subsequent analysis of the sewage transmission path. In the operation of projecting the sensing data of multi-point sewage treatment facilities onto the sewage pipeline topological structure data, first, the spatial positions of the sensing sites under the GIS coordinates and the line segment and length information of each section of the pipeline in the topological structure are obtained to construct a three-dimensional projection matrix model. The spatial position points of the sensing nodes are mapped to the corresponding pipeline segment projection points using the shortest Euclidean distance criterion. The maximum allowable projection error threshold is set to 0.5 meters, and the cosine angle calibration algorithm based on the coordinate point distance ratio is used to adjust the projection error. Data with errors exceeding the threshold range are removed. During the process of marking the positions according to the mapped data based on the position information, the data points retained after the projection operation are marked with position numbers in the sewage pipeline structure. The marking method uses a three-layer nested structure. The first level is the main pipeline number, the second level is the pipeline segment number, and the third level is the sensing site number. The numbering rule is represented in the form of "M-P-V", where M is the main number, P is the segment number, and V is the sensor number. At the same time, the numbering information and the corresponding standardized sensing data are written into the data index table to support the spatial node association call in the subsequent model parameter correction. The marking result finally generates the corresponding position information form of the multi-point sewage sensing projection data, with the format requirement being the CSV format. All fields are required to contain the site number, three-dimensional projection coordinates, pipe segment number, and mapping error value fields. The error is retained to three decimal places, with the unit being meters.

[0023] Preferably, step S2 includes the following steps: Step S21: Match the position data of the corresponding position information and the sensing data of the multi-point sewage facilities. Set the matching tolerance range to ±0.5 meters to obtain the position information matching data; Step S22: Perform process section attribution division processing on the position information matching data. According to the node distribution in the sewage pipeline topological structure, each matching point is attributed to the process section closest to it. Set the attribution distance threshold to 1.0 meters to obtain the process section allocation data; Step S23: Screen the processing characteristics of the process section allocation data. Select data samples with water quality parameters of pH 6.5–8.5, dissolved oxygen greater than 2 mg / L, and COD less than 300 mg / L to obtain the screened sensing data; Step S24: Perform time series splicing on the screened sensing data. Splice adjacent data in the order of the acquisition timestamps. Set the maximum time interval not to exceed 10 seconds to obtain the spliced sensing data sequence; Step S25: Construct an initial sewage treatment process model based on the spliced sensing data sequence; extract the process section time-series segment data from the initial sewage treatment process model, where the minimum time span of the treatment unit is set to 30 seconds; extract the sewage data set of the treatment process according to the process section time-series segment data.

[0024] In this embodiment, the position information matching process is performed on each record in the multi-point sewage sensing projection data generated in the previous stage. First, the three-dimensional space coordinates (x, y, z) and timestamp information in the record are extracted, and the Euclidean distance is calculated between them and the node space coordinates in the sewage pipeline topology structure diagram. The matching tolerance range is set to ±0.5 meters, and the space matching formula , where ( , , ) is the position of the sensing data point, ( , , ) is the topological node coordinate. If the calculation result If it is less than or equal to 0.5 meters, the matching of the sensing data record with the corresponding topological node is successful and output as position information matching data. The matching operation uses the KDTree method in the scipy.spatial library in Python to construct an index tree structure for efficient matching operations. Each data output includes the sensor measurement value, measurement time, measurement space coordinates, and the matching node number. Import the position information matching data into the process section attribution module. This module is based on the sewage treatment units defined in the topological structure diagram. All nodes in the topological structure are divided into several process section areas according to the process unit section numbers. Calculate the Euclidean distance between the matching data point and the central nodes of all process sections, and take the central node of the process section with the smallest distance and not exceeding 1.0 meters as the attributed process section. The attribution distance calculation uses the same three-dimensional space distance calculation formula as described above. Set the distance threshold d ≤ 1.0 meters. If multiple process sections meet the conditions, the process section corresponding to the pipe section number where the current point is located is preferentially matched. The attribution operation is completed in the PostGIS spatial database, and the ST_Distance function is used for spatial matching and screening. The attribution result is output as process section allocation data, including fields: original sensing data number, attributed process section number, matching distance, attribution confirmation flag. Perform processing characteristic screening on each sensing parameter in the process section allocation data. Read the pH value, dissolved oxygen DO concentration, and chemical oxygen demand COD value in each data. Samples with a pH range not between 6.5 and 8.5 are excluded. Samples with a dissolved oxygen DO lower than 2 mg / L are excluded. Samples with a COD higher than 300 mg / L are excluded. The screening logic is implemented using NumPy array logical operations, and the expression is set as: (pH ≥ 6.5) ∧ (pH ≤ 8.(5) ∧ (DO > 2) ∧ (COD < 300). All samples that satisfy the above conditions simultaneously are retained. After screening, the data structure retains the original field structure and adds a screening label field to record whether each sample meets the screening conditions. Finally, the screened sensing data set is output. The screened sensing data is sorted by the acquisition timestamp, and the data sample sets within each process section are grouped. The data within each group is arranged in chronological order. The time interval between adjacent samples is calculated. The maximum interval threshold is set to 10 seconds. Adjacent samples with a time interval less than or equal to 10 seconds are spliced to form a continuous data segment. Samples with a time interval greater than 10 seconds are treated as the starting points of new sequences. The splicing operation uses the groupby method in the pandas library in Python to group by the process section number, and then applies the rolling method to traverse the window for time difference calculation and splicing marking processing. Finally, a spliced sensing data sequence is formed. Each sequence contains data points of three parameters, pH, DO, and COD, within a continuous time period. The data structure is adjusted to a list structure, including fields: sequence number, start time, end time, number of data points, and sequence data content. Each spliced sensing data sequence is used as an input sample to construct an initial sewage treatment process model. The modeling dimension is set to the time series dimension and the parameter dimension. An LSTM (Long Short-Term Memory) network model is used to construct a processing process sequence prediction model. The Sequential model in the TensorFlow framework is used to build a three-layer LSTM network structure. The input dimension is (N, 3), where N is the number of time points in each data sequence and 3 is the parameter dimension (pH, DO, COD). The output is the values of the three parameters predicted for the next time point. The minimum time span for each segment in the training data set is set to 30 seconds. If the length of the sample sequence is less than 30 data points, it does not participate in the modeling. The time change trends of pH, DO, and COD in each segment are extracted as the model training target. After the initial model is output, the data in the model input sequence is uniformly classified into process section time series segment data. The data structure remains consistent with the sequence splicing, and a processed sewage data set for the process is uniformly constructed.

[0025] Preferably, step S3 includes the following steps: Step S31: Perform multi-zone pressure simulation on the processed sewage data set and record the pressure response to obtain the unit partial pressure characteristics; Step S32: Perform watershed continuous infiltration stratification processing on the unit partial pressure characteristics to obtain a stratified infiltration structure; Based on the stratified infiltration structure, perform a flow state channel configuration simulation on the processed sewage data set to generate a simulated sewage treatment process environment; Step S33: Perform distributed pressure control mapping based on the simulated sewage treatment process environment to obtain unit hydraulic distribution data; reconstruct the hydraulic time-series path based on the unit hydraulic distribution data; Step S34: Conduct path logic stitching simulation according to the hydraulic time-series path, and construct a boundary loop structure based on the path logic stitching data; determine the flow boundary conditions based on the boundary loop structure and the hydraulic time-series path, and simulate virtual flow field data based on the flow boundary conditions and the sewage data set of the treatment process.

[0026] In this embodiment, the sewage data set of the treatment process generated in the previous stage is imported into a three-dimensional finite volume multi-region pressure simulation system to construct a simulation model based on the spatial process section division. The model uses the interFoam solver module in the OpenFOAM framework. The size of the simulation domain is set according to the actual size of each process section. A three-dimensional grid is constructed in units of 0.1 meters. The boundary conditions are set as a stable flow velocity inlet (1.5 m / s) and a constant pressure outlet (101325 Pa). Each treatment unit area is used as an independent simulation area. The partition definition is completed through the grid partition control dictionary file blockMeshDict. The hydraulic parameters in the sewage data set are used as the initial field variables and input into the model. The PISO pressure iteration algorithm is used for transient solution. The pressure change data of each unit area at each time step (the set time step is 0.1 s) is recorded, and the pressure response curve of each unit within the simulation time is exported. The average pressure gradient, response peak value, and response delay time of each unit are output as characteristic indicators, generating the unit partial pressure characteristics including unit number, spatial position, average response value, maximum pressure, and response time. The unit partial pressure characteristics are input into a three-dimensional infiltration layer analysis system. A flow direction vector field is constructed according to the average pressure gradient direction between units. The entire pressure field is divided into multiple continuous infiltration layers using a three-dimensional layer interpolation algorithm. The layer division threshold is set to 0.2 Pa / m, and the continuous region of the pressure gradient change is divided into one layer at this threshold. The layer division operation uses the piecewise linear interpolation module in MATLAB. At least three or more infiltration layer structure data are output in each process section. The structure data includes layer number, start and end Z coordinates, average pressure value, and pressure change rate. The layer division structure is imported into the fluid volume element simulation platform. Combining the flow velocity and flow rate information in the sewage data set, a flow pattern channel configuration model is constructed using the Ansys Fluent simulation module. The turbulence model is set as The model has the inlet conditions uniformly set as a fixed flow rate (1.5 m / s) and a temperature constant (298 Kelvin). The channel configuration boundary is defined based on the pressure distribution and flow direction data inside different permeable layers, and a complete simulated sewage treatment process environment is output. The simulation output data includes a velocity field distribution map, a channel structure element model, and a node connection table. The boundary structures and internal flow velocity information of each treatment unit in the simulated sewage treatment process environment are extracted, and a distributed pressure-controlled mapping model is constructed. The volScalarField structure in OpenFOAM is used to generate the volume pressure mapping value of each unit, and each volume unit is mapped to the corresponding treatment unit number, and a unit hydraulic distribution data structure is constructed. The fields include unit number, unit volume pressure, pressure difference vector, and flow direction unit vector. The hydraulic path reconstruction operation uses the Dijkstra shortest path algorithm to construct a weight matrix in a directed graph with each treatment unit as a node. The weight value is the ratio of the unit pressure difference between two units to the path length. After constructing the path graph, the maximum hydraulic flux path is iteratively output starting from the inlet unit, and the connection nodes, pressure change values, and time tags at each time step in the output path sequence are output to complete the hydraulic time-series path reconstruction. The connection order and unit boundary topology structure in the hydraulic time-series path are read, and the path logic stitching simulation operation is performed. The connection determination of the interface surfaces between adjacent path nodes is carried out. The judgment criterion is that the included angle of the coplanar boundary normal vectors between two nodes is less than 10 degrees and the relative position distance is less than 0.1 m. The stitching logic uses a triangular mesh splicing algorithm to insert connection surface nodes in the coplanar area to form a boundary closed loop. All closed loop areas are extracted as a set of boundary loop structures. The structure records the closed loop path number, the number of interface surfaces, the stitching point coordinates, and the set of surface normal vectors. After constructing the geometric model of each boundary loop, the flow boundary conditions of each boundary are generated by combining the flow rate and pressure data in the hydraulic path. The boundary conditions are mainly fixed pressure boundaries and fixed flow rate boundaries. The specific condition values are extracted from the sewage data set in the treatment process as the set values. All boundary conditions and the sewage data set are imported into the virtual flow field simulation module, and a three-dimensional transient flow field simulation process is performed. The simulation module uses the PISO pressure iteration method based on the control volume method, the time step is set to 0.05 s, and the total simulation time is set to 60 s. The exported virtual flow field data includes the velocity vector, pressure scalar, and flow direction unit vector of each grid point at each time step.

[0027] Particularly importantly, step S34 includes: Perform segment association on the hydraulic time-series path to obtain path stitching candidate nodes; Perform stitching logic fitting based on the path stitching candidate nodes to generate path logic stitching data; Construct adjacent boundary loops according to the path logic stitching data to obtain the boundary loop structure; Perform boundary state mapping on the boundary loop structure data and the hydraulic time-series path to generate flow boundary conditions; Perform virtual fluid field configuration processing based on the flow boundary conditions and the sewage data set of the treatment process to obtain virtual flow field data.

[0028] In this embodiment, the hydraulic time-sequence path is imported into the path segment analysis module, and the segmentation operation is performed on the node sequence in each path. The length of each segment is set to 1.5 meters. The continuous node sequence in the path is divided into multiple segments according to the distance condition. The spatial proximity determination is performed on the end nodes of adjacent segments. The determination threshold is set to 0.15 meters. If the Euclidean distance between the termination nodes of two segments is less than the threshold, the node is marked as a candidate node for path stitching. At the same time, the degree value of the node in the path topology graph is calculated, and the nodes with a degree value less than 2 are filtered out to avoid interference from isolated endpoints. The numbers and coordinate information of all qualified nodes are output as a path stitching candidate node table. The data structure includes candidate node numbers, spatial coordinates, adjacent path numbers, and segment vector directions. The path stitching candidate node table is input into the logic fitting module, and the stitching logic is constructed using a fitting algorithm based on the included angle of three-dimensional vector directions. For the path segment direction vectors between each pair of candidate nodes, the included angle value is calculated and it is determined whether the included angle is less than 30 degrees. If the angle requirement is met and the distance between the nodes does not exceed 0.If the distance between two nodes is 15 meters, it is determined that there is a potential logical stitching relationship between the pair of nodes, and a stitching logical connection edge is generated, which is recorded using graph structure data. After constructing the logical stitching graph, the Kruskal minimum spanning tree algorithm is used to traverse the edge set in the graph, and the minimum path set forming a loop structure or a cross path is extracted from it. The starting and ending coordinates of the nodes of each stitching edge, the stitching angle, and the stitching distance are marked. At the same time, a path logical stitching data set is generated, and the data structure includes logical edge number, starting node, target node, stitching vector, and spatial angle. The path logical stitching data is imported into the boundary loop generation module, and each logical stitching edge is traversed in turn. According to its connection relationship, the closed-loop path is extracted, and the boundary connection surface of each formed closed-loop path structure is spliced. In the three-dimensional space, a plane connection surface is constructed based on the positions of the two end points of the stitching edge, and the Delaunay triangulation method is used to generate the surface mesh. The minimum number of faces of each boundary loop is set to 8, and the maximum normal vector difference threshold is set to 20 degrees. Those exceeding this value do not participate in the construction of the closed-loop structure. The topological nodes, boundary grids, triangulation surfaces, and boundary vector direction information of each closed-loop structure are uniformly encapsulated into boundary loop structure data, which includes boundary loop number, the number of stitching logical edges included, node set, boundary grid file path, and direction tensor matrix. The pressure gradient change values at the upstream and downstream of the path segment are extracted from the boundary loop structure data and the hydraulic time series path data. The flow direction attribute is assigned to each boundary loop according to the principle that the flow direction is from the high-pressure area to the low-pressure area. The type of the boundary entrance is judged by the angle between the velocity vector of each node and the normal vector of the boundary loop. If the angle is less than 45 degrees, it is defined as the entrance boundary; otherwise, it is defined as the exit boundary. The entrance boundary is set with a constant pressure boundary condition, and the pressure value is set to the maximum pressure of the node in the path. The exit boundary is set as a constant flow velocity boundary, and the flow velocity value is taken as the average of the flow velocities of the downstream path nodes. All boundary conditions are uniformly stored in the flow boundary condition dictionary, which includes boundary number, boundary type, pressure value or flow velocity value, boundary normal vector, and corresponding path number. The sewage data set during the processing process and the above flow boundary conditions are input into the virtual fluid field configuration platform. The platform uses the Fluent, a multi-physics field joint modeling environment based on CFD, to construct a complete sewage treatment channel structural element model. The Tet mesh division method is used to generate the computational domain elements, the minimum element size is set to 0.05 meters, the maximum element size does not exceed 0.3 meters, the simulation time step is set to 0.05 seconds, and the total simulation time is 120 seconds. The fluid property is set as an incompressible Newtonian fluid, the density is set to 998 kg / m³, and the dynamic viscosity is 0.001 Pa·s. The k-ω SST turbulence model is used to process the flow in the near-wall region. The dissolved oxygen, pH, and COD values in the sewage dataset are introduced as physical variables in the model to participate in the flow field solution. After the boundary conditions are loaded, a three-dimensional transient simulation is performed, and the velocity vector, pressure distribution, and pollutant diffusion fields in the flow field at each time step are exported and uniformly output as structured data in HDF5 format, including the velocity, pressure, and sewage composition index values of each voxel at each moment.

[0029] Preferably, the particle motion tracking based on the sewage dataset and virtual flow field data in step S4, and the drawing of the particle settlement curve include: Analyze the fluid shear field based on the virtual flow field data, and infer the shear stress distribution data based on the fluid shear field; Identify the sewage particle parameters in the sewage dataset during the treatment process; Combine the shear stress distribution data and the sewage particle parameters, and establish the particle-fluid interaction relationship; Identify the particle force data according to the particle-fluid interaction relationship and the shear stress distribution data; Perform a time-domain advancement simulation based on the particle force data, and record the particle motion trajectory in the simulated advancement time-domain data; Extract the settlement rate characteristics in the particle motion trajectory, and analyze the time-varying settlement rate of the particles; Draw the particle settlement rate curve based on the time-varying settlement rate of the particles; Convert the particle settlement rate curve to displacement representation to obtain the particle settlement displacement curve; Identify the time relationship in the particle settlement displacement curve, and draw the particle settlement curve based on the time relationship and the particle settlement displacement curve.

[0030] In this embodiment, using the three-dimensional virtual flow field data, the entire processing interval is divided into equidistant cubic grids at an interval of 20 cm. The central difference method is used to perform directional difference operations on the velocity field data in each grid to obtain the velocity change rate of the fluid in each dimension direction. The numerical flow boundary modeling module in the finite volume method is called, and combined with the set dynamic viscosity parameter of the fluid (set to 0.001 Pa·s in this embodiment), the shear field is partitioned in the velocity change region, and a shear stress point distribution mapping matrix is established for all regions with large velocity gradients. Each matrix point is bound to the current position coordinates, the corresponding velocity change value, and the deduced local fluid stress value. Finally, a multi-dimensional data array containing three-dimensional coordinate indices and shear stress data values is output. The sewage particle image data and the particle physical parameter record table collected at the experimental site are read. The particle image data is obtained by the microscopic image acquisition system. The area covered by each frame of the image is 2 square millimeters, and the system collects 5 frames of data per second. After using the edge detection method to extract the particle contour boundary, the projected area and contour perimeter of each particle are obtained based on the pixel distribution in the binary image, and the particle shape factor is obtained by the image ratio method. At the same time, a high-precision laser particle size analyzer is used to statistically measure the particle size in the corresponding sample. The measured particle size range is between 1 micron and 15 microns, and the average density is determined by the particle sedimentation experiment. After the buoyancy method is measured, the value range is from 1.05 grams per cubic centimeter to 1.4 g. All data are uniformly numbered and stored in the particle basic attribute table, which includes particle size, density, shape factor, and image number. Taking the particle size, density, and shape factor of each particle as input parameters, according to its position in the shear stress field, the local stress value at the corresponding grid position is extracted from the shear stress distribution data. Then, based on the interaction law between the stress value and particle properties in space, a data model of three types of force terms is constructed, corresponding to the fluid resistance force, buoyancy force, and sinking tendency of the particle respectively. The data structure nesting method is used to construct an independent force vector data structure for each particle. At the same time, the position of the particle and its local stress environment are updated at each simulation time step, and the evolution path of its interaction is recorded. The interaction results between all particles and the fluid are stored in a three-dimensional array, which includes the relative force state, position index, and current shear field environment number of each particle at each time point. Indexed by the particle number, the position data of each particle at each simulation time step is traversed, and the shear stress value of the grid where it is located and the velocity difference of the adjacent grid are extracted by the spatial position correspondence method. Using these two data and the combination of particle size and density in the particle property data table, the current fluid action intensity term of the particle is retrieved in the data interaction model, and at the same time, the gravity influence term and the buoyancy vector term are loaded, so as to construct a complete three-dimensional force data vector, recording the direction and intensity of all force vectors acting on the particle at this time step. Finally, it is output as a continuous-time force table for each particle, including the total force magnitude, direction unit vector, shear field number, and historical path information at each time step, which serves as the driving source for the next trajectory advancement. Set each simulation step size to 0.In 0.02 seconds, read the force direction and magnitude of each particle at the initial moment from the force data table. According to the force state and velocity direction at the previous time point, calculate the new position of the particle step by step. Record the new position each time it is advanced and store it together with the force value in the trajectory table. The system sets the maximum advancement time for each particle during the entire simulation process to 30 seconds, generating no more than 1500 trajectory points for each particle. The trajectory is output in CSV format, with each line being a time step containing the coordinate points, velocity direction, and force intensity identifier at that time. The trajectories of all particles form a set of three-dimensional space motion paths. Extract the displacement change amount of the particles in the vertical direction from the above trajectory set, sample at each time step, calculate the distance difference in the vertical direction between the coordinate points of adjacent two frames, obtain the sinking velocity for each time step, and then perform a sliding window smoothing process on the velocity sequence. Average the values once every five frames as a window to reduce the impact caused by single-point abnormal fluctuations. Organize the sinking velocity sequence of each particle over the entire time period. Analyze the trend of the particle sinking velocity changing with time through the maximum value, minimum value, and change slope of this sequence. Finally, output the time series data table of the sinking rate of each particle, mark the start and end times of the acceleration section, stable section, and fluctuation section, import the sinking rate data table into the graph drawing program, set the abscissa as the time step length, and the ordinate as the sinking velocity at the corresponding moment. Use a line graph to draw the sinking velocity change curve of each particle, display the curves in groups, and use different colors to distinguish particles in different particle size ranges. Mark the legend as "particle size less than 5 microns", "particle size between 5 and 10 microns", "particle size greater than 10 microns", set the image size to 12 centimeters per side, and the resolution to 300 dots per inch. Read the sinking rate curve data of each particle, perform an accumulative calculation on the velocity data and time step length for each time step, obtain the vertical displacement of the particle from the initial moment to the current time point, arrange the displacement values in chronological order to generate a displacement sequence, and store it as the particle displacement time table. Draw the particle sinking displacement curve, with the abscissa being time and the ordinate being the displacement distance. The curve trend shows the vertical distance change trajectory of the particle during the entire sinking process, and the unit is unified as millimeters. Identify three types of key time points from the particle sinking displacement curve, namely the starting point where continuous sinking begins, the starting point of entering the velocity stable section, and the termination point of the sinking process. Mark these time points and the corresponding displacement amounts on the curve graph, and cut the curve into segments to draw a complete sinking trend graph. The sinking curve is presented in a smooth curve manner, marking the start and end times and sinking distances of each segment. Compare the sinking curves of particles with different particle sizes and output them as an overlay chart containing multiple curves for use in sinking behavior clustering, feature extraction, or model fitting.

[0031] Preferably, the sedimentation offset mapping fitting based on the preset ideal sedimentation situation and particle sedimentation curve in step S4 includes: Drawing an ideal sedimentation curve based on the preset ideal sedimentation situation; Project and compare the ideal deposition curve and the particle sedimentation curve, and construct a sedimentation difference matrix; Calculate the offset of each particle trajectory based on the sedimentation difference matrix, and integrate it into a set of particle trajectory offsets; Determine the standard particle motion trajectory according to the ideal deposition curve; Overlay the set of particle trajectory offsets and the standard particle motion trajectory, and perform offset correction to obtain a set of offset-corrected trajectories; Perform spatio-temporal distribution clustering on the set of offset-corrected trajectories to obtain trajectory correction point data; Identify key offset patterns based on the trajectory correction point data; Predict the offset of the particle sedimentation curve through the key offset patterns, and generate particle sedimentation trajectory offset data.

[0032] In this embodiment, according to the preset ideal deposition situation, the basic parameters required for the ideal deposition curve are first defined. These parameters include the particle size, density of the particles, and the dynamic characteristics of the fluid. In this embodiment, the particle size range is set to be from 5 microns to 10 microns, the density is 1.2 grams per cubic centimeter, and the dynamic viscosity of the fluid is set to 0.0012 Pascal-seconds. In addition, a deposition velocity model needs to be set. This model assumes that the deposition velocity of particles in a static fluid environment follows Stokes' law. During the calculation of the particle velocity, it is assumed that the temperature of the fluid is 25°C and the density of water is 1 gram per cubic centimeter. Using these set physical parameters, calculations are carried out according to the deposition rate formula to draw an ideal particle deposition curve. This curve shows the relationship between the sedimentation velocity of particles and time under ideal conditions and is used as a reference benchmark in the subsequent correction process. Through numerical simulation, the sedimentation curves of each particle in the actual sewage treatment process are obtained. These sedimentation curves are constructed based on the aforementioned particle force data and the fluid shear field environment. The time span of the sedimentation curve and the change in the sedimentation position of the particles are recorded within 30 seconds. For each time step, data on the relationship between the sedimentation displacement of the particles in the vertical direction and the time point is generated. In the particle sedimentation curve, the deviation is measured by comparing the real-time position of each particle with the ideal deposition curve. A sedimentation difference matrix is constructed. The rows and columns of this matrix represent the particle numbers and time steps respectively. The elements in the matrix are the difference values between the actual sedimentation displacement of the particles at this time step and the corresponding displacement of the ideal deposition curve. The matrix details the offset of each particle at each moment. Each row in the difference matrix shows the offset state of the particle at different time points. The larger the value in the matrix, the more obvious the difference between the actual sedimentation and the ideal deposition; conversely, it indicates a smaller difference. The results of the sedimentation difference matrix are corresponded to the trajectories measured for each particle in the actual treatment process. The difference between the actual sedimentation position of each particle at each time point and the ideal position on the ideal deposition curve is compared one by one, and the offset of each particle is calculated. The offset is obtained by subtracting the corresponding position on the ideal deposition curve from the actual sedimentation position. The offsets of all particles at each time step will form a set of particle trajectory offsets. Each offset data point consists of the particle number, time step, and offset. The offset set will record the trajectory correction data of all particles during the simulation process. The offset data of all particles is integrated. According to the ideal deposition curve, the ideal motion trajectory of each particle is calculated and determined at each time step. A standard trajectory during the deposition process is set. The standard trajectory assumes that all particles sink stably in the same fluid environment according to Stokes' law and the dynamic characteristics of the particles. Based on this standard trajectory, the ideal motion trajectory of each particle is defined. This trajectory serves as a specification for the sinking path of the particles, unaffected by external disturbances and only considering the natural sedimentation velocity in the fluid environment. Each time point on the ideal deposition curve corresponds to a specific particle sinking position. In this process, the offset data of the particles is only determined by the fluid dynamics and the physical characteristics of the particles and is independent of the minor disturbances in the external flow field.The obtained canonical particle motion trajectory is the sinking path of each particle under ideal conditions, which is the reference standard for subsequent offset correction and spatiotemporal distribution cluster analysis. By matching the particle trajectory offset set with the canonical particle motion trajectory one by one, the offset is added to the position in the canonical trajectory at each time step to obtain the corrected trajectory of each particle. The calculation formula of the offset corrected trajectory is: corrected position = canonical position + offset. The corrected trajectory set will include the corrected position of each particle at each time step. The corrected position represents the offset correction effect of the particle in the actual sewage treatment process due to external factors. The offset corrected trajectory set of all particles will serve as the basis for subsequent analysis to further calibrate the particle motion trajectory to ensure that the particle behavior in the sewage treatment process conforms to the actual fluid environment. The corrected particle trajectory data will be subjected to spatiotemporal distribution analysis. The cluster analysis method is used to spatially cluster the corrected positions of all particles at different time steps, analyze the distribution law of the particle corrected position, and use K-mean The s (K-means) clustering algorithm divides the particle trajectory correction points into several groups according to their similarity. Each group of particles shows similar deviation trends under similar spatiotemporal conditions. The clustering algorithm automatically groups the spatiotemporal distribution data of particles by setting a predetermined number of clusters. The particles in each cluster show relatively consistent sedimentation behavior. The clustering result forms a trajectory correction point data set, which records the trajectory correction points of particles belonging to the same spatiotemporal cluster in the corrected motion trajectory of the particles. The clustered data set will provide trend information of the particle trajectory. According to the results of clustering analysis, for each trajectory correction point data, Pattern recognition uses data mining technology to extract the spatiotemporal characteristics of each particle trajectory and identify common offset patterns of particle motion trajectories under different sedimentation conditions. During the pattern recognition process, the data of each corrected trajectory is first smoothed to remove noise and outliers. Then, the regularity of the corrected trajectory is analyzed to identify time periods or regions with significant offsets, and these key offset patterns are recorded. Offset patterns can be classified based on factors such as particle settling rate changes and settling stability. Each pattern corresponds to a specific type of settling behavior. Identifying these key offset patterns helps to accurately predict and control the subsequent particle settling process. The identified key offset patterns are used to predict the offset of the particle settling curve. At each time period of particle settling, the particle settling curve is corrected based on the identified offset patterns, and the particle settling position offset in the future time step is estimated to generate new particle settling trajectory offset data. The prediction results are calculated based on the initial state of the particle, the fluid environment, and the key offset patterns. The generated particle settling trajectory offset data provides a correction value for the particle position at each time step. Based on this correction data, the settling behavior of particles in the sewage treatment process can be adjusted and optimized in real time. ,

[0033] Preferably, reconstructing the sedimentation tank architecture according to the sedimentation tank morphological parameters in step S5 and determining the reference inclination angle of the inlet deflector based on the sedimentation tank architecture includes: Constructing an initial sedimentation tank framework based on the sedimentation tank morphological parameters; Identifying key structural nodes in the initial sedimentation tank framework and connecting the nodes based on the key structural nodes to generate a sedimentation tank skeleton network; Identifying the geometric skeleton of the inlet area in the sedimentation tank skeleton network; Calculating the position and angle of the deflector according to the geometric skeleton of the inlet area to generate a deflector parameter set; Determining the reference inclination angle of the inlet deflector based on the geometric framework set of the inlet area and the deflector parameter set.

[0034] In this embodiment, the size of the sedimentation tank is set according to the specific morphological parameters of the sedimentation tank, including basic information such as the length, width, height of the tank body and the slope of the tank bottom. In this embodiment, it is assumed that the sedimentation tank is 20 meters long and 10 meters wide, the slope of the tank bottom is 5 degrees, and the water depth in the tank is 3 meters. Using these parameters, the initial three-dimensional framework of the sedimentation tank is first constructed, and the geometric shape of the tank body is modeled using CAD (computer-aided design) software. By setting the inclination of the tank bottom and the boundary of the tank wall, a geometric model of the sedimentation tank is established. This model provides the basis for the overall shape of the sedimentation tank. At the same time, it is necessary to ensure that each boundary condition is consistent with the given The physical parameters of the sedimentation tank are matched, and the stability of the pool body and the fluid flow characteristics are taken into consideration during the construction process. In the initial framework of the sedimentation tank, the nodes with key structural significance are identified. These nodes mainly include the pool wall connection points, the intersection points between the pool bottom and the pool wall, and the interface points of the guide plate installation position. In this embodiment, the detailed three-dimensional point cloud data of the sedimentation tank frame is obtained by 3D scanning technology, the point cloud data is processed, the key structural nodes are screened out, and the relationship between the nodes is analyzed by geometric modeling tools to determine the connection method and relative position of each node. Then, based on these key nodes, the nodes are connected to construct the sedimentation tank. The skeleton network of the sedimentation tank is composed of various nodes and connecting lines, which represent the structural framework of the sedimentation tank. By analyzing the geometric features in the skeleton network of the sedimentation tank, especially the areas related to the inlet flow rate and flow direction, the geometric skeleton of the inlet area in the skeleton network of the sedimentation tank is identified. In this step, the position of the water inlet is first set. It is assumed that the water inlet is at the entrance of the sedimentation tank. Then, according to the geometric structure of the sedimentation tank and the principle of fluid dynamics, the fluid flow direction of the water inlet area is calculated. The fluid flow is simulated using CFD (computational fluid dynamics) simulation software to obtain the fluid streamlines, and then the factors affecting the inlet flow rate and flow direction are identified. Key geometric features: The geometric skeleton of the water inlet area is determined by outlining the water inlet streamlines and area boundaries in three-dimensional space to determine the actual geometric shape of the water inlet area. The geometric skeleton reflects the main path of fluid flow and the shape of the area. Based on the geometric skeleton of the water inlet area and the requirements of fluid flow, the appropriate deflector position and angle are calculated. In this step, the geometric skeleton information of the water inlet area is used to first determine the installation position of the deflector. The selection of the installation position is based on the flow direction and flow velocity requirements of the water inlet. Then, the optimal angle between the deflector and the water inlet flow is determined through fluid dynamics calculations. Fluid simulation software (such as ANSYS Fluent) is used to simulate the deflector design. By simulating the fluid flow at different angles of the deflector, the optimal angle is obtained to make the fluid flow more uniform, thereby achieving an ideal fluid distribution effect. Finally, based on the simulation results, the specific parameters of the deflector, including its position, angle, size, etc., are set. These deflector parameter data constitute the deflector parameter set. By analyzing the geometric framework of the water inlet area and the deflector parameter set, based on the fluid dynamics calculation and simulation results, the base inclination angle of the water inlet deflector is determined.First, comprehensively analyze the geometric framework of the water inlet area and the parameter set of the flow deflector, calculate the velocity and flow direction changes of the fluid before and after the flow deflector, and use the CFD simulation results to optimize the angle of the flow deflector so that the fluid can be evenly distributed into the sedimentation tank along a predetermined trajectory. Finally, determine the reference inclination angle. In this embodiment, through experiments and simulations, it is obtained that the reference inclination angle of the flow deflector should be 30 degrees. This angle can effectively guide the fluid into the sedimentation tank in practical applications and minimize unnecessary turbulence. Then, through multiple simulations and adjustments, confirm the rationality and effect of this reference inclination angle, and obtain the final design value of the reference inclination angle of the flow deflector.

[0035] Preferably, predicting the sedimentation imbalance abnormal state based on the reference inclination angle of the inlet flow deflector and the particle sedimentation offset data in step S5 includes: Simulate the long-term sedimentation process through the particle sedimentation trajectory offset data and the sedimentation tank morphology parameters, and determine the cumulative sediment distribution based on the long-term sedimentation process; Calculate the sediment load distribution according to the cumulative sediment distribution, and generate the structural load distribution data; Map the structural load distribution data into the initial sedimentation tank framework, and perform finite element stress analysis to generate the stress distribution data; Identify the stress concentration points of the flow deflector based on the stress distribution data and the sedimentation tank morphology parameters, and analyze the stress state of the flow deflector; Predict the deformation trend of the flow deflector according to the stress state of the flow deflector; Calculate the change rate of the inclination angle over time according to the deformation trend of the flow deflector, and map it into the dynamic evolution curve of the flow deflector inclination angle; Predict the future inclination angle change data through the dynamic evolution curve of the flow deflector inclination angle and the reference inclination angle of the inlet flow deflector; Perform surface fitting on the future inclination angle change data to obtain the inclination angle change fitting surface; Segment the inclination angle change fitting surface by time series stability to generate the change trend fracture segments; Map the corresponding flow deflector structure response area based on the change trend fracture segments; Deduce the flow field deformation state based on the corresponding flow deflector structure response area and the particle sedimentation offset data; Perform gradient analysis on the flow field deformation state, calculate the flow velocity change rate of the corresponding flow deflector area, and generate the flow field disturbance intensity; Compare the intensity based on the preset critical sedimentation condition and the flow field disturbance intensity, and predict the sedimentation imbalance abnormal state based on the intensity comparison result.

[0036] In this embodiment, the long-term simulation of the sedimentation process is carried out by using the morphological parameters of the sedimentation tank and the offset data of the particle sedimentation trajectory. Assuming that the total depth of the sedimentation tank is 3 meters and the width is 10 meters, physical parameters such as the particle size and density (e.g., the particle diameter is 0.5 mm and the density is 2.5 g / cm³), as well as information such as the water flow velocity in the sedimentation tank, are combined with the sedimentation kinetic equation. Using the Particle Tracking Simulation technology, the sedimentation process of particles at different time points is simulated. The sedimentation trajectory of particles in the tank is obtained through the hydrodynamic model, and the distribution of each particle at the bottom of the tank is gradually accumulated to obtain the cumulative distribution data of the sediment. This data shows the thickness and distribution density of the sediment in different areas of the tank. The simulation results help to accurately reflect the sedimentation dynamics of the sedimentation tank and ensure that the simulation matches the actual sedimentation environment. The time span of the simulation is 6 months, and the particle position is updated once a day. The distribution of the sediment is obtained through multiple iterative simulations. Based on the cumulative distribution of the sediment, the distribution of the sediment load is calculated. Assuming that the density of the sediment is 1.5 g / cm³, the thickness of the sediment in different areas at the bottom of the tank is 0.5 meters, 1.2 meters, 0.8 meters, the sediment load calculation formula is Load = Density × Acceleration due to gravity × Sediment volume. By applying this formula in different areas of the sedimentation tank, the sediment load of each area is obtained. Further, based on the load distribution data, the structural load distribution data is generated. Considering the influence of the sediment load on the sedimentation tank structure and combining the structural characteristics of the tank body, the load distribution data is matched with the stress states at positions such as the pool wall and the pool bottom to generate a sediment load distribution map. This map can be used for further analysis of the structural stress conditions and provide necessary data for subsequent finite element analysis. During the calculation process, the load distribution not only depends on the thickness of the sediment but is also closely related to the flow velocity of the water and the shape of the tank body. The load changes caused by the irregular shape of the tank bottom are considered during actual calculation. The structural load distribution data generated in the previous step is mapped into the initial sedimentation tank framework, and the finite element analysis (Finite Element Analysis, FEA) method is used to perform stress analysis on the sedimentation tank. First, the load data is input into the finite element analysis model to establish a finite element model of the sedimentation tank. Assuming the thickness of the pool wall of the sedimentation tank is 30 cm and the thickness of the pool bottom is 50 cm, the sediment load distribution is applied to each unit by refining the mesh, and the stress calculation is performed to obtain the stress distribution data. Stress simulation is carried out using software such as ANSYS. By performing a detailed mesh division on the structure of the sedimentation tank, the stress conditions of the sedimentation tank under different loads are simulated. The results show the stress variation conditions in different areas of the sedimentation tank structure, especially the stress concentration areas at the pool wall and the pool bottom. The simulated stress data provides an accurate basis for subsequent analysis. Based on the stress distribution data and the sedimentation tank shape parameters, the stress concentration points of the guide plate are identified. First, the geometric shape and position of the guide plate are input into the stress analysis model. Assuming the guide plate is located in the inlet area and the inclination angle is 30 degrees, the stress distribution on the guide plate under different loading conditions is calculated by performing mesh division on the surface of the guide plate, and the areas with the maximum stress are identified. These areas are the stress concentration points. During this process, the stress field data is used to determine the maximum stress area of the guide plate. Assuming the maximum stress value is 200 MPa, the stress field is used to analyze the stress conditions of the guide plate, and further analyze whether there is a risk of damage or deformation of the guide plate. In response to the changes in the stress concentration points of the guide plate, the design is further optimized to reduce the risk brought by stress concentration. According to the aforementioned stress state of the guide plate, the deformation trend of the guide plate is predicted. First, through the stress-strain relationship formula (for example... , where is the stress, is the elastic modulus, To calculate the deformation degree of the deflector under different stresses (for strain), the maximum stress value obtained from the previous calculation and the elastic modulus of the deflector material are used to obtain the deformation degree. Assuming the material of the deflector is steel with an elastic modulus of 210 GPa, the calculation results show that under the maximum stress, the deformation of the deflector is 0.001 mm. Based on this data, the deformation trend that the deflector will exhibit during long-term use can be speculated. The step-by-step cumulative stress analysis method is adopted to predict the long-term deformation of the deflector, and finally the change trend of its deformation degree over time is calculated. According to the deformation trend of the deflector, the change rate of its tilt angle is calculated. Assuming the initial tilt angle of the deflector is 30 degrees, based on the deformation trend, the tilt angle change rate is set to 0.01 degrees per hour. After a certain period of time (such as 12 hours) of calculation, the tilt angle change of the deflector is 0.12 degrees. By accumulating these change rate data, the dynamic evolution curve of the deflector tilt angle is obtained. This curve shows the tilt angle change of the deflector during long-term use. Using this curve, the state of the deflector can be monitored in real time to predict its change trend within a certain period in the future. According to the dynamic evolution curve of the deflector tilt angle and the reference tilt angle of the inlet deflector, the change data of the future tilt angle is predicted. First, the reference tilt angle is compared with the dynamic evolution curve. The reference tilt angle is set to 30 degrees. Using the calculated change rate and the accumulated change data, the expected change of the deflector tilt angle within a certain period in the future is obtained. Assuming that after 12 hours, the expected tilt angle change of the deflector is 0.1 degree, then it is predicted that within the next 12 hours, the tilt angle of the deflector will change to 30.1 degrees. Based on this prediction, the angle of the deflector can be adjusted to avoid excessive deviation. The change data of the future tilt angle is input into the surface fitting algorithm, and polynomial fitting (such as quadratic polynomial) is used to process the data. The fitting function is set as , where is the time variable, , , are the fitting coefficients. The fitting coefficients are calculated by the least squares method. Assuming the obtained fitting surface is , this surface represents the variation trend of the tilt angle over time. Using this fitted surface, the tilt angle of the deflector at a certain future moment can be accurately predicted. According to the fitted tilt angle change surface, time-series stability analysis is carried out. First, the fitted surface is divided into multiple time periods (for example, each hour is a time period), and the variation trend within each time period is analyzed. If the tilt angle changes significantly within a certain time period, it is marked as a fracture segment of the variation trend. Suppose that between the 5th hour and the 6th hour, the tilt angle change rate is high, and this time period is determined as a fracture segment of the variation trend. Based on these fracture segments, the operation strategy of the deflector can be adjusted to avoid excessive tilt or uneven flow field. According to the fracture segments of the variation trend, the structural response area of the deflector is mapped. First, according to the fracture segments of the tilt angle variation trend, the pressure distribution area on the deflector is determined. By analyzing the pressure distribution and deformation data, the response area of the deflector is deduced. Suppose that between the 6th hour and the 7th hour of the variation trend, the force application points and deformation points of the deflector are concentrated in the central area of the deflector, and this area should be regarded as the key response area of the deflector. Through further monitoring and analysis, ensure that this area does not exceed the set structural bearing range. Using the structural response area of the deflector and the particle sedimentation offset data, the deformation state of the flow field is deduced. First, based on the known particle sedimentation trajectory offset data, combined with the stress and deformation data of the deflector, using a fluid dynamics model (such as the Navier-Stokes equation), the variation of the flow field is simulated. Suppose the sedimentation velocity of the particles in the flow field is 0.2 m / s. Through fluid simulation calculation, the degree of flow field deformation is obtained, the stability of the flow field is analyzed, and the angle of the deflector is adjusted according to this deformation state to maintain the uniformity and stability of the flow field. Gradient analysis is carried out on the deformation state of the flow field, and the velocity change rate is calculated. Using the velocity field data of the flow field, the velocity change rate is set to 0.1 m / s². Based on this data, a flow field disturbance intensity map is generated. The intensity of the velocity disturbance is obtained through fluid simulation calculation, and combined with the force data of the deflector, the distribution of the disturbance intensity is generated. This data can be used for further flow field optimization and deflector adjustment. Based on the preset critical sedimentation conditions (such as the sediment thickness exceeding 0.5 meters) and the flow field disturbance intensity data, intensity comparison is carried out. Suppose that under the critical sedimentation conditions, the load of the sediment reaches a specific value, and the sediment imbalance abnormal state is predicted by combining the flow field disturbance intensity. If the comparison result shows that the flow field disturbance intensity exceeds the set threshold and the sediment thickness exceeds 0.5 meters, it indicates the occurrence of the sediment imbalance state. Through this prediction mechanism, measures can be taken in advance for adjustment. Based on the above-mentioned prediction results of the sediment imbalance abnormal state, sediment imbalance warning data is generated. Suppose the warning condition is that the flow field disturbance intensity exceeds a certain threshold and the sediment thickness reaches the preset condition. The system will automatically generate warning data and display a warning graph through a visualization interface, display potential abnormal areas, and provide specific adjustment plans, such as adjusting the inlet flow rate, changing the tilt angle of the deflector, etc.

[0037] Particularly important is to perform an intensity comparison based on preset critical sedimentation conditions and the intensity of flow field disturbances, and predict abnormal sedimentation imbalance states based on the intensity comparison results, including: Perform a point-to-point mapping comparison between the preset critical sedimentation conditions and the intensity of flow field disturbances, construct a spatial distribution correspondence relationship, and thus determine the intensity ratio of each region; Set a threshold range of disturbance intensity based on the intensity ratio of each region, and screen high-risk regions that exceed the threshold based on the threshold range of disturbance intensity; Conduct spatial clustering analysis on high-risk regions and determine potential imbalance source points; Simulate the state evolution of potential imbalance points according to the intensity of flow field disturbances to generate a simulated state evolution path; Predict abnormal sedimentation imbalance states in the sewage treatment process based on the simulated state evolution path.

[0038] In this embodiment, critical settlement conditions are set, including the maximum allowable thickness of the sediment, the minimum stable settlement velocity, etc. These conditions are obtained from the historical data and experimental data of actual sewage treatment facilities. Then, spatial distribution data of the flow field disturbance intensity is acquired. The calculation of the flow field disturbance intensity is based on the difference between the velocity field and the flow variation. For example, the flow field disturbance values in different regions are obtained through a CFD (Computational Fluid Dynamics) model. The comparison between the flow field disturbance intensity of each region and the critical settlement conditions is carried out by establishing a point-to-point mapping relationship. The comparison is made according to the magnitude of the velocity change and the settlement rate to obtain the intensity ratio of each region. This ratio represents the degree of adaptation between the flow field disturbance and the settlement conditions. The higher the ratio, the greater the risk of sediment deposition imbalance in that region. A threshold range is set, which is obtained by analyzing historical data and experimental data. Usually, the upper and lower limits of the flow field disturbance intensity are determined through statistical analysis methods of the velocity field (such as cluster analysis). For example, the threshold of the flow field disturbance intensity is set to 0.5 - 1.5 m / s (meters per second). By comparing the intensity ratio of each region with the set threshold range for screening, those regions that exceed the threshold range are identified. These regions are considered high-risk regions, indicating that in these regions, the flow field disturbance intensity will be too high, resulting in excessive sediment settlement or imbalance. Through a data visualization tool, such as a GIS (Geographic Information System) platform, these high-risk regions are marked with different colors. Using a spatial clustering analysis method, such as the K-means algorithm or DBSCAN (Density-Based Spatial Clustering of Applications with Noise), spatial clustering processing is performed on the high-risk regions. The purpose of this process is to identify potential imbalance source points within the region. Through the spatial clustering method, the regions with excessive flow field disturbance intensity are clustered and analyzed to find the aggregation points in these regions. The aggregation points are the potential imbalance source points. During the clustering process, certain parameters need to be set, such as the clustering radius (for example, 50 meters) and the minimum number of clustering points (for example, 3 data points). The output of the clustering result is a set of potential imbalance source points, which will be used as the basis for the next state evolution simulation. A numerical simulation method is used to simulate the state evolution of the potential imbalance points. During the simulation process, first, the initial conditions of the potential imbalance points are set, such as the sediment thickness, the flow field disturbance intensity, the settlement rate, etc. By establishing a mathematical model based on physical laws (such as the Navier-Stokes equation), combined with the influence of the disturbance intensity and the settlement conditions, the state change process of the imbalance points is simulated. For example, in the simulation, an initial flow field disturbance intensity of 1.2 m / s is set, and it is assumed that the initial value of the sediment thickness is 0.3 meters. Using a numerical calculation tool for dynamic evolution simulation, the state change path of this point within a future time period (such as within 24 hours) is calculated. The simulation results provide the change trajectory of the sediment thickness and the flow field disturbance intensity of the potential imbalance points at different time points, forming a clear state evolution path. Based on the generated simulated state evolution path, combined with the preset critical settlement conditions, the prediction of the abnormal state of sediment deposition imbalance is carried out.Suppose that in the simulation path, the sediment thickness at a potential imbalance point exceeds the set critical thickness (e.g., 0.5 m) within 24 hours, and the flow field disturbance intensity exceeds 0.8 m / s, then this point is determined to be in an abnormal state of sediment imbalance. According to the evolution results of the simulation path, the system will output a risk prediction report, indicating the time and area where sediment imbalance is likely to occur, and then provide adjustment plans, such as increasing or decreasing the inflow water flow rate, adjusting the pump intensity, etc., to ensure the stable operation of the sewage treatment process. The data in the simulation path will be updated and adjusted through the real-time monitoring system.

[0039] Preferably, step S6 includes the following steps: Step S61: Perform time series decomposition on the abnormal state of sediment imbalance and extract key time nodes to determine the abnormal occurrence time series; Step S62: Based on the abnormal occurrence time series, perform retrospective analysis on the abnormal over-limit parameters in the initial sewage treatment process model; Step S63: Perform historical trend analysis on the abnormal over-limit parameters through the preset historical model parameters and identify the abnormal change pattern; Infer the parameter drift characteristics according to the abnormal change pattern; Step S64: Based on the parameter drift characteristics, perform model parameter correction on the abnormal over-limit parameters, and perform parameter reconfiguration on the initial sewage treatment process model based on the corrected model parameters to generate a corrected sewage treatment process model.

[0040] In this embodiment, based on data such as sediment thickness and flow field disturbance intensity monitored during the sewage treatment process, time series decomposition is performed on the time series data of the entire sewage treatment process. The time series decomposition method uses the Empirical Mode Decomposition (EMD) algorithm. The EMD algorithm decomposes the original time series data into multiple Intrinsic Mode Functions (IMFs), and each IMF reflects the signal fluctuations at different time scales. For the abnormal state of sediment imbalance, the time window of the abnormal signal is identified by extracting the IMF components that change drastically at specific time points, and the key time nodes when the abnormality occurs are identified. These time nodes are the moments when the sediment thickness exceeds the set critical value or the flow field disturbance intensity shows abnormal fluctuations. By comparing the IMF components of multiple time series with historical data, the core time points affecting sediment imbalance can be accurately extracted. For example, if the monitored flow field disturbance intensity suddenly jumps from 0.5 m / s to 1.5 m / s and lasts for a certain period of time, it can be calibrated as a potential abnormal occurrence moment. According to the abnormal occurrence time series extracted in the previous step, trace back to the relevant parameters in the sewage treatment process model, and use the regression analysis method to evaluate the model input parameters at each moment, focusing on key parameters such as flow rate, flow velocity, and sediment thickness in the treatment tank. By comparing with historical data, analyze the change trends of these parameters and identify which parameters exceed the normal fluctuation range before or after the abnormal time node. For example, during the process of the flow velocity fluctuating from 0.2 m / s to 0.7 m / s, if it is accompanied by abnormal thickening of the sediment thickness, it indicates the impact of the increase in flow velocity on sediment imbalance. In the backtracking process, by comparing the set warning threshold and the actual parameter change value, preliminary abnormal over-limit parameters are obtained. For example, if the sediment settlement rate exceeds the set maximum settlement rate during a certain period, it is confirmed that this parameter is the root cause of the imbalance. When performing historical trend analysis on the abnormal over-limit parameters, the sliding window method is used to smooth the historical data. The window size is set to 10 days, and the mean and standard deviation of each parameter are calculated within the window to analyze the fluctuation range of the parameter in the historical data. By comparing the trends in historical data, identify whether there is an abnormal change pattern in the parameter value before the moment of abnormality. For example, if the sediment settlement rate has been steadily increasing in the past week and reaches the preset critical value during the abnormality, it indicates that there is a persistent drift in this parameter. By identifying the drift pattern and combining historical data, the drift characteristics of the parameter can be inferred. Usually, the drift pattern can be described by mathematical models such as linear trend and exponential trend. Suppose a parameter such as the settlement rate shows a linear increase, then this trend can be represented by the equation (where is the parameter value to be predicted, is the initial value of the parameter at = 0, is the growth rate, To fit with time and then predict the future change trend of parameters. When correcting abnormal over-limit parameters according to the parameter drift characteristics obtained in the previous step, the Least Squares Method is used to fit and correct the drift parameters. If the sedimentation rate parameter has a linear growth drift trend, the trend line obtained by fitting is corrected through the Least Squares Method to obtain new corrected parameters. For example, a (growth rate) in the drift trend is adjusted to make it closer to the actual observed value. The corrected parameters are used as the input of the initial sewage treatment process model, and then parameter reconfiguration is performed through an optimization algorithm. The optimization algorithm uses the Genetic Algorithm or the Particle Swarm Optimization (PSO). According to the historical data and the corrected parameter values, key variables in the model, such as flow velocity, sediment thickness, oxygen concentration in the reaction tank, etc., are adjusted. Finally, a new sewage treatment process model is generated, which can accurately reflect the actual performance of the sewage treatment system according to the historical parameter correction and ensure its optimal operating state during the treatment process.

[0041] The present invention also provides a system for correcting parameters of a sewage treatment process model, which is used to execute the method for correcting parameters of a sewage treatment process model as described above. The system for correcting parameters of a sewage treatment process model includes: A data acquisition module, which is used to collect multi-point sewage treatment facility sensing data through a sewage treatment sensing network; extract the sewage pipeline topological structure of the sewage treatment sensing network; project the multi-point sewage facility sensing data into the sewage pipeline topological structure, and mark the corresponding position information; A model construction module, which is used to construct an initial sewage treatment process model based on the corresponding position information and multi-point sewage facility sensing data; extract the sewage treatment process sewage data set in the initial sewage treatment process model; An environmental simulation module, which is used to perform sewage treatment environmental simulation according to the sewage treatment process sewage data set to generate a simulated sewage treatment process environment; simulate virtual flow field data based on the simulated sewage treatment process environment and the sewage treatment process sewage data set; A particle tracking module, which is used to perform particle motion tracking based on the sewage treatment process sewage data set and the virtual flow field data, and draw a particle sedimentation curve; perform sedimentation offset mapping fitting based on the preset ideal sedimentation situation and the particle sedimentation curve to generate particle sedimentation trajectory offset data; A sediment analysis module, which is used to obtain sedimentation tank shape parameters; reconstruct the sedimentation tank architecture according to the sedimentation tank shape parameters, and determine the reference tilt angle of the inlet guide plate based on the sedimentation tank architecture; predict the sediment imbalance abnormal state based on the reference tilt angle of the inlet guide plate and the particle sedimentation offset data; A parameter correction module is used to retrospectively analyze abnormal over-limit parameters in the initial sewage treatment process model through the abnormal state of sedimentation imbalance; based on the abnormal over-limit parameters, the model parameters of the initial sewage treatment process model are corrected to generate a corrected sewage treatment process model.

[0042] Through the application of the data acquisition module, the present invention realizes the comprehensive monitoring of sewage treatment facilities and the acquisition of real-time data, ensures the accuracy and timeliness of the data. The extracted sewage pipeline topological structure provides a clear spatial framework for subsequent analysis. The data projection and position information marking enhance the traceability of the data, effectively supporting the basis of model construction. The model construction module generates an initial sewage treatment process model based on multi-point sensing data, providing an accurate description of the treatment process. The extracted sewage data set provides a necessary basis for model optimization. The simulated sewage treatment process environment generated by the environmental simulation module ensures the effective simulation of the actual treatment situation. The simulation of virtual flow field data provides important support for particle motion analysis. The particle tracking module reveals the behavioral characteristics of particles during the treatment process by plotting the sedimentation curve. The generation of sedimentation offset mapping fitting provides a data basis for the optimization of the sedimentation process. The sedimentation tank shape parameters obtained by the sedimentation analysis module provide a scientific basis for the reconstruction of the sedimentation tank architecture. Based on the prediction of the reference tilt angle of the baffle plate, the abnormal state of sedimentation imbalance can be identified in a timely manner. The parameter correction module clarifies the direction of model parameter adjustment through the retrospective analysis of abnormal over-limit parameters. The sewage treatment process model after model parameter correction significantly improves the accuracy and applicability of the model, provides reliable theoretical support for the optimized operation of sewage treatment facilities, overall improves the sewage treatment effect and treatment efficiency, and promotes the intelligent and refined development of sewage treatment technology.

[0043] Therefore, from any perspective, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, it is intended to cover all changes falling within the meaning and scope of the equivalent elements of the application documents within the present invention.

[0044] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for correcting parameters of a sewage treatment process model, characterized in that: The following steps are involved: Step S1: collecting sensor data of multiple sewage treatment facilities through a sewage treatment sensor network; extracting the sewage pipe topology of the sewage treatment sensor network; projecting the sensor data of multiple sewage facilities into the sewage pipe topology and marking the corresponding location information; Step S2: constructing an initial sewage treatment process model based on corresponding location information and multi-point sewage facility sensor data; extracting a treatment process sewage data set in the initial sewage treatment process model; Step S3: performing sewage treatment environment simulation based on the sewage data set during treatment to generate a simulated sewage treatment process environment; Simulate virtual flow field data based on the simulated sewage treatment process environment and sewage data set; Step S4: Tracking particle motion based on the wastewater dataset and virtual flow field data during treatment, and drawing a particle sedimentation curve; performing sedimentation offset mapping fitting based on the preset ideal sedimentation conditions and the particle sedimentation curve, and generating particle sedimentation trajectory offset data; Step S5: Obtaining sedimentation tank morphological parameters; reconstructing the sedimentation tank structure according to the sedimentation tank morphological parameters, and determining the reference inclination angle of the water inlet guide plate based on the sedimentation tank structure; predicting the abnormal state of sedimentation imbalance based on the reference inclination angle of the water inlet guide plate and the particle sedimentation offset data; Step S6: analyzing abnormal exceeding limit parameters in the initial sewage treatment process model through retrospective analysis of the abnormal state of sedimentation imbalance; and correcting the model parameters of the initial sewage treatment process model based on the abnormal exceeding limit parameters to generate a corrected sewage treatment process model.

2. The method for correcting parameters of a sewage treatment process model according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: Designing the layout of key nodes of the sewage treatment facility to obtain sensor deployment site data; configuring a multi-modal acquisition unit for each sensor deployment site to construct a multi-point composite acquisition unit; Step S12: constructing a sewage treatment sensor network based on the multi-point composite acquisition unit; performing multi-point synchronous sampling based on the sewage treatment sensor network to obtain multi-point sewage treatment facility sensor data, wherein the sampling frequency is set to 1 Hz to 5 Hz, and the sampling time is not less than 10 minutes; Step S13: performing structured decoding processing on the sensor data of multiple sewage treatment facilities to obtain standardized sewage sensor node data; Step S14: Analyze the sewage pipe topology data based on the sewage treatment sensor network, wherein the number of nodes is not less than 20 and the number of connected edges is not less than the number of nodes - 1; Step S15: Projecting the multi-point sewage treatment facility sensor data onto the sewage pipe topology data to obtain location information mapping data, with the projection error threshold limited to ≤0.5m; Step S16: performing site marking processing according to the location information mapping data to obtain corresponding location information of the multi-point sewage sensor projection data.

3. The method for correcting parameters of a sewage treatment process model according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: performing position data matching on the corresponding position information and the multi-point sewage facility sensor data, setting the matching tolerance range to ±0.5 meters, and obtaining position information matching data; Step S22: Process segment attribution is performed on the location information matching data. According to the node distribution in the sewage pipe topology, each matching point is assigned to the process segment closest to it. The attribution distance threshold is set to 1.0 meter to obtain process segment allocation data. Step S23: screening the process section allocation data for processing characteristics, selecting data samples with water quality parameters in the range of pH 6.5–8.5, dissolved oxygen greater than 2 mg / L, and COD less than 300 mg / L, to obtain screening sensor data; Step S24: performing time-series splicing on the filtered sensor data, splicing adjacent data in the order of acquisition timestamps, setting the maximum time interval to no more than 10 seconds, and obtaining a spliced sensor data sequence; Step S25: construct an initial sewage treatment process model based on the spliced sensor data sequence; extract the process segment time series fragment data in the initial sewage treatment process model, wherein the minimum time span of the treatment unit is set to 30 seconds; extract the treatment process sewage data set based on the process segment time series fragment data.

4. The method for correcting parameters of a sewage treatment process model according to claim 1, wherein: Step S3 includes the following steps: Step S31: performing multi-zone pressure simulation on the sewage data set during treatment, and recording the pressure response to obtain the unit partial pressure characteristics; Step S32: performing continuous permeation stratification processing on the unit partial pressure characteristics to obtain a layered permeation structure; performing flow channel configuration simulation on the wastewater data set during treatment based on the layered permeation structure to generate a simulated wastewater treatment process environment; Step S33: performing distributed pressure control mapping based on the simulated sewage treatment process environment to obtain unit hydraulic distribution data; and reconstructing the hydraulic time series path based on the unit hydraulic distribution data; Step S34: Perform path logic stitching simulation according to the hydraulic timing path, and construct a boundary ring structure based on the path logic stitching data; determine the flow boundary conditions based on the boundary ring structure and the hydraulic timing path, and simulate virtual flow field data based on the flow boundary conditions and the sewage data set of the treatment process.

5. The method for correcting parameters of a sewage treatment process model according to claim 1, wherein: Tracking particle motion based on the wastewater dataset and virtual flow field data during treatment and drawing a particle settling curve in step S4 includes: Analyze the fluid shear field based on virtual flow field data, and infer shear stress distribution data based on the fluid shear field; Identify wastewater particle parameters in a process wastewater dataset; Combine shear stress distribution data and wastewater particle parameters and establish particle-fluid interaction relationships; Identify particle force data based on particle-fluid interaction relationship and shear stress distribution data; Perform time-domain propulsion simulation based on particle force data and record the particle motion trajectory in the simulated propulsion time-domain data; Extract sedimentation rate characteristics from particle motion trajectories and analyze the time-varying sedimentation rate of particles; Draw a particle sedimentation velocity curve based on the time-varying sedimentation velocity of the particles; The particle settling rate curve is converted into a displacement representation, thereby obtaining a particle settling displacement curve; Identify the time relationship in the particle settling displacement curve, and draw the particle settling curve based on the time relationship and the particle settling displacement curve.

6. The method for correcting parameters of a sewage treatment process model according to claim 1, characterized in that: The sedimentation offset mapping fitting based on the preset ideal sedimentation condition and the particle sedimentation curve in step S4 includes: Draw an ideal deposition curve based on the preset ideal deposition conditions; Project and compare the ideal sedimentation curve and the particle sedimentation curve, and construct the sedimentation difference matrix; The offset of each particle trajectory is calculated according to the sedimentation difference matrix and integrated into a particle trajectory offset set; Determine the standard particle motion trajectory based on the ideal deposition curve; The particle trajectory offset set is superimposed on the standard particle trajectory, and an offset correction is performed to obtain an offset-corrected trajectory set. Perform spatiotemporal clustering on the offset-corrected trajectory set to obtain trajectory correction point data; Identify key excursion patterns based on trajectory correction point data; The particle sedimentation curve is predicted to be offset through the key offset mode to generate particle sedimentation trajectory offset data.

7. The method for correcting parameters of a sewage treatment process model according to claim 1, characterized in that: Reconstructing the sedimentation tank structure according to the sedimentation tank morphological parameters in step S5 and determining the reference inclination angle of the water inlet guide plate based on the sedimentation tank structure includes: Construct the initial sedimentation tank framework based on the sedimentation tank morphological parameters; Identify the key structural nodes in the initial sedimentation tank framework and connect the nodes based on the key structural nodes to generate the sedimentation tank skeleton network; Identify the geometric skeleton of the inlet area in the sedimentation tank skeleton network; The guide plate position and angle are calculated based on the geometric skeleton of the water inlet area, thereby generating a guide plate parameter set; The reference inclination angle of the water inlet guide plate is determined based on the guide plate parameter set of the water inlet area geometric framework.

8. The method for correcting parameters of a sewage treatment process model according to claim 7, characterized in that: Predicting the abnormal sedimentation imbalance state based on the water inlet guide plate reference inclination angle and the particle sedimentation offset data in step S5 includes: The long-term sedimentation process is simulated by using the particle settling trajectory migration data and sediment pool morphology parameters, and the cumulative distribution of sediments is determined based on the long-term sedimentation process; Calculate sediment load distribution based on the cumulative distribution of sediments and generate structural load distribution data; Map the structural load distribution data to the initial sedimentation tank framework and perform finite element stress analysis to generate stress distribution data; Identify the stress concentration points of the guide plate based on stress distribution data and sedimentation pool morphological parameters, and analyze the stress state of the guide plate; Predict the deformation trend of the guide plate according to its stress state; The rate of change of the inclination angle over time is calculated based on the deformation trend of the guide plate and mapped into a dynamic evolution curve of the guide plate inclination angle; The future tilt angle change data is predicted through the dynamic evolution curve of the guide plate tilt angle and the reference tilt angle of the water inlet guide plate; Performing surface fitting processing on the future tilt angle change data to obtain a tilt angle change fitting surface; The fitting surface of the tilt angle change is segmented for temporal stability to generate change trend fracture segments; Based on the fracture fragment mapping of the change trend, the corresponding deflector structure response area; The flow field deformation state is deduced based on the corresponding guide plate structure response area and particle sedimentation offset data; Perform gradient analysis on the deformation state of the flow field, calculate the corresponding velocity change rate of the guide plate area, and generate the flow field disturbance intensity; An intensity comparison is performed based on the preset critical sedimentation conditions and the flow field disturbance intensity, and the abnormal state of sedimentation imbalance is predicted based on the intensity comparison results.

9. The method for correcting parameters of a sewage treatment process model according to claim 1, wherein: Step S6 includes the following steps: Step S61: performing time series decomposition on the abnormal state of deposition imbalance and extracting key time nodes to determine the abnormal occurrence time series; Step S62: Back-analyzing abnormal exceeding limit parameters in the initial sewage treatment process model based on the abnormal occurrence time sequence; Step S63: performing historical trend analysis on abnormal out-of-limit parameters using preset historical model parameters, and identifying abnormal change patterns; inferring parameter drift characteristics based on the abnormal change patterns; Step S64: Correcting the model parameters of the abnormal and out-of-limit parameters based on the parameter drift characteristics, and reconfiguring the parameters of the initial sewage treatment process model based on the corrected model parameters, thereby generating a corrected sewage treatment process model.

10. A sewage treatment process model parameter correction system, characterized in that: For executing the sewage treatment process model parameter correction method according to claim 1, the sewage treatment process model parameter correction system comprises: A data acquisition module is used to collect sensor data of multiple sewage treatment facilities through the sewage treatment sensor network; extract the sewage pipe topology of the sewage treatment sensor network; project the sensor data of multiple sewage facilities into the sewage pipe topology and mark the corresponding location information; A model building module is used to build an initial sewage treatment process model based on corresponding location information and multi-point sewage facility sensor data; extract the treatment process sewage data set in the initial sewage treatment process model; The environmental simulation module is used to simulate the sewage treatment environment based on the sewage treatment process data set and generate a simulated sewage treatment process environment; simulate virtual flow field data based on the simulated sewage treatment process environment and the sewage treatment process data set; The particle tracking module is used to track particle motion based on the wastewater data set and virtual flow field data during treatment and draw particle sedimentation curves. It also performs sedimentation offset mapping fitting based on the preset ideal sedimentation conditions and particle sedimentation curves to generate particle sedimentation trajectory offset data. The sedimentation analysis module is used to obtain sedimentation pond morphological parameters; reconstruct the sedimentation pond structure based on the sedimentation pond morphological parameters, and determine the reference inclination angle of the water inlet guide plate based on the sedimentation pond structure; and predict the abnormal state of sedimentation imbalance based on the reference inclination angle of the water inlet guide plate and particle settling offset data; The parameter correction module is used to analyze the abnormal and excessive parameters in the initial sewage treatment process model through retrospective analysis of the abnormal state of sedimentation imbalance; based on the abnormal and excessive parameters, the model parameters of the initial sewage treatment process model are corrected to generate a corrected sewage treatment process model.

Images