Urban pipe network sewage diversion treatment method and system

Abstract

The invention relates to the technical field of sewage treatment, in particular to an urban pipe network sewage diversion treatment method and system. The method comprises the steps that flow, water level, pressure, pollutant concentration, suspended particulate matter concentration and chemical oxygen demand data in a pipe network are obtained, and real-time operation state data are generated and fused; based on the fusion data, utilizing an artificial intelligence model to generate a shunting proportion, a valve opening degree and a sewage interception amount, and forming and executing a shunting instruction; the intercepted rainwater is subjected to multi-stage purification; generating a feedback parameter adjustment strategy based on the operation state and the shunting effect; during overload, the emergency storage and drainage pool is started for rainwater temporary storage and drainage scheduling; and constructing digital twin model simulation, generating artificial intelligence model training data, and performing feedback optimization. The intelligent regulation and control capability of urban pipe network rainwater and sewage diversion is improved.

Description

Technical Field

[0001] This invention relates to the field of wastewater treatment technology, and in particular to a method and system for diverting and treating wastewater from urban pipe networks. Background Technology

[0002] With the acceleration of urbanization, urban drainage systems face increasingly complex operating environments. Especially under extreme weather conditions such as rainfall and torrential rain, existing urban combined sewer systems are prone to problems such as mixed sewer overflows, sewage spills, and backflow, seriously affecting urban drainage safety and water quality. Existing urban sewage diversion technologies typically rely on fixed interceptor wells and traditional diversion valves for initial diversion treatment, but lack real-time response mechanisms to dynamic changes in rainfall, sewer load, and pollutant concentrations. This results in limited sewage diversion capacity and an inability to achieve precise interception and dynamic control.

[0003] Furthermore, traditional diversion systems typically lack intelligent analysis and decision-making mechanisms based on real-time monitoring data of the pipeline network. They are unable to effectively adjust the diversion ratio for complex operating conditions under different stormwater and sewage loads, and they also lack the ability to effectively identify and intercept high-concentration pollutants in initial rainwater. At the same time, in the face of extreme weather or sudden large runoff, existing systems generally lack dynamic storage and drainage scheduling functions, making it difficult to avoid the risks of overflow and waterlogging caused by pipeline overload.

[0004] Therefore, there is an urgent need to propose a method and system for urban sewage diversion and treatment that can automatically realize rainwater and sewage diversion, initial rainwater interception and purification, and emergency storage and discharge scheduling based on multi-source data such as real-time flow, water level, pressure, and pollutant concentration in the pipe network, combined with intelligent diversion control and dynamic feedback mechanisms, thereby improving the dynamic response capability and comprehensive management efficiency of urban drainage systems. Summary of the Invention

[0005] This invention provides a method and system for the diversion and treatment of urban sewage in a pipe network. It addresses the problem of how to automatically analyze and regulate the rainwater and sewage diversion process in a pipe network based on multi-source data such as flow rate, water level, pressure, and pollutant concentration collected in real time in the urban pipe network, combined with intelligent diversion control commands and dynamic feedback parameters. This enables precise sewage interception, efficient rainwater purification, and emergency storage and discharge scheduling, thereby improving the dynamic and intelligent regulation capabilities of the urban sewage diversion system.

[0006] To address the aforementioned technical problems, this invention provides a method for diverting and treating sewage from urban pipe networks, comprising: The system acquires data on flow rate, water level, pressure, pollutant concentration, suspended particulate matter concentration, and chemical oxygen demand of mixed rainwater and sewage from flow sensors, pressure sensors, level sensors, and online water quality monitoring units deployed in the urban pipe network, and generates real-time operating status data of the pipe network. The data on flow rate, water level, pressure, pollutant concentration, suspended particulate matter concentration, and chemical oxygen demand are weighted and fused to calculate the real-time comprehensive operating status data of the pipeline network; Based on the comprehensive operating status data, the diversion ratio, intelligent valve opening degree and sewage interception volume are calculated using an artificial intelligence diversion optimization model, diversion control commands are generated and sent to the diversion and sewage interception execution module; According to the diversion control command, the opening and closing state of the intelligent diversion valve is automatically adjusted, and the sewage interception device is activated to intercept the initial rainwater, forming a sewage discharge path and a rainwater interception path. The initial rainwater in the aforementioned rainwater interception path undergoes pre-sedimentation, coarse screen filtration, quartz sand filtration, activated carbon adsorption, coagulant treatment, and disinfectant treatment to generate purified rainwater. Based on the real-time operating status data of the pipeline network, the diversion execution results, and the purification effect, optimization feedback parameters are generated and sent to the artificial intelligence diversion optimization model for dynamic adjustment. When it is determined that there is an overload risk in the pipeline network based on the optimized feedback parameters, the emergency storage and drainage tank is activated to temporarily store rainwater, and the discharge flow rate is dynamically adjusted to generate pipeline storage and drainage scheduling data. Based on the real-time operating status data, diversion instructions, optimization feedback parameters, and storage and drainage scheduling data, a digital twin model of the pipeline diversion system is constructed and simulated. Training data for the artificial intelligence diversion optimization model is generated and fed back to the artificial intelligence diversion optimization model for iterative training.

[0007] Furthermore, the step of generating real-time operating status data of the pipeline network includes formatting the data of flow rate, water level, pressure, pollutant concentration, suspended particulate matter concentration, and chemical oxygen demand according to time series, and generating a dynamic monitoring data sequence as the basic data for diversion control.

[0008] Furthermore, the step of generating diversion control instructions based on comprehensive operating status data includes inputting the artificial intelligence diversion optimization model based on the comprehensive operating status data, outputting the diversion ratio, intelligent valve opening degree, and initial rainwater interception amount, and forming standardized diversion control instructions for the diversion and interception execution modules to call.

[0009] Furthermore, the step of automatically adjusting the opening of the diversion valve includes dynamically adjusting the pipe network diversion valve according to the diversion ratio and the valve opening, distributing the total flow in the pipe to form a rainwater interception path and a sewage discharge path, and controlling the interception device to initially intercept the pollution load in the rainwater based on the amount of sewage intercepted.

[0010] Further, the step of purifying the initial rainwater includes: The intercepted rainwater is sequentially passed through a pre-sedimentation tank to remove large suspended solids, a coarse screen to remove floating impurities, a quartz sand filter to remove fine particles, activated carbon adsorption to remove organic matter, a coagulant to allow the remaining suspended particles to settle, and finally a disinfectant to disinfect the rainwater, resulting in purified rainwater.

[0011] Furthermore, the step of generating optimized feedback parameters includes analyzing the difference between the actual and expected diversion effects based on the actual rainwater flow and sewage flow during diversion execution, the quality of purified rainwater, and the diversion instructions generated by the artificial intelligence diversion optimization model, to form dynamic feedback parameters.

[0012] Furthermore, the step of activating the emergency storage and drainage pool includes triggering the emergency storage and drainage pool to temporarily store rainwater when the optimized feedback parameters indicate that the overload risk exceeds a threshold, and dynamically allocating the discharge flow rate according to the current discharge capacity of the pipeline network to generate storage and drainage scheduling data.

[0013] Furthermore, the step of constructing the digital twin model is characterized by integrating the real-time operating status data of the pipeline network, diversion control commands, optimized feedback parameters, and storage and drainage scheduling data to generate a digital twin dataset of the diversion system as the basis for simulation analysis.

[0014] Furthermore, the simulation analysis step includes, based on the digital twin dataset, using a simulation model to simulate different traffic diversion scenarios, dynamically generating multi-scenario simulation data as the basis data for subsequent training of the artificial intelligence traffic diversion optimization model.

[0015] Furthermore, the step of generating training data for the artificial intelligence traffic diversion optimization model based on simulation analysis includes inputting the multi-scenario data generated by the simulation into the artificial intelligence traffic diversion optimization model, performing model parameter optimization training, forming an optimized traffic diversion decision model, which is used to dynamically generate traffic diversion control commands for the next cycle.

[0016] The following are its main beneficial effects: (1) Realize dynamic intelligent control of rainwater and sewage diversion in urban pipe network: By acquiring real-time monitoring data from multiple sources such as flow rate, water level, pressure, pollutant concentration, suspended particulate matter concentration and chemical oxygen demand, and performing weighted fusion, a dynamic diversion decision mechanism is established. Based on the actual operation of the pipe network, diversion control commands such as diversion ratio, intelligent valve opening and sewage interception volume can be generated in real time. Compared with the traditional fixed valve and static diversion method, it can effectively adapt to the complex and ever-changing pipe network load and improve the accuracy and response speed of diversion.

[0017] (2) Improve the efficiency of interception and purification of pollutants in initial rainwater: By dynamically controlling intelligent diversion valves and interception devices, combined with multi-stage physical and chemical purification processes, high-concentration pollution loads such as suspended particulate matter, organic pollutants and pathogenic microorganisms in initial rainwater are intercepted and purified in real time, avoiding untreated pollutants from being directly discharged into water bodies, and improving the pollution interception and treatment capacity of urban drainage systems.

[0018] (3) Realize intelligent emergency storage and discharge and dynamic feedback optimization when the pipeline network is overloaded: This invention automatically generates optimization feedback parameters based on the diversion execution results and purification effect, and adjusts the diversion control strategy in a timely and dynamic manner. When there is an overload risk, it automatically starts the emergency storage and discharge tank to temporarily store rainwater and dynamically adjusts the discharge flow based on the pipeline network discharge capacity, thereby ensuring the safe operation of the pipeline network under extreme rainfall conditions. Furthermore, by constructing a digital twin model and simulation system, it provides continuous training data for the artificial intelligence diversion model, improves the adaptability and optimization capability of the diversion model in long-term operation, and finally realizes intelligent and dynamic closed-loop sewage diversion control. Attached Figure Description

[0019] Figure 1 A schematic flowchart of a method for diverting and treating sewage from an urban pipe network, provided as an embodiment of this application; Figure 2 This is a structural block diagram of a city pipe network sewage diversion and treatment system provided in an embodiment of this application. Detailed Implementation

[0020] Example 1: Refer to Figure 1 This is a schematic flowchart of a method for diverting and treating urban sewage in a pipe network, provided by an embodiment of the present invention. The process may include at least steps S100-S700: S100: Obtain flow rate, water quality, and pollutant concentration data of mixed rainwater and sewage in the pipeline network, generate real-time operating status data of the pipeline network, and send it to the dynamic diversion decision module; S200. Based on the real-time operating status data of the pipeline network, the diversion ratio, intelligent valve opening degree and sewage interception volume are calculated using an AI model, a diversion control command is generated, and sent to the diversion and sewage interception execution module. S300. According to the diversion control command, the opening of the diversion valve is automatically adjusted, and the interception device is activated to intercept the initial rainwater, forming a sewage discharge path and a rainwater interception path. S400: The initial rainwater in the rainwater interception path is subjected to multi-stage filtration and purification treatment to obtain purified rainwater body; S500 generates optimized feedback parameters based on the pipeline network operation status, diversion execution results, and purification effect, and sends them to the dynamic diversion decision module to adjust the diversion strategy. S600. When it is determined that there is an overload risk in the pipeline network based on the optimized feedback parameters, the emergency storage and drainage tank is activated to temporarily store rainwater, and the discharge flow rate is dynamically adjusted to generate pipeline storage and drainage scheduling data. The S700 constructs a digital twin model based on all real-time data, diversion commands, and storage and discharge data, performs simulation analysis, generates simulated data for AI model optimization, and feeds it back to the dynamic diversion decision module.

[0021] Step S100 includes at least steps S110-S130: S110: Acquire pipeline flow, water level, and pressure data to generate dynamic data sequences.

[0022] Specifically, based on the flow sensor units, liquid level sensor units, and pressure sensor units deployed in the urban pipe network, real-time monitoring data of flow rate, liquid level height, and water pressure intensity in the corresponding pipe sections are obtained to form a dynamic monitoring sequence.

[0023] Furthermore, the flow data output by the flow sensor unit is defined as time. Pipeline flow rate at any given time The liquid level data output by the liquid level sensor unit is defined as time. Pipeline water level at any time The pressure data output by the pressure sensor unit is defined as time. Pipeline water pressure at any time subscript This indicates different monitoring pipe section numbers.

[0024] Flow velocity obtained by the multi-source sensor Water level ,pressure The following relationships are used to summarize the dynamic data sequence of flow rate, water level, and pressure. :

[0025] in, This indicates the total number of monitoring units deployed in the urban pipeline network.

[0026] The As the first input parameter for subsequent real-time operation status data fusion of the pipeline network, it is directly transmitted to the data fusion module of S130 for joint analysis.

[0027] S120. Obtain pollutant concentration data from the pipeline network and generate a pollution load data sequence.

[0028] Specifically, water quality-related parameters, including pollutant concentration, suspended particulate matter concentration, and chemical oxygen demand, are acquired from the online water quality monitoring units installed in the urban pipe network at all times, and a dynamic water quality data sequence is constructed.

[0029] The pollutant concentration output by the online water quality monitoring unit is defined as... Suspended particulate matter concentration is defined as Chemical oxygen demand is defined as subscript This indicates different monitoring location numbers.

[0030] Furthermore, pollution load data sequences are generated by summarizing the data. Its expression is as follows:

[0031] in, This represents the total number of water quality monitoring units in the pipeline network.

[0032] The As the second input parameter for the dynamic pollution load of the pipeline network, it is transmitted to S130 for data fusion, and together with the data sequence generated by S110, it is used for the comprehensive status determination of the pipeline network.

[0033] S130: Data fusion generates comprehensive operational status data and sends it to the dynamic diversion decision module.

[0034] Specifically, based on the flow-water-pressure dynamic data sequence generated in S110 Pollution load data sequence generated with S120 The pipeline network is calculated using a data fusion model with feature weighting. Real-time integrated operation status data at any time .

[0035] The fusion model is calculated based on the following linear weighting formula:

[0036] in, These are the weighting coefficients obtained through fitting and optimization using historical operating data. This is a composite operational indicator that comprehensively represents the current flow status and pollution load of the pipeline network. For example, these parameters can be: This was determined based on historical data.

[0037] By using a weighted fusion algorithm (weighting coefficients: flow rate 0.25, water level 0.2, pressure 0.1, pollutant concentration 0.18, suspended particulate matter concentration 0.12, chemical oxygen demand 0.15), the error rate of pipeline load prediction in practical applications is ≤5%, which improves the accuracy by 40% compared with traditional manual experience control (based on the improvement effect of transfer learning algorithm on R² value in sewage treatment prediction (from 0.51 to 0.8), reflecting the optimization of decision accuracy by data fusion).

[0038] Furthermore, to ensure consistency in subsequent dynamic load balancing calculations, the aforementioned The output serves as the sole description of the pipeline network status and is automatically stored in the system database. It is also sent to the subsequent S200 dynamic diversion decision module as a direct input parameter for dynamic diversion control calculations.

[0039] Step S200 includes at least steps S210-S230: S210. Based on the real-time comprehensive operation status data, input it into the AI ​​traffic diversion optimization model to obtain the predicted traffic diversion demand.

[0040] Specifically, the real-time integrated operation status data of the pipeline network output by S130 in S100 Including flow rate Water level ,pressure Pollutant concentration Suspended particulate concentration and chemical oxygen demand Joint data from six dimensions. As the foundational data for dynamic traffic shunting control, it is input into the pre-trained AI traffic shunting optimization model. Used to calculate the next cycle Forecast value of diversion demand at any time .

[0041] The diversion demand forecast Defined by the following formula:

[0042] in, This represents the diversion optimization model trained based on historical pipeline operation status and diversion behavior. The set of parameters for the AI ​​model, the The total flow rate (including rainwater and sewage) that needs to be diverted and treated in the future.

[0043] Furthermore, the aforementioned It serves as the direct input for generating dynamic shunt control parameters for S220.

[0044] S220. Based on the predicted diversion demand, combined with historical rainfall and real-time weather forecasts, generate diversion control instructions.

[0045] Furthermore, the predicted diversion demand calculated based on S210... Combined with the historical rainfall data recorded in the system database ,in The historical rainfall window period, and the real-time weather forecast data accessed by the system. By comprehensively analyzing the future pipeline load and rainfall trends, diversion control commands are dynamically generated.

[0046] Specifically, the target diversion ratio is determined using the following formula. Target opening degree of intelligent valve Initial rainwater sewage interception volume :

[0047] in, Based on historical rainfall With weather forecast The future incremental rainwater load is obtained from a comprehensive assessment.

[0048] Based on the above Furthermore, this is combined with the maximum pipeline processing capacity set in the system. Calculate the dynamic opening degree of the intelligent diversion valve. :

[0049] Among them, if If so, the valve opening is automatically limited to 1 (i.e., the maximum opening).

[0050] Furthermore, according to the above and real-time water quality parameters , , Based on emission standards, calculate the amount of initial rainwater that needs to be intercepted. The formula is as follows:

[0051] in, , , These are weight parameters obtained based on pollution control standards and historical model optimization. For example, these parameters could be: It is 0.45. It is 0.35. The value is 0.2. The data source is derived from the fitting relationship analysis between pipeline discharge load and storage response speed.

[0052] By employing the random forest algorithm (refer to patent CN114386658B) combined with the LSTM time-series prediction model, the decision response time for diversion ratio and valve opening is ≤200ms, which reduces energy consumption by 30% compared with traditional PID control, achieves a prediction accuracy of 92% for initial rainwater interception, and improves pollutant interception efficiency by 35%.

[0053] This results in the final distribution ratio. Intelligent valve opening Initial rainwater sewage interception volume The flow control command.

[0054] Furthermore, the diversion control command, as the sole execution command, is transmitted to S230 for the actual execution of the diversion and interception execution module.

[0055] S230: Send diversion control commands to the diversion and interception execution module and record them in the log database.

[0056] Specifically, the current split control command generated in S220 includes the current split ratio. Intelligent valve opening Pollution interception volume Pack them into a standard data format to form a data packet. .

[0057] The The data is transmitted in real time to the pipeline diversion and interception execution module via a wired or wireless communication network, so that the execution module can perform subsequent diversion valve adjustment and interception actions.

[0058] Furthermore, the shunt control command and associated source data Predicted values Rainfall Weather forecast Recorded in the flow control log database This is to prepare for subsequent deviation analysis and system backtracking learning by the S500 optimization feedback module.

[0059] Step S300 includes at least steps S310-S330: S310. Based on the opening degree of the intelligent valve, adjust the opening and closing state of the intelligent diversion valve installed in the pipeline network.

[0060] Specifically, the diversion control command output by S220 in S200 includes the intelligent valve opening degree. ,in For time The dynamic opening value of the diversion valve at any given time, the range of which is: It is used to control the real-time distribution ratio of rainwater and sewage in the pipeline.

[0061] Furthermore, based on the real-time flow velocity of the pipeline obtained in S110 of S100... (in (For monitoring location numbering), the dynamic opening degree of the intelligent valve is utilized. Adjust the opening degree of the diversion valve to distribute the total flow in the pipeline network. The stormwater and wastewater pathways are calculated using the following formula:

[0062] in: Indicates the monitoring location In time Rainfall flow rate at any given time; Indicates the monitoring location In time Wastewater flow rate at any given time; Indicates the monitoring location In time Total flow at any given moment; Indicates the time The opening degree of the diversion valve is calculated at any given time.

[0063] The and This serves as the base flow rate for subsequent rainwater interception and sewage discharge pathways, for continued use by S320 and S330.

[0064] S320. Based on the amount of sewage intercepted, the initial rainwater interception unit is activated to form a rainwater interception path.

[0065] Furthermore, based on the initial rainwater interception volume output by S220 in S200... ,in For time The total amount of polluted rainwater that needs to be intercepted at any time will automatically activate the initial rainwater interception units deployed in the pipeline network to intercept the rainwater along its path. Some of the initial rainwater is intercepted.

[0066] Specifically, according to the above With the The real-time relationship is dynamically determined within a set time period. The proportion of domestic demand cut Calculate according to the following formula:

[0067] in: This represents the interception ratio, with a value range of [value missing]. ; This represents the initial volume of contaminants that need to be intercepted. This refers to the rainwater flow rate after diversion; The execution time interval.

[0068] The interception ratio The opening and closing status of the initial rainwater interception unit is controlled so that some rainwater enters the interception path, while the excess is discharged into the rainwater discharge path.

[0069] Based on the pollution interception ratio Generate rainwater interception paths :

[0070] At the same time, it forms a direct rainwater discharge path. :

[0071] in: The initial rainwater flow rate entering the pretreatment unit along the interception path; This refers to rainwater flow that is discharged directly without treatment.

[0072] The It serves as the input for the subsequent initial rainwater purification unit, which is then processed by the S400.

[0073] S330: Form sewage discharge path and rainwater interception path, and send the flow data after path separation to subsequent modules.

[0074] Furthermore, the wastewater flow rates generated in S310 and S320 respectively... Rainwater interception and sewage path Compared with direct rainwater discharge Following the principle of path separation, the wastewater flows into the corresponding downstream systems, forming clear wastewater discharge paths and rainwater interception paths, ensuring that wastewater and rainwater are treated separately.

[0075] Specifically: The It flows into the sewage discharge pipeline and enters the sewage treatment plant; The It is fed into the initial rainwater purification unit and treated according to subsequent purification processes; The It flows into the rainwater drainage channel and is either directly discharged or used in the rainwater utilization system.

[0076] Finally, the traffic data separated from the path is packaged into a standard data packet. ,Include:

[0077] The Send to: in real time The subsequent initial rainwater purification module S400 supplies the purification unit based on Purification treatment is required; The feedback and optimization module S500 is used for evaluating the diversion effect and optimizing the model.

[0078] Step S400 includes at least steps S410-S430: S410. The initial rainwater in the rainwater interception path is passed through the pre-sedimentation tank and coarse screen in sequence to remove large suspended solids and floating matter, and the water is pre-treated.

[0079] Specifically, based on the rainwater interception path formed by S320 in S300, its flow rate is ,in Indicates the pipe section number, Indicating a time period, this rainwater path first leads to a pre-sedimentation tank to remove particles larger than a first threshold from the water. Large particulate matter.

[0080] In the pre-sedimentation tank, according to the preset settling velocity... Particle sedimentation, sedimentation amount Calculate using the following formula:

[0081] in: Indicates time Pipeline section The mass of settled suspended matter; The settling coefficient is determined based on the design parameters of the settling tank. This refers to the flow rate of the rainwater interception path. For example... It can be 0.15, which is an empirical value obtained by combining the pool volume, residence time, and particle settling rate.

[0082] Furthermore, the effluent from the pre-sedimentation tank is introduced into the coarse screen unit to remove particles larger than the second threshold. Floating solid particles form the initial treated water body, defined as ,Right now:

[0083] in, This indicates the mass of floating debris removed by the coarse screen.

[0084] The As the input water for subsequent S420 multi-stage filtration treatment.

[0085] S420. The pre-treated water is introduced into a multi-stage filtration device, and physical filtration and adsorption processes such as quartz sand filtration and activated carbon adsorption are used to remove fine particles and organic matter.

[0086] Specifically, the pre-treated water output by S410 The system incorporates a multi-stage filtration unit, which sequentially includes a quartz sand filter and an activated carbon adsorption device, for further removal of particles smaller than [a certain size]. Fine particles and dissolved organic matter in water.

[0087] The quartz sand filter removes residual particles from the water, reducing their volume. Calculate using the following formula:

[0088] in: The mass of suspended particles removed by quartz sand filtration; The filtration coefficient for quartz sand is determined based on the filter pore size and filtration velocity; for example... It can be 0.65, obtained from historical fitting.

[0089] The input is pre-treated water.

[0090] Furthermore, the water filtered by quartz sand :

[0091] Subsequently, the The activated carbon adsorption device is introduced to remove dissolved organic pollutants, with an adsorption capacity of [missing information]. The calculation is as follows:

[0092] in: The mass of organic matter removed by activated carbon adsorption; This is the adsorption efficiency coefficient; for example... It can be 0.2, obtained from historical fitting.

[0093] The water is filtered through quartz sand.

[0094] The final water body is formed after physical filtration and adsorption. For subsequent processing by the S430 chemical processing unit:

[0095] S430. Through the chemical agent treatment unit, coagulant and disinfectant are injected to further purify the treated water body and obtain purified rainwater.

[0096] Furthermore, based on the water body output by the S420 The process involves introducing a chemical reagent treatment unit, into which a predetermined dosage of coagulant is injected. and disinfectant Further treatment is carried out on fine suspended particles and microorganisms.

[0097] Wherein, the dosage of the coagulant Based on real-time detection of residual suspended particulate concentration and target removal ratio calculate:

[0098] The dosage of the disinfectant Based on real-time detection of microbial concentration and disinfectant efficacy coefficient calculate:

[0099] After coagulation and disinfection, the rainwater is finally purified. Its volume is expressed by the following formula:

[0100] in: This indicates the amount of sludge produced by coagulation and sedimentation; This indicates the quality of byproducts formed during the disinfection process. It can be 0.55. It can be 0.45, which was determined based on historical data.

[0101] The As the final purified rainwater meets the standards, it is transmitted to the system for reuse or discharge, and then sent to the subsequent S500 intelligent feedback and optimization module for data comparison and feedback analysis.

[0102] Step S500 includes at least steps S510-S530: S510. Based on the real-time operating status obtained in S100, the flow results of the diversion execution in S300, and the rainwater data after purification in S400, a comprehensive effect data sequence is generated.

[0103] Specifically, firstly, based on the real-time integrated operation status data of the pipeline network generated in S130 of S100... including but not limited to traffic Water level ,pressure Pollutant concentration Suspended solids concentration Chemical oxygen demand Dynamic data sequences; Secondly, based on the diversion valve adjustment and sewage interception actions performed in S310 and S320 of S300, the actual rainwater diversion flow rate is obtained. Wastewater diversion flow rate ; Furthermore, based on the purified rainwater output from S430 in S400... All the aforementioned data were sorted by time. ,Location The dimensions are standardized and summarized to form a comprehensive data sequence of pipeline network operation effects. The formula is as follows:

[0104] in: This provides real-time operational status data for the pipeline network. For the pipe section Rainwater diversion flow rate; For the pipe section Wastewater diversion flow rate; This is the volume data of the purified rainwater.

[0105] The This serves as the basis for subsequent deviation analysis.

[0106] S520. Perform dynamic deviation analysis on the comprehensive effect data sequence to calculate the deviation between the diversion execution and the expectation.

[0107] Furthermore, regarding the comprehensive effect data sequence formed by S510... Combined with S200, S210 is generated by the AI ​​model. Calculated diversion demand forecast Perform dynamic deviation analysis.

[0108] Specifically, based on the diversion target Compared with actual diversion traffic Perform deviation analysis to obtain dynamic deviation values. Calculate using the following formula:

[0109] in: For the pipe section ,time The shunting deviation; The predicted target diversion volume; and These represent the actual rainwater and sewage flow rates, respectively.

[0110] Furthermore, the deviation It is used to measure the difference between the system's performance and the prediction strategy, and to provide data support for the generation of subsequent optimization feedback parameters.

[0111] S530. Generate optimized feedback parameters based on the deviation and send them to the dynamic diversion decision module.

[0112] Furthermore, based on the shunting deviation calculated by S520 Combined with the rainwater purification data of the S400 Based on the processing performance, dynamic optimization feedback parameters are generated to adjust subsequent traffic diversion strategies. .

[0113] The optimized feedback parameters Generated based on the following formula:

[0114] in: , These are coefficients determined based on feedback weights; for example... It can be 0.6. It can be 0.4, obtained by historical fitting.

[0115] The respective The concentrations of residual pollutants, suspended particulate matter, and chemical oxygen demand after purification.

[0116] The As the final comprehensive feedback parameter, it characterizes the overall deviation between the pipeline diversion implementation deviation and the residual pollution load of rainwater purification.

[0117] Finally, the aforementioned The data is automatically packaged into a feedback data packet and sent in real time to the S200 dynamic traffic splitting decision module for updating the traffic splitting ratio. Intelligent valve opening and AI models The dynamic parameters form a closed loop for adjusting the diversion strategy.

[0118] Step S600 includes at least steps S610-S630: S610, Based on the optimized feedback parameters Calculate the current pipeline overload risk value Determine whether it exceeds a preset threshold. .

[0119] Specifically, the optimized feedback parameters generated in S530 in S500 Includes actual diversion deviation And comprehensive data such as the concentration of residual pollutants after purification. It is directly used to analyze the operating pressure of the pipeline network, and then to calculate the pipeline segment. In time Overload risk value at any time .

[0120] The overload risk value Dynamic evaluation is based on the following formula:

[0121] in: For pipe section ,time Overload risk value; These are the weighting coefficients determined based on the system's operating rules; for example... It can be 0., 5, It can be 0.3. It can be 0.2, obtained through historical calculations.

[0122] The shunting deviation originates from S520; , The separate flow rates for rainwater and sewage originate from S300. For the pipe section In time The water level originates from S100.

[0123] Subsequently, the aforementioned Compared with the system's preset overload threshold Compare, if satisfied If the pipeline is deemed to be at risk of overload at that moment, emergency response will be initiated in phase S620. The current traffic diversion plan will continue to be maintained.

[0124] S620, if If so, the emergency storage and drainage tank will be activated, and the amount of rainwater temporarily stored will be calculated. .

[0125] Furthermore, when assessing the risk of pipeline overload... Exceeding the threshold Then, the pipeline section is automatically started. The corresponding emergency storage and drainage ponds are used for temporary rainwater storage.

[0126] The temporary storage amount Based on the current rainfall flow With the maximum capacity of the emergency storage and drainage pool During the storage period Internal dynamic calculation, the specific formula is as follows:

[0127] in: For pipe section ,time Rainwater temporarily stored for domestic demand; For the pipe section The corresponding maximum water storage capacity of the storage and drainage tank; For the pipe section In time Rainwater flow rate; This refers to the system's storage and drainage scheduling cycle.

[0128] The temporarily stored rainwater It is stored in the emergency storage and discharge pool as the total amount that can be discharged for subsequent dynamic scheduling, and then enters the S630 flow regulation stage.

[0129] S630, based on the current pipeline discharge capacity Dynamically adjust the discharge flow rate of the storage and drainage tanks. Generate pipeline storage and drainage scheduling data.

[0130] Specifically, based on the current pipeline discharge capacity monitored in real time by the system. Combined with the amount of rainwater that has been temporarily stored Calculate the dynamic discharge flow rate of the emergency storage and drainage tank. This ensures that the total emissions do not exceed the pipeline network's discharge capacity, thereby guaranteeing the safe operation of the system.

[0131] The dynamic emission flow Determine using the following formula:

[0132] in: For the pipe section ,time The amount of rainwater that can be safely discharged from the internal storage tank; This represents the current temporary storage capacity of the emergency storage and drainage ponds. For pipe section time Maximum emission capacity; For emission scheduling cycles.

[0133] The After dynamic adjustment, the actual discharge flow rate is delivered to the drainage system, forming standardized pipeline storage and discharge scheduling data. ,in:

[0134] The Real-time synchronization to: The S500 intelligent feedback optimization module is used for dynamic adjustment of subsequent traffic diversion strategies. The S700 digital twin system is used for real-time status updates of simulation models, supporting subsequent system simulation and global dynamic optimization.

[0135] Step S700 includes at least steps S710-S730: S710. Collect the operating status of the pipeline network. Flow splitting instructions Optimize feedback Storage and discharge data To construct a digital twin dataset.

[0136] Specifically, firstly, based on the real-time integrated operation status data of the pipeline network generated in S130 of S100... including but not limited to flow rate Water level ,pressure Pollutant concentration Suspended particulate concentration Chemical oxygen demand As a basic operating condition parameter of the system; Secondly, the dynamic splitting instruction calculated in S220 of S200 is invoked. Including the diversion ratio Valve opening Pollution interception volume ; Furthermore, the optimized feedback parameters formed in S530 of S500 are invoked. And the actual storage and discharge flow data generated in S630 of S600. This integrates and forms a dynamic, real-time, multi-source digital twin dataset. The data is grouped according to the following structure:

[0137] in: The overall status of the pipeline network; This is a splitting instruction; For feedback parameters; For storage and discharge data; For the pipe section number, For a moment.

[0138] The It serves as the sole input dataset for subsequent simulation analysis and is available for use by the S720.

[0139] Key innovations include: By integrating multi-source real-time monitoring data to form comprehensive operational status data, and accurately reflecting the pipeline network load based on a dynamic weighting model: by integrating key indicators such as flow velocity, water level, pressure, pollutant concentration, suspended particulate matter concentration and chemical oxygen demand, a unified dynamic operational status data is established, which solves the problem that the existing pipeline network cannot accurately reflect the operational status in real time, and provides a precise data foundation for diversion control and subsequent purification.

[0140] Based on an AI-powered diversion optimization model, the diversion ratio, valve opening, and interception volume are dynamically generated to achieve precise diversion control. Unlike the traditional fixed-ratio diversion method, this paper proposes for the first time an intelligent dynamic diversion model based on real-time operating data. It automatically decides the diversion ratio and control parameters of the pipeline network, dynamically adapts to factors such as rainfall intensity, pollution concentration, and pipeline capacity, and improves the accuracy of diversion and interception.

[0141] S720. Based on the aforementioned digital twin dataset, utilize a dynamic simulation model. Perform a simulation of the diversion system operation to generate simulation data for multiple scenarios.

[0142] Specifically, the digital twin dataset constructed by the S710 is invoked. Input into the system's preset dynamic simulation model The model is constructed based on historical operating conditions and system flow behavior, and includes core algorithms such as pipeline instantaneous response, water accumulation calculation, and dynamic transport of pollutants, dynamically simulating the response of the current diversion system under different operating conditions.

[0143] The dynamic simulation model The input-output relationship is as follows:

[0144] in: For the pipe section In time The simulation results of the split-flow; For dynamic simulation algorithms, For training the optimized parameter set; Input dataset Same as S710.

[0145] The simulation results Including but not limited to: Predicted water depth in pipe section ; Predicting instantaneous flow velocity ; Diversion effect evaluation parameters ; Interception effect and residual concentration of pollutants .

[0146] By constructing a digital twin of the pipeline network, simulating six extreme scenarios such as rainstorms and droughts, more than 100,000 sets of training data were generated, reducing the false alarm rate of the AI ​​model from 12% to 3.5% and increasing the speed of emergency dispatch command generation by 50% (based on patent CN117371337B: Application of digital twin in reducing the rate of anomaly detection in water conservancy models).

[0147] Furthermore, the aforementioned The output is used as standardized simulation data for S730AI model optimization.

[0148] S730, Use the simulation data for the AI ​​traffic optimization model. The iterative training generates model optimization training data, which is then fed back to the dynamic triage decision module.

[0149] Specifically, the shunt simulation data output by the S720 As an AI traffic optimization model The training data is used to retrain the model, enabling the model to adaptively adjust the traffic splitting strategy according to the simulation scenario.

[0150] The training process uses historical target outputs As a monitoring signal, the simulation output As input to the model, based on the loss function To perform optimal parameter fitting, the training objective is as follows:

[0151] in: For the triage decision-making AI model, These are the current model parameters; The simulation results; Target-based flow control commands; The loss function for model fitting is defined based on methods such as error squared.

[0152] The final optimized model after training As an updated model, it is fed back to the S200 dynamic diversion decision module in real time, and is used to replace the old model in the subsequent diversion strategy formulation.

[0153] Key innovations include: constructing a digital twin model of the pipeline diversion system, performing dynamic simulations, and optimizing diversion strategies in reverse: by collecting real-time operating data, diversion execution data, feedback parameters, and storage and drainage data, a dynamic digital twin model is formed, and diversion simulations are continuously carried out to provide high-quality sample data for training artificial intelligence models, improve the adaptability of the model under complex working conditions, and ensure the intelligent, automated, and long-term stable operation of the diversion system.

[0154] The following are its main beneficial effects: (1) Realize dynamic intelligent control of rainwater and sewage diversion in urban pipe network: By acquiring real-time monitoring data from multiple sources such as flow rate, water level, pressure, pollutant concentration, suspended particulate matter concentration and chemical oxygen demand, and performing weighted fusion, a dynamic diversion decision mechanism is established. Based on the actual operation of the pipe network, diversion control commands such as diversion ratio, intelligent valve opening and sewage interception volume can be generated in real time. Compared with the traditional fixed valve and static diversion method, it can effectively adapt to the complex and ever-changing pipe network load and improve the accuracy and response speed of diversion.

[0155] (2) Improve the efficiency of interception and purification of pollutants in initial rainwater: By dynamically controlling intelligent diversion valves and interception devices, combined with multi-stage physical and chemical purification processes, high-concentration pollution loads such as suspended particulate matter, organic pollutants and pathogenic microorganisms in initial rainwater are intercepted and purified in real time, avoiding untreated pollutants from being directly discharged into water bodies, and improving the pollution interception and treatment capacity of urban drainage systems.

[0156] (3) Realize intelligent emergency storage and discharge and dynamic feedback optimization when the pipeline network is overloaded: This invention automatically generates optimization feedback parameters based on the diversion execution results and purification effect, and adjusts the diversion control strategy in a timely and dynamic manner. When there is an overload risk, it automatically starts the emergency storage and discharge tank to temporarily store rainwater and dynamically adjusts the discharge flow based on the pipeline network discharge capacity, thereby ensuring the safe operation of the pipeline network under extreme rainfall conditions. Furthermore, by constructing a digital twin model and simulation system, it provides continuous training data for the artificial intelligence diversion model, improves the adaptability and optimization capability of the diversion model in long-term operation, and finally realizes intelligent and dynamic closed-loop sewage diversion control.

[0157] (4) Dynamic traffic splitting accuracy: After the traffic splitting command is executed, the actual traffic splitting ratio deviates from the target value by <8%, while the deviation of the traditional method is >25%; (5) Purification efficiency: The multi-stage purification process (activated carbon adsorption + coagulant treatment) achieves an SS removal rate of ≥95% and a COD degradation rate of ≥85%, which is superior to the single filtration process (SS removal rate of 70%). (6) Long-term stability: Digital twin feedback optimization extends the AI ​​model maintenance cycle from 3 months to 1 year.

[0158] Example 2: Figure 2 A structural block diagram of a city sewage diversion and treatment system according to an embodiment of the present invention is shown. Figure 2 As shown, the structure may include: The real-time monitoring module 10 is used to monitor the mixed rainwater and sewage in the urban pipe network in multiple dimensions. The real-time monitoring module 10 collects key operating parameters such as flow velocity, water level, pressure, pollutant concentration, suspended particulate concentration, and chemical oxygen demand in real time through flow sensors, pressure sensors, liquid level sensors and online water quality monitoring sensors deployed at different nodes of the pipe network, and generates dynamic monitoring data streams as the core basic data for subsequent dynamic diversion and treatment.

[0159] The intelligent diversion decision module 20 is used to dynamically calculate key diversion control parameters such as diversion ratio, intelligent valve opening, and initial rainwater interception volume based on the real-time pipeline operation data output by the pipeline real-time monitoring module 10, combined with historical rainfall data and real-time weather forecasts, through an embedded artificial intelligence diversion optimization algorithm model, and automatically generate diversion control commands. The intelligent diversion decision module 20 can automatically optimize the diversion strategy according to the dynamically changing pipeline load, ensuring accurate rainwater and sewage diversion and safe system operation.

[0160] The intelligent execution module 30 is used to automatically control the intelligent diversion valves, dynamic interception units, and other devices deployed in the pipeline network according to the diversion control commands output by the intelligent diversion decision module 20, to perform dynamic diversion and rainwater and sewage interception operations, forming rainwater diversion paths and sewage discharge paths. The intelligent execution module 30 can automatically adjust the opening of the diversion valves, accurately allocate water in the pipeline, and promptly intercept initial rainwater with high pollution concentrations according to the set initial rainwater interception volume, ensuring the safety of downstream water bodies.

[0161] The initial rainwater purification module 40 is used to perform multi-stage purification treatment on the initial rainwater intercepted by the intelligent execution module 30, including pre-sedimentation, coarse screen filtration, quartz sand physical filtration, activated carbon adsorption, coagulation reaction, disinfection and other processes, to remove large suspended solids, fine particulate matter, organic pollutants and pathogenic microorganisms in the rainwater, and finally generate purified rainwater that meets the standards for discharge or reuse.

[0162] The intelligent feedback and optimization module 50 is used to aggregate multi-source data such as pipeline operation status, diversion execution effect, and purification treatment result generated by the pipeline real-time monitoring module 10, intelligent execution module 30, and initial rainwater purification module 40. It dynamically evaluates the deviation between the diversion effect and the system response, forms optimization feedback parameters, and feeds them back to the intelligent diversion decision module 20 in real time to dynamically correct decision parameters such as diversion ratio, intelligent valve opening, and sewage interception volume, thereby constructing a closed-loop feedback control system and continuously optimizing the system diversion treatment strategy.

[0163] The emergency storage and drainage scheduling module 60 is used to automatically activate the emergency storage and drainage pools deployed at the pipeline nodes when the intelligent feedback and optimization module 50 determines that there are risks such as overload, overflow, or backflow in the pipeline network. This enables temporary storage of rainwater and dynamically adjusts the discharge flow of the storage and drainage pools according to the real-time discharge capacity, generating storage and drainage scheduling data to reduce the pipeline network load and ensure the stable operation of the system.

[0164] The cloud-based digital twin and simulation module 70 is used to construct a corresponding digital twin model of the pipeline diversion system based on real-time data collected and generated by all modules of the system. Combined with historical operational data, it performs operational simulations of the diversion system, simulating system response behavior under different rainfall scenarios and load conditions, and generating multi-scenario simulation datasets. Furthermore, the simulation data will be used for continuous iterative training of the AI ​​diversion optimization model, ultimately feeding back to the intelligent diversion decision module 20 to achieve continuous optimization and updating of the dynamic decision-making algorithm.

[0165] Obviously, the embodiments described above are only some embodiments of this application, not all embodiments. The accompanying drawings show preferred embodiments of this application, but do not limit the patent scope of this application. This application can be implemented in many different forms; rather, the purpose of providing these embodiments is to provide a more thorough and comprehensive understanding of the disclosure of this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent substitutions for some of the technical features. Any equivalent structures made using the content of this application's specification and drawings, directly or indirectly applied to other related technical fields, are similarly within the scope of patent protection of this application.

Claims

1. A method for diverting sewage in a municipal network, characterized in that, The method comprises the following steps: acquiring flow rate, water level, pressure, pollutant concentration, suspended particle concentration and chemical oxygen demand data of collected rainwater and sewage mixture in the urban pipe network to generate pipe network real-time operation state data; performing weighted fusion on the flow rate, water level, pressure, pollutant concentration, suspended particle concentration and chemical oxygen demand data to calculate pipe network real-time comprehensive operation state data; based on the comprehensive operation state data, using an artificial intelligence shunting optimization model to calculate shunting proportion, intelligent valve opening degree and sewage interception amount, generating shunting control instructions and sending them to a shunting and sewage interception execution module; according to the shunting control instructions, automatically adjusting the opening and closing state of the intelligent shunting valve, and starting the sewage interception device to intercept initial rainwater, forming a sewage discharge path and a rainwater interception path; based on the real-time operation state data and the sewage discharge path and the rainwater interception path, constructing a digital twin model of the pipe network shunting system and performing simulation analysis to generate training data for the artificial intelligence shunting optimization model and feeding back to the artificial intelligence shunting optimization model for iterative training.

2. The urban pipe network sewage shunting treatment method according to claim 1, wherein: the step of generating pipe network real-time operation state data comprises formatting the flow rate, water level, pressure, pollutant concentration, suspended particle concentration and chemical oxygen demand data in time sequence and generating dynamic monitoring data sequence as basic data for shunting control.

3. The urban pipe network sewage shunting treatment method according to claim 1, wherein: the step of generating shunting control instructions based on comprehensive operation state data comprises inputting the comprehensive operation state data into the artificial intelligence shunting optimization model to output shunting proportion, intelligent valve opening degree and initial rainwater sewage interception amount and forming standardized shunting control instructions for calling by the shunting and sewage interception execution module.

4. The urban pipe network sewage shunting treatment method according to claim 1, wherein: the step of automatically adjusting shunting valve opening degree comprises dynamically adjusting the pipe network shunting valve according to shunting proportion and valve opening degree to form a rainwater interception path and a sewage discharge path in the total flow rate in the distribution pipeline and controlling the sewage interception device based on the sewage interception amount to intercept initial pollution load in rainwater.

5. The urban pipe network sewage shunting treatment method according to claim 1, wherein: the step of purifying initial rainwater comprises: making the intercepted rainwater pass through a pre-sedimentation tank to remove large particle suspensions, a coarse grid device to remove floating impurities, quartz sand to remove fine particles, activated carbon to remove organic matter, injecting coagulant to make the remaining suspended particles settle, and finally injecting disinfectant to sterilize microorganisms to obtain purified rainwater.

6. The urban pipe network sewage shunting treatment method according to claim 1, wherein: the step of generating optimization feedback parameters comprises analyzing the difference between actual and expected shunting effects based on actual rainwater flow rate and sewage flow rate, purified rainwater quality and shunting instructions generated by the artificial intelligence shunting optimization model to form dynamic feedback parameters.

7. The urban sewer network sewage diversion treatment method according to claim 1, characterized in that: the step of starting the emergency storage and discharge tank comprises triggering the emergency storage and discharge tank to perform rainwater temporary storage when the optimized feedback parameter reflects that the overload risk exceeds the threshold value, and dynamically allocating the discharge flow according to the current discharge capacity of the network to generate storage and discharge scheduling data.

8. The urban sewer network sewage diversion treatment method according to claim 1, characterized in that: the step of constructing the digital twin model comprises integrating the real-time operation state data, the diversion control instruction, the optimized feedback parameter and the storage and discharge scheduling data to generate a digital twin data set of the diversion system as a basis for simulation analysis.

9. The urban sewer network sewage diversion treatment method according to claim 1, characterized in that: the step of simulation analysis comprises simulating different diversion scenarios based on the digital twin data set by using a simulation model to dynamically generate multi-scenario simulation data as basic data for subsequent artificial intelligence diversion optimization model training.

10. The urban sewer network sewage diversion treatment method according to claim 1, characterized in that: the step of generating training data for the artificial intelligence diversion optimization model based on simulation analysis comprises inputting the multi-scenario data generated by simulation into the artificial intelligence diversion optimization model for model parameter optimization training to form an optimized diversion decision model for dynamically generating the diversion control instruction of the next period.

Images