Intelligent joint control system and adaptive optimization method for urban pipe network multi-pump station based on dynamic coupling model of swmm

Abstract

The application discloses an intelligent joint control system and an adaptive optimization method for a city pipe network multi-pump station based on a SWMM dynamic coupling model, and the system is aimed at the topological characteristics of the city pipe network, takes the SWMM model as the core, and comprises the following five modules: a global perception module, an SWMM model prediction control module, a collaborative optimization module, an intelligent joint control execution module and a health assessment module. The application provides an intelligent joint control system and an adaptive optimization method for a city pipe network multi-pump station based on a SWMM dynamic coupling model, which is suitable for the collaborative operation optimization and dynamic regulation of the multi-pump station and the multi-gate in the city pipe network, and can realize the global hydraulic collaboration, energy consumption optimization and facility protection of the pipe network.

Description

Technical Field

[0001] This invention relates to the field of urban drainage technology, and specifically to an intelligent multi-pump station control system and adaptive optimization method for urban pipe networks based on SWMM (storm management model). Background Technology

[0002] Urban pipe networks are the core infrastructure for collecting and transporting domestic sewage and rainwater. Through the synergistic effect of multi-stage pumping stations and pipe network transportation, sewage is safely delivered to downstream treatment facilities. These networks are characterized by "dense nodes, complex topology, strong hydraulic coupling, and large fluctuations in operating conditions," making traditional pumping station control models ill-suited to their complex operational needs. Existing technologies, while some research focuses on energy saving at single pumping stations or simple pipe network collaborative optimization, have not developed an integrated solution covering "water volume prediction, global coordination, operating condition adaptation, and pipe network protection," and lack accurate simulation of complex hydraulic processes, thus failing to meet the dynamic operational requirements of urban pipe networks.

[0003] SWMM (Storm Water Management Model), as a comprehensive drainage system simulation tool, can simultaneously simulate processes such as rainfall runoff, surface runoff, pipeline transmission, pump station operation, and reservoir regulation, thus providing a more comprehensive reflection of the coupling characteristics of the pipeline network in terms of "hydraulic-water quality-operating conditions." However, existing research has not fully utilized SWMM's dynamic simulation capabilities to achieve coordinated optimization control of multiple pump stations. Therefore, it is necessary to construct an intelligent multi-pump station joint control system based on the SWMM dynamic coupling model to address the problems of insufficient coordination, poor adaptability to operating conditions, and damage to pipeline health inherent in traditional control modes.

[0004] Therefore, based on the above situation, this application conducts further research on intelligent joint control systems for multiple pumping stations. Summary of the Invention

[0005] To address the shortcomings of the existing technologies, this invention provides an intelligent joint control system and adaptive optimization method for multiple pumping stations in urban pipe networks based on the SWMM dynamic coupling model. This system is applicable to the coordinated operation optimization and dynamic control of multiple pumping stations and gates in urban pipe networks, and can realize global hydraulic coordination, energy consumption optimization, and facility protection of the pipe network.

[0006] The present invention is solved by the following technical solution.

[0007] The intelligent joint control system for multiple pumping stations in urban pipeline networks based on the SWMM dynamic coupling model is designed for the topological characteristics of urban pipeline networks. It is centered on the SWMM model and includes the following five modules.

[0008] Global perception module: It integrates water level of pipeline nodes, flow rate of pipeline section, energy consumption of pumping station, flow rate of rainwater inlet and short-term rainfall forecast data to build a state perception network with multi-source data fusion.

[0009] SWMM Model Predictive Control Module: Based on the SWMM model, it constructs a hydraulic coupling prediction framework, combining a hot start mechanism and dynamic time step to achieve accurate simulation and prediction of pipeline flow, water level, water depth and siltation risk in the next 3-6 hours.

[0010] Collaborative optimization module: It integrates adaptive genetic algorithm and Bayesian optimization algorithm to solve the multi-objective optimization problem of "energy saving-drainage-pipeline protection" for different working conditions, and generate the globally optimal pump station start-up / stop / frequency and gate opening control strategy.

[0011] Intelligent joint control execution module: It converts the optimization strategy into execution commands for pump station start / stop / frequency and gate opening, and sends them to each device through the industrial bus, and verifies the feasibility of the commands in real time.

[0012] Health assessment module: Based on SWMM simulation results, the system operating status is assessed from four dimensions: drainage efficiency, energy consumption level, pipeline health, and operating condition adaptability.

[0013] In a preferred embodiment, the collaborative optimization module employs a multi-objective optimization function, combining genetic algorithms and Bayesian optimization algorithms, to dynamically adjust the optimization strategy according to the working condition type; the working condition type includes dry season, rainy season, and special periods.

[0014] In a preferred embodiment, the intelligent joint control execution module: based on the pipeline topology relationship simulated by SWMM, a "hierarchical-regional" collaborative execution strategy is designed, including hierarchical collaborative constraints, regional collaborative constraints, and safety constraint verification; the hierarchical collaborative constraint condition is: upstream drainage rate ≤ downstream remaining capacity; the regional collaborative constraint condition is: unified target water level within the region, water level difference between regions ≤ 0.3m; the safety constraint verification condition is: continuous operation of pumping station ≤ 8 hours, pipe section flow velocity ∈ [0.6, 2.5] m / s.

[0015] The adaptive optimization method in the intelligent joint control system of multiple pumping stations in urban pipe networks based on the SWMM dynamic coupling model in this application is characterized by the following steps: S10: Hydraulic coupling prediction model based on SWMM: For the "hydraulic-siltation" coupling characteristics of urban pipe networks, a prediction framework is constructed with the SWMM model as the core. This prediction framework includes SWMM hydraulic simulation, siltation risk prediction, hot start and dynamic step size; S20: Adaptive collaborative optimization algorithm for operating conditions: For the differences in operating conditions of urban pipe networks during "dry season-rainy season-special periods", a multi-objective optimization mechanism based on SWMM simulation results is designed. This multi-objective optimization mechanism includes a multi-objective optimization function and an algorithm collaboration mechanism; S30: Intelligent joint control execution strategy: For the hydraulic coupling characteristics of upstream and downstream pumping stations in drainage pipe networks, combined with the pipe network topology relationship simulated by SWMM, a "hierarchical-regional" collaborative execution strategy is designed. The collaborative execution strategy includes: hierarchical collaboration, regional collaboration, and safety constraint verification; S40: Health assessment: Based on the SWMM simulation results, a four-dimensional evaluation index is constructed to quantify the intelligent joint control effect.

[0016] In a preferred embodiment, the intelligent joint control execution strategy in step S30 specifically includes the following steps: S31: Data acquisition at the perception layer: Real-time acquisition of water level, pipe flow, pump station current, rainwater inlet flow, and meteorological radar / satellite rainfall forecast data at each node through a sensor network, inputting them into the SWMM model as initial conditions for simulation; S32: Operating condition identification: Based on the water volume fluctuation rate and flow velocity change rate simulated by SWMM, combined with human experience thresholds, determining whether the current period is the dry season, rainy season, or a special period; S33: Predictive simulation: Loading the hot start state, running the SWMM model to predict the pipe network status for the next 3-6 hours, and outputting a dynamic hydraulic scenario; S34: Collaborative optimization: Based on... Adjust the SWMM prediction time domain according to the operating condition type, call the corresponding algorithm to solve the multi-objective optimization problem, and generate control strategies for pump station start / stop / frequency and gate opening; S35: Strategy decomposition and safety verification: decompose the global strategy into execution instructions for each pump station and gate, verify whether the instructions meet the safety constraints, and re-optimize the instructions that do not meet the constraints through a genetic algorithm; S36: Joint execution: send instructions to the pump station PLC and gate controller through industrial Ethernet to synchronously execute the actions of upstream and downstream pump stations and trigger real-time calibration of the SWMM model; S37: Effect evaluation and feedback: calculate four-dimensional evaluation indicators based on the execution results of SWMM simulation, and feed them back to the optimization module to adjust the algorithm parameters for the next round.

[0017] In a preferred embodiment, in step S34, the following multi-objective optimization function is used for optimization:

[0018] J=ω1·Cenergy + ω2·Cdelay + ω3·Cpipe

[0019] Where: Cenergy is the energy consumption cost; Cdelay is the drainage delay cost; Cpipe is the pipeline health cost; ω1, ω2, ω3 are dynamic weights.

[0020] In a preferred embodiment, in step S34, a multi-objective optimization function is used, combining genetic algorithm and Bayesian optimization algorithm, to dynamically adjust the optimization strategy according to the working condition type.

[0021] In a preferred embodiment, the genetic algorithm employs a "pump station-gate" joint encoding, achieves global search through adaptive crossover and mutation, and evaluates individual fitness using SWMM simulation results; the Bayesian optimization is based on Gaussian process modeling of the objective function, employs a "desired improvement + constraint handling" strategy for local fine-tuning, and updates the surrogate model with real-time data from SWMM simulation; the switching logic between the genetic algorithm and Bayesian optimization is based on automatic switching according to the working condition complexity index α, using Bayesian optimization when α < 0.4, and enabling a fusion strategy of genetic algorithm + Bayesian optimization when α ≥ 0.4.

[0022] Compared with the prior art, the present invention has the following beneficial effects: it provides an intelligent joint control system and adaptive optimization method for multiple pumping stations in urban pipe networks based on the SWMM dynamic coupling model, which is applicable to the coordinated operation optimization and dynamic control of multiple pumping stations and multiple gates in urban pipe networks, and can realize global hydraulic coordination, energy consumption optimization and facility protection of the pipe network. Detailed Implementation

[0023] This application relates to an intelligent joint control system and adaptive optimization method for multiple pumping stations in urban pipe networks based on the SWMM dynamic coupling model. The system is designed for the topological characteristics of urban pipe networks and is based on the SWMM model, comprising the following five modules.

[0024] Global perception module: It integrates water level of pipeline nodes, flow rate of pipeline section, energy consumption of pumping station, flow rate of rainwater inlet and short-term rainfall forecast data to build a state perception network of multi-source data fusion, and provides real-time input parameters (such as surface runoff coefficient, pipeline roughness, pumping station characteristic curve, etc.) for SWMM model to support subsequent prediction and optimization.

[0025] SWMM Model Predictive Control Module: Based on the SWMM model, it constructs a hydraulic coupling prediction framework, combines a hot start mechanism with dynamic time steps, and realizes accurate simulation of pipeline flow, water level, water depth and siltation risk in the next 3-6 hours, supporting forward-looking optimization decisions.

[0026] Collaborative optimization module: It integrates adaptive genetic algorithm and Bayesian optimization algorithm to solve the multi-objective optimization problem of "energy saving-drainage-pipeline protection" for different operating conditions (dry season, rainy season, special period) and generate the globally optimal pump station start-up / stop / frequency and gate opening control strategy.

[0027] Intelligent joint control execution module: It converts the optimization strategy into execution commands for pump station start / stop / frequency and gate opening, and sends them to each device through the industrial bus, and verifies the feasibility of the commands in real time (such as continuous operation time of pump station and flow velocity constraints of pipe section) to ensure that the control is implemented.

[0028] Health assessment module: Based on SWMM simulation results, it assesses the system's operating status from four dimensions: drainage efficiency, energy consumption level, pipe network health (frequency of water accumulation, degree of flow velocity deviation, wear risk), and operating condition adaptability, providing feedback to the optimization module and forming a continuous improvement closed loop.

[0029] The following section provides a detailed explanation of the intelligent joint control system for multiple pumping stations in urban pipe networks based on the SWMM dynamic coupling model, using specific adaptive optimization methods.

[0030] (1) Hydraulic coupling prediction model based on SWMM: A prediction framework is constructed with the SWMM model as the core, taking into account the "hydraulic-siltation" coupling characteristics of urban pipe networks.

[0031] SWMM hydraulic simulation: Utilizing SWMM's dynamic flow simulation function, it simulates water levels at pipe network nodes, flow rates in pipe sections, pumping volumes at pumping stations, and surface water accumulation processes. Its core equations include the continuity equation and the momentum equation (SWMM's built-in hydraulic formulas), specifically in the following form:

[0032] Continuity equation:

[0033] Momentum equation:

[0034] Where Q is the flow rate, A is the flow area, h is the water level, q(x,t) is the side inflow (such as rainwater inflow), S0 is the pipe bottom slope, and S f Let g be the friction slope, and g be the acceleration due to gravity.

[0035] Siltation risk prediction: Based on the long-term deviation between the pipe segment flow velocity (v) simulated by SWMM and the critical flow velocity (vcritical, usually taken as 0.6-0.9 m / s), the pipe network siltation risk index is predicted, as shown in the following formula:

[0036]

[0037] Where k is the sedimentation coefficient, τ is the time window, and the integral term reflects the cumulative effect of the flow velocity being higher than the critical value for a long period of time.

[0038] Hot start and dynamic step size:

[0039] Hot start: Saves parameters such as node water level, pipe flow, and pump station operating status of the current SWMM simulation, avoiding recalculation from the initial state for each prediction, improving calculation efficiency by more than 50%;

[0040] Dynamic step size: A 600-second step size is used when the water volume is stable during the dry season, and the step size is automatically switched to 120 seconds when the water volume suddenly increases during the rainy season, balancing prediction accuracy and computational efficiency.

[0041] (2) Working condition adaptive collaborative optimization algorithm

[0042] To address the differences in operating conditions of urban pipeline networks during the dry season, rainy season, and special periods, a multi-objective optimization mechanism based on SWMM simulation results is designed:

[0043] Operating conditions Core Objectives Prediction Time Domain Optimization Algorithm Key constraints (based on SWMM simulation results) Dry season (pure sewage) Energy conservation priority 4 hours Bayesian optimization The water level fluctuation at the node is ≤ ±0.3m, and the pumping station efficiency is ≥ 65%. Rainy season (mixed rainwater) Drainage priority 6 hours Genetic Algorithm + Bayesian Optimization Water depth ≤ 0.15m (no flooding), pipe flow velocity ≤ 2.5m / s (wear prevention) Special period Stability First 3 hours Enhanced Genetic Algorithm Flow velocity ≥ 0.6 m / s (to prevent siltation); node water level ≤ 90% of design water level (to prevent overflow).

[0044] Multi-objective optimization function:

[0045]

[0046] in:

[0047] • Cenergy represents the energy cost (power consumption of the pumping station, calculated based on the SWMM pumping station power curve).

[0048] • Cdelay is the cost of drainage delay (pipeline residence time, calculated by SWMM simulation of water flow path length and velocity).

[0049] • Cpipe represents the health cost of the pipeline network (the degree to which the flow velocity deviates from the critical value, calculated based on the flow velocity distribution from long-term SWMM simulations).

[0050] ω1, ω2, and ω3 are dynamic weights (ω2 increases to 0.6 during the rainy season and ω1 increases to 0.5 during the dry season).

[0051] Algorithm collaboration mechanism:

[0052] Genetic Algorithm: A joint encoding of "pump station-gate" is adopted (1 bit of pump station indicates start and stop, and 3 bits of gate indicates opening degree). Global search is achieved through adaptive crossover (Pc=0.7−0.9) and mutation (Pm=0.01−0.05). The fitness of individuals (such as water depth and energy consumption) is evaluated using SWMM simulation results.

[0053] Bayesian optimization: Based on Gaussian process modeling of the objective function, a "expected improvement + constraint handling" strategy (EIconstrained) is used for local fine-tuning optimization, and the surrogate model is updated with real-time data simulated by SWMM.

[0054] Switching logic: Automatic switching is performed based on the working condition complexity index α (calculated based on the water volume fluctuation rate and flow velocity change rate simulated by SWMM). When α < 0.4, Bayesian optimization is used, and when α ≥ 0.4, a fusion strategy of genetic algorithm + Bayesian optimization is enabled.

[0055] (3) Intelligent joint control execution strategy: Based on the hydraulic coupling characteristics of upstream and downstream pumping stations in the drainage pipeline network and combined with the pipeline network topology relationship simulated by SWMM, a "hierarchical-regional" collaborative execution strategy is designed:

[0056] Hierarchical collaboration:

[0057] The upstream pumping station dynamically adjusts the drainage rate based on the remaining drainage capacity of the downstream pumping station to avoid downstream overload: Q_upstream adjustment = min(Q_upstream current, Q_downstream remaining capacity).

[0058] Where, Q downstream remaining capacity = Q downstream design - Q downstream current (calculated by SWMM simulation of the maximum transmission capacity of the downstream pipeline network).

[0059] Regional coordination: The pipeline network is divided into several regions, with pump stations within each region operating synchronously and water levels between regions linked through boundary nodes.

[0060] Within the region: a unified target water level (2.0m in the dry season and 1.5m in the rainy season) is used to verify the hydraulic balance of the regional pipe network through SWMM simulation;

[0061] Between regions: the outlet water level of the upstream region ≤ the inlet water level of the downstream region +0.3m (to prevent backflow), and the hydraulic connection between the upstream and downstream regions is verified by SWMM simulation.

[0062] Safety constraint verification:

[0063] Pumping stations: Single pump continuous operation ≤8 hours (overheat protection), start-stop interval between adjacent pumping stations ≥3 minutes (water impact protection, verified by SWMM simulation of water hammer effect);

[0064] Pipeline network: Node water level ≤ 90% of design water level (overflow prevention, SWMM simulated water depth), pipe section flow velocity ∈ [0.6, 2.5] m / s (siltation / wear prevention, based on SWMM long-term flow velocity statistics).

[0065] (4) The health assessment system is based on SWMM simulation results, constructs four-dimensional assessment indicators, and quantifies the effect of intelligent joint control:

[0066] Evaluation Dimensions Key Indicators Optimization objective (based on SWMM simulation) Drainage efficiency Pipeline utilization rate (actual flow / design flow), drainage delay time Utilization rate ≥ 80%, delay ≤ 30 minutes (SWMM simulated dwell time) Energy consumption level Energy consumption per unit water volume (kWh / m³), pump station efficiency Energy consumption ≤ 0.25 kWh / m³, efficiency ≥ 65% (calculated from the pump station power curve) Pipeline health Flow velocity deviation from critical value index, pipe wear rate Deviation index ≤ 0.3 (standard deviation of flow velocity from critical value), wear rate ≤ 1% / year (based on long-term flow velocity distribution) Operating condition adaptability Target water level compliance rate and parameter adjustment response speed Compliance rate ≥ 90% (SWMM simulated water level tracking), response time ≤ 5 minutes

[0067] In this application, the system control flow is a closed-loop cycle, and the specific steps are as follows:

[0068] 1. Data acquisition at the perception layer: Real-time data on water level at each node, flow rate in pipe section, pump station current, rainwater inlet flow rate, and meteorological radar / satellite rainfall forecast (1-hour resolution) are acquired through the sensor network and input into the SWMM model as initial conditions.

[0069] 2. Operating Condition Identification: Based on the water flow fluctuation rate (ΔQ / Q average) and flow velocity change rate simulated by SWMM, combined with human experience thresholds, it is determined whether the current period is the dry season, rainy season, or a special period (such as centralized discharge of industrial wastewater).

[0070] 3. Predictive Simulation: Load the hot start state (save the node water level, pipe flow rate, and pump station operation status of the previous SWMM simulation), run the SWMM model to predict the pipe network status (flow rate, water level, water depth, and flow velocity) for the next 3-6 hours, and output the dynamic hydraulic scenario.

[0071] 4. Collaborative optimization: Adjust the SWMM prediction time domain according to the operating condition type (6 hours in the rainy season / 4 hours in the dry season), call the corresponding algorithm to solve the multi-objective optimization problem (the objective function combines the energy consumption, drainage delay and pipeline health indicators simulated by SWMM), and generate control strategies for pump station start / stop / frequency and gate opening.

[0072] 5. Strategy Decomposition and Safety Verification: The global strategy is decomposed into execution instructions for each pump station and gate (e.g., "Pump station P1: Run; Gate G3: 60% opening"). The instructions are verified to meet safety constraints (pump station start-stop interval, pipe section flow velocity range). Instructions that do not meet the requirements are re-optimized using a genetic algorithm.

[0073] 6. Joint Control Execution: Sends instructions to the pump station PLC and gate controller via industrial Ethernet to synchronously execute the actions of upstream and downstream pump stations and triggers real-time calibration of the SWMM model (correcting model parameters using measured data 15 minutes after execution).

[0074] 7. Effect evaluation and feedback: Based on the execution results of SWMM simulation, calculate four-dimensional evaluation indicators and feed them back to the optimization module to adjust the algorithm parameters for the next round (such as dynamic weights and working condition complexity index α).

[0075] As can be seen from the above description, this application provides an intelligent joint control system and adaptive optimization method for multiple pumping stations in urban pipe networks based on the SWMM dynamic coupling model. It is applicable to the coordinated operation optimization and dynamic control of multiple pumping stations and multiple gates in urban pipe networks, and can realize global hydraulic coordination, energy consumption optimization and facility protection of the pipe network.

[0076] The scope of protection of this invention includes, but is not limited to, the above embodiments. The scope of protection of this invention is defined by the claims. Any substitutions, modifications, or improvements to this technology that are easily conceived by those skilled in the art fall within the scope of protection of this invention.

Claims

1. An urban pipe network multi-pump station intelligent joint control system based on a SWMM dynamic coupling model, characterized in that, The system is aimed at the topological characteristics of urban pipe network, and includes the following five modules based on the SWMM model: Global perception module: It integrates pipe network node water level, pipe section flow, pump station energy consumption, rainwater inlet flow and short-term rainfall forecast data to build a state perception network of multi-source data fusion. SWMM model prediction control module: It builds a hydraulic coupling prediction framework based on the SWMM model, combines the hot start mechanism and dynamic time step, and realizes the accurate simulation of pipe network flow, water level, waterlogging depth and siltation risk in the next 3-6 hours. Collaborative optimization module: It integrates adaptive genetic algorithm and Bayesian optimization algorithm to solve the multi-objective optimization problem of "energy saving-drainage-pipe network protection" for different working conditions, and generates the globally optimal pump station start-stop / frequency and gate opening control strategy. The working conditions include dry season, rainy season and special period. According to the type of working condition, the SWMM prediction time domain is adjusted, and the corresponding algorithm is called to solve the multi-objective optimization problem to generate the control strategy of pump station start-stop / frequency and gate opening. Intelligent joint control execution module: It converts the optimization strategy into execution instructions of pump station start-stop / frequency and gate opening, and sends them to each device through industrial bus, and checks the feasibility of the instructions in real time. Health assessment module: Based on the SWMM simulation results, the system running state is evaluated from four dimensions of drainage efficiency, energy consumption level, pipe network health and working condition adaptation degree.

2. The urban pipe network multi-pump station intelligent joint control system based on the SWMM dynamic coupling model according to claim 1, characterized in that, In the collaborative optimization module, a multi-objective optimization function is used, combined with genetic algorithm and Bayesian optimization algorithm, to dynamically adjust the optimization strategy according to the type of working condition.

3. The urban pipe network multi-pump station intelligent joint control system based on the SWMM dynamic coupling model according to claim 2, characterized in that, In the intelligent joint control execution module, based on the pipe network topology relationship simulated by SWMM, a "hierarchical-regional" collaborative execution strategy is designed, which includes hierarchical collaborative constraints, regional collaborative constraints and safety constraint verification. The hierarchical collaborative constraint condition is: upstream drainage rate ≤ downstream residual capacity. The regional collaborative constraint condition is: unified target water level in the region, and water level difference between regions ≤ 0.3 m. The safety constraint verification condition is: pump station continuous operation ≤ 8 hours, pipe section flow rate ∈ [0.6, 2.5] m / s.

4. The adaptive optimization method in the intelligent joint control system of urban pipe network multi-pump stations based on the SWMM dynamic coupling model according to any one of claims 1 to 3, characterized in that, The steps include: S10: SWMM-based hydraulic coupling prediction model: A prediction framework is built based on the SWMM model for the "hydraulic-siltation" coupling characteristics of urban pipe network, which includes SWMM hydraulic simulation, siltation risk prediction, hot start and dynamic step length. S20: Working condition adaptive collaborative optimization algorithm: A multi-objective optimization mechanism based on SWMM simulation results is designed for the differences in "dry season-rainy season-special period" working conditions of urban pipe network, which includes a multi-objective optimization function and an algorithm collaboration mechanism. S30: Intelligent joint control execution strategy: A "hierarchical-regional" collaborative execution strategy is designed based on the hydraulic coupling characteristics of upstream and downstream pump stations of the drainage pipe network and the pipe network topology relationship simulated by SWMM. The collaborative execution strategy includes hierarchical collaboration, regional collaboration and safety constraint verification. S40: Health assessment: Based on the SWMM simulation results, four-dimensional evaluation indicators are built to quantify the effect of intelligent joint control.

5. The urban pipe network multi-pump station intelligent joint control system and adaptive optimization method based on the SWMM dynamic coupling model according to claim 4, characterized in that, The intelligent joint control execution strategy in step S30 specifically includes the following steps: S31: Perception layer data collection: Real-time acquisition of water level, pipe segment flow, pump station current, rainwater inlet flow, and weather radar / satellite rainfall forecast data by a sensor network, input to the SWMM model as initial conditions for simulation; S32: Working condition identification: Based on the water fluctuation rate and flow rate change rate of SWMM simulation, combined with artificial experience threshold, determine whether it is a dry season, a rainy season, or a special period; S33: Prediction simulation: Load the hot start state and run the SWMM model to predict the pipe network state in the next 3-6 hours, output the dynamic hydraulic scenario; S35: Strategy decomposition and safety check: Decompose the global strategy into execution instructions for each pump station and gate, check if the instructions meet the safety constraints, and re-optimize the instructions that do not meet the safety constraints through genetic algorithm; S36: Joint control execution: Issue instructions to pump station PLC and gate controller through industrial Ethernet, synchronize the actions of upstream and downstream pump stations, and trigger real-time calibration of the SWMM model; S37: Effect evaluation and feedback: Based on the execution results of SWMM simulation, calculate four-dimensional evaluation indicators and feedback to the optimization module to adjust the algorithm parameters for the next round.

6. The urban pipe network multi-pump station intelligent joint control system and adaptive optimization method based on the SWMM dynamic coupling model according to claim 5, characterized in that, In step S34, the following multi-objective optimization function is used for optimization: J = ω1 · Cenergy + ω2 · Cdelay + ω3 · Cpipe Where: Cenergy is the energy cost; Cdelay is the drainage delay cost; Cpipe is the pipe network health cost; ω1, ω2, ω3 are dynamic weights.

7. The urban pipe network multi-pump station intelligent joint control system and adaptive optimization method based on the SWMM dynamic coupling model according to claim 6, characterized in that, In step S34, a multi-objective optimization function is used, combined with genetic algorithm and Bayesian optimization algorithm, to dynamically adjust the optimization strategy according to the working condition type.

8. The urban pipe network multi-pump station intelligent joint control system and adaptive optimization method based on the SWMM dynamic coupling model according to claim 7, characterized in that, The genetic algorithm: Adopting "pump station-gate" joint coding, realizing global search through adaptive crossover and mutation, and using SWMM simulation results to evaluate individual fitness; The Bayesian optimization: Based on Gaussian process modeling of the objective function, using the "expected improvement + constraint processing" strategy for local fine optimization, and updating the surrogate model with real-time data from SWMM simulation; Switching logic between genetic algorithm and Bayesian optimization: Automatically switch according to the working condition complexity index α, use Bayesian optimization when α < 0.4, and enable the fusion strategy of genetic algorithm + Bayesian optimization when α ≥ 0.4.

Images