Intelligent city pipe network overflow emergency supervision internet of things large model system and method

Abstract

The present application provides a smart city pipe network overflow emergency supervision Internet of Things large model system and method, and relates to the field of gas pipeline supervision.The method comprises the following steps: obtaining pipe network monitoring data based on sensors deployed in the drainage pipe network through an emergency supervision object platform; generating overflow estimation data through a prediction model based on the pipe network monitoring data and future weather data; in response to the overflow estimation data meeting overflow conditions, generating overflow control parameters based on the overflow estimation data, wherein the overflow control parameters comprise at least one of the following: catch basin parameters, pump station parameters, and storage and regulation facility parameters, and the pump station parameters comprise the operating power and pumping direction of a target water pump; sending the overflow control parameters to the emergency supervision object platform and controlling the working parameters of the overflow emergency equipment.The method realizes the rational scheduling of water quantity, reduces the frequency and influence of overflow occurrence, and improves the management level of urban drainage pipe networks.

Description

Technical Field

[0001] This invention belongs to the field of gas pipeline supervision, and specifically relates to a large-scale IoT model system and method for emergency monitoring of gas pipeline overflow in smart cities. Background Technology

[0002] In some urban areas, combined sewer systems are used, where domestic sewage and rainwater flow into the same pipe network. On sunny days, this system can normally transport sewage to wastewater treatment plants for processing. However, during heavy rains, large amounts of rainwater flood the network, causing the total volume to far exceed the design capacity of existing pipe networks and wastewater treatment plants. To prevent urban flooding and sewage backflow, combined sewer systems typically have overflow outlets at locations such as interceptor wells, allowing the excess mixed rainwater and sewage to be directly discharged into nearby rivers, lakes, or other receiving water bodies. This phenomenon is called a combined sewer overflow (CSO).

[0003] CSO incidents can cause serious environmental problems. Untreated overflowing stormwater and sewage contain large amounts of suspended solids, organic pollutants, heavy metals, and pathogenic microorganisms, posing a serious pollution and threat to the water quality of receiving water bodies, aquatic ecosystems, and public health.

[0004] How to accurately predict the occurrence time and total overflow volume of CSO events, and achieve refined and proactive control of overflow emergency equipment in drainage pipe networks to reduce overflow volume and environmental pollution risks, is a key technical problem that urgently needs to be solved in the fields of smart city construction and water environment management. Summary of the Invention

[0005] The invention includes a smart city pipeline overflow emergency monitoring IoT big data model system. The system includes an emergency monitoring management platform configured to: acquire pipeline monitoring data based on sensors deployed within the drainage pipeline network through an emergency monitoring object platform; generate overflow prediction data using a predictive model based on the pipeline monitoring data and future weather data; and generate overflow control parameters based on the overflow prediction data if the overflow prediction data meets overflow conditions. The overflow control parameters include at least one of interceptor well parameters, pump station parameters, and storage facility parameters. This includes determining the operating power and extraction direction of the target water pump; sending the overflow control parameters to the emergency monitoring platform and controlling the operating parameters of the overflow emergency equipment, including: controlling the opening degree of the gate based on the interception well parameters, wherein the gate is deployed at the overflow outlet of the interception well; controlling the target water pump to operate with its corresponding extraction direction and operating power based on the pump station parameters; controlling the opening of the target pipeline valve to discharge the water extracted by the target water pump to the drainage area; and controlling the opening degree of the inlet valve and outlet valve of the storage facility based on the storage facility parameters; wherein the storage facility is connected to the drainage network.

[0006] The invention includes a method for emergency monitoring of overflow in smart city pipe networks. The method is executed by an emergency monitoring management platform and includes: acquiring pipe network monitoring data based on sensors deployed within the drainage pipe network through an emergency monitoring platform; generating overflow prediction data using a predictive model based on the pipe network monitoring data and future weather data; and generating overflow control parameters based on the overflow prediction data in response to the overflow prediction data meeting overflow conditions. The overflow control parameters include at least one of interceptor well parameters, pump station parameters, and storage facility parameters, wherein the pump station parameters include the parameters of the target water pump. The system controls the operating power and extraction direction; it sends the overflow control parameters to the emergency monitoring platform and controls the operating parameters of the overflow emergency equipment, including: controlling the opening degree of the gate based on the interception well parameters, wherein the gate is deployed at the overflow outlet of the interception well; controlling the target water pump to operate with its corresponding extraction direction and operating power based on the pump station parameters; controlling the opening of the target pipeline valve to discharge the water extracted by the target water pump to the drainage area; and controlling the opening degree of the inlet valve and outlet valve of the storage facility based on the storage facility parameters; wherein the storage facility is connected to the drainage network.

[0007] This invention integrates sensors deployed within drainage pipe networks to acquire multi-dimensional network monitoring data in real time. Combined with future weather data, it uses predictive models to accurately estimate potential overflow events. Based on the predicted overflow situation, it dynamically adjusts the hydraulic conditions of the pipe network to achieve reasonable scheduling and allocation of water, thereby reducing the frequency and impact of overflows, effectively mitigating environmental pollution caused by overflows, and improving the resilience and refined management level of urban drainage pipe networks. Attached Figure Description

[0008] This specification will be further described by way of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting; in these embodiments, the same reference numerals denote the same structures, wherein:

[0009] Figure 1 This is a schematic diagram of the platform structure of a large-scale IoT model system for emergency monitoring of overflow in smart city pipeline networks, as shown in some embodiments of this specification. Figure 2 This is an exemplary flowchart of a smart city pipeline overflow emergency monitoring method according to some embodiments of this specification; Figure 3 This is a schematic diagram illustrating the generation of overflow prediction data according to some embodiments of this specification; Figure 4 This is an exemplary flowchart illustrating the determination of the prediction model type according to some embodiments of this specification; Figure 5 This is an exemplary flowchart illustrating the generation of overflow control parameters according to some embodiments of this specification. Detailed Implementation

[0010] Figure 1 This is a schematic diagram of the platform structure of a smart city pipeline overflow emergency monitoring IoT big data model system according to some embodiments of this specification.

[0011] In some embodiments, the smart city pipeline overflow emergency monitoring IoT big data model system 100 can be used to monitor, predict, warn of, and respond to overflow events in urban drainage pipelines.

[0012] like Figure 1 As shown, the smart city pipeline overflow emergency monitoring IoT big data model system 100 includes an emergency monitoring user platform 110, an emergency monitoring service platform 120, an emergency monitoring management platform 130, an emergency monitoring sensor network platform 140, and an emergency monitoring object platform 150 connected in sequence. In some embodiments, one or more platforms in the smart city pipeline overflow emergency monitoring IoT big data model system 100 can exchange information and / or data through a network.

[0013] Emergency monitoring user platform 110 refers to a platform used for interacting with users. In some embodiments, emergency monitoring user platform 110 can be used to provide data visualization and push notifications. In some embodiments, emergency monitoring user platform 110 may include various mobile terminal devices.

[0014] In some embodiments, the emergency monitoring user platform 110 can interact with the emergency monitoring management platform 130 through the emergency monitoring service platform 120. For example, the emergency monitoring user platform 110 can obtain monitoring data information uploaded by the emergency monitoring management platform 130, or the emergency monitoring user platform 110 can issue monitoring request instructions to the emergency monitoring management platform 130.

[0015] The emergency monitoring service platform 120 refers to a platform used to transmit user instructions and control information. The emergency monitoring service platform 120 can interact with the emergency monitoring user platform 110 and the emergency monitoring management platform 130.

[0016] The emergency monitoring and management platform 130 refers to a platform used to monitor and manage data related to the smart city pipeline overflow emergency monitoring IoT big data model system 100. The emergency monitoring and management platform 130 can interact with the emergency monitoring service platform 120 and the emergency monitoring sensor network platform 140.

[0017] The Emergency Monitoring and Management Platform 130 refers to the central platform responsible for the overall management, data processing, decision generation, and instruction issuance of emergency monitoring of urban pipeline overflows. For example, the Emergency Monitoring and Management Platform 130 can receive data from the platform of the emergency monitoring object to perform overflow prediction and generate control strategies.

[0018] In some embodiments, the emergency monitoring and management platform 130 can be implemented using a variety of hardware and software components. For example, the hardware may include memory, computing hardware (such as servers, processors, etc.), and gateways.

[0019] In some embodiments, the emergency monitoring and management platform 130 may further include a data center, which may include a database, a data processing model library, and a computing unit. The database stores various types of data required for system operation, such as pipeline monitoring data, future weather data, overflow prediction data, and historical control parameters. The data processing model library stores various emergency monitoring data processing models and prediction models. The computing unit performs data calculations and model inference; it can call corresponding data processing models or prediction models from the data processing model library as needed, and retrieve corresponding data from the database for processing.

[0020] The emergency monitoring sensor network platform 140 refers to a functional platform for sensor communication. In some embodiments, the emergency monitoring sensor network platform 140 can be configured as a communication network and gateway, etc.

[0021] In some embodiments, the emergency monitoring sensor network platform 140 can interact with the emergency monitoring management platform 130 and the emergency monitoring object platform 150.

[0022] Emergency monitoring platform 150 refers to the platform responsible for collecting emergency monitoring data and executing emergency control commands. For example, emergency monitoring platform 150 can connect to sensors deployed within the drainage network to obtain network monitoring data and control overflow emergency equipment such as intercepting wells, pumping stations, and storage facilities. For more information on sensors, please refer to [link to relevant documentation]. Figure 2 The corresponding content.

[0023] Some embodiments of this specification, the smart city pipeline overflow emergency monitoring IoT big data model system 100, can form an information operation closed loop between various platforms, unified management for coordinated operation, and realize the informatization and intelligentization of urban overflow management.

[0024] Figure 2 This is an exemplary flowchart of a smart city pipeline overflow emergency monitoring method according to some embodiments of this specification.

[0025] Step 210: Obtain pipeline monitoring data through the emergency monitoring platform based on sensors deployed in the drainage pipeline network.

[0026] A drainage network refers to an underground pipeline system used to collect, transport, and discharge urban rainwater, sewage, or mixed rainwater and sewage. For example, a drainage network includes facilities such as pipes, manholes, intercepting wells, and pumping stations.

[0027] Sensors are used to monitor the operational status of drainage pipe networks in real time. Examples of sensors include rain gauges, level gauges, flow meters, gate opening sensors, and pump station status sensors.

[0028] Pipeline monitoring data refers to various types of data regarding the operational status of drainage pipeline networks. For example, pipeline monitoring data includes one or more of the following: rainfall data, water level data, flow rate data, or equipment operational status data. Equipment operational status data may include operational status data such as the on / off status, current, power, and operating time of equipment like pumps.

[0029] In some embodiments, the emergency monitoring and management platform acquires network monitoring data from multiple sensors deployed within the drainage network through the emergency monitoring object platform.

[0030] For example, the emergency monitoring and management platform uses the emergency monitoring object platform to monitor rainfall intensity and cumulative rainfall from rain gauges deployed within the drainage network, and rationally distributes these gauges according to terrain to comprehensively understand the regional rainfall situation. As another example, the emergency monitoring and management platform uses the emergency monitoring object platform to monitor information such as pump current, power, and operating time from pump station status sensors deployed within the drainage network.

[0031] Step 220: Based on pipeline monitoring data and future weather data, obtain overflow prediction data through a prediction model.

[0032] Future weather data refers to various types of weather conditions over a period of time in the future. For example, future weather data includes future rainfall intensity, rainfall area, rainfall duration, and temperature. Future weather data can be provided by weather radar and other sources.

[0033] A predictive model is a model used to estimate or infer future events, states, or trends based on input data. In some embodiments, predictive models include hydrological simulation models for simulating hydrodynamic processes and hydrographic model models for prediction based on graph structures. More detailed descriptions can be found in [link to relevant documentation]. Figure 3 The corresponding explanation is provided in the text.

[0034] Overflow prediction data refers to data related to forecasting possible future overflow events. For example, overflow prediction data includes one or more of the following: predicted water level curves for future time points, predicted overflow locations, and overflow flow rate curves. A predicted water level curve reflects the water level at various locations (such as drain pipes, overflow outlets, etc.). An overflow flow rate curve reflects the overflow flow rate at each time point from the start to the end of an overflow at a given location.

[0035] In some embodiments, the emergency monitoring and management platform inputs current pipeline monitoring data as initial conditions and future weather data as driving conditions into the prediction model, thereby driving the prediction model to perform calculations to obtain overflow prediction data. More information on obtaining overflow prediction data can be found in the relevant descriptions below.

[0036] Step 230: In response to the overflow prediction data meeting the overflow conditions, overflow control parameters are generated based on the overflow prediction data.

[0037] Overflow conditions refer to preset conditions or thresholds used to determine whether an overflow event is about to occur or has already occurred. For example, overflow conditions could be that the estimated overflow volume exceeds a certain threshold, or that the water level in the pipeline network reaches a preset warning height.

[0038] Overflow control parameters are control commands or setpoints used to adjust overflow emergency equipment to manage or reduce overflow events. Overflow control parameters include at least one of the following: interceptor well parameters, pump station parameters, and storage facility parameters.

[0039] Interception well parameters refer to the parameters used to control the gates or other components within the interception well to regulate water flow. For example, interception well parameters include the opening degree of the interception well gates.

[0040] Pump station parameters refer to parameters used to control the operating status of pumps and the direction of water flow within a pump station. For example, pump station parameters include the operating power, start / stop status, and pumping direction of the target pump.

[0041] The target pump refers to the specific pump selected in a pumping station to perform pumping or drainage tasks.

[0042] Operating power refers to the rate at which a device consumes or outputs energy during operation. For example, operating power can be the electrical power consumed by a water pump when drawing water, or it can be a percentage of its rated power.

[0043] The extraction direction refers to the direction in which the water pump draws water from one area and delivers it to another.

[0044] Storage facility parameters refer to parameters used to control the influent and effluent conditions of the storage facility. For example, storage facility parameters include the opening degree of the influent valve and the opening degree of the effluent valve.

[0045] In some embodiments, the emergency monitoring and management platform determines whether the acquired overflow prediction data meets preset overflow conditions. For example, if it is predicted that the water level in the pipeline network will exceed the height of its overflow weir, or the estimated overflow volume will reach or exceed a preset warning threshold, then the overflow conditions are considered met.

[0046] In some embodiments, the emergency monitoring and management platform can generate overflow control parameters based on preset rules. For example, the platform can pre-divide the urban drainage network into several independent or semi-independent control sub-regions, and preset a series of "IF-THEN" control rules for the overflow emergency equipment (such as interceptor gates, pumping station pumps, and storage facility valves) within each sub-region. The triggering conditions of the preset rules are associated with overflow prediction data. For example, one rule could be: "If the predicted water level exceeds threshold A, then set the interceptor gate opening to parameter X"; or "If the predicted water level rise rate exceeds threshold B, then activate target pumping station Y." The emergency monitoring and management platform can directly convert the prediction results into specific control parameters through table lookup or mapping.

[0047] Step 240: Send the overflow control parameters to the emergency monitoring platform and control the operating parameters of the overflow emergency equipment.

[0048] Overflow emergency equipment refers to on-site facilities used to manage water flow or reduce the impact of overflows by adjusting their operating status in the event of an overflow or potential overflow incident. Examples of overflow emergency equipment include intercepting wells, pumping stations, storage facilities, and related gates and valves.

[0049] Operating parameters refer to the specific operational settings used to define or adjust the equipment when it performs its functions. For example, the operating parameters of overflow emergency equipment include the opening degree of the intercepting well gate, the operating power and pumping direction of the pumping station pumps, and the opening degree of the inlet and outlet valves of the storage facility.

[0050] Step 241: Based on the parameters of the interception well, control the opening of the gate, and deploy the gate at the overflow outlet of the interception well.

[0051] An intercepting well is a structure used to intercept and transport sewage to a wastewater treatment plant when it is not raining, but to allow some of the mixed rainwater and sewage to overflow when the water volume exceeds the carrying capacity during rainy days. For example, an intercepting well typically includes a pipe connecting to the wastewater treatment plant and an overflow outlet located on higher ground, with the water flow controlled by a gate.

[0052] Opening degree refers to the extent to which a gate, valve, or other regulating device is opened, usually expressed as a percentage or angle. For example, the opening degree of a gate can be 50%, indicating that it is opened to the halfway position.

[0053] In some embodiments, the emergency monitoring and management platform issues control commands to the intercepting well on-site through the emergency monitoring object platform, based on the gate number and corresponding opening value specified in the intercepting well parameters. Upon receiving the command, the corresponding gate's electric actuator adjusts the gate's opening degree according to the opening degree carried in the command. For example, when a brief but rapid rainfall peak is approaching upstream, the emergency monitoring and management platform automatically closes the gate of the intercepting well, ensuring that the downstream sewage treatment plant is not overloaded.

[0054] Step 242: Based on the pump station parameters, control the target water pump to operate in its corresponding pumping direction and operating power.

[0055] A pumping station is a device that pumps water from a lower elevation to a higher elevation, or transfers water from one area to another. Pumping stations allow for proactive intervention and scheduling of water flow within a pipe network system.

[0056] In some embodiments, the emergency monitoring and management platform, based on the target pump number, operating power, and extraction direction specified in the pump station parameters, issues control commands to the target pump within the pump station through the emergency monitoring object platform, thereby controlling the pump's start and stop, adjusting its operating frequency or power, and setting the water extraction direction.

[0057] Step 243: Control the opening of the target pipeline valve to discharge the water pumped by the target water pump to the drainage area.

[0058] A target pipeline valve is a specific valve used to control the connection or disconnection of water flow between different pipelines or areas. For example, when a pumping station draws water from pipeline A to pipeline B, the target pipeline valve is the valve between pipeline A and pipeline B.

[0059] A drainage area refers to a designated geographical area or facility to which water is directed or discharged. For example, a drainage area could be another section of drainage pipe, a storage facility, or a nearby receiving water body.

[0060] In some embodiments, when the emergency monitoring and management platform determines that water needs to be transferred via a target pump, it searches the database based on the pumping and drainage areas specified in the pump station parameters to determine the target pipeline valve that needs to be opened to connect the pumping and drainage areas. This target pipeline valve is typically closed when not pumping or draining water. The emergency monitoring and management platform issues an opening command to this valve to create a water flow channel, ensuring that the water pumped by the target pump can be smoothly discharged to the designated drainage area.

[0061] Step 244: Based on the parameters of the water storage facility, control the opening degree of the inlet valve and the outlet valve of the water storage facility.

[0062] Storage and regulation facilities are structures used to temporarily store rainwater or mixed rainwater and sewage to regulate the water volume in drainage networks. These facilities are connected to the drainage network. They can reduce flood peaks and minimize overflow pollution. For example, storage and regulation facilities may include underground reservoirs or deep tunnels, connected to the drainage network via inlet and outlet valves.

[0063] The opening degree of the inlet valve refers to the degree to which the valve used by the regulating and storage facility to control the inflow of water is opened.

[0064] The opening degree of the outlet valve refers to the degree to which the valve of the water storage facility is opened to control the discharge of water.

[0065] In some embodiments, storage facilities are used to store rainwater far exceeding the capacity of the drainage network system, thereby achieving "peak shaving" and pollution reduction. Storage facilities are typically connected to different pipes in the drainage network via multiple inlets and outlets.

[0066] In some embodiments, the emergency monitoring and management platform, based on the storage facility number, inlet valve opening, and outlet valve opening specified in the storage facility parameters, issues control commands to the electric valves of the on-site storage facility through the emergency monitoring object platform, thereby adjusting the opening degree of the inlet and outlet valves to control the amount of water entering and leaving the storage facility.

[0067] For example, when extremely heavy rainfall is predicted for the next hour, and severe overflow cannot be avoided regardless of adjustments to gates and pumping stations, storage facilities can be activated. The emergency monitoring and management platform calculates the optimal activation time and inlet valve opening, directing the initial rainwater with the highest pollution concentration (the "first wave of dirty water" flushing the surface and pipes when it first starts raining) into the pool.

[0068] This specification describes a smart city pipeline overflow emergency monitoring IoT big data system in some embodiments. By integrating sensors deployed within the drainage pipeline network, it acquires multi-dimensional pipeline monitoring data in real time. Combined with future weather data, it uses a predictive model to accurately predict potential overflow events. Based on the predicted overflow situation, it dynamically adjusts the hydraulic conditions of the pipeline network to achieve reasonable scheduling and allocation of water, thereby reducing the frequency and impact of overflows, effectively reducing environmental pollution caused by overflows, and improving the resilience and refined management level of the urban drainage pipeline network.

[0069] Figure 3 This is a schematic diagram illustrating the generation of overflow prediction data according to some embodiments of this specification.

[0070] In some embodiments, the prediction model may include a hydrological simulation model and a hydrographic model. The hydrological simulation model may be a physical model, and the hydrographic model may be a neural network model.

[0071] In some embodiments, such as Figure 3 As shown, the emergency monitoring and management platform can determine the duration to be predicted; based on the duration to be predicted, determine the prediction model type; in response to the determined prediction model type being a hydrological map model 330, construct an urban pipe network map 320 based on the structural data 311 of the drainage pipe network, the pipe network monitoring data 312, and the future weather data 313; and use the hydrological map model 330 to generate overflow prediction data 340 based on the urban pipe network map 320.

[0072] Hydrological simulation models are models that simulate the movement, distribution, and transformation of water bodies in geographic space and time based on physical laws and hydrological cycle processes. Examples include the Storm Water Management Model (SWMM) and the Hydrological Simulation Program (HSPF).

[0073] Hydrological graph model 330 refers to a model that represents a hydrological system (such as a drainage network) as a graph structure and uses graph analysis or graph neural networks for data processing and prediction.

[0074] In some embodiments, the emergency monitoring and management platform can determine the duration to be predicted.

[0075] The duration to be predicted refers to the length of the future time range that needs to be predicted. For example, the duration to be predicted could be 30 minutes, 1 hour, or 6 hours in the future.

[0076] In some embodiments, the emergency monitoring and management platform can determine the forecast duration by analyzing future weather data. For example, the platform can acquire rainfall intensity data for a specific future time period and calculate the regional average rainfall intensity. A higher average rainfall intensity indicates a more severe future rainfall situation, and a shorter forecast duration is preferred for a faster and more timely response. Conversely, a lower average rainfall intensity allows for a longer forecast duration to obtain more comprehensive information. In some embodiments, the platform can determine the forecast duration based on rainfall intensity by querying a preset table. The preset table is constructed based on rainfall intensity and forecast duration.

[0077] In some embodiments, the emergency monitoring and management platform can determine the overflow prototype pattern based on future weather data; determine the type of equipment to be controlled based on the overflow prototype pattern; and determine the forecast duration based on the type of equipment to be controlled.

[0078] Overflow prototype patterns refer to overflow event types with typical characteristics formed after summarizing and classifying historical overflow events. For example, overflow prototype patterns include "single-point sudden thunderstorm pattern in the west of the city" and "rain belt moving from south to north pattern".

[0079] In some embodiments, the emergency monitoring and management platform can collect historical overflow event data, summarize and classify the historical overflow event data, and determine overflow prototype patterns. For example, the emergency monitoring and management platform can collect historical records of overflow events that have occurred over a long period of time, which include rainfall data at the time, response data of the drainage network (i.e., operating parameters of overflow emergency equipment), etc.

[0080] For example, the emergency monitoring and management platform can use clustering algorithms to cluster the collected historical overflow event data to obtain overflow prototype patterns with typical characteristics. For instance, clustering algorithms such as K-Means, Density-Based Spatial Clustering of Applications with Noise (DBSCAN), or methods combining autoencoders with clustering can be used.

[0081] In some embodiments, the overflow prototype pattern can be determined by an expert experience rule base, by identification based on a deep learning model, or by judgment based on preset thresholds and conditions.

[0082] In some embodiments, the emergency monitoring and management platform can determine the type of device to be controlled based on a defined overflow prototype pattern.

[0083] The type of equipment to be controlled refers to the type of emergency overflow equipment that is predicted to require regulation in a specific overflow scenario. For example, the type of equipment to be controlled could be a gate in a diversion well, a pumping station pump, or a valve in a storage facility. In the event of an overflow incident in the future, the emergency monitoring and management platform intends to use specific types of emergency overflow equipment for regulation.

[0084] In some embodiments, the emergency monitoring and management platform can determine the type of equipment to be controlled based on the overflow prototype pattern corresponding to the current or predicted future weather, and identify the most suitable type of overflow emergency equipment for regulation. For example, the emergency monitoring and management platform can use classification algorithms (e.g., decision trees, support vector machines, K-nearest neighbors, etc.) to analyze future weather data and determine which identified overflow prototype pattern the future weather scenario belongs to.

[0085] For example, once the current overflow prototype pattern is identified, the emergency monitoring and management platform can extract the reference controllable equipment type and its quantity corresponding to the pattern. For instance, if the current scenario is determined to be a "single-point sudden thunderstorm pattern in the west of the city", and historical statistics show that this pattern usually requires the control of interception well gates, with an average number of 30, the emergency monitoring and management platform will identify "interception well gates" as the controllable equipment type and record its corresponding quantity.

[0086] In some embodiments, the type of device to be controlled may be determined by rules preset by experts, based on risk assessment results, or by selecting from multiple device types through a multi-objective optimization algorithm.

[0087] In some embodiments, the emergency monitoring and management platform can determine the duration to be predicted based on the identified type of device to be controlled.

[0088] In some embodiments, the emergency monitoring and management platform can set a suitable future prediction time range based on the type and quantity of the devices to be controlled, thereby determining the prediction duration. For example, the emergency monitoring and management platform can determine the prediction duration by querying a preset table, which includes the type of device to be controlled and its corresponding control response time or the required prediction duration.

[0089] Some embodiments in this specification determine the overflow prototype pattern based on future weather data, determine the type of equipment to be controlled based on the overflow prototype pattern, and determine the forecast duration based on the type of equipment to be controlled. This makes the forecasting process more purposeful and targeted, improves the efficiency and real-time nature of emergency decision-making, optimizes the utilization of computing resources, and ensures that sufficient, timely and focused overflow forecast data is provided at critical moments. This further enhances the emergency monitoring and management platform's ability to cope with urban flooding and overflows.

[0090] In some embodiments, the emergency monitoring and management platform may determine the prediction model type based on the determined duration to be predicted.

[0091] The prediction model type refers to the specific type or combination of prediction models that can be used to generate overflow forecast data. For example, the prediction model type can be a hydrological simulation model, a hydrological map model, or a combination of both.

[0092] In some embodiments, the emergency monitoring and management platform can select between a hydrological simulation model and a hydrological map model based on the length of the forecast duration. For example, when the forecast duration is less than or equal to a duration threshold, a more computationally efficient hydrological map model can be selected for prediction; when the forecast duration is greater than the duration threshold, a hydrological simulation model can be selected. The forecast duration can be preset by the system.

[0093] In some embodiments, the prediction model type can also be determined in various other ways. For example, in addition to considering the duration of the prediction, factors such as the availability of computing resources, the required prediction accuracy, historical prediction results, and the complexity of the current pipeline network can be comprehensively considered to determine the prediction model type. More information on determining the prediction model can be found in [link to relevant documentation]. Figure 4 The corresponding description.

[0094] Structural data 311 of drainage pipe networks refers to various types of data that describe static information such as the physical layout, topological connections, facility dimensions, and materials of drainage pipe networks. For example, structural data of drainage pipe networks includes pipe length, diameter, slope, material, and the geographical location and specifications of facilities such as inspection wells, intercepting wells, and pumping stations.

[0095] The urban pipe network map 320 refers to the graphical structure data that reflects the facilities (such as manholes, pumping stations, and overflow outlets) and pipe connections in the urban drainage pipe network. The urban pipe network map 320 consists of several nodes 321 and several edges 322.

[0096] In some embodiments, node 321 may correspond to a manhole, pumping station, gate or storage facility valve, etc., and edge 322 may correspond to a drainage pipe.

[0097] In some embodiments, the node characteristics of node 321 include the geographical location (e.g., latitude and longitude), altitude information, real-time water level data, and real-time flow data of the facility corresponding to node 321. In some embodiments, if node 321 is a pumping station, gate, or regulating valve, its corresponding node characteristics also include its current operating parameters. In some embodiments, the node characteristics also include the current rainfall intensity and the predicted rainfall intensity at various future time points associated with the area where node 321 is located.

[0098] In some embodiments, the edge features of edge 322 include the length, diameter, slope, and material roughness of the pipe corresponding to the edge. The material roughness may include the material type of the pipe, such as concrete or polyvinyl chloride (PVC).

[0099] like Figure 3 As shown, the emergency monitoring and management platform constructs an urban pipeline network map 320 based on the structural data 311 of the drainage pipeline network, the pipeline network monitoring data 312, and the future weather data 313.

[0100] For example, the structural data 311 of the drainage network can include construction drawings, engineering drawings, and Geographic Information System (GIS) data of the urban pipeline network. The structural data describes the length, diameter, slope, and material of the pipelines, as well as the geographical location and specifications of facilities such as inspection wells, intercepting wells, pumping stations, overflow outlets, and storage facilities. The emergency monitoring and management platform can read the structural data from the database.

[0101] In some embodiments, the emergency monitoring and management platform can traverse the aforementioned structural data or GIS data to abstract the main facilities in the drainage network, such as each pumping station, intercepting well, overflow outlet, storage facility, and key inspection well, into nodes of the urban network map. Each node is assigned a unique identifier (ID).

[0102] In some embodiments, the emergency monitoring and management platform can establish connections between abstracted nodes based on the structural data of the drainage network or the pipeline connection relationships described in GIS data, forming edges in the graph. For example, if water can flow from facility A to facility B through a pipeline, a directed edge from A to B is established.

[0103] In some embodiments, the emergency monitoring and management platform can determine pipeline health data based on pipeline monitoring data; and construct an urban pipeline map based on pipeline health data, drainage pipeline structural data, and pipeline monitoring data.

[0104] Pipeline health data refers to various types of data used to assess the structural integrity, operational efficiency, and potential failure risks of pipes or equipment in a drainage network. For example, pipeline health data includes the degree of siltation, damage, or blockage risk level of pipes.

[0105] In some embodiments, determining pipeline health data may include based on real-time or periodic analysis of pipeline monitoring data. Further details regarding pipeline monitoring data can be found in the corresponding descriptions above.

[0106] For example, an emergency monitoring and management platform can use a pre-defined rule base to identify anomalies, thereby determining pipeline health data. For instance, if the nighttime flow rate of a certain pipeline segment consistently exceeds a preset threshold, it is determined that there may be pipeline damage or groundwater infiltration. The emergency monitoring and management platform can then upgrade the damage risk level of that pipeline segment (e.g., from low risk to medium risk) as part of the pipeline health data.

[0107] In some embodiments, constructing an urban pipeline network map involves integrating multiple data sources into a unified, visualized, or queryable data model.

[0108] For example, in one embodiment, the emergency monitoring and management platform first uses the structural data of the drainage network (e.g., the length, diameter, slope, material, and connection relationship of the pipes) to establish the basic topology of the urban network, including nodes (e.g., manholes, pumping stations) and edges (e.g., drainage pipes).

[0109] Based on this, the emergency monitoring and management platform further attaches pipeline health data and pipeline monitoring data (such as real-time water level, flow rate, rainfall, etc.) as dynamic attributes or related information to the corresponding nodes or edges of the pipeline map.

[0110] In some embodiments of this specification, by determining the health data of the pipe network based on the pipe network monitoring data, and constructing an urban pipe network map together with the structural data and monitoring data, the real-time operation and health status of the urban drainage pipe network can be dynamically and comprehensively reflected. By dynamically updating the pipe network map, the long-term accuracy of the subsequent prediction model is improved, and the long-term reliability of the prediction results is guaranteed.

[0111] like Figure 3 As shown, the inputs to the hydrological map model 330 include the urban pipe network map 320, and the generated overflow prediction data 340.

[0112] In some embodiments, the hydrological graph model includes a graph convolutional network (GCN), a graph attention network (GAN), or other graph neural network (GNN) models.

[0113] In some embodiments, the emergency monitoring and management platform can acquire a training dataset to train a hydrological map model. The training dataset includes training samples and their corresponding labels. Training samples can be historical urban pipe network maps, while labels represent water level curves and overflow locations and flow rates at subsequent time points collected through actual monitoring. Subsequent time points refer to the time points within the predicted duration, starting from the current time point. Water level curves include water level changes at various drainage pipes, overflow outlets, etc. Overflow flow rate curves include the predicted overflow flow rate at each time point from the start to the end of the overflow. Historical urban pipe network maps and labels can be obtained based on historical data.

[0114] In some embodiments, the training process of the hydrographic map model includes performing multiple iterations. At least one iteration includes: selecting one or more training samples from the training dataset, inputting the one or more training samples into the initial hydrographic map model to obtain the predicted outputs corresponding to the one or more training samples; substituting the predicted outputs corresponding to the one or more training samples and the labels of these one or more samples into a predefined loss function formula to calculate the value of the loss function; and updating the model parameters in the initial model in reverse using methods such as gradient descent based on the value of the loss function. When the iteration termination condition is met, the iteration ends, and the trained model is obtained.

[0115] Some embodiments in this specification significantly improve the efficiency and adaptability of urban drainage network overflow prediction by introducing hydrological map models as the prediction model type and dynamically selecting the prediction model according to the prediction duration. The strategy of intelligently selecting models according to prediction needs optimizes the utilization of computing resources and provides reliable and timely overflow prediction data, thereby providing more flexible and accurate decision support for emergency monitoring and management platforms and reducing the risks caused by urban flooding and overflow.

[0116] In some embodiments, the emergency monitoring and management platform may, in response to a pipeline health data indication that a drainage pipeline meets the dredging conditions, determine a flushing time based on the future weather data; at the flushing time, generate a cleaning control command and send the cleaning control command to the flushing valve associated with the drainage pipeline to perform the pipeline flushing operation.

[0117] Drainage pipes refer to specific pipe sections within a drainage network used to transport water. For example, a drainage pipe can be a main pipe for collecting domestic sewage or a branch pipe for transporting rainwater.

[0118] Dredging conditions refer to preset standards or thresholds used to determine whether silt, debris, or other contaminants have accumulated in drainage pipes and whether dredging and maintenance are required. For example, dredging conditions could include an abnormal rise in water level within the pipe at low flow rates, or pipe network health data indicating a blockage exceeding a certain risk level. Another example is that the degree of blockage in the pipe network exceeds a preset threshold.

[0119] Flushing time refers to a pre-set time point or period of time suitable for performing flushing operations on drainage pipes. For example, flushing time could be during off-peak hours at night or during periods when no rainfall is predicted.

[0120] In some embodiments, when pipeline health data indicates that a drainage pipe meets dredging conditions, the emergency monitoring and management platform determines an appropriate flushing time based on future weather data. This dredging condition may be that the drainage pipe's blockage score exceeds a preset threshold in the pipeline health data.

[0121] For example, when pipeline health data indicates that drainage pipes meet the dredging conditions, the emergency monitoring and management platform will select the period of no rainfall predicted in future weather data as the preferred flushing time.

[0122] Cleaning control commands are instructions issued to flushing valves or other related equipment to start, regulate, or stop pipeline flushing operations. For example, cleaning control commands may include the opening time of flushing valves, flushing pressure, and flushing duration.

[0123] A flushing valve is a specialized valve installed on drainage pipes to control the water supply for flushing the pipes and removing internal debris. For example, a flushing valve can be connected to an offline flushing well or utilize water from the upstream main pipeline network for pipe flushing.

[0124] Pipe flushing is a maintenance activity that involves injecting a high flow rate of water into drainage pipes to remove accumulated silt and debris. For example, pipe flushing is typically performed during off-peak flow periods to restore the pipes' water flow capacity.

[0125] In some embodiments, the emergency monitoring and management platform generates corresponding cleaning control instructions during the flushing process. These instructions typically include parameters for operating the flushing valves, such as the valve opening time, flushing duration, and flushing pressure.

[0126] For example, the emergency monitoring and management platform dynamically adjusts the flushing parameters in the cleaning control command based on the degree of blockage in the drainage pipe. If the pipeline health data shows that the pipe is severely clogged, the command can set a longer flushing duration and / or a higher flushing pressure to ensure effective dredging.

[0127] Flushing valves associated with drainage pipes include special valves installed at specific locations in the pipe network (e.g., upstream of a gentle section of pipe prone to siltation).

[0128] In another embodiment, the emergency monitoring and management platform sends the cleaning control command to a flushing valve associated with a designated drainage pipe. Upon receiving the command, the flushing valve executes a pipe flushing operation according to the command parameters. For example, the valve may automatically open, allowing a high flow rate of water to enter the drainage pipe, thereby flushing away silt and debris inside the pipe and restoring its flow capacity. The flushing operation is typically performed during off-peak flow periods to minimize the impact on the daily drainage system.

[0129] In some embodiments, in addition to directly controlling the valve opening, the cleaning control command can also be used to adjust the flow rate or start / stop frequency of the flushing valve to adapt to different pipeline conditions and environmental requirements.

[0130] In some embodiments of this specification, the flushing time is intelligently determined by combining future weather data, and a cleaning control command is generated and sent to execute the pipeline flushing operation. This not only enables a timely and accurate response to the problem of siltation in drainage pipelines, but also optimizes the timing of flushing operations by taking into account weather and flow factors, reducing the impact on the daily operation of the drainage system, improving the efficiency and safety of dredging and maintenance, and thus significantly enhancing the overall emergency management capability and resilience of the urban pipe network.

[0131] Figure 4 This is an exemplary flowchart illustrating the determination of the prediction model type according to some embodiments of this specification.

[0132] Step 410: Based on future weather data, obtain future weather characteristics.

[0133] Future weather characteristics refer to quantitative attributes or indicators extracted from future weather data to describe future meteorological conditions. For example, future weather characteristics include rainfall coverage, peak flood intensity, rainfall center mobility, and rainfall volatility.

[0134] In some embodiments, the emergency monitoring and management platform can divide the urban pipeline service area into multiple grids and calculate future weather characteristics based on the rainfall intensity value of each grid at each future time step or point in time.

[0135] In some embodiments, the emergency monitoring and management platform can also obtain future weather characteristics through other means. For example, the platform can obtain more future weather characteristics by analyzing other meteorological elements such as temperature, wind speed, and humidity in future weather data, or by combining them with Geographic Information System (GIS) data.

[0136] Step 420: Determine the network status characteristics based on the network monitoring data and the structural data of the drainage network.

[0137] Network status characteristics refer to quantitative attributes or indicators used to describe the current operating status of a drainage network. For example, network status characteristics include network connectivity, network saturation, hydraulic gradient consistency, and the activity level of control equipment.

[0138] In some embodiments, the emergency monitoring and management platform can extract information such as water level, flow rate, and pump station operating status from real-time or historical pipeline monitoring data, and combine it with structural data such as the topology, pipe size, and slope of the drainage pipeline network to perform calculations to determine the network status characteristics.

[0139] For example, the emergency monitoring and management platform can determine network connectivity by calculating the average nodal degree of the urban pipe network map (reflecting the tightness of the connection between different parts of the drainage pipe network); determine pipe network saturation by calculating the proportion of monitoring nodes whose water level exceeds the warning line (e.g., exceeding 80% of the pipe diameter); determine hydraulic gradient consistency by calculating the proportion of pipes whose actual hydraulic gradient direction is opposite to the physical gradient direction; and determine the activity level of control equipment by calculating the total number of pump stations in "open" or "working" state and partially closed gates and valves (i.e., commissioning facilities).

[0140] In some embodiments, the emergency monitoring and management platform can also determine the pipeline network status characteristics through other means. For example, it can determine more dimensions of pipeline network status characteristics based on water quality data, foreign object accumulation, and pipeline aging.

[0141] Step 430: Determine the model call parameters based on future weather characteristics, pipeline status characteristics, and the duration to be predicted.

[0142] Model invocation parameters refer to the set of parameters used to guide the selection of prediction models and the execution plan. For example, model invocation parameters include a specific invocation time and its corresponding prediction model type.

[0143] The invocation time refers to the specific point in time or time period during the forecasting process when the forecasting model is activated or its forecast results are used. For example, the invocation time for a hydrological simulation model is 0-10 minutes in the future.

[0144] In some embodiments, the emergency monitoring and management platform can generate model invocation parameters by predicting and selecting a model. This predictive selection model can be a machine learning model, such as a Multilayer Perceptron (MLP), a Gradient Boosting Decision Tree (XGBoost), or a Light Gradient Boosting Machine (LightGBM).

[0145] The inputs to the predictive selection model include future weather characteristics (such as rainfall coverage and flood peak intensity), pipeline network status characteristics (such as network connectivity and pipeline network saturation), and the duration to be predicted.

[0146] The output of the prediction selection model includes model invocation parameters, i.e., the invocation scheme of the prediction model. For example, the output can be a standalone invocation of a graph neural network (GNN) model, a standalone invocation of a hydrological simulation model, or a combination of invocation schemes.

[0147] The training samples for the prediction selection model include sample pipeline state features, sample future weather features, and the duration of the sample to be predicted. The labels for the prediction selection model are the corresponding model call parameters.

[0148] The training samples and labels for the prediction selection model can be obtained based on historical data. For example, historical events that successfully suppressed overflows or achieved satisfactory overflow effects can be used to extract the corresponding sample network status features, sample future weather features, and the predicted duration (which can be the actual historical predicted duration). Backtesting can then be performed on the historical events using both GNN and SWMM to obtain the errors between the model's prediction results and the actual recorded data at different time points under different predicted durations. These two error curves can be plotted, and their intersection point can be found. The time corresponding to the intersection point is the optimal switching point for that historical event. The label can be either "Use a model with smaller error before the optimal switching point" or "Use another model with smaller error after this point."

[0149] The training process for the prediction selection model can be similar to that for the reference hydrographic model.

[0150] In some embodiments, the model invocation parameters can also be determined through preset rules, expert systems, or reinforcement learning. For example, a specific model can be directly invoked based on preset rainfall intensity levels and pipeline saturation thresholds.

[0151] In some embodiments, the emergency monitoring and management platform may, in response to a number of determined prediction model types being greater than one, combine the overflow prediction data predicted by the determined prediction models according to the call time to obtain updated overflow prediction data.

[0152] In some embodiments, the emergency monitoring and management platform can determine the number of currently available or activated prediction model types based on preset configurations or real-time system status. For example, multiple prediction models may be configured, such as hydrological map models, hydrological simulation models, and deep learning-based graph neural network models. When at least two of these models are activated or selected for prediction, the number of prediction model types is considered to be greater than one.

[0153] In some other embodiments, the emergency monitoring and management platform can obtain a list of currently running or about-to-run predictive models by querying data in a processor or database. If the list contains predictive models of different types and the number exceeds one, then the number of predictive model types is greater than one. For example, when a GNN model and an SWMM model are used for prediction simultaneously, the number of model types is two.

[0154] In some embodiments, the system may also dynamically determine whether to enable multiple different types of prediction models to improve prediction accuracy based on historical data or expert experience.

[0155] In some embodiments, in response to a number of determined prediction model types being greater than one, the emergency monitoring and management platform combines the overflow prediction data predicted by the determined prediction models according to the call time to obtain updated overflow prediction data.

[0156] In some embodiments, the emergency monitoring and management platform receives overflow prediction data output from different prediction models. This data may be for different time periods or different prediction targets. The emergency monitoring and management platform integrates this data according to the call timestamp of each model and a preset combination strategy. For example, if the GNN model is for short-term prediction and the SWMM model is for medium- to long-term prediction, the emergency monitoring and management platform will concatenate the short-term prediction results of the GNN model with the medium- to long-term prediction results of the SWMM model according to the timeline.

[0157] In some embodiments, the emergency monitoring and management platform can combine the short-term prediction results of the hydrological map model with the medium- to long-term prediction results of the hydrological simulation model.

[0158] In some embodiments, at the “seams” between prediction results from different models, the emergency monitoring and management platform can use a smoothing algorithm to process the data to avoid abrupt changes in the prediction curve.

[0159] The emergency monitoring and management platform can also combine the overflow prediction data predicted by the established prediction model in various other ways, including but not limited to linear combination, nonlinear combination, decision tree combination, etc.

[0160] In some embodiments, after performing the above-described combined operations, the emergency monitoring and management platform can generate and obtain updated overflow prediction data. This data is a prediction result that integrates the advantages of multiple models, including overflow prediction information over a longer time range or with higher accuracy.

[0161] For example, after combining and smoothing the GNN and SWMM models, the emergency monitoring and management platform will store and provide a complete, smooth, time-span predicted water level curve or overflow flow curve, which is the updated overflow prediction data.

[0162] In some embodiments of this specification, when the number of prediction model types applied is greater than one, the prediction results of different prediction models (such as GNN models and SWMM models) in their respective advantageous time periods can be dynamically integrated. This not only improves the quality and naturalness of the overflow prediction results, but also enhances the stability of the downstream optimization algorithm, providing more reliable decision support for the emergency monitoring and management platform.

[0163] Step 440: Determine the prediction model type based on the model call parameters.

[0164] In some embodiments, the emergency monitoring and management platform will determine the prediction model type based on the prediction model type corresponding to the model call parameters.

[0165] For more information on prediction model types, please refer to [link / reference]. Figure 2 The corresponding content.

[0166] In some embodiments of this specification, by dynamically determining the model calling parameters based on future weather characteristics, pipeline network status characteristics, and the duration to be predicted, the emergency monitoring and management platform can intelligently select or combine prediction models according to specific scenarios, thereby significantly improving the accuracy and timeliness of drainage pipeline overflow prediction and providing more accurate and timely decision support for emergency management.

[0167] Figure 5 This is an exemplary flowchart illustrating the generation of overflow control parameters according to some embodiments of this specification.

[0168] Step 510: Determine candidate control parameters based on overflow prediction data.

[0169] More information on overflow prediction data can be found at [link to relevant documentation]. Figure 2 The corresponding description.

[0170] Candidate control parameters refer to parameter configurations that are evaluated during the optimization process and may be used to adjust the operating status of overflow emergency equipment. For example, candidate control parameters may be a combination of a set of gate openings of intercepting wells, pump station operating power, and storage facility valve openings randomly generated at the start of iterative optimization.

[0171] In some embodiments, the emergency monitoring and management platform generates candidate control parameters based on preset rules or historical data analysis. For example, it can generate one or more initial combinations of equipment control parameters as candidate control parameters based on equipment operation strategies under historical rainfall events or on preliminary rules of thumb.

[0172] Step 520: Perform an iterative process based on the candidate control parameters and the prediction model.

[0173] An iterative process is an optimization method that repeatedly executes specific computational steps with the goal of minimizing overall risk and cost. Each step improves upon the result of the previous step until a pre-defined stopping condition is met. For example, in particle swarm optimization, the iterative process includes evaluating the fitness of each candidate control parameter, updating the individual and global optimal positions, and adjusting the candidate control parameters accordingly.

[0174] In some embodiments, the iterative process can be implemented using Particle Swarm Optimization (PSO) or other similar swarm intelligence optimization algorithms, such as Genetic Algorithm (GA), Simulated Annealing (SA), etc.

[0175] The iterative process includes steps 521 to 523 as described below.

[0176] Step 521: Input the candidate control parameters for this round into the prediction model to generate the comprehensive risk cost corresponding to the candidate control parameters for this round.

[0177] The candidate control parameters for this round refer to a specific set of candidate control parameters that are input into the prediction model for evaluation and calculation at the current stage of the iteration process. For example, in the Nth iteration of the particle swarm optimization algorithm, the candidate control parameters for this round are the set of device parameter configurations represented by the current particle swarm.

[0178] In some embodiments, the emergency monitoring and management platform can input the candidate control parameters for each round into the prediction model, and the prediction model can output the overflow prediction data corresponding to the candidate control parameters for that round. The predicted overflow for each round can be calculated based on the overflow prediction data output for that round. In some embodiments, the comprehensive risk cost is calculated as an evaluation of the fitness of the candidate control parameters.

[0179] The overall risk cost refers to an indicator that quantifies the combined environmental, operational, and economic impacts of a specific overflow control scheme. A lower overall risk cost generally indicates a more effective overflow control scheme.

[0180] In some embodiments, the emergency monitoring and management platform can calculate the comprehensive risk cost by weighting factors such as predicted overflow, environmental risk weight, equipment operating costs, and pollutant concentration.

[0181] In some embodiments, the emergency monitoring and management platform determines the comprehensive risk cost based on the predicted overflow volume of the overflow outlet, environmental risk weights, and equipment operating costs output by the prediction model.

[0182] Predicted overflow refers to an estimate of the total volume of water discharged from an overflow outlet. For example, predicted overflow may include the volume of mixed stormwater and sewage.

[0183] Environmental risk weights refer to the relative risk coefficients assigned to an overflow event at a spillway based on factors such as the importance, environmental sensitivity, and environmental capacity of the receiving water body into which it flows. For example, an overflow outlet flowing into a drinking water source protection area may have a higher environmental risk weight, while an overflow outlet flowing into an ordinary river may have a lower weight. Environmental risk weights can be preset based on experience.

[0184] Equipment operating costs refer to the energy consumption, maintenance, or other related expenses incurred by overflow emergency equipment during the execution of control operations.

[0185] In some embodiments, equipment operating costs may also include equipment wear and tear, maintenance costs, and labor costs. Equipment operating costs can be calculated based on the equipment's operating power, operating time, and the opening and closing frequency of gates and valves.

[0186] In some embodiments, the overall risk cost can also be determined through a multi-objective optimization function, which includes multiple sub-objectives such as minimizing overflow, minimizing equipment operating costs, and minimizing environmental risks.

[0187] In some embodiments, the overall risk cost can be positively correlated with the overflow volume of each overflow outlet, the corresponding environmental risk weight, and the operating cost of each device. For example, the overall risk cost can be determined by the following formula (1).

[0188] (1) In the above formula (1), C To take into account the overall risk costs, w 1 represents the adjustable weighting coefficient corresponding to the overflow flow. w 2 represents the adjustable weighting coefficient corresponding to the equipment operating cost. Vi For the first i The overflow capacity of each overflow outlet ERWi For the first i Environmental risk weights for each overflow outlet C equipmentFor equipment operating costs. If the focus is more on minimizing the total overflow, then... w 1 is set to a relatively large value.

[0189] In some embodiments, the emergency monitoring and management platform can assess environmental risk weights based on predefined rules. For example, if the overflow from the spillway flows into a river, its risk weight can be less than that of flowing into a lake.

[0190] Step 522: In response to the failure to meet the stopping condition, based on the candidate control parameters of this round and the comprehensive risk cost, obtain the updated candidate control parameters as the candidate control parameters for the next round of iteration.

[0191] A stopping condition refers to a preset standard or rule used to terminate an iterative process or the execution of an optimization algorithm. For example, a stopping condition could be reaching the maximum number of iterations, the overall risk cost converging to a certain level, or falling below a preset threshold.

[0192] In some embodiments, during the iteration process, the emergency monitoring and management platform first determines whether the stopping condition is met. If the stopping condition is not met, the iteration continues to find a better control scheme. Obtaining updated candidate control parameters refers to adjusting the position and velocity of each particle in the particle swarm based on the results of the current iteration, thereby generating candidate control schemes for the next round of iteration.

[0193] Step 523: In response to the meeting of the stopping condition, stop the iteration and determine the overflow control parameters based on the candidate control parameters of this round.

[0194] For more information on overflow control parameters, please refer to [link / reference]. Figure 2 The corresponding description.

[0195] In some embodiments, the emergency monitoring and management platform can be continuously iterated until a preset stopping condition is met.

[0196] In some embodiments, the emergency monitoring and management platform determines the optimal solution obtained by the iterative optimization algorithm as the overflow control parameter. In the particle swarm optimization algorithm, the global optimal position Gbest of the output particle swarm is used as the final overflow control parameter.

[0197] In some embodiments, the sensor further includes a water quality sensor for acquiring water quality sensing data. In some embodiments, the emergency monitoring and management platform determines the pollutant concentration based on the water quality sensing data and the predicted overflow rate using a water quality prediction algorithm; and determines the comprehensive risk cost based on the pollutant concentration.

[0198] A water quality sensor is a device used to detect various chemical, physical, or biological indicators in water. For example, a water quality sensor can be used to measure the concentration of chemical oxygen demand (COD), total suspended solids (TSS), or ammonia nitrogen in water.

[0199] Water quality sensing data refers to various data about water quality parameters acquired in real time through water quality sensors. For example, water quality sensing data includes COD concentration, TSS concentration, and ammonia nitrogen concentration in the water.

[0200] In some embodiments, water quality sensors can be deployed at key nodes of the pipeline system, such as in intercepting wells upstream of overflow outlets, at the inlet of pumping stations, or at the outlet of storage facilities, to monitor water quality in real time.

[0201] Water quality prediction algorithms are computational methods used to estimate or infer the concentration of pollutants or other water quality indicators in water bodies based on input data. For example, water quality prediction algorithms can use historical water quality sensor data, pipeline health data, and predicted overflow rates to predict pollutant concentrations through machine learning models or vector database matching.

[0202] Pollutant concentration refers to the level of a specific harmful substance in a body of water. For example, pollutant concentration can be the level of chemical oxygen demand (COD), total suspended solids (TSS), or ammonia nitrogen in the water at an overflow outlet.

[0203] The emergency monitoring and management platform can construct a pollution feature vector by using real-time or near-real-time water quality sensor data from the overflow outlet, the total predicted overflow volume from the wellhead output by the prediction model, and the average rainfall intensity of this rainfall.

[0204] In some embodiments, the vector database pre-stores multiple sets of reference pollution feature vectors and corresponding multiple reference pollutant concentrations. The reference pollutant concentrations are obtained based on historical overflow events.

[0205] In some embodiments, the emergency monitoring and management platform first filters out all combined sewer overflow (CSO) events that have occurred and have complete monitoring data from the system's data center; based on the filtered CSO overflow events, it constructs corresponding reference pollution feature vectors, which include information such as water quality sensor data, overflow volume, and rainfall intensity at the time; the actual pollutant concentrations monitored in these CSO overflow events (which can be the mean or weighted mean, with higher weights for greater hazards) are used as the reference pollutant concentrations for the corresponding reference pollution feature vectors; finally, multiple reference pollution feature vectors and their corresponding reference pollutant concentrations are placed into a vector database (e.g., Milvus, Faiss, etc.) to complete the construction of the vector database.

[0206] When it is necessary to determine the current pollutant concentration at the overflow outlet, the similarity between the current pollution feature vector and each reference pollution feature vector in the vector database is calculated to find one or more most similar reference vectors, and the current pollutant concentration at the overflow outlet is inferred based on their corresponding reference pollutant concentrations.

[0207] In some embodiments, the pollutant concentration can be the average or weighted average of the concentrations of these harmful substances. For example, pollutants that are more toxic or pose a greater environmental hazard can be given a higher weight. In some embodiments, the pollutant concentration can also be dynamically corrected based on real-time environmental monitoring data, upstream pollution source emission information, etc.

[0208] In some embodiments, the emergency monitoring and management platform can also determine pollutant concentrations based on water quality sensor data and pipeline health data, using water quality prediction algorithms.

[0209] That is, the pollution feature vector includes the pipeline health data of the pipe corresponding to the overflow outlet. The corresponding pipe can be a pipe directly connected to the overflow outlet.

[0210] For more information on water quality sensor data, pipeline health data, water quality prediction algorithms, pollution feature vectors, and pollutant concentrations, please refer to the corresponding content above.

[0211] In some embodiments of this specification, by introducing the health data of the pipeline corresponding to the overflow outlet into the pollution feature vector, systematic deviations in scenarios such as heavy rainfall, sudden erosion due to siltation, or fluctuations in groundwater levels are reduced, thereby mitigating scheduling errors and environmental risks caused by water quality prediction errors.

[0212] In some embodiments, in order to more comprehensively assess the combined impact of the overflow, a pollutant concentration factor is further introduced on the basis of the combined risk cost determined by the predicted overflow volume, ERW, and equipment operating costs of the overflow outlet based on the output of the prediction model.

[0213] In some embodiments, the overall risk cost can be positively correlated with the overflow volume of each overflow outlet, the corresponding environmental risk weight, the operating cost of each device, the overflow volume of each overflow outlet, and the corresponding pollutant concentration. For example, the overall risk cost can be determined by the following formula (2).

[0214] (2) In the above formula (2), w 3 represents the adjustable weighting coefficient corresponding to the pollutant concentration. If greater emphasis is placed on pollution control, this coefficient can be adjusted. w The value of 3 is set to be larger; C i For the estimated first iThe pollutant concentration at each overflow outlet during the overflow period can be the individual concentration of harmful substances such as COD, TSS, and ammonia nitrogen, or their weighted average; the relevant explanations of the other coefficients can be found in the description corresponding to formula (1) above.

[0215] Some embodiments in this specification acquire water quality sensing data by introducing water quality sensors, and determine the pollutant concentration at the overflow outlet by combining the predicted overflow flow rate with a water quality prediction algorithm. By incorporating the pollutant concentration into the determination of comprehensive risk costs, the environmental relevance and economic effectiveness of the overflow control strategy are significantly improved, enabling the emergency monitoring and management platform to make smarter and more environmentally responsible decisions.

[0216] Some embodiments in this specification combine the overflow prediction data output by the prediction model with environmental risk weights and equipment operating costs to construct a comprehensive risk cost function. This function can comprehensively and quantitatively evaluate the advantages and disadvantages of different overflow control schemes, effectively improving the efficiency, robustness, and control effect of the optimization algorithm. It can effectively cope with the complex and variable situations of urban stormwater pipe networks under different rainfall scenarios. While reducing the environmental risk of overflow, it also takes into account the optimization of equipment operating costs, which helps to reduce the risk of urban flooding and environmental pollution, and improve the intelligence level of emergency response.

[0217] The basic concepts have been described above. Obviously, for those skilled in the art, the detailed disclosure above is merely illustrative and does not constitute a limitation of this specification. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this specification. Such modifications, improvements, and corrections are suggested in this specification and therefore remain within the spirit and scope of the exemplary embodiments described herein.

[0218] Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments described herein are intended to be illustrative rather than limiting, and should be considered consistent with the teachings of this specification. Accordingly, the embodiments described herein are not limited to those explicitly introduced and described herein.

Claims

1. A large-scale IoT model system for emergency monitoring of overflow in smart city pipeline networks, characterized in that, The system includes an emergency monitoring and management platform, which is configured as follows: The emergency monitoring platform acquires network monitoring data based on sensors deployed within the drainage network. Based on the pipeline monitoring data and future weather data, overflow prediction data is generated through a prediction model. The prediction model includes a hydrological simulation model and a hydrological map model. The hydrological simulation model is a physical model, and the hydrological map model is a neural network model. In response to the overflow prediction data satisfying the overflow conditions, overflow control parameters are generated based on the overflow prediction data. These overflow control parameters include at least one of interceptor well parameters, pump station parameters, and storage facility parameters. The pump station parameters include the operating power and pumping direction of the target pump. The overflow control parameters are sent to the emergency monitoring platform, and the operating parameters of the overflow emergency equipment are controlled, including: Based on the parameters of the intercepting well, the opening degree of the gate is controlled, and the gate is deployed at the overflow outlet of the intercepting well; Based on the pump station parameters, the target water pump is controlled to operate in its corresponding extraction direction and operating power. Control the opening of the target pipeline valve to discharge the water pumped by the target water pump to the drainage area; and, Based on the parameters of the water storage facility, the opening degree of the inlet valve and the outlet valve of the water storage facility are controlled; wherein, the water storage facility is connected to the drainage pipe network; The process of generating overflow prediction data based on the pipeline monitoring data and future weather data through a prediction model includes: Based on the aforementioned future weather data, a prototype overflow model was determined; Based on the overflow prototype pattern, determine the type of device to be controlled; Based on the type of device to be controlled, the duration to be predicted is determined; Based on the aforementioned future weather data, future weather characteristics are obtained; Based on the pipeline network monitoring data and the drainage pipeline network structural data, the pipeline network status characteristics are determined. Based on the future weather characteristics, the pipeline network status characteristics, and the forecast duration, the model calling parameters are determined, wherein the model calling parameters include the calling time and the corresponding prediction model type; The prediction model type is determined based on the model call parameters; In response to the prediction model type being the hydrographic model, the pipeline health data is determined based on the pipeline monitoring data; Based on the pipeline health data, the drainage pipeline structure data, and the pipeline monitoring data, an urban pipeline network map is constructed. Using the aforementioned hydrological map model, based on the urban pipe network map, the overflow prediction data is generated; and, In response to the fact that the number of the prediction model types is greater than one, the overflow prediction data is combined according to the call time to obtain the updated overflow prediction data; The emergency monitoring and management platform is further configured as follows: In response to the pipeline health data indicating that a drainage pipe meets the dredging conditions, the flushing time is determined based on the future weather data; During the flushing time, a cleaning control command is generated and sent to the flushing valve associated with the drain pipe to perform the pipe flushing operation.

2. The system according to claim 1, characterized in that, The emergency monitoring and management platform is further configured as follows: Based on the overflow prediction data, candidate control parameters are determined; and, An iterative process is performed based on the candidate control parameters and the prediction model. The iterative process includes: The candidate control parameters for this round are input into the prediction model to generate the comprehensive risk cost corresponding to the candidate control parameters for this round; wherein, the comprehensive risk cost is determined based on the predicted overflow volume of the overflow outlet, the environmental risk weight, and the equipment operating cost output by the prediction model; In response to the failure to meet the stopping condition, based on the candidate control parameters of this round and the comprehensive risk cost, updated candidate control parameters are obtained and used as candidate control parameters for the next iteration; and, In response to the satisfaction of the stopping condition, the iteration is stopped and the overflow control parameters are determined based on the candidate control parameters of the current round.

3. The system according to claim 2, characterized in that, The sensor also includes a water quality sensor, which is used to acquire water quality sensing data; The emergency monitoring and management platform is further configured as follows: Based on the water quality sensor data and the predicted overflow flow, the pollutant concentration at the overflow outlet is determined using a water quality prediction algorithm; and, The overall risk cost is determined based on the concentration of the pollutant.

4. A method for emergency monitoring of overflow in smart city pipeline networks, characterized in that, The method is executed by the emergency monitoring and management platform, and the method includes: The emergency monitoring platform acquires network monitoring data based on sensors deployed within the drainage network. Based on the pipeline monitoring data and future weather data, overflow prediction data is generated through a prediction model. The prediction model includes a hydrological simulation model and a hydrological map model. The hydrological simulation model is a physical model, and the hydrological map model is a neural network model. In response to the overflow prediction data satisfying the overflow conditions, overflow control parameters are generated based on the overflow prediction data. These overflow control parameters include at least one of interceptor well parameters, pump station parameters, and storage facility parameters. The pump station parameters include the operating power and pumping direction of the target pump. The overflow control parameters are sent to the emergency monitoring platform, and the operating parameters of the overflow emergency equipment are controlled, including: Based on the parameters of the intercepting well, the opening degree of the gate is controlled, and the gate is deployed at the overflow outlet of the intercepting well; Based on the pump station parameters, the target water pump is controlled to operate in its corresponding extraction direction and operating power. Control the opening of the target pipeline valve to discharge the water pumped by the target water pump to the drainage area; and, Based on the parameters of the water storage facility, the opening degree of the inlet valve and the outlet valve of the water storage facility are controlled; wherein, the water storage facility is connected to the drainage pipe network; The process of generating overflow prediction data based on the pipeline monitoring data and future weather data through a prediction model includes: Based on the aforementioned future weather data, a prototype overflow model was determined; Based on the overflow prototype pattern, determine the type of device to be controlled; Based on the type of device to be controlled, the duration to be predicted is determined; Based on the aforementioned future weather data, future weather characteristics are obtained; Based on the pipeline network monitoring data and the drainage pipeline network structural data, the pipeline network status characteristics are determined. Based on the future weather characteristics, the pipeline network status characteristics, and the forecast duration, the model calling parameters are determined, wherein the model calling parameters include the calling time and the corresponding prediction model type; The prediction model type is determined based on the model call parameters; In response to the prediction model type being the hydrographic model, the pipeline health data is determined based on the pipeline monitoring data; Based on the pipeline health data, the drainage pipeline structure data, and the pipeline monitoring data, an urban pipeline network map is constructed. Using the aforementioned hydrological map model, based on the urban pipe network map, the overflow prediction data is generated; and, In response to the fact that the number of the prediction model types is greater than one, the overflow prediction data is combined according to the call time to obtain the updated overflow prediction data; The method further includes: In response to the pipeline health data indicating that a drainage pipe meets the dredging conditions, the flushing time is determined based on the future weather data; During the flushing time, a cleaning control command is generated and sent to the flushing valve associated with the drain pipe to perform the pipe flushing operation.

5. The method according to claim 4, characterized in that, The process of generating overflow control parameters based on the overflow prediction data includes: Based on the overflow prediction data, candidate control parameters are determined; and, An iterative process is performed based on the candidate control parameters and the prediction model. The iterative process includes: The candidate control parameters for this round are input into the prediction model to generate the comprehensive risk cost corresponding to the candidate control parameters for this round; wherein, the comprehensive risk cost is determined based on the predicted overflow volume of the overflow outlet, the environmental risk weight, and the equipment operating cost output by the prediction model; In response to the failure to meet the stopping condition, based on the candidate control parameters of this round and the comprehensive risk cost, updated candidate control parameters are obtained and used as candidate control parameters for the next iteration; and, In response to the satisfaction of the stopping condition, the iteration is stopped and the overflow control parameters are determined based on the candidate control parameters of the current round.

6. The method according to claim 5, characterized in that, The sensor also includes a water quality sensor, which is used to acquire water quality sensing data; The method further includes: Based on the water quality sensor data and the predicted overflow flow, the pollutant concentration at the overflow outlet is determined using a water quality prediction algorithm; and, The overall risk cost is determined based on the concentration of the pollutant.

Images