Method and system for monitoring the operating state of a substation rainwater drainage system

Abstract

The application provides a method and system for monitoring the operation state of a substation rainwater drainage system, and relates to the technical field of data processing. The method comprises the following steps: collecting original monitoring data of the substation rainwater drainage system; sending the original monitoring data to a monitoring platform; determining a set of effective monitoring points by performing communication availability determination on the installation positions of sensors; converting the monitoring data in the set of effective monitoring points into real-time hydraulic state quantities of corresponding nodes; constructing an online hydraulic digital twin model; inputting the real-time hydraulic state quantities into the online hydraulic digital twin model to output prior hydraulic states; correcting the prior hydraulic states to obtain continuous hydraulic operation state estimation results; extracting state deviation characteristic parameters according to the continuous hydraulic operation state estimation results; determining whether the operation state of the rainwater drainage system is abnormal according to the state deviation characteristic parameters; if yes, outputting the operation state monitoring result of the rainwater drainage system; otherwise, continuing to monitor.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for monitoring the operational status of a substation rainwater drainage system. Background Technology

[0002] Substation rainwater drainage systems typically consist of collection wells, inspection wells, underground pipe sections, gate / overflow structures, and discharge outlets. Operating in a concealed underground environment and influenced by factors such as rainfall intensity, catchment boundary conditions, and substation topography, their operational status exhibits significant time-varying and uncertainties. In cases of heavy rainfall, short-duration downpours, or obstructed drainage channels, the drainage system may experience rapid water level rises, localized backflow, and reduced flow capacity, leading to risks such as substation flooding, equipment foundation inundation, channel flooding, and moisture damage to secondary facilities.

[0003] On the one hand, by monitoring key node water levels, flow rates, and boundary conditions online, timely identification of changes in drainage capacity, the location of discharge bottlenecks, and the evolution trend of risks can be achieved. This provides quantitative data for flood control and drainage scheduling, gate control, and emergency response, reducing the impact of substation water accumulation on critical components such as primary and secondary equipment and cable trenches. Simultaneously, if the monitoring results can further output anomaly types and risk levels, they can be used for optimized allocation of operation and maintenance resources, graded management of hidden dangers, and the development of lean maintenance plans, thereby improving the resilience and reliability of substations under extreme rainfall conditions.

[0004] However, underground wireless communication is significantly affected by installation depth, shielding materials of the well chamber structure, and humid environments. Existing methods often fail to model the correlation between signal strength and transmission cost and energy consumption characteristics, making it difficult to guarantee the location of monitoring points and data reliability, thus affecting the quality of continuous monitoring. In addition, facing non-stationary conditions such as sudden inflow of rainstorms and rapid changes in boundary conditions, traditional methods based on static rules or offline calibration models are difficult to achieve online updates and real-time corrections, and cannot form continuous state estimation and trend judgment capabilities. They are also unable to provide stable and reliable risk level outputs in the early stages of anomalies, limiting their engineering applicability and promotion value in substation flood control operations. Summary of the Invention

[0005] In view of the shortcomings of the prior art, the purpose of this invention is to provide a method for monitoring the operation status of a substation rainwater drainage system. This method can solve the technical problems that traditional methods based on static rules or offline calibration models are difficult to achieve online updates and real-time corrections, cannot form continuous state estimation and trend judgment capabilities, and cannot provide stable and reliable risk level outputs in the early stages of anomalies, thus limiting their engineering applicability and promotion value in substation flood control operations.

[0006] A first aspect of this invention provides a method for monitoring the operational status of a substation rainwater drainage system, comprising:

[0007] S1: Collect raw monitoring data of the substation's rainwater drainage system;

[0008] S2: The raw monitoring data is sent to the monitoring platform via low-power wide-area wireless communication.

[0009] S3: Based on the wireless signal strength parameters, determine the communication availability of the sensor installation location and identify the set of effective monitoring points;

[0010] S4: Based on the sensor calibration relationship, the monitoring data in the effective monitoring point set is converted into the real-time hydraulic state quantity of the corresponding node;

[0011] S5: Based on the topology and hydraulic parameters of the substation rainwater drainage system, construct an online hydraulic digital twin model;

[0012] S6: Input the real-time hydraulic state variables into the online hydraulic digital twin model and output the prior hydraulic state;

[0013] S7: The prior hydraulic state is corrected by the data assimilation algorithm to obtain the continuous hydraulic operating state estimation result;

[0014] S8: Based on the continuous hydraulic operating state estimation results, extract the state deviation characteristic parameters that reflect the operating characteristics of the substation rainwater drainage system;

[0015] S9: Based on the state deviation characteristic parameters, determine whether there is an abnormality in the operating status of the substation rainwater drainage system; if yes, proceed to step S10; otherwise, return to step S1.

[0016] S10: Output the operational status monitoring results of the substation's rainwater drainage system.

[0017] A second aspect of the present invention provides an operational status monitoring system for a substation rainwater drainage system, comprising: a processor and a memory;

[0018] The memory stores programs or instructions that can run on a processor, and when the program or instructions are executed by the processor, they implement the steps of the substation rainwater drainage system operation status monitoring method as described in the first aspect.

[0019] A third aspect of the present invention provides a readable storage medium on which a program or instructions are stored, and when the program or instructions are executed by a processor, the steps of the substation rainwater drainage system operation status monitoring method as described in the first aspect are implemented.

[0020] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0021] In this embodiment of the invention, a closed-loop operation mechanism is constructed in the substation stormwater drainage system, consisting of communication availability screening, hydraulic digital twin modeling, data assimilation correction, multi-dimensional deviation feature extraction, and anomaly judgment output. This mechanism enables continuous, interpretable, and quantifiable monitoring of the underground drainage network's operational status. By determining communication availability based on wireless signal strength and screening effective monitoring points, the stability of monitoring data upload and long-term operational reliability are improved. By converting monitoring data into real-time hydraulic state quantities and embedding them into an online digital twin model based on topology and hydraulic parameters, physical consistency prediction of the evolution of water level and flow in the drainage system is achieved. Dynamic correction of prior hydraulic states using a data assimilation algorithm enhances the model's accuracy and robustness under sudden inflow and complex boundary conditions. This enables the differentiation between local and system-level anomalies, providing tiered early warning and decision support for stormwater drainage system operational risks. Ultimately, this improves the substation's flood control safety and operational intelligence under extreme conditions such as heavy rainfall. Attached Figure Description

[0022] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0023] Figure 1 This is a flowchart illustrating a method for monitoring the operational status of a substation rainwater drainage system, as provided in an embodiment of the present invention.

[0024] Figure 2 This is a schematic diagram of the operation status monitoring system of a substation rainwater drainage system provided in an embodiment of the present invention. Detailed Implementation

[0025] To enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. It should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0026] The following description, in conjunction with the accompanying drawings, details the method for monitoring the operational status of a substation rainwater drainage system provided by the present invention through specific embodiments and application scenarios.

[0027] Reference manual attached Figure 1 The diagram shows a flowchart of a method for monitoring the operational status of a substation rainwater drainage system according to an embodiment of the present invention.

[0028] This invention provides a method for monitoring the operational status of a substation rainwater drainage system, which may include the following steps:

[0029] S1: Collect raw monitoring data of the substation's rainwater drainage system.

[0030] The substation rainwater drainage system refers to a network of underground or semi-underground hydraulic facilities deployed within and around the substation area for collecting, converging, transporting, and discharging rainwater. This includes units such as collection wells, inspection wells, drainage pipe sections, gates, overflow structures, and discharge outlets. Raw monitoring data refers to basic data directly collected by sensors deployed at key nodes or pipe sections of the aforementioned system without model processing. This data typically includes water level-related data, flow-related data, timestamp data, equipment operating status data, environmental and boundary condition data, and communication and equipment health data, used to objectively reflect the system's true operating status at a given moment.

[0031] It should be noted that S1 enables continuous acquisition of key hydraulic parameters of the stormwater drainage system, providing reliable basic data support for subsequent hydraulic modeling, state prediction, and anomaly identification. Its advantages are twofold: firstly, it avoids the timeliness and information gaps inherent in relying solely on manual inspections, improving the real-time performance and completeness of the data; secondly, through standardized acquisition and unified time stamping of raw data, it creates conditions for multi-node data fusion and dynamic analysis, improving the accuracy and traceability of monitoring results from the source, while providing a high-quality input foundation for subsequent physical consistency calculations and data assimilation correction in digital twin models.

[0032] In one possible implementation, the raw monitoring data includes: water level-related data, timestamp data, flow rate-related data, equipment operating status data, environmental and boundary condition data, and communication and equipment health data.

[0033] S2: The raw monitoring data is sent to the monitoring platform via low-power wide-area wireless communication.

[0034] Low-power wide-area wireless communication refers to wireless communication technologies with long-distance transmission capabilities and low terminal power consumption. Examples include communication methods based on narrowband IoT (NB-IoT), LoRa, or other cellular IoT protocols. These technologies reduce terminal transmission power and standby power consumption while ensuring a large coverage area, making them suitable for data backhaul in underground or semi-underground scenarios. A monitoring platform refers to a centralized data processing and management system deployed on a local server in a substation or in a remote operation and maintenance center, used to receive, store, analyze, and display monitoring data. Low-power wide-area wireless communication is a mature existing technology, and will not be elaborated upon here.

[0035] It should be noted that, on the one hand, the low-power design extends sensor battery life, reduces maintenance frequency and operating costs, and improves the long-term sustainability of the system. On the other hand, the wide-area coverage capability enhances communication adaptability in complex underground environments, enabling monitoring data to be uploaded to the platform in real time for unified processing and analysis, thereby improving the timeliness, completeness, and centralized management level of the entire rainwater drainage system's operational status monitoring.

[0036] S3: Based on the wireless signal strength parameters, determine the communication availability of the sensor installation location and identify the set of effective monitoring points.

[0037] Wireless signal strength parameters refer to physical quantities that characterize the ability of a sensor terminal to receive or transmit wireless signals in its current installation environment. Examples include Received Signal Strength Indication (RSSI) or Signal-to-Noise Ratio (SNR), which reflect the quality of the communication link. Sensor installation location refers to the specific location of the monitoring terminal within the substation's rainwater drainage system—whether it is in an underground well, pipe section node, or boundary structure. Its installation depth, well structure, and surrounding materials all affect wireless propagation. The effective monitoring point set refers to the set of sensor nodes that have been determined to meet communication reliability and stability requirements and are capable of consistently uploading monitoring data over a long period.

[0038] It should be noted that, on the one hand, quantitatively assessing communication quality improves the overall data integrity and reliability of the monitoring network. On the other hand, eliminating low-reliability monitoring points reduces the interference of invalid data on subsequent hydraulic modeling and state estimation, thereby improving the accuracy of anomaly identification from the source. Simultaneously, optimizing terminal energy consumption allocation and maintenance strategies enhances the system's long-term stable operation capability.

[0039] In one possible implementation, S3 specifically includes:

[0040] S301: Obtain wireless signal strength parameters of multiple candidate monitoring points under underground installation conditions.

[0041] Candidate monitoring points refer to the set of locations where sensors are pre-selected for deployment in the substation's rainwater drainage system, typically including inspection wells, collection wells, or key pipe section nodes.

[0042] S302: Based on the underground propagation environment, additional attenuation correction calculations are performed on each wireless signal strength parameter to obtain the equivalent received signal strength.

[0043] The equivalent received signal strength refers to the corrected signal strength value used to characterize the actual underground communication capability after deducting the aforementioned additional attenuation.

[0044] Specifically, when evaluating the wireless communication capabilities of underground monitoring points, the wireless signal strength parameter at the corresponding ground location is first used as the baseline signal strength to characterize the basic coverage level of the area under unobstructed underground conditions. Subsequently, considering the actual installation conditions of the sensors in the underground environment, the additional attenuation generated during the underground propagation of the wireless signal is corrected. This additional attenuation includes propagation path loss caused by installation depth and shielding loss caused by structural materials such as manhole covers and walls. Specifically, the depth-related propagation attenuation is determined based on the sensor's underground installation depth, and the corresponding material attenuation is determined based on the manhole structure and material type. These attenuation values ​​are then superimposed and subtracted from the baseline signal strength to obtain the equivalent received signal strength, reflecting the actual underground communication conditions. The equivalent received signal strength is used to characterize the sensor's ability to stably receive and transmit wireless signals under the current underground propagation environment.

[0045] S303: Based on the equivalent received signal strength and the energy consumption characteristics of the wireless terminal under different coverage enhancement levels, calculate the communication availability score for each candidate monitoring point:

[0046]

[0047] in, This represents the communication availability score for the i-th candidate monitoring point. P represents the equivalent received signal strength calculated for the i-th candidate monitoring point in an underground propagation environment. min P represents the minimum permissible received signal strength threshold for a stormwater drainage system. max N represents the maximum permissible received signal strength threshold for the stormwater drainage system. i denoted by , w1 represents the coverage enhancement level index required for the i-th candidate monitoring point to complete effective data transmission within a unit time, w2 represents the weight coefficient of signal strength factor in communication availability evaluation, and w2 represents the weight coefficient of communication repetition number factor in communication availability evaluation.

[0048] Specifically, N iDefined as the number of retransmissions required to complete one effective data transmission in the current wireless environment, it is the actual number of retransmissions corresponding to the NB-IoT coverage enhancement level, and its value is a discrete value such as 1, 2, 4, 8, etc. This parameter can be obtained by mapping the coverage enhancement level returned by the communication module, or by statistically analyzing the average number of retransmissions during actual communication.

[0049] For example, in engineering implementation, the following mapping relationship can be used: coverage enhancement level CE0 corresponds to N. i =1, coverage enhancement level CE1, corresponding to N i =8, coverage enhancement level CE2, corresponding to N i =32. Additionally, the upper limit of coverage enhancement level metrics is typically limited by communication protocols or device capabilities, for example, no more than 2048 times, due to N i The physical meaning of N dictates that at least one transmission is required to complete communication, therefore N i The minimum value is 1, and there is no case where it is less than 1.

[0050] In this embodiment of the invention, in the calculation of communication availability score, N i It participates in the calculation in its reciprocal form, that is, through Characterizes communication cost factors. N i The larger the value, the more repeated transmissions are needed to ensure communication reliability, resulting in increased communication latency and energy consumption. The smaller the value, the lower its contribution to the communication availability score. For example, when N... i When =1, it means that communication can be completed in a single transmission, corresponding to =1. When N i When =32, the corresponding =0.03125. Through the above method, communication reliability and communication cost can be uniformly incorporated into the evaluation system. Simultaneously, by setting weighting coefficients w1 and w2, the signal strength factor and communication cost factor are weighted and integrated, which avoids the excessive influence of a single parameter on the scoring results, thereby ensuring the stability and rationality of the scoring results.

[0051] In one specific embodiment, for a candidate monitoring point: equivalent received signal strength =-100dBm, the allowable received signal strength range is P min =-120dBm, P max =-80dBm, then the normalized signal strength is 0.5, when the monitoring point corresponds to N i When N = 1, it indicates good communication conditions. i When N = 32, it indicates poor communication conditions. Let the weighting coefficients w1 = 0.7 and w2 = 0.3, then the communication availability scores are as follows: when N...i When N = 1, the score is 0.65; when N = 1, the score is 0.65. i When N = 32, the score is 0.359, which shows that as N increases... i By increasing the communication availability score, a reasonable decrease can be achieved, thereby enabling a comprehensive evaluation of communication reliability and communication cost.

[0052] It should be noted that those skilled in the art can set the received signal strength (P) according to actual needs. min and P max The magnitudes of the weighting coefficients (w1 and w2) are not limited in this invention.

[0053] Specifically, the above formula normalizes the equivalent received signal strength and, in combination with the impact of the number of communication repetitions on energy consumption and stability, introduces weighting coefficients to weight and fuse different communication performance indicators, thereby forming a comprehensive communication availability evaluation value that can simultaneously reflect the quality of wireless signal coverage and communication costs. This value is used to determine whether underground monitoring points are suitable for long-term, stable data collection and transmission.

[0054] S304: Combining the communication availability score and the communication availability score threshold, determine whether each candidate monitoring point meets the upload reliability requirements. If yes, proceed to step S305. Otherwise, mark the candidate monitoring points as low-reliability monitoring points and remove them.

[0055] In this embodiment of the invention, the communication availability score threshold can be determined based on engineering experience, and the threshold range is set to [0.4, 0.7] based on engineering experience.

[0056] It should be noted that those skilled in the art can set the communication availability score threshold according to actual needs, and this invention does not impose any limitations on it.

[0057] Specifically, if the communication availability score is not less than the communication availability score threshold, the corresponding candidate monitoring point is determined as a valid monitoring point and included in the set of valid monitoring points. Otherwise, the candidate monitoring point is determined not to meet the long-term stable communication conditions and is removed from the set of valid monitoring points or marked as a low-confidence monitoring point.

[0058] S305: Select candidate monitoring points that meet the upload reliability requirements as valid monitoring points and output the set of valid monitoring points.

[0059] Specifically, five underground inspection wells in the drainage network of a substation were pre-selected as candidate monitoring points. First, the baseline RSSI was measured at the corresponding ground location of each well, finding it to be -85 dBm. Then, based on the installation depth (e.g., 2 meters, 3 meters) and the well cover material (e.g., cast iron, composite material), the depth attenuation (e.g., 8 dB, 12 dB) and material attenuation (e.g., 5 dB, 3 dB) were estimated. These values ​​were then summed and subtracted from the baseline signal strength to obtain the equivalent received signal strength for each well. Subsequently, the required coverage enhancement level N for the wireless terminal under the current signal conditions was determined. i (For example, the number of repeated transmissions is 1 or 4) Substitute the data into the communication availability scoring formula to calculate the communication availability score, and then compare the calculation result with the preset threshold. If the communication availability score of a well is lower than the threshold, it is marked as a low-confidence monitoring point and removed. Finally, only the well locations that meet the conditions are retained to form a set of effective monitoring points for subsequent data acquisition and modeling.

[0060] It should be noted that by introducing an underground propagation attenuation correction and energy consumption coupling scoring mechanism, a quantitative assessment of the communication capabilities of candidate monitoring points can be achieved, thereby avoiding communication instability issues caused by simply relying on ground signals or experience for site selection. Simultaneously, by comprehensively balancing the energy costs associated with signal quality and coverage enhancement levels, monitoring points that possess both stable communication capabilities and low operating energy consumption can be selected, improving the long-term reliability and economy of the entire monitoring network. Furthermore, by eliminating low-reliability monitoring points, the interference of data packet loss and abnormal fluctuations on subsequent digital twin modeling and state assimilation processes can be reduced, improving the overall monitoring accuracy and anomaly identification reliability of the system from the source.

[0061] S4: Based on the sensor calibration relationship, the monitoring data in the effective monitoring point set is converted into the real-time hydraulic state quantity of the corresponding node.

[0062] Among them, sensor calibration relationship refers to the physical quantity conversion model established after sensor installation through experimental calibration or field comparison. It is used to describe the correspondence between sensor output signals (such as voltage, current, digital quantities, or raw water level readings) and actual hydraulic physical quantities, such as water level-voltage linear relationship, flow rate-water level curve relationship, or flow rate conversion model based on weir flow formula. Real-time hydraulic state quantities refer to physical state parameters that can be directly used for hydraulic model calculations, such as nodal head, pipe section flow rate, cross-sectional area of ​​water passage, or inflow rate.

[0063] Specifically, firstly, based on the calibration curves or equations established during the installation and commissioning phase of various sensors, the functional relationship between their output signals and actual hydraulic physical quantities is determined. For example, the voltage value or digital water level reading output by the level gauge is converted into the actual water level height using linear or nonlinear calibration formulas. For monitoring points where flow rate is indirectly calculated from water level, the water level value can be converted into the corresponding flow rate by combining weir flow formulas, orifice flow formulas, or pre-calibrated water level-flow relationship curves. Subsequently, according to the topology of the stormwater drainage system, the converted water level or flow rate data of each monitoring point is mapped to the corresponding computing nodes or supernodes and synchronized according to a unified timestamp to form node head vectors and inflow vectors that correspond one-to-one with the model state variables. For data with outliers or missing values, preprocessing can be performed through filtering or interpolation. The final output is a real-time hydraulic state quantity that can be directly input into the digital twin model, thus completing the conversion process from the original monitoring signal to physically consistent state variables.

[0064] It should be noted that, on the one hand, this ensures the physical consistency between the data input and the hydraulic model variables, allowing subsequent digital twin model calculations to be based on real hydraulic parameters. On the other hand, it achieves standardized conversion of data from different types of sensors through unified calibration relationships, eliminating dimensional differences and systematic errors, and improving data comparability and accuracy. Simultaneously, the construction of real-time hydraulic state quantities provides a direct and standardized input interface for subsequent prediction calculations and data assimilation, thereby improving the overall accuracy, stability, and engineering feasibility of state estimation.

[0065] S5: Based on the topology and hydraulic parameters of the substation rainwater drainage system, construct an online hydraulic digital twin model.

[0066] The topology of a stormwater drainage system refers to the network structure formed by interconnected units such as drainage pipe sections, manholes, collection wells, gates, and discharge outlets, used to describe the spatial connections and confluence paths of water flow within the system. Hydraulic parameters are physical parameters used to characterize the flow motion, including pipe diameter, length, bottom slope, roughness coefficient, cross-sectional shape, local loss coefficient, and additional storage area at nodes. A hydraulic digital twin model is a virtual hydraulic simulation model built on a computer platform based on physical mechanism equations (such as the mass conservation equation) that corresponds to the structure and operating characteristics of the actual stormwater drainage system, used to reflect and predict the system's operating status in real time.

[0067] It should be noted that, on the one hand, through topological structure modeling and hydraulic parameter embedding, the model has clear physical constraints and spatial coupling relationships, improving the physical consistency and interpretability of the calculation results. On the other hand, through the online operation mechanism, the model can be updated in real time according to the changes in the inflow and boundary conditions, realizing continuous prediction of the node head and pipe segment flow. At the same time, the digital twin model provides a unified calculation framework for subsequent state assimilation and anomaly identification, enabling the system to upgrade from single-point monitoring to network-wide linkage analysis, significantly enhancing the integrity, accuracy, and intelligent level of the operation monitoring of the substation rainwater drainage system.

[0068] In the embodiment of the present invention, the online hydraulic digital twin model is constructed once based on the topological structure and hydraulic parameters of the substation rainwater drainage system during the system initialization stage and is continuously used during the subsequent operation process, rather than being repeatedly constructed in each monitoring cycle. The topological structure and hydraulic parameters are derived from the engineering design data or on-site calibration data of the substation drainage system, remaining relatively stable during the system operation and only being adjusted accordingly when the structure changes or the parameters are updated. Further, "online" means that the hydraulic digital twin model can receive real-time hydraulic state variables during the operation process and perform dynamic calculations and state updates, that is, the model state is corrected in real time through data assimilation or recursive calculation methods, rather than referring to the model being reconstructed at each moment. In other words, the model has a "pre-constructed, continuously operating, online updated" model structure.

[0069] In a possible implementation manner, S5 specifically includes:

[0070] S501: Uniformly encode the original pipe segment units and node units of the substation rainwater drainage system to determine the original pipe segment-node topological network.

[0071] Among them, the original pipe segment unit refers to the smallest physical pipe segment divided according to the actual engineering structure in the rainwater drainage system. The node unit refers to the position unit where different pipe segments are connected, confluence occurs, or the water level is defined, such as inspection wells, sump wells, etc.

[0072] S502: Define the structure operator for online update, and map the original pipe segment-node topological network to a set of computational units of super pipe segments-super nodes based on the structure operator.

[0073] Among them, the structure operator refers to the mapping rule or transformation function used to compress, merge, or reconstruct the original pipe segment-node network. The super pipe segment refers to the computational unit formed by merging multiple consecutive pipe segments without bifurcation or control structures. The super node refers to the core computational node used to represent the head state after topological mapping.

[0074] Specifically, in constructing the online hydraulic digital twin model, the original topology of the stormwater drainage system is first uniformly encoded. The original topology includes multiple pipe segment units and multiple node units, where pipe segment units represent water flow channels, and node units represent pipe segment connections and confluence locations. Subsequently, a structural operator is constructed to map and transform the original pipe segment units and node units. Continuous pipe segments without bifurcations or control structures in the topology are merged into super pipe segment units, while nodes with clearly defined connections and water level states are merged or reconstructed into super node units. Through the structural operator, the mapping process from the original "pipe segment-node" network to the "super pipe segment-super node" network is realized, allowing the topology to be compressed and reorganized while maintaining hydraulic physical consistency. The mapped topology is used for subsequent discrete equation construction and state-space modeling, thus forming a more computationally efficient and structurally clear online hydraulic digital twin network skeleton.

[0075] S503: Combining the super-segment-super-node computational unit set, a continuous hydraulic control model describing the coupling relationship between mass conservation and momentum conservation is established based on the one-dimensional Saint-Venant equations.

[0076]

[0077]

[0078] Where A represents the cross-sectional area of ​​the water passage, i.e., the instantaneous water storage capacity of the pipe section or channel at a certain location, t represents the time index, Q represents the flow rate, i.e., the volume of water passing through the cross-section per unit time, x represents the spatial coordinate along the length of the pipe or channel, and q represents the flow rate. in Represents the external inflow term per unit length. This represents the rate of change of cross-sectional water storage over time. This represents the rate of change of flow along the spatial direction. The flow rate is represented by u, which represents the rate of change of flow rate over time. It represents the change in the momentum of a body of water along a spatial direction. This represents the rate of change of flow over time. Represents gravitational acceleration. This represents the spatial gradient of water level h along the pipe direction. Indicates the bottom slope of a pipe or channel. This refers to the frictional gradient, which is the frictional resistance between the water flow and the pipe wall or channel bottom. This represents local loss items, which describe energy losses caused by local structures such as elbows, manholes, and gates.

[0079] Specifically, the mass conservation equation describes the spatial and temporal continuity of water bodies in a stormwater drainage system. By balancing changes in cross-sectional water storage, flow rate, and external inflow, it ensures that the model satisfies the water conservation constraint at any given time. It is the fundamental governing equation describing the processes of stormwater collection, transport, and temporary storage. The momentum conservation equation describes the dynamic behavior of stormwater flowing in pipes or channels. By comprehensively considering factors such as water level gradient, bottom slope, friction loss, and local energy loss, it characterizes the dynamic evolution of flow rate and velocity, thereby reflecting the true hydraulic response characteristics of the stormwater drainage system under different operating conditions.

[0080] Among them, the one-dimensional Saint-Venant equations are the governing equations for mass and momentum conservation describing unsteady flow in open channels or pipes.

[0081] S504: Combining staggered mesh arrangement and backward Euler implicit scheme, the continuous hydraulic control model is discretized in time and space to determine the discrete update equation set that can be iterated online:

[0082]

[0083]

[0084] in, express The flow rate at the i-th position of the k-th pipe at time k. express The flow rate at the (i-1)th position of the k-th pipe at time k, i.e., the upstream flow rate. This represents the width of the water surface at the i-th discrete cross-section within the k-th pipe. This represents the spatial length of the i-th discrete unit within the k-th pipe. This represents the width of the water surface at the (i-1)th discrete section within the k-th pipe. This represents the spatial length of the (i-1)th discrete unit within the k-th pipe. This represents the additional water storage area of ​​the l-th node connected to the k-th pipeline. This indicates that the i-th node connected to the k-th pipeline is at time i. water level, This represents the water level at time t of the I-th node connected to the k-th pipeline. Indicates the preset time step. This represents the external inflow rate at the i-th node connected to the k-th pipeline. This represents the flow rate at the i-th position of the k-th pipe at time t. Let represent the flow velocity at the downstream node of the (I+1)th discrete unit within the k-th pipe. express The flow rate at position I+1 of the k-th pipe at time k. This represents the flow velocity of the k-th pipe at node I. express The flow rate at the I-th position of the k-th pipe at time k. This represents the cross-sectional area of ​​the k-th pipeline at the i-th discrete section. express The water level at downstream node I+1 of the k-th pipeline at time k. This represents the bottom slope of the k-th pipeline at the i-th discrete unit. Let represent the friction gradient of the k-th pipeline at the i-th discrete unit. This represents the local loss term of the k-th pipeline at the i-th discrete unit. This indicates the length deviation.

[0085] Among them, the backward Euler implicit scheme refers to a stable numerical scheme that uses future time-time variables for solution in time discretization.

[0086] Specifically, This refers to the horizontal cross-sectional area of ​​inspection wells, sump wells, or node shafts, reflecting the additional water storage capacity at the node due to the well chamber structure, measured in square meters. Its introduction allows the node water storage effect to be accurately represented in the discrete model.

[0087] S505: The supernode head is used as the boundary variable carrier, and boundary flow constraints are constructed based on the orifice flow relationship.

[0088] Specifically, when constructing the boundary conditions of the online hydraulic digital twin model, the head height at the supernode and the head height of the boundary nodes connected to it are first determined, and the head difference between them is calculated. The head difference is used to characterize the effective driving force at the boundary. When the head of the supernode is higher than that of the boundary node, a positive head difference is formed, indicating that the water flows outward. When the head of the boundary node is higher than that of the supernode, a reverse head difference is formed, indicating that there is a backflow tendency. Subsequently, based on the magnitude of the head difference and combined with the gravitational acceleration factor, the theoretical flow velocity per unit area is calculated, and a flow coefficient related to the characteristics of the boundary structure is introduced to correct the theoretical flow velocity. Finally, the corrected flow velocity is multiplied by the equivalent flow area at the boundary to obtain the actual flow value at the boundary. Through the above process, dynamic constraint calculation of the flow at boundary locations such as discharge outlets, gates, or overflow weirs is realized, enabling the digital twin model to automatically adjust the outflow or backflow state at the boundary according to real-time head changes.

[0089] S506: Embed the boundary flow constraints into the discrete update equations to determine the constrained discrete update equations.

[0090] S507: Constrained discrete update equations are used to construct a sparse linear algebraic solution system with supernode head as the core state variable.

[0091] Among them, the sparse linear algebra solution system refers to the solution structure of a system of linear equations where most of the elements of the coefficient matrix are zero.

[0092] Specifically, firstly, after completing the topological mapping of super-segment to super-node and constructing the discrete hydraulic equations, the head value of each supernode at the current time step is taken as the core unknown of the system. All supernode heads are arranged in order of node number to form a state vector, which serves as the main variable in the linear algebra solution system. Then, for each k-th super-segment, the boundary flow expression for that super-segment at the current time step is established based on its connection relationships with upstream and downstream supernodes. Specifically, the upstream boundary flow of the k-th super-segment is expressed as a linear combination of the upstream and downstream supernode heads, superimposed with a compensation term consisting of factors such as the flow rate, bottom slope, friction term, and local loss term from the previous time step. Similarly, the downstream boundary flow of the super-segment is expressed as a linear combination of the heads of the corresponding two supernodes plus a compensation term. Through these steps, the upstream and downstream boundary flows of each super-segment can be uniformly expressed as a "linear function of node heads plus a constant term." That is, each super-segment is only related to the heads of its two directly connected supernodes. Next, the boundary flow expressions for all super-segments are substituted into the discrete mass conservation equations at the nodes. Since each supernode is connected to only a few adjacent super-segments, the node mass conservation equations only contain the head variables of a few nodes directly connected to that node. By establishing the above mass conservation equations for all supernodes and substituting the flow expressions for each segment, a set of linear algebraic equations concerning the head of all supernodes is obtained. Therefore, the resulting system coefficient matrix is ​​a typical sparse matrix structure, with its non-zero elements corresponding only to node pairs connected in the topology. Finally, the sparse linear algebraic equations are uniformly assembled to form a linear system of the following form: "System coefficient matrix × Supernode head state vector = Constant term vector". Here, the system coefficient matrix consists of the linear coefficient parameters of each super-segment, and the constant term vector consists of compensation terms and external inflow factors.

[0093] It should be noted that by using a linear solution method applicable to sparse matrices to solve this linear system, the solution for the head of all supernodes at the current time step can be obtained, thus completing the construction and calculation of a sparse linear algebra solution system with the head of supernodes as the core state variable.

[0094] S508: Based on the sparse linear algebra solution system, define the online assimilable state equations for digital twins.

[0095] Specifically, in constructing the online hydraulic digital twin model of the stormwater drainage system, the water level of each supernode is first used as the core state variable of the system, and all node water levels are arranged into a state vector according to their numerical order. Within each computational time step, a recursive relationship between the node water level at the current moment and the node water level at the previous moment is established based on the node water level state at the previous moment, the external inflow at the current time step, and the system coefficient matrix derived from the discrete hydraulic equations. Specifically, a diagonal state coefficient matrix is ​​constructed using the proportional relationship between the node's additional water storage area and the time step, making the node water level change directly related to the node's water storage capacity. At the same time, the inflow of each node at the current time step is combined into an input vector to reflect external driving factors such as rainfall confluence or branch pipe inflow. Combined with the coupling terms obtained from the conservation of momentum and the discretization of the continuity equation, a complete system of linear algebraic equations is formed. By solving this system of equations, the node water level state at the current time step can be obtained. Since the state representation is a standard discrete-time recursive structure, it can be directly embedded with a data assimilation algorithm to correct the state vector after obtaining the observed water level data, thereby realizing online updating and real-time correction of the digital twin model.

[0096] S509: Based on the online assimilable state equation, a Kalman assimilation loop with twin-observation consistency is constructed to achieve online correction of the head state of supernodes.

[0097] Specifically, after completing the state prediction of the digital twin model of the stormwater drainage system, the predicted head state value for the current time step is first calculated based on the head state vector of the previous time step and the external inflow at the current time step through the state transition relationship. This prediction process comprehensively considers the system dynamics model, structural parameters, and input driving factors, and introduces system process noise to characterize the model uncertainty. Subsequently, the predicted head state is converted into predicted observation values ​​through the observation mapping relationship and compared with the water level observation values ​​collected by actual sensors to calculate the observation residuals. Based on the magnitude of the observation residuals and the relationship between the system state error covariance and the observation noise covariance, the gain coefficient for the current time step is calculated. The gain coefficient is used to measure the weight distribution between the prediction model results and the actual observation data. When the reliability of the observation data is high, the gain coefficient increases, enhancing the influence of the observation on the state correction. When the observation noise is large, the gain coefficient decreases, making the model prediction result dominant. Finally, the observation residuals are multiplied by the gain coefficient and superimposed on the predicted head state to obtain the corrected head state estimate, thereby achieving real-time updates of the node head state. Through the recursive process of prediction-residual calculation-gain adjustment-state correction described above, the digital twin model can continuously fuse observation information at each time step to achieve online correction and dynamic optimization of the head state.

[0098] S510: Based on the online correction of the supernode head state and combined with hydraulic parameters, an online hydraulic digital twin model is constructed.

[0099] Specifically, after obtaining the corrected head state vectors for each supernode, the corrected head is first substituted into the pre-established boundary flow relationship to calculate the boundary flow at each discharge outlet, gate, or external connection node. Then, the corrected head is substituted into the recursive relationship of the superpipe segment to calculate the upstream and downstream boundary flow of each superpipe segment. Based on this boundary flow, the flow rate, velocity, and cross-sectional area distribution at each discrete section within the pipe segment are recovered along the pipe segment direction. Finally, the updated node head, pipe segment flow, and cross-sectional hydraulic parameters are used as the system state output for the current time step and as the initial state for the next time step calculation, thus completing the construction and dynamic updating of the online hydraulic digital twin model.

[0100] It should be noted that by using topology compression and structural operator mapping, the original complex drainage network is transformed into a computationally more efficient super-segment-supernode system, significantly reducing the computational scale and improving the online solution speed while maintaining hydraulic-physical consistency. Introducing the one-dimensional Saint-Venant equation ensures the model satisfies mass and momentum conservation constraints, improving the physical reliability of the model results. Staggered meshes and implicit discretization schemes enhance numerical stability, enabling stable computation even under conditions of sudden inflow during heavy rainfall or rapid boundary changes. Constructing a sparse linear algebraic solution system improves the real-time computational capability of large-scale networks. Online assimilable state equations and Kalman closures enable dynamic fusion of model predictions and observational data, reducing model error accumulation. Finally, by updating node head and segment flow rates in real time, continuous, network-wide dynamic simulation and state estimation of the entire substation rainwater drainage system are achieved, significantly improving monitoring accuracy, response speed, and the engineering practicality of anomaly identification.

[0101] S6: Input the real-time hydraulic state variables into the online hydraulic digital twin model and output the prior hydraulic state.

[0102] Among them, the prior hydraulic state refers to the prediction results of the nodal head, pipe section flow and related hydraulic variables obtained by the model based on the state at the previous moment and the current inflow conditions before the current observation is integrated and corrected. It is also called the predicted state or background field.

[0103] It should be noted that, on the one hand, by introducing real-time inflow and the state at the previous moment, dynamic prediction of the evolution trend of water level and flow can be achieved, providing a reasonable physical background field for subsequent assimilation correction. On the other hand, model prediction can compensate for the lack of information caused by the limited spatial distribution of monitoring points, realizing the expansion from "discrete monitoring" to "continuous state estimation of the entire network".

[0104] In one possible implementation, S6 specifically includes:

[0105] S601: Arrange the external inflow data in the real-time hydraulic state variables according to the node number order to obtain the inflow input vector of the current time step.

[0106] External inflow data refers to the flow rate input from rainfall runoff, branch pipe inflow, or external connection pipe sections.

[0107] S602: Sort the head states of each supernode after online correction in the previous time step according to the node number order, and determine the initial state vector of the current time step.

[0108] The initial state vector refers to the state column vector composed of the supernode head after assimilation correction in the previous time step.

[0109] S603: Based on the inflow input vector and the initial state vector, and combined with the discrete mass conservation equation and the momentum conservation equation, an implicit recursive relationship of the supernode head within a preset time step is established.

[0110] It should be noted that those skilled in the art can set the size of the preset time step according to actual needs, and this invention does not limit it.

[0111] Implicit recursive relations refer to the update relations that are obtained by solving a system of equations where the current state appears on both sides of the equation.

[0112] Specifically, within a preset time step Δt, based on the discretized mass conservation equation, the change in water storage at a node is expressed as the ratio of the node's additional water storage area to the change in head. This water storage term, along with the external inflow and pipe coupling terms for each node, are uniformly incorporated into a system of linear equations. Specifically, the node head state from the previous time step is weighted and combined with the inflow vector at the current time step, where the weights are determined by the ratio of the node's additional water storage area to the preset time step Δt, thus forming a diagonal state coefficient matrix to reflect the water storage inertia characteristics of each node. Simultaneously, the coupling term obtained from the discretized pipe momentum equation is added as a compensation vector to the right-hand side of the equations, constructing an implicit linear algebraic equation regarding the node head at the current time step. By solving this system of equations, a new node head state can be obtained within the preset time step Δt, thereby achieving recursive updates of the node head between adjacent time steps.

[0113] S604: Solve the implicit recurrence relation linearly to obtain the predicted supernode head state vector at the current time step:

[0114]

[0115] in, express The nodal head state vector at time Zt Y represents the system coefficient matrix at time t, -1 indicates the inverse operation of the matrix, and Y represents the system coefficient matrix at time t. t Let x represent the diagonal matrix at time t. t This represents the nodal head state vector at time t, i.e., the corrected head state from the previous time step. express External inflow data at any given time This represents the compensation term vector obtained at time t from the discrete decomposition of the momentum equation and the derivation of the boundary conditions.

[0116] In this matrix, the diagonal elements are the ratio of the additional water storage area of ​​the node to the time step Δt.

[0117] Specifically, this formula is used to solve for the predicted head state at the current time step within a preset time step Δt, based on the nodal head state at the previous moment, the current external inflow conditions, and the coupling compensation term obtained from the discretization of the hydraulic equations. This enables the time-recursive update of the hydraulic state of the stormwater drainage system. Its advantages are twofold: firstly, by uniformly characterizing the hydraulic coupling relationship between nodes through a system matrix, it ensures that the calculation results satisfy the constraints of mass and momentum conservation, exhibiting good physical consistency; secondly, by employing an implicit solution form, the model maintains high numerical stability and computational robustness even under larger time steps or sudden inflow conditions, supporting real-time online prediction and continuous operation monitoring of substation stormwater drainage systems.

[0118] S605: Substitute the predicted supernode head state vector into the super-segment recursive relationship to determine the upstream and downstream boundary flows of each super-segment.

[0119] Specifically, after obtaining the predicted head state vector for the current time step, the upstream and downstream node heads corresponding to each k-th super-pipe segment are first extracted based on the connection relationship between the super-pipe segments and nodes. Then, the upstream and downstream node heads are substituted into the pre-established super-pipe segment recursive formula, and combined with the corresponding flow coefficients and compensation terms, the upstream and downstream boundary flows of the super-pipe segment at the current time step are calculated. This calculation process is performed sequentially for all super-pipe segments to determine the upstream and downstream boundary flow distribution of each super-pipe segment in the entire drainage network, providing boundary conditions for subsequent flow recovery within pipe segments and updates to the hydraulic state of the entire network.

[0120] S606: The predicted supernode head state vector and upstream and downstream boundary flows are output as prior hydraulic states.

[0121] It should be noted that by constructing an implicit recursive mechanism based on mass and momentum conservation constraints, continuous physical consistency updates of the head state over time are achieved, ensuring that the prediction results satisfy hydraulic equilibrium conditions and improving model reliability. By incorporating the additional storage area of ​​nodes into the state coefficient matrix, the model's ability to characterize the inertia of well chamber water storage is enhanced, improving the response accuracy under sudden rainstorm conditions. By employing a unified matrix solution method, the coupling relationships between nodes are systematically described, enabling the overall network hydraulic response to be calculated holistically rather than in isolation. Implicit formatting improves numerical stability, allowing the model to maintain stable convergence even under large time steps and strong inflow conditions. By outputting the predicted head and boundary flow to form a complete prior state, a high-quality background field is provided for subsequent data assimilation, thereby improving the overall accuracy and engineering reliability of online prediction, risk assessment, and anomaly identification for substation stormwater drainage systems.

[0122] S7: The prior hydraulic state is corrected by a data assimilation algorithm to obtain the continuous hydraulic operating state estimation result.

[0123] Data assimilation algorithms are mathematical methods that fuse model predictions with real-time observation data. They typically employ Kalman filtering or its extensions, using error covariance propagation and gain calculation to achieve state correction. Continuous hydraulic operating state estimation results refer to the optimal estimates of nodal head, pipe flow, etc., obtained at each time step after the fusion of prediction and observation. These estimates are recursively updated over time, continuously reflecting changes in system operation.

[0124] Specifically, "continuous hydraulic operating state estimation results" refers to hydraulic state estimation results that are continuously updated in a recursive manner over a time series. Its "continuous" characteristic is mainly reflected in the continuity in the time dimension, rather than the continuous distribution in the spatial dimension.

[0125] In this embodiment of the invention, the prior hydraulic state refers to the predicted hydraulic state calculated within the current time step based on the state of the previous time step and the current inflow conditions, using an online hydraulic digital twin model. This predicted state serves as the prediction input during the data assimilation process. Correction refers to the process of weightedly fusing and updating the prior hydraulic state with the observation data at the corresponding time point. The fusion weights are jointly determined by the prediction error covariance and the observation noise covariance, thereby obtaining a corrected hydraulic state that more closely approximates the true state.

[0126] Furthermore, the continuous hydraulic operating state estimation result refers to the hydraulic state sequence obtained by progressively updating through the aforementioned "prediction-correction" recursive process within each discrete time step. Its "continuous" characteristic is reflected in the continuous recursive update in the time dimension, rather than the continuity of the prior hydraulic state itself. In other words, the prior hydraulic state is the prediction result of a single time step, which, after data assimilation and correction, forms the optimal state estimate for the current time step and serves as the input for the next time step, thus constituting a continuously evolving hydraulic operating state estimation result over time.

[0127] In one possible implementation, S7 specifically includes:

[0128] S701: Collect water level observation data from each monitoring point and construct observation vectors based on the water level observation data.

[0129] Among them, water level observation data refers to the real-time measured values ​​of node water levels collected by sensors at each monitoring point.

[0130] S702: Obtain the wireless signal strength parameters of each monitoring point and construct the observation noise covariance matrix based on the wireless signal strength parameters.

[0131] The observation noise covariance matrix is ​​a matrix used to characterize the magnitude and uncertainty of observation errors, and its diagonal elements represent the observation noise variance of each monitoring point.

[0132] S703: Using the predicted supernode head state vector from the prior hydraulic state as the predicted state, and combining it with the observation mapping matrix, calculate the predicted observation vector:

[0133]

[0134]

[0135] in, express The predicted head state estimation vector at time A t This represents the state transition matrix at time t. B represents the head state estimation vector at time t. t This represents the input interaction matrix at time t. This represents the compensation term derived from the discrete hydraulic equations of the system at time t. express The predicted observation vector at time, i.e. (the sensor observables derived from the predicted head state). express The observation matrix at each time point.

[0136] Specifically, the first formula predicts the nodal head state at the current time step based on the head state of the previous time step, the external inflow conditions at the current time step, and the hydraulic coupling relationship of the system. Its essential function is to complete the time progression of the online hydraulic digital twin model and realize the recursive prediction of the nodal head state. The second formula maps the predicted head state to the observation space, generating predicted observations that can be directly compared with sensor measurements. Its essential function is to establish a bridge between the model prediction results and the actual observation data.

[0137] S704: Calculate the observation residuals based on the observed vector and the predicted observation vector.

[0138] Specifically, firstly, the actual water level observations collected at each monitoring point are obtained and arranged into an observation vector according to their node numbers. Simultaneously, based on the prediction results of the online hydraulic digital twin model, the predicted observation values ​​for the corresponding nodes at that moment are calculated and arranged into a prediction observation vector in the same order. Then, for each corresponding node, the actual observation value is subtracted from the predicted observation value to obtain the observation bias for that node. Combining the observation biases of all nodes in numerical order yields the observation residual vector. The observation residual is used to characterize the degree of difference between the model's prediction results and the actual observation data, and serves as the basis for subsequent state correction calculations.

[0139] S705: Based on the observation residuals, and combining the prediction error covariance matrix and the observation noise covariance matrix, calculate the Kalman gain matrix:

[0140]

[0141] in, express The Kalman gain matrix at time 10:00. express The prediction error covariance matrix at time 1. T This indicates the transpose operation. express The observation noise covariance matrix at time t.

[0142] S706: Based on the Kalman gain matrix, the predicted supernode head state vector is corrected to obtain the continuous hydraulic operation state estimation result.

[0143] It should be noted that by constructing a complete Kalman assimilation mechanism, the dynamic fusion of model prediction results and real-time observation data is achieved, improving the accuracy of nodal head estimation. By introducing an observation noise covariance matrix, monitoring points under different communication quality conditions are assigned differentiated weights, reducing the interference of low-reliability monitoring points on system state estimation. Adaptive adjustment of the trade-off between the model and observations through Kalman gain ensures rapid convergence when observations are reliable, while maintaining model dominance when observation noise is high, improving overall robustness. Recursive update structures prevent error accumulation, achieving continuous and smooth estimation of the hydraulic state. Ultimately, this enhances the stability and prediction reliability of the stormwater drainage system's state identification under conditions of sudden inflow, local blockage, or communication fluctuations, providing a high-precision state foundation for subsequent anomaly detection and risk classification.

[0144] S8: Based on the continuous hydraulic operating state estimation results, extract the state deviation characteristic parameters that reflect the operating characteristics of the substation rainwater drainage system.

[0145] Among them, state deviation characteristic parameters refer to quantitative indicators used to characterize the degree of difference between the current system operating state and the model predicted state, historical normal state or physical equilibrium state. They typically include normalized observation residuals, state correction magnitude, time evolution difference components, network-level deviation energy indicators, and communication reliability weighted deviation indicators.

[0146] In one possible implementation, S8 specifically includes:

[0147] S801: Combining observation residuals and innovative covariance matrices, construct dimensionless normalized bias characteristics.

[0148] Among them, the innovative covariance matrix refers to the covariance matrix used to characterize the statistical properties of observation residuals.

[0149] Specifically, by normalizing the variance of the observation residuals, the deviations of different nodes under different dimensions and noise levels are made comparable, thus forming a dimensionless deviation index with a unified scale.

[0150] For example, the calculation process of the dimensionless normalized bias feature is as follows: At the current time t, firstly, the actual observed water level value of the i-th monitoring point is obtained, and combined with the prediction results of the online hydraulic digital twin model, the predicted observed value corresponding to this monitoring point is obtained through the observation mapping relationship. Subsequently, the actual observed value and the predicted observed value are differencing to obtain the observation residual of this monitoring point, which is used to characterize the degree of deviation between the model prediction and the actual observation. Based on this, the diagonal element corresponding to this monitoring point in the innovation covariance matrix at the current time is further obtained. This diagonal element reflects the uncertainty or noise level of the observation residual of this monitoring point. Finally, the observation residual is divided by the square root of this diagonal element to standardize the residual, thereby obtaining the dimensionless normalized bias feature.

[0151] In this embodiment of the invention, the dimensionless normalized bias feature is correlated with the observation residuals and the innovation covariance matrix. Specifically, the innovation covariance matrix is ​​used to characterize the uncertainty level of the current observation residuals, and its diagonal elements represent the variance of the residuals at the corresponding monitoring points. By dividing the observation residuals by the square root of the variance, the residuals are standardized to obtain the dimensionless normalized bias feature. Thus, the bias not only reflects the magnitude of the difference between the prediction and the observation, but also the significance of that difference under the current uncertainty conditions, enabling a unified comparison between different monitoring points.

[0152] S802: Construct state correction intensity features.

[0153] Specifically, the state correction intensity feature is used to reflect the adjustment magnitude between the model prediction and the observed correction during the Kalman assimilation process, so as to characterize the degree of state correction of the system at the current time step.

[0154] In this embodiment of the invention, at the current time t, for the i-th supernode, its predicted head state before data assimilation and its head state after data assimilation correction are obtained respectively. Subsequently, a difference operation is performed on the predicted head state and the corrected head state, and the absolute value is taken to obtain the state correction magnitude of the node at the current time, which is used as the state correction intensity feature.

[0155] S803: Perform differential operations on the dimensionless normalized deviation characteristics to determine the time evolution deviation characteristics.

[0156] Specifically, by calculating the change in normalized deviation between adjacent time steps, the growth trend or abrupt change characteristics of the deviation can be characterized, which can be used to identify sudden anomalies or continuously accumulating anomalies.

[0157] S804: Calculate the network-level deviation energy index based on the predicted supernode head state vector and the continuous hydraulic operation state estimation results.

[0158] Specifically, within the current time step, based on the corrected head state estimation results and the predicted head state before correction, the flow rate corresponding to each k-th super-segment is calculated according to the established super-segment recursive relationship. Then, for the k-th super-segment, the absolute value of the flow rate difference between the two states before and after correction is calculated to obtain the flow rate deviation of that super-segment. The flow rate deviation is used to characterize the degree of influence on the flow rate of that segment during the assimilation correction process. Next, the flow rate deviations of all super-segments are squared to eliminate differences in positive and negative directions and amplify the impact of larger deviations. Finally, the squared deviation values ​​of all super-segments are summed to obtain the network-level deviation energy index across the entire network. This index reflects the overall deviation of the hydraulic state of the entire stormwater drainage system before and after assimilation correction within the current time step. When multiple segments experience large deviations simultaneously, the network-level deviation energy index increases significantly, thus characterizing the abnormal risk level at the system level.

[0159] S805: Construct a weighted deviation feature for communication reliability based on wireless signal strength parameters.

[0160] Specifically, by mapping wireless signal strength to a weighting factor, the deviation of monitoring points with poor communication quality is suppressed, thereby reducing the risk of abnormal misjudgment.

[0161] In this embodiment of the invention, the wireless signal strength parameter of the i-th monitoring point at the current time is obtained, and the wireless signal strength parameter is normalized based on a preset range of minimum and maximum received signal strength to obtain the corresponding communication reliability weight factor. Specifically, when the wireless signal strength is lower than the minimum received signal strength, the weight factor is 0; when the wireless signal strength is higher than the maximum received signal strength, the weight factor is 1. Subsequently, the communication reliability weight factor and the deviation characteristics of the corresponding monitoring point are weighted and calculated to obtain the communication reliability weighted deviation characteristics.

[0162] S806: Combine dimensionless normalized deviation characteristics, state correction intensity characteristics, time evolution deviation characteristics, network-level deviation energy index, and communication reliability weighted deviation characteristics to obtain state deviation characteristic parameters.

[0163] It should be noted that by constructing a multi-dimensional deviation feature system, an upgrade from single-point residual analysis to system-level comprehensive anomaly characterization is achieved. Normalization eliminates differences in dimensions and noise levels across different nodes, improving the comparability of deviation indicators. State correction intensity features reflect the adjustment magnitude between the model and observations, enhancing sensitivity to potential anomalies. Temporal difference features characterize the deviation growth trend, enabling the system to identify both sudden and gradual anomalies. Network-level deviation energy indicators quantify network-wide coupled risks, improving the ability to identify systemic anomalies. A communication credibility weighting mechanism reduces the interference of low-quality observations on judgment results, enhancing the robustness of anomaly determination. Ultimately, a state deviation feature parameter system with a clear structure, physical consistency, and strong noise resistance is formed, providing a high-precision, multi-scale, and engineering-interpretable decision-making basis for subsequent anomaly classification and risk level determination.

[0164] S9: Based on the state deviation characteristic parameters, determine whether there is an abnormality in the operating status of the substation rainwater drainage system. If yes, proceed to step S10. Otherwise, return to step S1.

[0165] Specifically, firstly, statistical thresholds are set for the dimensionless normalized deviation characteristic, state correction intensity characteristic, and time evolution deviation characteristic. When the normalized deviation of any node continuously exceeds the preset confidence interval or its time difference shows a continuous increasing trend, a local anomaly is identified. Secondly, a system-level threshold is set for the network-level deviation energy index. When this index is significantly higher than the historical normal operation interval in the current time step or within several consecutive time steps, a risk of network-wide coupling anomaly is identified. Simultaneously, the communication reliability weighted deviation characteristic is combined to adjust the weight of the anomaly contribution of nodes with poor communication quality to avoid false alarms. Finally, the above multi-dimensional features are fused and calculated using logical rules or a weighted scoring model to obtain a comprehensive anomaly judgment result. When the comprehensive score exceeds the anomaly judgment threshold, the system operation status is considered abnormal; otherwise, it is considered normal operation and enters the next monitoring cycle. This achieves quantitative, hierarchical, and robust judgment of anomalies in the rainwater drainage system.

[0166] In this embodiment of the invention, anomaly determination is based on multi-dimensional state deviation characteristics for quantitative judgment. The specific rules are as follows: First, a statistical threshold is set for the dimensionless normalized deviation characteristics. When the absolute value of the dimensionless normalized deviation characteristic of any monitoring point exceeds the preset threshold, the monitoring point is determined to have an abnormal deviation. Second, a change threshold is set for the time evolution deviation characteristics. When the time evolution deviation characteristics continuously increase and exceed the preset threshold over multiple consecutive time steps, a continuous evolution anomaly is determined to exist. Third, a system-level threshold is set for the network-level deviation energy index. When the network-level deviation energy index exceeds the threshold obtained based on statistics from historical normal operating states, a system-level anomaly risk is determined to exist.

[0167] In one implementation, the aforementioned deviation characteristics can be fused according to preset weights to obtain a comprehensive anomaly evaluation index. This comprehensive anomaly evaluation index is then compared with a preset comprehensive threshold to obtain the final anomaly determination result. Each threshold can be determined based on historical normal operation data through statistical methods or set according to engineering experience, thereby ensuring that the anomaly determination process has clear quantitative rules and feasibility.

[0168] Specifically, the weighted scoring model takes the form of "normalizing each feature first, and then linearly summing them according to preset weights," and its output is a comprehensive anomaly score. For example, firstly, various state deviation features are processed to unify their dimensions or normalize them, so that they are within a comparable numerical range. Then, the normalized feature quantities are linearly weighted and fused according to preset weights to obtain the comprehensive anomaly score. The comprehensive anomaly score can be expressed as: the comprehensive anomaly score is the result of multiplying each feature quantity by its corresponding weight coefficient and then summing the results, where each weight coefficient is a non-negative value and the sum of the weight coefficients is 1. Furthermore, the weight coefficients have a clear method of determination: on the one hand, they can be preset based on engineering experience, for example, assigning relatively high weights to dimensionless normalized deviation features that reflect the degree of instantaneous deviation and network-level deviation energy indicators that reflect the overall anomaly degree of the system, and assigning auxiliary weights to time-evolution deviation features and communication reliability weighted deviation features. On the other hand, the weight coefficients can also be determined based on historical operating data, through statistical analysis of normal and abnormal samples, so that the model has data-driven adaptability.

[0169] In one specific embodiment, the weighting coefficients can be: dimensionless normalized deviation feature weight is 0.30, state correction intensity feature weight is 0.20, time evolution deviation feature weight is 0.15, network-level deviation energy index weight is 0.25, and communication reliability weighted deviation feature weight is 0.10.

[0170] It should be noted that by using multi-dimensional state deviation characteristic parameters as a unified judgment criterion, a decision-making mechanism of "deviation analysis - anomaly judgment - closed-loop feedback" is established, enabling automatic identification and dynamic response of the stormwater drainage system's operating status within a continuous monitoring framework. When the comprehensive deviation index exceeds a set threshold, the system immediately enters the anomaly output process, which helps shorten the anomaly detection time and improve the timeliness of risk warnings. When the anomaly conditions are not met, the system returns to step S1 to continue data collection and cyclical updates, forming a continuous rolling monitoring closed loop. This avoids unnecessary false alarms and frequent alarms, improving system stability and operational efficiency.

[0171] S10: Output the operational status monitoring results of the substation's rainwater drainage system.

[0172] In one possible implementation, the operational status monitoring results include: the anomaly type and the risk level corresponding to the anomaly type.

[0173] Among them, "anomaly type" refers to the anomaly category classified according to different deviation characteristic patterns, such as local node water level anomaly, pipeline flow anomaly, network-wide coupling anomaly, continuous accumulation anomaly, or sudden impact anomaly. "Risk level" refers to the indicator that classifies the severity of anomalies based on factors such as anomaly intensity, scope of impact, and duration. For example, it can be divided into general risk, high risk, and severe risk levels, which are used to guide the priority of operation and maintenance response.

[0174] It should be noted that, on the one hand, by clearly defining the anomaly type, maintenance personnel can quickly locate the nature and possible location of the problem, improving handling efficiency. On the other hand, by outputting risk level classifications, differentiated alarm management is achieved, avoiding resource waste or response delays caused by a "one-size-fits-all" approach to alarms. Simultaneously, presenting technical analysis results in an intuitive hierarchical format helps management decision-makers quickly understand the system's risk level, thereby improving the precision, visualization, and intelligence of flood control management for substation rainwater drainage systems.

[0175] In this embodiment of the invention, a closed-loop operation mechanism is constructed in the substation stormwater drainage system, consisting of communication availability screening, hydraulic digital twin modeling, data assimilation correction, multi-dimensional deviation feature extraction, and anomaly judgment output. This mechanism enables continuous, interpretable, and quantifiable monitoring of the underground drainage network's operational status. By determining communication availability based on wireless signal strength and screening effective monitoring points, the stability of monitoring data upload and long-term operational reliability are improved. By converting monitoring data into real-time hydraulic state quantities and embedding them into an online digital twin model based on topology and hydraulic parameters, physical consistency prediction of the evolution of water level and flow in the drainage system is achieved. Dynamic correction of prior hydraulic states using a data assimilation algorithm enhances the model's accuracy and robustness under sudden inflow and complex boundary conditions. This enables the differentiation between local and system-level anomalies, providing tiered early warning and decision support for stormwater drainage system operational risks. Ultimately, this improves the substation's flood control safety and operational intelligence under extreme conditions such as heavy rainfall.

[0176] Reference manual attached Figure 2 The diagram shows a schematic representation of the operation status monitoring system for a substation rainwater drainage system provided in an embodiment of the present invention.

[0177] This invention provides an operational status monitoring system 20 for a substation rainwater drainage system, comprising: a processor 201 and a memory 202;

[0178] The memory 202 stores programs or instructions that can run on the processor 201. When the program or instructions are executed by the processor 201, they implement the steps of the above-described method for monitoring the operating status of the substation rainwater drainage system and achieve the same technical effect. To avoid repetition, the present invention will not elaborate further.

[0179] It should be understood that the processor 201 in this embodiment of the invention may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0180] It should also be understood that the memory 202 in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus RAM (DR RAM).

[0181] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0182] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0183] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0184] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0185] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0186] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0187] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0188] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0189] This invention provides a readable storage medium comprising: storing a program or instructions on the readable storage medium, wherein when the program or instructions are executed by a processor, the program or instructions implement the steps of the above-described method for monitoring the operating status of a substation rainwater drainage system, and can achieve the same technical effect. To avoid repetition, this invention will not elaborate further.

[0190] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the protection scope of the present invention.

Claims

1. A method for monitoring the operational status of a substation rainwater drainage system, characterized in that, include: S1: Collect raw monitoring data of the substation's rainwater drainage system; S2: The raw monitoring data is sent to the monitoring platform via low-power wide-area wireless communication. S3: Based on the wireless signal strength parameters, determine the communication availability of the sensor installation location and identify the set of effective monitoring points; S4: Based on the sensor calibration relationship, convert the monitoring data in the set of effective monitoring points into the real-time hydraulic state quantities of the corresponding nodes; S5: Based on the topology and hydraulic parameters of the substation rainwater drainage system, construct an online hydraulic digital twin model; S6: Input the real-time hydraulic state variables into the online hydraulic digital twin model and output the prior hydraulic state; S7: The prior hydraulic state is corrected by a data assimilation algorithm to obtain the continuous hydraulic operating state estimation result; S8: Based on the continuous hydraulic operating state estimation results, extract the state deviation characteristic parameters that reflect the operating characteristics of the substation rainwater drainage system; S9: Based on the state deviation characteristic parameters, determine whether there is an abnormality in the operating status of the substation rainwater drainage system; If so, proceed to step S10; Otherwise, return to step S1; S10: Output the operation status monitoring results of the substation rainwater drainage system; Specifically, S5 includes: S501: Unify the coding of the original pipe segment units and node units of the substation rainwater drainage system to determine the original pipe segment-node topology network; S502: Define a structural operator for online updates, and map the original pipe segment-node topology network into a set of super pipe segment-super node computing units based on the structural operator; S503: Combining the aforementioned super-segment-super-node computing unit set, a continuous hydraulic control model describing the coupling relationship between mass conservation and momentum conservation is established based on the one-dimensional Saint-Venant equations. S504: Combining staggered grid arrangement and backward Euler implicit scheme, the continuous hydraulic control model is spatiotemporally discretized to determine a discrete update equation set that can be iterated online; S505: Using the supernode head as the boundary variable carrier, and constructing boundary flow constraints based on the orifice flow relationship; S506: Embed the boundary flow constraint into the discrete update equation set to determine the constraint discrete update equation set; S507: Discretely update the set of constraint equations to construct a sparse linear algebra solution system with the supernode head as the core state variable; S508: Based on the aforementioned sparse linear algebra solution system, define the online assimilable state equation for digital twins; S509: Based on the online assimilable state equation, construct a twin-observation consistent Kalman assimilation loop to achieve online correction of the supernode head state; S510: Based on the online corrected head state of the supernode and combined with the hydraulic parameters, construct the online hydraulic digital twin model.

2. The method for monitoring the operational status of a substation rainwater drainage system according to claim 1, characterized in that, The raw monitoring data includes: water level related data, timestamp data, flow rate related data, equipment operating status data, environmental and boundary condition data, and communication and equipment health data.

3. The method for monitoring the operational status of a substation rainwater drainage system according to claim 1, characterized in that, S3 specifically includes: S301: Acquire wireless signal strength parameters of multiple candidate monitoring points under underground installation conditions; S302: Based on the underground propagation environment, additional attenuation correction calculations are performed on each of the wireless signal strength parameters to obtain the equivalent received signal strength; S303: Based on the equivalent received signal strength and combined with the energy consumption characteristics of the wireless terminal under different coverage enhancement levels, calculate the communication availability score of each candidate monitoring point; S304: Combining the communication availability score and the communication availability score threshold, determine whether each candidate monitoring point meets the upload reliability requirements; if yes, proceed to step S305; otherwise, mark the candidate monitoring point as a low-confidence monitoring point and remove the low-confidence monitoring point. S305: Select the candidate monitoring points that meet the upload reliability requirements as valid monitoring points and output the set of valid monitoring points.

4. The method for monitoring the operational status of a substation rainwater drainage system according to claim 1, characterized in that, S6 specifically includes: S601: Arrange the external inflow data in the real-time hydraulic state variables according to the node number order to obtain the inflow input vector of the current time step; S602: Sort the head states of each supernode after online correction in the previous time step according to the node number order, and determine the initial state vector of the current time step; S603: Based on the inflow input vector and the initial state vector, and combining the discrete mass conservation equation and the momentum conservation equation, establish the implicit recursive relationship of the supernode head within a preset time step; S604: Solve the implicit recursive relationship linearly to obtain the predicted supernode head state vector at the current time step; S605: Substitute the predicted supernode head state vector into the super pipe segment recursive relationship to determine the upstream and downstream boundary flow of each super pipe segment; S606: Output the predicted supernode head state vector and the upstream and downstream boundary flows as the prior hydraulic state.

5. The method for monitoring the operational status of a substation rainwater drainage system according to claim 1, characterized in that, Specifically, S7 includes: S701: Collect water level observation data from each monitoring point and construct an observation vector based on the water level observation data; S702: Obtain the wireless signal strength parameters of each monitoring point, and construct the observation noise covariance matrix based on the wireless signal strength parameters; S703: Take the predicted supernode head state vector in the prior hydraulic state as the predicted state, and calculate the predicted observation vector by combining it with the observation mapping matrix. S704: Calculate the observation residual based on the observation vector and the predicted observation vector; S705: Based on the observation residuals, and combining the prediction error covariance matrix and the observation noise covariance matrix, calculate the Kalman gain matrix; S706: Based on the Kalman gain matrix, the predicted supernode head state vector is corrected to obtain the continuous hydraulic operation state estimation result.

6. The method for monitoring the operational status of a substation rainwater drainage system according to claim 5, characterized in that, S8 specifically includes: S801: Combining the observed residuals and the innovative covariance matrix, construct a dimensionless normalized bias characteristic; S802: Construct state correction intensity features; S803: Perform a difference operation on the dimensionless normalized deviation characteristics to determine the time evolution deviation characteristics; S804: Calculate the network-level deviation energy index based on the predicted supernode head state vector and the continuous hydraulic operation state estimation results; S805: Based on the wireless signal strength parameters, construct a communication reliability weighted deviation feature; S806: Combine the dimensionless normalized deviation feature, the state correction intensity feature, the time evolution deviation feature, the network-level deviation energy index, and the communication reliability weighted deviation feature to obtain the state deviation feature parameter.

7. The method for monitoring the operational status of a substation rainwater drainage system according to claim 1, characterized in that, The operational status monitoring results include: the anomaly type and the risk level corresponding to the anomaly type.

8. A monitoring system for the operational status of a substation rainwater drainage system, characterized in that, include: Processor and memory; The memory stores programs or instructions that can run on the processor, which, when executed by the processor, implement the steps of the substation rainwater drainage system operation status monitoring method as described in any one of claims 1 to 7.

9. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the substation rainwater drainage system operation status monitoring method as described in any one of claims 1 to 7.

Images