A mall pipe network data progressive acquisition and dynamic optimization method

Abstract

The application discloses a kind of mall pipe network data progressive acquisition and dynamic optimization method, it is related to mall building electromechanical and pipe network operation digitalization technical field, specifically related to a kind of mall pipe network data progressive acquisition and dynamic optimization method.The method is with response anchor point table as minimum feasible topological starting condition, according to the pipe network operating object in operation and maintenance task, operation type and business influence boundary generate low amplitude safe perturbation event, in action time window, the response data and unresponsive data of response observation point are collected, form the response signature including response arrival time, response amplitude, recovery curve and unresponsive boundary, and generate candidate topology according to this, again combining conservation residual, response arrival time, recovery curve and unresponsive boundary carry out library check, according to physical response confidence, residual confidence and retest confidence determine topological hierarchy and hierarchical library, for not passing the check or below official topological layer candidate topology generates retest task according to weak evidence item.

Description

Technical Field

[0001] This invention relates to the field of digital technology for the operation and maintenance of electromechanical systems and pipeline networks in shopping malls, specifically a method for the progressive acquisition and dynamic optimization of shopping mall pipeline network data. Background Technology

[0002] The water supply, fire protection, air conditioning, and drainage pipe networks within shopping malls are typically distributed within ceilings, shafts, equipment mezzanines, and shop interior spaces, making them highly concealed. Frequent subsequent modifications can easily lead to discrepancies between blueprints and the actual site conditions. During operation and maintenance, it is necessary to understand valve control ranges, branch connection relationships, and impact boundaries, but these are often limited by business continuity, the intensity of construction intervention, and site accessibility. Therefore, the continuous supplementation and verification of pipe network data has always been a real challenge in shopping mall operation and maintenance. Related industry surveys indicate that a significant proportion of completed commercial complexes lack complete digital pipe network archives; in handling pipe network faults, a large amount of time is spent confirming pipeline routes and valve locations, reflecting the substantial impact of missing basic pipe network data on operation and maintenance efficiency.

[0003] Existing technologies offer solutions to improve pipeline network data through drawings, as-built documentation, equipment ledgers, BIM, GIS, IoT point locations, or model calibration. These solutions integrate node, pipe segment, valve, equipment, and work order records into a unified model, and then enhance data completeness through rule comparison, model correction, or on-site data entry. While this approach is beneficial for establishing a foundational data base, its effectiveness heavily relies on the completeness of the original data, the quality of continuous data entry, and sensor coverage. For scenarios like shopping malls, which involve frequent concealed modifications, numerous local connections, and dispersed end-point locations, relying solely on drawings, ledgers, or model calibration often fails to promptly confirm actual control boundaries and concealed connectivity relationships, and it is also difficult to generate directly verifiable topological evidence. Specifically: Detailed modeling solutions, such as BIM, are difficult to update synchronously after modeling due to changes such as merchant renovations and equipment replacements, leading to a decline in data accuracy over time; operational data acquisition solutions, such as IoT monitoring, focus on collecting operational parameters but do not address the lack of basic data such as topological connectivity and valve control range; and equipment file solutions, such as ledgers and work order management, lack topological connections between equipment records and pipe sections and nodes, failing to support intelligent operation and maintenance needs such as valve control range queries and fault impact analysis.

[0004] Other technologies utilize pressure drop testing, pressure transient analysis, high-frequency sensing, leak or transient location, and dynamic identification routes combined with hydraulic models. These technologies identify anomalies, leak sources, or network conditions by using pressure, flow, temperature difference, or liquid level changes caused by valve, pump, or end-point testing actions, based on propagation delay, waveform characteristics, fitting results, or model comparison. While these methods can reflect the dynamic response of the pipeline network, they typically rely on specialized testing, high-frequency acquisition, relatively stable sensor deployment, or pre-calibrated model conditions. Their application focuses primarily on leak location, transient source identification, or model correction. For concealed pipeline networks in operating shopping malls, directly applying the above dynamic identification approach is susceptible to limitations such as short testing windows, limited disturbance amplitude, sparse observable points, manual readings, and constraints from multi-zone operations. This makes it difficult to continuously and progressively obtain verifiable topological evidence without disrupting operations.

[0005] Therefore, while existing technologies can improve understanding of pipe networks through data integration and dynamic testing, they still have certain limitations. For example, they struggle to balance the boundaries of operational impact, low-intervention implementation conditions, and the continuous verification of topological evidence. In particular, they are difficult to progressively obtain verifiable topological evidence in the context of concealed pipe networks in shopping malls without affecting business operations. Based on this, there is an urgent need for a progressive data collection and dynamic optimization method for shopping mall pipe networks to provide technical means for the continuous verification and updating of the actual connectivity, control range, and impact range of shopping mall pipe networks. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and propose a progressive data acquisition and dynamic optimization method for shopping mall network to solve the above-mentioned problems.

[0007] The objective of this invention is achieved through the following technical solution: a progressive data acquisition and dynamic optimization method for shopping mall pipeline networks based on operation and maintenance perturbation response signatures, comprising: S1, establishing a response anchor point table based on control action points and response observation points of the shopping mall pipeline network, and using the response anchor point table as the starting condition for the minimum feasible topology; the response anchor point table records the control action points, response observation points, and business impact boundaries corresponding to each operation zone; S2, receiving operation and maintenance tasks, and generating low-amplitude safety perturbation events according to the pipeline operation objects, operation types, and business impact boundaries in the operation and maintenance tasks; the low-amplitude safety perturbation events have action amplitude, action time window, and recovery conditions; S3, collecting response data and non-response data of response observation points within the action time window, and generating response signatures based on the response data and non-response data; the response signature includes response arrival time, response amplitude, recovery curve, and non-response boundary; the non-response boundary is formed by response observation points that have not produced response changes within the action time window and have observable parameters; S4, generating candidate topologies based on the response signatures; the candidate topologies include candidate connection relationships and candidate control... The scope and candidate influence range are defined, where the candidate control range is defined by the response observation point that produces the response change, and by the non-response boundary, defining the outer range; S5, generate conserved residuals based on low-amplitude security perturbation events and response signatures, and perform a write-to-validation of the candidate topology based on the conserved residuals, response arrival time, recovery curve, and non-response boundary; S6, when the write-to-validation passes, determine the topology level of the candidate topology based on the physical response confidence, residual confidence, and re-validation confidence. The topology level includes the unverified layer, the restricted use layer, and the formal topology. The system first creates a topology layer and writes candidate topologies into the topology library according to the topology level. If any physical response confidence, residual confidence, or verification confidence is lower than the corresponding hard threshold, the candidate topology is prohibited from entering the formal topology layer. In step S7, a frozen layer verification task is generated if the library write verification fails. An upgraded verification task is generated if the topology level of the candidate topology is lower than that of the formal topology layer. Both the frozen layer verification task and the upgraded verification task are generated based on the weak evidence items of the candidate topology and are used to specify the operation object, response observation point, and signature component to be verified for the next low-amplitude security perturbation event.

[0008] The response anchor point table includes control action points, response observation points, observable parameters, reading methods, allowable disturbance amplitude, and business impact boundaries corresponding to the operating zone. Each operating zone has at least one control action point and at least one response observation point.

[0009] Low-amplitude safety disturbance events are caused by one of the following pipeline operations: valve opening adjustment, end-point water discharge, pump frequency adjustment, branch flushing, and air conditioning water temperature difference disturbance. The amplitude of the action is lower than the equipment's allowable fluctuation boundary, and the action time window is shorter than the allowable duration corresponding to the business impact boundary.

[0010] When generating a response signature, response observation points that produce response changes within the action time window are recorded as the response point set, and response observation points that form non-response boundaries are recorded as the non-response point set. Boundary clarity is then generated based on the response point set and the non-response point set.

[0011] When the response arrival time is inconsistent with the spatial order of the response observation points along the candidate propagation path, the candidate topology is marked as an anomalous candidate topology, and the weak evidence item of the anomalous candidate topology is set as an unstable response arrival time order.

[0012] The conserved residual is generated from the operational amplitude of the low-amplitude safety perturbation event, the response amplitude of the response observation point, and at least one physical quantity among the flow rate, pressure, temperature, and liquid level of the pipeline medium. The conserved residual is used to determine the degree to which the candidate topology explains the changes in physical quantities before and after the perturbation.

[0013] Write verification includes residual verification, delay verification, recovery verification, and boundary verification. Residual verification is used to determine whether the conserved residual falls within the residual boundary. Delay verification is used to determine whether the response arrival time meets the candidate propagation path. Recovery verification is used to determine whether the recovery curve meets the recovery conditions of low-amplitude safety perturbation events. Boundary verification is used to determine whether the unresponded boundary matches the candidate control range.

[0014] When a candidate topology is in the frozen layer, it is prohibited from participating in operation and maintenance queries. When it is in the pending verification layer, it is only used to generate re-verification tasks. When it is in the restricted use layer, it participates in the scope of impact prompts. When it is in the formal topology layer, it participates in the formal queries and updates of the topology library.

[0015] When determining the topology level, the physical response confidence is generated by the integrity of the response signature, the residual confidence is generated by the degree of deviation of the conserved residual from the residual boundary, and the verification confidence is generated by the support results of two or more low-amplitude security perturbation events for the same candidate topology. Based on the physical response confidence, residual confidence, and verification confidence, it is determined whether the candidate topology enters the verification layer and the restricted use layer.

[0016] Weak evidence items include one of the following: unclear non-response boundaries, unstable response arrival time order, conserved residuals exceeding residual boundaries, and mismatch between recovery curves and recovery conditions. The re-verification task specifies the boundary response observation point, sampling time window, operation object to be adjusted, and signature component to be verified based on the weak evidence items.

[0017] The beneficial effects of this invention are: By using a response anchor point table as the initiation condition for the minimum feasible topology, and requiring each operational zone to have at least control action points, response observation points, and operational impact boundaries, this invention can still initiate topology identification from operable and observable locations even when initial drawings are incomplete or existing topology data has gaps. Compared to methods that rely on a one-time complete survey or large-scale preliminary data collection, this scheme ensures that each operational action has a clear incentive entry point and observation exit point, preventing subsequent data collection from falling into disorderly expansion. This facilitates the establishment of a unified evidence standard under lower initial conditions and provides a stable reference for the subsequent generation and verification of candidate topologies.

[0018] By standardizing maintenance actions such as valve opening adjustments, end-point water discharge, pump frequency adjustments, branch flushing, or air conditioning water temperature difference disturbances into low-amplitude safety disturbance events, and by constraining the action amplitude, action time window, and recovery conditions, this invention enables on-site operations that were originally only for maintenance or inspection to also possess controllable physical excitation functions. Existing methods, if relying on additional testing arrangements, often increase on-site coordination costs; if they simply wait for natural operational fluctuations, it is difficult to form comparable evidence. This invention, however, can form a repeatable response process within the operational impact boundary. This not only helps reduce reliance on additional specialized surveying actions but also improves the executability of on-site implementation and the consistency of subsequent verification.

[0019] By constructing response signatures using both response and non-response data within the same action time window, and using non-response boundaries as evidence of the outer edge of candidate control ranges, this invention prevents unchanged observations from being simply considered missing or low-value records. Response observations that produce response changes are used to define the inner edge of the candidate control range, while response observations that do not produce response changes but have observable parameters are used to define the outer edge of the candidate control range. This places the candidate control range and candidate influence range under bilateral constraints. This process improves the accuracy and reliability of boundary determination and can promptly form anomalous candidate topologies when the response arrival time and spatial location order are inconsistent, reducing the risk of misjudging occasional responses as stable connectivity relationships.

[0020] By introducing conserved residuals and using them, along with residual verification, time delay verification, recovery verification, and boundary verification, as the basis for database writing verification, this invention ensures that candidate topologies not only conform to local response phenomena but also explain the changes in physical quantities, propagation timing, and recovery process before and after the perturbation. Existing digital completion paths, if primarily based on drawing adjacency, empirical judgment, or single response results, easily write physically difficult-to-close relationships directly into the topology database. This invention, however, uses multiple verification cross-constraints on candidate topologies, allowing only results with higher interpretability to continue. This reduces the risk of incorrect database writing and improves the traceability, verifiability, and engineering reliability of topology writing results.

[0021] By establishing a frozen layer, a verification layer, a restricted-use layer, and a formal topology layer, and using physical response confidence, residual confidence, and verification confidence as gating conditions, this invention clearly distinguishes the usage permissions of candidate topologies with different maturity levels within the system. Compared to the approach of allowing low-confidence relationships to directly enter the formal database and thus affect queries, updates, and impact scope judgments, this layered mechanism prohibits the frozen layer from participating in operational queries, the verification layer is only used to generate verification tasks, the restricted-use layer only participates in impact scope hints, and only the formal topology layer participates in formal queries and updates. This helps to limit uncertain information within a controlled scope, maintain the stability of the dynamic optimization process of the topology database, and reduce the possibility of erroneous relationships spreading to subsequent business chains.

[0022] For weak evidence items such as unclear non-response boundaries, unstable response arrival times, residuals exceeding residual boundaries, and mismatches between recovery curves and recovery conditions, this invention does not simply require continued supplementary sampling. Instead, it generates frozen layer verification tasks or upgraded verification tasks to clarify the operational objects, response observation points, and signature components to be verified for the next low-amplitude security perturbation event. Existing supplementary sampling methods often only continue to accumulate records and may not be able to form effective verification for decisive doubts. This invention, however, directly targets the verification tasks to the weak links in the evidence chain, ensuring that each new maintenance action serves to clarify specific uncertainties. This improves the asymptotic convergence efficiency of the topology evidence chain, enabling the dynamic optimization of the topology library to be based on a closed loop of continuous verification, hierarchical release, and evidence accumulation. Attached Figure Description

[0023] Figure 1 The process of this invention Figure 1 ; Figure 2 The process of this invention Figure 2 ; Figure 3 The process of this invention Figure 3 . Detailed Implementation

[0024] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] Example 1 like Figure 1As shown, this embodiment uses the water supply network of the second-floor dining area of ​​a shopping mall as an example to illustrate a progressive data acquisition and dynamic optimization method for shopping mall water supply networks based on maintenance perturbation response signatures. The water supply network of this dining area includes one zone valve, three terminal water points, and one adjacent but different zone terminal water point. The system pre-establishes a topology library, which records at least the network nodes, candidate connection relationships, candidate control ranges, candidate influence ranges, topology levels, and verification tasks. The system is also connected to the maintenance task system to receive maintenance, inspection, terminal water release, and valve adjustment tasks, and can record manually read pressure gauge readings, electronic pressure gauge readings, or observation results entered by inspection terminals.

[0026] In this embodiment, the system first establishes a response anchor point table. The response anchor point table records the control action points, response observation points, and business impact boundaries corresponding to each operational zone, and serves as the activation condition for the minimum feasible topology. Taking the second-floor catering area as an example, the control action points include the catering area water supply zone valves and end-point water outlets. The response observation points include the first, second, and third end-point pressure observation points within the catering area, and the fourth end-point pressure observation point adjacent to the retail area. The response anchor point table also records the observable parameters, reading methods, allowable disturbance amplitude, and upper limit of the action time window for each response observation point. Preset parameter values ​​can be set as follows: the adjustment amplitude of the zone valve opening should not exceed 10% of the current opening; the single action time window is 30 to 120 seconds; and the recovery observation time window is 30 to 180 seconds after the action ends. These parameter values ​​are determined based on the equipment specifications of the valves and end-point water outlets, the business period impact boundaries, and the time available for maintenance personnel to complete the readings. After project deployment, updates via calibration tasks are allowed, but the updated action amplitude must not exceed the equipment's allowable fluctuation boundaries.

[0027] When the maintenance task system receives a maintenance task titled "Investigation of Low Water Pressure at the Second Floor Catering Area Terminal," the system generates a low-amplitude safety disturbance event based on the pipeline operation object, operation type, and business impact boundary in the maintenance task. In this embodiment, the pipeline operation object is the water supply zone valve in the catering area, the operation type is a short-term reduction in valve opening, the action amplitude is set to a 5% reduction in the current valve opening, the action time window is set to 60 seconds, and the recovery condition is that the valve returns to its pre-action opening position and each response observation point returns to near the pre-action baseline pressure within the recovery observation time window. The event record of this low-amplitude safety disturbance event includes at least the event number, associated maintenance task number, pipeline medium, pipeline operation object, action amplitude, action start time, action end time, recovery condition, associated response observation point, business impact boundary, and executor identification.

[0028] To establish a pre-action baseline, the system collects the initial pressure at each response observation point within a baseline sampling window prior to the execution of a low-amplitude safety disturbance event. The engineering default value for the baseline sampling window is 60 seconds before the start of the action, and the system takes the average of manual or electronic pressure gauge readings within this time window as the baseline pressure. During the execution of the low-amplitude safety disturbance event, the system collects pressure changes, response arrival time, and recovery curves at each response observation point within the action time window, and simultaneously records non-response data. Non-response data is not the same as uncollected data; in this embodiment, non-response data refers to data where the response observation point has observable parameters, is within the same action time window, and has completed reading recording, but the change in pressure relative to the baseline pressure has not reached the response judgment boundary. The response judgment boundary can be the larger of three times the standard deviation of the baseline pressure and the engineering minimum resolvable pressure change value, which can be set to 5 kPa.

[0029] The system generates a response signature based on the response data and non-response data. The response signature includes the response arrival time, response amplitude, recovery curve, and non-response boundary. The response arrival time is the difference between the time when the pressure change at the response observation point first reaches the response judgment boundary and the start time of the low-amplitude safety perturbation event; the response amplitude is the maximum deviation of the pressure relative to the baseline pressure within the action time window; the recovery curve is the time series of pressure returning to near the baseline after the action ends; the non-response boundary is formed by response observation points that did not produce a response change within the action time window and have observable parameters. To ensure a unified calculation method for the response signature, this embodiment uses the following response amplitude formula: in, For the first The response amplitude at each observation point For action time window, For the first Each response observation point at time... Pressure value, For the first The baseline pressure at each response observation point within the baseline sampling window. If If the response determination boundary is reached, the response observation point is added to the response point set; if If the response determination boundary is not reached and the response observation point is validly read within the action time window, then the response observation point is added to the non-response point set.

[0030] In this embodiment, the first and second end-pressure observation points show pressure changes in the same direction within the action time window, the third end-pressure observation point changes close to the response judgment boundary, and the fourth end-pressure observation point does not show a response change and its reading is valid. The system defines the inner range of the candidate control range by the first and second end-pressure observation points, and defines the outer range of the candidate control range by the non-response boundary formed by the fourth end-pressure observation point. Because the change of the third end-pressure observation point is close to the response judgment boundary, it is retained as an observation point to be verified and is not directly used as the boundary of the formal control range. The above processing is used to illustrate the judgment relationship between the set of response points, the set of non-response points, and the observation points to be verified, and is not written into the measured result values.

[0031] The system generates candidate topologies based on response signatures. Candidate topologies include candidate connectivity relationships, candidate control ranges, and candidate influence ranges. Candidate connectivity relationships indicate a control relationship to be verified between the zone valve and the response observation point where a response change occurs; the candidate control range is defined by the inner boundary of the response observation point where the response change occurs and the outer boundary of the non-responding boundary; the candidate influence range records the merchants, equipment, or end-points that the candidate control range may affect. After generation, the candidate topology is not directly written into the topology library as a formal topology but instead enters the library writing and verification process.

[0032] Before writing the database for verification, the system generates a conserved residual based on low-amplitude safety perturbation events and response signatures. For water supply networks, this embodiment uses the degree of explanation for pressure changes as the basic caliber for the conserved residual. The system calculates the predicted pressure change based on the adjustment range of the zone valve opening, the response amplitude at each response observation point, the candidate propagation distance of the response observation point relative to the zone valve, and the preset attenuation coefficient, and then compares it with the measured response amplitude. The candidate propagation distance is calculated from the node sequence length, pipe segment length, or number of floor zone spans corresponding to the candidate connection relationships in the topology library; the preset attenuation coefficient is determined by the calibration task of the same network medium, similar valves, and pipe segments with similar diameters; the predicted response amplitude is obtained by looking up a table or linear interpolation using the action amplitude, candidate propagation distance, and preset attenuation coefficient. The conserved residual can be calculated using the following formula: in, To conserve residuals, For the set of response points, For the first The response amplitude at each observation point The predicted response amplitude is obtained based on the action amplitude of a low-amplitude security perturbation event and candidate propagation paths. This represents the minimum resolvable pressure variation value in the engineering process. The default engineering value for the residual boundary can be 0.35; when... When the residual is not greater than the residual boundary, the residual verification output passes the marker; when When the value exceeds the residual boundary, the residual check outputs a residual out-of-bounds flag.

[0033] In addition to residual verification, the system also performs time delay verification, recovery verification, and boundary verification. Time delay verification determines whether the response arrival time conforms to the proximity relationship of the candidate propagation path; if the response observation point closer to the partition valve is consistently later than the distant response observation point and cannot be explained by reading delay, a time delay anomaly flag is output. Recovery verification determines whether the recovery curve meets the recovery conditions for low-amplitude safety disturbance events; in this embodiment, the recovery condition can be set to the response observation point pressure returning to within ±2 times the baseline pressure within 180 seconds after the action ends, within the response judgment boundary range. Boundary verification determines whether the non-response boundary matches the candidate control range; if the observation point in the non-response point set is spatially located outside the response point set and the reading is valid, the boundary verification outputs a boundary matching flag; if the non-response point set is insufficient or the reading is invalid, an unclear boundary flag is output.

[0034] The system determines the topology level of candidate topologies based on the write verification results and confidence levels. Physical response confidence is generated from the integrity of the response signature, residual confidence is generated from the deviation of the conserved residuals from the residual boundaries, and verification confidence is generated from the support results of low-amplitude security perturbation events for the same candidate topology. To ensure that the non-response boundary can serve as an executable criterion outside the candidate control range, the system first calculates the clarity of the non-response boundary: in, For unresponsive boundary clarity, This represents the number of response observation points that complete a valid read within the action time window without producing any response change. This represents the number of response observations that are near the candidate control range boundary but have not completed a valid read. The number of response observation points whose response amplitude is close to the response determination boundary but fail to be stably included in the response point set or the non-response point set. The closer it is to one, the clearer the boundary of the non-response; when When the boundary clarity threshold is below a preset threshold, the boundary check outputs an unclear boundary marker and writes the unclear unresponsive boundary into the weak evidence item. In this embodiment, the engineering default value of the boundary clarity threshold is set to 0.6, which is derived from the engineering requirement that there is at least one valid unresponsive observation point and the number of pending observation points near the boundary is controlled.

[0035] In this embodiment, the physical response confidence level is calculated using the following formula: in, For the confidence level of the physical response, In response to arrival time integrity, In response to amplitude integrity, For unresponsive boundary clarity, To restore curve integrity, each integrity value ranges from zero to one, and is determined by the number of valid records within the action time window, whether the response judgment boundary has been reached, whether the set of non-responding points has valid readings, and whether there are continuous records within the recovery observation time window. The above weights are engineering default values, derived from the strength of the influence of each piece of information on the candidate control range judgment. They can be updated by calibration samples after project deployment, but the updated weights remain non-negative and sum to one.

[0036] The residual confidence level can be calculated using the following formula: in, For residual confidence level, To conserve residuals, This represents the residual boundary. The verification confidence level is set to an initial value of 0.5 after the first low-amplitude security perturbation event; it increases when subsequent low-amplitude security perturbation events provide consistent support for the same candidate topology, and decreases when subsequent low-amplitude security perturbation events provide conflicting results. The system calculates the gating criterion accordingly: in, The gating criterion for entering the formal topology layer. For verification confidence levels, the default hard thresholds can be set as follows: physical response confidence level not lower than 0.7, residual confidence level not lower than 0.65, verification confidence level not lower than 0.7, and gating criterion not lower than 0.7. If any physical response confidence level, residual confidence level, or verification confidence level is lower than the corresponding hard threshold, the system will prohibit the candidate topology from entering the formal topology layer.

[0037] In the single-cycle operation of this embodiment, if the residual verification, delay verification, recovery verification, and boundary verification all pass, but the re-verification confidence level is still the initial value calibrated after the first event, the system writes the candidate topology into the verification layer or the restricted use layer, instead of entering the formal topology layer. Candidate topologies in the verification layer are only used to generate re-verification tasks; candidate topologies in the restricted use layer can participate in impact scope hints, but must not cover existing connections in the formal topology layer. Candidate topologies in the frozen layer are prohibited from participating in operation and maintenance queries; candidate topologies in the formal topology layer participate in formal queries and updates of the topology library.

[0038] When the write verification fails, the system generates a frozen layer verification task; when the candidate topology's topology level is lower than the formal topology level, the system generates an upgraded verification task. Both the frozen layer verification task and the upgraded verification task are generated based on the weak evidence items of the candidate topology. Weak evidence items can include unclear non-response boundaries, unstable response arrival times, conserved residuals exceeding residual boundaries, and mismatches between the recovery curve and recovery conditions. In this embodiment, if the third end pressure observation point is close to the response judgment boundary but has not stably entered the response point set, the system records the weak evidence item as unclear non-response boundaries and generates an upgraded verification task, specifying that the next low-amplitude safety perturbation event will still use the catering area water supply zone valve as the operation object, the third end pressure observation point and the fourth end pressure observation point as the key response observation points, and the non-response boundary and recovery curve as the signature components to be verified.

[0039] Through the above process, the system transforms a low-amplitude safety perturbation event into a response signature without requiring a complete one-time mapping of the pipeline network. It then uses non-response boundaries to limit the outer edge of the candidate control range and controls the write permissions of the candidate topology through conserved residuals and topology-level gating. Before meeting the hard threshold of the formal topology layer, this candidate topology can only be used for generating restricted prompts or verification tasks, thus allowing the topology library to gradually converge with subsequent maintenance tasks.

[0040] Example 2 like Figure 1 and Figure 2 As shown in Example 1, this embodiment, based on the water supply network and fire-fighting terminal test network of a shopping mall with multiple operating zones, illustrates the cross-zone establishment method of the response anchor point table, the enhanced judgment method of the non-response boundary, and the marking method of the abnormal candidate topology. The shopping mall network includes at least a catering area, a retail area, a cinema area, and an underground equipment area. Each operating zone has an independent business impact boundary and forms control action points through zone valves, terminal test points, branch water discharge points, or pump control action points. Each operating zone also has at least one response observation point among pressure gauges, electronic pressure gauges, terminal reading points, liquid level observation points, or inspection terminal input points. Before executing low-amplitude safety disturbance events, the system first establishes a response anchor point table on a per-operating-zone basis, ensuring that each operating zone has at least one control action point and one response observation point. Each response observation point is associated with observable parameters, reading methods, allowable disturbance amplitude, upper limit of the action time window, and business impact boundary.

[0041] The response anchor point table does not record the complete surveyed pipeline drawings, but rather a set of engineering anchor points used to initiate minimum feasible topology acquisition. For any operational zone, the system first reads the zone identifier, pipeline medium, known nodes, candidate connection relationships, and historical maintenance tasks from the existing topology library, and then reads the inspection routes, valve operation permissions, end-point test permissions, and business hours restrictions from the maintenance task system. If an operational zone lacks control action points, the system does not dispatch low-amplitude safety disturbance events to that operational zone; if an operational zone lacks response observation points, the system marks that operational zone as an anchor point missing zone and generates an anchor point supplementation task. The anchor point supplementation task specifies the control action points, response observation points, reading methods, and observable parameters to be supplemented, and does not directly generate candidate topologies.

[0042] To ensure a consistent set of criteria for triggering the response anchor table, the system calculates the availability of anchor points for each operational zone: in, For the first Availability of partition anchors for each operational partition This is a complete flag for the basic field. It is set to one if the operation partition has recorded the operation partition identifier, pipeline medium, business impact boundary and topology level, otherwise it is set to zero. The number of control action points within the operational zone that meet the requirements for operating permissions and security boundaries; The number of response observation points within this operational zone that meet the requirements for reading method and observable parameters; This is an executable flag for the operational impact boundary. It is set to one when the upper limit of the action time window is shorter than the allowed impact duration for that operational zone and the allowed disturbance amplitude does not exceed the equipment's allowed fluctuation boundary; otherwise, it is set to zero. Only when... At one point, the operational zone entered the range for generating low-amplitude security disturbance events; when If the value is less than one, the system freezes the candidate topology generation entry for that operational partition and retains the reason for the missing anchor point.

[0043] In this embodiment, low-amplitude safety disturbance events are generated by one of the following pipeline operations: fire-fighting terminal water testing, water supply terminal water release, short-term adjustment of zone valve opening, or pump frequency fine-tuning. The system determines the action time window based on the business impact boundary of the operating zone and the action amplitude based on the equipment's allowable fluctuation boundary. The engineering default range of the action time window is 30 to 120 seconds, and the engineering default range of the recovery observation time window is 30 to 180 seconds after the action ends. If the selected pipeline operation is fire-fighting terminal water testing, the system restricts the testing action to the test period allowed by fire protection maintenance and limits the response observation point to points with observable pressure, flow, or terminal readings in the same fire-fighting zone and adjacent operating zones. If the selected pipeline operation is water supply terminal water release, the system restricts the action amplitude to the allowable fluctuation boundary that does not affect the normal water use of merchants and restores the system to the state before the action ends.

[0044] When the system executes a low-amplitude safety perturbation event, it simultaneously collects response data and non-response data within the same action time window. Response data includes the response arrival time, response amplitude, and recovery curve of the response observation point; non-response data includes observation records that were effectively read within the same action time window, have observable parameters, and have not reached the response judgment boundary. The system distinguishes between non-response data and uncollected data: non-response data must have the reading time, reading source, observable parameters, and baseline comparison caliber; uncollected data is only recorded as anchor point missing or reading failure records and is not included in the non-response point set. The response judgment boundary is the larger of the baseline fluctuation boundary and the minimum resolvable engineering change value. The baseline fluctuation boundary is determined by the baseline sampling window before the action begins, and the minimum resolvable engineering change value is determined by the accuracy of pressure gauges, flow meters, level observation devices, or manual readings.

[0045] For the first For each response observation point, the system calculates the response amplitude using the following formula: in, For the first The response amplitude at each observation point For action time window, For the first Each response observation point at time... observable parameter values, This is the baseline value of the observation point within the baseline sampling window. The observable parameter can be one of pressure, flow rate, level, or end-point reading. If... If the response determination boundary is reached, the response observation point is added to the response point set; if If the response determination boundary is not reached and the read is valid, the response observation point is added to the unresponded point set; if the read is missing or the read source does not meet the requirements, the response observation point is added to the pending observation point set.

[0046] The system forms cross-zone non-response boundaries based on the set of response points, the set of non-response points, and the set of pending observation points. For candidate control areas, the set of response points defines the inner side of the candidate control area, the set of non-response points defines the outer side, and the set of pending observation points is not directly used as boundary evidence. If a non-response point and a response point are located in adjacent operating zones along the same candidate propagation path, and the non-response point is outside the response point, then the non-response point is used as evidence of the outer boundary. If the non-response point is spatially inside the response point, or if the response point spatially crosses the non-response point and still produces a response change, then the system marks the corresponding candidate topology as a spatial order conflict.

[0047] The system calculates the sharpness of the unresponsive boundary: in, For unresponsive boundary clarity, This represents the number of response observation points that complete a valid read within the action time window without producing any response change. This represents the number of response observations that are near the candidate control range boundary but have not completed a valid read. This refers to the number of response observation points whose response amplitude is close to the response determination boundary but fail to stably be included in the response point set or the non-response point set. The engineering default value for the boundary sharpness threshold is set to 0.6; when... When the threshold is below this threshold, the system does not release the set of unresponsive points as evidence outside the candidate control range, but instead writes the unresponsive boundary as a weak evidence item.

[0048] The system also calculates cross-partition spatial order consistency to determine whether the spatial distribution of the response point set and the non-response point set conforms to the candidate propagation path: in, For spatial order consistency, This represents the number of spatial order conflicts. The number of response observation points in the response point set. This represents the number of response observations in the set of non-responding points. This represents the number of unresolved observations whose response amplitude is close to the response decision boundary. Spatial order conflicts include situations where an unresponding point is inside a response point, a response point crosses the unresponding boundary, and a far-end response point on the same candidate propagation path reaches the response decision boundary before a near-end response point and cannot be explained by read latency. The engineering default value for the spatial order consistency threshold is set to 0.7; when When the value is below this threshold, the system marks the candidate topology as an abnormal candidate topology and sets the weak evidence item as an unstable response arrival time order or a spatial order conflict.

[0049] When generating candidate topologies, the system does not directly write all unresponsive points outside the candidate control range, but first calculates the boundary release criterion: in, Release criteria for unresponsive boundaries To determine the availability of operational zone anchor points corresponding to this low-amplitude security disturbance event, For unresponsive boundary clarity, This represents the spatial order consistency. When... When the threshold value is not lower than the boundary release threshold, the system releases the set of unresponsive points as evidence outside the candidate control range and generates candidate connectivity relationships, candidate control ranges, and candidate influence ranges; when When the threshold is below the boundary release threshold, the system only saves the set of response points and the set of pending observation points, and does not use the set of unresponsive points outside the formally defined candidate control range. The engineering default value for the boundary release threshold is 0.65, and it is updated by the calibration task after project deployment, but the updated threshold must not be lower than the unresponsive boundary clarity threshold.

[0050] When the response arrival time is inconsistent with the spatial order of the response observation points, the system performs anomaly candidate topology marking. The response arrival time order is determined by the time when each response observation point first reaches the response judgment boundary; the spatial order is determined by the candidate propagation path, operational zone adjacency relationship, floor direction, and node sequence in the response anchor point table. If the near-end response observation point is consistently later than the far-end response observation point, and the difference cannot be explained by the reading method, sampling time, or manual input delay, the system marks the candidate topology as anomaly candidate topology. Anomaly candidate topology does not enter the formal topology layer; its candidate connection relationship is written to the verification layer or frozen layer, and the unstable response arrival time order is recorded in the weak evidence item. This anomaly marking is only used to constrain the use rights and re-verification tasks of candidate topologies, and does not write bypasses, valve internal leakage, incorrect drawing direction, or concealed connections as definite facts into the topology library.

[0051] For candidate topologies marked as having spatial order conflicts or unstable response arrival time sequences, the system generates a boundary enhancement verification task. This verification task records at least the task number, associated candidate topology, network operation object, action type, action time window, key response observation points, key non-response observation points, signature components to be verified, weak evidence items, and recovery conditions. If the weak evidence item is an unclear non-response boundary, the verification task designates response observation points near the boundary as key reading points; if the weak evidence item is an unstable response arrival time sequence, the verification task designates near-end and far-end response observation points to be read synchronously within the same action time window; if the weak evidence item is a spatial order conflict, the verification task designates operational partitions on both sides of the candidate propagation path for boundary reconfirmation. After the verification task is generated, the system freezes the formal write permissions for the corresponding candidate topology, allowing it to be used for limited impact scope indication or subsequent verification scheduling.

[0052] Example 3 like Figures 1 to 3 As shown, this embodiment, based on Embodiments 1 and 2, focuses on the dynamic optimization of branch circuits in a shopping mall's air conditioning water system, illustrating the implementation methods for conserved residuals, database write verification, topology-level gating, and verification task generation. The air conditioning water system includes chilled water supply and return mains, floor branch circuits, shop terminal coils, branch regulating valves, variable frequency pumps, and several supply and return water temperature observation points, pressure observation points, and flow observation points. The system pre-establishes a response anchor point table, which records the branch regulating valves, pump frequency adjustment points, supply and return water temperature observation points, pressure observation points, flow observation points, reading methods, allowable disturbance amplitude, upper limit of the action time window, and business impact boundary corresponding to each operating zone. The response anchor point table serves as a prerequisite for the initiation of low-amplitude safety disturbance events; operating zones that do not meet the anchor point availability requirements do not generate candidate topologies.

[0053] In this embodiment, low-amplitude safety disturbance events are caused by pump frequency fine-tuning or short-term branch regulating valve opening adjustment. The system determines the action amplitude, action time window, and recovery conditions based on the pipeline operation object, operation type, and business impact boundary in the operation and maintenance task. The action amplitude of pump frequency fine-tuning is lower than the allowable fluctuation boundary of the variable frequency pump, and the action amplitude of short-term branch regulating valve opening adjustment is lower than the allowable fluctuation boundary of normal operation of the terminal coil. The engineering default range of the action time window is 30 seconds to 120 seconds, and the engineering default range of the recovery observation time window is 30 seconds to 300 seconds after the action ends. After the action ends, when the pump frequency or branch regulating valve opening returns to the state before the action, and the supply and return water temperature difference, pressure, or flow rate returns to near the baseline before the action, the system records that the recovery conditions are met.

[0054] The system collects data on changes in supply and return water temperature difference, pressure, and flow rate, as well as response arrival time and recovery curves, at response observation points within the action time window, and simultaneously records non-response data. Non-response data refers to data where the response observation point has observable parameters, is within the same action time window, and has completed a valid read, but the changes in its supply and return water temperature difference, pressure, or flow rate do not reach the response judgment boundary. Non-response data is not equivalent to a read failure; observation points with read failures are added to the pending observation point set and are not included in the external limitation of the candidate control range. The system generates a response signature based on the response point set, the non-response point set, and the pending observation point set, and compares the response signature with the candidate propagation path.

[0055] For air conditioning water systems, the system incorporates at least one physical quantity from the supply and return water temperature difference, pressure, and flow rate into the conserved residuals. For the first... For each response observation point, the system first calculates the normalized response amplitude: in, For the first Normalized response amplitude at each response observation point For action time window, For the first Each response observation point at time... The observed values ​​of supply and return water temperature difference, pressure, or flow rate. The baseline value of the observation point within the baseline sampling window. This refers to the engineering dimensional quantity of the physical quantity corresponding to the observation point of the response. The engineering dimensional quantity is determined by the equipment nameplate range, historical stable operating range, or calibration task, and is used to bring different physical quantities into the same comparison caliber.

[0056] The system generates a predicted normalized response amplitude based on the action amplitude, candidate propagation paths, and physical quantity type, and then compares it with the actual normalized response amplitude to obtain the conserved residual of the air conditioning water system. in, For the conservation residual of the air conditioning water system, For the set of response points, For the first The weights of physical quantities at each response observation point This represents the actual normalized response amplitude. The predicted normalized response amplitude is obtained based on the action amplitude, candidate propagation paths, and calibrated attenuation relationship. This is the minimum resolvable change normalized. The weights of physical quantities are determined by the observable parameters recorded in the response anchor table. When the same response observation point simultaneously has records of temperature difference, pressure, and flow rate, the system uses the weight combination determined by the calibration task to form a single comparison value, where each weight is non-negative and their sum is one. The engineering default value for the residual boundary is set to 0.35 and is updated after calibration of similar air conditioning water branches.

[0057] Write verification includes residual verification, delay verification, recovery verification, and boundary verification. Residual verification takes the conserved residuals and residual boundaries of the air conditioning water system as input. When the residual is not greater than the residual boundary, the output residual passes through the marker. When the residual exceeds the residual boundary, a residual out-of-bounds flag is output. Delay verification takes the response arrival time and candidate propagation path as input to determine if the proximity of the near-end response observation point, the far-end response observation point, and the candidate propagation path is consistent. If the far-end response observation point consistently reaches the response decision boundary earlier than the near-end response observation point, and the reading method and sampling time cannot explain this difference, a delay order instability flag is output. Recovery verification takes the recovery curve and recovery conditions as input to determine if each response observation point returns to the vicinity of the pre-action baseline after the action ends. If the recovery curve does not meet the recovery conditions within the recovery observation time window, a recovery mismatch flag is output. Boundary verification takes the non-response boundary and candidate control range as input to determine if the set of non-response points is located outside the candidate control range and the reading is valid. If the set of non-response points is insufficient, there are too many unresolved observation points, or the boundary clarity is insufficient, a non-response boundary unclear flag is output.

[0058] The system summarizes the output of the four write verifications into a verification pass rate: in, To verify the pass rate of the write operation, This is used as a residual check flag; it is set to one when the residual check passes and to zero when the residual exceeds the limit. This is a delay check flag; it is set to one when the delay order is stable and to zero when the delay order is unstable. To restore the check flag, a value of 1 is used if the restoration condition is met, and a value of zero is used if the restoration does not match. This is used as a boundary verification flag. It is set to 1 when the non-responding boundary matches the candidate control range, and to 0 when the non-responding boundary is unclear. The default value of the write library verification pass threshold is set to 1, meaning that the candidate topology is only allowed to enter the topology level gate when all four verifications pass; if any verification fails, the candidate topology enters the frozen layer and generates a frozen layer re-verification task.

[0059] After the write-to-database verification passes, the system determines the topology level of the candidate topology based on the physical response confidence, residual confidence, and verification confidence. The physical response confidence is generated from the response signature integrity, which includes response arrival time integrity, response amplitude integrity, non-response boundary clarity, and recovery curve integrity. The residual confidence is generated from the deviation of the conserved residual from the residual boundary. The verification confidence is generated from the support results of two or more low-amplitude security perturbation events for the same candidate topology; the verification confidence increases when subsequent low-amplitude security perturbation events provide consistent support with the existing candidate topology, and decreases when subsequent low-amplitude security perturbation events provide conflicting results.

[0060] The system determines the topology level release criterion using the following formula: in, The topology-level release criterion For the confidence level of the physical response, For residual confidence level, To retest the confidence level, This is for the write-to-library validation pass rate. The engineering default value for the physical response confidence hard threshold is set to 0.7, the engineering default value for the residual confidence hard threshold is set to 0.65, the engineering default value for the revalidation confidence hard threshold is set to 0.7, and the engineering default value for the topology-level release criterion hard threshold is set to 0.7. If any confidence level is lower than the corresponding hard threshold, or... When the candidate topology falls below the hard threshold of the topology layer release criterion, the system prohibits the candidate topology from entering the formal topology layer.

[0061] The system executes four-layer database writes based on topology-level release criteria and write verification results. When a candidate topology is in the frozen layer, it is prohibited from participating in maintenance queries, impact scope prompts, and formal topology updates; it is only allowed as an associated object for re-verification tasks. When a candidate topology is in the pending verification layer, it is only used to generate re-verification tasks and does not participate in impact scope prompts. When a candidate topology is in the restricted usage layer, it is allowed to participate in impact scope prompts, but it must not overwrite existing connections in the formal topology layer, nor can it be used as a basis for automatic dispatching or determining the water outage range. When a candidate topology is in the formal topology layer, it participates in formal topology database queries and updates and can serve as a source of candidate propagation paths for subsequent low-amplitude security perturbation events. If subsequent re-verification tasks produce conflicting results, the corresponding connections in the formal topology layer are marked as pending verification, and the system rolls back to the previous stable version for querying.

[0062] When the write verification fails, the system generates a frozen layer verification task; when the write verification passes but the candidate topology does not reach the hard threshold of the formal topology layer, the system generates an upgraded verification task. The frozen layer verification task is used to eliminate candidate topology errors caused by residual out-of-bounds errors, unstable response arrival time order, recovery mismatch, or unclear non-response boundaries; the upgraded verification task is used to supplement verification confidence, boundary clarity, or response signature integrity. Both types of verification tasks are generated based on weak evidence items, but their task objectives and usage permissions differ: the frozen layer verification task does not allow candidate topologies to participate in any operational queries, while the upgraded verification task allows candidate topologies to participate in impact scope hints within the restricted usage layer.

[0063] To ensure the mapping from weak evidence items to re-examination tasks is executable, the system calculates the priority of re-examination tasks: in, Prioritize the verification task. Marked as an unresponsive boundary with unclear boundaries. This is a marker indicating unstable time delay order. Marking for residual out-of-bounds errors. To restore mismatch markers, a value of 1 is assigned if the corresponding weak evidence item exists, and 0 if it does not. The weights are default values ​​derived from the impact of each weak evidence item on the risk of erroneous writes to the formal topology layer. These weights are updated via a calibration task after project deployment, ensuring they remain non-negative and sum to 1. The system determines the execution order of re-verification tasks based on task priority and operational impact boundaries. When multiple re-verification tasks have the same priority, those in the frozen layer and with fewer associated operational partitions are executed first.

[0064] Different weak evidence items correspond to different retesting tasks. When the weak evidence item is an unclear non-response boundary, the retesting task specifies boundary response observation points, non-response observation points, and pending observation points, and requires that the same pipeline operation object be maintained for the next low-amplitude safety disturbance event, extending the recovery observation time window or supplementing synchronous readings. When the weak evidence item is an unstable response arrival time sequence, the retesting task specifies near-end response observation points, far-end response observation points, and a unified reading time reference, and requires synchronous acquisition within the same action time window for the next low-amplitude safety disturbance event. When the weak evidence item is a residual exceeding the limit, the retesting task specifies the operation object to be adjusted, candidate propagation paths, and calibrated attenuation relationships, and requires that the action amplitude, response amplitude, and candidate propagation distance be reacquired during the retest. When the weak evidence item is a recovery mismatch, the retesting task specifies the recovery curve, recovery conditions, and recovery observation time window, and prohibits releasing candidate topology levels before the recovery conditions are met.

[0065] In one implementation, after receiving the air conditioning water branch balance adjustment task, the system converts pump frequency fine-tuning or short-term branch regulating valve opening adjustment into low-amplitude safety disturbance events; within the action time window, it collects supply and return water temperature difference, pressure, or flow rate changes, and forms a set of response points, a set of non-response points, and a set of unresolved observation points; the system then generates response signatures, candidate topologies, and air conditioning water system conservation residuals, and performs residual verification, delay verification, recovery verification, and boundary verification. If all four verifications pass and the topology layer release criterion reaches the hard threshold, the candidate topology enters the formal topology layer; if all four verifications pass but the re-verification confidence is insufficient, the candidate topology enters the unverified layer or the restricted use layer, and an upgrade re-verification task is generated; if any verification fails, the candidate topology enters the frozen layer, and a frozen layer re-verification task is generated.

[0066] Through the above process, the candidate topology is not directly written into the formal topology layer after a single response. Instead, it is released layer by layer after being constrained by the conservation residual, four write verifications, confidence hard threshold, and re-verification task. This allows the shopping mall network topology to be progressively collected and dynamically optimized during continuous operation and maintenance.

[0067] The above description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be modified within the scope of the concept described herein by means of the above teachings or the technology or knowledge in related fields.

Claims

1. A method for progressive data collection and dynamic optimization of shopping mall pipeline networks based on operation and maintenance perturbation response signatures, characterized in that, include: S1. Establish a response anchor point table based on the control action points and response observation points of the shopping mall's pipeline network, and use the response anchor point table as the starting condition for the minimum feasible topology. The response anchor point table records the control action points, response observation points, and business impact boundaries corresponding to each operating zone. S2. Receive maintenance tasks and generate low-amplitude safety disturbance events based on the pipeline operation objects, operation types, and business impact boundaries in the maintenance tasks. The low-amplitude safety disturbance events have action amplitude, action time windows, and recovery conditions. S3. Collect response data and non-response data from the response observation points within the action time window, and generate a response signature based on the response data and non-response data. The response signature includes response arrival time, response amplitude, recovery curve, and non-response boundary. The non-response boundary is formed by response observation points that have not produced response changes within the action time window and have observable parameters. S4. Generate candidate topologies based on the response signatures. The candidate topologies include candidate connection relationships, candidate control ranges, and candidate impact ranges. The candidate control range is defined by the response observation points that produce response changes. S5, a conserved residual is generated based on the low-amplitude security perturbation event and the response signature, and a write-to-library verification is performed on the candidate topology based on the conserved residual, the response arrival time, the recovery curve, and the unresponding boundary; S6, when the write-to-library verification passes, the topology level of the candidate topology is determined based on the physical response confidence, residual confidence, and re-verification confidence. The topology level includes a layer to be verified, a restricted-use layer, and a formal topology layer, and the candidate topology is written to the library according to the topology level. Enter the topology library, where if any of the physical response confidence, residual confidence, and verification confidence is lower than the corresponding hard threshold, the candidate topology is prohibited from entering the formal topology layer; S7, when the write library verification fails, a frozen layer verification task is generated; when the topology level of the candidate topology is lower than the formal topology layer, an upgrade verification task is generated. Both the frozen layer verification task and the upgrade verification task are generated based on the weak evidence items of the candidate topology and are used to specify the operation object, response observation point, and signature component to be verified for the next low-amplitude security perturbation event.

2. The method according to claim 1, characterized in that, The response anchor point table includes control action points, response observation points, observable parameters, reading methods, allowable disturbance amplitude, and business impact boundaries corresponding to the operating zone, and each operating zone has at least one control action point and at least one response observation point.

3. The method according to claim 1, characterized in that, The low-amplitude safety disturbance event is formed by one of the following pipeline operations: valve opening adjustment, end water discharge, pump frequency adjustment, branch flushing, and air conditioning water temperature difference disturbance. The amplitude of the action is lower than the equipment's allowable fluctuation boundary, and the action time window is shorter than the allowable duration corresponding to the business impact boundary.

4. The method according to claim 1, characterized in that, When generating the response signature, the response observation points that produce response changes within the action time window are recorded as the response point set, the response observation points that form the non-response boundary are recorded as the non-response point set, and the boundary clarity is generated based on the response point set and the non-response point set.

5. The method according to claim 4, characterized in that, When the response arrival time is inconsistent with the spatial order of the response observation points along the candidate propagation path, the candidate topology is marked as an anomalous candidate topology, and the weak evidence item of the anomalous candidate topology is set as an unstable response arrival time order.

6. The method according to claim 1, characterized in that, The conserved residual is generated from the operational amplitude of the low-amplitude safety perturbation event, the response amplitude of the response observation point, and at least one physical quantity among the flow rate, pressure, temperature, and liquid level of the pipeline medium. The conserved residual is used to determine the degree to which the candidate topology explains the changes in physical quantities before and after the perturbation.

7. The method according to claim 1, characterized in that, The write verification includes residual verification, delay verification, recovery verification, and boundary verification. The residual verification is used to determine whether the conserved residual falls within the residual boundary. The delay verification is used to determine whether the response arrival time conforms to the candidate propagation path. The recovery verification is used to determine whether the recovery curve conforms to the recovery conditions of the low-amplitude security perturbation event. The boundary verification is used to determine whether the non-responding boundary matches the candidate control range.

8. The method according to claim 1, characterized in that, When a candidate topology is in the frozen layer, it is prohibited from participating in operation and maintenance queries; when it is in the pending verification layer, it is only used to generate re-verification tasks; when it is in the restricted use layer, it participates in the scope of impact prompts; and when it is in the formal topology layer, it participates in the formal query and update of the topology library.

9. The method according to claim 8, characterized in that, When determining the topology level, the physical response confidence is generated by the integrity of the response signature, the residual confidence is generated by the deviation of the conserved residual from the residual boundary, and the verification confidence is generated by the support results of two or more low-amplitude security perturbation events for the same candidate topology. Based on the physical response confidence, the residual confidence, and the verification confidence, it is determined whether the candidate topology enters the verification layer and the restricted use layer.

10. The method according to claim 1, characterized in that, The weak evidence items include one of the following: unclear non-response boundary, unstable response arrival time sequence, conservation residual exceeding residual boundary, and mismatch between recovery curve and recovery condition. The re-verification task specifies the boundary response observation point, sampling time window, operation object to be adjusted, and signature component to be verified based on the weak evidence items.

Images