Urban facility comprehensive operation and maintenance management system and method based on digital twinning

Abstract

The invention discloses an urban facility comprehensive operation and maintenance management system and method based on digital twinning, and relates to the technical field of digital twinning, and the method comprises the steps: collecting feature parameters when a feature event occurs and a safety evaluation result is abnormal, and building a feature parameter dynamic threshold prediction model; characteristic parameters of the associated facilities and the cables in the characteristic time period are collected, and a dynamic time interval formula is constructed; when a characteristic event occurs in the associated facility, collecting a real-time characteristic parameter and inputting the real-time characteristic parameter into the characteristic parameter dynamic threshold prediction model, generating a dynamic threshold and calculating a real-time deviation rate; if the real-time characteristic parameters exceed a dynamic threshold value, early warning is given out, otherwise, a dynamic time interval is obtained through calculation, collection is carried out according to the dynamic time interval, self-adaptive early warning threshold value adjustment is achieved, the early warning accuracy is improved, the monitoring frequency is intelligently adjusted according to the running state change, and the resource utilization rate is improved.

Description

Technical Field

[0001] This invention relates to the field of digital twin technology, specifically to a comprehensive operation and maintenance management system and method for urban facilities based on digital twins. Background Technology

[0002] Existing urban infrastructure operation and maintenance management methods rely heavily on manual inspections and fixed-rule scheduling, which suffer from high subjectivity, inconsistent work, and low efficiency. Static thresholds are difficult to cope with complex urban operation and maintenance scenarios, especially when the operating status of urban infrastructure is complex and changeable. The patent CN120634821A, "Management Method and System for Urban Underground Cable Network Based on Digital Twin," proposes a new management approach. By integrating the basic attribute data and real-time sensing data of the urban underground cable network, a digital twin mapping model of the network is constructed. This enables real-time monitoring of the urban underground cable network, identification of key network nodes with abnormal status and their abnormal transmission links, and generation of maintenance instructions based on risk assessment. Despite progress in real-time monitoring and anomaly identification, some limitations remain. The methods still rely heavily on statically set thresholds and lack adaptability to complex and dynamic environmental factors. At the same time, traditional management methods have failed to achieve deep data fusion among various facilities, resulting in an inability to respond flexibly and make dynamic decisions in the face of changing urban operating conditions. Therefore, in order to improve the comprehensive operation and maintenance management level of urban facilities, this invention proposes a comprehensive operation and maintenance management system and method for urban facilities based on digital twins. Through real-time dynamic data analysis, it can more flexibly adapt to environmental factors and realize intelligent decision-making. Based on the data-driven dynamic scheduling mechanism, it can accurately identify and handle potential risks, and significantly improve the overall sustainability and operational efficiency of urban infrastructure. Summary of the Invention

[0003] The purpose of this invention is to provide a digital twin-based integrated operation and maintenance management system and method for urban facilities to solve the problems raised in the prior art.

[0004] To address the aforementioned technical problems, this invention provides the following technical solution: a method for integrated operation and maintenance management of urban facilities based on digital twins, the method comprising: Step S100: Collect cable operation parameters, construct cable data sets and upload them to the urban facility management platform based on the digital twin model to form a comprehensive operation record of urban facilities; Step S200: Filter records containing abnormal cable data groups in the historical comprehensive operation records, extract abnormal operation data within the sliding time window, calculate the correlation index of urban facilities based on the abnormal operation data, and determine the related facilities based on the correlation index. Step S300: Statistically analyze the frequency of events occurring in the historical abnormal operation records of related facilities, extract feature events, construct a parameter set for related facilities and a cable parameter set, take the value of the feature event as the target variable, calculate the importance coefficient of the input variable through a random forest model, and filter to obtain the feature parameters under different feature events; Step S400: Collect feature parameters when feature events occur and when the safety assessment results are abnormal during the training period, and establish a dynamic threshold prediction model for feature parameters; collect feature parameters of related facilities and cables within the feature time period, and construct a dynamic time interval formula; Step S500: When a characteristic event occurs in the associated facility, real-time characteristic parameters are collected and input into the dynamic threshold prediction model of the characteristic parameters to generate a dynamic threshold and calculate the real-time deviation rate; if the real-time characteristic parameters exceed the dynamic threshold, an early warning is issued; otherwise, the dynamic time interval is calculated and data is collected according to the dynamic time interval.

[0005] Furthermore, step S100 includes: Step S101: Install various types of operation monitoring sensors at the location of the underground cable to monitor the cable's operation parameters, preset the data acquisition time interval, synchronize the acquired operation parameters in time, generate cable data sets with timestamps, and upload them to the urban facility management platform built on a digital twin model. Step S102: Collect the connection relationship between underground cables and various urban facilities, construct the cable topology, collect the operation data corresponding to each type of urban facility based on the topology, and upload the operation data with timestamps to the urban facility management platform to form a comprehensive operation record of urban facilities. Step S103: Set an initial threshold for cable operating parameters, compare the operating parameters in each collected cable data set with the initial threshold, mark operating parameters that exceed the initial threshold as abnormal operating parameters, and mark cable data sets with abnormal operating parameters as abnormal cable data sets. By deploying various types of operation monitoring sensors in underground cables, it is possible to collect multi-dimensional operation parameters such as temperature, current, voltage, and humidity of the cables in real time. Through time synchronization and timestamp marking, the data can be accurately aligned and traceable, providing a reliable data foundation for subsequent condition assessment and fault diagnosis. By collecting data on the connection between cables and urban facilities, constructing the cable topology, and combining it with the operational data of urban facilities, a digital twin model of the cable system and related equipment can be formed in the urban facility management platform. This enables panoramic and dynamic management of urban infrastructure and improves the systematicness and intelligence of operation and maintenance work.

[0006] Furthermore, step S200 includes: Step S201: In the historical integrated operation record set, filter the records containing abnormal cable data groups and mark them as historical abnormal integrated operation records. Collect the collection time point corresponding to the abnormal cable data group in each historical abnormal integrated operation record. The preset sliding time window is a, forming a sliding time range of [Aa, A+a], where A represents the collection time point. Step S202: Obtain the topology of the cable corresponding to each historical abnormal comprehensive operation record, obtain the set of urban facilities that are connected to the cable, collect the operation data of each urban facility within the sliding time range, if a certain operation data exceeds a preset threshold and the duration reaches the preset duration, then mark the operation data as abnormal operation data, collect the numerical change of the abnormal operation data of a certain historical abnormal comprehensive operation record within the sliding time range, calculate the numerical change rate of the abnormal operation data, summarize all historical abnormal comprehensive operation records, and calculate the average numerical change rate of the abnormal operation data; Step S203: Based on the cable topology, calculate the topological distance between the cable and urban facilities, and based on GIS coordinates, calculate the spatial distance between the cable and urban facilities. Normalize the topological distance and spatial distance, and perform a weighted sum of the normalized topological distance and spatial distance to calculate the proximity score. Step S204: Count the number of times abnormal operation data of a certain city facility occurs, calculate the frequency of occurrence of the city facility, normalize the average rate of change of the value, the proximity score and the frequency of occurrence, and calculate the correlation index according to the following formula: ; Where S represents the correlation index, s1 represents the normalized frequency of occurrence, and s2 represents the frequency of occurrence. b s1 represents the average rate of change of the normalized value of the b-th abnormal data point, s2 represents the normalized proximity score, and w1 and w2 represent the average rate of change of the normalized value. b w3 and w3 represent the weights of the frequency of occurrence, the b-th abnormal running data, and the proximity score, respectively. Step S205: Preset a correlation index threshold, compare the correlation index of each city facility with the correlation index threshold, and if it exceeds the correlation index threshold, mark the corresponding city facility as a related facility. By filtering abnormal cable data groups from historical integrated operation records and constructing a sliding time window based on the collection time point, the facility operation status before and after the cable abnormality can be included in the analysis scope, revealing the potential correlation between abnormal events from the time dimension, avoiding the omissions and misjudgments caused by traditional methods that only analyze a single time point; The connection relationship is determined by the topology of cables and urban facilities, and the spatial distance is calculated by combining GIS coordinates. The two are then normalized and weighted to obtain a comprehensive proximity score. Through the dual constraints of topological distance and spatial distance, the accuracy of abnormal propagation path modeling is significantly improved, and the physical and logical relationship between cables and urban facilities can be more realistically reflected. By calculating the numerical and average rate of change of abnormal operating data, the dynamic evolution trend of facility operating status can be reflected, thereby judging the intensity and persistence of the anomaly, providing a quantitative basis for the root cause analysis of cable anomalies, and improving the scientificity and objectivity of anomaly identification. The three key parameters of frequency of occurrence, rate of change of abnormal data, and proximity score are normalized and weighted and fused to build a unified correlation index model. By setting the correlation index threshold, urban facilities that are highly correlated with abnormal cable height can be automatically identified, realizing the transformation from "manual judgment" to "algorithm recognition" and greatly improving the efficiency of operation and maintenance analysis. Furthermore, step S300 includes: Step S301: Obtain the events that occur in the historical abnormal operation records of each associated facility, count the frequency of occurrence of each event, preset the frequency threshold, and mark the events that exceed the frequency threshold as feature events; Step S302: In a certain historical abnormal comprehensive operation record, collect the operation data of related facilities and construct a set of related facility parameters, collect the operation parameters of the cable and construct a set of cable parameters. If a feature event exists in a certain time window in the historical abnormal comprehensive operation record, the value is 1; otherwise, the value is 0. Step S303: Using the set of parameters of associated facilities and the set of parameters of cables as input variables, and the value of feature events as target variables, a sample set is constructed by selecting monitoring data within the time window range. The input variables are normalized and trained by a random forest model. The feature importance coefficient of each input variable in the model is calculated and filtered according to the preset importance coefficient threshold to determine the feature parameters of associated facilities and cables in different feature events. By statistically analyzing the frequency of events in each associated facility and setting thresholds to identify characteristic events, it is possible to effectively uncover frequently occurring abnormal patterns in multiple historical abnormal operation records. This provides important abnormal characteristic information for further analysis, reduces noise interference, focuses on key events that truly affect system safety, and improves the accuracy and sensitivity of fault identification. In the historical anomaly comprehensive operation record, multi-dimensional operation data of related facilities and cables are collected and integrated into a parameter set, which provides a comprehensive data foundation for subsequent modeling and analysis. This enables the analysis of equipment from multiple dimensions, thereby improving the accuracy of fault location and correlation analysis. By training input variables using a random forest model, key parameters related to feature events can be automatically identified based on a large amount of historical data, and the feature importance coefficient of each parameter in the model can be calculated. Compared with traditional manual screening methods, machine learning models can more comprehensively and accurately uncover potential fault features, providing a scientific basis for the selection of feature parameters and avoiding the bias that may be caused by manual intervention.

[0007] Furthermore, step S400 includes: Step S401: Select several consecutive days as the training period. When a certain characteristic event occurs in a certain related facility, arrange staff to conduct a safety assessment of the related facility and cable. When the safety assessment result is abnormal, collect the characteristic parameters of the related facility and cable at the current time point. Step S402: Obtain the type, temporal features and environmental information of the current feature event, assign a number to the type, normalize the temporal features and environmental information, use the number, the normalized temporal features and environmental information as input, use the feature parameters of the associated facilities and cables at the current time point as input, train through the XGBoost regression model, and generate a dynamic threshold prediction model for feature parameters. Step S403: Collect the time points corresponding to the occurrence of the characteristic event and the abnormality of the safety assessment result. Take the time point when the characteristic event occurs as the starting point and the time point when the safety assessment result is abnormal as the ending point to obtain the characteristic time period. Obtain the characteristic parameters of the associated facilities and cables corresponding to each collection time point in the characteristic time period. Step S404: Calculate the dynamic threshold of the characteristic parameters of the associated facilities and cables corresponding to each collection time point using the characteristic parameter threshold prediction model, and calculate the difference between the threshold and the characteristic parameters of the associated facilities and cables collected at the collection time point to obtain the characteristic parameter deviation value of the associated facilities and cables. Step S405: Obtain the feature parameter deviation value between every two adjacent acquisition time points, calculate the feature parameter deviation rate, and establish the dynamic time interval formula as follows: ; Where T represents the characteristic time period, P ie Let Q be the deviation rate of the e-th feature parameter in the i-th time interval. ie Let t represent the weight of the deviation rate of the e-th feature parameter in the i-th time interval. i Let f represent the weight of the i-th time interval, f represent the total number of feature parameters, and r represent the total number of time intervals. The abnormal comprehensive operation records within the training period are summarized, and the dynamic time interval formula is trained by the linear regression method. The updated feature parameter deviation rate weight and time interval weight are applied to the dynamic time interval formula. By assigning staff to conduct safety assessments of related facilities and cables, and collecting relevant characteristic parameters when anomalies are found in the safety assessments, potential safety hazards can be detected in a timely manner, equipment safety risks can be identified in advance, the incidence of sudden failures can be effectively reduced, and an accurate data foundation can be provided for subsequent predictive models, thereby improving the safety of equipment operation. By utilizing the type, temporal characteristics, and environmental information of feature events through the XGBoost regression model, a dynamic threshold prediction model for feature parameters is established. Through comprehensive analysis of multi-dimensional data, the safety threshold for equipment operation can be dynamically adjusted, making the prediction more flexible and accurate, and avoiding the misjudgment or omission that may be caused by traditional static thresholds. By comparing the occurrence time of characteristic events with the time points of safety assessment anomalies, characteristic time periods are formed and the deviation values ​​of characteristic parameters are calculated. Dynamic thresholds are set for each time point and the deviation values ​​are calculated. This allows for accurate judgment of the equipment's operational deviations during abnormal events. Compared with traditional fixed thresholds, the application of dynamic thresholds makes monitoring more consistent with the actual operating status of the equipment and improves the sensitivity to minor anomalies.

[0008] By combining the XGBoost regression model with the dynamic time interval formula, an efficient and intelligent dynamic threshold prediction system was established, which can realize real-time monitoring and early warning. Through automated model training and dynamic adjustment, the system response speed is improved, and intelligent decision-making based on real-time data is possible, further enhancing operation and maintenance efficiency and security.

[0009] Furthermore, step S500 includes: Step S501: When a characteristic event occurs in the associated facility, obtain the characteristic parameters corresponding to the occurrence of the characteristic event, collect the real-time characteristic parameters of the associated facility and cable, input the characteristic event number, the normalized time sequence characteristics and environmental information into the characteristic parameter dynamic threshold prediction model, generate the characteristic parameter dynamic threshold of the associated facility and cable, if there is a real-time characteristic parameter greater than the characteristic parameter dynamic threshold, issue an early warning to the staff to check, if there is no such parameter, proceed to step S502; Step S502: Calculate the real-time deviation rate of the characteristic parameters and input it into the dynamic time interval formula to obtain the dynamic time interval, and collect the characteristic parameters of the associated facilities and cables according to the dynamic time interval; By acquiring real-time characteristic parameters of associated facilities and cables when characteristic events occur, and using a trained dynamic threshold prediction model of characteristic parameters for real-time comparison, an early warning is automatically triggered when the real-time parameters exceed the threshold. This can effectively identify potential abnormal states. Compared with the traditional fixed threshold judgment method, the dynamic threshold model can adaptively adjust the threshold according to time series characteristics, environmental factors and event types, thereby reducing false alarms and false negatives and significantly improving the accuracy and reliability of anomaly identification. When the system detects that the real-time characteristic parameters exceed the limits, it can immediately issue an early warning notification to the staff, realizing the transformation from "passive response" to "proactive prevention and control". It can intervene in the inspection in advance when the equipment shows signs of failure, shorten the time cycle from risk discovery to handling, and effectively reduce the probability of sudden equipment failure. By calculating the real-time deviation rate of characteristic parameters and inputting it into the dynamic time interval formula, the system achieves adaptive adjustment of the data sampling frequency. When the equipment's operating status is relatively stable, the system automatically extends the sampling interval to reduce invalid data collection. When the deviation rate increases or the operating status fluctuates, the system automatically shortens the sampling interval to achieve higher frequency monitoring. The dynamic sampling strategy can significantly reduce the system's computational burden and data redundancy while ensuring monitoring accuracy.

[0010] To better implement the above methods, a comprehensive urban facility operation and maintenance management system based on digital twins is also proposed. The system includes a comprehensive operation record module, a related facility module, a feature parameter module, a dynamic time interval module, and a real-time analysis module. Integrated Operation Record Module: Collects cable operation parameters, constructs cable data sets, and uploads them to the urban facility management platform based on a digital twin model to form an integrated operation record of urban facilities; Related facilities module: Filter records containing abnormal cable data groups in historical comprehensive operation records, extract abnormal operation data within the sliding time window, calculate the correlation index of urban facilities based on the abnormal operation data, and determine related facilities based on the correlation index; Feature Parameter Module: Statistically analyzes the frequency of events occurring in the historical abnormal operation records of related facilities, extracts feature events, constructs a set of parameters for related facilities and a set of parameters for cables, takes the value of the feature event as the target variable, calculates the importance coefficient of the input variable through a random forest model, and selects feature parameters under different feature events; Dynamic time interval module: Collects feature parameters when feature events occur and when safety assessment results are abnormal during the training period, and establishes a dynamic threshold prediction model for feature parameters; collects feature parameters of associated facilities and cables within the feature time period, and constructs a dynamic time interval formula; Real-time analysis module: When a characteristic event occurs in the associated facility, it collects real-time characteristic parameters and inputs them into the dynamic threshold prediction model of the characteristic parameters to generate a dynamic threshold and calculate the real-time deviation rate; if the real-time characteristic parameters exceed the dynamic threshold, an early warning is issued; otherwise, the dynamic time interval is calculated and data is collected according to the dynamic time interval.

[0011] Furthermore, the integrated operation record module includes an integrated operation record unit and an abnormal cable data group unit: Establish a comprehensive operation record unit: Install various types of operation monitoring sensors at the location of underground cables to monitor the cable's operation parameters, preset the data acquisition time interval, synchronize the acquired operation parameters in time, generate cable data sets with timestamps, and upload them to the urban facility management platform built based on a digital twin model. Collect the connection relationship between underground cables and various urban facilities, construct the cable topology, and based on the topology, collect the operation data corresponding to each urban facility, and upload the time-stamped operation data to the urban facility management platform to form a comprehensive operation record of urban facilities. Abnormal Cable Data Group Unit: Sets an initial threshold for cable operating parameters, compares the operating parameters in each collected cable data group with the initial threshold, marks operating parameters that exceed the initial threshold as abnormal operating parameters, and marks cable data groups with abnormal operating parameters as abnormal cable data groups.

[0012] Furthermore, the associated facility module includes an abnormal operation data analysis unit and an associated facility identification unit: Abnormal Operation Data Analysis Unit: In the historical comprehensive operation record set, records containing abnormal cable data groups are filtered and marked as historical abnormal comprehensive operation records. The collection time point corresponding to the abnormal cable data group in each historical abnormal comprehensive operation record is collected. A sliding time window is preset to form a sliding time range. The topology of the cable corresponding to each historical abnormal comprehensive operation record is obtained, and the set of urban facilities connected to the cables is obtained. The operation data of each urban facility is collected within the sliding time range. If a certain operation data exceeds a preset threshold and the duration reaches a preset duration, the operation data is marked as abnormal operation data. The numerical change of the abnormal operation data of a certain historical abnormal comprehensive operation record within the sliding time range is collected, and the numerical change rate of the abnormal operation data is calculated. All historical abnormal comprehensive operation records are summarized, and the average numerical change rate of the abnormal operation data is calculated. Identify associated facility units: Count the number of times abnormal operation data of a certain city facility occurs, calculate the occurrence frequency of the city facility, normalize the average rate of change of the value, the proximity score and the occurrence frequency to calculate the correlation index, preset the correlation index threshold, compare the correlation index of each city facility with the correlation index threshold, if it exceeds the correlation index threshold, the corresponding city facility is marked as an associated facility.

[0013] Furthermore, the real-time analysis module includes an anomaly detection unit and a dynamic data acquisition unit: Anomaly detection unit: When a characteristic event occurs in the associated facility, the characteristic parameters corresponding to the characteristic event are obtained, the real-time characteristic parameters of the associated facility and cable are collected, the number of the characteristic event, the normalized time sequence characteristics and environmental information are input into the characteristic parameter dynamic threshold prediction model, and the dynamic threshold of the characteristic parameters of the associated facility and cable is generated. If there is a real-time characteristic parameter that is greater than the characteristic parameter dynamic threshold, an early warning notice is issued to the staff to check. Dynamic acquisition unit: Calculates the real-time deviation rate of characteristic parameters and inputs it into the dynamic time interval formula to obtain the dynamic time interval, and acquires the characteristic parameters of associated facilities and cables according to the dynamic time interval.

[0014] Compared with existing technologies, the beneficial effects of this invention are as follows: Real-time mapping and multi-source data fusion of urban facility operation status are achieved through a digital twin model, improving the comprehensiveness of monitoring; a sliding time window and multi-dimensional correlation index calculation method are used to accurately identify related facilities affecting cable anomalies; key feature parameters are automatically filtered using a random forest model to achieve intelligent feature event identification; a dynamic threshold prediction model is built based on XGBoost to achieve adaptive adjustment of early warning thresholds, improving early warning accuracy; a dynamic time interval formula is introduced to intelligently adjust the monitoring frequency according to changes in operation status, improving resource utilization; and a closed-loop intelligent operation and maintenance system is constructed from data collection, anomaly identification, feature extraction to dynamic early warning, realizing predictive operation and maintenance and proactive management of urban facilities. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating a comprehensive operation and maintenance management method for urban facilities based on digital twins, according to the present invention. Figure 2 This is a schematic diagram of the structure of a digital twin-based integrated operation and maintenance management system for urban facilities according to the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments 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, 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.

[0017] Please see Figure 1 and Figure 2 This invention provides a technical solution: a method for integrated operation and maintenance management of urban facilities based on digital twins, the method comprising: Step S100: Collect cable operation parameters, construct cable data sets and upload them to the urban facility management platform based on the digital twin model to form a comprehensive operation record of urban facilities; Step S100 includes: Step S101: Install various types of operation monitoring sensors at the location of the underground cable to monitor the cable's operation parameters, preset the data acquisition time interval, synchronize the acquired operation parameters in time, generate cable data sets with timestamps, and upload them to the urban facility management platform built on a digital twin model. Step S102: Collect the connection relationship between underground cables and various urban facilities, construct the cable topology, collect the operation data corresponding to each type of urban facility based on the topology, and upload the operation data with timestamps to the urban facility management platform to form a comprehensive operation record of urban facilities. Step S103: Set an initial threshold for cable operating parameters, compare the operating parameters in each collected cable data set with the initial threshold, mark operating parameters that exceed the initial threshold as abnormal operating parameters, and mark cable data sets with abnormal operating parameters as abnormal cable data sets. For example, taking the operation monitoring of urban underground power distribution networks as an example, there are multiple 10kV underground cable lines in a certain municipal area to supply power to commercial and residential areas. Multiple types of operational monitoring sensors are installed at key cable nodes (such as cable wells, junction boxes, and branch boxes) within the area, including: Fiber optic temperature sensors are used to monitor the temperature of the cable's outer sheath and conductors in real time. Humidity sensors are used to monitor the humidity in cable trenches or underground environments; Current and voltage sensors are used to monitor the electrical operating parameters of cables; Partial discharge sensors are used to identify characteristics of cable insulation degradation. Vibration and tilt sensors are used to detect abnormalities in manhole covers or supports; Data on the connection relationships between underground cables and various urban facilities (such as substations, street light control cabinets, drainage pumping stations, traffic signal systems, etc.) is collected to construct a complete cable topology model. For example, the main power supply cable in Area A connects to two branch nodes, leading to the power supply ring network of the commercial area and the power supply ring network of the residential area, respectively.

[0018] At the same time, operational data for facilities in various cities are collected, such as: Substation outgoing line current and voltage; Load rate of the street light control system; Start-up and shutdown status and power consumption of drainage pumping stations; Power supply status of the traffic signal system.

[0019] After all data is synchronized in time, the timestamped operational data is uploaded to the urban facilities management platform to form a comprehensive urban facilities operation record, realizing the correlation mapping between cable operation status and urban facilities operation load; Based on historical operating experience and standard specifications, initial threshold values ​​for cable operating parameters are set, for example: Conductor temperature threshold: 90℃; Ambient humidity threshold: 95%; Current threshold: 110% of rated current; Partial discharge signal amplitude threshold: 30mV; When a cable data set is collected at a certain moment and the temperature parameter reaches 95℃ and the partial discharge signal exceeds 35mV, the system marks the data set as an abnormal cable data set and records the abnormality type as "temperature over-limit + insulation discharge abnormality".

[0020] Step S200: Filter records containing abnormal cable data groups in the historical comprehensive operation records, extract abnormal operation data within the sliding time window, calculate the correlation index of urban facilities based on the abnormal operation data, and determine the related facilities based on the correlation index. Step S200 includes: Step S201: In the historical integrated operation record set, filter the records containing abnormal cable data groups and mark them as historical abnormal integrated operation records. Collect the collection time point corresponding to the abnormal cable data group in each historical abnormal integrated operation record. The preset sliding time window is a, forming a sliding time range of [Aa, A+a], where A represents the collection time point. Step S202: Obtain the topology of the cable corresponding to each historical abnormal comprehensive operation record, obtain the set of urban facilities that are connected to the cable, collect the operation data of each urban facility within the sliding time range, if a certain operation data exceeds a preset threshold and the duration reaches the preset duration, then mark the operation data as abnormal operation data, collect the numerical change of the abnormal operation data of a certain historical abnormal comprehensive operation record within the sliding time range, calculate the numerical change rate of the abnormal operation data, summarize all historical abnormal comprehensive operation records, and calculate the average numerical change rate of the abnormal operation data; Step S203: Based on the cable topology, calculate the topological distance between the cable and urban facilities, and based on GIS coordinates, calculate the spatial distance between the cable and urban facilities. Normalize the topological distance and spatial distance, and perform a weighted sum of the normalized topological distance and spatial distance to calculate the proximity score. Step S204: Count the number of times abnormal operation data of a certain city facility occurs, calculate the frequency of occurrence of the city facility, normalize the average rate of change of the value, the proximity score and the frequency of occurrence, and calculate the correlation index according to the following formula: ; Where S represents the correlation index, s1 represents the normalized frequency of occurrence, and s2 represents the frequency of occurrence. b s1 represents the average rate of change of the normalized value of the b-th abnormal data point, s2 represents the normalized proximity score, and w1 and w2 represent the average rate of change of the normalized value. b w3 and w3 represent the weights of the frequency of occurrence, the b-th abnormal running data, and the proximity score, respectively. Step S205: Preset a correlation index threshold, compare the correlation index of each city facility with the correlation index threshold, and if it exceeds the correlation index threshold, mark the corresponding city facility as a related facility. For example, based on the topology of the abnormal cable, the system automatically extracts a set of urban facilities that are connected to it, including: commercial area power distribution ring network; underground drainage pumping station; street light control box; Within the sliding time range, the load current of the commercial area power distribution ring network is 345-420, and the threshold is 350, which is abnormal; the start-stop frequency of the underground drainage pump station is 6, and the threshold is 4, which is abnormal; the voltage of the street light control box is 214-223, and the threshold is 220±10%, which is not abnormal. The calculated average rate of change for abnormal load current was 18.5%; the average rate of change for abnormal start / stop frequency was 22.3%. The topological distance of the commercial area's power distribution ring network is 1, and the spatial distance is 35. The underground drainage pumping station has a topological distance of 2 and a spatial distance of 80. The topological distance of the street light control box is 3, and the spatial distance is 120. After normalization, the topological distance of the commercial area's power distribution ring network is 0.1, and the spatial distance is 0.2. The topological distance between the underground drainage pumping stations is 0.4, and the spatial distance is 0.5. The topological distance of the street light control box is 0.8, and the spatial distance is 0.9. Proximity = 0.6 × topological distance + 0.4 × spatial distance. The calculated proximity of the commercial area's power distribution ring network is 0.14; the proximity of the underground drainage pump station is 0.44; and the proximity of the street light control box is 0.84. The preset correlation threshold is 0.5, the correlation index of the commercial area power distribution ring network is 0.59, the correlation index of the underground drainage pumping station is 0.56, and the correlation index of the street light control box is 0.25. The power distribution ring network and underground drainage pumping station in the commercial area are interconnected.

[0021] Step S300: Statistically analyze the frequency of events occurring in the historical abnormal operation records of related facilities, extract feature events, construct a parameter set for related facilities and a cable parameter set, take the value of the feature event as the target variable, calculate the importance coefficient of the input variable through a random forest model, and filter to obtain the feature parameters under different feature events; Step S300 includes: Step S301: Obtain the events that occur in the historical abnormal operation records of each associated facility, count the frequency of occurrence of each event, preset the frequency threshold, and mark the events that exceed the frequency threshold as feature events; Step S302: In a certain historical abnormal comprehensive operation record, collect the operation data of related facilities and construct a set of related facility parameters, collect the operation parameters of the cable and construct a set of cable parameters. If a feature event exists in a certain time window in the historical abnormal comprehensive operation record, the value is 1; otherwise, the value is 0. Step S303: Using the set of parameters of associated facilities and the set of parameters of cables as input variables, and the value of feature events as target variables, a sample set is constructed by selecting monitoring data within the time window range. The input variables are normalized and trained by a random forest model. The feature importance coefficient of each input variable in the model is calculated and filtered according to the preset importance coefficient threshold to determine the feature parameters of associated facilities and cables in different feature events. For example, a characteristic event of a commercial area's power distribution ring network is a sudden increase in load current, and a characteristic event of an underground drainage pumping station is an excessively high start-stop frequency. During time window T1, the current of the commercial area distribution ring network is 360, the power factor of the commercial area distribution ring network is 0.91, the start-stop frequency of the underground drainage pump station is 3, the cable temperature is 88, the partial discharge intensity is 45, the characteristic event - load surge is 1, and the characteristic event - excessive start-stop frequency is 0. During time window T2, the current of the commercial area distribution ring network is 340, the power factor of the commercial area distribution ring network is 0.93, the start-stop frequency of the underground drainage pump station is 5, the cable temperature is 86, the partial discharge intensity is 38, the characteristic event - load surge is 0, and the characteristic event - excessive start-stop frequency is 1. In time window T3, the current of the commercial area distribution ring network is 370, the power factor of the commercial area distribution ring network is 0.9, the start-stop frequency of the underground drainage pump station is 2, the cable temperature is 92, the partial discharge intensity is 60, the characteristic event - load surge is 1, and the characteristic event - excessive start-stop frequency is 0. In time window T4, the current of the commercial area distribution ring network is 330, the power factor of the commercial area distribution ring network is 0.95, the start-stop frequency of the underground drainage pump station is 1, the cable temperature is 80, the partial discharge intensity is 30, the characteristic event - load surge is 0, and the characteristic event - excessive start-stop frequency is 0. The system inputs the above sample set into the machine learning model for training. The model selected is the random forest algorithm, with the following parameters: number of decision trees: 200; maximum depth: 8; minimum number of sample splits: 5; random seed: 42. After training, the system calculates the feature importance coefficients of each input variable in the model, and the results are as follows: In the case of the "sudden increase in load current" event, The current of the commercial area power distribution ring network is 0.32, the power factor of the commercial area power distribution ring network is 0.18, the start-stop frequency of the underground drainage pump station is 0.05, the cable temperature is 0.24, and the partial discharge intensity is 0.21. The system presets the feature importance coefficient threshold to 0.15, and the screening results are the current of the commercial area power distribution ring network, the power factor of the commercial area power distribution ring network, the cable temperature, and the partial discharge intensity. In the case of "excessive start-stop frequency", The start-stop frequency of the underground drainage pump station is 0.33, the motor temperature is 0.26, the cable current carrying capacity is 0.2, the cable temperature is 0.16, and the partial discharge intensity is 0.05. The screening results show that the start-stop frequency, motor temperature, cable current carrying capacity, and cable temperature are the key parameters for the underground drainage pump station.

[0022] Step S400: Collect feature parameters when feature events occur and when the safety assessment results are abnormal during the training period, and establish a dynamic threshold prediction model for feature parameters; collect feature parameters of related facilities and cables within the feature time period, and construct a dynamic time interval formula; Step S400 includes: Step S401: Select several consecutive days as the training period. When a certain characteristic event occurs in a certain related facility, arrange staff to conduct a safety assessment of the related facility and cable. When the safety assessment result is abnormal, collect the characteristic parameters of the related facility and cable at the current time point. Step S402: Obtain the type, temporal features and environmental information of the current feature event, assign a number to the type, normalize the temporal features and environmental information, use the number, the normalized temporal features and environmental information as input, use the feature parameters of the associated facilities and cables at the current time point as input, train through the XGBoost regression model, and generate a dynamic threshold prediction model for feature parameters. Step S403: Collect the time points corresponding to the occurrence of the characteristic event and the abnormality of the safety assessment result. Take the time point when the characteristic event occurs as the starting point and the time point when the safety assessment result is abnormal as the ending point to obtain the characteristic time period. Obtain the characteristic parameters of the associated facilities and cables corresponding to each collection time point in the characteristic time period. Step S404: Calculate the dynamic threshold of the characteristic parameters of the associated facilities and cables corresponding to each collection time point using the characteristic parameter threshold prediction model, and calculate the difference between the threshold and the characteristic parameters of the associated facilities and cables collected at the collection time point to obtain the characteristic parameter deviation value of the associated facilities and cables. Step S405: Obtain the feature parameter deviation value between every two adjacent acquisition time points, calculate the feature parameter deviation rate, and establish the dynamic time interval formula as follows: ; Where T represents the characteristic time period, P ie Let Q be the deviation rate of the e-th feature parameter in the i-th time interval. ie Let t represent the weight of the deviation rate of the e-th feature parameter in the i-th time interval. i Let f represent the weight of the i-th time interval, f represent the total number of feature parameters, and r represent the total number of time intervals. The abnormal comprehensive operation records within the training period are summarized, and the dynamic time interval formula is trained by the linear regression method. The updated feature parameter deviation rate weight and time interval weight are applied to the dynamic time interval formula. For example, the duration of event 1 is 8, the ambient temperature is 35, the humidity is 60, the external interference is 45, the cable temperature is 92, the partial discharge intensity is 60, and the motor temperature is -. The duration of Event 2 was 15, the ambient temperature was 33, the humidity was 65%, the external interference was 40%, the cable temperature was 88%, the partial discharge intensity was 42%, and the motor temperature was 85%. The system normalizes the above input variables and then outputs: input: event number + normalized time series characteristics + normalized environmental information; Output: Characteristic parameters of the associated facilities and cables collected; Both are used to train the XGBoost regression model, with the following training parameter settings: Number of trees: 300; Learning rate: 0.05; Maximum depth: 6; Subsampling rate: 0.8; Evaluation metric: RMSE.

[0023] Step S500: When a characteristic event occurs in the associated facility, real-time characteristic parameters are collected and input into the characteristic parameter dynamic threshold prediction model to generate a dynamic threshold and calculate the real-time deviation rate; if the real-time characteristic parameters exceed the dynamic threshold, an early warning is issued; otherwise, the dynamic time interval is calculated and data is collected according to the dynamic time interval. Step S500 includes: Step S501: When a characteristic event occurs in the associated facility, obtain the characteristic parameters corresponding to the occurrence of the characteristic event, collect the real-time characteristic parameters of the associated facility and cable, input the characteristic event number, the normalized time sequence characteristics and environmental information into the characteristic parameter dynamic threshold prediction model, generate the characteristic parameter dynamic threshold of the associated facility and cable, if there is a real-time characteristic parameter greater than the characteristic parameter dynamic threshold, issue an early warning to the staff to check, if there is no such parameter, proceed to step S502; Step S502: Calculate the real-time deviation rate of the characteristic parameters and input it into the dynamic time interval formula to obtain the dynamic time interval, and collect the characteristic parameters of the associated facilities and cables according to the dynamic time interval; To better implement the above methods, a comprehensive urban facility operation and maintenance management system based on digital twins is also proposed. The system includes a comprehensive operation record module, a related facility module, a feature parameter module, a dynamic time interval module, and a real-time analysis module. Integrated Operation Record Module: Collects cable operation parameters, constructs cable data sets, and uploads them to the urban facility management platform based on a digital twin model to form an integrated operation record of urban facilities; The integrated operation record module includes an integrated operation record establishment unit and an abnormal cable data group unit: Establish a comprehensive operation record unit: Install various types of operation monitoring sensors at the location of underground cables to monitor the cable's operation parameters, preset the data acquisition time interval, synchronize the acquired operation parameters in time, generate cable data sets with timestamps, and upload them to the urban facility management platform built based on a digital twin model. Collect the connection relationship between underground cables and various urban facilities, construct the cable topology, and based on the topology, collect the operation data corresponding to each urban facility, and upload the time-stamped operation data to the urban facility management platform to form a comprehensive operation record of urban facilities. Abnormal Cable Data Group Unit: Sets an initial threshold for cable operating parameters, compares the operating parameters in each collected cable data group with the initial threshold, marks operating parameters that exceed the initial threshold as abnormal operating parameters, and marks cable data groups with abnormal operating parameters as abnormal cable data groups.

[0024] Related facilities module: Filter records containing abnormal cable data groups in historical comprehensive operation records, extract abnormal operation data within the sliding time window, calculate the correlation index of urban facilities based on the abnormal operation data, and determine related facilities based on the correlation index; The associated facilities module includes an abnormal operation data analysis unit and an associated facilities identification unit: Abnormal Operation Data Analysis Unit: In the historical comprehensive operation record set, records containing abnormal cable data groups are filtered and marked as historical abnormal comprehensive operation records. The collection time point corresponding to the abnormal cable data group in each historical abnormal comprehensive operation record is collected. A sliding time window is preset to form a sliding time range. The topology of the cable corresponding to each historical abnormal comprehensive operation record is obtained, and the set of urban facilities connected to the cables is obtained. The operation data of each urban facility is collected within the sliding time range. If a certain operation data exceeds a preset threshold and the duration reaches a preset duration, the operation data is marked as abnormal operation data. The numerical change of the abnormal operation data of a certain historical abnormal comprehensive operation record within the sliding time range is collected, and the numerical change rate of the abnormal operation data is calculated. All historical abnormal comprehensive operation records are summarized, and the average numerical change rate of the abnormal operation data is calculated. Identify associated facility units: Count the number of times abnormal operation data of a certain city facility occurs, calculate the occurrence frequency of the city facility, normalize the average rate of change of the value, the proximity score and the occurrence frequency to calculate the correlation index, preset the correlation index threshold, compare the correlation index of each city facility with the correlation index threshold, if it exceeds the correlation index threshold, the corresponding city facility is marked as an associated facility.

[0025] Feature Parameter Module: Statistically analyzes the frequency of events occurring in the historical abnormal operation records of related facilities, extracts feature events, constructs a set of parameters for related facilities and a set of parameters for cables, takes the value of the feature event as the target variable, calculates the importance coefficient of the input variable through a random forest model, and selects feature parameters under different feature events; Dynamic time interval module: Collects feature parameters when feature events occur and when safety assessment results are abnormal during the training period, and establishes a dynamic threshold prediction model for feature parameters; collects feature parameters of associated facilities and cables within the feature time period, and constructs a dynamic time interval formula; Real-time analysis module: When a characteristic event occurs in the associated facility, it collects real-time characteristic parameters and inputs them into the characteristic parameter dynamic threshold prediction model to generate a dynamic threshold and calculate the real-time deviation rate; if the real-time characteristic parameter exceeds the dynamic threshold, an early warning is issued; otherwise, the dynamic time interval is calculated and data is collected according to the dynamic time interval. The real-time analysis module includes an anomaly detection unit and a dynamic data acquisition unit. Anomaly detection unit: When a characteristic event occurs in the associated facility, the characteristic parameters corresponding to the characteristic event are obtained, the real-time characteristic parameters of the associated facility and cable are collected, the number of the characteristic event, the normalized time sequence characteristics and environmental information are input into the characteristic parameter dynamic threshold prediction model, and the dynamic threshold of the characteristic parameters of the associated facility and cable is generated. If there is a real-time characteristic parameter that is greater than the characteristic parameter dynamic threshold, an early warning notice is issued to the staff to check. Dynamic acquisition unit: Calculates the real-time deviation rate of characteristic parameters and inputs it into the dynamic time interval formula to obtain the dynamic time interval, and acquires the characteristic parameters of associated facilities and cables according to the dynamic time interval.

[0026] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for integrated operation and maintenance management of urban facilities based on digital twins, characterized in that, The methods include: Step S100: Collect cable operation parameters, construct cable data sets and upload them to the urban facility management platform based on the digital twin model to form a comprehensive operation record of urban facilities; Step S200: Filter records containing abnormal cable data groups in the historical comprehensive operation records, extract abnormal operation data within the sliding time window, calculate the correlation index of urban facilities based on the abnormal operation data, and determine the related facilities based on the correlation index. Step S300: Statistically analyze the frequency of events occurring in the historical abnormal operation records of related facilities, extract feature events, construct a parameter set for related facilities and a cable parameter set, take the value of the feature event as the target variable, calculate the importance coefficient of the input variable through a random forest model, and filter to obtain the feature parameters under different feature events; Step S400: Collect feature parameters when feature events occur and when the safety assessment results are abnormal during the training period, and establish a dynamic threshold prediction model for feature parameters; collect feature parameters of related facilities and cables within the feature time period, and construct a dynamic time interval formula; Step S500: When a characteristic event occurs in the associated facility, real-time characteristic parameters are collected and input into the dynamic threshold prediction model of the characteristic parameters to generate a dynamic threshold and calculate the real-time deviation rate; if the real-time characteristic parameters exceed the dynamic threshold, an early warning is issued; otherwise, the dynamic time interval is calculated and data is collected according to the dynamic time interval.

2. The method for integrated operation and maintenance management of urban facilities based on digital twins according to claim 1, characterized in that, Step S100 includes the following steps: Step S101: Install various types of operation monitoring sensors at the location of the underground cable to monitor the cable's operation parameters, preset the data acquisition time interval, synchronize the acquired operation parameters in time, generate cable data sets with timestamps, and upload them to the urban facility management platform built on a digital twin model. Step S102: Collect the connection relationship between underground cables and various urban facilities, construct the cable topology, collect the operation data corresponding to each type of urban facility based on the topology, and upload the operation data with timestamps to the urban facility management platform to form a comprehensive operation record of urban facilities. Step S103: Set the initial threshold for cable operating parameters, compare the operating parameters in each collected cable data set with the initial threshold, mark the operating parameters that exceed the initial threshold as abnormal operating parameters, and mark the cable data set with abnormal operating parameters as abnormal cable data set.

3. The method for integrated operation and maintenance management of urban facilities based on digital twins according to claim 2, characterized in that, Step S200 includes the following steps: Step S201: In the historical integrated operation record set, filter the records containing abnormal cable data groups and mark them as historical abnormal integrated operation records. Collect the collection time point corresponding to the abnormal cable data group in each historical abnormal integrated operation record. The preset sliding time window is a, forming a sliding time range of [Aa, A+a], where A represents the collection time point. Step S202: Obtain the topology of the cable corresponding to each historical abnormal comprehensive operation record, obtain the set of urban facilities that are connected to the cable, collect the operation data of each urban facility within the sliding time range, if a certain operation data exceeds a preset threshold and the duration reaches the preset duration, then mark the operation data as abnormal operation data, collect the numerical change of the abnormal operation data of a certain historical abnormal comprehensive operation record within the sliding time range, calculate the numerical change rate of the abnormal operation data, summarize all historical abnormal comprehensive operation records, and calculate the average numerical change rate of the abnormal operation data; Step S203: Based on the cable topology, calculate the topological distance between the cable and urban facilities, and based on GIS coordinates, calculate the spatial distance between the cable and urban facilities. Normalize the topological distance and spatial distance, and perform a weighted sum of the normalized topological distance and spatial distance to calculate the proximity score. Step S204: Count the number of times abnormal operation data of a certain city facility occurs, calculate the frequency of occurrence of the city facility, normalize the average rate of change of the value, the proximity score and the frequency of occurrence, and calculate the correlation index according to the following formula: ； Where S represents the correlation index, s1 represents the normalized frequency of occurrence, and s2 represents the frequency of occurrence. b s1 represents the average rate of change of the normalized value of the b-th abnormal data point, s2 represents the normalized proximity score, and w1 and w2 represent the average rate of change of the normalized value. b w3 and w3 represent the weights of the frequency of occurrence, the b-th abnormal running data, and the proximity score, respectively. Step S205: Preset a correlation index threshold, compare the correlation index of each city facility with the correlation index threshold, and if it exceeds the correlation index threshold, mark the corresponding city facility as a related facility.

4. The method for integrated operation and maintenance management of urban facilities based on digital twins according to claim 3, characterized in that, Step S300 includes the following steps: Step S301: Obtain the events that occur in the historical abnormal operation records of each associated facility, count the frequency of occurrence of each event, preset the frequency threshold, and mark the events that exceed the frequency threshold as feature events; Step S302: In a certain historical abnormal comprehensive operation record, collect the operation data of related facilities and construct a set of related facility parameters, collect the operation parameters of the cable and construct a set of cable parameters. If a feature event exists in a certain time window in the historical abnormal comprehensive operation record, the value is 1; otherwise, the value is 0. Step S303: Using the set of associated facility parameters and the set of cable parameters as input variables, and the value of the feature event as the target variable, a sample set is constructed by selecting monitoring data within the time window range. The input variables are normalized and trained using a random forest model. The feature importance coefficient of each input variable in the model is calculated, and the model is filtered according to the preset importance coefficient threshold to determine the feature parameters of associated facilities and cables in different feature events.

5. A method for integrated operation and maintenance management of urban facilities based on digital twins according to claim 4, characterized in that, Step S400 includes the following steps: Step S401: Select several consecutive days as the training period. When a certain characteristic event occurs in a certain related facility, arrange staff to conduct a safety assessment of the related facility and cable. When the safety assessment result is abnormal, collect the characteristic parameters of the related facility and cable at the current time point. Step S402: Obtain the type, temporal features and environmental information of the current feature event, assign a number to the type, normalize the temporal features and environmental information, use the number, the normalized temporal features and environmental information as input, use the feature parameters of the associated facilities and cables at the current time point as input, train through the XGBoost regression model, and generate a dynamic threshold prediction model for feature parameters. Step S403: Collect the time points corresponding to the occurrence of the characteristic event and the abnormality of the safety assessment result. Take the time point when the characteristic event occurs as the starting point and the time point when the safety assessment result is abnormal as the ending point to obtain the characteristic time period. Obtain the characteristic parameters of the associated facilities and cables corresponding to each collection time point in the characteristic time period. Step S404: Calculate the dynamic threshold of the characteristic parameters of the associated facilities and cables corresponding to each collection time point using the characteristic parameter threshold prediction model, and calculate the difference between the threshold and the characteristic parameters of the associated facilities and cables collected at the collection time point to obtain the characteristic parameter deviation value of the associated facilities and cables. Step S405: Obtain the feature parameter deviation value between every two adjacent acquisition time points, calculate the feature parameter deviation rate, and establish the dynamic time interval formula as follows: ； Where T represents the characteristic time period, P ie Let Q be the deviation rate of the e-th feature parameter in the i-th time interval. ie Let t represent the weight of the deviation rate of the e-th feature parameter in the i-th time interval. i Let f represent the weight of the i-th time interval, f represent the total number of feature parameters, and r represent the total number of time intervals. The abnormal comprehensive operation records within the training period are summarized, and the dynamic time interval formula is trained by the linear regression method. The updated feature parameter deviation rate weight and time interval weight are applied to the dynamic time interval formula.

6. A method for integrated operation and maintenance management of urban facilities based on digital twins according to claim 5, characterized in that, Step S500 includes the following steps: Step S501: When a characteristic event occurs in the associated facility, obtain the characteristic parameters corresponding to the occurrence of the characteristic event, collect the real-time characteristic parameters of the associated facility and cable, input the characteristic event number, the normalized time sequence characteristics and environmental information into the characteristic parameter dynamic threshold prediction model, generate the characteristic parameter dynamic threshold of the associated facility and cable, if there is a real-time characteristic parameter greater than the characteristic parameter dynamic threshold, issue an early warning to the staff to check, if there is no such parameter, proceed to step S502; Step S502: Calculate the real-time deviation rate of the characteristic parameters and input it into the dynamic time interval formula to obtain the dynamic time interval. Then, collect the characteristic parameters of the associated facilities and cables according to the dynamic time interval.

7. A digital twin-based integrated operation and maintenance management system for urban facilities, used to implement the digital twin-based integrated operation and maintenance management system for urban facilities as described in any one of claims 1-6, characterized in that, The system includes a comprehensive operation record module, an associated facility module, a feature parameter module, a dynamic time interval module, and a real-time analysis module; The integrated operation record module collects cable operation parameters, constructs cable data sets, and uploads them to the urban facility management platform based on a digital twin model to form an integrated operation record of urban facilities. The associated facilities module: filters records containing abnormal cable data groups in historical comprehensive operation records, extracts abnormal operation data within a sliding time window, calculates the correlation index of urban facilities based on the abnormal operation data, and determines associated facilities based on the correlation index. The feature parameter module: statistically analyzes the frequency of events occurring in the historical abnormal comprehensive operation records of related facilities, extracts feature events, constructs a set of parameters for related facilities and a set of cable parameters, takes the value of the feature event as the target variable, calculates the importance coefficient of the input variable through a random forest model, and filters to obtain feature parameters under different feature events; The dynamic time interval module: collects feature parameters when feature events occur and when the safety assessment results are abnormal during the training period, and establishes a dynamic threshold prediction model for feature parameters; collects feature parameters of associated facilities and cables within the feature time period, and constructs a dynamic time interval formula. The real-time analysis module: when a characteristic event occurs in the associated facility, it collects real-time characteristic parameters and inputs them into the dynamic threshold prediction model of the characteristic parameters, generates a dynamic threshold and calculates the real-time deviation rate; if the real-time characteristic parameters exceed the dynamic threshold, it issues an early warning; otherwise, it calculates the dynamic time interval and collects data according to the dynamic time interval.

8. A comprehensive urban infrastructure operation and maintenance management system based on digital twins according to claim 7, characterized in that, The integrated operation record module includes an integrated operation record establishment unit and an abnormal cable data group unit: The establishment of the comprehensive operation record unit involves: installing multiple types of operation monitoring sensors at the location of the underground cable to monitor the cable's operation parameters; pre-setting the data acquisition time interval; synchronizing the acquired operation parameters in time; generating cable data sets with timestamps; and uploading them to the urban facility management platform built on a digital twin model; collecting the connection relationships between underground cables and various urban facilities; constructing the cable topology; and based on the topology, collecting the operation data corresponding to each type of urban facility and uploading the timestamped operation data to the urban facility management platform to form a comprehensive operation record of urban facilities. The abnormal cable data group unit: sets an initial threshold for cable operating parameters, compares the operating parameters in each collected cable data group with the initial threshold, marks operating parameters that exceed the initial threshold as abnormal operating parameters, and marks cable data groups with abnormal operating parameters as abnormal cable data groups.

9. A comprehensive urban infrastructure operation and maintenance management system based on digital twins according to claim 7, characterized in that, The associated facility module includes an abnormal operation data analysis unit and an associated facility identification unit: The abnormal operation data analysis unit: In the historical comprehensive operation record set, it filters records containing abnormal cable data groups and marks them as historical abnormal comprehensive operation records. It collects the collection time point corresponding to the abnormal cable data group in each historical abnormal comprehensive operation record, presets a sliding time window to form a sliding time range, obtains the topology of the cable corresponding to each historical abnormal comprehensive operation record, obtains the set of urban facilities that are connected to the cable, collects the operation data of each urban facility within the sliding time range, and if a certain operation data exceeds a preset threshold and lasts for a preset duration, it marks the operation data as abnormal operation data. It collects the numerical change of the abnormal operation data of a certain historical abnormal comprehensive operation record within the sliding time range, calculates the numerical change rate of the abnormal operation data, summarizes all historical abnormal comprehensive operation records, and calculates the average numerical change rate of the abnormal operation data. The process of determining associated facility units involves: counting the number of times abnormal operation data of a certain city facility occurs, calculating the occurrence frequency of the city facility, normalizing the average rate of change of the value, the proximity score, and the occurrence frequency to calculate the correlation index, setting a preset correlation index threshold, comparing the correlation index of each city facility with the correlation index threshold, and if it exceeds the correlation index threshold, then the corresponding city facility is marked as an associated facility.

10. A comprehensive urban infrastructure operation and maintenance management system based on digital twins according to claim 7, characterized in that, The real-time analysis module includes an anomaly detection unit and a dynamic acquisition unit: The anomaly detection unit: when a characteristic event occurs in the associated facility, it acquires the characteristic parameters corresponding to the characteristic event, collects the real-time characteristic parameters of the associated facility and cable, inputs the characteristic event number, the normalized time sequence characteristics and environmental information into the characteristic parameter dynamic threshold prediction model, generates the characteristic parameter dynamic threshold of the associated facility and cable, and if there is a real-time characteristic parameter greater than the characteristic parameter dynamic threshold, it issues an early warning to notify the staff to check. The dynamic acquisition unit calculates the real-time deviation rate of the characteristic parameters and inputs it into the dynamic time interval formula to obtain the dynamic time interval. Then, it acquires the characteristic parameters of the associated facilities and cables according to the dynamic time interval.

Images