Urban well lid abnormity real-time monitoring and alarming method based on Internet of Things

Abstract

The invention provides an urban well lid abnormity real-time monitoring and alarming method based on the Internet of Things, and the method comprises the steps: obtaining a real-time displacement parameter and a vibration parameter of a monitoring sensor, triggering an abnormity detection signal if the parameters exceed a preset normal range threshold value, and generating a preliminary abnormity judgment result; performing similarity matching on the abnormal feature parameters and a historical case library, and if the matching degree exceeds a second preset threshold value, determining an abnormal event state; according to the abnormal event state, adjusting the acquisition frequency of a non-abnormal area to a low-frequency mode, and generating a global energy consumption optimization scheme; updating a time rule parameter of the regional attribute database, recording a current frequency configuration state, and generating a dynamic frequency management scheme; and if the abnormal confirmation frequency of the monitoring point continuously exceeds a third preset threshold value, recalculating the initial frequency reference value and the time dimension adjustment coefficient, and generating a periodic evaluation result.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a real-time monitoring and alarm method for abnormalities of urban manhole covers based on the Internet of Things. Background Art

[0002] Problem background:

[0003] Abnormal monitoring of urban manhole covers is an integral part of smart city development, crucial for ensuring public safety and the stability of urban operations. Displacement, loss, or damage to manhole covers can lead to traffic accidents or pedestrian injuries, making real-time monitoring and prompt alerting a pressing need in urban management. Monitoring systems built using IoT technology can provide strong technical support for urban safety and hold irreplaceable value.

[0004] However, many current manhole cover monitoring solutions suffer from significant shortcomings in practical application. Many systems rely on fixed-frequency data collection, unable to flexibly adjust to the actual conditions of different regions or time periods, resulting in wasted resources and poor monitoring results. In particular, the lack of effective optimization methods for data processing and energy management leads to high operating costs and makes it difficult to meet the diverse needs of complex urban environments.

[0005] Against this backdrop, this field faces prominent technical challenges. The first and foremost challenge is how to reasonably determine the frequency of data collection in different scenarios to balance the timeliness of monitoring and the energy consumption of the system. If the collection frequency is too high, it will not only increase the energy consumption of the equipment and the burden of data transmission, but may also result in a large amount of redundant data, affecting system efficiency; while if the frequency is too low, key abnormal events may be missed, reducing the reliability of monitoring. This problem further extends to how to dynamically adjust the collection strategy based on historical data and real-time feedback, so that the system can adapt to low-frequency monitoring under normal conditions, and can quickly switch to high-frequency mode when an abnormality is suspected to occur, ensuring the accuracy of anomaly confirmation. This dynamic switching requirement from normal to abnormal conditions makes system design encounter unique difficulties in the intelligent control of data collection intervals.

[0006] Therefore, how to design a dynamic data collection interval adjustment mechanism based on regional characteristics, time patterns and abnormal signals for urban manhole cover monitoring scenarios has become a key issue that needs to be solved urgently. Summary of the Invention

[0007] The present invention provides a real-time monitoring and alarm method for abnormalities of urban manhole covers based on the Internet of Things, which mainly includes:

[0008] Obtain a pre-established regional attribute database, extract traffic flow parameters and historical abnormal frequency parameters corresponding to the monitoring points, generate regional differentiated initial acquisition frequency reference values based on the parameters, and form an initial frequency configuration table;

[0009] According to the initial frequency configuration table, combined with historical time regularity data, the abnormal probability distribution parameters of each time period are extracted. If the abnormal probability of the peak period exceeds the first preset threshold, the acquisition frequency of the period is increased to the first high frequency mode, and the time dimension frequency adjustment coefficient is determined;

[0010] Acquire real-time displacement parameters and vibration parameters of the monitoring sensor. If the parameters exceed the preset normal range threshold, trigger an abnormality detection signal and generate a preliminary abnormality judgment result;

[0011] According to the preliminary abnormality judgment result, the dynamic acquisition strategy module is activated, the acquisition mode of the corresponding monitoring point is switched to the second high-frequency mode, high-density displacement vibration data is acquired, and abnormal characteristic parameters are extracted;

[0012] Performing similarity matching between the abnormal feature parameters and the historical case library, and confirming the abnormal event status if the matching degree exceeds a second preset threshold;

[0013] According to the abnormal event status, the acquisition frequency of the non-abnormal area is adjusted to a low-frequency mode to generate a global energy consumption optimization plan;

[0014] Updating the time regularity parameters of the regional attribute database, recording the current frequency configuration status, and generating a dynamic frequency management plan;

[0015] If the abnormality confirmation frequency of the monitoring point continues to exceed the third preset threshold, the initial frequency reference value and the time dimension adjustment coefficient are recalculated to generate a periodic evaluation result.

[0016] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0017] The present invention discloses a real-time monitoring and alarm method for urban manhole cover anomalies based on the Internet of Things. By establishing a regional attribute database, an initial acquisition frequency configuration table is generated in combination with traffic flow and historical abnormal frequency parameters, and the acquisition frequency is adjusted according to time regularity data. When an anomaly is detected, the present invention activates a dynamic acquisition strategy, increases the acquisition frequency to obtain high-density data, and performs abnormal feature matching and confirmation. At the same time, the present invention can also adjust the acquisition frequency of non-abnormal areas according to the abnormal state to achieve global energy consumption optimization. In addition, the present invention also has periodic evaluation and adaptive adjustment capabilities, and can recalculate frequency parameters according to the abnormal confirmation frequency. This method realizes the intelligent management of acquisition frequency, optimizes system energy consumption while ensuring the abnormality detection effect, and improves monitoring efficiency and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 The present invention is a flow chart of a method for real-time monitoring and alarming of abnormalities of urban manhole covers based on the Internet of Things.

[0019] Figure 2 The figure is a schematic diagram of a method for real-time monitoring and alarming of abnormalities of urban manhole covers based on the Internet of Things of the present invention.

[0020] Figure 3 This is another schematic diagram of a method for real-time monitoring and alarming of abnormalities of urban manhole covers based on the Internet of Things of the present invention. DETAILED DESCRIPTION

[0021] To further understand the content of the present invention, the present invention is described in detail with reference to the accompanying drawings and examples. The present application is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are intended only to illustrate the relevant invention and are not intended to limit the invention. It should also be noted that, for ease of description, only portions relevant to the invention are shown in the accompanying drawings.

[0022] like Figure 1-3 In this embodiment, a method for real-time monitoring and alarming of abnormalities of urban manhole covers based on the Internet of Things may specifically include:

[0023] Step S101: obtain a pre-established regional attribute database, extract traffic flow parameters and historical abnormal frequency parameters corresponding to monitoring points, generate regional differentiated initial acquisition frequency reference values based on the parameters, and form an initial frequency configuration table.

[0024] Relevant data records for monitoring points are obtained from a pre-established regional attribute database. These data records contain traffic flow information and historical anomaly records and are organized into a structured dataset. Within this structured dataset, statistical analysis methods are used to extract traffic flow parameters and historical anomaly frequency parameters for each monitoring point, generating a parameter set. Based on this parameter set, a linear regression model is used to perform a weighted calculation on the traffic flow parameters and historical anomaly frequency parameters to generate differentiated initial acquisition frequency benchmark values, thereby determining a frequency benchmark dataset. If the benchmark value for a monitoring point in the frequency benchmark dataset exceeds a preset threshold, a secondary check is performed on the traffic flow parameters and historical anomaly frequency parameters for that point. The benchmark value is then adjusted based on the regional attribute information to generate an adjusted frequency benchmark dataset. Using this adjusted frequency benchmark dataset, the initial acquisition frequency benchmark values are mapped to the corresponding points based on the regional attributes and monitoring point information, forming an initial frequency configuration table. The frequency values in the initial frequency configuration table are then compared with the historical anomaly frequency parameters. If significant deviations are found, the frequency values are fine-tuned based on the difference calculation results to generate a final frequency configuration table. According to the final frequency configuration table, associated data acquisition and parameter extraction logic are used to format the frequency values in the table and output standardized configuration data records.

[0025] For example, consider a city traffic monitoring system that acquires monitoring point data from a pre-established regional attribute database. This database stores traffic flow information and historical anomaly records for multiple intersections. Suppose that intersection A has an average daily traffic volume of 5,000 vehicles, and historical anomaly records indicate three congestion events per month. By organizing this data into a structured dataset, it facilitates subsequent analysis. This step ensures data standardization and lays the foundation for subsequent statistical analysis.

[0026] In one possible implementation, when performing statistical analysis on a structured data set, traffic flow parameters for intersection A can be extracted, such as an average daily flow of 5,000 vehicles, and historical anomaly frequency parameters, such as a monthly anomaly frequency of 0.1. By performing similar analyses on multiple points, a parameter set is formed. The key to this step is to quantify the characteristics of the points, providing data support for subsequent modeling.

[0027] Specifically, when using a linear regression model to weight parameters, traffic flow parameters can be given a higher weight, as traffic flow directly influences the required frequency of data collection. Abnormal frequency, as an auxiliary parameter, is weighted slightly lower. Assume that the initial baseline frequency for intersection A is calculated to be 2 collections per hour and included in the frequency baseline dataset. If the preset threshold range is 1 to 5 times per hour, the baseline value for intersection A meets the requirements and does not require adjustment. This step ensures the rationality of the frequency baseline value and avoids wasted resources or insufficient data collection.

[0028] For example, when performing secondary verification on points exceeding the threshold, suppose the baseline value for Intersection B is calculated to be 6 times per hour, exceeding the upper threshold. Taking into account regional attributes (e.g., Intersection B is located in a commercial area with large traffic fluctuations and a high historical anomaly frequency), the baseline value is adjusted to 5 times per hour, forming an adjusted frequency baseline dataset. This not only optimizes resource allocation but also improves the targeted nature of monitoring.

[0029] In one possible implementation, when the adjusted frequency reference value is mapped to a point, an initial frequency configuration table can be generated, with intersection A set at 2 times per hour and intersection B at 5 times per hour. The historical abnormal frequency parameters are then compared. If the historical abnormal frequency of intersection B is higher, while the configured frequency is lower, the frequency configuration table is fine-tuned to 5.5 times per hour, forming the final frequency configuration table. This step ensures that the frequency configuration matches actual needs through dynamic adjustment.

[0030] Specifically, when the final frequency configuration table outputs standardized configuration data records, the frequency values can be formatted into uniform units, such as the number of collections per hour, and associated with data acquisition logic to ensure that the system can directly apply the configuration. This not only improves system compatibility but also reduces subsequent maintenance costs. Through the above process, the collection frequency of monitoring points is optimized, meeting real-time monitoring needs while avoiding excessive resource consumption, achieving a balance between efficiency and effectiveness.

[0031] Step S102: According to the initial frequency configuration table and combined with historical time regularity data, the abnormal probability distribution parameters of each time period are extracted. If the abnormal probability of the peak period exceeds the first preset threshold, the acquisition frequency of the period is increased to the first high frequency mode, and the time dimension frequency adjustment coefficient is determined.

[0032] Initial collection frequency and historical data are obtained, and changes in anomaly probability across time dimensions are analyzed to obtain anomaly probability distribution parameters for each time period. Based on these anomaly probability distribution parameters, data records from peak hours are evaluated. If the anomaly probability exceeds a preset probability threshold, the collection frequency adjustment logic is triggered, determining the conditions for enabling high-frequency mode. Using this high-frequency mode, the collection frequency during peak hours is dynamically adjusted. The adjustment coefficient for the time dimension is calculated based on the time period information to obtain adjusted frequency configuration data. Based on this adjusted frequency configuration data, historical anomaly probability patterns are incorporated to analyze the changing trends of the distribution parameters for each time period to determine whether there are persistent fluctuations in anomaly probability. If the fluctuations exceed a preset range, the collection frequency is recalibrated based on the time dimension and the adjustment coefficient to obtain an optimized frequency configuration record. Using this optimized frequency configuration record, the adjustment coefficient is validated using a logistic regression model against the time period and peak hour anomaly probability data to determine the final frequency configuration. Based on this final frequency configuration, corresponding execution parameters are generated for the collection frequency in each time dimension to obtain standardized frequency adjustment data.

[0033] For example, in urban traffic monitoring, when obtaining initial data collection frequency and historical patterns, one could imagine a road network monitoring system that compiles traffic flow and historical anomaly data for different road sections. Suppose the initial collection frequency for a main road is three times per hour on weekdays, and historical patterns indicate a higher probability of anomalies during peak hours in the morning and evening. By analyzing this data, a basic dataset with a time dimension is formed, supporting subsequent anomaly probability analysis.

[0034] For example, when analyzing changes in anomaly probability across time dimensions, a day can be divided into multiple periods, such as the morning rush hour (7:00-9:00), the off-peak (9:00-17:00), and the evening rush hour (17:00-19:00). Assuming the anomaly probability during the morning rush hour is 0.3, exceeding the preset threshold of 0.2, the collection frequency adjustment logic is triggered, determining the specific conditions for activating high-frequency mode, such as increasing the frequency to 5 times per hour. This adjustment logic can better capture traffic changes during peak hours.

[0035] For example, when dynamically adjusting the peak-hour collection frequency using high-frequency mode, the time-based adjustment coefficient can be calculated based on the time period division information. Assuming the adjustment coefficient for the morning peak is 1.5, based on the initial frequency of 3 times per hour, the adjusted frequency is 4.5 times per hour. This approach ensures that frequency adjustments match actual traffic demand.

[0036] For example, when integrating historical analysis of anomaly probability patterns into the configuration data after frequency adjustment, we can observe the changing trends of distribution parameters across time periods. For example, if the anomaly probability during the evening peak period increases for three consecutive days, with fluctuations exceeding the preset range of 0.1-0.3, a secondary calibration combining the time dimension and the adjustment coefficient is performed, adjusting the frequency from 4 to 5 times per hour, thus generating the optimized frequency configuration record. This calibration mechanism enables timely response to abnormal fluctuations.

[0037] For example, when correlating time period divisions and abnormal probability data with optimized frequency configuration records, a logistic regression model can be used to verify the rationality of the adjustment coefficient. Assuming the verification results indicate that a morning peak adjustment coefficient of 1.5 is reasonable, the final frequency configuration plan is determined to be 4.5 times per hour. This verification method ensures the scientific validity of the configuration plan.

[0038] For example, when generating execution parameters for each time dimension based on the final frequency configuration plan, the frequency during the morning peak can be standardized to 4.5 times per hour, while the frequency during off-peak hours can be maintained at 3 times per hour, generating standardized frequency adjustment data. This standardized processing facilitates direct system application of the configuration and improves execution efficiency. Through this approach, dynamic adjustment of the collection frequency can accurately respond to traffic anomalies, ensuring timely monitoring and rational resource utilization.

[0039] Step S103: acquiring real-time displacement parameters and vibration parameters of the monitoring sensor. If the parameters exceed a preset normal range threshold, an abnormality detection signal is triggered to generate a preliminary abnormality judgment result.

[0040] The displacement and vibration parameters of the monitoring sensor are acquired and recorded in real time using a data acquisition module to determine an original parameter dataset. Within the original parameter dataset, the displacement and vibration parameters are range-determined using preset thresholds. If the parameters exceed the normal range, an anomaly detection signal is generated, generating a preliminary anomaly indicator. Based on the preliminary anomaly indicator, the trigger time and duration of the anomaly detection signal are determined. Signal features are extracted using a time series analysis tool to determine the time window in which the anomaly occurred. Within the time window in which the anomaly occurred, monitoring sensor data records are acquired for the relevant period. The data comparison module analyzes the changing trends of the displacement and vibration parameters to determine whether persistent anomaly characteristics exist. If persistent anomaly characteristics are confirmed, the parameter fluctuation amplitude corresponding to the trend is acquired. The fluctuation amplitude is compared with the preset threshold to determine the severity of the anomaly. Based on the severity of the anomaly, a logistic regression model is used to validate the anomaly detection signal. The model output is used to calibrate the preliminary anomaly indicator to obtain final anomaly determination data. For the final anomaly determination data, the corresponding monitoring sensor identifier is obtained. The anomaly records are archived using a data storage module to provide basic information for subsequent data queries.

[0041] For example, in the field of urban traffic monitoring, it is particularly important to obtain the displacement parameters and vibration parameters of the monitoring sensors for the safety monitoring of bridge structures at key nodes of the road network. Assuming that a bridge is equipped with multiple sets of sensors, the displacement parameters are used to detect the deformation of the bridge structure, while the vibration parameters reflect the dynamic response caused by vehicle traffic or wind loads. Through the data acquisition module, the system records data once a minute to form an original parameter data set. If the normal range of the displacement parameter is set to -5 mm to 5 mm and the vibration frequency range is 0.5 Hz to 2 Hz, if a certain record shows that the displacement reaches 6 mm, it exceeds the preset threshold, and the system generates an abnormality detection signal and marks it as a preliminary abnormality.

[0042] For example, for preliminary anomaly identification, the system will extract the trigger time and duration of the anomaly detection signal. Assuming that the anomaly signal is triggered at 8 a.m. and lasts for 10 minutes, the time series analysis tool is used to determine that the anomaly time window is from 8 a.m. to 8:10 a.m. The determination of this time window helps to focus the analysis scope and avoid data redundancy. Next, the system extracts all sensor data within this period and analyzes the changing trends of displacement and vibration parameters. If the displacement parameter gradually increases from 6 mm to 7 mm within 10 minutes, and the vibration frequency also increases from 1.5 Hz to 2.2 Hz, showing a persistent abnormal characteristic, the existence of the anomaly is further confirmed.

[0043] For example, when determining the severity of an anomaly, the system compares the fluctuation amplitude with a preset threshold. Assuming a displacement fluctuation of 2 mm, exceeding the normal range by 40%, and a vibration frequency fluctuation of 0.7 Hz, exceeding the normal range by 35%, the system classifies the anomaly severity as medium. Subsequently, a logistic regression model is used to verify the anomaly signal. By inputting historical data on displacement and vibration parameters and current fluctuation characteristics, the model outputs an anomaly probability of 0.85, which is higher than the preset threshold of 0.7. This calibrates the preliminary anomaly flag and produces the final anomaly determination data. This verification process ensures the accuracy of the determination.

[0044] For example, for the final abnormality determination data, the system records the corresponding sensor ID, such as sensor number S001, and archives the abnormality record through the data storage module, including information such as the abnormality time, parameter value, and severity. This archived data provides the foundation for subsequent query and analysis, facilitating the tracing of the cause and consequences of the abnormality, and also provides a data basis for bridge maintenance decisions. In this way, the system can promptly identify potential risks to the bridge structure and safeguard traffic safety.

[0045] Step S104 , based on the preliminary abnormality judgment result, activate the dynamic acquisition strategy module, switch the acquisition mode of the corresponding monitoring point to the second high-frequency mode, obtain high-density displacement vibration data, and extract abnormal characteristic parameters.

[0046] By starting the dynamic acquisition strategy module, the acquisition mode of the monitoring node is adjusted to the high-frequency mode based on the abnormal judgment results to obtain displacement vibration data. Based on the acquired displacement vibration data, a data preprocessing method is used to remove noise interference and obtain a cleaned vibration data stream. If there are fluctuations exceeding the preset threshold in the cleaned vibration data stream, the abnormal feature parameters are determined through feature extraction methods. Based on the determined abnormal feature parameters, the support vector machine algorithm is used to classify the abnormal features and determine the type of abnormality. If the determined abnormality type belongs to the high-risk category, the data density enhancement mechanism is triggered to obtain higher-density data samples. Through in-depth analysis of higher-density data samples, potential abnormal patterns are extracted and the changing trend of abnormal development is determined. Based on the changing trend of abnormal development, the operating parameters of the dynamic acquisition strategy module are adjusted to optimize the frequency and range of subsequent data acquisition.

[0047] For example, in the scenario of collecting displacement and vibration data from monitoring sensors, it is particularly important to activate the dynamic collection strategy module based on abnormal judgment results. The core of the dynamic collection strategy module is to adjust the collection mode of the monitoring node according to the abnormal situation, such as switching from the conventional low-frequency mode to the high-frequency mode to capture more detailed data changes. Assume that in a bridge monitoring project, the regular collection frequency is once per minute. When an abnormal vibration signal is detected, the system automatically adjusts to a high-frequency collection of once per second to obtain more continuous displacement and vibration data. This adjustment can help capture high-frequency fluctuations in a short period of time and ensure data integrity.

[0048] For example, for the acquired displacement vibration data, data preprocessing methods are used to remove noise interference.

[0049] In one possible implementation, the system might use sliding average filtering. Assuming the raw vibration data contains random noise due to ambient wind, averaging five consecutive data points effectively smooths the data curve, resulting in a cleaned vibration data stream. This approach can reduce the possibility of misjudgment and improve the accuracy of subsequent analysis.

[0050] For example, if fluctuations exceeding a preset threshold are found in the cleaned vibration data stream, such as a vibration amplitude exceeding the normal range of 0.5 mm, the system will determine the abnormal feature parameters through feature extraction methods.

[0051] Specifically, key indicators such as fluctuation frequency and peak amplitude can be extracted for subsequent classification. This extraction method helps to accurately locate the specific manifestations of anomalies.

[0052] For example, during the abnormality feature classification stage, support vector machine algorithms can be used to determine the type of abnormality. If the extracted features show that the vibration frequency abnormality is concentrated in a certain high-frequency band, the system might classify it as a high-risk condition caused by equipment fatigue. This classification method provides important evidence for subsequent decision-making.

[0053] For example, when an anomaly is identified as high-risk, triggering the data density enhancement mechanism becomes particularly necessary. Assuming data is collected once per second in the original high-frequency mode, the enhancement mechanism could increase this to once every 0.2 seconds, generating a higher density of data samples. This approach can more comprehensively reflect the dynamic changes in anomalies.

[0054] For example, by deeply analyzing higher-density data samples, the system can extract potential abnormal patterns, such as discovering that vibration amplitude shows a periodic increase within a specific time period. This helps determine whether the abnormality has the risk of continued deterioration and provides data support for subsequent maintenance.

[0055] For example, adjusting the operating parameters of the dynamic collection strategy module based on the changing trends of anomalies is crucial. If analysis reveals that anomalies are more pronounced at night, the system might optimize the collection frequency, maintaining a high frequency at night and reverting to a low frequency during the day to balance resource consumption and monitoring effectiveness. This optimization approach can significantly improve the relevance and efficiency of data collection.

[0056] Step S105 , performing similarity matching between the abnormal feature parameters and the historical case library, and if the matching degree exceeds a second preset threshold, confirming the abnormal event status.

[0057] By extracting abnormal features from the input data and performing preliminary data processing using feature decomposition technology, an abnormal feature set is obtained. Based on the abnormal feature set, relevant records in the historical case library are obtained, and the similarity value between the two is calculated using the cosine similarity algorithm. If the similarity value is higher than the preset threshold, the current data is determined to be in an abnormal determination state, and the abnormal event status identifier is output. Based on the abnormal event status identifier, several historical cases with the highest degree of matching are screened from the case library, and the corresponding solution records are obtained. By integrating and analyzing the screened solution records, the processing priority and related strategies of the current abnormal features are determined. Based on the processing priority and related strategies, a targeted repair plan is generated, and the final event status confirmation result is output. After obtaining the event status confirmation result, the historical case library is updated to incorporate the current abnormal features and processing strategies, forming a closed-loop data supplement.

[0058] For example, when monitoring the operating status of equipment, extracting abnormal features from input data can be achieved through time-frequency analysis of the equipment's vibration signals. During initial processing, feature decomposition techniques can be used to break down complex signals into multiple frequency band features, thereby forming a set of abnormal features. For example, suppose a device detects an abnormally high vibration frequency during operation. After decomposition, it is found to be primarily concentrated in a specific frequency band, which constitutes the core content of the abnormal feature set.

[0059] For example, based on the aforementioned abnormal feature set, relevant records in the historical case library can be retrieved by comparing vibration frequency band data from similar devices. Using the cosine similarity algorithm, the degree of match between the current feature and the historical record is calculated. For example, if the similarity between the current feature vector and a historical case is 0.85, which is higher than the preset threshold of 0.8, an abnormality is determined and an abnormal event status indicator is output. This approach can quickly identify potential problems.

[0060] For example, when screening historical cases and obtaining solution records for equipment experiencing abnormal vibration frequency, the three most closely matching cases can be extracted from the case library. These cases correspond to different solutions, such as adjusting the equipment load, checking for component wear, or replacing lubricants. By integrating and analyzing these solutions, the priority for addressing the current abnormality, such as prioritizing component wear, can be determined, and relevant strategies can be developed. This approach helps quickly focus on the core issue.

[0061] For example, when generating targeted repair plans, based on priorities and policies, a recommendation might be made to first inspect equipment components. If wear exceeds a certain level, replacement of the components is arranged, and the results of the event status confirmation are recorded. This approach ensures targeted processing and avoids wasted resources.

[0062] For example, when updating a historical case library, information about current anomaly characteristics, such as vibration frequency band data, and handling strategies, such as component replacement, is added to the library, forming a closed-loop data supplement. The next time a similar anomaly occurs, the system can directly call upon the updated case data, shortening response time. This closed-loop mechanism continuously optimizes the accuracy of subsequent judgments.

[0063] For example, prioritizing and formulating strategies for handling abnormal characteristics are particularly critical throughout the entire process. For equipment operation scenarios, priorities may be determined based on the impact of the abnormality on overall equipment operation. For example, high-frequency vibration, which may cause equipment downtime, has the highest priority. This classification helps to rationally allocate resources and ensure that critical issues are addressed first.

[0064] For example, from a business perspective, the extraction of abnormal features and case library updates during equipment monitoring are caused by potential risks that may arise during equipment operation. If not promptly addressed, the consequences could lead to equipment failure or even production interruption. This approach effectively identifies risks and provides solutions, ensuring stable equipment operation. This approach has high practical value in real-world applications.

[0065] Step S106 , adjusting the acquisition frequency of the non-abnormal area to a low-frequency mode according to the abnormal event state, and generating a global energy consumption optimization solution.

[0066] The monitoring system obtains real-time data from each area, analyzes whether there are any abnormal events, and obtains preliminary event status classification results. Based on the preliminary event status classification results, non-abnormal areas are identified and the scope of the area where the collection frequency needs to be adjusted is determined. Using the preset threshold standard, if the data flow of the identified non-abnormal area is lower than the preset threshold, its collection frequency is adjusted to low-frequency mode to obtain the adjusted frequency configuration. Based on the adjusted frequency configuration, the energy consumption change trend of each area is calculated to determine whether the global energy consumption has reached the optimization target. If the global energy consumption does not reach the optimization target, the frequency control parameters of the non-abnormal area are readjusted based on the energy consumption change trend to obtain a new frequency configuration plan. Based on the new frequency configuration plan, the collection frequency settings of each area are updated to obtain the updated energy consumption data distribution. Based on the updated energy consumption data distribution, the global energy consumption is predicted and analyzed in combination with the support vector machine algorithm to determine the final optimization strategy.

[0067] For example, consider a scenario involving monitoring a city's water supply network, where a monitoring system acquires real-time data and analyzes abnormal events. The system collects data from water pressure and flow sensors in various areas to preliminarily determine whether there are pipeline leaks or abnormal water usage. During this process, the system compares the data with historical normal ranges, classifying areas with potential problems and normal areas. For example, if water pressure data in a particular area consistently falls 10% below normal, it is preliminarily classified as a potential abnormality area, while other areas are identified as normal. This classification helps ensure the rationality of subsequent resource allocation.

[0068] For example, the identification of non-anomalous areas and the adjustment of data collection frequency can be implemented based on a specific standard. For example, if the preset threshold is 100 units of data flow per hour and the flow rate in a non-anomalous area is only 50 units, the system will adjust the data collection frequency from once per minute to once every five minutes to reduce unnecessary energy consumption. After the adjustment, the system will record the new frequency configuration and observe the impact of this change on overall energy consumption. This approach can effectively reduce the operating burden of equipment while ensuring basic data collection needs.

[0069] For example, when calculating energy consumption trends in each region and determining the global energy consumption optimization target, analysis can be performed by comparing energy consumption data before and after adjustments. Assuming that after adjusting the frequency, energy consumption in non-abnormal areas drops from 1,000 units per day to 800 units, the system will further evaluate whether global energy consumption has reached the preset optimization target, such as total daily energy consumption below 5,000 units. If the target is not reached, the system will further adjust the frequency parameters to a lower frequency mode based on trend analysis, such as discovering that the energy consumption of certain areas has decreased less significantly. This dynamic adjustment helps gradually approach the optimization target.

[0070] For example, the effectiveness of updating frequency configurations and collecting energy consumption data distribution can be verified through real-time monitoring. Suppose a new frequency configuration reduces the collection frequency for a certain area to every 10 minutes. The system records the updated energy consumption data and finds that energy consumption in that area has further decreased to 600 units per day. This distribution data provides a basis for subsequent forecasting and also reflects the direct impact of the frequency adjustment.

[0071] For example, when using a support vector machine algorithm for global energy consumption forecasting and analysis, it is possible to infer energy consumption trends over the next 24 hours based on historical energy consumption data and current distribution data. For example, the system analysis indicates that if the current frequency configuration is maintained, total energy consumption the following day may be 4,800 units, close to the optimization target. Based on this, the system determines the final strategy, such as maintaining the low-frequency mode or fine-tuning the frequency in certain areas. This predictive analysis provides data support for decision-making and ensures maximum resource utilization. The various implementation examples above demonstrate the complete logical chain from data collection and frequency adjustment to energy consumption optimization. The detailed operations at each link are closely centered around the core area of water supply network monitoring, ensuring the targeted and practical nature of the strategy. At the same time, through dynamic adjustments and predictive analysis, the system's operational efficiency and resource conservation capabilities are significantly improved.

[0072] Step S107: updating the time regularity parameters of the regional attribute database, recording the current frequency configuration status, and generating a dynamic frequency management solution.

[0073] Historical data from the regional attribute database is retrieved and features extracted based on temporal patterns to determine the temporal distribution pattern. Based on this temporal distribution pattern, the current frequency configuration is analyzed to determine whether the configuration parameters meet the preset threshold range. If they fall below the threshold, a parameter update process is triggered to determine the updated parameter values. Based on the updated parameter values and the dynamic frequency adjustment strategy, a preliminary frequency adjustment plan is generated and the key indicator values for the adjustment plan are obtained. The data records for key indicator values are analyzed for completeness. If any records are missing or abnormal, a pre-established support vector machine model is used to complete the data and obtain a complete indicator dataset. Based on the complete indicator dataset and the business objectives of the management plan, an optimized dynamic frequency configuration is generated to determine whether the optimized configuration meets the requirements of the regional attributes. If the optimized configuration meets the requirements of the regional attributes, the optimized configuration is written to the database, the current frequency configuration status is updated, and the latest configuration status information is obtained. Based on this latest configuration status information, a final dynamic frequency management plan is generated and the implementation time and scope of the plan are determined.

[0074] For example, when acquiring historical data from a regional attribute database, one can first focus on data records from the past year and extract time-related features, such as daily traffic changes during peak and off-peak hours. For example, suppose a region experiences a surge in data traffic between 8:00 AM and 10:00 AM on weekdays, while data activity is almost nonexistent between midnight and 5:00 PM. This temporal distribution pattern forms the basis for subsequent analysis. This pattern can be used to initially determine whether the current frequency configuration is appropriate. For example, if high-frequency data collection is maintained at night, it is clear that the configuration parameters are not meeting energy-saving targets and are outside the preset threshold range.

[0075] For example, when triggering a parameter update, the nighttime data collection frequency can be adjusted from once per minute to once per hour based on historical data analysis. The updated parameter values must ensure basic data collection needs while reducing unnecessary resource consumption. Incorporating a dynamic frequency adjustment strategy, a preliminary plan might include maintaining a high frequency during weekday peak hours, reducing to a medium frequency during off-peak hours, and maintaining a low frequency throughout the day on weekends. Key metrics for this approach include frequency switching response time and data collection coverage.

[0076] For example, when analyzing data record integrity, if missing data is found for a particular night, a support vector machine model can be used to fill in the missing data using historical data patterns from similar time periods. For example, based on the average value for the same time period over the past week, it can be estimated that the missing data should be 10 requests per hour. This complete indicator dataset is then formed after the missing data is filled in. This approach ensures the accuracy of subsequent analysis and provides a reliable basis for configuration optimization.

[0077] For example, when generating a dynamic frequency optimization configuration, consider business objectives, such as ensuring real-time data availability and balancing energy consumption. If a region's attributes require focused monitoring of peak hours, the optimization configuration will prioritize high-frequency data collection during these hours, while appropriately reducing the frequency during other periods. For example, assuming a peak frequency of 5 times per minute and a non-peak frequency of 1 time per hour, this configuration meets demand while effectively controlling resource usage.

[0078] For example, after writing the optimized configuration to the database, the current frequency configuration status needs to be updated and the latest status information obtained. Suppose, after the update, the average daily frequency in a certain area drops from 3 times per minute to 1.5 times per minute. The status information shows a reduction of approximately 30% in energy consumption, providing data support for subsequent management plans. Ultimately, the dynamic frequency management plan can be set to update every Monday at dawn, covering all non-abnormal areas, ensuring the timeliness and applicability of configuration adjustments. This approach can significantly improve resource utilization efficiency and extend equipment life.

[0079] For example, when implementing a plan, the time range can be further refined based on different regional attributes. For example, frequency adjustment strategies can be customized separately for commercial and residential areas, with commercial areas focusing more on daytime and residential areas focusing more on nighttime. This differentiated management can better meet actual needs and improve the adaptability of the overall plan.

[0080] Step S108: If the abnormality confirmation frequency of the monitoring point continues to exceed the third preset threshold, the initial frequency reference value and the time dimension adjustment coefficient are recalculated to generate a periodic evaluation result.

[0081] By collecting real-time data from monitoring points, the system captures dynamic changes in abnormal and confirmed frequencies to determine whether the current monitoring status exceeds a preset threshold. If the confirmed frequency consistently exceeds the threshold, the initial benchmark recalculation process is triggered. Frequency benchmark parameters are extracted from historical data to obtain an updated initial frequency benchmark value. Based on this updated initial frequency benchmark value and the time-dependent variation characteristics, the corresponding adjustment coefficient is calculated to determine whether the adjustment coefficient meets the requirements for periodic evaluation. If the adjustment coefficient meets the requirements, the system optimizes the frequency benchmark and adjustment coefficient using a pre-established regression model to obtain the optimized time-dependent adjustment parameters. Using the optimized time-dependent adjustment parameters, the confirmed abnormal frequency data is calibrated to determine whether the calibrated abnormal frequency still exceeds the preset threshold. If the calibrated abnormal frequency still exceeds the preset threshold, a periodic evaluation result is generated. The monitoring strategy for the monitoring point is updated based on the abnormal point information in the evaluation results. Based on the updated monitoring strategy, the abnormal and confirmed frequency data for the monitoring point are recollected. The above process is repeated to continuously optimize the calculated frequency benchmark and adjustment coefficient.

[0082] For example, during real-time data collection at monitoring points, sensors deployed throughout the area can continuously capture dynamic frequency-related data. Suppose the abnormal frequency value at a monitoring point consistently reaches 50 times per minute for a period of time, significantly exceeding the preset threshold of 30 times per minute. At this point, the system automatically triggers a recalculation of the initial baseline, extracting the frequency baseline parameters for the past 30 days from historical data. The updated initial frequency baseline value might be adjusted to 35 times per minute. This approach dynamically adjusts the baseline based on actual operating conditions, ensuring the rationality of subsequent adjustments.

[0083] Specifically, when calculating the adjustment coefficient, we can consider the changing characteristics of the time dimension, such as the difference in frequency fluctuations between weekdays and non-weekdays. Assuming that the frequency peak on weekdays typically occurs between 9:00 AM and 11:00 AM, while the peak on non-weekdays is more dispersed, the system would use this pattern to calculate an adjustment coefficient of 1.2 for weekdays and 0.8 for non-weekdays. If the adjustment coefficient meets the requirements of the periodic evaluation, for example, within the range of 0.5 to 1.5, it is further optimized through a regression model to obtain a more realistic adjustment parameter for the time dimension. This approach effectively adapts to the frequency fluctuation characteristics of different time periods.

[0084] For example, when calibrating the frequency data for anomaly confirmation, if the calibrated anomaly frequency remains at 45 per minute, exceeding the preset threshold, the system will generate periodic assessment results. Based on the anomaly point information in the assessment results, the monitoring strategy can be updated, such as increasing the sampling frequency at that point from once per minute to once every 30 seconds, to more accurately capture abnormal fluctuations. This approach allows for timely identification of potential problems and the implementation of targeted measures.

[0085] Specifically, after implementing the updated monitoring strategy, the system will recollect abnormal frequency and confirmed frequency data, cyclically optimizing the frequency benchmark and adjustment factor. For example, if the abnormal frequency drops to 38 times per minute, approaching the benchmark value, during the next round of data collection, the adjustment strategy is effective. This cyclic optimization mechanism continuously improves monitoring accuracy and ensures the stability of frequency configuration within the region.

[0086] For example, in practical applications, auxiliary parameters can be introduced to account for the characteristics of different locations, such as the impact of environmental factors on frequency. If a location is located in a high-interference area, the system will additionally consider the impact of interference intensity on the frequency reference and dynamically adjust the calibration parameters. This multi-dimensional approach can further enhance the adaptability of monitoring strategies and ensure that they maintain optimal operation even in complex environments.

[0087] The above embodiments are intended to illustrate the technical solutions of the present invention and are not intended to limit the present invention. The present invention is described in detail with reference to the preferred embodiments only. It should be understood by those skilled in the art that the technical solutions of the present invention may be modified or replaced with equivalents without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or equivalents should be included within the scope of the claims of the present invention.

Claims

1. A real-time monitoring and alarm method for abnormalities of urban manhole covers based on the Internet of Things, characterized in that: The method comprises: Obtain a pre-established regional attribute database, extract traffic flow parameters and historical abnormal frequency parameters corresponding to the monitoring points, generate regional differentiated initial acquisition frequency reference values based on the parameters, and form an initial frequency configuration table; According to the initial frequency configuration table, combined with historical time regularity data, the abnormal probability distribution parameters of each time period are extracted. If the abnormal probability of the peak period exceeds the first preset threshold, the acquisition frequency of the period is increased to the first high frequency mode, and the time dimension frequency adjustment coefficient is determined; Acquire real-time displacement parameters and vibration parameters of the monitoring sensor. If the parameters exceed the preset normal range threshold, trigger an abnormality detection signal and generate a preliminary abnormality judgment result; According to the preliminary abnormality judgment result, the dynamic acquisition strategy module is activated, the acquisition mode of the corresponding monitoring point is switched to the second high-frequency mode, high-density displacement vibration data is acquired, and abnormal characteristic parameters are extracted; Performing similarity matching between the abnormal feature parameters and the historical case library, and confirming the abnormal event status if the matching degree exceeds a second preset threshold; According to the abnormal event status, the acquisition frequency of the non-abnormal area is adjusted to a low-frequency mode to generate a global energy consumption optimization plan; Updating the time regularity parameters of the regional attribute database, recording the current frequency configuration status, and generating a dynamic frequency management plan; If the abnormality confirmation frequency of the monitoring point continues to exceed the third preset threshold, the initial frequency reference value and the time dimension adjustment coefficient are recalculated to generate a periodic evaluation result.

2. The method for real-time monitoring and alarming of abnormalities of urban manhole covers based on the Internet of Things according to claim 1 is characterized in that: The method of obtaining a pre-established regional attribute database, extracting traffic flow parameters and historical abnormal frequency parameters corresponding to monitoring points, and generating regional differentiated initial acquisition frequency reference values based on the parameters to form an initial frequency configuration table includes: Obtain relevant data records of monitoring points from a pre-established regional attribute database, wherein the data records include traffic flow information and historical abnormality records, and organize them into a structured data set; For the structured data set, a statistical analysis method is used to extract traffic flow parameters and historical abnormal frequency parameters of each monitoring point to obtain a parameter set; Based on the parameter set, a linear regression model is used to perform weighted calculation on traffic flow parameters and historical abnormal frequency parameters to generate differentiated initial acquisition frequency benchmark values and determine a frequency benchmark data set; If the benchmark value of a monitoring point in the frequency benchmark data set exceeds the preset threshold range, the traffic flow parameters and historical abnormal frequency parameters of the point are double-checked, and the benchmark value is adjusted in combination with the regional attribute information to obtain an adjusted frequency benchmark data set; By using the adjusted frequency reference data set, the initial collected frequency reference values are mapped to corresponding points according to regional attributes and monitoring point information to form an initial frequency configuration table; Comparing the frequency values in the initial frequency configuration table with the historical abnormal frequency parameters, if there is a significant deviation, fine-tuning the frequency values based on the difference calculation results to generate a final frequency configuration table; According to the final frequency configuration table, associated data acquisition and parameter extraction logic are used to format the frequency values in the table and output standardized configuration data records.

3. The method for real-time monitoring and alarming of abnormalities of urban manhole covers based on the Internet of Things according to claim 1 is characterized in that: The method extracts the abnormal probability distribution parameters of each time period based on the initial frequency configuration table and in combination with historical time regularity data. If the abnormal probability of a peak period exceeds a first preset threshold, the acquisition frequency of the period is increased to a first high frequency mode, and the time dimension frequency adjustment coefficient is determined, including: Obtain the initial acquisition frequency and historical regularity data, analyze the changes in abnormal probability in each time dimension, and obtain the abnormal probability distribution parameters under time period division; According to the abnormal probability distribution parameters, the data records during the peak period are evaluated. If the abnormal probability exceeds the preset probability threshold, the acquisition frequency adjustment logic is triggered to determine the activation conditions of the high-frequency mode; Adopting the high-frequency mode, dynamically adjusting the collection frequency during peak hours, calculating the adjustment coefficient of the time dimension in combination with the time period division information, and obtaining the configuration data after frequency adjustment; For the configuration data after the frequency adjustment, the historical laws of abnormal probability are integrated, and the trend of distribution parameter changes in each time period is analyzed to determine whether there is a continuous abnormal probability fluctuation; If the abnormal probability fluctuation exceeds the preset fluctuation range, the acquisition frequency is recalibrated in combination with the time dimension and the adjustment coefficient to obtain an optimized frequency configuration record; By using the optimized frequency configuration records, correlating the time period division and the abnormal probability data of the peak period, the adjustment coefficient is verified using a logistic regression model to determine the final frequency configuration plan; According to the final frequency configuration scheme, corresponding execution parameters are generated for the acquisition frequency of each time dimension to obtain standardized frequency adjustment data.

4. The method for real-time monitoring and alarming of abnormalities of urban manhole covers based on the Internet of Things according to claim 1 is characterized in that: The real-time displacement parameters and vibration parameters of the monitoring sensor are obtained. If the parameters exceed the preset normal range threshold, an abnormality detection signal is triggered to generate a preliminary abnormality judgment result, including: Acquire displacement parameters and vibration parameters from the monitoring sensor, record the parameters in real time through a data acquisition module, and determine an original parameter data set; For the original parameter data set, the preset threshold is used to determine the range of the displacement parameters and vibration parameters. If the parameters exceed the normal range, an abnormality detection signal is generated to obtain a preliminary abnormality identification; Based on the preliminary anomaly identification, the trigger time and duration of the anomaly detection signal are obtained. The signal features are extracted using time series analysis tools to determine the time window in which the anomaly occurs. For the time window when the anomaly occurs, obtain the monitoring sensor data records within the relevant period, analyze the changing trends of displacement parameters and vibration parameters through the data comparison module, and determine whether there are persistent abnormal characteristics; If persistent abnormal characteristics are confirmed, the parameter fluctuation amplitude corresponding to the change trend is obtained, and the severity classification of the abnormality is determined by comparing the fluctuation amplitude with the preset threshold; According to the severity of the anomaly classification, the logistic regression model is used to verify the anomaly detection signal, and the preliminary anomaly identification is calibrated through the model output results to obtain the final anomaly judgment data; For the final abnormality judgment data, obtain the monitoring sensor identification corresponding to the judgment data, archive the abnormality records through the data storage module, and determine the basic information for subsequent data queries.

5. The method for real-time monitoring and alarming of abnormalities of urban manhole covers based on the Internet of Things according to claim 1 is characterized in that: According to the preliminary abnormality judgment result, the dynamic acquisition strategy module is activated, the acquisition mode of the corresponding monitoring point is switched to the second high-frequency mode, high-density displacement vibration data is acquired, and abnormal characteristic parameters are extracted, including: By starting the dynamic acquisition strategy module and adjusting the acquisition mode of the monitoring node to the high-frequency mode based on the abnormal judgment results, displacement and vibration data are obtained; According to the acquired displacement vibration data, a data preprocessing method is used to remove noise interference and obtain a cleaned vibration data stream; If there are fluctuations exceeding a preset threshold in the cleaned vibration data stream, the abnormal characteristic parameters are determined through feature extraction methods; According to the determined abnormal feature parameters, the support vector machine algorithm is used to classify the abnormal features and determine the abnormal type; If the abnormal type determined belongs to the high-risk category, the data density enhancement mechanism is triggered to obtain higher-density data samples; Through in-depth analysis of higher-density data samples, potential abnormal patterns can be extracted and the changing trend of abnormal development can be determined; According to the changing trend of abnormal development, adjust the operating parameters of the dynamic collection strategy module to optimize the frequency and scope of subsequent data collection.

6. The method for real-time monitoring and alarming of abnormalities of urban manhole covers based on the Internet of Things according to claim 1 is characterized in that: The abnormal feature parameters are matched with the historical case library for similarity, and if the matching degree exceeds a second preset threshold, the abnormal event status is confirmed, including: By extracting abnormal features from the input data and using feature decomposition technology to preliminarily process the data, an abnormal feature set is obtained; According to the abnormal feature set, relevant records in the historical case library are obtained, and the similarity value between the two is calculated using the cosine similarity algorithm; If the similarity value is higher than the preset threshold, the current data is determined to be in an abnormal determination state and the abnormal event state identifier is output; Based on the abnormal event status identification, several historical cases with the highest matching degree are selected from the case library to obtain the corresponding solution records; Determine the processing priority and related strategies for the current abnormal characteristics by integrating and analyzing the screened solution records; Generate targeted repair plans based on processing priorities and relevant strategies, and output the final event status confirmation results; After obtaining the event status confirmation result, the historical case library is updated, and the current abnormal characteristics and processing strategies are added to form a closed-loop data supplement.

7. The method for real-time monitoring and alarming of abnormalities of urban manhole covers based on the Internet of Things according to claim 1 is characterized in that: The step of adjusting the acquisition frequency of the non-abnormal area to a low-frequency mode according to the abnormal event state and generating a global energy consumption optimization solution includes: The monitoring system obtains real-time data from each area, analyzes whether there are abnormal events, and obtains preliminary event status classification results; Based on the preliminary event status classification results, identify the non-abnormal areas and determine the areas where the acquisition frequency needs to be adjusted; Using a preset threshold standard, if the data flow of the identified non-abnormal area is lower than the preset threshold, the collection frequency is adjusted to a low-frequency mode to obtain the adjusted frequency configuration; By adjusting the frequency configuration, the energy consumption trend of each area is calculated to determine whether the global energy consumption reaches the optimization target; If the global energy consumption does not reach the optimization target, the frequency control parameters of the non-abnormal area are readjusted according to the energy consumption change trend to obtain a new frequency configuration plan; According to the new frequency configuration plan, update the collection frequency settings of each area and obtain the updated energy consumption data distribution; The updated energy consumption data distribution is combined with the support vector machine algorithm to predict and analyze the global energy consumption and determine the final optimization strategy.

8. The method for real-time monitoring and alarming of abnormalities of urban manhole covers based on the Internet of Things according to claim 1 is characterized in that: The updating of the time regularity parameters of the regional attribute database, recording the current frequency configuration status, and generating a dynamic frequency management solution includes: Obtain historical data from the regional attribute database, perform feature extraction based on temporal patterns, and obtain the temporal distribution pattern; Analyze the frequency configuration status based on the time-based distribution pattern and determine whether the configuration parameters meet the preset threshold range. If they are below the threshold range, trigger the parameter update process and determine the updated parameter value. Generate a preliminary frequency adjustment plan based on the updated parameter values and the dynamic frequency adjustment strategy, and obtain key indicator values in the adjustment plan; Analyze the integrity of data records for key indicator values. If records are missing or abnormal, call the pre-established support vector machine model to complete the data and obtain a complete indicator data set. Based on the complete indicator data set and the business objectives of the management plan, an optimized configuration of dynamic frequency is generated to determine whether the optimized configuration meets the requirements of regional attributes; If the optimized configuration meets the requirements of the regional attributes, the optimized configuration is written to the database, the frequency configuration status is updated, and the latest configuration status information is obtained; Generate the final dynamic frequency management plan based on the latest configuration status information and determine the execution time and scope of the plan.

9. The method for real-time monitoring and alarming of abnormalities of urban manhole covers based on the Internet of Things according to claim 1 is characterized in that: If the abnormal confirmation frequency of the monitoring point continues to exceed the third preset threshold, the initial frequency reference value and the time dimension adjustment coefficient are recalculated to generate a periodic evaluation result, including: By collecting real-time data from monitoring points, the dynamic change data of abnormal frequency and confirmation frequency is obtained to determine whether the current monitoring status exceeds the preset threshold; If the frequency is confirmed to have continuously exceeded the preset threshold, the initial reference recalculation process is triggered, and the relevant parameters of the frequency reference are extracted from the historical data to obtain the updated initial frequency reference value; Based on the updated initial frequency reference value and the change characteristics of the time dimension, the corresponding adjustment coefficient is calculated to determine whether the adjustment coefficient meets the requirements of the periodic evaluation; If the adjustment coefficient meets the requirements of the periodic evaluation, the frequency benchmark and adjustment coefficient are optimized using a pre-established regression model to obtain the optimized time dimension adjustment parameters; By adjusting the parameters in the optimized time dimension, the frequency data of the abnormal confirmation is calibrated to determine whether the abnormal frequency after calibration still exceeds the preset threshold; If the abnormal frequency after calibration still exceeds the preset threshold, periodic evaluation result data is generated, and the monitoring strategy of the monitoring point is updated based on the abnormal point information in the evaluation result; According to the updated monitoring strategy, the abnormal frequency and confirmed frequency data of the monitoring points are collected again, and the above process is executed repeatedly to continuously optimize the calculation results of the frequency reference and adjustment coefficient.

Images