Well lid state real-time monitoring and intelligent early warning system based on Internet of Things

Abstract

The invention discloses a well lid state real-time monitoring and intelligent early warning system based on the Internet of Things, and belongs to the technical field of monitoring and alarming, and the system specifically comprises the steps: integrating a plurality of sensors at a well lid monitoring terminal, and synchronously collecting the operation state data of equipment; map marking is conducted on the well lids according to the equipment operation state, and the well lids with the normal equipment state are determined; classifying well lids with normal equipment states, and determining an inclination angle threshold value, a vibration frequency threshold value and a vibration noise threshold value of a well lid vibration noise risk, a well lid sinking risk or a road surface flatness risk according to different well lid types; judging whether the well lid with the abnormal fault precursor has a well lid vibration noise risk, a well lid sinking risk or a road surface flatness risk or not; meanwhile, abnormal displacement and water immersion conditions of the well lid are monitored; performing graded early warning on the intelligent well lid according to a judgment result, and correspondingly triggering different grades of early warning signals; and pushing early warning information and an operation and maintenance work order according to the triggered early warning signals of different levels.

Description

Technical Field

[0001] This invention relates to the field of monitoring and alarm technology, specifically to a real-time monitoring and intelligent early warning system for manhole cover status based on the Internet of Things. Background Technology

[0002] In the field of urban infrastructure operation and maintenance, manhole covers, as key nodes in underground pipe networks, are characterized by their large number, wide distribution, and complex operating conditions. Their operational status directly affects road traffic safety and the living environment of residents. With the penetration of IoT technology in infrastructure monitoring, existing IoT-based manhole cover monitoring systems are gradually replacing traditional models. Their core architecture comprises three main modules: a front-end monitoring terminal, a low-power data transmission network, and a back-end management platform. The front-end monitoring terminal integrates sensing devices such as vibration sensors, tilt sensors, and noise sensors, enabling real-time collection of operational status data such as vibration frequency, tilt angle, and vibration noise of the manhole cover. The data transmission link utilizes low-power wide-area network technologies such as NB-IoT and LoRa to achieve stable uploading of monitoring data to the back-end platform, adapting to the needs of dispersed deployment of urban manhole covers. The back-end management platform has basic data storage and threshold comparison functions, capable of determining whether a manhole cover is abnormal based on preset simple thresholds. Some systems also support marking the location of manhole covers and their "normal / abnormal" basic status on electronic maps and pushing simple maintenance work orders to maintenance personnel. This has initially realized the transformation of manhole cover monitoring from "passive maintenance" to "proactive early warning," improving maintenance response speed to some extent.

[0003] However, existing IoT-based manhole cover monitoring systems still have significant technical shortcomings in addressing the three core issues of manhole cover vibration and noise, road surface unevenness, and manhole cover subsidence, as well as the need for accurate identification of equipment status, making it difficult to meet the requirements of refined operation and maintenance. The existing systems lack specificity in their warning threshold settings and fail to integrate differences in manhole cover types and installation scenarios—most systems use uniform and fixed thresholds without considering the characteristics of different manhole cover materials and different installation environments. Summary of the Invention

[0004] The purpose of this invention is to provide a real-time monitoring and intelligent early warning system for manhole cover status based on the Internet of Things, and to solve the following technical problems: Existing IoT-based manhole cover monitoring systems, which address three core issues—manhole cover vibration and noise, road surface smoothness, and manhole cover subsidence—as well as the need for accurate identification of equipment status, still have significant technical shortcomings and are difficult to meet the requirements of refined operation and maintenance.

[0005] The objective of this invention can be achieved through the following technical solutions: The IoT-based real-time monitoring and intelligent early warning system for manhole cover status includes: The data acquisition module is used to integrate multiple sensors into the manhole cover monitoring terminal to synchronously collect equipment operating status data, including: vibration frequency, vibration noise, and tilt angle. The early warning equipment marking module is used to mark the location of the manhole cover on the map according to the equipment's operating status, and to identify manhole covers with normal equipment status. The manhole cover fault pre-diagnosis module is used to analyze the time-series changes in the long-term operating status of manhole covers, uncover the gradual abnormal characteristics before the occurrence of faults, and pre-diagnose and mark abnormal manhole covers with signs of faults in real time. The warning threshold determination module is used to classify manhole covers with normal equipment status and determine the tilt angle threshold, vibration frequency threshold, and vibration noise threshold for different types of manhole covers, such as manhole cover vibration and noise risk, manhole cover sinking risk, or road surface unevenness risk. The graded early warning signal triggering module is used to determine whether the abnormal manhole cover is at risk of vibration and noise, sinking, or road surface unevenness. Based on the determination result, the module provides graded early warnings for the smart manhole cover and triggers different levels of early warning signals accordingly. It also monitors abnormal displacement and water immersion of the manhole cover. The manhole cover alarm module is used to push early warning information and maintenance work orders based on the triggering of different levels of early warning signals, and to notify staff through corresponding alarm methods.

[0006] As a further aspect of the present invention: in the early warning equipment marking module, the process of marking the location of the manhole cover on the map according to the equipment operating status is as follows: The early warning device marking module receives real-time device operation status data output by the data acquisition module, as well as sensor heartbeat data and the real-time upload status of the acquired data; If the equipment operating status data is within the normal range defined for equipment operating status data and the sensor heartbeat data and collected data are uploaded in real time without interruption, the corresponding manhole cover will be marked with a blue label on the map. If the real-time upload of sensor heartbeat data or acquired data is interrupted, the module will trigger a manual opening and inspection process: When a man performs the opening operation manually, if the device simultaneously uploads the collected data, the corresponding manhole cover will be marked with an orange label on the map; If the equipment still fails to upload heart rate data and collected data after the manual opening operation is completed, it is determined that the equipment is in an abnormal state, and the corresponding manhole cover will be marked with a red mark on the map.

[0007] As a further aspect of the present invention: in the manhole cover fault pre-diagnosis module, the process of real-time pre-diagnosis and marking of abnormal manhole covers with early signs of fault is as follows: Collect historical equipment operation data of the target manhole cover monitoring terminal for n consecutive days, define the fault precursor period by associating it with the fault records of the same period, associate the historical operation data with the precursor label and label positive and negative samples, where n is a preset value; The daily rate of change of the equipment operating status data and the difference between the trend of the set sliding window are calculated and fused into a time-series progressive feature set. A fault precursor recognition model for a single manhole cover is obtained by using a temporal convolutional network as the architecture, training with a temporal progressive feature set and precursor labels, and optimizing through backpropagation. The system collects real-time data on the equipment's operating status from the manhole cover monitoring terminal to generate a time-series progressive feature. This feature is then input into a fault precursor identification model to obtain a probability value. If the probability value meets the criteria, the manhole cover is marked as having a fault precursor.

[0008] As a further aspect of the present invention: in the early warning threshold determination module, the process of determining the tilt angle threshold, vibration frequency threshold, and vibration noise threshold for the risks of manhole cover vibration noise, manhole cover subsidence, or road surface unevenness is as follows: Extract the installation environment attributes and physical specification attributes of manhole covers in normal condition to form a classification raw dataset; Based on the original dataset for classification, a hierarchical clustering algorithm combined with domain rules is used to classify manhole covers into heavy traffic type, regular road type, and slow traffic area type. Based on the classification results, filter the historical monitoring data of vibration frequency, tilt angle, and vibration noise during the normal operating period of the equipment under the corresponding category, and build a category-specific database. Data from the past x months of normal monitoring periods for different types of manhole covers were collected from a dedicated database. Based on the installation environment and physical specifications of the manhole covers, dedicated safety redundancy coefficients, finite element model mechanical parameters, and flatness analysis windows were set. The parameters were then substituted into the unified derivation logic in the order of manhole cover vibration and noise risk, manhole cover subsidence risk, and road surface flatness risk to obtain the tilt angle, vibration frequency, and vibration and noise threshold for each type of risk, where x is a preset value.

[0009] As a further aspect of the present invention: the process of classifying manhole covers into heavy-duty traffic type, regular road type, and slow-traffic area type based on the original classification dataset and using a hierarchical clustering algorithm combined with domain rules is as follows: By connecting with urban geographic information systems, traffic flow monitoring databases, and vibration source mapping systems, the road type, average daily traffic volume level, and surrounding vibration source intensity values ​​of manhole covers are extracted. Access the manhole cover production archive database to extract the load-bearing capacity, manhole cover material, and manhole cover diameter; By associating the installation environment attributes with the physical specification attributes, removing samples with missing values, performing one-hot encoding on non-numerical attributes and standardization on numerical attributes, a classification raw dataset is generated. Using the original dataset for classification as input, the Ward clustering criterion and Euclidean distance metric are adopted, and the clustering termination condition is set as the maximum intra-class distance threshold to obtain the initial classification clusters; Based on the urban manhole cover operation and maintenance standards, the initial clusters that meet the requirements of main roads + heavy load-bearing capacity + high daily traffic volume, secondary roads + standard load-bearing capacity + medium daily traffic volume, and sidewalks + light load-bearing capacity + low daily traffic volume are respectively divided into heavy traffic type, regular road type, and slow traffic area type. The remaining samples are redistributed according to the weighted similarity of installation environment attributes and physical specification attributes to determine the final classification.

[0010] As a further aspect of the present invention: the process of filtering historical monitoring data of vibration frequency, tilt angle, and vibration noise during normal operating periods of equipment under the corresponding category according to the classification results, and constructing a category-specific database is as follows: By using the unique identifier of the manhole cover, the classification results of the manhole cover are linked with the historical monitoring data stored in the data acquisition module, including vibration frequency, tilt angle, vibration noise, and acquisition timestamp, to generate a linked dataset; Based on the warning threshold, determine the normal range of the module's parameters, verify the associated dataset, and select the period when the three parameters under the same manhole cover's unique identifier are within the normal range for 12 consecutive collection cycles without data loss or interruption, and mark it as the period when the equipment is operating normally. The datasets are grouped and associated according to the classification results of manhole covers. In each group, the monitoring data collected during the normal operation period of the equipment are retained to obtain different types of manhole cover classification and filtering datasets. Create a dedicated database with specified fields for different types of manhole covers to classify and filter datasets, import the corresponding data, and set up a timed synchronization mechanism to receive newly added normal data.

[0011] As a further aspect of the present invention: the process of obtaining the tilt angle, vibration frequency, and vibration noise threshold of various risks is as follows: Data from nearly x months of normal-time monitoring data of heavy-duty traffic type, regular road type, and slow-traffic area type manhole covers were collected from a classification-specific database to generate various target datasets. Based on the installation environment and physical specifications of different types of manhole covers, a dedicated safety redundancy coefficient, finite element model mechanical parameters, and road surface smoothness analysis time window are set. The data preprocessing, finite element simulation, and spatiotemporal feature analysis logic were reused in the order of vibration and noise risk, manhole cover sinking risk, and road surface smoothness risk. The tilt angle, vibration frequency, and vibration and noise threshold of each type of risk were obtained by substituting specific parameters. The data is compiled into a three-threshold mapping table of manhole cover category, risk type, and associated attribute identification, and stored in the threshold database of the early warning threshold determination module for use by the graded early warning signal triggering module.

[0012] As a further aspect of the present invention: in the graded early warning signal triggering module, the process of determining whether the abnormal manhole cover has a risk of vibration and noise, a risk of sinking, or a risk of uneven road surface is as follows: The target manhole cover type is obtained, and the tilt angle threshold, vibration frequency threshold, and vibration noise threshold for vibration and noise risk of the target manhole cover in that type are obtained; the tilt angle threshold, vibration frequency threshold, and vibration noise threshold for manhole cover subsidence risk; and the tilt angle threshold, vibration frequency threshold, and vibration noise threshold for road surface unevenness risk. The vibration frequency, vibration noise, and tilt angle of the target manhole cover are acquired in real time. If any of these parameters exceeds the corresponding threshold, a corresponding risk warning is issued for the manhole cover.

[0013] As a further aspect of the present invention: the specific method for providing graded early warnings for smart manhole covers based on the judgment results, and for triggering different levels of early warning signals accordingly, is as follows: If any of the vibration frequency, vibration noise, or tilt angle exceeds the corresponding threshold, the manhole cover is determined to be a primary risk warning. If any two of the vibration frequency, vibration noise, or tilt angle exceed the corresponding threshold, the manhole cover is judged to be at a medium-risk warning level. If the vibration frequency, vibration noise, and tilt angle all exceed the corresponding thresholds, the manhole cover is judged to be a serious risk warning.

[0014] The beneficial effects of this invention are: This invention precisely addresses the shortcomings of existing systems in accurately monitoring three core issues: manhole cover vibration and noise, road surface smoothness, and manhole cover subsidence, thus meeting the requirements of refined operation and maintenance. The data acquisition module integrates multiple sensors in the manhole cover monitoring terminal to simultaneously collect three key data points: vibration frequency, vibration noise, and tilt angle. This comprehensively captures the characteristic signals of these three issues—vibration noise data directly correlates with noise pollution, tilt angle data reflects the trend of manhole cover subsidence, and the combined data of vibration frequency and tilt angle accurately determines road surface smoothness, avoiding the limitations of single-parameter monitoring in existing systems. Simultaneously, the warning threshold determination module first classifies manhole covers in normal condition and then customizes specific thresholds for different types of manhole covers. This completely solves the problem of mismatch between the existing system's uniform threshold and the manhole cover material and installation environment, significantly reducing false alarms and missed alarms in the absence of noise, minimizing ineffective maintenance actions, lowering maintenance costs, and ensuring that the three core issues are identified promptly and accurately.

[0015] This invention overcomes the limitations of existing systems, such as ambiguous equipment status identification and inefficient early warning response, improving the accuracy and timeliness of maintenance responses. The early warning equipment marking module identifies manhole covers in normal condition using map markings, providing accurate basic data support for subsequent classification and threshold determination, and clearly distinguishing them from abnormal states, avoiding the shortcomings of the existing system's vague "normal / abnormal" classification. The tiered early warning signal triggering module implements primary, intermediate, and severe early warnings based on the number of parameters exceeding limits, allowing staff to prioritize high-risk issues according to the warning level, avoiding resource misallocation. The manhole cover alarm module pushes early warning information and maintenance work orders based on different levels of warning signals, ensuring that all three types of risks and equipment anomalies are quickly transmitted to staff, shortening fault response time—avoiding over-response for minor risks and delays in handling severe risks, ultimately reducing safety accidents and public complaints caused by manhole cover problems, and protecting road traffic safety and the living environment of residents. Attached Figure Description

[0016] The invention will now be further described with reference to the accompanying drawings.

[0017] Figure 1 This is a schematic diagram of the structure of the Internet of Things-based real-time monitoring and intelligent early warning system for manhole cover status according to the present invention. Detailed Implementation

[0018] 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.

[0019] Please see Figure 1 As shown, this invention is an IoT-based real-time monitoring and intelligent early warning system for manhole cover status, comprising: The data acquisition module is used to integrate multiple sensors into the manhole cover monitoring terminal to synchronously collect equipment operating status data, including: vibration frequency, vibration noise, and tilt angle. The early warning equipment marking module is used to mark the location of the manhole cover on the map according to the equipment's operating status, and to identify manhole covers with normal equipment status. The manhole cover fault pre-diagnosis module is used to analyze the time-series changes in the long-term operating status of manhole covers, uncover the gradual abnormal characteristics before the occurrence of faults, and pre-diagnose and mark abnormal manhole covers with signs of faults in real time. The warning threshold determination module is used to classify manhole covers with normal equipment status and determine the tilt angle threshold, vibration frequency threshold, and vibration noise threshold for different types of manhole covers, such as manhole cover vibration and noise risk, manhole cover sinking risk, or road surface unevenness risk. The graded early warning signal triggering module is used to determine whether the abnormal manhole cover is at risk of vibration and noise, sinking, or road surface unevenness. Based on the determination result, the module provides graded early warnings for the smart manhole cover and triggers different levels of early warning signals accordingly. It also monitors abnormal displacement and water immersion of the manhole cover. The manhole cover alarm module is used to push early warning information and maintenance work orders based on the triggering of different levels of early warning signals, and to notify staff through corresponding alarm methods.

[0020] The data acquisition module is used to integrate multiple sensors into the manhole cover monitoring terminal to synchronously collect equipment operating status data, including vibration frequency, vibration noise, and tilt angle.

[0021] In manhole cover monitoring, the acquisition of three parameters—vibration frequency, vibration noise, and tilt angle—is achieved through a high-precision triaxial accelerometer built into the intelligent manhole cover safety monitoring terminal. The vibration frequency is indirectly calculated from the collected raw acceleration data using signal processing algorithms. The tilt angle is indirectly calculated using the sensed gravitational acceleration component. Simultaneously, vibration noise is inferred by monitoring the vibration frequency: the higher the vibration frequency, the more intense the vibration, and the more intense the vibration, the greater the noise. Specifically, after acquiring the vibration frequency, the triaxial accelerometer simultaneously obtains the vibration frequency spectrum and the acceleration amplitude corresponding to each frequency. The vibration energy is then calculated using both frequency and amplitude. Subsequently, based on an acoustic radiation model, the vibration energy is mapped to the sound pressure level parameter in the air, thus obtaining the characteristic parameters of the vibration noise. In the early warning device marking module, the process of marking the location of the manhole cover on the map according to the device's operating status and determining the manhole cover with a normal device status is as follows: The early warning equipment marking module receives real-time equipment operation status data (including operating parameters such as vibration frequency and tilt angle) output by the data acquisition module, as well as sensor heartbeat data and the real-time upload status of the acquired data.

[0022] If the device's operating status data falls within the defined normal range (vibration frequency, tilt angle, and other parameters are all within the threshold), and the heartbeat data and collected data are uploaded in real time without interruption, then the corresponding smart manhole cover will be marked with a blue label on the map (the corresponding device is in normal condition).

[0023] If the real-time upload of heartbeat data or collected data is interrupted (signal interruption), the module will trigger a manual opening and inspection procedure: When a manual operation is performed to open the cover, if the device simultaneously uploads the collected data (data is reported when the cover is opened), the corresponding smart manhole cover will be marked with an orange label on the map (the corresponding device signal is lost).

[0024] If the device still fails to upload heartbeat data and collected data after the manual opening operation, it is determined that the device is in an abnormal state, and the corresponding smart manhole cover will be marked with a red mark on the map (corresponding to abnormal device state).

[0025] In the manhole cover fault pre-diagnosis module, the process of real-time pre-diagnosis and marking of abnormal manhole covers with early signs of fault is as follows: Collect historical data on the equipment operation status of the target manhole cover monitoring terminal for 60 consecutive days. The data dimensions include the daily average value of vibration frequency, the daily average value of vibration noise, and the daily average value of tilt angle, ensuring that the data sampling time interval is uniform (collected once at a fixed time every day). By linking the historical maintenance records of the manhole cover during the same period, the confirmed fault records (such as subsidence repair records and vibration and noise exceeding standard handling records) are filtered out, and the 14 days before the fault occurred are defined as the time of fault precursor; By associating historical operational status data with fault precursor time period labels, a single manhole cover time-series operational dataset containing operational parameters, timestamps, and fault precursor labels is constructed. Data within the fault precursor time period is labeled as positive samples, while normal operation data outside the fault precursor time period is labeled as negative samples.

[0026] Based on the time-series operation dataset of a single manhole cover, the daily rate of change of the three operating parameters is calculated. The calculation formula is: (parameter value of the day - parameter value of the previous day) / parameter value of the previous day × 100%, which is used to reflect the relative change of the parameters each day. The trend difference of the three operating parameters is calculated using a 7-day sliding window. Specifically, with the current day as the end of the window, the difference between the average value of the parameters in the previous 7 days and the average value of the parameters in the next 7 days is calculated to capture the gradual trend change of the parameters in a two-week cycle. The six features—the daily variation rate of vibration frequency, the daily variation rate of vibration noise, the daily variation rate of tilt angle, the 7-day sliding window trend difference of vibration frequency, the 7-day sliding window trend difference of vibration noise, and the 7-day sliding window trend difference of tilt angle—are fused together to form a time-series progressive feature set.

[0027] Temporal convolutional networks are used as the basic model architecture, which has the ability to capture long-term temporal dependencies and can effectively identify the gradual patterns of parameter changes. The generated time-series progressive feature set was used as the input layer data of the model, and the fault precursor labels (positive samples / negative samples) in the single manhole cover time-series running dataset were used as the output layer labels of the model. 80% of the dataset was divided into training set and 20% of the dataset was divided into validation set. The model parameters are iteratively optimized using the backpropagation algorithm to minimize the model's misjudgment rate for the early signs of failure until the model's accuracy in identifying early signs on the validation set stabilizes, thus obtaining a fault early sign identification model specific to a single manhole cover.

[0028] The equipment operation status data output by the manhole cover monitoring terminal is collected in real time (once every 2 hours). The daily change rate and 7-day sliding window trend difference corresponding to the real-time data are generated synchronously according to the same calculation method to form real-time time-series progressive features. The real-time progressive features are input into the trained single-manhole cover-specific fault precursor recognition model, and the model outputs the probability value that the current moment belongs to the fault precursor period. If the probability value output by the model reaches the preset judgment standard, the manhole cover will be marked as an abnormal manhole cover with a precursor to failure, and the marking time and the corresponding real-time operating parameters will be recorded simultaneously.

[0029] In the aforementioned warning threshold determination module, manhole covers with normal equipment status are classified, and the tilt angle threshold, vibration frequency threshold, and vibration noise threshold are determined according to different manhole cover types. Extract the installation environment attributes and physical specification attributes of manhole covers in normal condition to form a classification raw dataset; Based on the original dataset for classification, a hierarchical clustering algorithm combined with domain rules is used to classify manhole covers into heavy traffic type, regular road type, and slow traffic area type. Based on the classification results, filter the historical monitoring data of vibration frequency, tilt angle, and vibration noise during the normal operating period of the equipment under the corresponding category, and build a category-specific database. Data from the past x months of normal monitoring of three types of manhole covers were collected from a dedicated database. Based on their installation environment and physical specifications, dedicated safety redundancy coefficients, finite element model mechanical parameters, and flatness analysis windows were set. The parameters were then substituted into the dedicated parameters using a unified derivation logic, following the order of manhole cover vibration and noise risk, manhole cover subsidence risk, and road surface flatness risk. The tilt angle, vibration frequency, and vibration and noise threshold for each type of risk were obtained, where x is a preset value.

[0030] The process of extracting the installation environment attributes and physical specifications of manhole covers in normal condition to form a classification dataset; based on this classification dataset, a hierarchical clustering algorithm combined with domain rules is used to classify manhole covers into heavy-duty traffic type, regular road type, and slow-traffic area type, specifically including the following: By connecting with urban geographic information systems, traffic flow monitoring databases, and vibration source mapping systems, the system extracts information such as road type, average daily traffic volume level, and surrounding vibration source intensity values ​​for manhole covers. The urban geographic information system, based on urban road planning and construction archives, systematically records the official classification of each road segment (e.g., main roads, secondary roads, sidewalks), allowing for accurate identification of the road type where the manhole cover is located. The traffic flow monitoring database continuously tracks vehicle traffic volume per unit time using monitoring equipment deployed along the road segment, summarizing it into traffic volume levels according to a preset cycle (e.g., daily average), directly reflecting the frequency and intensity of traffic loads borne by the manhole cover daily. The vibration source mapping system determines the specific location of vibration sources such as rail transit lines and construction areas through on-site surveys, then converts the distance into intensity values ​​based on vibration propagation laws, reflecting the degree of external vibration interference faced by the manhole cover. These three types of information collectively constitute the installation environment attributes of the manhole cover, providing external conditions for subsequent classification.

[0031] By accessing the manhole cover production archive database, we can extract the load-bearing rating, manhole cover material, and manhole cover diameter. This is because manufacturers conduct performance tests and record parameters for each batch of products according to national standards before they leave the factory. The load-bearing rating specifies the maximum load limit that the manhole cover can withstand, the manhole cover material determines its mechanical properties (such as compressive and flexural strength), and the manhole cover diameter affects the stress distribution under load. These physical specifications directly reflect the manhole cover's own risk resistance capability and are indispensable internal conditions for classification.

[0032] By associating installation environment attributes with physical specifications, and removing samples with missing values, one-hot encoding is performed on non-numerical attributes, and standardization is performed on numerical attributes to generate a raw classification dataset. The association operation is achieved through the unique identifier of the manhole cover (such as the factory number) to ensure that the external environment information and its own physical information correspond one-to-one, avoiding data misalignment. Samples with missing values ​​are removed because the absence of any attribute (such as the lack of load-bearing rating) will lead to incomplete classification criteria, which may cause clustering bias. Non-numerical attributes (such as road type and manhole cover material) cannot be directly involved in the algorithm calculation. One-hot encoding can convert them into discrete numerical forms while preserving the category differences of the attributes. Numerical attributes (such as the intensity value of surrounding vibration sources and the diameter of the manhole cover) have different dimensions and numerical ranges (such as intensity value being a coefficient and diameter being a length). Standardization can eliminate the influence of dimensions, so that each attribute has a fair weight in the clustering process. The data after these processing can meet the input requirements of the hierarchical clustering algorithm and form a raw classification dataset.

[0033] Using the original classification dataset as input, the Ward clustering criterion and Euclidean distance metric were employed. The clustering termination condition was set as the maximum intra-cluster distance threshold to obtain initial classification clusters. The Ward clustering criterion was chosen because it merges clusters by minimizing the intra-cluster sum of squares, ensuring high similarity of sample attributes within the merged clusters, which meets the core requirement of similar attributes in manhole cover classification. The Euclidean distance metric is used to calculate the degree of difference between two manhole cover samples across multiple attribute dimensions. It is a commonly used and effective method for measuring sample similarity and is suitable for standardized numerical data. The clustering termination condition was set as the maximum intra-cluster distance threshold because when the maximum attribute distance of samples within a cluster exceeds this threshold, continued merging will lead to excessive differences within the clusters, losing the meaning of classification. The initial classification clusters obtained after stopping merging can ensure the consistency of attributes of samples within the clusters, providing a basis for subsequent rule corrections.

[0034] Based on urban manhole cover operation and maintenance standards, initial clusters meeting the following criteria are categorized into heavy-duty traffic type, regular road type, and slow-traffic area type: main roads + heavy load-bearing capacity + high daily traffic volume; secondary roads + standard load-bearing capacity + medium daily traffic volume; and sidewalks + light load-bearing capacity + low daily traffic volume. The remaining samples are then reassigned based on a weighted similarity of 60% for installation environment attributes and 40% for physical specifications to determine the final classification. The urban manhole cover operation and maintenance standards are based on long-term municipal operation and maintenance practices, clearly defining the required load-bearing capacity of manhole covers and the corresponding traffic flow range for different road types (e.g., main roads require heavy-duty manhole covers to withstand high traffic volumes). Therefore, the initial clusters meeting the above combination conditions are fully compatible. The three typical application scenarios in actual operation and maintenance can be directly classified into corresponding types. The remaining samples are special cases that do not meet the typical combination in the initial cluster (such as light load-bearing manhole covers on main roads). The principle of weighted similarity allocation is that the installation environment attributes (road type, traffic flow, etc.) determine the external load strength that the manhole cover can bear, which has a more critical impact on the classification. Therefore, it is given a weight of 60%. The physical specification attributes are the basis for the manhole cover to adapt to the external load, and are given a weight of 40%. By calculating the weighted similarity between the remaining samples and the three typical types, they are assigned to the type with the best matching attributes. Finally, it is ensured that all manhole cover samples can be classified into the category that conforms to the actual application scenario, thus completing the classification.

[0035] The process of filtering historical monitoring data on vibration frequency, tilt angle, and vibration noise during normal operating periods of equipment under the corresponding category, based on the classification results, and constructing a category-specific database, specifically includes the following: By using the unique identifier of each manhole cover, the classification results are linked to historical monitoring data (including vibration frequency, tilt angle, vibration noise, and collection timestamp) stored in the data acquisition module to generate a related dataset. The unique identifier of each manhole cover is a unique code set at the factory, which ensures that the classification results (heavy traffic type, regular road type, slow traffic area type) accurately correspond to the historical monitoring data of that manhole cover, avoiding data confusion between different manhole covers. The data acquisition module continuously collects the vibration frequency, tilt angle, and vibration noise of the manhole cover at fixed intervals and records the collection timestamp. This data can fully reflect the operating status of the manhole cover at different times. By linking the classification results with this monitoring data through the unique identifier, each monitoring data can be clearly assigned to the corresponding category of manhole cover, providing a clear category label for subsequent screening, thereby generating a related dataset.

[0036] The normal range of the module's parameters is determined based on the early warning threshold. The associated dataset is verified, and periods where all three parameters are within the normal range for 12 consecutive collection cycles under the same unique manhole cover identifier, without data loss or interruption, are marked as normal operating periods. The early warning threshold for determining the normal range of the module's parameters is based on statistical analysis of a large amount of manhole cover monitoring data in normal operating condition. This accurately defines the reasonable fluctuation range of the three parameters during normal operation of the manhole cover and can serve as a standard for judging whether the manhole cover is normal. The verification period is 12 consecutive collection cycles because single-time data collection may be affected by accidental factors (such as temporarily passing heavy vehicles), resulting in brief fluctuations. Stable data from multiple consecutive cycles better reflects the true normal operating status of the manhole cover. Periods with missing or interrupted data are excluded because such data cannot fully reflect the manhole cover's operating status and may lead to misjudgment. Only periods where all three parameters are within the normal range and the data is continuous and complete can be confirmed as normal operating periods.

[0037] The datasets are grouped and associated based on the classification results. Within each group, monitoring data whose collection timestamps fall within the normal operating period of the equipment are retained, resulting in three categories of filtered datasets. Grouping by classification results means that the manhole cover data belonging to the heavy traffic type, regular road type, and slow traffic area type in the associated dataset are divided into three data groups, ensuring that each group of data corresponds to only a single category of manhole cover and avoiding the mixing of different categories of data from affecting subsequent processing. The collection timestamp can accurately locate the generation period of the monitoring data. By comparing the timestamp with the marked normal operating period of the equipment, data within the normal period is retained, and monitoring data from abnormal periods (such as parameters out of range or data interruption) can be removed. Finally, three categories of filtered datasets containing only the monitoring data of the normal operating status of each type of manhole cover are obtained, providing a clean data foundation for building a dedicated database.

[0038] A dedicated database containing specified information was created for each of the three classification and screening datasets. The corresponding data was imported, and a timed synchronization mechanism was set up to receive newly added normal data. An independent database was created for each classification and screening dataset because the normal operating parameters of different categories of manhole covers differ; independent storage avoids data cross-interference and facilitates subsequent operations such as threshold derivation for single categories of manhole covers. Importing the three classification and screening datasets into their respective databases ensures a one-to-one correspondence between data storage and category, achieving orderly data management. The timed synchronization mechanism allows the data acquisition module to continuously generate new manhole cover monitoring data; timed synchronization automatically filters the portion of new data that matches the normal operating period of the equipment and imports it into the dedicated classification database, ensuring that the data in the database is always the latest normal status data, providing timely support for subsequent threshold updates and other tasks.

[0039] The process involves retrieving monitoring data from a dedicated database of three types of manhole covers over a period of x months during normal periods; setting dedicated safety redundancy coefficients, finite element model mechanical parameters, and flatness analysis windows based on their installation environment and physical specifications; and reusing unified derivation logic and substituting the dedicated parameters to obtain the tilt angle, vibration frequency, and vibration noise threshold for each type of risk, in the order of manhole cover vibration and noise risk, manhole cover subsidence risk, and road surface flatness risk. The specific details include the following: Monitoring data from the past x months for heavy-duty traffic type, regular road type, and slow-traffic area type manhole covers were extracted from a dedicated classification database to generate various target datasets. The dedicated classification database independently stores the corresponding normal-period monitoring data for each manhole cover type, allowing direct location and extraction of specific data for each type, avoiding data mixing between different categories. Data from the past x months was chosen because recent monitoring data accurately reflects the current traffic environment (e.g., recent traffic flow changes, dynamics of surrounding vibration sources) and the current mechanical performance of the manhole covers themselves, preventing data from becoming outdated due to the passage of time and ensuring the timeliness and accuracy of the data used for threshold derivation. Generating various target datasets involves separately aggregating the extracted data for each type of manhole cover, providing a precise and clean data foundation for subsequent differentiated threshold derivation for different types of manhole covers. This ensures that each threshold derivation is based solely on the normal operation data of that specific type of manhole cover, avoiding interference from cross-category data.

[0040] Based on the installation environment and physical specifications of the three types of manhole covers, specific safety redundancy coefficients, finite element model mechanical parameters, and road surface smoothness analysis time windows are set. The installation environment of heavy-duty traffic manhole covers is "main roads + high daily traffic volume," and the physical specifications are "heavy load-bearing + high-strength material (such as ductile iron)." These covers are subjected to frequent impacts and strong vibrations from heavy vehicles, therefore a high safety redundancy coefficient is required to reserve sufficient risk buffer space. The finite element model mechanical parameters need to match the compressive and flexural strength of the high-strength material to simulate the actual stress state. The road surface smoothness analysis time window is set to a longer duration to cover parameter fluctuations during periods of high traffic volume. The installation environment of conventional road manhole covers is "minor roads + daily traffic volume," and the physical specifications are "standard load-bearing + conventional material." (e.g., reinforced concrete)”, the load and vibration it bears are between heavy load and slow traffic area, so the safety redundancy coefficient, finite element model mechanical parameters, and flatness analysis time window are all set to medium level; the installation environment attribute of slow traffic area type manhole cover is "sidewalk + low daily traffic flow", and the physical specification attribute is "lightweight load-bearing + lightweight material (e.g. composite material)", the load and vibration it bears are the smallest, so the safety redundancy coefficient is the lowest, the finite element model mechanical parameters match the performance of lightweight material, and the flatness analysis time window is the shortest to adapt to the low frequency parameter changes of pedestrian traffic. The reason for setting such exclusive parameters is that there are essential differences in the external load strength and risk resistance of different types of manhole covers. Exclusive parameters can ensure that the subsequent threshold derivation fits the actual working conditions of each type of manhole cover, and avoid the threshold being too high or too low due to uniform parameters.

[0041] Following the order of vibration and noise risk, manhole cover subsidence risk, and road surface unevenness risk, the data preprocessing, finite element simulation, and spatiotemporal feature analysis logic were reused. Specific parameters were then input to obtain the tilt angle, vibration frequency, and vibration and noise thresholds for each risk category. Reusing the unified logic ensures consistency in the threshold derivation methods for the three types of manhole covers under the same risk dimension, guaranteeing the comparability and reliability of thresholds for different types of manhole covers and avoiding threshold deviations due to methodological differences. Data preprocessing logic was used for vibration and noise risk because short-term random fluctuations (such as instantaneous vibrations caused by passing small vehicles) may exist in monitoring data during normal periods. Data preprocessing can eliminate these fluctuations, highlighting the true correlation between vibration and noise and vibration frequency and tilt angle. After inputting a specific safety redundancy coefficient, a reasonable noise risk threshold can be determined based on the normal data range. Finite element simulation logic was used for manhole cover subsidence risk because... Since manhole cover subsidence is essentially a structural deformation problem, finite element simulation can be used to construct a mechanical model of the manhole cover and simulate the stress distribution at different tilt angles. When the stress reaches the material limit, the corresponding parameter is the critical risk value. After substituting the mechanical parameters of the dedicated finite element model, the material properties of each type of manhole cover can be accurately matched, and the subsidence risk threshold that conforms to the actual load-bearing capacity can be derived. For road surface smoothness risk, a spatiotemporal feature analysis logic is adopted because road surface smoothness needs to be reflected by parameter fluctuations over a period of time (such as angle and frequency changes caused by continuous pedestrian traffic or small vehicle passage). Data is divided according to a dedicated road surface smoothness analysis time window, and the parameter fluctuation characteristics within the window can be calculated. After associating with municipal smoothness standards, the risk threshold can be determined. After substituting into the dedicated time window, it can be adapted to the usage scenarios of each type of manhole cover (such as periods of high traffic or sparse pedestrian traffic), and the threshold that fits the actual smoothness requirements can be derived.

[0042] A comprehensive mapping table is created, summarizing the three thresholds: manhole cover category, risk type, and associated attributes. This table, after being labeled with relevant attributes, is stored in the threshold database of the early warning threshold determination module for use by the tiered early warning signal triggering module. The summarized mapping table organizes the tilt angle, vibration frequency, and vibration noise thresholds for each type of manhole cover corresponding to the three risk categories according to the "category-risk" dimension, ensuring a clear and unambiguous correspondence between thresholds and categories / risks, avoiding threshold confusion. The associated attribute labels are used to trace the manhole cover's installation environment and physical specifications corresponding to the thresholds. When manhole cover attributes change (such as material replacement), the corresponding threshold can be quickly located and updated. Storing the thresholds in the early warning threshold determination module's database ensures the timeliness and accuracy of threshold retrieval, providing stable threshold support for subsequent tiered early warnings, as the tiered early warning signal triggering module needs to retrieve the corresponding category and risk thresholds for comparison in real time when determining manhole cover risks.

[0043] In the graded early warning signal triggering module, the process of determining whether the manhole cover has a risk warning for vibration and noise, a risk warning for sinking, or a risk warning for road surface unevenness is as follows: The process involves obtaining the target manhole cover's type and then determining the tilt angle threshold, vibration frequency threshold, and vibration noise threshold for vibration and noise risk warnings, manhole cover subsidence risk warnings, or road surface unevenness risk warnings for that type. Specifically, this is achieved by first linking the target manhole cover's unique identifier (such as its factory serial number or installation serial number) to a manhole cover category record stored during the initial classification phase. This record clearly identifies each manhole cover as belonging to the heavy-duty traffic type, regular road type, or slow-traffic area type. Because the unique identifier of each manhole cover is permanently bound to its corresponding category information after classification, the specific type of the target manhole cover can be accurately located and extracted through the unique identifier, avoiding category matching errors. This process yields the corresponding three... The risk threshold is defined because the threshold database of the early warning threshold determination module stores data according to a fixed correspondence of "manhole cover category - risk type - threshold". That is, each type of manhole cover has its own exclusive threshold for vibration and noise risk, manhole cover subsidence risk, and road surface unevenness risk. After the target manhole cover type is determined, it is only necessary to retrieve the tilt angle threshold, vibration frequency threshold, and vibration and noise threshold corresponding to the three types of risks under that type from the database. This is because the threshold derivation process in the early stage has formulated exclusive thresholds that fit the actual working conditions of different types of manhole covers based on their installation environment and physical characteristics. There is a unique matching relationship between type and threshold, so the accurate corresponding threshold can be obtained directly through type.

[0044] The system acquires the vibration frequency, vibration noise, and tilt angle of the target manhole cover in real time. If any of these parameters exceeds the corresponding threshold, a risk warning is triggered. Real-time parameter acquisition is achieved through dedicated sensors installed on the target manhole cover monitoring terminal: vibration sensors continuously capture the vibration frequency generated by external forces (such as vehicle traffic or pedestrian trampling); noise sensors collect the vibration noise generated by friction or vibration between the manhole cover and its base; and tilt sensors monitor the tilt angle of the manhole cover relative to the horizontal road surface. These sensors transmit the collected parameters to the tiered warning signal triggering module at preset intervals. Because sensors directly perceive changes in the physical state of the manhole cover, they are the core components for acquiring real-time operational data, ensuring the timeliness and accuracy of parameter acquisition. The logic for determining a risk warning involves comparing the real-time parameters with corresponding thresholds one by one: if the vibration frequency, vibration noise, or tilt angle exceeds the threshold, a risk warning is triggered. If the vibration frequency exceeds the vibration frequency threshold corresponding to the risk of vibration and noise, manhole cover subsidence, or road surface unevenness, it indicates that the current vibration intensity of the manhole cover has exceeded the safe range for that risk. If the vibration and noise exceed the vibration and noise threshold corresponding to the risk of vibration and noise, it indicates that the noise generated by the manhole cover has reached the level of interference that requires warning. If the tilt angle exceeds the tilt angle threshold corresponding to the risk of manhole cover subsidence or road surface unevenness, it means that the manhole cover may have a tendency to subside or the road surface unevenness is not up to standard. Since each parameter directly corresponds to the core characteristics of a certain type of risk, if any parameter exceeds the threshold, it means that the risk condition of that parameter has been met. If a warning is not issued in time, it may lead to the expansion of the risk (such as the tilt angle exceeding the threshold may cause the manhole cover to fall off). Therefore, as long as any parameter exceeds the corresponding threshold, it is determined that the manhole cover has issued a warning for the risk type of that parameter, ensuring that the risk can be identified in a timely manner without omission.

[0045] The manhole cover abnormal displacement monitoring system integrates a triaxial accelerometer into the manhole cover monitoring terminal to collect displacement and attitude signals in real time. Edge nodes compare these real-time signals with the equipment baseline model, identifying abnormal displacement based on temporal changes and triggering tiered early warnings. The water immersion monitoring function uses conductivity and relative humidity sensors to detect sudden changes in liquid level and humidity within the manhole cover cavity. Edge nodes instantly mark water immersion events and generate maintenance work orders. Abnormal displacement and water immersion events are handled in a coordinated manner, with map markers updated synchronously and maintenance notifications pushed out.

[0046] In the graded early warning signal triggering module, the smart manhole cover is given graded early warnings based on the judgment result. The specific method for triggering different levels of early warning signals is as follows: If any of the vibration frequency, vibration noise, or tilt angle exceeds the corresponding threshold, the manhole cover is determined to be a primary risk warning. If any two of the vibration frequency, vibration noise, or tilt angle exceed the corresponding threshold, the manhole cover is judged to be at a medium-risk warning level. If the vibration frequency, vibration noise, and tilt angle all exceed the corresponding thresholds, the manhole cover is judged to be a serious risk warning.

[0047] The core logic of the entire graded early warning process is that the number of parameters exceeding the standard is positively correlated with the severity of the risk. This is because the vibration frequency, vibration noise, and tilt angle of the manhole cover correspond to different key dimensions of its operating status. An abnormality in a single dimension only represents a minor local problem, while an abnormality in multiple dimensions reflects a deterioration in the overall status. By statistically analyzing the number of parameters exceeding the standard to classify the early warning level, we can ensure that the early warning level is accurately matched with the actual degree of risk and hazard. This avoids both the waste of resources caused by over-responding to minor risks and the safety accidents caused by delayed responses to serious risks.

[0048] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A real-time monitoring and intelligent early warning system for manhole cover status based on the Internet of Things, characterized in that, Includes the following steps: The data acquisition module is used to integrate multiple sensors into the manhole cover monitoring terminal to synchronously collect equipment operating status data, including: vibration frequency, vibration noise, and tilt angle. The early warning equipment marking module is used to mark the location of the manhole cover on the map according to the equipment's operating status, and to identify manhole covers with normal equipment status. The manhole cover fault pre-diagnosis module is used to analyze the time-series changes in the long-term operating status of manhole covers, uncover the gradual abnormal characteristics before the occurrence of faults, and pre-diagnose and mark abnormal manhole covers with signs of faults in real time. The warning threshold determination module is used to classify manhole covers with normal equipment status and determine the tilt angle threshold, vibration frequency threshold, and vibration noise threshold for different types of manhole covers, such as manhole cover vibration and noise risk, manhole cover sinking risk, or road surface unevenness risk. The graded early warning signal triggering module is used to determine whether the abnormal manhole cover is at risk of vibration and noise, sinking, or road surface unevenness. Based on the determination result, the module provides graded early warnings for the smart manhole cover and triggers different levels of early warning signals accordingly. It also monitors abnormal displacement and water immersion of the manhole cover. The manhole cover alarm module is used to push early warning information and maintenance work orders based on the triggering of different levels of early warning signals, and to notify staff through corresponding alarm methods.

2. The IoT-based real-time monitoring and intelligent early warning system for manhole cover status as described in claim 1, characterized in that, In the early warning device marking module, the process of marking the location of the manhole cover on the map according to the device's operating status is as follows: The early warning device marking module receives real-time device operation status data output by the data acquisition module, as well as sensor heartbeat data and the real-time upload status of the acquired data; If the equipment operating status data is within the normal range defined for equipment operating status data and the sensor heartbeat data and collected data are uploaded in real time without interruption, the corresponding manhole cover will be marked with a blue label on the map. If the real-time upload of sensor heartbeat data or acquired data is interrupted, the module will trigger a manual opening and inspection process: When a man performs the opening operation manually, if the device simultaneously uploads the collected data, the corresponding manhole cover will be marked with an orange label on the map; If the equipment still fails to upload heart rate data and collected data after the manual opening operation is completed, it is determined that the equipment is in an abnormal state, and the corresponding manhole cover will be marked with a red mark on the map.

3. The IoT-based real-time monitoring and intelligent early warning system for manhole cover status according to claim 1, characterized in that, In the manhole cover fault pre-diagnosis module, the process of real-time pre-diagnosis and marking of abnormal manhole covers with early signs of fault is as follows: Collect historical equipment operation data of the target manhole cover monitoring terminal for n consecutive days, define the fault precursor period by associating it with the fault records of the same period, associate the historical operation data with the precursor label and label positive and negative samples, where n is a preset value; The daily rate of change of the equipment operating status data and the difference between the trend of the set sliding window are calculated and fused into a time-series progressive feature set. A fault precursor recognition model for a single manhole cover is obtained by using a temporal convolutional network as the architecture, training with a temporal progressive feature set and precursor labels, and optimizing through backpropagation. The system collects real-time data on the equipment's operating status from the manhole cover monitoring terminal to generate a time-series progressive feature. This feature is then input into a fault precursor identification model to obtain a probability value. If the probability value meets the criteria, the manhole cover is marked as having a fault precursor.

4. The IoT-based real-time monitoring and intelligent early warning system for manhole cover status according to claim 1, characterized in that, In the early warning threshold determination module, the process of determining the tilt angle threshold, vibration frequency threshold, and vibration noise threshold for the risk of manhole cover vibration noise, the risk of manhole cover sinking, or the risk of road surface unevenness is as follows: Extract the installation environment attributes and physical specification attributes of manhole covers in normal condition to form a classification raw dataset; Based on the original dataset for classification, a hierarchical clustering algorithm combined with domain rules is used to classify manhole covers into heavy traffic type, regular road type, and slow traffic area type. Based on the classification results, filter the historical monitoring data of vibration frequency, tilt angle, and vibration noise during the normal operating period of the equipment under the corresponding category, and build a category-specific database. Data from the past x months of normal monitoring periods for different types of manhole covers were collected from a dedicated database. Based on the installation environment and physical specifications of the manhole covers, dedicated safety redundancy coefficients, finite element model mechanical parameters, and flatness analysis windows were set. The parameters were then substituted into the unified derivation logic in the order of manhole cover vibration and noise risk, manhole cover subsidence risk, and road surface flatness risk to obtain the tilt angle, vibration frequency, and vibration and noise threshold for each type of risk, where x is a preset value.

5. The IoT-based real-time monitoring and intelligent early warning system for manhole cover status according to claim 4, characterized in that, The process of classifying manhole covers into heavy-duty traffic type, regular road type, and slow-traffic zone type based on the original classification dataset and using a hierarchical clustering algorithm combined with domain rules is as follows: By connecting with urban geographic information systems, traffic flow monitoring databases, and vibration source mapping systems, the road type, average daily traffic volume level, and surrounding vibration source intensity values ​​of manhole covers are extracted. Access the manhole cover production archive database to extract the load-bearing capacity, manhole cover material, and manhole cover diameter; By associating the installation environment attributes with the physical specification attributes, removing samples with missing values, performing one-hot encoding on non-numerical attributes and standardization on numerical attributes, a classification raw dataset is generated. Using the original dataset for classification as input, the Ward clustering criterion and Euclidean distance metric are adopted, and the clustering termination condition is set as the maximum intra-class distance threshold to obtain the initial classification clusters; Based on the urban manhole cover operation and maintenance standards, the initial clusters that meet the requirements of main roads + heavy load-bearing capacity + high daily traffic volume, secondary roads + standard load-bearing capacity + medium daily traffic volume, and sidewalks + light load-bearing capacity + low daily traffic volume are respectively divided into heavy traffic type, regular road type, and slow traffic area type. The remaining samples are redistributed according to the weighted similarity of installation environment attributes and physical specification attributes to determine the final classification.

6. The IoT-based real-time monitoring and intelligent early warning system for manhole cover status according to claim 4, characterized in that, The process of filtering historical monitoring data on vibration frequency, tilt angle, and vibration noise during normal operating periods of equipment under the corresponding category according to the classification results, and constructing a category-specific database is as follows: By using the unique identifier of the manhole cover, the classification results of the manhole cover are linked with the historical monitoring data stored in the data acquisition module, including vibration frequency, tilt angle, vibration noise, and acquisition timestamp, to generate a linked dataset; Based on the warning threshold, determine the normal range of the module's parameters, verify the associated dataset, and select the period when the three parameters under the same manhole cover's unique identifier are within the normal range for 12 consecutive collection cycles without data loss or interruption, and mark it as the period when the equipment is operating normally. The datasets are grouped and associated according to the classification results of manhole covers. In each group, the monitoring data collected during the normal operation period of the equipment are retained to obtain different types of manhole cover classification and filtering datasets. Create a dedicated database with specified fields for different types of manhole covers to classify and filter datasets, import the corresponding data, and set up a timed synchronization mechanism to receive newly added normal data.

7. The IoT-based real-time monitoring and intelligent early warning system for manhole cover status according to claim 4, characterized in that, The process of obtaining the tilt angle, vibration frequency, and vibration noise threshold for various risks is as follows: Data from nearly x months of normal-time monitoring data of heavy-duty traffic type, regular road type, and slow-traffic area type manhole covers were collected from a classification-specific database to generate various target datasets. Based on the installation environment and physical specifications of different types of manhole covers, a dedicated safety redundancy coefficient, finite element model mechanical parameters, and road surface smoothness analysis time window are set. The data preprocessing, finite element simulation, and spatiotemporal feature analysis logic were reused in the order of vibration and noise risk, manhole cover sinking risk, and road surface smoothness risk. The tilt angle, vibration frequency, and vibration and noise threshold of each type of risk were obtained by substituting specific parameters. The data is compiled into a three-threshold mapping table of manhole cover category, risk type, and associated attribute identification, and stored in the threshold database of the early warning threshold determination module for use by the graded early warning signal triggering module.

8. The IoT-based real-time monitoring and intelligent early warning system for manhole cover status according to claim 1, characterized in that, In the graded early warning signal triggering module, the process of determining whether the abnormal manhole cover, which is a precursor to a fault, poses a risk of vibration and noise, a risk of sinking, or a risk of road surface unevenness is as follows: Obtain the type of the target manhole cover, and then obtain the tilt angle threshold, vibration frequency threshold, and vibration noise threshold for the vibration and noise risk of the target manhole cover in that type. Thresholds for tilt angle, vibration frequency, and vibration noise of manhole cover subsidence risk; In addition, the tilt angle threshold, vibration frequency threshold, and vibration noise threshold for road surface unevenness risk; The vibration frequency, vibration noise, and tilt angle of the target manhole cover are acquired in real time. If any of these parameters exceeds the corresponding threshold, a corresponding risk warning is issued for the manhole cover.

9. The IoT-based real-time monitoring and intelligent early warning system for manhole cover status according to claim 8, characterized in that, The specific method for issuing graded early warnings for smart manhole covers based on the judgment results, and for triggering different levels of early warning signals accordingly, is as follows: If any of the vibration frequency, vibration noise, or tilt angle exceeds the corresponding threshold, the manhole cover is determined to be a primary risk warning. If any two of the vibration frequency, vibration noise, or tilt angle exceed the corresponding threshold, the manhole cover is judged to be at a medium-risk warning level. If the vibration frequency, vibration noise, and tilt angle all exceed the corresponding thresholds, the manhole cover is judged to be a serious risk warning.

Images