Abnormity identification method for drainage pipe network monitoring equipment

Abstract

The invention relates to the technical field of anomaly recognition, and discloses a drainage pipe network monitoring equipment anomaly recognition method, which comprises the steps of dividing all to-be-recognized monitoring equipment into a plurality of monitoring groups based on a pipeline topological structure of a drainage pipe network and mounting positions of all the to-be-recognized monitoring equipment; monitoring data of to-be-identified monitoring equipment is acquired in real time, and an acquired data slice sub-table is constructed; calculating an abnormal jump influence factor, an abnormal value influence factor and a fixed abnormal influence factor of each collected data slice sub-table, constructing a comprehensive influence factor, correcting an initial health coefficient of the to-be-identified monitoring equipment, and obtaining a current health coefficient; and obtaining the total number of the to-be-identified monitoring devices of which the current health coefficients are not located in the health coefficient allowable range in each monitoring group, and if the total number is smaller than a preset threshold value, determining that the to-be-identified monitoring devices of which the current health coefficients are not located in the health coefficient allowable range are abnormal devices.

Description

Technical Field

[0001] The present invention relates to the technical field of anomaly recognition, and particularly to a method for recognizing anomalies in drainage network monitoring equipment. Background Art

[0002] In the current intelligent construction of urban drainage systems, the deployment of drainage monitoring equipment such as level gauges, flow meters, water quality and level integrated machines, etc. has become the norm, aiming to support the efficient operation of drainage services through real-time data collection.

[0003] Currently, most drainage systems lack effective means for recognizing the abnormal states of monitoring equipment. The data collected by monitoring equipment often contains noise, errors, or distortion, and cannot effectively support the analysis and decision-making of downstream services. When equipment fails, deviates, or has abnormal data, the system cannot capture and respond in a timely manner, resulting in data deviating from the actual situation for a long time, thereby affecting the accuracy and timeliness of services. In this case, even if the equipment has been put into use, it is difficult to effectively confirm whether it is operating normally and whether useful data has been generated, resulting in a waste of resources and inefficiency of services.

[0004] Currently, the commonly used methods for abnormal identification of monitoring devices include manual curve viewing, cut-off point method, filtering method, three-fold standard method, and sliding window method. The method of manual curve viewing requires technicians to identify abnormal fluctuations or mutations by observing the real-time curve of device data. This method relies on experienced technicians, but for large-scale and complex systems, the efficiency and accuracy of manual identification are relatively low, and it cannot achieve real-time response. The cut-off point method sets a fixed threshold. When the monitored data exceeds or is lower than this threshold, it is judged as abnormal. However, due to the complexity of the drainage system environment, the fixed threshold cannot cope with dynamic changes. For example, during rainfall, the liquid level in the drainage system will naturally change significantly. At this time, the cut-off point method is prone to misjudging normal mutations as abnormal, resulting in a high false alarm rate. The filtering method eliminates noise and short-term fluctuations by smoothing the data curve to identify persistent abnormalities. Although the filtering method can effectively remove short-term interference, in the face of sudden events in the drainage system, such as unstable sensor interfaces or long-term abnormalities caused by the probe being entangled by garbage, the filtering may ignore important abnormal signals, resulting in inaccurate detection. The three-fold standard method is based on the principle of three standard deviations in statistics and considers data outside the normal range as abnormal. However, the distribution of drainage monitoring data often does not conform to the normal distribution, and coupled with the complexity of environmental changes, the three-fold standard method is prone to large deviations when judging abnormalities and cannot effectively cope with non-normal distribution data. The sliding window method analyzes the data change trend within the window by setting a time window to identify abnormalities. The sliding window method is suitable for real-time data monitoring, but the setting of the window size has a great impact on the detection result, and it is difficult to flexibly adjust in different scenarios. Especially when facing sudden long-term abnormalities (such as the probe being entangled by foreign objects), the sliding window method may not be able to identify in time or misjudge as normal fluctuations.

[0005] In summary, due to the complex and changeable actual working environment of the drainage pipe network, the data collected by monitoring devices often contain noise, errors, or distortion. Existing methods for abnormal identification of monitoring devices are difficult to comprehensively identify complex abnormal types in the actual working environment of the drainage pipe network when facing dynamic changes and non-normal distribution of monitoring data. When monitoring data is abnormal, it is also impossible to accurately judge whether it is a natural change in the liquid level in the drainage pipe network or an abnormality of the monitoring device, which is prone to misjudgment or missed judgment, and thus leads to poor accuracy and timeliness of drainage pipe network monitoring. Summary of the Invention

[0006] Therefore, the technical problem to be solved by the present invention is to overcome the problem in the prior art that abnormal devices in the drainage pipe network cannot be identified in a timely and effective manner.

[0007] To solve the above technical problem, the present invention provides a method for abnormal identification of drainage pipe network monitoring devices, including: Based on the pipeline topology of the drainage pipeline network and the installation locations of all monitoring devices to be identified therein, all the monitoring devices to be identified are divided into multiple monitoring groups; Based on a preset data acquisition frequency, the monitoring data of the monitoring devices to be identified is obtained in real time; For the monitoring data at the sampling moment, obtain the monitoring data within the sampling moment range to form a sub-table of the acquisition data slice; Based on the local fluctuation coefficient, standard deviation and interval level standard line of each monitoring data and its corresponding sub-table of the acquisition data slice, the deviation degree of each monitoring data is obtained; Based on the deviation degree of each monitoring data and a preset deviation threshold, the abnormal jump impact factor of each monitoring data is calculated; Sum up the abnormal jump impact factors of all the monitoring data in the sub-table of the acquisition data slice to obtain the abnormal jump impact factor of the sub-table of the acquisition data slice; Obtain the total number of data that exceed the fixed normal range interval among the monitoring data in the sub-table of the acquisition data slice and the total number of data in the sub-table of the acquisition data slice, and calculate the non-normal value impact factor of the sub-table of the acquisition data slice; Obtain the duration of the fixed value abnormal segment among the monitoring data in the sub-table of the acquisition data slice and the total time of the sliced data, and calculate the fixed abnormal impact factor of the sub-table of the acquisition data slice at the current moment; Based on the abnormal jump impact factor, non-normal value impact factor and fixed abnormal impact factor of the sub-table of the acquisition data slice, a comprehensive impact factor is constructed to correct the initial health coefficient of the monitoring device to be identified, and the current health coefficient of the monitoring device to be identified is obtained; Obtain the total number of monitoring devices to be identified whose current health coefficients are not within the allowable range of the health coefficient in each monitoring group. If the total number is less than the preset threshold, determine that the monitoring devices to be identified whose current health coefficients are not within the allowable range of the health coefficient are abnormal devices.

[0008] Preferably, based on the season, the spatial type of the pipeline and the business type, the length of the corresponding sub-table of the acquisition data slice is constructed, including: Initialize the length of the data slice sub-table to be ; Set corresponding seasonal adjustment coefficients for the dry season, rainy season and transition season ; Set corresponding adjustment coefficients for the spatial type of the pipeline for industrial areas, commercial areas and residential areas ; Set corresponding adjustment coefficients for the business type for daily operation and maintenance, leakage analysis, siltation analysis, flood control warning and equipment calibration services ; Based on the seasonal adjustment coefficient, the space type adjustment coefficient of the pipeline, and the service type adjustment coefficient corresponding to the current moment, adjust the length of the initial data slice sub-table to obtain the length of the collected data slice sub-table corresponding to the current moment, denoted as .

[0009] Preferably, based on the local fluctuation coefficient, the standard deviation, and the interval level standard line of each monitoring data and its corresponding collected data slice sub-table, obtain the deviation degree of each monitoring data, including: Based on the local weighted average value of the monitoring data in the collected data slice sub-table corresponding to the monitoring data at the th sampling moment, obtain the interval level standard line of the monitoring data at the th sampling moment , denoted as: ; Calculate the first local fluctuation coefficient of the monitoring data at the th sampling moment, denoted as: ; Calculate the second local fluctuation coefficient of the monitoring data at the th sampling moment, denoted as: ; Calculate the standard deviation of the monitoring data at the th sampling moment, denoted as: ; Based on the first local fluctuation coefficient, the second local fluctuation coefficient, and the standard deviation, calculate the deviation index of the monitoring data at the th sampling moment, denoted as: ; Calculate the ratio of the absolute value of the difference between the monitoring data at the th sampling moment and the interval level standard line to the deviation index, to obtain the deviation degree of the monitoring data at the th sampling moment, denoted as: ; where represents the weight of the monitoring data at the th sampling moment, and the expression is , represents the attenuation coefficient; is the monitoring data at the th sampling moment, representing the true monitoring horizontal line; Represents the arithmetic mean of the acquisition data slice sub - table of the monitoring data at the sampling moment, and the expression is , represents the total number of data in the acquisition data slice sub - table; and respectively represent the preset weights of the first local fluctuation coefficient and the second local fluctuation coefficient.

[0010] Preferably, sum the abnormal jump impact factors of all the monitoring data in the acquisition data slice sub - table to obtain the abnormal jump impact factor of the acquisition data slice sub - table, including: Based on the deviation degree of the monitoring data at the sampling moment and the preset deviation threshold, calculate the abnormal jump impact factor of the monitoring data at the sampling moment , which is expressed as: ; Sum the abnormal jump impact factors of the monitoring data in the acquisition data slice sub - table to obtain the abnormal jump impact factor of the acquisition data slice sub - table , which is expressed as: ; Among them, represents the abnormal jump weight, represents the preset deviation threshold.

[0011] Preferably, obtain the total number of data in the monitoring data of the acquisition data slice sub - table that exceeds the fixed normal range interval, and the total number of data in the acquisition data slice sub - table, and calculate the non - normal value impact factor of the acquisition data slice sub - table, which is expressed as: ; Among them, represents the non - normal value impact factor of the acquisition data slice sub - table, represents the non - normal value weight, represents the total number of data in the monitoring data of the acquisition data slice sub - table that exceeds the fixed normal range interval.

[0012] Preferably, obtain the duration of the fixed - value abnormal segment in the monitoring data of the acquisition data slice sub - table, and the total time of the slice data, and calculate the fixed - anomaly impact factor of the acquisition data slice sub - table at the current moment, which is expressed as: ; Among them, represents the fixed - anomaly impact factor of the acquisition data slice sub - table, represents the fixed - anomaly weight, Indicates the duration of the fixed-value abnormal segment in the monitoring data of the collected data slice sub-table. Indicates the total time of the slice data.

[0013] Preferably, based on the abnormal jump impact factor, non-normal value impact factor, and fixed abnormal impact factor of the collected data slice sub-table, a comprehensive impact factor is constructed to correct the initial health coefficient of the monitoring device to be identified, and the current health coefficient of the monitoring device to be identified is obtained, including: Based on the abnormal jump impact factor, non-normal value impact factor, and fixed abnormal impact factor of the collected data slice sub-table, a comprehensive impact factor is constructed , which is expressed as: ; Using the comprehensive impact factor to correct the initial health coefficient of the monitoring device to be identified to obtain the current health coefficient of the monitoring device to be identified , which is expressed as: .

[0014] Preferably, obtaining the current health coefficient of the monitoring device to be identified further includes: If the current health coefficient of the monitoring device to be identified is within the allowable range of the health coefficient, it is determined that the monitoring device to be identified at the current moment is not an abnormal device, and the sampling moment is set, and based on the monitoring data at the next sampling moment, the monitoring device to be identified at the next moment is identified.

[0015] Preferably, after determining that the monitoring device to be identified is an abnormal device, it further includes: Trigger the monitoring device to be identified to initiate an abnormal alarm, generate an operation and maintenance work order, and send it to the control center; After the operation and maintenance work order is processed, recalculate the current health coefficient of the monitoring device to be identified until it is detected that the current health coefficient of the monitoring device to be identified is within the allowable range of the health coefficient, and then the monitoring device to be identified is set to the service-supported state.

[0016] Preferably, after determining that the monitoring device to be identified is an abnormal device, it further includes: Reduce the preset data collection frequency of the monitoring device to be identified to the collection frequency threshold until the operation and maintenance work order is processed, and after the device enters the service-supported state, restore the preset data collection frequency.

[0017] The above technical solutions of the present invention have the following beneficial effects compared with the prior art: The method for identifying anomalies in drainage network monitoring equipment according to the present invention collects the monitoring data of the monitoring equipment to be identified in real time, calculates the interval level standard line and deviation index of each monitoring data based on the local data before and after each monitoring data, and captures the abnormal jumps of the monitoring data in real time. At the same time, combined with abnormal non-normal values and fixed-value abnormal segments, a multi-level anomaly marking mechanism is used to comprehensively capture different types of abnormal signals, avoid missed reports and false alarms, and improve the accuracy of anomaly identification. Moreover, through multi-dimensional data analysis and dynamic adjustment, combined with the real-time correction of the health coefficient of the monitoring equipment to be identified, it better adapts to the complex and changeable monitoring environment and the non-normal distribution of monitoring data, improving the accuracy and robustness of anomaly identification for the monitoring equipment to be identified.

[0018] Based on the season and business type at the time of identifying the monitoring equipment to be identified, the present invention adjusts the length of the sliced sub-table of the collected data, realizes multi-level data slicing analysis, and can adjust the window size and analysis range on different time scales, so as to accurately identify long-term anomalies and avoid the misjudgment risk brought by fixed windows. At the same time, preset deviation thresholds and fixed normal range intervals can both be adjusted according to the working environment and business type of the monitoring equipment to be identified, effectively avoiding the misjudgment problem caused by fixed thresholds and ensuring that anomalies in the monitoring equipment can still be accurately and timely identified in a complex and changeable working environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to make the content of the present invention easier to be clearly understood, the following further details the present invention according to specific embodiments of the present invention in conjunction with the drawings, where: Figure 1 is a flowchart of the steps of the method for identifying anomalies in drainage network monitoring equipment provided by the present invention; Figure 2 is a flowchart for calculating the deviation degree of monitoring data. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] The following further illustrates the present invention in conjunction with the drawings and specific embodiments, so that those skilled in the art can better understand the present invention and be able to implement it, but the embodiments cited do not limit the present invention.

[0021] Refer to Figure 1 As shown, the flowchart of the steps of the method for identifying anomalies in drainage network monitoring equipment provided by the present invention specifically includes the following steps: S101: Based on the pipeline topology of the drainage network and the installation positions of all monitoring equipment to be identified therein, all monitoring equipment to be identified is divided into multiple monitoring groups; S102: Based on a preset data collection frequency, the monitoring data of the monitoring equipment to be identified is obtained in real time; S103: For the Monitoring data at the sampling moment is obtained The monitoring data within the sampling moment range forms a sub-table of the acquisition data slice S104: Based on each monitoring data and the local fluctuation coefficient, standard deviation, and interval level standard line of its corresponding acquisition data slice sub-table, the deviation degree of each monitoring data is obtained S105: Based on the deviation degree of each monitoring data and a preset deviation threshold, the abnormal jump impact factor of each monitoring data is calculated S106: The abnormal jump impact factors of all monitoring data in the acquisition data slice sub-table are summed to obtain the abnormal jump impact factor of the acquisition data slice sub-table S107: The total number of data that exceeds the fixed normal range interval among the monitoring data in the acquisition data slice sub-table and the total number of data in the acquisition data slice sub-table are used to calculate the non-normal value impact factor of the acquisition data slice sub-table, expressed as: ; Wherein represents the non-normal value impact factor of the acquisition data slice sub-table represents the non-normal value weight represents the total number of data that exceeds the fixed normal range interval among the monitoring data in the acquisition data slice sub-table S108: The duration of the fixed value abnormal segment among the monitoring data in the acquisition data slice sub-table and the total time of the slice data are used to calculate the fixed abnormal impact factor of the acquisition data slice sub-table at the current moment, expressed as: ; Wherein represents the fixed abnormal impact factor of the acquisition data slice sub-table represents the fixed abnormal weight represents the duration of the fixed value abnormal segment among the monitoring data in the acquisition data slice sub-table represents the total time of the slice data S109: Based on the abnormal jump impact factor, non-normal value impact factor, and fixed abnormal impact factor of the acquisition data slice sub-table, a comprehensive impact factor is constructed to correct the initial health coefficient of the monitoring device to be identified, and the current health coefficient of the monitoring device to be identified is obtained, including: S109-1: Based on the abnormal jump impact factor, non-normal value impact factor, and fixed abnormal impact factor of the acquisition data slice sub-table, a comprehensive impact factor is constructed which is expressed as: ; S109-2: Use the comprehensive impact factor to correct the initial health coefficient Make corrections to obtain the current health coefficient of the monitoring device to be identified , expressed as: ; S110: Obtain the total number of monitoring devices to be identified whose current health coefficients are not within the allowable range of health coefficients in each monitoring group. If the total number is less than the preset threshold, determine that the monitoring devices to be identified whose current health coefficients are not within the allowable range of health coefficients are abnormal devices.

[0022] Specifically, if the total number of monitoring devices to be identified whose current health coefficients are not within the allowable range of health coefficients in the monitoring group is not less than the preset threshold, it is determined that the liquid level at the drainage pipe network where the monitoring group is located has undergone a natural change, rather than the monitoring device being abnormal.

[0023] Specifically, in the embodiments of the present invention, based on the season, the type of space where the pipeline is located, and the service type, the length of the corresponding sub-table of the collected data slices is constructed, including: Initialize the length of the sub-table of the data slices to ; Set corresponding seasonal adjustment coefficients for the dry season, rainy season, and transition season ; Set corresponding adjustment coefficients for the type of space where the pipeline is located for industrial areas, commercial areas, and residential areas ; Set corresponding adjustment coefficients for the service type for daily operation and maintenance, leakage analysis, blockage analysis, flood control warning, and equipment calibration services ; Based on the seasonal adjustment coefficient, the adjustment coefficient for the type of space where the pipeline is located, and the adjustment coefficient for the service type corresponding to the current moment, adjust the initial length of the sub-table of the data slices to obtain the length of the sub-table of the collected data slices corresponding to the current moment, expressed as .

[0024] The present invention adjusts the length of the sub-table of the collected data slices based on the season, the type of space where the pipeline is located, and the service type at the time of identifying the monitoring device to be identified, realizes multi-level data slice analysis, can adjust the window size and analysis range on different time scales, so as to accurately identify long-term anomalies and avoid the misjudgment risk brought by a fixed window. At the same time, the preset deviation threshold and the fixed normal range interval can both be adjusted according to the working environment and service type of the monitoring device to be identified, effectively avoiding the misjudgment problem brought by a fixed threshold, and ensuring that the anomalies of the monitoring device can be accurately and timely identified in a complex and changeable working environment.

[0025] Specifically, after determining that the monitoring device to be identified at the current moment is an abnormal device, it further includes: triggering the monitoring device to be identified to initiate an abnormal alarm, generating an operation and maintenance work order, and sending it to the control center; after the operation and maintenance work order is processed, recalculate the current health coefficient of the monitoring device to be identified until it is detected that the current health coefficient of the monitoring device to be identified is within the allowable range of the health coefficient, and then make the monitoring device to be identified enter the service-supported state.

[0026] Specifically, after determining that the monitoring device to be identified is an abnormal device, it further includes: reducing the preset data acquisition frequency of the monitoring device to be identified to the acquisition frequency threshold until the operation and maintenance work order is processed and the device enters the service-supported state, and then restoring the preset data acquisition frequency.

[0027] In this embodiment, by adjusting the data sampling frequency of the abnormal device, the operation cost of the monitoring device is reduced, and at the same time, the storage of unnecessary monitoring data is also reduced.

[0028] Specifically, if the current health coefficient of the monitoring device to be identified is within the allowable range of the health coefficient, it is determined that the monitoring device to be identified at the current moment is not an abnormal device, and let the sampling moment , and based on the monitoring data at the next sampling moment, identify the monitoring device to be identified at the next moment.

[0029] Refer to Figure 2 As shown, it is a flowchart for calculating the deviation degree of monitoring data; specifically, in step S104, based on each monitoring data and the local fluctuation coefficient, standard deviation, and interval level standard line of its corresponding acquisition data slice sub-table, obtain the deviation degree of each monitoring data, including: S104-1: Based on the local weighted average value of the monitoring data in the acquisition data slice sub-table corresponding to the monitoring data at the th sampling moment, obtain the interval level standard line of the monitoring data at the th sampling moment , which is expressed as: ; S104-2: Calculate the first local fluctuation coefficient of the monitoring data at the th sampling moment , which is expressed as: S104-3: Calculate the second local fluctuation coefficient of the monitoring data at the th sampling moment , which is expressed as: S104-4: Calculate the standard deviation of the monitoring data at the th sampling moment, which is expressed as: ; S104-5: Calculate the deviation index of the monitoring data at the sampling moment based on the first local fluctuation coefficient, the second local fluctuation coefficient, and the standard deviation , expressed as: ; S104-6: Calculate the ratio of the absolute value of the difference between the monitoring data at the sampling moment and the interval horizontal standard line to the deviation index, and obtain the deviation degree of the monitoring data at the sampling moment, expressed as: ; wherein, represents the weight of the monitoring data at the sampling moment, and the expression is , represents the attenuation coefficient; is the monitoring data at the sampling moment, representing the true monitoring horizontal line; represents the arithmetic mean of the data slice sub-table of the monitoring data at the sampling moment, and the expression is , represents the total number of data in the data slice sub-table; and respectively represent the preset weights of the first local fluctuation coefficient and the second local fluctuation coefficient.

[0030] Specifically, after calculating the abnormal jump influence factors of all the monitoring data in the data slice sub-table of the collected data in this embodiment, sum them up to obtain the abnormal jump influence factor of the data slice sub-table of the collected data, including: Based on the deviation degree of the monitoring data at the sampling moment and the preset deviation threshold, calculate the abnormal jump influence factor of the monitoring data at the sampling moment , expressed as: ; Sum up the abnormal jump influence factors of the monitoring data in the data slice sub-table of the collected data to obtain the abnormal jump influence factor of the data slice sub-table of the collected data , expressed as: ; wherein, represents the abnormal jump weight, represents the preset deviation threshold.

[0031] This embodiment adopts technical means of dynamically calculating and judging the device health coefficient, effectively solving the technical problems of the inability to respond to device anomalies in a timely manner and the lag in health status update in traditional methods, and thus realizing real-time alarm for the operating status of the device to be identified and monitored. Based on the dynamic adjustment mechanism, this embodiment adjusts the detection parameters according to real-time data and environmental changes, effectively avoiding the misjudgment problem caused by fixed thresholds and ensuring accurate anomaly identification in various complex environments. At the same time, a multi-level anomaly marking mechanism is introduced, including abnormal jump data points YE, non-normal value data points FZ, and fixed value abnormal segments FES, which can comprehensively capture different types of abnormal signals, avoid missed reports and false alarms, and thus improve the overall detection accuracy. Through multi-dimensional data analysis and dynamic adjustment, combined with the real-time correction of the device health coefficient, the present invention can better adapt to non-normal distributions and complex environmental changes, improving the accuracy and robustness of anomaly detection.

[0032] The monitoring devices described in the present invention include a liquid level gauge, a flow meter, a multi-parameter water quality sensor, a water pressure gauge, and a rain gauge. This embodiment can identify anomalies by combining the monitoring data of multiple different monitoring devices; for example, after using the liquid level gauge to identify the abnormal device in the monitoring group, the flow meter is used for secondary verification.

[0033] Based on the above embodiment, the embodiment of the present invention uses the above method for identifying anomalies in the drainage network monitoring device, and takes the liquid level gauge as the device to be identified and monitored for device anomaly identification, specifically including: S201: According to the service support requirements of the liquid level data, determine the initial data slice range n of the monitoring device, determine the data acquisition frequency m, determine the device anomaly analysis acquisition time range S, and set the allowable range ZG of the device health coefficient G; The initial data slice range n refers to the segmented processing of the monitoring data within a certain period of time, and each slice contains several data points; by slicing the data, it is possible to better analyze and process the changes in the liquid level data, identify abnormal situations, and selecting an appropriate slice range n helps to capture the abnormal behavior of the device on different time scales.

[0034] The setting of the slice range should be determined according to the service support requirements; for example, whether the data is used for early warning of overflow or for analyzing pipeline blockage; if the service requires close monitoring of the liquid level changes within a certain period of time, then n should be set smaller to more finely capture the changes; if the service focuses on long-term trend analysis, a larger n value can be selected.

[0035] Anomaly analysis needs to be based on data within a certain time period to judge the operating status of the device; the selection of the time range S determines the depth and breadth of the analysis and affects the ability to capture short-term and long-term anomalies. The acquisition time range should be set according to business requirements; for business scenarios with high real-time requirements, S should be shorter to quickly detect anomalies; for scenarios that require analyzing long-term trends, S can be set to a longer time period to capture the long-term behavior patterns of the device.

[0036] Specifically, in the embodiment of the present invention, based on the season, space type, and business type in which the device to be identified is currently operating, the length of the data slice sub-table is adjusted in real time to adapt to the requirements for the analysis granularity under different operating environments and business needs. This embodiment initializes the length of the data slice sub-table based on the default preset configuration to standardize the data processing processes of different devices.

[0037] Specifically, this embodiment sets the seasonal adjustment coefficients for the dry season, transition season, and rainy season to 0.8, 1.0, and 1.5 respectively; the value of the seasonal adjustment coefficient increases with the increase in the rainfall corresponding to each type of season. In the dry season with little hydraulic disturbance, a shorter time window corresponding to the length of the data slice sub-table is used to analyze weak changes; in the transition season, the state of the water conservancy system is unstable, and a relatively standard window is maintained; in the rainy season with large flow fluctuations, the window is expanded to capture stable patterns.

[0038] Similarly, for the adjustment coefficients of the space types where the pipelines are located in industrial areas, commercial areas, and residential areas , they are set to 1.2, 1.1, and 0.9 respectively; they are also set correspondingly based on the amount of water used in the space types where the pipelines are located. For the adjustment coefficients of business types , the adjustment coefficient for daily operation and maintenance is set to 1.0 to implement standard monitoring; the adjustment coefficient for leakage analysis services is set to 1.2 to observe stable small offsets; the adjustment coefficient for blockage analysis services is set to 1.5 to observe long-term cumulative change trends; the adjustment coefficient for flood control warning services is set to 0.8 to achieve high-frequency sampling and ensure the timeliness of warning responses; the adjustment coefficient for equipment calibration services is set to 0.5 to focus on short-term high-precision signals and achieve accurate calibration.

[0039] The health coefficient G is a comprehensive indicator used to evaluate the overall operating condition of the device. The initial value of G is usually 100 and will be dynamically adjusted according to the detected anomalies.

[0040] S202: Establish a device anomaly analysis database. The database includes the data acquisition targets of each device and the data slice sub-tables of S / (m·n). Each device includes a health coefficient G with an initial value set to 100; The device anomaly analysis database is a database specifically designed to store and process the data collected by monitoring devices. It not only supports the storage of raw data but also operations such as slicing, analyzing, and health assessment of the data. In the monitoring of the drainage system, devices often generate a large amount of data. A dedicated anomaly analysis database can ensure the systematic management of data, enabling the data to be accessed, analyzed, and updated efficiently. At the same time, the reasonable design of the database structure can improve the overall performance of the system, reduce data redundancy, and processing latency.

[0041] The data acquisition slice sub-table is a subset obtained by dividing the data in the parent table according to the time range and acquisition frequency. The slice sub-table is used for segmental analysis of the data, allowing the system to accurately detect anomalies within a smaller time range. The function of the slice sub-table is to slice the large-scale monitoring data by time so that the system can perform more refined anomaly analysis within a specific time period. This structure helps to improve the efficiency of the system in processing big data and can quickly locate and analyze abnormal behaviors within a specific time period.

[0042] S203: Retrieve the physical parameters of the device installation. According to the installation location and parameters such as well depth, calculate the fixed normal range interval Z; The fixed normal range interval Z is a numerical range set according to the physical parameters of the device, the installation environment, and historical data. This range defines the interval within which the liquid level data should be in the normal operating state of the device, usually represented in the form of upper and lower limit values.

[0043] S204: For each data in the data acquisition slice sub-table, perform dynamic calculation and judgment, and correct the device health coefficient G; Dynamic calculation and judgment refer to evaluating whether each data point in the data slice sub-table deviates from the expected normal range through real-time analysis, and determining whether there is an anomaly. The judgment logic is as follows: S204-1: Remove the data to be judged and calculate the interval horizontal standard line , expressed as: ; Among them, ; S204-2: Calculate the deviation index, including: Calculate the first local fluctuation coefficient, expressed as: ; Calculate the second local fluctuation coefficient, expressed as: ; Calculate the standard deviation, expressed as: ; Based on the first local fluctuation coefficient, the second local fluctuation coefficient, and the standard deviation, obtain the deviation index, expressed as: ; The first local fluctuation coefficient is used to measure the skewness of the data, the second local fluctuation coefficient is used to measure the kurtosis of the data, and the standard deviation is used to measure the dispersion degree of the data. Combining with the weight coefficient, the influence of the first local fluctuation coefficient and the second local fluctuation coefficient on the deviation index is adjusted to obtain the deviation index of the data.

[0044] S204-3: Determine the degree to which the judgment value deviates from the interval data level standard line; Based on the absolute value of the difference between the interval data level standard line and the true horizontal line and the deviation index, calculate the deviation degree, expressed as: ; If the deviation degree exceeds the allowable threshold, mark the monitoring data as an abnormal jump data point YE; S205: Determine whether each data in the sliced sub-table exceeds the set fixed normal range interval Z. If it exceeds, mark it as a non-normal value data point FZ; S206: Determine whether all the data in the sliced sub-table remains unchanged. If it remains unchanged, mark the sliced sub-table as a fixed value abnormal segment FES; S207: Update the slice range to n + a, where a is a preset judgment sliding interval, and continue to repeat step S202; S208: Calculate the device health coefficient G, including: S208-1: Calculate the abnormal jump impact factor, expressed as: ; ; Among them, is the abnormal jump weight, indicating the influence degree of the abnormal jump data point on the health coefficient; represents the abnormal judgment allowable threshold; S208-2: Calculate the non-normal value impact factor, expressed as: ; Among them, is the non-normal value weight, indicating the influence degree of the non-normal value data on the health coefficient; represents the total number of non-normal value data points, represents the data slice range; S208-3: Calculate the fixed abnormal impact factor, expressed as: ; Among them, is the fixed abnormal weight, indicating the influence degree of the fixed value abnormal segment on the health coefficient; represents the duration of the fixed value abnormal segment, represents the total time of the data slice; S208-4: Weighted sum the abnormal jump impact factor, non-normal value impact factor, and fixed abnormal impact factor to obtain the comprehensive impact factor, expressed as: ; S208-5: Calculate and obtain the updated health coefficient based on the initial health coefficient and the comprehensive impact factor, expressed as: ; S209: Determine whether the updated device health coefficient G is within the allowable range ZG; If it is not within the allowable range, the device initiates an abnormal alarm, determines that the device is an abnormal device and cannot support the drainage service, and simultaneously generates an operation and maintenance work order; S210: Set the acquisition frequency of the device marked as abnormal to 1 / 10 of the time-sharing, and reset the device health score to 100; S211: When the device maintenance work order is accepted and the device is repaired and reinstalled, restore the acquisition frequency to the default value and return to step S202; When within the buffer period ZQ, the device score meets the normal range, the device issues a return-to-normal alarm and re-enters the analyzable and business-supportable state.

[0045] Based on the above embodiments, use PL / pgSQL code to implement the abnormal identification method for the drainage network monitoring device provided by the present invention. The specific code is as follows: DECLARE v_error_msg VARCHAR(255); v_create_count INT; v_G INT; -- Health coefficient G v_threshold FLOAT := 2.0; -- Allowable threshold for abnormal judgment v_FZ_threshold INT := 20; -- Threshold for non-normal value data points v_FES_threshold INT := 10; -- Threshold for fixed value abnormal segments v_FES_duration INTERVAL; -- Duration of fixed value abnormal segments v_total_data_count INT; -- Total data volume of the sliced sub-table v_FZ_count INT; -- Count of non-normal value data points v_FES_count INT; -- Count of fixed value abnormal segments v_initial_time TIMESTAMP; -- Initial time of collection v_final_time TIMESTAMP; -- End time of collection v_total_time INTERVAL; -- Total time interval v_IF_total FLOAT; -- Total impact factor BEGIN -- Initialize the health coefficient G v_G := 100; -- Delete old data TRUNCATE TABLE water_level_may_abnormal; -- Step S1: Initialize the data slice range n, collection frequency m, and collection time range S SELECT MIN(sample_time), MAX(sample_time) INTO v_initial_time, v_final_time FROM water_level_may; v_total_time := v_final_time - v_initial_time; -- Calculate the total data volume of the sliced sub - table SELECT COUNT(*) INTO v_total_data_count FROM water_level_may; -- Step S2: Establish and initialize the equipment anomaly analysis database -- Here it is assumed that the database table already exists, and data slicing is performed -- Step S3: Calculate the fixed normal range interval z -- Here z is calculated from the installation physical parameters of the equipment -- Example: z_min = 0, z_max = buried_depth -- For subsequent anomaly detection -- Dynamically calculate and judge each data point, and correct the equipment health coefficient G WITH data_stats AS ( SELECT ID, location_id, sample_time, real_level, AVG(real_level) OVER (PARTITION BY location_id ORDERBY sample_time) AS avg_real_level, STDDEV_POP(real_level) OVER (PARTITION BY location_idORDER BY sample_time) AS stddev_real_level, ABS(real_level - AVG(real_level) OVER (PARTITION BYlocation_id ORDER BY sample_time)) / STDDEV_POP(real_level) OVER (PARTITIONBY location_id ORDER BY sample_time) AS deviation_score FROM water_level_may ), marked_abnormal AS ( SELECT ID, location_id, sample_time, real_level, CASE WHEN real_level>buried_depth OR real_level<0 ORreal_level IS NULL THEN 0 ELSE 1 END AS abnormal_1, -- 实际液位超出物理范围 CASE WHEN deviation_score>v_threshold THEN 0 ELSE 1 END AS abnormal_2 -- 超出动态阈值范围的偏差 FROM data_stats ) -- 插入异常标记结果到 water_level_may_abnormal 表 INSERT INTO water_level_may_abnormal (ID, abnormal_1, abnormal_2) SELECT ID, abnormal_1, abnormal_2 FROM marked_abnormal; -- Step S4: Comprehensively judge all conditions and update the final abnormal flag UPDATE water_level_may_abnormal SET abnormal_all = CASE WHEN abnormal_1 = 1 AND abnormal_2 = 1 THEN 1 ELSE 0 END; -- Calculate the non - normal value data points FZ SELECT COUNT(*) INTO v_FZ_count FROM water_level_may_abnormal WHERE abnormal_2 = 0; IF v_FZ_count>v_FZ_threshold THEN v_G := v_G - (v_FZ_count * 2); -- Reduce the G value by each non - normal value data point, 2 is the influence factor END IF; -- Step S5: Judge whether there is a fixed - value abnormal segment FES in the sliced sub - table WITH fixed_value_segments AS ( SELECT location_id, COUNT(*) AS segment_length, MIN(sample_time) AS start_time, MAX(sample_time) AS end_time, MAX(real_level) - MIN(real_level) AS value_range FROM water_level_may GROUP BY location_id, real_level HAVING MAX(real_level) - MIN(real_level) = 0 ) SELECT COUNT(*) INTO v_FES_count FROM fixed_value_segments WHEREsegment_length>v_FES_threshold; -- If there is a fixed value abnormal segment, further reduce the health factor G IF v_FES_count>0 THEN v_G := v_G - (v_FES_count * 10); -- Each fixed value abnormal segment reduces the G value, 10 is the impact factor END IF; -- Step S6: Determine whether all data in the slice subtable remains unchanged -- If it remains unchanged, the slice subtable is recorded as a fixed value abnormal segment, which has been implemented in the previous step -- Step S7: Update the slice range and repeat the steps -- Here we simulate the update of slice range n and acquisition frequency m -- Example: n += 1, m += 1; This part can be implemented according to specific business rules -- Step S8: Update the calculation of equipment health factor G v_IF_total := (v_FZ_count * 2) + (v_FES_count * 10); -- Total of impact factors v_G := v_G - v_IF_total; -- Update the health factor G based on the total impact factor -- Step S9: Determine whether the health factor G is within the allowable range IF v_G<50 THEN -- Example: If G is less than 50, trigger an alarm RETURN QUERY SELECT 'alert'::VARCHAR, 'The health factor is lower than the safe range, and the device may be abnormal'::VARCHAR; END IF; -- Returns the success status and amount of new data RETURN QUERY SELECT 'success'::VARCHAR, v_create_count::VARCHAR; --Exception handling part EXCEPTION WHEN OTHERS THEN GET STACKED DIAGNOSTICS v_error_msg = MESSAGE_TEXT; RETURN QUERY SELECT 'error'::VARCHAR, REPLACE(v_error_msg,'"', '')::VARCHAR; END; $BODY$ LANGUAGE plpgsql VOLATILE COST 100 ROWS 1000; Based on the above embodiments, in the embodiments of the present invention, the method for identifying anomalies in drainage network monitoring devices provided by the present invention is used to verify and test 105 drainage level devices in a certain city. Among them, if the health score is lower than 91 points, the device is determined to be abnormal and cannot support the business. Referring to Table 1 shown below, which is the device score table, a total of 84 devices are rated as healthy and 21 devices are abnormal. Table 1 Device Score Table Site Name Collection Volume Abnormal Volume Health Score G Rainwater Level of Linhu Avenue 1 1397 1102 21.12 Sewage Level of Shuangzhu Road 2 1485 756 49.09 Sewage Level of Xinli Road 1332 666 50 Sewage Level of Beipu Road 1426 636 55.4 Sewage Level of Xinta Street 2 1397 611 56.26 Sewage Level of Xiongfeng Road 1324 568 57.1 Rainwater Level of Fenhu Avenue 1398 492 64.81 Sewage Level of Luxin Avenue 3 1413 339 76.01 Sewage Level of Hainan Road 1439 345 76.03 Sewage Level of She'nan Road 1482 325 78.07 Sewage Level of Zhenxi Road 1400 295 78.93 Sewage Level of Renmin East Road 1 1396 261 81.3 Sewage Level of Zhennan Road 1213 218 82.03 Sewage Level of Pugang South Road 1387 235 83.06 Sewage Level of Xinshe Road 239 35 85.36 Sewage Level of Kangli Avenue 2 1388 197 85.81 Rainwater Level of Xincun Road 1308 168 87.16 Sewage Level of Linhu Avenue 1 1383 168 87.85 Sewage Level of Dasheng Road 1422 170 88.05 Sewage Level of Fuxin Road 1356 135 90.04 Sewage Level of Yuexiu Road 1294 119 90.8 Sewage Level of Fenyang Road 3 1057 94 91.11 Sewage Level of Renmin East Road 2 1349 107 92.07 Sewage Level of Hujing North Road 1442 112 92.23 Sewage Level of Chengsi Road 1 1145 86 92.49 Rainwater Level of Linhu Avenue 3 1386 103 92.57 Sewage Level of Luxin Avenue 1 1096 54 95.07 Sewage Level of Jinxian Road 1223 60 95.09 Sewage Level of Xinta Street 1 1225 58 95.27 Sewage Level of Yucai Road 1 1401 62 95.57 Sewage Level of Donggang Road 846 37 95.63 River Level of Beiyuegang 1309 56 95.72 Sewage Level of Laixiu Road 1 1332 54 95.95 Sewage Level of Shuangzhu Road 1325 53 96 Sewage Level of Pubei Road 1402 54 96.15 Sewage Level of Kangli Avenue 1 1048 39 96.28 River Level of Dongnangang 1348 47 96.51 Sewage Level of Songyang Road 1468 50 96.59 Sewage Level of Linhu Avenue 6 1421 42 97.04 Rainwater Level of Hunie Line 1299 35 97.31 Sewage Level of Chengsi Road 2 1291 33 97.44 Sewage Level of Shexing Road 1 250 6 97.6 Sewage Level of Yangsha Road 1209 29 97.6 Sewage Level of Meilan Road 2 1399 33 97.64 Sewage Level of South Ring Road 1 1447 34 97.65 Meilan Road Pumping Station 1382 32 97.68 Sewage Level of Linghou Road 1487 34 97.71 Sewage Level of Laixiu Road 3 948 21 97.78 Sewage Level of Linhu Avenue 4 1388 30 97.84 Rainwater Level of Luxin Avenue 2 1423 30 97.89 Sewage Level of Linhu Avenue 3 1332 27 97.97 Sewage Level of Datong Road 1339 26 98.06 Rainwater Level of Limin North Road 2 1397 27 98.07 Sewage Level of South Ring Road 2 904 17 98.12 Sewage Level of Xinyou Road 2 1408 26 98.15 Rainwater Level of Limin North Road 1 1444 25 98.27 Sewage Level of Nanjing Road 2 1338 23 98.28 Rainwater Level of Renmin East Road 1342 23 98.29 Sewage Level of Pugang Road 1 1481 25 98.31 Sewage Level of Xinyou Road 1 1447 23 98.41 Sewage Level of Renmin East Road 3 1481 21 98.58 Sewage Level of Meilan Road 1 788 11 98.6 Rainwater Level of Pubei Road 1317 18 98.63 Sewage Level of Luxin Avenue 2 1368 18 98.68 River Level of Shexingdang 237 3 98.73 Rainwater Level of Xinyou Road 1 1429 18 98.74 Sewage Level of Chengsi Road 3 1387 17 98.77 Rainwater Level of Luxin Avenue 1 1352 16 98.82 Sewage Level of Limin South Road 1267 15 98.82 Sewage Level of Nanjing Road 1 1384 16 98.84 Sewage Level of Yuejiang Road 1398 16 98.86 Sewage Level of Linhu Avenue 5 1487 16 98.92 River Level of Xishashi River 1317 14 98.94 Rainwater Level of Fenyang Road 1 1315 14 98.94 Pumping Station Level of Nanshangang 696 7 98.99 Rainwater Level of Kangli Avenue 1382 14 98.99 Sewage Level of Linhu Avenue 2 1395 14 99 Sewage Level of Jiangsu Road 1063 10 99.06 Rainwater Level of Xinli Road 1514 13 99.14 Sewage Level of Jinxin East Road 1394 12 99.14 Rainwater Level of Linhu Avenue 4 1415 11 99.22 Sewage Level of Xinchuan Road 1407 11 99.22 River Level of Dayanggang 1347 10 99.26 Rainwater Level of Sanbaidang Road 1349 10 99.26 Rainwater Level of Jiangsu Road 1443 10 99.31 Rainwater Level of Pugang Road 1333 9 99.32 Rainwater Level of Chengsi Road 1 1008 5 99.5 Sewage Level of Kangli Avenue 3 1019 5 99.51 Rainwater Level of Yuexiu Road 1485 7 99.53 Rainwater Level of Linhu Avenue 2 1372 6 99.56 Sewage Level of Jinkang Road 983 4 99.59 Sewage Level of Binhe South Road 1454 6 99.59 Rainwater Level of Nanjing Road 1409 5 99.65 Rainwater Level of Fenyang Road 2 1498 5 99.67 Rainwater Level of Pugang South Road 990 3 99.7 Rainwater Level of Chengsi Road 2 1382 4 99.71 Rainwater Pipe Level of Datong Road 1492 4 99.73 In the Grass of Laixiu Road 2 1433 3 99.79 Sewage Level of Fenhu Avenue 1368 2 99.85 Rainwater Level of Laixiu Road 1419 2 99.86 Sewage Level of East Pugang Road 3 1486 2 99.87 Sewage Level of Yucai Road 2 1146 1 99.91 River Level of Beilinggang 1197 1 99.92 Rainwater Level of Xinyou Road 2 1180 0 100 Level of Xinkai River 778 0 100 The method for identifying anomalies in drainage network monitoring devices according to the present invention collects the liquid level data of the monitoring devices to be identified in real time. Based on the local data before and after each liquid level data, it calculates the interval horizontal standard line and deviation index of each liquid level data, and captures the abnormal jumps of the liquid level data in real time. At the same time, by combining non-normal values and fixed-value abnormal segments, a multi-level anomaly marking mechanism is used to comprehensively capture different types of abnormal signals, avoid missed reports and false alarms, and improve the accuracy of anomaly identification. Moreover, through multi-dimensional data analysis and dynamic adjustment, combined with the real-time correction of the health coefficient of the monitoring devices to be identified, it better adapts to the complex and changeable monitoring environment and the non-normal distribution of monitoring data, and improves the accuracy and robustness of anomaly identification for the monitoring devices to be identified. The present invention adjusts the parameter n of the sliced sub-table of the collected data based on seasons and business types to achieve multi-level data slice analysis, and can adjust the window size and analysis range on different time scales, so as to accurately identify long-term anomalies and avoid the misjudgment risk brought by fixed windows. At the same time, the preset deviation threshold and fixed normal range interval can both be adjusted according to the working environment and business type of the monitoring devices to be identified, effectively avoiding the misjudgment problem brought by fixed thresholds and ensuring that the anomalies of the monitoring devices can still be accurately and timely identified in a complex and changeable working environment.

[0046] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) that contain computer-usable program code.

[0047] The present application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0048] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0049] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0050] Obviously, the above embodiments are only examples given for clear illustration and are not limitations on the implementation manners. For those of ordinary skill in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation manners here. And the obvious changes or modifications derived therefrom are still within the protection scope of the present invention.

Claims

1. A method for identifying anomalies in a drainage pipe network monitoring device, characterized in that, Including: Based on the pipeline topology of the drainage pipeline network and the installation locations of all monitoring devices to be identified, all the monitoring devices to be identified are divided into multiple monitoring groups; Based on a preset data acquisition frequency, the monitoring data of the monitoring devices to be identified is obtained in real time; For the monitoring data at the sampling moment, obtain the monitoring data within the sampling moment range to form a sub-table of the collected data slices; Based on the local fluctuation coefficient, standard deviation and interval level standard line of each monitoring data and its corresponding sub-table of acquisition data slices, the deviation degree of each monitoring data is obtained; Based on the deviation degree of each monitoring data and a preset deviation threshold, the abnormal jump impact factor of each monitoring data is calculated; Sum up the abnormal jump impact factors of all monitoring data in the sub-table of acquisition data slices to obtain the abnormal jump impact factor of the sub-table of acquisition data slices; Obtain the total number of data that exceed the fixed normal range interval in the monitoring data of the sub-table of acquisition data slices and the total number of data in the sub-table of acquisition data slices, and calculate the non-normal value impact factor of the sub-table of acquisition data slices; Obtain the duration of the fixed value abnormal segment in the monitoring data of the sub-table of acquisition data slices and the total time of the sliced data, and calculate the fixed abnormal impact factor of the sub-table of acquisition data slices at the current moment; Based on the abnormal jump impact factor, non-normal value impact factor and fixed abnormal impact factor of the sub-table of acquisition data slices, construct a comprehensive impact factor to correct the initial health coefficient of the monitoring device to be identified, and obtain the current health coefficient of the monitoring device to be identified; Obtain the total number of monitoring devices to be identified whose current health coefficients are not within the allowable range of health coefficients in each monitoring group. If the total number is less than the preset threshold, determine that the monitoring devices to be identified whose current health coefficients are not within the allowable range of health coefficients are abnormal devices.

2. The method for identifying abnormalities of the drainage network monitoring device according to claim 1, wherein Based on seasons, the spatial type of the pipeline and the business type, construct the corresponding length of the sub-table of acquisition data slices, including: Initialize the length of the data slice sub-table to ; Set corresponding seasonal adjustment coefficients for the dry season, rainy season, and transition season respectively ; Set corresponding adjustment coefficients for the spatial types of the pipelines for industrial areas, commercial areas, and residential areas respectively ; Set corresponding business type adjustment coefficients for daily operation and maintenance, leakage analysis, siltation analysis, flood control warning, and equipment calibration services respectively ; Based on the seasonal adjustment coefficient, the spatial type adjustment coefficient of the pipeline, and the business type adjustment coefficient corresponding to the current moment, adjust the length of the initial data slice sub-table to obtain the length of the collected data slice sub-table corresponding to the current moment, denoted as .

3. The method for identifying anomalies in a drainage network monitoring device according to claim 1, characterized in that, Based on the local fluctuation coefficient, standard deviation and interval level standard line of each monitoring data and its corresponding sub-table of acquisition data slices, obtain the deviation degree of each monitoring data, including: Based on the local weighted average of the monitoring data in the acquisition data slice sub-table corresponding to the monitoring data at the sampling moment, obtain the interval horizontal standard line of the monitoring data at the sampling moment , expressed as: ； Calculate the first local fluctuation coefficient of the monitoring data at the sampling moment , which is expressed as: ； Calculate the second local fluctuation coefficient of the monitoring data at the sampling moment , which is expressed as: ； Calculate the standard deviation of the monitored data at the sampling moment , which is expressed as: ； Based on the first local fluctuation coefficient, the second local fluctuation coefficient, and the standard deviation, calculate the deviation index of the monitoring data at the sampling moment , which is expressed as: ； Calculate the ratio of the absolute value of the difference between the monitored data at the sampling moment and the interval horizontal standard line to the deviation index to obtain the degree of deviation of the monitored data at the sampling moment, expressed as: ； Among them, represents the weight of the monitoring data at the sampling moment, and the expression is , where represents the attenuation coefficient; is the monitoring data at the sampling moment, representing the true monitoring horizontal line; represents the arithmetic mean of the acquisition data slice sub - table of the monitoring data at the sampling moment, and the expression is , where represents the total number of data in the acquisition data slice sub - table; and respectively represent the preset weights of the first local fluctuation coefficient and the second local fluctuation coefficient.

4. The method for identifying anomalies in a drainage network monitoring device according to claim 3, characterized in that, Sum up the abnormal jump impact factors of all monitoring data in the sub-table of acquisition data slices to obtain the abnormal jump impact factor of the sub-table of acquisition data slices, including: Based on the deviation degree between the monitoring data at the sampling moment and the preset deviation threshold, calculate the abnormal jump influence factor of the monitoring data at the sampling moment, which is expressed as: ; Sum the abnormal jump impact factors of the monitoring data in the sliced sub-table of the collected data, and obtain the abnormal jump impact factor of the sliced sub-table of the collected data , which is expressed as: ； Among them, represents the abnormal jump weight, represents the preset deviation threshold.

5. The abnormal identification method of the drainage network monitoring device according to claim 4, characterized in that, Obtain the total number of data that exceed the fixed normal range interval in the monitoring data of the sub-table of acquisition data slices and the total number of data in the sub-table of acquisition data slices, and calculate the non-normal value impact factor of the sub-table of acquisition data slices, expressed as: ； Among them, represents the abnormal value influencing factor of the acquisition data slice sub-table, represents the abnormal value weight, represents the total number of data that exceed the fixed normal range in the monitoring data of the acquisition data slice sub-table.

6. The method for identifying anomalies in a drainage network monitoring device according to claim 5, characterized in that, Obtain the duration of the fixed value abnormal segment in the monitoring data of the sub-table of acquisition data slices and the total time of the sliced data, and calculate the fixed abnormal impact factor of the sub-table of acquisition data slices at the current moment, expressed as: ； Among them, represents the fixed abnormal influence factor of the sub-table of the collected data slices, represents the fixed abnormal weight, represents the duration of the fixed-value abnormal segment in the monitoring data in the sub-table of the collected data slices, represents the total time of the sliced data.

7. The method for identifying abnormalities in a drainage network monitoring device according to claim 6, characterized in that, Based on the abnormal jump impact factor, non-normal value impact factor and fixed abnormal impact factor of the sub-table of acquisition data slices, construct a comprehensive impact factor to correct the initial health coefficient of the monitoring device to be identified, and obtain the current health coefficient of the monitoring device to be identified, including: Construct a comprehensive impact factor based on the abnormal jump impact factor, non-normal value impact factor, and fixed abnormal impact factor of the collected data slice sub-table , expressed as: ； Using the comprehensive impact factor to correct the initial health coefficient of the monitoring device to be identified and obtain the current health coefficient of the monitoring device to be identified , which is expressed as: 。 8. The method for identifying anomalies in a drainage network monitoring device according to claim 1, characterized in that, Obtain the current health coefficient of the monitoring device to be identified, and also include: If the current health coefficient of the monitoring device to be recognized is within the allowable range of the health coefficient, it is determined that the monitoring device to be recognized at the current moment is not an abnormal device, and the sampling moment is used to identify the monitoring device to be recognized at the next moment based on the monitoring data at the next sampling moment.

9. The method for identifying anomalies in a drainage network monitoring device according to claim 1, characterized in that, After determining that the monitoring device to be identified is an abnormal device, it also includes: Trigger the monitoring device to be identified to initiate an abnormal alarm and generate an operation and maintenance work order, and send it to the control center; After the operation and maintenance work order to be processed, recalculate the current health coefficient of the monitoring device to be identified until it is detected that the current health coefficient of the monitoring device to be identified is within the allowable range of the health coefficient, and then make the monitoring device to be identified enter the business supportable state.

10. The method for abnormity recognition of a drainage pipe network monitoring device according to claim 9, characterized in that, After determining that the monitoring device to be identified is an abnormal device, it further includes: Reduce the preset data collection frequency of the monitoring device to be identified to the collection frequency threshold, and resume the preset data collection frequency after the operation and maintenance work order is processed and the device enters the business supportable state.

Images