Drainage pipeline risk early warning method, system, equipment and medium

Abstract

The invention belongs to the technical field of intelligent water affair and urban disaster prevention, and provides a drainage pipeline risk early warning method, system and device and a medium, and the method comprises the following steps: dividing risk areas through combining pipeline GIS data and historical faults, and carrying out the differentiated arrangement of temperature and strain sensing optical fibers to form a double-parameter network; synchronously acquiring data by using a demodulator, processing the data by using a dynamic self-adaptive sliding window of a preset threshold value, and identifying abnormal segments; after multi-source fusion correction, inputting a pre-trained CNN-LSTM model to preliminarily evaluate the event type and risk; carrying out digital twinning simulation on consequences of moderate and above risks by using pipelines, and judging fourth-level risks; early warning is pushed according to grades, response is triggered, disposal feedback is collected, and model parameters and risk thresholds are dynamically adjusted through incremental learning; according to the invention, accurate capture of micro-anomaly of the drainage pipeline and real-time quantitative assessment of risk are realized, the life cycle of the system is prolonged, the operation and maintenance cost is reduced, and intelligent and safe operation and maintenance of the drainage pipeline are achieved.

Description

Technical Field

[0001] This invention belongs to the field of smart water management and urban disaster prevention technology, specifically a method, system, equipment and medium for early warning of drainage pipeline risks. Background Technology

[0002] Urban drainage pipelines are core infrastructure for ensuring urban water circulation and public safety, undertaking the core functions of rainwater and sewage transportation, flood control and drainage. However, their characteristics of being buried deep underground and having complex operating conditions make it difficult for existing monitoring technologies to meet the needs of intelligent operation and maintenance.

[0003] Current mainstream monitoring solutions mostly adopt a uniform deployment logic: either by attaching dual optical fibers at equal intervals throughout the pipeline, or by installing pressure and flow sensors at fixed points, without considering the risk differences in different sections of the pipeline. For example, in old urban areas, the bends and joints of cast iron pipes are prone to micro-strain due to material aging and concentrated stress, but uniformly deployed optical fibers are unable to capture such micro-anomalies, resulting in a high rate of missed detection, and eventually small leaks develop into pipeline ruptures. On the other hand, in newly built urban areas, straight sections of PE pipes are deployed with optical fibers according to the standards for high-risk areas, even though there is no risk, which is more than the actual need and increases the overall cost of the monitoring project.

[0004] Furthermore, existing monitoring technologies primarily employ fixed time windows for data processing. When faced with sudden anomalies, a full window period must be waited for identification, resulting in long response times and an expansion of the leakage range during this period. For slow-moving anomalies, the fixed window can easily overwhelm trending anomalies with normal data fluctuations, leading to a high false negative rate. In addition, risk assessment relies solely on a single parameter threshold without integrating multi-source information, making ineffective maintenance and operations prone to occur. Moreover, the lack of quantification capabilities for risk consequences forces maintenance personnel to passively attempt repairs, impacting residents' travel time. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method, system, device, and medium for early warning of risks in drainage pipelines, thereby resolving the issues raised in the background art regarding insufficient sensing accuracy and ambiguous risk assessment in traditional drainage pipeline monitoring, which leads to unquantifiable early warnings.

[0006] According to the first aspect of this disclosure, a risk early warning method for drainage pipelines is proposed, comprising the following steps:

[0007] S1. Based on pipeline GIS data and historical fault records, the pipeline is divided into risk zones, and temperature sensing fiber optics and strain sensing fiber optics are deployed in different risk zones in a differentiated manner to form a dual-parameter sensing network.

[0008] S2. The temperature and strain data of the dual-parameter sensor network are synchronously acquired by the demodulator, and the data stream is processed in real time using a dynamic adaptive sliding window analysis method based on a preset threshold to identify abnormal data segments.

[0009] S3. Perform multi-source data fusion correction on the identified abnormal data segments, and input them into the pre-trained CNN-LSTM fusion model for event type identification and preliminary risk assessment;

[0010] S4. For events assessed as having moderate or higher risks, drive the pipeline digital twin to perform consequence simulation and determine the four-level risk level based on the simulation results.

[0011] S5. Generate graded early warning information based on the determined risk level and trigger the corresponding response process. At the same time, collect feedback data on early warning handling and dynamically adjust the parameters and risk judgment threshold of the CNN-LSTM fusion model through incremental learning algorithm.

[0012] Preferably, the risk zoning in step S1 includes a high-risk zone, a medium-risk zone, and a low-risk zone;

[0013] The differentiated deployment methods include:

[0014] In high-risk areas, the temperature sensing fiber and strain sensing fiber are laid out in a dense spiral pattern, and the strain sensing fiber at the pipe interface is reinforced by a cross-shaped arrangement.

[0015] In the medium-risk area, the temperature sensing fiber and the strain sensing fiber are arranged in parallel with a spacing of 10cm.

[0016] In low-risk areas, the temperature sensing fiber and the strain sensing fiber are arranged in parallel with a spacing of 20cm.

[0017] Preferably, the dynamic adaptive sliding window analysis method in step S2 specifically includes:

[0018] When the fluctuation range of the monitored data is lower than the first preset threshold, the first time window is used for data analysis.

[0019] When the fluctuation range of the monitored data exceeds the second preset threshold or the rate of change exceeds the third preset threshold, the system automatically switches to a shorter second time window for data analysis.

[0020] Preferably, the multi-source data fusion correction in step S3 specifically includes:

[0021] Access geological monitoring data around the pipeline and operational data inside the pipeline;

[0022] Risk weighting factors are set based on the pipe material and service life to calculate the risk level of abnormal data.

[0023] Preferably, the CNN-LSTM fusion model in step S3 extracts the spatial distribution features of abnormal data through a convolutional neural network (CNN) and extracts the time series features of abnormal data through a long short-term memory network (LSTM), together completing the classification of event types and the output of risk confidence.

[0024] Preferably, the four risk levels in step S4 include at least the normal level based on the fluctuation state of temperature and strain data, the slightly abnormal level based on the abnormality of a single parameter in temperature or strain data, the moderate risk level based on the continuous deterioration of two parameters based on temperature and strain data and simulation confirmation, and the severe risk level based on the severe exceedance of parameters based on temperature or strain data and short-term failure prediction.

[0025] The specific criteria for determining the four risk levels are as follows:

[0026] Level I: Normal level; fluctuations in temperature and strain data are within the normal threshold range.

[0027] Level II: This is a minor anomaly; a single parameter briefly exceeds the threshold, but the trend is stable and there is no derivative risk.

[0028] Level III: Moderate risk; both parameters continuously exceed the threshold and the trend is deteriorating, and digital twin simulation shows that there is local risk;

[0029] Level IV: This is a severe risk level; parameters are severely exceeded, and digital twin simulation predicts that the pipeline structure has a risk of failure in the short term.

[0030] According to the second aspect of this disclosure, a drainage pipeline risk early warning system is also proposed, applied to the first aspect, comprising:

[0031] The sensing module includes temperature sensing optical fibers and strain sensing optical fibers that are deployed differently based on risk zones, and is used to collect temperature and strain data along the pipeline.

[0032] The data acquisition module includes a demodulator and an edge computing unit, wherein the edge computing unit is used to execute the dynamic adaptive sliding window analysis method to identify abnormal data fragments;

[0033] The intelligent analysis module is used to receive the abnormal data fragments, process them using a pre-trained CNN-LSTM fusion model and pipeline digital twin, and finally output the judgment result of the four-level risk level.

[0034] The early warning response module is used to generate graded early warning information based on the determined risk level and trigger the corresponding response process;

[0035] The feedback optimization module is used to collect actual handling result data of the early warning event from the early warning response module, and to incrementally learn and optimize the internal parameters and risk judgment threshold of the pre-trained CNN-LSTM fusion model.

[0036] Preferably, the intelligent analysis module includes:

[0037] A multi-source fusion unit is used to perform multi-source data fusion correction on the abnormal data fragments and output valid abnormal data.

[0038] The CNN-LSTM analysis unit is used to input the valid outlier data into the pre-trained CNN-LSTM fusion model and output the event type and initial risk confidence.

[0039] The digital twin simulation unit is used to drive the pipeline digital twin to perform consequence simulation for events with an initial risk confidence level of medium or higher, and generate simulation results.

[0040] The risk decision-making unit is used to determine the four-level risk level by combining the initial risk confidence level and the simulation results.

[0041] According to a third aspect of this disclosure, a computer device is also provided, the device comprising:

[0042] One or more processors;

[0043] Storage device for storing one or more programs;

[0044] When the one or more programs are executed by the one or more processors, the one or more processors implement the drainage pipeline risk warning method proposed in the first aspect.

[0045] According to a fourth aspect of this disclosure, a storage medium is provided on which a computer program is stored, which, when executed by a processor, implements the drainage pipeline risk warning method proposed in the first aspect.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] 1. This invention deploys dual optical fibers in a way that is more differentiated by pipeline risk, and uses encryption and cross reinforcement methods in high-risk areas, while controlling the distance as needed in medium and low-risk areas. This improves the ability to detect micro-anomalies in high-risk areas and reduces the amount of optical fiber used in low-risk areas, thereby reducing the overall monitoring cost.

[0048] 2. This invention uses a dynamic window to filter out redundant data in real time, performs multi-source fusion to correct interference, and simulates the consequences of risks through digital twins, thereby reducing abnormal response time and false alarm rate, and improving the efficiency of operation and maintenance personnel in formulating handling plans.

[0049] 3. This invention collects early warning and response feedback data, uses incremental learning to fine-tune the CNN-LSTM model parameters and risk thresholds, stabilizes the model's accuracy, extends the system's lifespan, and reduces long-term operation and maintenance costs. Attached Figure Description

[0050] Figure 1 This is a flowchart of the drainage pipeline risk early warning method of the present invention;

[0051] Figure 2 This is a block diagram of the drainage pipeline risk early warning system of the present invention;

[0052] Figure 3 This is a diagram illustrating the fiber optic cable layout within the pipe according to the present invention. Detailed Implementation

[0053] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention.

[0054] Example 1: This invention provides a method for early warning of risks in drainage pipelines, as shown in the attached figure. Figure 1 As shown, it includes the following steps:

[0055] S1. Based on pipeline GIS data and historical fault records, pipelines are classified into risk zones. The pipeline GIS data includes basic attributes such as pipe diameter, material, service life, and burial depth. Historical fault records are obtained by statistically analyzing the frequency of pipeline faults over the past 3-5 years. Pipe sections with a fault frequency of >3 times / year are directly classified into high-risk zones. Risk zones include high-risk zones, medium-risk zones, and low-risk zones.

[0056] The high-risk area is the core area of ​​the fault, including bends, interfaces, deeply buried sections, severely corroded sections, or sections with frequent historical faults; the medium-risk area is the area where the fault is easy to spread, that is, the transition section connecting with the high-risk area; the low-risk area is the area where the fault is infrequent, mainly the straight sections with intact structure, short service life, and no historical faults.

[0057] Temperature sensing fibers and strain sensing fibers are deployed in different risk areas using differentiated methods to form a dual-parameter sensing network; the differentiated deployment methods include:

[0058] In high-risk areas, temperature sensing fibers and strain sensing fibers are deployed in a dense spiral pattern, with the two fibers spirally wound around the pipe at a 5cm interval, effectively improving spatial coverage density; and strain sensing fibers at pipe interfaces are reinforced by a cross-shaped deployment, thereby enhancing the ability to capture micro-deformations and ensuring that no minute anomalies are missed, such as early strain changes caused by interface sealing failure.

[0059] In medium-risk areas, temperature sensing fibers and strain sensing fibers are laid out in parallel with a spacing of 10cm; this effectively balances monitoring accuracy and cost, and covers transitional areas where faults may spread.

[0060] In low-risk areas, temperature sensing fibers and strain sensing fibers are laid out in parallel with a spacing of 20cm; only basic monitoring capabilities are retained to reduce unnecessary resource investment.

[0061] This deployment method can improve the sensitivity of anomaly detection in high-risk areas, and the overall monitoring cost is reduced compared to uniform deployment, effectively improving resource utilization.

[0062] S2. Temperature and strain data from the dual-parameter sensor network are simultaneously acquired via a demodulator, converting the optical signal into a digital signal to provide raw data for subsequent analysis. A dynamic adaptive sliding window analysis method based on a preset threshold is then used to process the data stream in real time, identifying anomalous data segments. The dynamic adaptive sliding window analysis method specifically includes:

[0063] When the fluctuation range of the monitored data is lower than the first preset threshold, the first time window is used for data analysis.

[0064] When the fluctuation range of the monitored data exceeds the second preset threshold or the rate of change exceeds the third preset threshold, the system automatically switches to a shorter second time window for data analysis.

[0065] The first preset threshold is used when the monitored data fluctuation is within the normal range. At this time, the first time window is used to reduce data redundancy and reduce the processing pressure of the edge computing unit by using a large window. The second and third preset thresholds are used when abnormal signs appear in the data (such as strain fluctuation ≥5με, temperature fluctuation ≥2℃, or strain change rate >20με / s, temperature change rate >1℃ / s). At this time, the system automatically switches to the second time window, thereby quickly capturing the deterioration trend of sudden or slow abnormalities by using high-frequency sampling through a small window.

[0066] The identification of anomalous data segments involves using the Z-score outlier detection algorithm within each window to identify data points that exceed the normal range. Data segments consisting of three or more consecutive outliers are marked as anomalous, thus providing target objects for subsequent analysis. This step effectively shortens the response time to sudden anomalies and reduces the false negative rate for slow anomalies.

[0067] S3. Perform multi-source data fusion correction on the identified anomalous data segments, and input the correction into a pre-trained CNN-LSTM fusion model for event type identification and preliminary risk assessment; specifically, the multi-source data fusion correction includes:

[0068] External data is introduced to perform secondary verification of abnormal data fragments, eliminating environmental interference; the data types accessed are geological monitoring data around the pipeline and operational data inside the pipeline; the geological factors of the geological monitoring data around the pipeline are denoted as... The operating condition factors of the pipeline's internal operating data are denoted as... ;

[0069] Risk weighting factors are set based on the pipe material and service life, where the risk weighting factor for pipe material is denoted as... The risk weighting factor for the service life of the pipeline is denoted as The risk value of outlier data is corrected by weighted calculation, and valid outlier data is output.

[0070] Because the impact of various factors on risk is cumulative, the total risk weighting factor... The product of the individual factors is expressed as: .

[0071] The original anomaly risk value is calculated based on the deviation of temperature and strain data, and is expressed as:

[0072]

[0073] in, This represents the original anomaly risk value. This represents the absolute value of the temperature data deviating from the normal threshold. This is the upper limit of the temperature safety threshold. The absolute value of the strain data deviating from the normal threshold. This is the upper limit of the strain safety threshold. These are the weighting coefficients for temperature and strain, respectively.

[0074] Effective outlier risk value = original risk value × total weighting factor, i.e.: This value represents the valid abnormal data. Its physical meaning is the actual risk level of the abnormal data after considering the pipeline's own characteristics (material, age) and environmental conditions (geology, operating conditions).

[0075] The CNN-LSTM fusion model extracts the spatial distribution features of anomalous data through a convolutional neural network (CNN) and extracts the temporal series features of anomalous data through a long short-term memory network (LSTM). By combining the spatial distribution features and the temporal series features, it outputs the event type and the initial risk confidence level, providing a basis for subsequent classification.

[0076] This step eliminates environmental interference from a single sensor by fusing multi-source data. It combines the spatial and temporal feature extraction capabilities of the CNN-LSTM fusion model to accurately identify abnormal event types and assess initial risks. This solves the problems of high false alarm rates and ambiguous event type identification caused by single parameter threshold judgment, thus significantly improving data credibility.

[0077] S4. For events assessed as having moderate or higher risks, drive the pipeline digital twin to perform consequence simulation. The pipeline digital twin is built based on BIM and GIS. The simulation content includes risk diffusion paths (such as the penetration range of the leaking medium), structural failure time (such as predicting the pipeline rupture countdown when strain > 100με), and derivative disasters (such as the risk of road collapse). The output is quantified risk consequence data.

[0078] The simulation results are combined to determine the risk level of four levels. The four risk levels include at least the normal level based on the fluctuation state of temperature and strain data, the slightly abnormal level based on the abnormality of a single parameter in temperature or strain data, the moderate risk level based on the continuous deterioration of two parameters based on temperature and strain data and the simulation confirmation, and the severe risk level based on the severe exceedance of parameters based on temperature or strain data and the prediction of short-term failure.

[0079] The specific criteria for determining a Level IV risk level are as follows:

[0080] Level I: Normal level; fluctuations in temperature and strain data are within the normal threshold range.

[0081] Level II: This is a minor anomaly; a single parameter briefly exceeds the threshold, but the trend is stable and there is no derivative risk.

[0082] Level III: Moderate risk; both parameters continuously exceed the threshold and the trend is deteriorating, and digital twin simulation shows that there is local risk;

[0083] Level IV: This is a severe risk level; parameters are severely exceeded, and digital twin simulation predicts that the pipeline structure has a risk of failure in the short term.

[0084] This step uses a digital twin to reproduce the abnormal evolution process and combines it with the initial risk assessment results to achieve accurate quantification of the risk level. The degree of quantification of risk level determination is improved, and operation and maintenance personnel can directly formulate handling strategies based on the level.

[0085] S5. Generate tiered early warning information based on the determined risk level and trigger the corresponding response process; that is, match the push channels and handling procedures according to the risk level to ensure accurate delivery of early warning information and efficient response:

[0086] For Level II risks, a notification will only be displayed on the operations and maintenance platform interface, along with real-time data curves, so that operations and maintenance personnel can monitor subsequent changes.

[0087] For Level III risks, a notification is sent to the operations and maintenance manager's mobile app, along with a screenshot of the digital twin simulation and a checklist, requiring the inspection to be completed within 24 hours.

[0088] For Level IV risks, the municipal emergency command platform will be triggered, automatically pushing information to the repair team (including navigation and equipment list), and coordinating with the urban management department to release traffic control information.

[0089] Simultaneously, feedback data from early warning and response is collected, and the parameters and risk assessment threshold of the CNN-LSTM fusion model are dynamically adjusted through incremental learning algorithms. Specifically, the model parameters are fine-tuned every 7 days based on new data to ensure accuracy; the dynamic optimization and adjustment of the risk assessment threshold is to automatically raise the emergency alarm threshold for a certain area if there are 3 consecutive Level II false alarms, thereby reducing similar false alarms.

[0090] Example 2: The present invention also provides a drainage pipeline risk early warning system, applied in Example 1, as shown in the appendix. Figure 2 As shown, it includes:

[0091] The sensing module includes temperature-sensing optical fibers and strain-sensing optical fibers deployed differently based on risk zones, used to collect temperature and strain data along the pipeline. Through the dual optical fibers deployed differently in risk zones, temperature and strain data are collected directionally along the pipeline, ensuring sufficient data density in high-risk areas and preventing resource waste in low-risk areas. Specifically, the temperature-sensing optical fiber is used to collect temperature changes of the medium inside the pipeline, while the strain-sensing optical fiber is used to capture structural deformation of the pipeline.

[0092] By using encryption and cross-reinforcement in high-risk areas and deploying parallel arrays at intervals in medium- and low-risk areas, the monitoring needs of different regions can be matched. This allows the data acquisition accuracy to be focused on high-risk areas, improving the ability to capture micro-anomalies in those areas, while reducing the amount of fiber optic cables used in low-risk areas, lowering the overall monitoring cost, and optimizing resource utilization.

[0093] The data acquisition module includes a demodulator and an edge computing unit. The edge computing unit performs dynamic adaptive sliding window analysis to identify abnormal data segments. This module connects the sensing module and the intelligent analysis module, completing the conversion from optical signal to digital signal to abnormal segment, reducing invalid data transmission and improving system real-time performance. The demodulator converts the optical signal acquired from the fiber optic cable into a computable digital signal. The edge computing unit performs dynamic adaptive sliding window analysis to filter normal data and identify abnormal data segments in real time, uploading only the abnormal segments to the cloud-based intelligent analysis module.

[0094] The intelligent analysis module receives anomalous data segments and processes them using a pre-trained CNN-LSTM fusion model and a pipeline digital twin, ultimately outputting a four-level risk assessment result. The intelligent analysis module includes:

[0095] The multi-source fusion unit is used to perform multi-source data fusion correction on abnormal data fragments and output valid abnormal data. By combining the abnormal data fragments transmitted from the data acquisition module with multi-source external data and risk weighting factors, a secondary correction is performed to filter environmental interference and output valid abnormal data. The multi-source data access includes geological data around the pipeline and internal working condition data. The original abnormal data is weighted and calculated using the defined risk weighting factors to correct the risk value.

[0096] The CNN-LSTM analysis unit utilizes a pre-trained CNN-LSTM fusion model to extract spatial and temporal dual-dimensional features from valid outlier data, and outputs the event type and initial risk confidence.

[0097] The digital twin simulation unit is used to drive the pipeline digital twin to simulate the consequences of events with an initial risk confidence level of moderate or higher, reproduce the abnormal evolution process, and generate simulation results; the simulation content includes risk diffusion path, structural failure time and derivative disasters.

[0098] The risk decision-making unit is used to determine the four-level risk level by combining the initial risk confidence level and simulation results according to preset standards, clarify the priority of subsequent handling, and enable operation and maintenance personnel to directly formulate strategies according to the level, effectively improving the handling efficiency and avoiding low-risk events from occupying the repair resources of high-risk events.

[0099] The early warning response module is used to generate graded early warning information based on the determined risk level and trigger the corresponding response process;

[0100] The feedback optimization module is used to collect actual handling results data of early warning events from the early warning response module, and to incrementally learn and optimize the internal parameters and risk judgment threshold of the pre-trained CNN-LSTM fusion model.

[0101] Its optimization actions include model optimization and threshold optimization. Model optimization uses new feedback data to fine-tune the parameters of the CNN-LSTM model to ensure high accuracy. Threshold optimization automatically raises the response alarm threshold for a certain area after three consecutive Level II false alarms, thereby effectively extending the system life cycle and reducing long-term operation and maintenance costs.

[0102] Example 3: The present invention also provides a drainage pipeline risk early warning device, the device comprising:

[0103] One or more processors;

[0104] Storage device for storing one or more programs;

[0105] When one or more programs are executed by one or more processors, the one or more processors implement the drainage pipeline risk warning method proposed in Embodiment 1.

[0106] Example 4: The present invention also provides a storage medium for early warning of drainage pipeline risks, which stores a computer program that, when executed by a processor, implements the early warning method for drainage pipeline risks proposed in Example 1.

[0107] The computer storage medium of this invention can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. Computer-readable storage media include, but are not limited to: electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROMD), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0108] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0109] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0110] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0111] It should be understood that numerous specific implementation decisions can be made during the development of any actual implementation method, and in any engineering or design project. Such development efforts may be complex and time-consuming, but for those of ordinary skill in the art who benefit from this disclosure, the development effort will be a routine task in design, manufacturing, and production without requiring extensive experimentation.

[0112] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

[0113] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for early warning of risks in drainage pipelines, characterized in that, Includes the following steps: S1. Based on pipeline GIS data and historical fault records, the pipeline is divided into risk zones, and temperature sensing fiber optics and strain sensing fiber optics are deployed in different risk zones in a differentiated manner to form a dual-parameter sensing network. S2. The temperature and strain data of the dual-parameter sensor network are synchronously acquired by the demodulator, and the data stream is processed in real time using a dynamic adaptive sliding window analysis method based on a preset threshold to identify abnormal data segments. S3. Perform multi-source data fusion correction on the identified abnormal data segments, and input them into the pre-trained CNN-LSTM fusion model for event type identification and preliminary risk assessment; S4. For events assessed as having moderate or higher risks, drive the pipeline digital twin to perform consequence simulation and determine the four-level risk level based on the simulation results. S5. Generate graded early warning information based on the determined risk level and trigger the corresponding response process. At the same time, collect feedback data on early warning handling and dynamically adjust the parameters and risk judgment threshold of the CNN-LSTM fusion model through incremental learning algorithm.

2. The drainage pipeline risk early warning method as described in claim 1, characterized in that, The risk zoning in step S1 includes high-risk areas, medium-risk areas, and low-risk areas; The differentiated deployment methods include: In high-risk areas, the temperature sensing fiber and strain sensing fiber are laid out in a dense spiral pattern, and the strain sensing fiber at the pipe interface is reinforced by a cross-shaped arrangement. In the medium-risk area, the temperature sensing fiber and the strain sensing fiber are arranged in parallel with a spacing of 10cm. In low-risk areas, the temperature sensing fiber and the strain sensing fiber are arranged in parallel with a spacing of 20cm.

3. The drainage pipeline risk early warning method as described in claim 1, characterized in that, The dynamic adaptive sliding window analysis method in step S2 specifically includes: When the fluctuation range of the monitored data is lower than the first preset threshold, the first time window is used for data analysis. When the fluctuation range of the monitored data exceeds the second preset threshold or the rate of change exceeds the third preset threshold, the system automatically switches to a shorter second time window for data analysis.

4. The drainage pipeline risk early warning method as described in claim 1, characterized in that, The multi-source data fusion correction in step S3 specifically includes: Access geological monitoring data around the pipeline and operational data inside the pipeline; Risk weighting factors are set based on the pipe material and service life to calculate the risk level of abnormal data.

5. The drainage pipeline risk early warning method as described in claim 1, characterized in that, The CNN-LSTM fusion model in step S3 extracts the spatial distribution features of abnormal data through a convolutional neural network (CNN) and extracts the time series features of abnormal data through a long short-term memory network (LSTM), together completing the classification of event types and the output of risk confidence.

6. The drainage pipeline risk early warning method as described in claim 1, characterized in that, The four risk levels in step S4 include at least the normal level based on the fluctuation state of temperature and strain data, the slightly abnormal level based on the abnormality of a single parameter in temperature or strain data, the moderate risk level based on the continuous deterioration of two parameters based on temperature and strain data and simulation confirmation, and the severe risk level based on the severe exceedance of parameters based on temperature or strain data and short-term failure prediction. The specific criteria for determining the four risk levels are as follows: Level I: Normal level; fluctuations in temperature and strain data are within the normal threshold range. Level II: This is a minor anomaly; a single parameter briefly exceeds the threshold, but the trend is stable and there is no derivative risk. Level III: Moderate risk; both parameters continuously exceed the threshold and the trend is deteriorating, and digital twin simulation shows that there is local risk; Level IV: This is a severe risk level; parameters are severely exceeded, and digital twin simulation predicts that the pipeline structure has a risk of failure in the short term.

7. A drainage pipeline risk early warning system, applied in the method described in any one of claims 1-6, characterized in that, include: The sensing module includes temperature sensing optical fibers and strain sensing optical fibers that are deployed differently based on risk zones, and is used to collect temperature and strain data along the pipeline. The data acquisition module includes a demodulator and an edge computing unit, wherein the edge computing unit is used to execute the dynamic adaptive sliding window analysis method to identify abnormal data fragments; The intelligent analysis module is used to receive the abnormal data fragments, process them using a pre-trained CNN-LSTM fusion model and pipeline digital twin, and finally output the judgment result of the four-level risk level. The early warning response module is used to generate graded early warning information based on the determined risk level and trigger the corresponding response process; The feedback optimization module is used to collect actual handling result data of the early warning event from the early warning response module, and to incrementally learn and optimize the internal parameters and risk judgment threshold of the pre-trained CNN-LSTM fusion model.

8. The drainage pipeline risk early warning system as described in claim 7, characterized in that, The intelligent analysis module includes: A multi-source fusion unit is used to perform multi-source data fusion correction on the abnormal data fragments and output valid abnormal data. The CNN-LSTM analysis unit is used to input the valid outlier data into the pre-trained CNN-LSTM fusion model and output the event type and initial risk confidence. The digital twin simulation unit is used to drive the pipeline digital twin to perform consequence simulation for events with an initial risk confidence level of medium or higher, and generate simulation results. The risk decision-making unit is used to determine the four-level risk level by combining the initial risk confidence level and the simulation results.

9. A computer device, characterized in that, The device includes: One or more processors; Storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the drainage pipeline risk warning method as described in any one of claims 1-5.

10. A storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the drainage pipeline risk warning method as described in any one of claims 1-5.

Images