Electric power line fault intelligent early warning and monitoring system in tunnel environment

Abstract

The invention discloses an electric power line fault intelligent early warning and monitoring system in a tunnel environment, and relates to the field of electric power fault identification, and the system comprises an instruction management and control module which is used for accessing an electric power line general network of a target tunnel, and obtaining operation data access and control instruction issuing authority; the line planning module is used for dividing the whole line into different scales and outputting a plurality of sections of lines with different scales and physical coordinates and electrical parameters based on the topological structure and the historical fault point location distribution of the target tunnel power line by taking the historical fault distribution of the same point location data as a target; through dynamically adjustable multi-scale line division, a traditional global fixed monitoring mode is broken through, a line is divided into independent monitoring units of different scales, dynamic merging or slitting according to a data trend is supported, and the division scale of the line is dynamically adjusted according to a trend deterioration degree and historical data analysis. And an optimal segmentation point is flexibly selected under different conditions by using an optimization algorithm, and local refined monitoring is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of power failure identification, and specifically to an intelligent early warning and monitoring system for power line failures in a tunnel environment. Background Art

[0002] Power lines are a key part of the internal and external power supply in tunnels. Any failure may cause power interruption, seriously affecting the operational safety of the tunnels. Tunnels generally have narrow spaces, limited traffic capacity, and complex geographical conditions. The laying of power lines often faces many technical challenges. With the development of the Internet of Things and big data technologies, intelligent monitoring and early warning systems have gradually emerged, which can monitor various environmental parameters in real time and perform data analysis. The tunnel environment is affected by various factors such as geological activities, traffic loads, and construction vibrations. These factors can cause vibrations and settlements to the power lines;

[0003] In a tunnel environment, one of the factors with a marked interference effect is vibration. In most cases, due to the seismic design of the tunnel, ordinary vibrations do not cause too much impact. However, once their frequency or intensity exceeds the safety standard, it may, in certain situations, lead to sudden line failures, thereby affecting power transmission;

[0004] Traditional identification, early warning, and monitoring methods often focus on global general screening. The acquisition of abnormal data can be completed by collecting data through sensors and the like. However, for the early warning of future abnormal data, if global prediction is still carried out, not only is the accuracy insufficient, but it is also time-consuming, resulting in a waste of computing power and storage resources. It is difficult to accurately identify the designated areas with analytical value for targeted prediction and analysis, resulting in insufficient sensitivity of fault monitoring, easily missing local potential hazards. In the case of long tunnels and a large amount of data collection, the resource consumption is significant, and a lot of important abnormal trend information fails to be effectively utilized, and it is difficult to adaptively adjust the scales of several identifiable areas according to the quality of the predicted trends. Summary of the Invention

[0005] (I) Technical Problems to be Solved

[0006] In view of the above-mentioned drawbacks of the prior art, the present invention provides an intelligent early warning and monitoring system for power line failures in a tunnel environment, which can effectively solve the problems of the prior art.

[0007] (II) Technical Solutions

[0008] To achieve the above object, the present invention is realized through the following technical solutions:

[0009] The present invention discloses an intelligent early warning and monitoring system for power line failures in a tunnel environment, including:

[0010] The instruction control module is used to access the total power grid of the target tunnel and obtain the authority to issue operation data access and control instructions;

[0011] The line planning module, based on the topological structure of the target tunnel power line and the historical fault point distribution, aims at the historical fault distribution of the same point data, divides the overall line into different scales, and outputs several segments of different scale lines with physical coordinates and electrical parameters;

[0012] The data acquisition module is used to deploy distributed Internet of Things sensors to collect power fluctuation data, humidity, temperature and vibration data at specified intervals;

[0013] The anomaly correlation module is used to analyze the collected data, identify abnormal power fluctuations, humidity and temperature data exceeding the threshold, and synchronously obtain the vibration data of the planned scale line to which the abnormal fluctuations, humidity and temperature parameters belong. Based on the type amplitude of the vibration data, obtain the power fluctuation data, humidity, temperature data and line setting attribute data of all associated lines with the vibration data of this type amplitude;

[0014] The model establishment module is used to construct an identification model. The model inputs the extracted associated line data, and the model outputs the data change trend in the future preset period;

[0015] The trend prediction module is used to judge the positive and negative states of the data trend;

[0016] The alarm prompt module is used to be triggered when the data trend is negative and give an alarm prompt;

[0017] The line definition unit is used to calculate several cut-off points according to the data with a negative data change trend, cut the identified line according to the cut-off points, obtain several segments of scale lines, and provide separate acquisition data feedback as independent segments in the current period. If the data trend is positive, the adjacent scale lines are merged in the current period.

[0018] Furthermore, sub-modules are deployed under the anomaly correlation module. The sub-modules include: an anomaly detection module, a vibration analysis module and an association extraction module. The anomaly detection module and the vibration analysis module are connected through an electrical medium, and the vibration analysis module and the association extraction module are connected through an electrical medium. Among them:

[0019] The anomaly detection module is used to compare all the data participating in the detection according to a preset standard, detect abnormal values through statistical methods, record the time, location and specific values of the abnormal data, and mark the identified abnormal data as traceable items;

[0020] A vibration analysis module, which is used to collect vibration data when abnormal data occurs, classify the vibration data, identify its type and amplitude characteristics, classify the vibration data according to the characteristics, and correspondingly match several planned scale lines with relevant vibration characteristics;

[0021] An association extraction module, which is used to perform association processing on the abnormal data identified by the abnormal detection module and the vibration data provided by the vibration analysis module, and obtain power fluctuation, humidity, temperature and line setting attribute data belonging to the same planned scale line.

[0022] Furthermore, the process of association processing in the association extraction module is as follows:

[0023] S1: Based on the vibration data type and amplitude threshold, filter out the vibration characteristic data with the same line segment as the current abnormal power fluctuation, humidity or temperature data, and establish a mapping relationship between vibration and abnormal parameters;

[0024] S2: According to the acquisition timestamp of the data acquisition module and the line segment identifier, bind the abnormal power fluctuation, humidity, temperature data and vibration data according to the time synchronization and spatial same-scale line segment;

[0025] S3: Extract the line setting attribute data corresponding to the bound data, including line material, installation years and seismic resistance level, to form a multi-dimensional association data set;

[0026] S4: According to the association results of steps S1 - S3, only extract the historical and real-time data of the line segment with the vibration amplitude exceeding the standard and its adjacent preset range line segments as the input characteristic parameters of the trend prediction module, and mark the dynamic division identifier of the line segment.

[0027] Furthermore, the trend prediction module is based on the associated line data, uses time series prediction algorithms and multi-variable coupling analysis, inputs the power fluctuation, humidity, temperature and vibration characteristic parameters of the associated line segment in the current period, and outputs the predicted value of the abnormal fluctuation amplitude, the temperature and humidity change gradient and the trend score of the line segment in the future preset period.

[0028] Furthermore, the sensors deployed by the data acquisition module include three-axis vibration sensors, temperature and humidity sensors, and electric field distortion composite sensors. The data acquisition module performs downsampling storage on normal data, records the full frequency band of abnormal data, supports phase synchronous acquisition of vibration and electrical parameters, and performs preliminary filtering on the acquired data to remove noise processing.

[0029] Furthermore, sub-modules are deployed under the line definition unit. The sub-modules include: a slitting calculation module and a line integration module. The slitting calculation module and the line integration module are interconnected through a wireless network, where:

[0030] The slitting calculation module is used to calculate the optimal splitting point according to the degree of trend change. If the trend is negative, it identifies the splitting points of the scale line, splits the line according to the splitting points, forms independent scale lines, and records the characteristic data of each line segment.

[0031] The line integration module is used to merge adjacent scale lines when the trend is positive, integrate the power fluctuation, humidity, and temperature data of the corresponding independent lines, and generate new scale lines.

[0032] Furthermore, the working logic of the slitting calculation module is as follows:

[0033] Quantify and process the negative trend data of each line segment output by the trend prediction module to obtain the quantization parameters of voltage fluctuation variance, temperature rise gradient, and vibration energy spectral density change rate.

[0034] Establish a segmented optimization objective function for the line segment to be slit, and the formula is:

[0035]

[0036] In the formula, J(k) represents the line slitting quality evaluation function, and the smaller the value, the higher the comprehensive optimization degree of the segmentation scheme. k represents the number of candidate segments, w i represents the trend deterioration weight coefficient of the i-th segment, represents the comprehensive variance of the prediction data within this segment, Y represents the model complexity penalty factor, L avg represents the current average line segment length, and a represents the non-linear adjustment index;

[0037] Use the dynamic programming algorithm to solve the optimal number of segments k and its set of splitting points under the constraint conditions {k min ≤ k ≤ k max ∣ k max = f loor (1.5L / L b )}, where L is the total length of the line, and L b is the basic segment length threshold;

[0038] Apply a segmentation density constraint to the dangerous area with a bend radius less than R, and force overlapping monitoring buffers to be generated on both sides of this area. The buffer length L b = β · v · Δt, where v is the average vibration speed of the nearest three cycles, Δt is the data acquisition cycle, and β is the buffer coefficient;

[0039] When the difference degree of the deterioration trend between adjacent segments exceeds the threshold Y, activate the asymmetric splitting strategy, and adopt a denser splitting point layout on the side with a larger deterioration gradient.

[0040] Furthermore, the slitting calculation module and the line integration module are electrically connected to a configuration module through a dielectric medium. The configuration module is used to configure the division parameters to the line planning module according to the division performance of the slitting calculation module and the line integration module in the current cycle, and support manual active adjustment of the line division standard.

[0041] Furthermore, the instruction control module is wirelessly connected to a cloud storage module. The cloud storage module is used to store all the collected data and analysis data, perform cloud distributed storage, store the spatio-temporal evolution mode and disposal solutions of fault cases, automatically generate a fault disposal decision tree through cognitive map technology, and automatically eliminate obsolete fault mode data.

[0042] Furthermore, the instruction control module and the line planning module are wirelessly connected. The data acquisition module is connected to the instruction control module and the anomaly correlation module through a dielectric medium. The anomaly correlation module and the model establishment module are wirelessly connected. The model establishment module and the trend prediction module are wirelessly connected. The trend prediction module is wirelessly connected to the alarm prompt module and the line definition unit.

[0043] (III) Beneficial effects

[0044] Adopting the technical solution provided by the present invention, compared with the known prior art, it has the following beneficial effects:

[0045] 1. Through dynamically adjustable multi-scale line division, breaking through the traditional global fixed monitoring mode, based on historical fault distribution and topological structure, the line is divided into independent monitoring units of different scales, and supports dynamic merging or slitting according to data trends, dynamically adjusts the line division scale according to the degree of trend deterioration and historical data analysis, flexibly selects the optimal splitting point under different conditions using optimization algorithms, can update the accuracy of data processing in real time according to state changes, realizes local refined monitoring, and avoids waste of computing power.

[0046] 2. By introducing the classification of vibration data and the spatio-temporal binding of multi-dimensional parameters, establishing the mapping relationship between vibration types and abnormal parameters, through multi-dimensional associated data sets, accurately positioning the combined impact of vibration over-standard areas on power lines, reducing the false alarm rate, using a prediction model, combining the dynamic strategies of triggering slitting by negative trends and triggering merging by positive trends, realizing the adaptive optimization of monitoring granularity, and reducing redundant calculations while ensuring sensitivity by dynamically adjusting the monitoring range.

[0047] 3. By imposing segment density constraints and overlapping monitoring buffers on high-risk areas such as bends, and dynamically adjusting the buffer length in combination with vibration speed, through space-adaptive encrypted monitoring, early capture of potential line fatigue hazards caused by the accumulation of vibrations at bends. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0049] Figure 1 It is a framework schematic diagram of the present invention;

[0050] Figure 2 It is a framework schematic diagram of the abnormal association module in the present invention;

[0051] Figure 3 It is a framework schematic diagram of the line definition unit in the present invention.

[0052] The reference numerals in the figure respectively represent: 1. Instruction control module; 2. Line planning module; 3. Data acquisition module; 4. Abnormal association module; 41. Abnormal detection module; 42. Vibration analysis module; 43. Association extraction module; 5. Model establishment module; 6. Trend prediction module; 7. Alarm prompt module; 8. Line definition unit; 81. Slitting calculation module; 82. Line integration module; 83. Configuration module; 9. Cloud storage module. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] In order to make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0054] The following further describes the present invention with reference to the embodiments.

[0055] ① Embodiment 1

[0056] The intelligent early warning and monitoring system for power line faults in the tunnel environment of this embodiment, as Figure 1 shown, includes:

[0057] The instruction control module 1 is used to access the total power grid of the target tunnel and obtain the permission to issue operation data access and control instructions; the instruction control module 1 is connected to the cloud storage module 9 through wireless network interaction. The cloud storage module 9 is used to store all the collected data and analysis data, perform cloud distributed storage, store the spatio-temporal evolution pattern and disposal plan of fault cases, automatically generate a fault disposal decision tree through cognitive graph technology, and automatically eliminate outdated fault mode data;

[0058] The line planning module 2 divides the overall line into different scales based on the topological structure of the target tunnel power line and the historical fault point distribution, with the historical fault distribution of the same point data as the target, and outputs several different-scale lines with physical coordinates and electrical parameters;

[0059] The data acquisition module 3 is used to deploy distributed Internet of Things sensors to collect power fluctuation data, humidity, temperature and vibration data at specified intervals; the sensors deployed by the data acquisition module 3 include three-axis vibration sensors, temperature and humidity sensors, and electric field distortion composite sensors. The data acquisition module 3 downsamples and stores normal data, records all frequency bands of abnormal data, supports phase-synchronous acquisition of vibration and electrical parameters, and performs preliminary filtering on the collected data to remove noise;

[0060] The anomaly correlation module 4 is used to analyze the collected data, identify abnormal power fluctuations, humidity and temperature data exceeding the threshold, and simultaneously obtain the vibration data of the planned scale line to which the abnormal fluctuations, humidity and temperature parameters belong. Based on the type and amplitude of the vibration data, obtain the power fluctuation data, humidity, temperature data and line setting attribute data of all associated lines with the vibration data of this type and amplitude;

[0061] As Figure 2 shown, the anomaly correlation module 4 has sub-modules deployed at the lower level. The sub-modules include: the anomaly detection module 41, the vibration analysis module 42 and the correlation extraction module 43. The anomaly detection module 41 and the vibration analysis module 42 are connected through an electrical medium, and the vibration analysis module 42 and the correlation extraction module 43 are connected through an electrical medium, where:

[0062] The anomaly detection module 41 is used to compare all the data participating in the detection according to a preset standard, detect abnormal values through statistical methods, record the time, location and specific values of the abnormal data, and mark the identified abnormal data as traceable items;

[0063] The vibration analysis module 42 is used to collect the vibration data when the abnormal data occurs, classify the vibration data, identify its type and amplitude characteristics, classify the vibration data according to the characteristics, and correspondingly match several planned scale lines with relevant vibration characteristics;

[0064] The association extraction module 43 is used to perform an association process on the abnormal data identified by the anomaly detection module 41 and the vibration data provided by the vibration analysis module 42, and obtain the power fluctuation, humidity, temperature, and line setting attribute data belonging to the same planned scale line;

[0065] The model establishment module 5 is used to construct an identification model. The model inputs the extracted associated line data, and the model outputs the data change trend in a future preset period;

[0066] The trend prediction module 6 is used to judge the positive and negative states of the data trend; according to the deviation degree between the predicted value and the safety threshold, calculate the trend deterioration index. If the index exceeds the dynamically adjusted threshold, it is determined as "negative trend", otherwise it is determined as "positive trend"; the trend prediction module 6 is based on the associated line data, uses the time series prediction algorithm and multivariate coupling analysis, inputs the power fluctuation, humidity, temperature, and vibration characteristic parameters of the associated line segment in the current period, and outputs the predicted value of the abnormal fluctuation amplitude, the temperature and humidity change gradient, and the trend score of the line segment in the future preset period;

[0067] The alarm prompt module 7 is used to be triggered when the data trend changes negatively and perform an alarm prompt;

[0068] The line definition unit 8 is used to calculate several segmentation points according to the data with a negative change trend of the data, segment the identified line according to the segmentation points, obtain several scale lines, and provide the collected data feedback independently as an independent segment in the current period. If the data trend changes positively, the adjacent scale lines are merged in the current period;

[0069] As a preferred implementation manner in this embodiment, as Figure 3 shown, there are sub-modules deployed at the lower level of the line definition unit 8. The sub-modules include: a segmentation calculation module 81 and a line integration module 82. The segmentation calculation module 81 and the line integration module 82 are interconnected through a wireless network, where:

[0070] The segmentation calculation module 81 is used to calculate the optimal segmentation point according to the degree of trend change. If the trend is negative, identify the segmentation points of the scale line, segment the line according to the segmentation points, form independent scale lines, and record the characteristic data of each line segment;

[0071] The line integration module 82 is used to merge adjacent scale lines when the trend is positive, and integrate the power fluctuation, humidity, and temperature data of the corresponding independent lines to generate a new scale line;

[0072] The slitting calculation module 81 and the line integration module 82 are electrically connected to a configuration module 83 through a dielectric medium. The configuration module 83 is used to configure division parameters for application to the line planning module 2 according to the division performance of the slitting calculation module 81 and the line integration module 82 in the current cycle, and supports manual active adjustment of the line division standard;

[0073] The instruction control module 1 and the line planning module 2 are connected through a wireless network. The data acquisition module 3 is connected to the instruction control module 1 and the anomaly correlation module 4 through a dielectric medium. The anomaly correlation module 4 and the model establishment module 5 are connected through a wireless network. The model establishment module 5 and the trend prediction module 6 are connected through a wireless network. The trend prediction module 6 and the alarm prompt module 7 and the line definition unit 8 are connected through a wireless network.

[0074] Compared with the prior art, in this embodiment, the distributed Internet of Things sensors are used to collect power fluctuations, humidity, temperature, and vibration data in real time, which can detect potential faults more timely, improve the monitoring ability of the line state, and use the model to analyze the future data trend and dynamically adjust the threshold, so as to realize the early warning of faults, thus effectively avoiding the expansion of faults and the greater losses brought;

[0075] Through the effective association of abnormal data and vibration data, various parameters related to faults can be accurately identified, thus further improving the accuracy of fault identification. It has the ability to split and merge lines, dynamically adjusts according to the change of data trend, improves the flexibility and adaptability of the line, and better adapts to the line operation conditions in different large environments;

[0076] By monitoring power fluctuations and taking into account various factors such as humidity, temperature, and vibration, through multi-dimensional analysis, potential risks are more comprehensively evaluated. Through the alarm prompts and data trend analysis provided by the system, managers can quickly respond, optimize management decisions, thus improving the overall operation efficiency. Combining the time series prediction algorithm with multi-variable coupling analysis, it has higher technological advancement, can process complex line data and trend changes, and improves the scientificity and accuracy of fault prediction.

[0077] ② Embodiment 2

[0078] On other levels, this embodiment also provides another optimization mechanism based on Embodiment 1, specifically the working logic of a slitting calculation module 81, specifically:

[0079] Quantize the negative trend data of each line segment output by the trend prediction module 6 to obtain quantization parameters of voltage fluctuation variance, temperature rise gradient, and vibration energy spectral density change rate;

[0080] Establish a segmented optimization objective function for the line segment to be slit, and the formula is:

[0081]

[0082] In the formula, J(k) represents the line cutting quality evaluation function. The smaller the value, the higher the comprehensive optimization degree of the segmentation scheme. k represents the number of candidate segments, w i represents the trend deterioration weight coefficient of the i-th segment, represents the comprehensive variance of the predicted data within this segment, Y represents the model complexity penalty factor, which is dynamically adjusted according to the tunnel type: rock tunnel: 0.18; soft soil tunnel: 0.25; composite stratum: 0.22, L avg represents the current average segment length of the line, which is calculated as the ratio of the total effective line length within the current monitoring period to the historical average number of segments. a represents the non-linear adjustment index, and the value rule is: when the main vibration frequency is greater than 100 Hz, a = 1.2; when the main vibration frequency is between 50 - 100 Hz, a = 1.0; when the main vibration frequency is less than 50 Hz, a = 0.8;

[0083] The dynamic programming algorithm is used to solve the optimal number of segments k and its set of cut points under the constraint conditions {k min ≤ k ≤ k max ∣ k max = f loor (1.5L / L b )}, where L is the total line length and L b is the basic segment length threshold;

[0084] Segment density constraints are imposed on the dangerous area with a curve radius less than R, and overlapping monitoring buffers are forced to be generated on both sides of this area. The buffer length L b = β·v·Δt, where v is the average vibration velocity in the last three periods, Δt is the data acquisition period, and β is the buffer coefficient;

[0085] When the difference degree of the deterioration trend between adjacent segments exceeds the threshold Y, the asymmetric segmentation strategy is activated, and a denser cut point layout is adopted on the side with a larger deterioration gradient;

[0086] The non-linear combination of multi-dimensional deterioration indicators is introduced, and the coupling weighting of electrical parameters, mechanical vibration, and environmental changes is realized through w i A variable upper limit dynamic constraint mechanism is adopted to ensure that the segmentation granularity scales adaptively with the line length, avoid the waste of computing power caused by over-segmentation of short lines. Through the overlapping buffer technology, the edge effect in the vibration wave propagation process can be effectively captured, the false alarm rate can be reduced. With the asymmetric segmentation strategy, it is optimized for the anisotropic characteristics of vibration transmission in the tunnel to improve the monitoring resolution of key areas. The spatial attenuation effect is introduced to make the segmentation penalty of long lines increase super-linearly, effectively suppressing the mis-segmentation caused by the superposition of mechanical vibration waves.

[0087] ③ Example 3

[0088] This embodiment provides a process for association processing, specifically as follows:

[0089] S1: Based on the vibration data type and amplitude threshold, filter out the vibration characteristic data that belongs to the same line segment as the current abnormal power fluctuation, humidity or temperature data, and establish a mapping relationship between vibration and abnormal parameters;

[0090] S2: According to the acquisition timestamps and line segment identifiers of the data acquisition module 3, bind the abnormal power fluctuation, humidity, and temperature data with the vibration data according to time synchronization and spatially co-scaled line segments; during the spatio-temporal synchronization association process, taking the trigger moment of the vibration data as the reference, trace back a preset time window forward to obtain the fluctuation peaks and change rates of the abnormal power fluctuation, humidity, and temperature data within the same line segment. For the vibration data that exceeds the safety threshold, match its physically adjacent scaled line segments according to the line topology structure, and extract the sensor data and attribute parameters of the adjacent line segments;

[0091] S3: Extract the line setting attribute data corresponding to the bound data, including line material, installation years, and seismic resistance level, to form a multi-dimensional association data set;

[0092] S4: According to the association results of steps S1 - S3, only extract the historical and real-time data of the line segments whose vibration amplitudes exceed the standard and their adjacent preset range line segments as the input characteristic parameters of the trend prediction module 6, and mark the dynamic division identifiers of the line segments. The dynamic division identifiers are used for the adaptive slicing or merging operations of the cloud storage module 9. Specifically: when the association intensity between the vibration data and the abnormal data exceeds the preset risk threshold, start the independent monitoring mode of the line segment and generate the coordinate parameters of the cut point. Different from the traditional global scan, using the vibration data exceeding the standard as the trigger condition, only associate the abnormal line segment and the data of the adjacent area, reduce redundant calculations, and combine the timestamp, line segment identifier, and physical topology relationship to ensure the accuracy of data association. Map the association result directly to the line division operation to form a closed loop of data association and dynamic adjustment, reflecting the self-adaptability of the system.

[0093] Working principle: When the present invention is carried, the instruction control module 1 accesses the tunnel power main network through the Internet of Things and connects to the cloud database cloud storage module 9. The line planning module 2 divides the line into multi-scale physical coordinate segments based on the topological structure. The data acquisition module 3 deploys a composite sensor network to collect environmental data and preprocess it. The anomaly correlation module 4 identifies outliers through the sub-module anomaly detection module 41, analyzes vibration characteristics through the vibration analysis module 42, and correlates multi-dimensional data through the correlation extraction module 43. The model establishment module 5 constructs a prediction model, inputs the correlated data, and outputs the trend. The trend prediction module 6 judges the positive or negative of the trend based on the deterioration index. If it is negative, the alarm prompt module 7 gives an alarm, and triggers the line definition unit 8 to divide the fault segment through the sub-module segmentation calculation module 81, and the line integration module 82 merges the stable segments. At the same time, the configuration module 83 dynamically optimizes the division parameters and feeds them back to the line planning module 2 to form a closed-loop iteration;

[0094] By deploying a composite sensor, it realizes the phase-synchronized acquisition and noise filtering of power, temperature and humidity, and vibration parameters. Combining the fault decision tree generated by the cloud cognitive map and the historical spatio-temporal pattern analysis, it effectively improves the data correlation. Adopting the dynamic line segmentation technology, it adjusts the monitoring scale in real time according to the trend deterioration index, and realizes the leap from fixed threshold judgment to multi-variable prediction of fault warning;

[0095] At the same time, it optimizes the resource utilization rate through the distributed storage downsampling and full-band anomaly recording strategy. Through the refined diagnosis ability and the agile response characteristics of topological adaptability, it significantly improves the accuracy of power line fault identification and the timeliness of early warning in complex tunnel environments.

[0096] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements will not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. Intelligent early warning and monitoring system for power line faults in tunnel environment, characterized by: include: The command control module is used to access the power line network of the target tunnel and obtain the permission to access the operation data and issue control commands; The line planning module is based on the topological structure of the target tunnel power line and the historical fault point distribution. It takes the historical fault distribution of the same point data as the target, divides the overall line into different scales, and outputs several sections of different scale lines with physical coordinates and electrical parameters; Data collection module, used to deploy distributed IoT sensors to collect power fluctuation data, humidity, temperature and vibration data at specified periods; The abnormal association module is used to analyze the collected data, identify the abnormal power fluctuation, humidity and temperature data exceeding the threshold, and synchronously obtain the vibration data of the planned scale line based on the abnormal fluctuation, humidity and temperature parameters, and based on the type and amplitude of the vibration data, obtain the power fluctuation data, humidity, temperature data and line setting attribute data of all associated lines with the type and amplitude vibration data; The model building module is used to build the recognition model. The model inputs the extracted associated line data and outputs the data change trend of the future preset period. Trend prediction module, used to determine the positive and negative state of data trends; The alarm prompt module is used to trigger an alarm prompt when the data trend changes negatively; The line definition unit is used to calculate several cutting points according to the data with negative change trend, cut the identified line according to the cutting points, obtain several scale lines, and provide collection data feedback as independent segments in the current cycle. If the data trend is positive change, the adjacent scale lines will be merged in the current cycle.

2. The intelligent early warning and monitoring system for power line faults in tunnel environment according to claim 1 is characterized in that: The abnormal association module is deployed with submodules at the lower level, and the submodules include: an abnormality detection module, a vibration analysis module and an associated extraction module. The abnormality detection module and the vibration analysis module are interactively connected through an electrical medium, and the vibration analysis module and the associated extraction module are interactively connected through an electrical medium, wherein: The anomaly detection module is used to compare all the data involved in the detection according to the preset standards, detect abnormal values ​​through statistical methods, record the time, location and specific value of the abnormal data, and mark the identified abnormal data as traceable items; The vibration analysis module is used to collect vibration data when abnormal data occurs, classify the vibration data, identify its type and amplitude characteristics, classify the vibration data according to the characteristics, and match several planned scale routes with relevant vibration characteristics; The association extraction module is used to associate the abnormal data identified by the abnormal detection module with the vibration data provided by the vibration analysis module to obtain the power fluctuation, humidity, temperature and line setting attribute data belonging to the same planning scale line.

3. The intelligent early warning and monitoring system for power line faults in tunnel environment according to claim 2 is characterized in that: The process of association processing in the association extraction module is as follows: S1: Based on the vibration data type and amplitude threshold, the vibration characteristic data with the same line segment as the current abnormal power fluctuation, humidity or temperature data is screened out, and a mapping relationship between vibration and abnormal parameters is established; S2: According to the acquisition timestamp and line segment identification of the data acquisition module, the abnormal power fluctuation, humidity, temperature data and vibration data are bound according to time synchronization and spatial same-scale line segments; S3: extracting the line setting attribute data corresponding to the bound data, including line material, installation years and seismic resistance level, to form a multi-dimensional associated data set; S4: According to the correlation results of steps S1-S3, only the historical and real-time data of the line segment with excessive vibration amplitude and its adjacent preset range line segment are extracted as input feature parameters of the trend prediction module, and the dynamic division identification of the line segment is marked.

4. The intelligent early warning and monitoring system for power line faults in tunnel environment according to claim 1 is characterized in that: The trend prediction module is based on the associated line data, uses time series prediction algorithm and multivariate coupling analysis, inputs the power fluctuation, humidity, temperature and vibration characteristic parameters of the associated line segment in the current cycle, and outputs the predicted value of the abnormal fluctuation amplitude, temperature and humidity change gradient and trend score of the line segment in the future preset cycle.

5. The intelligent early warning and monitoring system for power line faults in tunnel environment according to claim 1 is characterized in that: The sensors deployed by the data acquisition module include a three-axis vibration sensor, a temperature and humidity sensor, and an electric field distortion composite sensor. The data acquisition module downsamples and stores normal data, records abnormal data in the full frequency band, supports phase synchronization acquisition of vibration and electrical parameters, performs preliminary filtering on the collected data, and removes noise.

6. The intelligent early warning and monitoring system for power line faults in tunnel environment according to claim 1 is characterized in that: The line definition unit has submodules deployed at the lower level, including: a segmentation calculation module and a line integration module, the segmentation calculation module and the line integration module are interactively connected via a wireless network, wherein: The segmentation calculation module is used to calculate the optimal segmentation point according to the trend change degree. If the trend is negative, the segmentation point of the scale line is identified, the line is segmented according to the segmentation point to form an independent scale line, and the characteristic data of each segment of the line is recorded; The line integration module is used to merge adjacent scale lines when the trend is positive, integrate the power fluctuation, humidity and temperature data of the corresponding independent lines, and generate a new scale line.

7. The intelligent early warning and monitoring system for power line faults in tunnel environment according to claim 6 is characterized in that: The working logic of the segmentation calculation module is: Quantitatively process the negative trend data of each line section output by the trend prediction module to obtain the quantitative parameters of voltage fluctuation variance, temperature rise gradient and vibration energy spectrum density change rate; A segment optimization objective function is established for the line segment to be cut, and the formula is: In the formula, J(k) represents the line segmentation quality evaluation function. The smaller the value, the higher the comprehensive optimization degree of the segmentation scheme. k represents the number of candidate segments, and w i represents the trend deterioration weight coefficient of the i-th segment, represents the comprehensive variance of the predicted data in the segment, Y represents the model complexity penalty factor, and L avg represents the average segment length of the current line, and a represents the nonlinear adjustment index; Using dynamic programming algorithm under the constraints {k min ≤k≤k max ∣k max =f loor (1.5L / L b )}, where L is the total length of the line, and L b is the basic segment length threshold; A segmented density constraint is imposed on the dangerous area with a curve radius less than R, forcing the generation of overlapping monitoring buffers on both sides of the area, with a buffer length of L b =β·v·Δt, where v is the average vibration velocity of the last three cycles, Δt is the data collection period, and β is the buffer coefficient; When the difference in deterioration trends between adjacent segments exceeds the threshold Y, the asymmetric segmentation strategy is activated, and an encrypted split point layout is adopted on the side with a larger deterioration gradient.

8. The intelligent early warning and monitoring system for power line faults in tunnel environment according to claim 6 is characterized in that: The segmentation calculation module and the line integration module are interactively connected to a configuration module through an electrical medium. The configuration module is used to configure the segmentation parameters to the line planning module application according to the segmentation performance of the segmentation calculation module and the line integration module in the current cycle, and supports manual active adjustment of the line segmentation standard.

9. The intelligent early warning and monitoring system for power line faults in tunnel environment according to claim 1 is characterized in that: The instruction control module is interactively connected to a cloud storage module via a wireless network. The cloud storage module is used to store all collected data and analysis data, perform cloud distributed storage, store the spatiotemporal evolution patterns and treatment plans of fault cases, automatically generate a fault treatment decision tree through cognitive graph technology, and automatically eliminate obsolete fault mode data.

10. The intelligent early warning and monitoring system for power line faults in tunnel environment according to claim 1, characterized in that: The instruction control module is interactively connected to the line planning module through a wireless network, the data acquisition module is connected to the instruction control module and the abnormal association module through an electrical medium, the abnormal association module is interactively connected to the model building module through a wireless network, the model building module is interactively connected to the trend prediction module through a wireless network, and the trend prediction module is interactively connected to the alarm prompt module and the line definition unit through a wireless network.

Images