A cloud-edge collaborative wake-up cable channel anti-external damage monitoring system

Abstract

The application discloses a kind of cable channel prevents outside breaking monitoring system of cloud edge coordination wake-up, belong to cable safety monitoring technical field, for solving the industry pain point of high false alarm rate of existing scheme, insufficient field endurance, slow early warning response.This system acquires triaxial vibration signal by multidirectional monitoring unit, and dynamic vibration threshold is generated by cloud in combination with geological parameters to determine suspected risk, and the visual perception end is awakened on demand to complete risk review, and local audible alarm and remote monitoring notification are triggered synchronously after identifying dangerous factors;Meanwhile, there is an energy-saving gradient scheduling mechanism, which can dynamically adjust the monitoring frequency according to the remaining power.The application can effectively reduce the false alarm rate, greatly improve the equipment endurance in the field without electricity, and realize early pre-warning of construction risk.

Description

Technical Field

[0001] This invention relates to the field of cable safety monitoring, and in particular to a cloud-edge collaborative wake-up cable channel anti-external damage monitoring system. Background Technology

[0002] In recent years, with the increasing coverage of urban power grid cables, higher demands have been placed on the adaptability of underground cables to complex environments. In particular, related operations in above-ground construction areas pose a significant threat to the safety of underground cables. To prevent damage from construction machinery and subsequent cable failures, existing technologies have proposed relevant monitoring schemes:

[0003] Chinese Patent Publication No. CN121686658A discloses a smart warning post and detection method for preventing external damage to power transmission lines. The method includes: collecting environmental data through an infrared sensor array and a micro-vibration sensor array; calculating wave velocity and propagation direction of the vibration data locally at the front end; filtering out interference vibrations from vehicles, pedestrians, etc.; and when suspected construction vibrations are detected, waking up the local intelligent control motherboard and initiating the front end's local visual recognition algorithm to analyze the construction behavior, combined with solar power for field monitoring. However, the decision-making logic of this scheme is entirely local at the front end, using static wave velocity and time thresholds, which cannot be dynamically adjusted according to different cable burial depths, soil types, and regional risks. This results in a high false alarm rate in soft soil areas and insufficient sensitivity in high-risk areas. Simultaneously, the local visual recognition algorithm significantly increases the front end's power consumption, leading to insufficient battery life during continuous rainy days and a high risk of equipment downtime, resulting in frequent manual maintenance. Furthermore, the slow response speed of the local front end's decision-making prevents early warning and protection.

[0004] Chinese Patent Publication No. CN119618368A discloses an intelligent vibration monitoring and adaptive optimization device based on neural networks. This device includes: a single vibration sensor collecting vibration data; local filtering and noise reduction of the data at the front end; extraction of variance, peak value, and power spectral characteristics; running a neural network model to calculate vibration triggering coefficients; local identification of vibration sources; and updating the local model and trigger threshold through historical samples to achieve self-optimization. However, all calculations in this solution are completed locally at the front end. The continuous operation of the neural network significantly increases the power consumption of the front-end processor, resulting in insufficient battery life. In outdoor environments without mains power, frequent battery replacements are required, leading to extremely high maintenance costs. Furthermore, this solution only optimizes vibration data without visual verification, resulting in a still relatively high false alarm rate, failing to meet the requirements for long-term stable monitoring of cable channels.

[0005] Chinese Patent Publication No. CN114360184A discloses a method and system for monitoring cable channels against external damage through multi-device linkage. This method includes distributed sensing via multiple devices such as vibration monitoring, cover plate movement monitoring, cable pile monitoring, and fiber optic monitoring. When a front-end device triggers an anomaly locally, it reports it to a server. The server issues an alarm command and simultaneously activates a front-end camera to capture video. The analysis results are then pushed to a human terminal, where a false alarm is identified and the local model is updated. However, this solution relies on a human terminal to identify false alarms, resulting in a slow response time and an inability to provide proactive protection. Often, confirmation is only achieved after the potential hazard has already occurred. Furthermore, the constant high-power communication of multiple devices leads to high energy consumption, poor long-term operational stability, low equipment online rate, and frequent manual debugging and maintenance, failing to meet the needs of long-term field monitoring.

[0006] Chinese Patent Publication No. CN114613540A discloses a method, system, device, and storage medium for monitoring cable damage from external sources. This solution collects vibration and image data at the front end, performs anomaly identification locally, and reports it to the cloud. It combines multi-source data to achieve early warning of external damage and supports remote configuration via the cloud. However, the front end of this solution still requires the identification algorithm to run continuously, which cannot solve the power consumption bottleneck in the field. Furthermore, the static trigger threshold used has poor adaptability to different geological regions, resulting in a high false alarm rate. Long-term operation still requires frequent manual debugging and maintenance.

[0007] In summary, existing cable duct external damage monitoring solutions are either limited by the architecture of front-end local decision-making or by static threshold design, and neither can fundamentally solve the industry pain point of insufficient long-term operational stability of current cable detection systems, leading to high frequency of manual debugging and maintenance. Specifically, this manifests in the following three aspects: First, the false alarm rate still has room for improvement. The static thresholds of existing solutions cannot adapt to different burial depths, soil conditions, and regional environments, resulting in a persistently high false alarm rate. Frequent false alarms significantly increase the troubleshooting burden on maintenance personnel and cannot meet the needs of long-term stable monitoring. Second, the availability and effectiveness of outdoor energy supply... The contradictions in monitoring are irreconcilable. Existing solutions require the front end to run high-power identification algorithms or use high-power communication links, resulting in insufficient battery life in environments without mains power. Equipment is prone to downtime during continuous rainy days, requiring frequent manual maintenance and battery replacement, which is extremely costly. Thirdly, the early warning response is slow and cannot achieve pre-protection. Existing solutions either rely on slow response decisions at the front end or on false alarm judgments from manual terminals. They cannot complete identification and early warning in the early stages when construction machinery approaches. Often, the hidden danger has already occurred before confirmation is completed, which cannot effectively prevent accidents caused by cable damage. Summary of the Invention

[0008] To address this, the present invention provides a cloud-edge collaborative wake-up cable channel external damage monitoring system, which can improve the long-term operational stability of the cable monitoring system and reduce the frequency of manual maintenance. To achieve this technical effect, the present invention makes improvements in three aspects: reducing false alarm rate, balancing energy supply with necessary monitoring power consumption requirements, and providing timely early warnings to reduce losses.

[0009] To achieve the above objectives, the present invention provides a cloud-edge collaborative wake-up cable channel external damage prevention monitoring system, which includes:

[0010] A multi-directional monitoring unit is provided, which includes a first-directional monitoring end, a second-directional monitoring end and a third-directional monitoring end. Each monitoring end is used to acquire the monitoring data in its respective direction, including the first-directional monitoring data, the second-directional monitoring data and the third-directional monitoring data, in order to collect real-time vibration signals.

[0011] The first cloud processing unit is used to receive the monitoring data in the first direction, the monitoring data in the second direction, and the monitoring data in the third direction, and convert them into a chief supervisor's measurement value and a two-dimensional direction vector through built-in data processing logic. The chief supervisor's measurement value is then compared with a predetermined threshold to determine whether a suspected dangerous state has been entered. The predetermined threshold is the maximum acceptable vibration value related to the geological environment obtained through a built-in function model.

[0012] A local processing unit, located at the monitoring site, is used to receive and transmit the suspected danger signal emitted by the first cloud processing unit;

[0013] The visual perception end is integrated into the local processing unit, and is controlled by the local processing unit to acquire and transmit images of the seismic source.

[0014] The second cloud processing unit receives the earthquake source image obtained by the visual sensing terminal, identifies the earthquake source image, determines whether the danger factor exists, and decides whether to enter the danger state based on the determination result.

[0015] When the third cloud processing unit is notified by the second cloud processing unit that it has entered the dangerous state, it issues a maintenance warning to the monitoring personnel until the monitoring personnel complete the manual confirmation, and then switches back to the non-suspected dangerous state and cancels the maintenance warning.

[0016] The local alarm terminal is used to receive the activation signal from the second cloud processing unit and activate the alarm at the monitored site.

[0017] The energy-saving gradient scheduling unit is used to dynamically adjust the system's operating mode based on the remaining power.

[0018] After the monitoring personnel complete the manual confirmation, the system is restored to a non-suspected danger state, the maintenance warning is lifted, and the local alarm terminal stops alarming.

[0019] Furthermore, the built-in data processing logic of the first cloud processing unit includes:

[0020] The first direction monitoring quantity, the second direction monitoring quantity, and the third direction monitoring quantity obtained at the same moment are used as the data set at that moment. Multiple data sets are obtained within the smallest unit time to form a data list of the smallest unit time.

[0021] Take out the first direction monitoring quantity of each data set in the minimum unit time data list to obtain multiple first direction monitoring quantities, and calculate the arithmetic average of the multiple first direction monitoring quantities to obtain the average point of the first direction monitoring quantity.

[0022] The method for obtaining the average point of the second-direction monitoring quantity and the average point of the third-direction monitoring quantity from the second-direction monitoring quantity and the third-direction monitoring quantity is the same as the method for obtaining the average point of the first-direction monitoring quantity from the first-direction monitoring quantity;

[0023] The average points of the monitoring data in the first direction, the second direction, and the third direction are each squared and then added together to obtain the measurement value of the chief supervisor.

[0024] The average point of the first direction monitoring, the average point of the second direction monitoring, and the average point of the third direction monitoring are taken as the three directional components of a three-dimensional vector, which is the three-dimensional direction vector. The three-dimensional direction vector is then projected onto the plane where the cable extends, which is the two-dimensional direction vector.

[0025] Furthermore, the predetermined threshold value is related to the soil medium, cable burial depth, and regional environment where the multi-directional monitoring unit is located. It is a parameter pre-set by a built-in function model during the collaborative installation process of cable route deployment. The built-in function model is as follows: ;

[0026] in, The predetermined threshold value is the desired value, in units of (The square of gravitational acceleration);

[0027] The system baseline threshold is the minimum effective value that can trigger an alarm under standard operating conditions. It is also the system's default universal threshold, and its unit is [unit missing]. (The square of gravitational acceleration);

[0028] e is a natural constant in mathematics, which has no unit.

[0029] h represents the cable burial depth, in meters (m).

[0030] This is a regional environmental tolerance coefficient used to balance the risk of underreporting in high-risk work areas with the risk of false reporting in wilderness areas. It is determined manually by assessing the required tolerance level under different regional environments. The smaller the value, the higher the system sensitivity; conversely, the larger the value, the higher the tolerance. (No unit.)

[0031] Configuration method: The power grid operation and maintenance personnel manually configure the configuration in the background monitoring system according to the actual risk situation of the area.

[0032] The soil medium damping attenuation coefficient, derived from the Rayleigh wave attenuation equation, characterizes the absorption and dissipation capacity of different soil types for vibrational stress waves. It can be found in the local soil survey report, and the unit is 1 / 2 oz. (Attenuation rate per meter).

[0033] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0034] Reduce system false alarm rate and improve monitoring reliability: This invention, through its built-in dynamic threshold function model, can adaptively adjust the trigger threshold according to the cable burial depth and soil medium of different monitoring points. This solves the defect of static thresholds in existing technologies that cannot adapt to complex geology and environment, effectively filters environmental interference vibrations in different scenarios, and greatly reduces the burden of ineffective troubleshooting for operation and maintenance personnel.

[0035] Overcoming the bottleneck of power supply in the field and significantly improving system endurance: This invention adopts a cloud-edge collaborative on-demand wake-up architecture. Under normal circumstances, only the low-power multi-directional monitoring unit, local processing unit, and 4G communication channel operate, while the high-power visual perception end and 5G communication channel remain completely powered off and in hibernation. At the same time, the high-load recognition computing power is handled by the cloud, avoiding the high power consumption caused by the local algorithm operation at the front end. Combined with dynamic energy-saving gradient scheduling based on the remaining power, it achieves long-term stable operation in environments without mains power, solves the problem of traditional equipment being prone to downtime during continuous rainy days, extends the equipment maintenance cycle from several weeks to several months, and significantly reduces operation and maintenance costs.

[0036] Achieving rapid response and pre-emptive protection: This invention employs a rapid wake-up and cloud-based automatic verification mechanism, utilizing a cloud-based visual AI model for automatic verification to replace traditional manual terminal false alarm judgments. This significantly shortens the risk confirmation response chain, enabling risk confirmation to be completed without human intervention. It overcomes the shortcomings of existing technologies, such as slow response and inability to intervene in advance. It can complete identification and on-site audio-visual warnings in the early stages when construction machinery touches the cable, realizing a closed loop from "post-event evidence collection" to "pre-event interception," effectively preventing cable damage accidents. Attached Figure Description

[0037] Figure 1This is a flowchart illustrating the main logic of an embodiment of the present invention.

[0038] Figure 2 This is a flowchart illustrating the internal data processing logic of an embodiment of the present invention.

[0039] Figure 3 This is a flowchart illustrating the energy-saving gradient in an embodiment of the present invention. Detailed Implementation

[0040] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0041] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles and / or logic of the present invention and are not intended to limit the scope of protection of the present invention.

[0042] refer to Figure 1 , Figure 2 and Figure 3 This invention provides a cloud-edge collaborative wake-up cable channel external damage prevention monitoring system. The following will combine... Figures 1 to 3 The following is a detailed description of the hardware selection, connection relationships, and workflow of the system of the present invention:

[0043] I. Specific Implementation and Connection of Hardware System

[0044] In this embodiment, to achieve the above technical solution, the following specific selections were made for each core component of the system. Those skilled in the art can select equivalent devices with the same function according to actual needs. The scope of protection of this invention is not limited to the following specific models:

[0045] 1. Multi-directional monitoring unit: Employs a high-precision triaxial digital accelerometer ADXL355, which features extremely low noise density (22.5 μg / √Hz) and can accurately acquire real-time vibration monitoring quantities in three orthogonal directions (X, Y, and Z), corresponding to the first, second, and third direction monitoring quantities in this invention. This sensor normally maintains a low-power operating state, providing the system with underlying sensing data.

[0046] 2. Local processing unit: The STM32L4 series low-power microcontroller unit (MCU) is used as the core scheduling node of the field device, responsible for the preliminary filtering of sensor data, level control of hardware pins, and status management of communication links.

[0047] 3. Visual perception end: It adopts a high-definition camera module with integrated Ingenic T31 chip. This module supports FastBoot fast startup mechanism and can complete hardware initialization and image output within 500ms, which solves the problem of missed shots caused by the slow startup of traditional camera modules.

[0048] 4. A dual-channel communication module, comprising:

[0049] (1) First low-power narrowband communication module: adopts EC801-E 4G communication module, which is normally kept in the listening state, responsible for transmitting small data, and the standby power consumption is only at the milliwatt level.

[0050] (2) Second high-bandwidth communication module: The RM500U-CN 5G communication module is used. Under normal circumstances, it is in a completely power-off sleep state. It is only physically powered on by the local processing unit after being triggered to wake up, and is used to carry high-speed transmission of high-definition video streams.

[0051] 5. Local alarm terminal: It adopts a high-power audible and visual alarm light, and its power supply circuit is controlled by a relay. It is normally disconnected to save power. When triggered, it can produce an alarm sound of more than 110dB and a bright flashing light to achieve on-site deterrence.

[0052] 6. Power supply system: A 10W monocrystalline silicon solar panel is used as the natural energy receiving end to charge the 12V / 10Ah lithium iron phosphate battery pack (energy storage end), providing continuous energy supply for the entire system in the field without mains power.

[0053] II. Routine Low-Power Monitoring and Internal Data Processing Logic

[0054] Under normal circumstances, the system is in a non-dangerous state. At this time, the camera module, 5G communication module and alarm light are all in a power-off sleep state. Only the multi-directional monitoring unit, local processing unit and 4G communication module maintain low power consumption. The overall power consumption of the system is controlled within 100mW to maximize battery life.

[0055] like Figure 2 As shown, during this stage, the multi-directional monitoring unit continuously collects real-time acceleration data in the X, Y, and Z directions at a sampling rate of 100Hz, and transmits it to the first cloud processing unit at a transmission frequency of 0.5 seconds per transmission. The first cloud processing unit follows... Figure 2 The internal data processing logic shown preprocesses the raw data, and the specific steps are as follows:

[0056] Step 1: Data Window Aggregation: Using 100ms as the smallest unit time window, collect 10 sets of continuous triaxial acceleration sampling data within this window to form a data list with the smallest unit time.

[0057] Step 2: Moving Average Noise Reduction: Apply mean filtering to the sampled data within the window, and calculate the arithmetic mean of the monitored quantities in the X, Y, and Z directions respectively to obtain the average point of the monitored quantity in the first direction. Average point of monitoring in the second direction Third-party average monitoring points This step effectively filters out high-frequency environmental noise and improves data stability.

[0058] Step 3: Calculation of Supervisor Measurement Value: Based on the principle of spatial vector synthesis, the average components in the three directions are squared and summed to calculate the supervisor measurement value G. The calculation formula is as follows:

[0059] Among them, the chief supervisor's measured value G represents the overall intensity of the current environmental vibration and can be directly compared with the predetermined threshold.

[0060] Step 4: Calculate the two-dimensional direction vector: , , As a component of the three-dimensional vector, it is projected onto the horizontal plane along the cable's extension direction to obtain a two-dimensional direction vector indicating the location of the seismic source. This vector is used to guide the camera module in adjusting its shooting angle during the subsequent wake-up phase, ensuring accurate capture of the seismic source image.

[0061] After completing the above processing, the first cloud processing unit transmits the calculated chief measurement value G and the two-dimensional direction vector to the local processing unit via the 4G communication channel at a frequency of 1 second / time.

[0062] III. Dynamic Threshold Determination and Wake-up Mechanism Based on Built-in Function Model

[0063] like Figure 1 As shown, after receiving the vibration data reported by the front end, the first cloud processing unit does not use the traditional static threshold for comparison, but instead calls the built-in threshold judgment function model to calculate the dynamic predetermined threshold for the current monitoring point. This is used to determine whether a suspected dangerous situation has been entered.

[0064] The core formula of this threshold determination model is as follows:

[0065] ;

[0066] The definitions, value selection methods, and engineering implementation methods for each parameter are as follows:

[0067] Parameter 1: System baseline threshold :

[0068] Physical units: (the square of gravitational acceleration, );

[0069] Parameter meaning: The basic sensitivity boundary of the system under standard test conditions;

[0070] Calibration method: Calibration shall be carried out at a standard test site before mass production deployment of the product;

[0071] The test site set standard working conditions: clay layer, cable buried at a depth of 1 meter, and a 20-ton tracked excavator was used to simulate construction operations at a distance of 2 meters from the vertical projection point of the sensor. Through repeated tests, the minimum effective square value of the composite acceleration that can reliably trigger the alarm of the system was extracted, and a 15% safety margin was reserved, which was finally solidified as the benchmark threshold.

[0072] Engineering Example: After calibration, the system baseline threshold in this embodiment is configured as follows: .

[0073] Parameter 2: Cable burial depth parameter :

[0074] Physical unit: m (meter);

[0075] Parameter meaning: It represents the thickness of the physical protective layer. The greater the burial depth, the greater the attenuation of ground vibration transmitted to the cable. The system should have a higher tolerance to ground vibration in order to filter out shallow environmental interference.

[0076] Value acquisition method: No real-time measurement by front-end sensors is required. During the system deployment phase, the cable design burial depth of the monitoring point is directly read by connecting to the State Grid GIS geographic information system ledger or by consulting the pipeline engineering as-built drawings, and configured as a static parameter in the cloud node database;

[0077] Engineering Example: In this embodiment, the pipeline drawing for a certain monitoring point shows that the cable burial depth is 2.0 meters, therefore, [the following is a list of parameters] is configured. .

[0078] Parameter 3: Soil medium damping attenuation coefficient :

[0079] Physical units: (Attenuation rate per meter);

[0080] Parameter meaning: Derived from the Rayleigh wave attenuation equation, it characterizes the absorption and dissipation capacity of different soil types for vibration stress waves. The formula contains... These items constitute the geological compensation factor, enabling adaptive threshold adjustment for different soil environments;

[0081] Value acquisition method: During deployment, retrieve the geological exploration report for this road section and look up the value in the system's pre-set geological damping parameter library;

[0082] Engineering example: If the geological conditions are hard rock, vibration attenuation is small; referring to the table, we can obtain... If the geology is conventional clay, refer to the table to obtain... If the geology consists of soft, backfilled sand, vibrations attenuate quickly; refer to the table for details. In this embodiment, the geological conditions at the monitoring point are clay, therefore, a configuration is used. .

[0083] Parameter 4: Regional Environmental Tolerance Coefficient :

[0084] Physical unit: dimensionless;

[0085] Parameter meaning: Spatial risk weighting factor, used to balance the underreporting risk of high-risk operation areas with the false reporting risk of wilderness areas; The smaller the value, the higher the system sensitivity; conversely, the larger the value, the higher the tolerance.

[0086] Configuration method: The power grid operation and maintenance personnel manually configure the system in the background monitoring system according to the actual risk situation of the area.

[0087] Engineering Example: If this point is located at the intersection of a main municipal road, and there is upcoming subway construction, the risk is extremely high. [The following appears to be a separate, unrelated sentence:] Configuration... Actively lower the threshold to increase alertness; if the location is in a remote, desolate area with minimal human activity, configure... To improve tolerance and prevent false triggering and power consumption; in this embodiment, the monitoring point is located in a suburban construction area, therefore, it is configured with .

[0088] Based on the above parameters, the cloud calculates the dynamic predetermined threshold for this monitoring point: ;

[0089] The first cloud processing unit will process the supervisor's measurements reported from the front end. Compare with this dynamic threshold:

[0090] like If the condition is determined to be non-suspected dangerous, the system will maintain routine monitoring.

[0091] like If the cloud determines that the system has entered a suspected dangerous state, it immediately sends a "high priority wake-up" signal to the local processing unit through the 4G downlink control link.

[0092] IV. Visual review and closed-loop linkage early warning stage

[0093] like Figure 1 As shown, after receiving the wake-up signal from the cloud, the local processing unit immediately performs the following operations:

[0094] Step 1: Hardware power-on wake-up: Simultaneously pull up the hardware enable pins of the camera module and the 5G communication module to physically power them up and wake them up from the zero-power sleep state.

[0095] Step 2: Camera Angle Adjustment: Based on the previously calculated two-dimensional direction vector, control the gimbal rotation of the camera module to aim the lens at the direction of the seismic source, ensuring accurate capture of the target image.

[0096] Step 3: Fast streaming and uploading: The camera module uses the FastBoot mechanism to complete initialization and output high-definition images within 500ms. At the same time, the 5G module completes network registration and pushes the H.265 encoded high-definition video stream to the second cloud processing unit in real time.

[0097] After receiving the video stream, the second cloud processing unit calls upon a high-precision visual AI model deployed in the cloud to analyze the image in real time.

[0098] The model is based on the YOLOv8 target detection architecture and has been specifically trained for heavy construction machinery such as excavators, drilling rigs, and loaders. It can achieve accurate target identification and skeleton extraction. At the same time, it analyzes the target's motion trajectory to identify whether it has destructive behavior characteristics such as digging or drilling, which are the hazard factors in this invention.

[0099] Based on the recognition results, the system executes different branch logic:

[0100] Branch 1: False Alarm Elimination: If visual verification shows only passing heavy trucks or unrelated personnel in the image, and no hazardous factors are identified, it is determined to be an environmental false trigger. A hibernation command is issued from the cloud, and the local processing unit controls the camera module and 5G module to power off, and the system returns to normal low-power monitoring status.

[0101] Branch Two: Hazard Confirmation and Closed-Loop Execution: If the visual model confirms the presence of heavy construction machinery and destructive behavior, the system is determined to be in a hazardous state. At this point, the system executes a dual closed-loop intervention:

[0102] Intervention 1: On-site early warning: The cloud sends an activation signal through the 4G link, the local processing unit drives the relay to activate, and connects the power supply circuit of the audible and visual alarm light. The site immediately emits a high-decibel alarm sound and flashes a bright light to deter construction personnel in real time and achieve pre-event intervention.

[0103] Intervention 2: Remote notification: The third cloud processing unit generates an anomaly alarm work order and continuously sends maintenance warnings to the monitoring personnel via SMS, APP push, etc., until the monitoring personnel complete the manual confirmation, at which point the system will deactivate the alarm and switch back to a non-suspected danger state.

[0104] V. Dynamic Energy-Saving Gradient Scheduling Based on Remaining Power

[0105] like Figure 3As shown, to ensure the system's survivability in extreme environments such as continuous rainy days, the system has a built-in energy-saving gradient scheduling logic that dynamically adjusts the system's operating parameters based on the remaining power at the energy storage end.

[0106] In this embodiment, two levels of charge gradients are defined:

[0107] First tier (remaining battery > 30%): The battery is determined to be sufficient, the system enters normal power consumption mode, the sampling rate of the multi-directional monitoring unit is maintained at 100Hz, the data reporting interval is 1 second, and after triggering wake-up, the camera module can maintain video streaming for a maximum of 30 seconds;

[0108] Second tier (remaining power ≤ 30%): if the power is insufficient, the system enters the power saving mode, the sampling rate of the multi-directional monitoring unit is reduced to 50Hz, and the sampling power consumption is reduced while ensuring the sensing capability; the data reporting interval is extended to 2 seconds to reduce the communication frequency; after triggering wake-up, the maximum push duration of the camera module is shortened to 5 seconds; and the system immediately shuts down and goes into sleep mode after completing the capture verification.

[0109] When the remaining battery power drops further to below 15%, the system enters extreme power-saving mode: at this time, the system will significantly increase the vibration trigger threshold and only trigger the local alarm directly when extremely strong vibration is detected, skipping the video review process, in order to save power to the maximum extent and ensure that the most critical early warning function does not fail.

[0110] Through the aforementioned phased dynamic scheduling mechanism, this system can achieve long-term stable operation in field environments without mains power access, completely solving the problem of insufficient battery life of traditional monitoring equipment.

Claims

1. A cloud-edge collaborative wake-up cable channel external damage prevention monitoring system, characterized in that, include: A multi-directional monitoring unit is used to acquire monitoring data in the first direction, the second direction, and the third direction to collect real-time vibration signals; The first cloud processing unit is used to receive the first direction monitoring data, the second direction monitoring data and the third direction monitoring data, convert them into a chief measurement value and a two-dimensional direction vector through built-in data processing logic, and compare the chief measurement value with a predetermined threshold to determine whether a suspected dangerous state has been entered. The predetermined threshold is the maximum acceptable vibration value related to the geological environment, obtained through a built-in function model. The local processing unit, located at the monitoring site, is used to receive and transmit the suspected danger status signal, the chief supervisor's measurement value, and the two-dimensional direction vector emitted by the first cloud processing unit; The visual perception end is integrated into the local processing unit and is controlled by the local processing unit. It adjusts the screen angle according to the orientation of the two-dimensional direction vector in order to acquire and transmit the source image. The second cloud processing unit is used to receive the earthquake source image obtained by the visual perception terminal, identify the earthquake source image and determine whether there are dangerous factors, and decide whether to enter a dangerous state based on the judgment result. The local alarm terminal is used to receive the activation signal from the second cloud processing unit and activate the alarm at the monitored site. When the third cloud processing unit is notified by the second cloud processing unit that it has entered the dangerous state, it issues a maintenance warning to the monitoring personnel. If the measured value of the chief supervisor is greater than the predetermined threshold, the first cloud processing unit determines that the suspected dangerous state has been entered and sends the measured value of the chief supervisor and the two-dimensional direction vector to the local processing unit; If the second cloud processing unit identifies the presence of the hazard factor, it enters the hazard state and activates the local alarm terminal and the third cloud processing unit; The energy-saving gradient scheduling unit is used to dynamically adjust the system's operating mode based on the remaining power. After the monitoring personnel complete the manual confirmation, the system is restored to a non-suspected danger state, the maintenance warning is lifted, and the local alarm terminal stops alarming.

2. The cloud-edge collaborative wake-up cable channel external damage prevention monitoring system according to claim 1, characterized in that, The built-in data processing logic includes: Acquire multiple datasets containing monitoring data in the first direction, the second direction, and the third direction within the minimum unit of time; The arithmetic mean of the monitoring data in each direction is calculated to obtain the average point of the monitoring data in the first direction, the average point of the monitoring data in the second direction, and the average point of the monitoring data in the third direction. The average points of the monitoring data in the first direction, the second direction, and the third direction are squared and then added together to obtain the measurement value of the chief supervisor. The average point of the first direction monitoring, the average point of the second direction monitoring, and the average point of the third direction monitoring are taken as the three directional components of the three-dimensional direction vector, and the three-dimensional direction vector is projected onto the plane where the cable extension direction is located to obtain the two-dimensional direction vector.

3. The cloud-edge collaborative wake-up cable channel external damage prevention monitoring system according to claim 1, characterized in that, The predetermined threshold is generated based on the built-in function model, which is: ； in, For the predetermined threshold, It is a natural constant in mathematics. This is the system's baseline threshold. The damping attenuation coefficient of the soil medium. For cable burial depth parameters, The system reference threshold is the minimum effective value that can trigger an alarm under standard operating conditions. The soil medium damping attenuation coefficient is the absorption and dissipation capacity of different soil types for vibration stress waves. The cable burial depth parameter is the depth at which the cable is buried underground. The regional environmental tolerance coefficient is a parameter that is manually configured under different regional environments.

4. The cloud-edge collaborative wake-up cable channel external damage prevention monitoring system according to claim 1, characterized in that, The energy-saving gradient scheduling unit includes a natural energy receiving end and an energy storage end; The energy storage terminal is used to receive and store the energy from the natural energy receiving terminal, and to provide energy to the local processing unit, the visual sensing terminal, the local alarm terminal and the multi-directional monitoring unit, while also feeding back the remaining power to the local processing unit; If the remaining power is within a preset first gradient, the local processing unit controls the multi-directional monitoring unit to enter the normal power consumption mode; If the remaining power is within the preset second gradient, the local processing unit controls the multi-directional monitoring unit to enter the energy-saving power consumption mode; The monitoring frequency of the normal power consumption mode is higher than that of the energy-saving power consumption mode.

5. The cloud-edge collaborative wake-up cable channel external damage prevention monitoring system according to claim 1, characterized in that, The multi-directional monitoring unit, the local processing unit, the local alarm terminal, and any two of the first cloud processing unit, the second cloud processing unit, and the third cloud processing unit communicate via a 4G communication channel. The visual perception terminal communicates with the second cloud processing unit via a 5G communication channel. Under the non-potentially dangerous conditions, the visual sensing terminal and the 5G communication channel are in a power-off sleep state; When the suspected dangerous state is entered, the local processing unit physically powers on and wakes up the visual sensing terminal.

6. A method for monitoring cable channel damage prevention based on cloud-edge collaborative wake-up of the system according to any one of claims 1 to 5, characterized in that, Includes the following steps: S1: The multi-directional monitoring unit collects the monitoring data in the first direction, the monitoring data in the second direction, and the monitoring data in the third direction, and uploads them to the first cloud processing unit; S2: The first cloud processing unit converts the monitored quantity into a supervisor's measurement value and a two-dimensional direction vector through its built-in data processing logic. It compares the supervisor's measurement value with a predetermined threshold. If it determines that a suspected dangerous state has been entered, it sends the supervisor's measurement value and the two-dimensional direction vector to the local processing unit at the monitoring site. S3: After receiving the measured value from the chief supervisor and the two-dimensional direction vector, the local processing unit activates the normally dormant visual sensing terminal, adjusts the screen angle of the visual sensing terminal according to the two-dimensional direction vector, and controls the visual sensing terminal to collect the seismic source image and upload it to the second cloud processing unit. S4: The second cloud processing unit identifies the seismic source image. If it determines that there is a dangerous factor or that a dangerous state has been entered, it activates the local alarm terminal at the monitoring site and notifies the third cloud processing unit. S5: The third cloud processing unit issues a maintenance warning to the monitoring personnel. After receiving a manual confirmation instruction from the monitoring personnel, the system switches back to a non-suspected danger state, clears the alarm, and resets all components to normal sleep mode.

Images