Intelligent monitoring loop and monitoring method for multi-source data fusion of drainage pipelines in cold regions

Abstract

This application provides a multi-source data fusion intelligent monitoring ring and monitoring method for drainage pipelines in cold regions. The monitoring ring includes: a mechanical body, a multi-parameter sensing module, an end-side intelligent unit, a hybrid power supply and energy management unit, a wireless communication unit, and a quick-installation sealing assembly. The mechanical body is formed by two semi-annular aluminum alloy frames interlocking together. The mechanical body includes a radial expansion wedge and a flexible sealing sleeve. The radial expansion wedge has an adjustable eccentric cam embedded within it. The radial expansion wedge is covered with a EPDM rubber toothed surface. The flexible sealing sleeve has an integrally injection-molded silicone corrugated layer. The fiber optic strain gauge chain is spirally attached to the inner side of the frame. This application achieves continuous, accurate, low-power, and intelligent fusion monitoring of multi-source information on temperature, stress, and corrosion in drainage pipelines in cold regions, providing reliable technical support for structural safety assessment, corrosion prevention and control, and operation and maintenance decisions.

Description

Technical Field

[0001] This application relates to the field of drainage pipeline monitoring technology, and more specifically, to a multi-source data fusion intelligent monitoring loop and monitoring method for drainage pipelines in cold regions. Background Technology

[0002] Drainage pipelines in cold regions are constantly exposed to complex environments characterized by low temperatures, freeze-thaw cycles, continuous humidity, and chemical corrosion, making them highly susceptible to defects such as segment cracks, joint leaks, steel reinforcement corrosion, and internal icing leading to pipe diameter reduction. To address these issues, various online monitoring and detection technologies have been researched both domestically and internationally, primarily falling into the following categories:

[0003] 1. Point-type or cable-type sensor nodes A typical solution involves attaching / burying strain gauges, thermometers, and electrode-type corrosion sensors on the inner or outer wall of the tunnel segment, and then transmitting the signals to a ground-based data acquisition box via wired connections.

[0004] Limitations: Large amount of wiring required; installation trenching can easily damage the waterproof layer; cables become brittle at low temperatures; water ingress into interfaces can cause signal drift; maintenance and replacement are extremely difficult.

[0005] 2. External wireless sensor unit MEMS strain / tilt, temperature and humidity, or pH sensors are packaged as battery-powered wireless nodes, fixed to the inner surface of the tube by magnetic force or expansion bolts, and communicate with the ground aggregation gateway via LoRa or NB-IoT.

[0006] Limitations: A single node can only monitor a single or a small number of physical quantities; the battery capacity drops by more than 40% below -20 ℃, making it difficult to meet the requirement of maintenance-free operation for many years; the node protrusions are easily knocked off by floating ice and debris, bringing the risk of secondary damage.

[0007] 3. Distributed fiber optic monitoring Fiber Bragg gratings (FBGs) or Brillouin scattering fibers are laid longitudinally along the tube segments to obtain strain or temperature field distribution.

[0008] Limitations: It can only measure "temperature" or "response", and cannot reflect electrochemical corrosion; the optical fiber needs to pass through multiple inspection wells, making construction complex; it is sensitive to bending and pulling, and its reliability is insufficient in old pipeline networks.

[0009] 4. Water quality and flow velocity monitoring probe Traditional water quality electrodes or ultrasonic Doppler current meters are often installed at pump stations or water outlets, making it difficult to obtain local corrosion environment data in icing sections and low-lying waterlogged sections.

[0010] 5. Data Fusion and Intelligent Analysis Most existing monitoring systems are based on a core architecture of "single parameter – single node – cloud analysis".

[0011] Limitations: Insufficient data fusion and intelligent analysis; lack of ability to synchronously fuse and provide real-time early warning of multimodal data on temperature, strain, and corrosion at the "edge"; high communication traffic and cloud computing overhead, resulting in limitations on real-time performance and cost.

[0012] 6. Energy supply and reliability issues The wireless sensors inside the pipeline are mainly powered by batteries, which have poor low-temperature discharge performance and are difficult to replace; the single working mode of micro energy harvesting (such as water turbines, piezoelectric devices, etc.) is difficult to provide continuous power in seasonal main stream / ice-bound environments.

[0013] Patent application CN105807020A, entitled "A Comprehensive Pipeline Monitoring System and Monitoring Method Thereof," discloses an online monitoring device for drainage pipelines in the form of a "monitoring sphere," such as... Figure 1 As shown, the device includes an external overall structure, an integrated sensing unit, a power and communication unit, and a data processing and positioning unit. The external overall structure adopts a sealed spherical shell (outer diameter approximately Φ120 mm), with the shell material being corrosion-resistant ABS and a stainless steel inner skeleton; the two halves of the shell are sealed by O-rings and connected by a threaded ring lock to ensure sealing performance under an underwater internal pressure of 0.4 MPa. The integrated sensing unit includes a water quality detection module, a leakage / seepage detection module, a temperature sensor, and a pressure / level sensor. The water quality detection module includes multi-parameter microprobes for pH, conductivity, turbidity, and electrode-type residual chlorine, arranged within the circumferential openings of the spherical shell. The leakage / seepage detection module uses a triaxial gyroscope and accelerometer to record the micro-deflection curves of the sphere's rolling trajectory within the pipe to determine flow field anomalies caused by localized leakage. The temperature sensor employs a single-point thermistor, primarily used to calibrate the water quality probes. The pressure / level sensor is used to determine whether the pipe is full or overloaded. The energy and communication unit is powered by a built-in lithium-ion battery (3.7 V / 6 Ah). It communicates wirelessly with a ground gateway using a LoRa module (470 MHz) with a data packet interval of 5 seconds. A magnetic induction switch and a wireless charging coil are located on the top of the shell for easy maintenance and charging. The data processing and positioning unit collects data from various sensors through an MCU (STM32 series) and stores it in a local 8 GB eMMC. Combining the pipeline flow velocity and three-dimensional attitude change curves, it uses a mileage-time estimation method to locate the relative position of the monitoring ball in the pipe. The ground terminal software performs water quality threshold judgment and leakage alarm on the received data.

[0014] The detection device is used by maintenance personnel who place a monitoring ball into the pipeline at the inspection well. The ball rolls forward with the water flow, collecting and uploading data in real time. The ball is then retrieved through a downstream inspection well or interception net to complete data retransmission and battery replacement. This device is primarily applicable for water quality monitoring and initial leakage detection in municipal drainage networks at ambient temperatures; it is particularly effective for pipe sections with an inner diameter ≥ 300 mm and a flow velocity of 0.1–1.0 m / s. However, this device has the following technical shortcomings:

[0015] 1. Inability to perform structural health monitoring (lack of strain monitoring capabilities) The existing solution only has the function of collecting environmental parameters such as water quality, conductivity and flow velocity. It lacks strain or stress sensors, cannot evaluate the structural response of the tunnel segments under loads such as frost heave, displacement and earthquake, and does not have the ability to provide structural early warning.

[0016] 2. Lack of corrosion electrochemical monitoring function The device does not integrate an electrode-type corrosion sensor or a rebar potential probe, and therefore cannot identify the rebar corrosion and concrete spalling process caused by freeze-thaw or acid-alkali erosion. Its monitoring capability is limited to water quality parameters, resulting in serious information gaps.

[0017] 3. Single-point patrol monitoring cannot achieve long-term, fixed-point continuous monitoring. The monitoring ball moves forward by water flow and cannot be fixed in key risk locations (such as deformed sections or areas of concentrated corrosion) for long-term monitoring; The monitoring duration is limited (power lasts for 2-3 hours), and manual retrieval and re-deployment are required, making it unsuitable for establishing a routine monitoring system.

[0018] 4. Not suitable for low-temperature operating conditions in cold regions, resulting in poor reliability. The internal battery and electronic modules experience a significant voltage drop below -10°C, making them prone to failure or false alarms in extremely cold regions. The device is not encapsulated with an antifreeze structure, and the spherical shell is easily blocked or cracked by icing substances, making it unsuitable for long-term operation in environments with alternating freezing and thawing.

[0019] 5. Lack of multi-parameter fusion and edge-side intelligent analysis capabilities This technology only enables data collection and transmission. All data needs to be uploaded to the cloud for post-processing and does not have edge inference capabilities. The data is highly redundant, communication bandwidth and server costs are high, and real-time intelligent early warning and local response are not possible.

[0020] 6. Insufficient battery life and frequent manual maintenance required. The built-in lithium battery has a limited capacity, and its discharge efficiency decreases by 40-60% under low temperature conditions, making it impossible to achieve maintenance-free operation for several months in cold regions. The lack of energy harvesting capability means that the device needs to be manually removed, recharged, and redeployed after each use, which severely restricts the efficiency of engineering deployment.

[0021] In summary, while existing "monitoring sphere" technologies possess some water quality monitoring capabilities, they have significant shortcomings in structural health assessment, corrosion monitoring, cold-region adaptability, intelligent integration, and deployment efficiency, making it difficult to meet the actual needs of multi-source integrated intelligent monitoring of "temperature-stress-corrosion" in cold-region drainage pipe networks. Cold-region drainage pipelines urgently require an integrated monitoring device that is structurally friendly, quick to install, cold and corrosion resistant, self-powered, integrates multiple parameters, and provides end-side intelligent analysis to achieve dynamic perception and timely early warning of the entire "temperature-stress-corrosion" lifecycle of the pipe network. Summary of the Invention

[0022] In view of this, this application provides a multi-source data fusion intelligent monitoring loop and monitoring method for drainage pipelines in cold regions, in order to overcome the limitations of existing monitoring devices for drainage pipelines in cold regions in terms of detection parameters, cold resistance, energy endurance, intelligent fusion and installation and maintenance.

[0023] To achieve the above objectives, the technical solution adopted in this application is as follows: The intelligent monitoring ring for multi-source data fusion of drainage pipelines in cold regions includes: a mechanical body, a multi-parameter sensing module, an end-side intelligent unit, a hybrid power supply and energy management unit, a wireless communication unit, and a quick-installation-sealing assembly; the multi-parameter sensing module, the end-side intelligent unit, the hybrid power supply and energy management unit, the wireless communication unit, and the quick-installation-sealing assembly are all fixed on the mechanical body.

[0024] Furthermore, the mechanical body is formed by two semi-annular aluminum alloy frames locking together; and the mechanical body includes a radial expansion wedge and a flexible sealing sleeve; the radial expansion wedge is embedded with an adjustable eccentric cam; the radial expansion wedge is covered with a EPDM rubber tooth surface; and the flexible sealing sleeve is integrally injection molded with a silicone corrugated layer. Furthermore, the multi-parameter sensing module includes a temperature sensor, a fiber optic strain gauge, an electrochemical corrosion probe, a humidity-conductivity composite probe, and an ultrasonic microfluidometer; the fiber optic strain gauge is spirally attached to the inner side of the skeleton.

[0025] Furthermore, the edge-side intelligent unit includes an ultra-low-power MCU, a 1 TOPS NPU chip, an intelligent module, an edge processor, a wireless communication module, and a memory module. The ultra-low-power MCU and the 1 TOPS NPU chip are responsible for sensor synchronous sampling and data standardization and fusion inference. The 1 TOPS NPU chip pre-embeds a lightweight LSTM-Attention model, which periodically outputs segment damage levels and over-limit threshold alarms. The memory module consists of eMMC and QSPI Flash, used for cyclically buffering 30 days of raw data and model parameters.

[0026] Furthermore, the hybrid power supply and energy management unit includes a micro turbine generator, a LiFePO4 cryogenic battery, and a supercapacitor array. The micro turbine generator provides self-generated energy when there is water flow; its output, after rectification and voltage limiting, prioritizes charging the supercapacitor array and supplying power to the monitoring ring load, while simultaneously maintaining charge of the LiFePO4 cryogenic battery when energy is abundant. The supercapacitor array handles the pulse power requirements for wireless transmission and local buzzer / flash alarms. The LiFePO4 cryogenic battery provides steady-state power supply under conditions of no water flow or low temperature, thereby enabling long-term continuous operation of the monitoring ring.

[0027] Furthermore, the temperature sensor, fiber optic strain gauge, electrochemical corrosion probe, humidity-conductivity composite probe, and ultrasonic microflow meter in the multi-parameter sensing module are all connected to the end-side intelligent unit via 4-in-1 FPC cables, and the connection ports are encapsulated with low-modulus epoxy resin.

[0028] Furthermore, the mechanical body, multi-parameter sensing module, end-side intelligent unit, hybrid power supply and energy management unit, wireless communication unit, and quick-installation sealing assembly are integrated into an integrated cold-resistant silicone outer sheath via a highly reliable quick-connect interface.

[0029] Furthermore, the wireless communication module adopts a LoRa+NB-IoT dual-mode module and an adaptive circuit selection algorithm; and has a built-in UWB-LoRa subnet collaboration protocol, enabling network relay between nodes within the pipeline segment.

[0030] Furthermore, the quick-connect-seal assembly includes a snap-fit ​​hinge and a stainless steel snap ring, as well as an electro-optical-hydraulic tee quick-connect socket, which facilitates external fiber optic extension or calibration connection without disassembling the entire ring.

[0031] A multi-source data fusion intelligent monitoring method for drainage pipelines in cold regions, the method being based on an end-side intelligent unit in any of the aforementioned monitoring loops, the method comprising: S1, Timescale Correction: Timing control of multiple ADCs / DACs and unified timescale correction of temperature-strain-corrosion potential; S2. Feature Extraction: The feature extraction module in the end-side intelligent unit performs preprocessing and feature construction on temperature, strain and corrosion potential signals. S3, Fusion Reasoning: The fusion layer in the edge-side intelligent unit adaptively fuses the temperature, strain, and corrosion potential features, and outputs a fixed-dimensional fusion feature vector; S4. Output Decision: A lightweight fully connected network + Sigmoid is used to output the pipeline damage level and the remaining pipeline life respectively; S5. When the pipeline damage level exceeds the threshold, a local buzzer and flashing alarm will be issued immediately. S6. Using LZF compression and predictive residual coding, the signal is transmitted to the remote control center via a wireless communication unit. Transmit key features, event fragments, and alarm messages to reduce communication volume and power consumption.

[0032] Compared with the prior art, the beneficial effects of this application are: This application's monitoring ring achieves long-term, accurate, and real-time monitoring of the "temperature-response-corrosion" status of drainage pipelines in cold regions through circumferential expansion non-destructive installation, multi-parameter collaborative measurement, end-side intelligent fusion analysis, and hybrid self-powered design. It comprehensively overcomes the problems of single monitoring parameters, poor cold resistance, frequent maintenance, and insufficient intelligence in existing technologies, and has significant engineering applicability and promotion value. Attached Figure Description

[0033] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0034] Figure 1 This is a schematic diagram of the overall three-dimensional structure of the monitoring ring in this application.

[0035] Figure 2 This is an exploded view of the monitoring ring assembly in this application.

[0036] Figure 3 This is a schematic diagram of the quick-release sealing assembly structure of the monitoring ring in this application.

[0037] Figure 4 This is a block diagram illustrating the working principle of the edge-side intelligent unit in this application, which involves data acquisition, fusion, and inference.

[0038] Figure 5This is a block diagram illustrating the working principle of the hybrid power supply and energy management unit of this application.

[0039] Figure 6 This is the flowchart for the installation, initialization, and online operation of the monitoring ring in this application.

[0040] Reference numerals: Mechanical body-1, Multi-parameter sensing module-2, End-side intelligent unit-3, Hybrid power supply and energy management unit-4, Wireless communication unit-5, Quick-release sealing assembly-6; Semi-annular aluminum alloy frame-11, Radial expansion wedge block-12, Flexible sealing sleeve-13; Temperature sensor-21, Fiber optic strain gauge chain-22, Electrochemical corrosion probe-23, Humidity-conductivity composite probe-24, Ultrasonic microflow meter-25; Ultra-low power MCU-31, 1 TOPS NPU chip-32, Edge processor-33, Wireless communication module-34, Memory module-35; Micro turbine generator-41, LiFePO4 cryogenic battery-42, Supercapacitor array-43; Locking hinge-61, Stainless steel snap ring-62, Electro-optical-hydraulic tee quick-connect socket-63. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of this application, but not all embodiments.

[0042] like Figures 1 to 3 As shown, the multi-source data fusion intelligent monitoring ring for drainage pipelines in cold regions includes: a mechanical body 1, a multi-parameter sensing module 2, an end-side intelligent unit 3, a hybrid power supply and energy management unit 4, a wireless communication unit 5, and a quick-installation-sealing assembly 6. The multi-parameter sensing module 2, the end-side intelligent unit 3, the hybrid power supply and energy management unit 4, the wireless communication unit 5, and the quick-installation-sealing assembly 6 are all fixed on the mechanical body 1.

[0043] Furthermore, the mechanical body 1 is formed by two semi-annular aluminum alloy frames 11 locked together; and the mechanical body 1 includes a radial expansion wedge 12 and a flexible sealing sleeve 13; the radial expansion wedge 12 is embedded with an adjustable eccentric cam; the radial expansion wedge is covered with a EPDM rubber toothed surface; the flexible sealing sleeve 13 (integrated injection-molded silicone corrugated layer) is used to seal the circumferential contact interface between the semi-annular aluminum alloy frame 11 and the inner wall of the drainage pipe, that is, the "circumferential bonding strip" formed in the pipe after the monitoring ring is installed.

[0044] In specific implementation, the semi-annular aluminum alloy frame 11 can be anodized with 6061-T6; the two halves are quickly locked together to form a complete ring. The adjustable eccentric cam can generate a 1–3 kN tightening force on the inner wall of the Φ300 mm-Φ800 mm pipe by rotating 90°; the radial tightening wedge is covered with EPDM rubber teeth to improve the low-temperature friction coefficient. The flexible sealing sleeve 13 has an integrally injection-molded silicone corrugated layer to prevent water seepage, ice buildup, and impact from debris.

[0045] The mechanical body 1 of this application adopts an integrated structure of circumferential expansion and flexible sealing sleeve, which can be quickly clamped to the inner wall of Φ300–Φ800 mm tube segments without drilling or slotting. It is also low in profile and resistant to ice impact and debris impact, solving the technical problems of traditional sensor node installation damaging the structure and poor protection performance.

[0046] Furthermore, the multi-parameter sensing module 2 includes a temperature sensor 21, a fiber optic strain gauge 22, an electrochemical corrosion probe 23, a humidity-conductivity composite probe 24, and an ultrasonic microfluidometer 25; the fiber optic strain gauge 22 is spirally attached to the inner side of the semi-annular aluminum alloy skeleton 11.

[0047] Specifically, the temperature sensor 21 can be a thermistor with an accuracy of ±0.2 ℃; the fiber optic strain gauge 22 has a range of ±2500 με and a spatial resolution of 10 mm. The electrochemical corrosion probe 23 is a three-electrode type (working-reference-auxiliary) used to measure the potential of the reinforcing steel and the corrosion rate, with an accuracy of ±5 mV. The humidity-conductivity composite probe 24 is used for monitoring water film thickness and salt concentration. The ultrasonic microflow meter 25 uses a dual-reflector time-difference measurement method with a range of 0-1 m / s. -1 It can withstand -40℃. The temperature sensor 21, fiber optic strain gauge 22, electrochemical corrosion probe 23, humidity-conductivity composite probe 24 and ultrasonic microflow meter 25 in the multi-parameter sensing module 2 are all connected to the end-side intelligent unit 3 by 4-in-1 FPC cable, and the connection port is potted with low modulus epoxy resin.

[0048] This application solves the technical problems of missing multi-parameter synchronous sensing and low reliability of sensing nodes in extreme cold and freeze-thaw environments. It enables the monitoring ring to simultaneously acquire key parameters such as temperature field, structural strain field, and electrochemical corrosion field within a single device, achieving comprehensive state characterization under the same spatial and temporal reference. Moreover, it ensures that the sensing, packaging, and connectors can operate stably for a long time under cycling conditions from -40 ℃ to +20 ℃ and high humidity and water immersion conditions without embrittlement, water ingress, or signal drift.

[0049] Furthermore, the edge-side intelligent unit 3 includes an ultra-low power MCU 31, a 1 TOPS NPU chip 32, an edge processor 33, a wireless communication module 34, and a memory module 35; the ultra-low power MCU 31 and the 1 TOPS NPU chip 32 are responsible for sensor synchronous sampling and data standardization and fusion inference; a lightweight LSTM-Attention model is pre-embedded to periodically output segment damage levels and over-limit threshold alarms; the memory module 35 consists of eMMC and QSPI Flash, used for cyclically buffering 30 days of raw data and model parameters.

[0050] Specifically, during monitoring loop operation, the ultra-low power MCU 31 (dual-core 48 MHz) and 1 TOPS NPU chip are responsible for sensor synchronization sampling, data standardization, and fusion inference. A lightweight LSTM-Attention model is pre-embedded, outputting the segment damage level (0-5) and over-limit threshold alarms at 1-minute intervals. The memory module 35 is configured with 8 GB eMMC + 32 MB QSPI Flash for cyclically buffering 30 days of raw data and model parameters.

[0051] As can be seen, this application realizes the standardization, fusion and machine learning inference of multimodal data locally in the monitoring ring, outputs damage level and threshold over-limit alarm, reduces the amount of uplink data and improves real-time performance, and solves the technical problem of lack of real-time fusion and intelligent early warning capabilities on the edge side.

[0052] Furthermore, the hybrid power supply and energy management unit 4 includes a micro turbine generator 41, a LiFePO4 cryogenic battery 42, and a supercapacitor array 43. The micro turbine generator 41 provides self-generated energy when there is water flow. Its output, after rectification and voltage limiting, is used to charge the supercapacitor array and supply power to the monitoring ring load. At the same time, when there is surplus energy, it sustains the charging of the LiFePO4 cryogenic battery 42. The supercapacitor array 43 is used to carry the pulse power requirements of wireless transmission and local buzzer and flash alarm. The LiFePO4 cryogenic battery 42 provides steady-state power supply when there is no water flow or low temperature, thereby realizing the long-term continuous operation of the monitoring ring.

[0053] Specifically, the flow velocity of the micro turbine generator 41 is ≥0.2 ms. -1 Output 0.3-0.8 W; The LiFePO4 low-temperature battery 42 (6.4 V / 12 Ah) maintains a discharge rate of ≥70% at -40 ℃, providing peak pulse power; the bidirectional DC-DC + MPPT controller 4-4 intelligently distributes generated energy and maintains battery float charging. The average daily power consumption of the entire monitoring loop is <0.25Wh, achieving ≥3 years of maintenance-free operation.

[0054] See Figure 5 Energy management and low-power strategies: When the flow velocity of the micro turbine generator 41 is ≥0.2 m / s -1 Furthermore, if the temperature is ≥ -5 ℃, the system will preferentially use hydroelectric power generation; the LiFePO4 low-temperature battery voltage is < 5.8 V or the flow rate is < 0.1 ms. -1 When the MCU enters deep sleep mode (wakes up once every 60 seconds), it retains only 10 Hz sampling; the supercapacitor array 43 provides an instantaneous 150 mA pulse for LoRa transmission / buzzer alarm; the average standby current of the whole system is 80 µA, and the peak operating current is 180 mA, meeting the 3-year maintenance-free target.

[0055] It is evident that this application can solve the technical problem of insufficient long-term maintenance-free energy supply. Under the condition of low temperature causing battery capacity decay, the self-powered water turbine-lithium iron battery hybrid power supply and ultra-low power consumption design can meet the energy demand of the monitoring ring for continuous operation for ≥3 years.

[0056] Furthermore, the wireless communication unit 5 adopts a LoRa+NB-IoT dual-mode module and an adaptive circuit selection algorithm; and has a built-in UWB-LoRa subnet collaboration protocol, enabling network relay between nodes within the pipeline segment.

[0057] Specifically, the wireless communication unit 5 adopts a LoRa (470 MHz) + NB-IoT (B8 / B20) dual-mode module; it has a built-in adaptive link selection algorithm that automatically switches to NB-IoT when the RSSI is below -115 dBm; and it has a built-in UWB-LoRa subnet cooperation protocol, which enables network relay between nodes within the pipe section, improving communication reliability in tunneling and icing attenuation environments.

[0058] Furthermore, the quick-connect-seal assembly 6 includes a snap-lock hinge 61, a stainless steel snap ring 62, and an electro-optical-hydraulic tee quick-connect socket 63, which facilitates external fiber optic extension or calibration connection without disassembling the entire ring.

[0059] In practice, the quick-release sealing assembly 7 can be closed by a single person within 30 seconds. The electro-optical-hydraulic tee quick-connect socket 73 can be an IP68 rated electro-optical-hydraulic tee quick-connect socket.

[0060] Furthermore, the mechanical body 1, multi-parameter sensing module 2, end-side intelligent unit 3, hybrid power supply and energy management unit 4, wireless communication unit 5, and quick-installation sealing assembly 6 are integrated into an integrated cold-resistant silicone outer sheath via a highly reliable quick-connect interface.

[0061] See Figure 4A multi-source data fusion intelligent monitoring method for drainage pipelines in cold regions, the method being based on an end-side intelligent unit in any of the above-mentioned monitoring loops, the method comprising: S1, Timescale Correction: Timing control of multiple ADCs / DACs and unified timescale correction of temperature-strain-corrosion potential; S2. Feature Extraction: The feature extraction module of the end-side intelligent unit 3 performs preprocessing and feature construction on temperature, strain and corrosion potential signals. Specifically, temperature characteristics can be statistically analyzed using a sliding window (mean / variance / slope), and further classified into temperature states using fuzzy hierarchical classification based on trapezoidal membership functions. Strain characteristics can be extracted by wavelet denoising, followed by the extraction of peak values, valley values, peak-to-peak values, upper / lower envelopes and their slopes. Corrosion potential characteristics can be extracted by multi-electrode differential and common-mode suppression filtering, and the potential mean, drift rate, and short-term fluctuation amplitude can be extracted.

[0062] S3. Fusion Inference: The fusion layer on the edge intelligent unit 3 adaptively fuses the temperature, strain, and corrosion potential characteristics, and outputs a fixed-dimensional fusion feature vector. Specifically, the fusion layer is a lightweight multi-source fusion neural network model LSTM + attention network (which can be accelerated by NPU-32) LSTM + attention weight structure.

[0063] S4. Output Decision: A lightweight fully connected network + Sigmoid structure is used to output the pipeline damage level and the pipeline remaining life respectively.

[0064] S5. When the pipeline damage level exceeds the threshold, a local buzzer and flashing alarm will be issued immediately.

[0065] Specifically, when the pipeline damage level is ≥3, a local buzzer will be triggered immediately.

[0066] S6. Using LZF compression and predictive residual coding, the signal is transmitted to the remote control center via wireless communication unit 5. Transmit key features, event fragments, and alarm messages to reduce communication volume and power consumption.

[0067] Installation, initialization, and operation process (see Figure 6): Installation: ① Open the inspection well → ② Unfold the semi-circular aluminum alloy frame and fit it into the inner wall of the segment → ③ Rotate the wedge to tighten → ④ Close the quick-release and check the sealing sleeve fit.

[0068] Initialization: ① Connect handheld terminal via Bluetooth → Set monitoring node ID and reference time → ② Scan sensor serial number → ③ Start self-test (including 4-hour low temperature simulation).

[0069] Running online: A complete data frame is collected every 60 seconds and fusion inference is performed. When the damage level is ≥3 or the temperature drop is >5 ℃ / h, rapid sampling (10 Hz) will be triggered and an alarm will be sent. Logs are written to the eMMC in a loop, and the complete waveform of the most recent 72 hours of data is preserved. maintain: When the battery expires after 3 years or when there are frequent abnormal alarms, loosen the wedge, remove the ring for inspection or replace the battery.

[0070] Key implementation details Cold-resistant packaging: All chips are packaged in ceramic or QFN packaging with conformal encapsulation, and are selected as devices operating at -55 ℃.

[0071] Corrosion-resistant design: PCB gold-plated + full Parylene C coating, electrode probes are made of 904L stainless steel.

[0072] Sensor calibration: Temperature / strain / potential three parameters are calibrated at the factory at -30 ℃ with two points and an error ≤2%.

[0073] Firmware upgrade: LoRa point-to-point repeater supports differential upgrade packages, avoiding loop disassembly.

[0074] Safety protection: Automatic shutdown under low voltage, overvoltage TVS + surge protection design; wireless communication encryption adopts... AES-128.

[0075] In summary, this application enables continuous, accurate, low-power, and intelligent integrated monitoring of multi-source information on temperature, stress, and corrosion in cold-region drainage networks, providing reliable technical support for structural safety assessment, corrosion prevention and control, and operation and maintenance decisions.

[0076] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A multi-source data fusion intelligent monitoring ring for cold region drainage pipeline, characterized in that, The utility model relates to a kind of integrated monitoring ring for long-distance pipeline, including: Mechanical body, multi-parameter sensing module, end-side intelligent unit, hybrid power supply and energy management unit, wireless communication unit and fast installation-sealing assembly; The multi-parameter sensing module, end-side intelligent unit, hybrid power supply and energy management unit, wireless communication unit and fast installation-sealing assembly are all fixed on the mechanical body.

2. The cold region sewer multi-source data fusion intelligent monitoring ring of claim 1, wherein, The mechanical body is formed by two half-ring aluminum alloy skeletons locked together;And the mechanical body includes radial expansion wedge and flexible sealing sleeve;The adjustable eccentric cam is embedded in the radial expansion wedge;The radial expansion wedge is covered with EPDM tooth surface outside;The flexible sealing sleeve is integrally injection molded with silicon rubber corrugated layer.

3. The cold region sewer multi-source data fusion intelligent monitoring ring of claim 2, wherein, The multi-parameter sensing module includes temperature sensor, fiber grating strain chain, electrochemical corrosion probe, humidity-conductivity composite probe and ultrasonic microflowmeter;The fiber grating strain chain is spirally attached to the inside of the skeleton.

4. The cold region sewer multi-source data fusion intelligent monitoring ring of claim 3, wherein, The end-side intelligent unit includes ultra-low power consumption MCU, 1 TOPS NPU chip, edge processor, wireless communication module and memory module;The ultra-low power consumption MCU and 1 TOPS NPU chip are responsible for sensing synchronous sampling and data standardization and fusion inference;The 1 TOPS NPU chip pre-embeds a lightweight LSTM-Attention model, which periodically outputs pipe segment damage level and out-of-limit threshold alarm;The memory module is eMMC and QSPI Flash, which is used for cyclic buffering of raw data and model parameters.

5. The cold region sewer multi-source data fusion intelligent monitoring ring of claim 4, wherein, The hybrid power supply and energy management unit includes micro-turbine water turbine generator, LiFePO4 low-temperature battery and super capacitor array;The micro-turbine water turbine generator is used to provide self-power generation energy under water flow condition, and its output is rectified and limited in voltage to charge the super capacitor array preferentially and supply power to the monitoring ring load, while maintaining the charge of the LiFePO4 low-temperature battery when there is excess energy;The super capacitor array is used to bear the pulse power demand of wireless transmission and local buzzer flashing alarm;The LiFePO4 low-temperature battery is used to provide steady-state power supply under no water flow or low temperature condition, thereby realizing long-term continuous operation of the monitoring ring.

6. The cold region sewer multi-source data fusion intelligent monitoring ring of claim 5, wherein, The temperature sensor, fiber grating strain chain, electrochemical corrosion probe, humidity-conductivity composite probe and ultrasonic microflowmeter in the multi-parameter sensing module all use 4-in-1 FPC flat cable to the end-side intelligent unit, and the connection port is filled with low-modulus epoxy resin.

7. The cold region sewer multi-source data fusion intelligent monitoring ring of claim 6, wherein, The mechanical body, multi-parameter sensing module, end-side intelligent unit, hybrid power supply and energy management unit, wireless communication unit and fast installation-sealing assembly are integrated in the integrated cold-resistant silica gel outer sheath through high-reliability fast connector.

8. The cold region sewer multi-source data fusion intelligent monitoring ring of claim 1, wherein, The wireless communication module uses LoRa+ NB-IoT dual-mode module and adaptive circuit selection algorithm;And it has built-in UWB-LoRa subnet coordination protocol, and nodes in the pipe segment can be networked and relayed.

9. The cold region sewer multi-source data fusion intelligent monitoring ring of claim 8, wherein, The fast installation-sealing assembly includes lock type hinge and stainless steel clasp spring, as well as electric-optical-liquid tee connector, which facilitates external fiber extension or calibration connection without disassembling the whole ring.

10. The intelligent monitoring method for multi-source data fusion of cold region drainage pipeline is characterized in that, The method is based on the end-side intelligent unit in any one of the monitoring rings, and the method comprises: S1, time scale correction: multi-channel ADC / DAC timing control and unified time scale correction of temperature-strain-corrosion potential; S2, feature extraction: pre-processing and feature construction of temperature, strain and corrosion potential signals in the feature extraction module in the end-side intelligent unit; S3, fusion reasoning: adaptive fusion of temperature, strain and corrosion potential features by the fusion layer in the end-side intelligent unit, to output a fixed-dimension fusion feature vector; S4, output decision: output of pipeline damage grade and pipeline remaining life by using a light-weight fully connected network + Sigmoid respectively; S5, when the pipeline damage grade exceeds a threshold, a buzzer flash warning is immediately given locally; S6, LZF compression and prediction residual coding are used, and key features, event fragments and warning messages are transmitted to a remote control center through a wireless communication unit, so as to reduce communication volume and power consumption. ​

Images