A thermal characteristic monitoring system and method for circular pipe equipment

By combining an attached sensing array component and an infrared thermal imaging acquisition module, the system solves the problem of all-weather, high-precision, and blind-spot-free thermal condition monitoring of circular pipe equipment, realizes continuous online monitoring of curved sections, and improves the equipment fault identification capability.

CN123111483APending Publication Date: 2026-11-17MAINTENANCE BRANCH COMPANY STATE GRID ZHEJIANG ELECTRIC POWER
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611412004.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-11
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing thermal condition monitoring technologies for circular pipe equipment are difficult to achieve all-weather, high-precision, and blind-spot-free monitoring, especially in curved sections where there is a risk of thermal failure. Traditional methods are easily affected by environmental interference and are complex to install, making it difficult to achieve continuous online monitoring.

Method used

The monitoring system, which combines an attached sensing array component with an infrared thermal imaging acquisition module, determines the number and spacing of sampling units through a curvature adaptive quantization array algorithm to achieve continuous monitoring in the entire circumference. The system also performs spatiotemporal registration and weighted fusion through a data processing module to output high-precision thermal feature monitoring results.

Benefits of technology

It enables all-weather, high-precision, and blind-spot-free thermal monitoring of circular pipe equipment, overcomes the problems of environmental interference and complex installation, improves the sensitivity and specificity of monitoring, and can identify potential equipment hazards at an early stage.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123111483A_ABST
    Figure CN123111483A_ABST
Patent Text Reader

Abstract

This invention relates to the field of power system equipment condition monitoring and fault diagnosis technology, and discloses a thermal characteristic monitoring system and method for circular pipe equipment. The system includes an attached sensing array assembly, an infrared thermal imaging acquisition module, and a data processing module. The attached sensing array assembly comprises a series array topology formed by multiple sampling units connected in series and mounted on a fixed hose clamp. The number and spacing of the sampling units in the series array topology are determined by a curvature adaptive quantization array algorithm. Each sampling unit integrates several sensors to collect thermal characteristic monitoring data of the circular pipe equipment. The infrared thermal imaging acquisition module includes an infrared thermal imager to collect infrared thermal images of the circular pipe equipment. The data processing module performs spatiotemporal registration and weighted fusion of the two types of data to obtain the thermal characteristic monitoring results of the circular pipe equipment, achieving all-weather, blind-spot-free, high-precision thermal characteristic monitoring and early fault warning for curved circular pipe equipment.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of power system equipment condition monitoring and fault diagnosis technology, and in particular to a thermal characteristic monitoring system and method for circular tube equipment, which is especially suitable for all-weather online thermal condition monitoring of curved circular tube power equipment such as gas-insulated metal-enclosed transmission lines (GIL). Background Technology

[0002] Circular tube structures are widely used in core power system equipment due to their excellent sealing performance, high structural strength, and strong adaptability. Gas-insulated metal-enclosed transmission lines (GILs) are one of the most representative circular tube power equipment. They seal high-voltage conductors within a grounded metal casing and transmit electrical energy by filling it with compressed insulating gas. Because GIL laying routes often need to traverse curved tunnels under rivers, urban underground utility tunnels, undulating mountainous terrain, and complex spaces within power plants, numerous bends are unavoidable. The risk of thermal failures in bends of GILs is significantly higher than in straight sections. This is mainly due to the combined effects of abnormally increased conductor contact resistance caused by bending deformation, uneven current distribution, and deteriorated heat dissipation, which seriously threaten the safe and stable operation of the power system.

[0003] Currently, thermal status monitoring methods for such equipment are mainly divided into two categories: infrared thermography and fiber Bragg grating (FBG) methods. Both methods have significant technical drawbacks. Infrared thermography is a non-contact monitoring method that uses an infrared thermal imager to receive infrared radiation emitted from the equipment surface and forms a thermal image through photoelectric conversion and signal processing. However, this method is easily affected by external factors such as outdoor sunlight, rain, dust, and changes in ambient temperature, leading to distortion of the equipment's thermal radiation characteristics. Furthermore, this method has stringent requirements for monitoring ambient temperature and monitoring duration, often requiring inspection personnel to handhold the equipment and take photos on-site, relying on manual experience to interpret the monitoring results. It is difficult to achieve continuous online monitoring and is only suitable for large-scale troubleshooting and rapid fault location. FBG, on the other hand, is a contact monitoring method that attaches or embeds a fiber Bragg grating sensor into the surface of the structure being measured. It obtains temperature information by measuring the drift of the reflected light wavelength. However, this method is highly sensitive to external environmental factors such as temperature, humidity, and vibration, and has stringent requirements for sensor accuracy and system operational stability. In addition, optical fibers are brittle and have weak resistance to breakage, making them difficult to lay in the circumference of curved pipes. The installation process is also complicated, and the laying method can easily create blind spots in monitoring, making it difficult to achieve full circumferential thermal state monitoring of curved pipes.

[0004] Therefore, how to solve the problem that existing thermal state monitoring technologies for circular pipe equipment cannot meet the requirements for all-weather, high-precision, and blind-spot-free thermal monitoring has become a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] This invention provides a thermal characteristic monitoring system and method for circular pipe equipment, solving the problem of how to achieve all-weather, high-precision, and blind-spot-free thermal monitoring of circular pipe equipment.

[0006] To address the aforementioned technical problems, the first aspect of this invention provides a thermal characteristic monitoring system for circular pipe equipment, comprising an attached sensing array assembly, an infrared thermal imaging acquisition module, and a data processing module; wherein, The attached sensing array assembly includes at least three fixing hose clamps sleeved on the outer wall of the circular tube device and a series array topology installed on each of the fixing hose clamps; the series array topology is formed by sequentially connecting several sampling units that are uniformly arranged along the axial direction of the circular tube device through aviation cables, and the number and spacing of the sampling units in the series array topology are determined by a curvature adaptive quantization array algorithm; each sampling unit integrates several sensors to collect thermal characteristic monitoring data of the circular tube device; The infrared thermal image acquisition module includes an infrared thermal imager to acquire infrared thermal images of the circular tube device; The data processing module is communicatively connected to the attached sensing array component and the infrared thermal image acquisition module, respectively, and is used to perform spatiotemporal registration and weighted fusion of the thermal feature monitoring data and the infrared thermal image to obtain the thermal feature monitoring results of the circular tube equipment.

[0007] A second aspect of the present invention provides a method for monitoring thermal characteristics of a system as described in the first aspect, characterized in that it includes: The thermal characteristic monitoring data of the circular tube device is collected in real time by the attached sensing array component, and the infrared thermal image of the circular tube device is collected in real time by the infrared thermal image acquisition module. The data processing module performs spatiotemporal registration and weighted fusion of the thermal feature monitoring data and the infrared thermal image to obtain the thermal feature monitoring results of the circular tube equipment.

[0008] Compared with the prior art, the beneficial effects of the embodiments of the present invention are as follows: By organically combining an attached sensing array component with an infrared thermal imaging acquisition module, the former enables high-precision online monitoring of key areas around the clock, while the latter enables rapid, large-scale inspections and surveys. The fusion of their data forms a continuous circumferential thermal feature map, solving the problems of pure infrared methods being susceptible to environmental interference and unable to monitor continuously, while also compensating for the shortcomings of purely contact-based measurement points being discrete and having blind spots. A curvature-adaptive quantization array algorithm is used to determine the number and spacing of sampling units, adapting to various specifications of circular pipe equipment. The array design is standardized and replicable, ensuring that spatial sampling satisfies the Nyquist sampling theorem, avoiding thermal signal aliasing, and significantly reducing redundant hardware costs and data processing burden. Using a fixed hose clamp as the mechanical load-bearing foundation, and in conjunction with aviation cables for series arraying, this structure can still operate under high temperature, high pressure, or strong convection industrial environments. Maintaining stable thermal coupling between the sampling unit and the pipe wall ensures the continuity and repeatability of data from long-term online monitoring. Through spatiotemporal registration, the temperature data sampled by the contact sensor and the infrared thermal image are mapped to the same spatiotemporal coordinate system. Then, through a weighted fusion algorithm, background noise and clutter in the infrared image can be effectively eliminated, while compensating for local measurement errors caused by the installation offset of the contact sensor. The final output thermal feature monitoring results are significantly better than those from a single data source in terms of sensitivity and specificity, and can identify early hidden dangers such as blockage, scaling, leakage, or local overheating inside the circular pipe earlier. It not only effectively overcomes the problems of traditional monitoring methods being greatly affected by environmental interference and unable to monitor continuously, but also effectively solves the blind spots and cost problems of empirical point placement by setting up a series array topology, accurately realizing all-weather, high-precision, and blind-spot-free thermal monitoring of circular pipe equipment. Attached Figure Description

[0009] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a structural diagram of a thermal characteristic monitoring system for circular tube equipment provided in a certain embodiment of the present invention; Figure 2 This is a schematic diagram of a sampling unit structure provided in a certain embodiment of the present invention; Figure 3 This is a schematic diagram of the installation of a sampling unit provided in a certain embodiment of the present invention; Figure 4 This is an infrared thermal image before preprocessing provided in a certain embodiment of the present invention; Figure 5 This is a preprocessed enhanced infrared thermal image provided in a certain embodiment of the present invention; Figure 6This is a flowchart of a method for monitoring the thermal characteristics of a circular tube device according to a certain embodiment of the present invention; Figure label: The components include: 1. Series sampling unit; 2. First aviation connector; 3. Threaded hole; 4. Fixing iron plate; 5. Heat flux sensor; 6. Annular sealing ring; 7. Heat flux sensor outlet; 8. Temperature sensor; 9. Master node sampling unit; 10. Second aviation connector; 11. Fixing hose clamp; 100. Attached sensing array assembly; 200. Infrared thermal imaging acquisition module; 300. Data processing module. Detailed Implementation

[0011] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings and examples. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0012] It should be understood that the step numbers used in the text are for ease of description only and are not intended to limit the order in which the steps are performed.

[0013] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0014] The terms “comprising” and “including” indicate the presence of the described feature, whole, step, operation, element and / or component, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or collections thereof.

[0015] The term “and / or” refers to any combination of one or more of the associated listed items, as well as all possible combinations, and includes these combinations.

[0016] In one embodiment, such as Figure 1 As shown, the first aspect of the present invention provides a thermal characteristic monitoring system for circular pipe equipment, including an attached sensing array assembly 100, an infrared thermal imaging acquisition module 200, and a data processing module 300; wherein, The attached sensing array assembly 100 includes at least three fixing hose clamps 11 sleeved on the outer wall of the circular tube device and a series array topology mounted on the fixing hose clamps 11. The series array topology is formed by sequentially connecting several sampling units that are uniformly arranged along the axial direction of the circular tube device through aviation cables. The number and spacing of the sampling units in the series array topology are determined by a curvature adaptive quantization array algorithm. Each sampling unit integrates several sensors to collect thermal characteristic monitoring data of the circular tube device. The infrared thermal image acquisition module 200 includes an infrared thermal imager to acquire infrared thermal images of the circular tube device; The data processing module 300 is communicatively connected to the attached sensing array component 100 and the infrared thermal image acquisition module 200, respectively, and is used to perform spatiotemporal registration and weighted fusion of the thermal feature monitoring data and the infrared thermal image to obtain the thermal feature monitoring results of the circular tube equipment.

[0017] Specifically, the thermal feature monitoring system provided by this invention combines an attached sensing array with infrared thermal imaging and is designed for thermal fault monitoring scenarios of circular pipe equipment such as GIL containing curved pipe sections. It includes an attached sensing array component 100, an infrared thermal imaging acquisition module 200, and a data processing module 300, realizing dual-modal collaboration of fixed-point online monitoring and full-area inspection and blind spot filling.

[0018] The attached sensing array assembly 100 is the core of the system's contact monitoring, used to acquire high-precision temperature and heat flux data of the circular pipe wall. It includes at least three fixed hose clamps 11 fitted onto the outer wall of the circular pipe device and a series array topology mounted on the fixed hose clamps 11. The fixed hose clamps 11 are made of 304 stainless steel and have a preferred width of 20 mm and a preferred thickness of 1.5 mm. They are equipped with a bolt locking mechanism. The hose clamps with the corresponding circumference specifications are selected according to the diameter of the circular pipe. During installation, they are wrapped around the outer wall of the circular pipe and locked. Two fixed hose clamps 11 are respectively fitted onto the starting end and the end end of the circular pipe device. The remaining fixed hose clamps 11 are evenly arranged between these two fixed hose clamps 11 and are fitted onto the middle section of the circular pipe device. These hose clamps are kept parallel in the pipe wall axis and have a consistent tightening torque (set to 15 N·m to prevent excessive tightness from damaging the pipe wall or excessive looseness from causing slippage).

[0019] In one embodiment, the attached sensing array assembly 100 further includes a fixing plate 4 and fasteners; wherein, the fixing plate 4 is an L-shaped stainless steel stamping part, one end of which has a through hole corresponding to the fixing threaded hole 3 of the sampling unit, and the other end is clamped on the upper and lower edges of the fixing hose clamp 11. During installation, after the fixing plate 4 is clamped into the hose clamp, the fastener (preferably an M3 screw) is passed through the through hole of the fixing plate 4 and tightened into the threaded hole 3 of the sampling unit, so that the sampling unit is firmly clamped on the hose clamp, forming a stable overall structure.

[0020] The serial array topology installed on the fixed hose clamp 11 is formed by sequentially connecting several sampling units that are uniformly arranged along the axial direction of the circular pipe equipment via aviation cables; the schematic diagram of the sampling unit structure is shown below. Figure 2 As shown in the diagram, the sampling unit installation diagram is as follows: Figure 3 As shown, each sampling unit has a square box structure, with the outer shell made of die-cast aluminum alloy and anodized for corrosion protection. The preferred dimensions are 60mm × 60mm × 20mm. The front of the box has two aviation connectors: a first aviation connector 2 (input / serial interface) on the left and a second aviation connector 10 (output / master node interface) on the right, both using 4-pin M12 aviation connectors with an IP67 protection rating. Adjacent sampling units are connected in series via the first aviation connector 2 and aviation cable adapters, forming a ring-shaped series array topology. The second aviation connector 10 of each master node sampling unit 9 serves as the convergence output port of its respective series array topology and communicates with the data processing module 300 via aviation cables, enabling electrical series connection and data exchange between the sampling units. Each side of the box extends a mounting lug with a threaded hole 3 for locking with a fixing plate 4.

[0021] The number and spacing of sampling cells in the cascaded array topology are determined by a curvature adaptive quantization array algorithm. In one embodiment, determining the number and spacing of sampling cells in the cascaded array topology using the curvature adaptive quantization array algorithm includes: The outer wall circumference is determined based on the diameter of the circular tube device, and the number of sampling units in the cascaded array topology is determined in combination with the minimum point density requirement. Calculate the curvature and circumferential angle of the bending plane of the circular tube equipment, and divide the circular tube equipment into standard circular tube segments and curved circular tube segments according to the curvature. The spacing of the sampling units in the series array topology is determined based on the outer circumference of the standard circular pipe section and the number of sampling units in the series array topology set on its outer wall. The bending deformation compensation factor is determined based on the bending curvature and circumferential angle of the bending plane of the curved circular pipe segment, and the spacing of the sampling units in the series array topology is determined by combining the outer wall perimeter of the curved circular pipe segment and the number of sampling units in the series array topology set on its outer wall.

[0022] Specifically, this invention calculates the outer wall circumference based on the nominal diameter of the curved circular tube, providing a basis for the sampling unit layout. That is, π is multiplied by the diameter of the circular tube (where π is 3.1416) to obtain the outer wall circumference. Considering the blind-zone-free requirement for thermal monitoring of curved circular tubes, the number of sampling units must meet the minimum layout density, while ensuring that the number of sampling units in the series array topology is ≥8. Therefore, the calculation process for the number of sampling units is as follows: In the formula, n is the number of sampling units in the cascaded array topology; C is the perimeter of the outer wall; The basic spacing coefficient is preferably 0.4m. Adjusting this coefficient can control the density of single-row placement of curved circular pipes and adapt to the placement density requirements of circular pipes of different diameters. This indicates rounding up. Since this invention uses a single-axis series chain as its basic topology, the aforementioned number of sampling units is considered as the number of units arranged circumferentially at the same axial position.

[0023] Based on the 3D design model of the pipe segment, the spatial centerline coordinates of the pipe are extracted, and the curvature (reciprocal of the radius) and circumferential angle of the bending plane of each micro-element segment are calculated. (Using the bending plane of the pipe segment, i.e., the vertical plane where the elbow is located, as the reference plane, a local cylindrical coordinate system is established, and the circumferential angle φ of the bending plane is defined: on the cross-section of the pipe wall, φ=0° points to the outside of the bend, i.e., the tension side, away from the center of the elbow; φ=180° points to the inside of the bend, i.e., the compression side, facing the center of the elbow; φ=90° and 270° are the neutral axis sides. It should be noted that the circumferential angle of the bending plane of the straight segment is empty and does not need to be calculated; this angle only applies to the bending segment.) If the calculated curvature is zero, it is determined to be a standard circular pipe segment (straight segment); otherwise, it is a bent circular pipe segment.

[0024] For a standard circular pipe segment, the horizontal spacing between sampling units in its tandem array topology is the arc length spacing between sampling units along the circumference of the curved pipe, and its value is equal to the ratio of the circumference to the number of sampling units; the column spacing (that is, the spacing between two sampling units located on the same axis in two adjacent tandem array topologies in a standard circular pipe segment) is the arc length between the intersection of the plane containing the two columns of sampling units and the central axis of the curved pipe. In the formula, b is the column spacing; L is the horizontal spacing; and d is the diameter. However, in actual installation, the column spacing can be adjusted within the range of [b, 1.5b] or [1.5b, 2.5b] according to the specifications of the round pipe to adapt to the layout requirements of round pipes of different diameters.

[0025] For a curved circular pipe segment, the calculation process for the transverse spacing between sampling units in its tandem array topology is the same as that for a standard circular pipe segment. However, considering the inconsistency between the inner and outer curved surfaces, the column spacing is first determined based on the curvature and circumferential angle of the curved circular pipe segment. Then, the bending deformation compensation factor is determined by combining the bending deformation compensation factor with the outer perimeter of the curved circular pipe segment and the number of sampling units in the tandem array topology set on its outer wall. This process is expressed by the following formula: In the formula, For deployment Bending deformation compensation factor; For deployment The bending deformation compensation factor in the previous column; For deployment The curvature of the curved circular pipe is defined by φ, which is the circumferential angle of the bending plane. In practice, the angle corresponding to the previous column of sampling units is typically used. In actual applications, the angle coinciding with the bending plane is defined as 0 degrees and 180 degrees, with 0 degrees on the outer side of the bend and 180 degrees on the inner side. The positive direction of the angle follows the right-hand screw rule with the axial extension direction of the circular pipe. The algorithm ultimately generates a coordinate table, including the number, location, and coordinates of sampling units in each circular pipe segment.

[0026] This invention introduces the spatial geometry of the pipe (bending curvature and circumferential angle of the bending plane) as the core independent variable, so that the physical field sampling conforms to the thermo-mechanical coupling distortion law and completely eliminates the "blind zone of the hot zone of the bend pipe". It automatically adapts to any bending angle and bending diameter ratio, has strong universality, can effectively control the total number of points, avoid excessive redundancy or insufficient, and provide prior geometric constraints for subsequent spatiotemporal registration.

[0027] The sampling units are divided into two types: serial sampling units 1 and master node sampling units 9. They have the same external dimensions but differ only in their internal circuit functions. Serial sampling units 1 only have serial forwarding functions and are used as intermediate nodes in the array link. Master node sampling units 9 additionally have bus aggregation and protocol conversion functions and serve as the sole output node for the entire data column. The total number of sampling units in each serial array topology is not less than 8, and at least one master node sampling unit 9 is configured in the serial array topology, with the rest being serial sampling units 1.

[0028] The sampling unit integrates a temperature sensor 8 and a heat flux sensor 6 in the central area of ​​its bottom surface, both with their sensing surfaces flush with the bottom surface. The temperature sensor 8 is a Class A PT1000 platinum resistance thermometer with a temperature range of -40℃ to +150℃ and an accuracy of ±0.15℃. It is fixed to the center of the bottom surface using thermally conductive adhesive. The probe is embedded in the center of the bottom contact surface of the housing, directly thermally coupled to the pipe wall, and is used to collect the surface temperature of the pipe wall. The heat flux sensor 6 is a differential thermopile type heat flux sensor with a sensitive area of ​​10mm × 10mm, a range of 0 to 5000W / m², and a sensitivity of 10μV / (W / m²). It is arranged around the temperature sensor 8 to collect the heat flux density normal to the pipe wall, reflecting the direction and intensity of heat transfer.

[0029] A ring-shaped sealing ring 6, made of fluororubber with a cross-sectional diameter of 2mm, is provided around the bottom periphery of the sampling unit. This ring surrounds the sensing areas of multiple sensors belonging to the sampling unit. When the sampling unit is pressed against the tube wall, the sealing ring deforms under pressure, forming a sealed cavity to protect the internal sensors from moisture and dust corrosion. In another embodiment, the ring-shaped sealing ring 6 is used to seal the contact surface between the sampling unit and the tube wall. The entire bottom surface area inside the sealing ring is covered with a layer of highly thermally conductive flexible silicone pad, 1mm thick, with a thermal conductivity of 4.0W / (m·K) and a Shore hardness of 35HA. This effectively fills the tiny air gaps between the housing of the sampling unit and the curved outer wall of the cylindrical device, reducing contact thermal resistance.

[0030] The sampling unit also integrates a main control subunit and a local communication subunit. The main control subunit uses an STM32L431 low-power microcontroller with a built-in 12-bit high-precision ADC to acquire the resistance signal from temperature sensor 8 and the millivolt-level voltage signal from heat flux sensor 6. The local communication subunit uses an RS485 differential bus, supports the Modbus-RTU protocol, and has a default communication rate of 9600bps, which is configurable. The main control subunit processes the thermal characteristic monitoring data, including temperature and heat flux data, collected by the sensors in its sampling unit via electrical connection and then sends it in real time to the background monitoring system.

[0031] In one embodiment, the main control subunit is configured to: The temperature data is sequentially subjected to moving average filtering, Kalman denoising, and factory calibration compensation to obtain the processed temperature data. The heat flux data collected by each sampling unit on the same circumferential column are sorted by polar angle to form a circumferential heat flux sequence. The total heat flux power is obtained by performing a line integral over the entire circumference of the circumference heat flux sequence. Based on the directional distribution characteristics of the circumferential heat flux sequence, the total heat flux power is decomposed using the least squares method to obtain the heat source components for fault diagnosis.

[0032] Specifically, this invention sets the sliding window length to 10 sampling points (i.e., a 100ms time window). For each temperature data collected, the arithmetic mean of the 10 data points within the window is calculated as the coarse-filtered output temperature at the current moment. This step effectively eliminates single-point abnormal jumps caused by occasional electromagnetic radiation on site.

[0033] The coarse-filtered output temperature after moving average is used as the observation input of the Kalman filter. The thermal time constant model of the pipe wall material (carbon steel) is solidified in the main control subunit. The state transition matrix A=1 (temperature slowly changing), the control input is 0, the process noise covariance Q=0.01, and the measurement noise covariance R=0.05 (obtained from the sensor factory test). The filter performs a prediction-update iteration every 10ms and outputs the optimal state estimate Tk. This step further compresses the residual white noise amplitude to less than 1 / 5 of the original value, and the response delay does not exceed two sampling periods (20ms), ensuring rapid tracking of the actual temperature rise.

[0034] The main control subunit's Flash memory pre-stores the third-order polynomial compensation coefficients (a0, a1, a2, a3) obtained during the factory calibration of this specific sampling unit. These coefficients are obtained through point-by-point calibration fitting in a constant-temperature oil bath from -20℃ to 300℃. The main control subunit substitutes the optimal state estimate Tk into this polynomial: T = a0 + a1·Tk + a2·Tk² + a3·Tk³, and outputs the final processed temperature data T. This step ensures that even under the same thermal environment, the measurement deviation of multiple sampling units does not exceed ±0.1℃.

[0035] The heat flux density sensor (unit: W / m²) of the sampling unit synchronously acquires the original heat flux signal. After passing through the same low-pass filter, the filtered heat flux is obtained. The main control subunit rearranges these filtered heat flux values ​​according to the physical polar angle from smallest to largest, forming a circumferential heat flux sequence. The sign of each q is recorded according to the following convention: a positive value indicates heat transfer from the inside to the outside of the pipe wall (heat dissipation), and a negative value indicates heat transfer from the outside to the inside (heat absorption). Subsequently, the trapezoidal rule is used to perform numerical line integration on this circumferential heat flux sequence to calculate the total heat flux power at the current cross-section. This value directly reflects the total heat power lost by the steam inside the pipe at that cross-section.

[0036] Establish a directional distribution model: It is assumed that the circumferential heat flux distribution consists of three superimposed components: (1) a uniform component (C0, representing isotropic basic thermal radiation / convective heat dissipation); (2) a first-order harmonic along the vertical direction (C1·cosθ+S1·sinθ, representing asymmetric heat transfer caused by gravity-induced medium stratification or bottom liquid accumulation, θ being the angle of the filtered heat flux); (3) a very small number of higher-order components along the horizontal direction (to simplify the calculation, only the first two orders are taken in this embodiment). Therefore, a fitting function is constructed: qfit(θ) = C0 + C1·cosθ + S1·sinθ. Substituting the measured values ​​of the circumferential heat flux sequence into this function, several equations are formed (three unknowns: C0, C1, and S1). The optimal solution is obtained using the least squares method. The main control subunit internally calls matrix inversion to calculate: C0 (the value of the uniform heat dissipation base); C1 (the amplitude of the vertical component, a positive value indicates that the heat dissipation at the top is stronger than at the bottom); and S1 (the amplitude of the horizontal component, which is almost negligible in practical applications). Subsequently, the main control subunit further decomposes the total heat flux power into heat source components with clear physical meanings: the first heat source component is the product of C0 and the corresponding circumference of the circular tube (this part is regarded as the overall heat loss of the insulation layer); and the second heat source component is the product of the square root of the sum of the squares of C1 and S1 and the diameter of the circular tube, which is regarded as the abnormal source component that enhances heat transfer outward due to local overheating caused by condensate deposition at the bottom of the tube). Both can be used for fault diagnosis.

[0037] This invention ensures the truth value of temperature data through triple purification filtering, transforms scalar integration into vector decomposition, reveals the directional fingerprint of heat flow, and enables the system to quantitatively distinguish between heat from the internal medium of the pipe center and interference heat from environmental radiation / convection. Filtering, integration, and least squares decomposition are all completed in the main control subunit of the sampling unit. Only lightweight characteristic parameters such as processed temperature value, total heat flow power, and decomposed heat source components are uploaded to the data processing module 300, rather than the original high sampling rate waveform. This reduces the storage and computing power requirements of the data processing module 300 by about 70%, making it particularly suitable for ultra-large-scale monitoring networks with up to hundreds of sampling units.

[0038] This invention determines the number and spacing of sampling units in each series array topology installed on the fixed hose clamp 11 based on the curvature adaptive quantization array algorithm. Then, according to the calculation results, an aviation cable (preferably F46-4 core with shielding, temperature resistant 200℃) is arranged along the axial direction of the circular pipe. The cable is secured to the pre-reserved lugs of the three hose clamps with nylon cable ties. At each preset axial position, a branch cable is led out from the main line of the aviation cable and connected to a sampling unit. Each sampling unit is a small circular metal-encapsulated probe (15mm in diameter, 5mm thick), which integrates a PT1000 high-precision platinum resistance thermometer (for measuring absolute temperature) and a heat flux density plate (for measuring radial heat flux). After applying thermal grease to the bottom surface of the probe, it is tightly fitted to the outer surface of the pipe wall by a spring clamp extending laterally from the hose clamp. All units are connected in series via the aviation cable and then connected to the field data acquisition station via a bus interface.

[0039] The infrared thermal imaging acquisition module 200 is used to acquire the surface temperature distribution of the circular tube, serving as a supplement and extension to the attached array. It includes an infrared thermal imager, which can be an uncooled focal plane infrared thermal imager with a detector resolution of 384×288 pixels, thermal sensitivity NETD≤50mK, temperature range -20℃~+150℃, and accuracy ±2℃ or ±2% of the reading. It can be used for portable handheld inspection or deployed on a tracked robot for automatic inspection. Its optical axis is perpendicular to the axial centerline of the tube segment, and the field of view is adjusted to completely cover the circular tube equipment for acquiring infrared thermal images of the tube equipment. First, the infrared thermal imager is calibrated on-site using a standard blackbody furnace. Then, the emissivity of the tube wall material is corrected using a thermocouple patch of known temperature (as a reference), setting the emissivity parameter to 0.92 (for oxidized carbon steel surfaces).

[0040] In one embodiment, the infrared thermal image acquisition module 200 further includes an infrared image optimization network unit, deployed in a backend server or edge computing gateway, for enhancing the infrared thermal images acquired by the infrared thermal imager.

[0041] The infrared image optimization network unit is used for: The infrared thermal image is integrated with a one-dimensional convolution kernel along the channel dimension to obtain infrared features. Then, the infrared features are extracted with a depthwise separable convolution along the spatial dimension to extract local texture features and output spatial enhancement features. After flattening the spatial enhancement features along the feature dimension, perform a fast Fourier transform to obtain a complex frequency domain feature sequence, which is then multiplied element-wise with the learnable complex mask matrix. The result of the multiplication is then processed by an inverse Fourier transform to obtain the real part feature map and the imaginary part feature map. The real feature map and the imaginary feature map are concatenated in complex space dimension to obtain fused features, which are then used for image reconstruction through pixel upsampling to obtain an enhanced infrared thermal image.

[0042] Specifically, this invention employs an infrared image optimization network to preprocess infrared thermal images: First, 1D convolution is used to integrate cross-channel information of the image to obtain infrared features. Then, depthwise convolution is used to enhance the correlation of its local spatial features, that is, to extract the local texture features of the image and output spatial enhancement features. The formula is as follows: In the formula, For network input, H , W , C These are the height, width, and number of channels of the infrared image, respectively. and Convolution kernels for 1D convolution and depthwise convolution, respectively. , These are the biases for 1D convolution and depthwise convolution, respectively. and These are the multidimensional feature outputs of 1D convolution and depthwise convolution, respectively. The PReLU activation function is used. This is an element-wise convolution operation.

[0043] Then, the GeLU activation function is used for processing. To avoid feature dead zones and improve nonlinear fitting ability, and to compare the results with... The formula for element-wise product is as follows: In the formula, To calculate the product, we pass it into a 1D convolution PCov2, adjusting the number of channels to accommodate frequency domain filtering, as shown in the following formula: In the formula, For filtering results; The convolution kernel is a 1D convolution PCov2; The bias is for the 1D convolution PCov2.

[0044] The selection results are then subjected to frequency domain filtering to improve the model's response to sparse high-frequency infrared features, as shown in the following formula: In the formula, for The flattened 1D sequence, For flattening operation, The complex feature sequence after Fourier transform. For Fast Fourier Transform; Features of the filtered complex sequences To and Learnable complex matrices of the same type; The complex frequency domain feature sequence after inverse Fourier transform. This is the inverse Fourier transform. To perform the real part operation, To perform the imaginary part operation, Folding operations are used to recover high-dimensional features in sequence order. , These are the real part feature map and the imaginary part feature map, respectively.

[0045] After completing the frequency domain filtering, the output features are concatenated. The imaginary part features can effectively characterize the periodic repetitive elements in the data spectrum. Instead of directly discarding the imaginary part features, we introduce the role of the imaginary part through methods such as feature decomposition, self-cross-matching operation, and local enhancement. This can avoid the loss of key features and broaden the convergence path to the complex space. The formula is as follows: In the formula, , These are the depthwise convolution outputs of the real part feature map and the imaginary part feature map, respectively. , , , These represent the convolution kernel and bias of the corresponding convolutional layer; , , After linear transformation Characteristic components, bond components, and state components. The softmax activation function is used. For tensor batch matrix multiplication, The maximum dimension length of the bond component. This is to introduce the imaginary part; For channel splicing output, This is to stitch two or more feature maps together along the channel dimension to form a larger feature map.

[0046] After channel stitching is completed, the feature map is reconstructed at high resolution by pixel upsampling and filtering, as shown in the following formula: In the formula, , These are the depthwise convolution outputs of DConv4 and DConv5, respectively. , , , These represent the convolution kernel and bias of the corresponding convolutional layer. Filter the output features of upsampled pixels. For pixel filtering upsampling operation, The output of the model is an enhanced infrared thermal image.

[0047] Infrared thermal image before preprocessing, as shown Figure 4 As shown, the preprocessed enhanced infrared thermal image is as follows: Figure 5 As shown, the preprocessed thermal image has higher accuracy and clarity. The infrared image optimization network is trained end-to-end. The training set contains a large number of real GIL infrared images under different environments and their corresponding high-quality reference images. The joint loss function is L=λ1·LMSE+λ2·(1-SSIM), where LMSE is the mean squared error loss, SSIM is the structural similarity, and λ1 and λ2 are weight coefficients. The optimizer is Adam, with an initial learning rate of 1e-4 and 100 training epochs.

[0048] This invention extracts "axial thermal gradient" and "circumferential texture" by decoupling the channel dimension and spatial dimension, thus reducing the computational load. It can learn complex frequency domain masks to achieve spectral separation of "noise-thermal signal", which is more physically interpretable than simple spatial domain filtering. It retains and outputs real and imaginary feature maps, and then performs complex spatial dimension splicing to ensure that the phase gradient of the temperature field is completely preserved. This is crucial for accurately determining whether heat is transferred from the inside to the outside or vice versa for curved pipe sections, avoiding misjudgments caused by directional ambiguity.

[0049] The data processing module 300 is deployed on the background monitoring server and is the data analysis and decision-making center of the entire device. In terms of hardware, it is an industrial-grade server or edge computing gateway. It is connected to the attached sensing array component 100 and the infrared thermal image acquisition module 200 respectively. It is used to perform spatiotemporal registration and weighted fusion of thermal feature monitoring data and infrared thermal images to obtain the thermal feature monitoring results of the circular tube device.

[0050] In one embodiment, the data processing module 300 is further configured to: The infrared thermal image is sequentially subjected to effective region segmentation and geometric correction to obtain a planar infrared image; The planar infrared image is sequentially subjected to cylindrical unfolding and cylindrical physical coordinate mapping processing, so that each infrared pixel in the planar infrared image corresponds to physical coordinates including axial position and circumferential arc length, forming an infrared mapping image. Using the infrared map as a reference, the thermal feature monitoring data are sequentially subjected to physical coordinate matching and time-based interpolation registration to obtain the registered array sampling point dataset. A partitioned adaptive weighting strategy is used to determine the fusion weights, and the infrared mapping map and the registered array sampling point dataset are weighted and fused using the fusion weights to obtain the fused temperature value. Fault data is determined based on the fusion temperature value and the heat source component, and the fault data is used as the monitoring result of the thermal characteristics of the circular tube equipment.

[0051] Specifically, the data processing module 300 receives real-time data packets from the attached sensing array via an RS485 to Ethernet gateway and receives image data from the infrared thermal imager via an FTP or HTTP interface, thus completing the unified access, parsing, and storage of the two types of data.

[0052] A threshold segmentation and edge detection algorithm was used to extract the effective area of ​​the outer wall of the circular tube in the infrared thermal image. Interfering pixels such as background walls, supports, and pipelines were removed to obtain a mask image containing only the surface of the circular tube. For the tube wall image within the mask, perspective geometry correction was performed using known thermal imager intrinsic parameters (focal length, principal point) and installation extrinsic parameters (thermal imager distance from the tube center 2.5m, optical axis perpendicular to the tube axis and aimed at the midpoint of the tube segment). This transformed the tilted projection into an orthographic projection—equivalent to moving the thermal imager directly above the tube and shooting vertically along the tube axis. The corrected tube wall image is rectangular, with the tube axis along the horizontal direction and the circumference along the vertical direction. The pixel size is proportional in both the axial and circumferential directions (each pixel corresponds to a physical length of 0.5mm), resulting in a planar infrared image.

[0053] The corrected planar infrared image is further subjected to cylindrical unfolding mapping. The height (vertical direction) of the rectangular image corresponds to the perimeter, and the width (horizontal direction) corresponds to the pipe segment length. The outer wall of the circular pipe is a cylindrical curved surface, and the infrared image is a planar perspective projection. Direct matching will result in circumferential edge distortion. Therefore, this invention establishes a mapping model of "pixel coordinates → cylindrical physical coordinates of the circular pipe": Axial dimension: The horizontal direction of the planar infrared image corresponds to the axial direction of the circular pipe. A linear scaling factor is determined by the pixel positions at both ends of the circular pipe, and the horizontal coordinate of each pixel can be directly converted into the axial physical position (meters away from the reference point); Circumferential dimension: The vertical direction of the planar infrared image corresponds to the circumferential projection of the circular pipe. Combining the three known parameters of shooting focal length, object distance, and circular pipe diameter, the circumferential angle corresponding to the vertical coordinate of each pixel is calculated through trigonometric functions, and then converted into the circumferential arc length physical coordinate. After the mapping is completed, the planar infrared image is converted into a cylindrical unfolded temperature map. Each infrared pixel corresponds to a unique physical coordinate of "axial position + circumferential arc length", forming an infrared mapping map.

[0054] The physical installation coordinates of the sampling units have been accurately recorded using a curvature adaptive algorithm (the z-value and circumferential arc length s-value of each unit along the axial direction, where s is calculated from the pre-determined circumferential installation angle on the fixed hose clamp 11). The data processing module 300 projects the coordinates of these discrete points onto the temperature map, finds the four nearest infrared pixels, and records their infrared temperature values; the contact sensor data upload frequency is 10Hz (period 100ms), and the infrared frame rate is 30Hz (period 33.3ms). Using the infrared frame time as a reference, the temperature and heat flux values ​​of each sensor at the two nearest sampling times before and after the reference are linearly interpolated to obtain the synchronous estimate of the sensor at the reference time. After this processing, at each infrared frame time, a registered array sampling point dataset with the same coordinate system at the same time is obtained, and each sampling point contains three types of information: physical coordinates, calibration temperature, and heat flux.

[0055] This invention employs a partitioned adaptive weighting strategy to determine the fusion weights. For any spatial location, the normalized distance dnorm to the nearest sampling unit is calculated. When dnorm < 0.3 (near the measurement point), the fusion weight is 0.75; when 0.3 ≤ dnorm < 0.7 (transition region), the weight transitions linearly according to distance; when dnorm ≥ 0.7 (gap region), the fusion weight is 0.15. The fusion weights and the difference between 1 and the fusion weights are used to weight the temperature in the infrared mapping and the temperature in the registered array sampling point dataset by pixel, respectively, to obtain the fused temperature value of that pixel. This forms a complete fused temperature distribution map with the same size as the cylindrical unfolded temperature map, and each point has a high-precision fused temperature value.

[0056] By combining the fused temperature value and heat source component, a comprehensive fault determination logic is executed to obtain fault data, which is then output as the thermal characteristic monitoring result of the circular tube equipment. First, check if there are any local temperature surge areas in the fused temperature distribution map: set the axial gradient threshold to 15℃ / m and the circumferential gradient threshold to 10℃ / m (based on arc length). If a continuous area (area ≥ 50mm²) exceeds the threshold, it is marked as a "suspected thermal anomaly area".

[0057] Then retrieve the sensor heat source component corresponding to the abnormal area: if the circumferential position of the area corresponds to the bottom of the pipe (around s≈πr) and the second heat source component at that location is greater than 15W / m and the direction points to the bottom of the pipe (θ=180°), then it is determined to be "internal condensation / liquid accumulation causing local overheating", and the fault type and fault location are output as fault data, along with the highest fusion temperature value and timestamp.

[0058] If the abnormal area is located at the top of the pipe (s≈0) and the direction of the second heat source component points to the top of the pipe (θ=0°), then the environmental data is combined to determine whether it is caused by sunlight or external heat source radiation. If the fusion temperature is abnormal but the heat source component is mainly a uniform component, the first heat source component increases while the directional component is not obvious, then it is determined that the external insulation layer is damaged or damp, which is fault data.

[0059] If the heat source component and fusion temperature in the abnormal area are both non-directional, and the ambient wind speed suddenly increases, it may be a false alarm due to infrared noise. The system will only record the information and will not trigger an alarm.

[0060] In addition, fault data can be determined by combining the fusion temperature value with the Gaussian thermal theorem. Based on Gauss's theorem for heat, for steady-state heat conduction in cylindrical coordinates, radial heat flux density is derived by inverting the circumferential temperature distribution and heat flux distribution. A discretized heat conduction equation is established, and the ill-conditioned inverse problem is solved using the Tikhonov regularized least squares method to obtain a sequence of radial heat flux density distributed along the circumference. When the value at a certain angular position is significantly higher than the average, it is determined that there is an internal abnormal heat source at that location, and its circumferential angle and corresponding axial position are recorded.

[0061] Three diagnostic features were extracted: maximum temperature rise, abnormal heat flux power rate, and hotspot diffusion rate. The weights were determined using the analytic hierarchy process (AHP) to be 0.4, 0.35, and 0.25, respectively. After normalization, each feature was weighted and summed to obtain a comprehensive health score S (0–100 points). Based on the S value, four levels were defined: S < 30 for normal, 30 ≤ S < 60 for caution (blue alert), 60 ≤ S < 80 for abnormal (yellow warning), and S ≥ 80 for severe (red emergency warning). Warning information was communicated to maintenance personnel via audible and visual alarms, SMS, and app push notifications.

[0062] The assembly process of a thermal characteristic monitoring system for circular pipe equipment includes: 1. Quantization calculation of array parameters: Based on the design drawings of the GIL device to be monitored, key parameters are obtained: nominal diameter of the circular tube, curvature of the bending section, monitoring accuracy level, etc. The number and spacing of sampling units in each series array topology are determined by the curvature adaptive quantization array algorithm. 2. Sensor array attachment and installation: (1) Pre-assembly: Lay the fixed hose clamp flat on the workbench and mark the position of each sampling unit according to the calculated unit spacing. Clip the fixing iron plate into the upper and lower edges of the hose clamp, align it with the fixing screw holes on both sides of the sampling unit, and use M3 stainless steel screws to lock the sampling unit and the fixing iron plate onto the hose clamp. Complete the pre-assembly of all sampling units in sequence to form a modular array belt.

[0063] (2) On-site installation: Wrap the pre-installed array belt around the target section of the GIL tube and rotate it circumferentially to align one of the sampling units with the 0° position (top) of the bending plane. After confirming that the silicone pads on the bottom of each sampling unit are fully attached to the tube wall without any warping, use a torque wrench to evenly tighten the hose clamp locking bolts. The tightening torque should be controlled at 8 to 10 N·m to ensure that the array does not slip or loosen.

[0064] (3) Multi-column layout: Install each column array sequentially along the axial direction according to the calculated column spacing, with adjacent columns staggered by half a unit spacing (column circumferential deflection angle 17), forming staggered layout points to further reduce the axial monitoring blind zone.

[0065] 3. Connecting the network to the backend: (1) Unit cascading: Starting from the main node type sampling unit at the beginning of the column, connect its first aviation connector to the first aviation connector of the next cascading sampling unit with aviation cables, and connect them one by one until the sampling unit at the end of the column. Make sure the positioning keys of the cable connectors are aligned and tighten the lock nuts to ensure a waterproof seal.

[0066] (2) Main node output: Connect the second aviation plug of the main node sampling unit to the nearest RS485-Ethernet gateway through the main communication cable. The gateway then connects to the background data fusion and diagnostic subsystem through the industrial Ethernet.

[0067] (3) Power supply debugging: Connect the 12V DC power supply and observe the status indicator lights of each sampling unit to confirm that all are powered on normally. Test the communication link in the background to verify that the data of each sampling unit can be uploaded normally and the address encoding is correct.

[0068] This application addresses the challenge of achieving all-weather, high-precision, and blind-spot-free thermal monitoring of circular pipe equipment. A thermal characteristic monitoring system for circular pipe equipment is designed, which organically combines an attached sensing array component with an infrared thermal imaging acquisition module. The former enables all-weather, high-precision online monitoring of key areas, while the latter enables large-scale, rapid inspection and surveying. The data from both are fused to form a continuous circumferential thermal characteristic map. This solves the problems of pure infrared methods being susceptible to environmental interference and unable to monitor continuously, while also compensating for the shortcomings of pure contact methods, such as discrete measurement points and blind spots. A curvature adaptive quantization array algorithm is used to determine the number and spacing of sampling units, adapting to various specifications of circular pipe equipment. The array design is standardized and replicable, ensuring that spatial sampling satisfies the Nyquist sampling theorem, avoiding thermal signal aliasing, and significantly reducing redundant hardware costs and data processing burden. A fixed hose clamp is used as the mechanical load-bearing foundation, in conjunction with aviation... The cables are connected in series and arrayed. This structure can maintain stable thermal coupling between the sampling unit and the pipe wall even in high-temperature, high-pressure, or strong convection industrial environments, ensuring the data continuity and repeatability of long-term online monitoring. Through spatiotemporal registration, the temperature data sampled by the contact sensor and the infrared thermal image are mapped to the same spatiotemporal coordinate system. Then, through a weighted fusion algorithm, background noise and clutter in the infrared image can be effectively eliminated, while compensating for local measurement errors caused by the installation offset of the contact sensor. The final output thermal feature monitoring results are significantly better than those from a single data source in terms of sensitivity and specificity, and can identify early hidden dangers such as blockage, scaling, leakage, or local overheating inside the circular pipe earlier. It not only effectively overcomes the problems of traditional monitoring methods being greatly affected by environmental interference and unable to monitor continuously, but also effectively solves the blind spots and cost problems of empirical placement, accurately realizing all-weather, high-precision, and blind-spot-free thermal monitoring of circular pipe equipment.

[0069] It should be noted that each module in the aforementioned thermal characteristic monitoring system for circular pipe equipment can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module. For specific limitations regarding the thermal characteristic monitoring system for circular pipe equipment, please refer to the limitations described above; both have the same function and role, and will not be repeated here.

[0070] In another embodiment, such as Figure 6 As shown, a second aspect of the present invention provides a method for monitoring thermal characteristics of a system as described in the first aspect, characterized in that it includes: S1. The thermal characteristic monitoring data of the circular tube device is collected in real time by the attached sensing array component, and the infrared thermal image of the circular tube device is collected in real time by the infrared thermal image acquisition module. S2. The data processing module performs spatiotemporal registration and weighted fusion of the thermal feature monitoring data and the infrared thermal image to obtain the thermal feature monitoring results of the circular tube equipment.

[0071] It should be noted that although the steps in the flowchart above are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order requirement for the execution of these steps, and they can be executed in other orders.

[0072] In summary, this invention relates to the field of power system equipment condition monitoring and fault diagnosis technology, and discloses a thermal characteristic monitoring system and method for circular pipe equipment. The system includes an attached sensing array assembly, an infrared thermal imaging acquisition module, and a data processing module. The attached sensing array assembly comprises a series array topology formed by multiple sampling units connected in series and mounted on a fixed hose clamp. The number and spacing of the sampling units in the series array topology are determined by a curvature adaptive quantization array algorithm. Each sampling unit integrates several sensors to collect thermal characteristic monitoring data of the circular pipe equipment. The infrared thermal imaging acquisition module includes an infrared thermal imager to collect infrared thermal images of the circular pipe equipment. The data processing module performs spatiotemporal registration and weighted fusion of the two types of data to obtain the thermal characteristic monitoring results of the circular pipe equipment, achieving all-weather, blind-spot-free, high-precision thermal characteristic monitoring and early fault warning for curved circular pipe equipment.

[0073] The various embodiments in this specification are described in a progressive manner. For directly identical or similar parts of the embodiments, refer to each other. Each embodiment focuses on its differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. It should be noted that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0074] The embodiments described above are merely preferred embodiments of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various improvements and substitutions without departing from the technical principles of this invention, and these improvements and substitutions should also be considered within the scope of protection of this application. Therefore, the scope of protection of this patent application should be determined by the scope of the claims.

Claims

1. A thermal characteristic monitoring system for circular pipe equipment, characterized in that, It includes an attached sensing array assembly, an infrared thermal imaging acquisition module, and a data processing module; among which, The attached sensing array assembly includes at least three fixing hose clamps sleeved on the outer wall of the circular tube device and a series array topology installed on each of the fixing hose clamps; the series array topology is formed by sequentially connecting several sampling units that are uniformly arranged along the axial direction of the circular tube device through aviation cables, and the number and spacing of the sampling units in the series array topology are determined by a curvature adaptive quantization array algorithm; each sampling unit integrates several sensors to collect thermal characteristic monitoring data of the circular tube device; The infrared thermal image acquisition module includes an infrared thermal imager to acquire infrared thermal images of the circular tube device; The data processing module is communicatively connected to the attached sensing array component and the infrared thermal image acquisition module, respectively, and is used to perform spatiotemporal registration and weighted fusion of the thermal feature monitoring data and the infrared thermal image to obtain the thermal feature monitoring results of the circular tube equipment.

2. The thermal characteristic monitoring system for circular pipe equipment according to claim 1, characterized in that, The attached sensing array assembly also includes a fixing iron plate and fasteners; the sampling unit is provided with a threaded hole, and the fixing iron plate is provided with a through hole adapted to the threaded hole; the fasteners pass through the through hole and the threaded hole to connect the fixing iron plate, the sampling unit and the fixing hose clamp.

3. The thermal characteristic monitoring system for circular pipe equipment according to claim 1, characterized in that, The sampling units are divided into two types: serial sampling units and master node sampling units. Each type of sampling unit is equipped with a first aviation plug and a second aviation plug. On any fixed hose clamp, adjacent sampling units are connected to the aviation cable through the first aviation plug and are connected in series to form a ring-shaped serial array topology. The second aviation plug of the master node sampling unit serves as the convergence output port of its serial array topology and is connected to the data processing module through the aviation cable.

4. The thermal characteristic monitoring system for circular pipe equipment according to claim 1, characterized in that, The bottom of the sampling unit is covered with a flexible silicone pad, which is disposed between the housing of the sampling unit and the outer wall of the cylindrical device. An annular sealing ring is provided at the contact point between the sampling unit and the outer wall of the cylindrical device, and the annular sealing ring is arranged around the sensing area of ​​multiple sensors of the sampling unit.

5. A thermal characteristic monitoring system for circular pipe equipment according to claim 3, characterized in that, The step of determining the number and spacing of sampling cells in the cascaded array topology using a curvature adaptive quantization array algorithm includes: The outer wall circumference is determined based on the diameter of the circular tube device, and the number of sampling units in the cascaded array topology is determined in combination with the minimum point density requirement. Calculate the curvature and circumferential angle of the bending plane of the circular tube equipment, and divide the circular tube equipment into standard circular tube segments and curved circular tube segments according to the curvature. The spacing of the sampling units in the series array topology is determined based on the outer circumference of the standard circular pipe section and the number of sampling units in the series array topology set on its outer wall. The bending deformation compensation factor is determined based on the bending curvature and circumferential angle of the bending plane of the curved circular pipe segment, and the spacing of the sampling units in the series array topology is determined by combining the outer wall perimeter of the curved circular pipe segment and the number of sampling units in the series array topology set on its outer wall.

6. A thermal characteristic monitoring system for circular pipe equipment according to claim 5, characterized in that, The total number of sampling units in each of the cascaded array topologies is not less than 8, and includes at least one of the master node type sampling units.

7. A thermal characteristic monitoring system for circular pipe equipment according to claim 1, characterized in that, The sampling unit further includes a main control subunit; the thermal characteristic monitoring data includes temperature data and heat flux data; The main control subunit is used for: The temperature data is sequentially subjected to moving average filtering, Kalman denoising, and factory calibration compensation to obtain the processed temperature data. The heat flux data collected by each sampling unit on the same circumferential column are sorted by polar angle to form a circumferential heat flux sequence. The total heat flux power is obtained by performing a line integral over the entire circumference of the circumference heat flux sequence. Based on the directional distribution characteristics of the circumferential heat flux sequence, the total heat flux power is decomposed using the least squares method to obtain the heat source components for fault diagnosis.

8. A thermal characteristic monitoring system for circular pipe equipment according to claim 1, characterized in that, The infrared thermal imaging acquisition module further includes an infrared image optimization network unit; the infrared image optimization network unit is used for: The infrared thermal image is integrated with a one-dimensional convolution kernel along the channel dimension to obtain infrared features. Then, the infrared features are extracted with a depthwise separable convolution along the spatial dimension to extract local texture features and output spatial enhancement features. After flattening the spatial enhancement features along the feature dimension, perform a fast Fourier transform to obtain a complex frequency domain feature sequence, which is then multiplied element-wise with the learnable complex mask matrix. The result of the multiplication is then processed by an inverse Fourier transform to obtain the real part feature map and the imaginary part feature map. The real feature map and the imaginary feature map are concatenated in complex space dimension to obtain fused features, which are then used for image reconstruction through pixel upsampling to obtain an enhanced infrared thermal image.

9. A thermal characteristic monitoring system for circular pipe equipment according to claim 7, characterized in that, The data processing module is also used for: The infrared thermal image is sequentially subjected to effective region segmentation and geometric correction to obtain a planar infrared image; The planar infrared image is sequentially subjected to cylindrical unfolding and cylindrical physical coordinate mapping processing, so that each infrared pixel in the planar infrared image corresponds to physical coordinates including axial position and circumferential arc length, forming an infrared mapping image. Using the infrared map as a reference, the thermal feature monitoring data are sequentially subjected to physical coordinate matching and time-based interpolation registration to obtain the registered array sampling point dataset. A partitioned adaptive weighting strategy is used to determine the fusion weights, and the infrared mapping map and the registered array sampling point dataset are weighted and fused using the fusion weights to obtain the fused temperature value. Fault data is determined based on the fusion temperature value and the heat source component, and the fault data is used as the monitoring result of the thermal characteristics of the circular tube equipment.

10. A method for monitoring the thermal characteristics of a system as described in any one of claims 1-9, characterized in that, include: The thermal characteristic monitoring data of the circular tube device is collected in real time by the attached sensing array component, and the infrared thermal image of the circular tube device is collected in real time by the infrared thermal image acquisition module. The data processing module performs spatiotemporal registration and weighted fusion of the thermal feature monitoring data and the infrared thermal image to obtain the thermal feature monitoring results of the circular tube equipment.