Distributed state monitoring and reconstruction method for overhead transmission line operation and maintenance protection

Abstract

This invention proposes a distributed state monitoring and reconstruction method for the operation and maintenance of overhead transmission lines. The method includes: injecting optical signals into a target optical fiber to obtain distributed strain data; mapping the spatial coordinates of the optical fiber to the geographical coordinates of the transmission line based on the feature points corresponding to the line structure in the data, and normalizing the relevant data to form a dataset; constructing a neural network model with geographical coordinates as input and conductor displacement as output; embedding the equations describing the conductor's dynamic behavior as physical constraints into the model loss function, using hard constraints to force the model output to satisfy the displacement boundary conditions at the line suspension points, and training the constrained model to obtain a state reconstruction model; inputting the spatiotemporal coordinates to be measured into the state reconstruction model to obtain a continuous displacement field and stress field over the entire span; and solving the inverse problem to retrieve the implicit physical parameters of the conductor. This invention achieves accurate reconstruction and health assessment of the entire physical state of the overhead transmission line during operation and maintenance.

Description

Technical Field

[0001] This invention belongs to the field of overhead transmission line operation and maintenance protection technology, and specifically relates to a distributed status monitoring and reconstruction method for overhead transmission line operation and maintenance protection. Background Technology

[0002] In the operation and maintenance of overhead power lines, some machinery may temporarily deviate from its operating route or illegally enter the protected area. If the distance to the line falls below the minimum safe distance, it can easily cause line tripping, equipment damage, and other accidents, posing significant risks to line safety and the personal safety of workers. Repairs also require specialized personnel and equipment and a considerable amount of time. Meanwhile, transmission lines are constantly exposed to the natural environment and are vulnerable to natural disasters such as conductor galloping, icing loads, strong winds, lightning strikes, and long-term fatigue damage. These factors have become significant hidden dangers affecting the safe and stable operation of the power grid. Traditional transmission line operation and maintenance mainly relies on manual inspections and periodic testing, which suffers from long inspection cycles, limited coverage, poor real-time performance, and high maintenance costs, making it difficult to promptly detect sudden hazards or dynamic risks. Therefore, building an intelligent monitoring system for transmission lines with real-time sensing, intelligent analysis, and automatic early warning capabilities to prevent mechanical damage to overhead lines and monitor natural disasters has become an important development direction and urgent need for the power industry.

[0003] Currently, the mainstream monitoring solutions for power transmission lines mainly rely on intelligent inspection systems based on visual perception. These systems deploy visible light or infrared cameras on transmission towers, combining edge computing and artificial intelligence image recognition technology to identify and issue warnings for external risk targets such as construction machinery, vehicles exceeding height limits, and smoke / fire within the transmission corridor. This approach achieves visualized monitoring of the transmission environment and is an important component of current intelligent operation and maintenance. However, existing visual monitoring solutions have the following significant shortcomings: First, they primarily target external environmental risks and cannot directly and quantitatively perceive the intrinsic physical states of the conductor itself, such as minute deformations, dynamic stresses, and structural fatigue. Second, they struggle to achieve high-precision, continuous, distributed measurements of micron-level deformations and dynamic stresses of long-distance conductors. Third, they are severely affected by weather conditions such as sunlight, rain, fog, and snow, as well as viewing angle obstructions, resulting in blind spots and the risk of misjudgment. Fourth, they primarily rely on data-driven models, lacking physical constraints on the conductor's mechanical behavior, and cannot convert localized sensing data into continuous, physically consistent state reconstruction results across the entire field, making it difficult to achieve the leap from superficial monitoring to fundamental diagnosis. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention proposes a distributed condition monitoring and reconstruction method for the operation and maintenance of overhead transmission lines. By reusing existing optical fiber composite overhead ground wires (OPGW) as the sensing medium, long-distance, distributed, and high-precision sensing of conductor conditions is achieved. Furthermore, through intelligent algorithms embedded with physical laws, accurate reconstruction and health assessment of the entire line's physical state are realized, thereby improving the depth, accuracy, and reliability of transmission line condition sensing and achieving accurate reconstruction and health assessment of the entire line's physical state in the operation and maintenance of overhead transmission lines.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention proposes a distributed condition monitoring and reconfiguration method for the operation and maintenance protection of overhead transmission lines, comprising the following steps: By injecting probe light signals into the optical fibers of overhead power transmission lines, acquiring and demodulating the back Rayleigh scattering signals at various points along the optical fiber, distributed strain data of the optical fiber can be obtained. Based on the feature points corresponding to the line structure in the distributed strain data, the spatial coordinates of the optical fiber are mapped to the geographical coordinates of the transmission line, and the geographical coordinates, spatial coordinates and distributed strain data are all normalized to form the input dataset. A neural network model is constructed with normalized geographic coordinates as input and conductor displacement as output. The control equations describing the dynamic behavior of the conductor are embedded as physical constraints into the loss function of the neural network model. Hard constraints are used to force the output of the neural network model to satisfy the displacement boundary conditions at the suspension point of the line. The neural network model with embedded physical constraints is trained using the input dataset to obtain the trained state reconstruction model. The spatiotemporal coordinates to be measured are input into the trained state reconstruction model to obtain the continuous displacement field and stress field of the transmission line over the entire span; and the implicit physical parameters of the conductor are obtained by solving the inverse problem.

[0006] Furthermore, the method also includes evaluating the structural health of the line based on the reconstructed displacement field, stress field, and inverted physical parameters. Specifically, based on the reconstructed displacement field and stress field, combined with the material's SN curve and linear cumulative damage theory, the fatigue damage index of the conductor is calculated and the remaining fatigue life is predicted.

[0007] Furthermore, a probe light signal is injected into the optical fiber composite overhead ground wire of the overhead transmission line to acquire and demodulate the backscattered Rayleigh signals at various points along the optical fiber, thereby obtaining the distributed strain data of the optical fiber composite overhead ground wire, specifically: Injecting a linearly swept-frequency chirped pulse optical signal into the fiber optic composite overhead ground wire; The backscattered Rayleigh signals at various points along the optical fiber are received and demodulated to obtain an interferometric intermediate frequency signal reflecting the strain change of the optical fiber. The intermediate frequency signal is processed to obtain the optical time delay variation distributed along the optical fiber; Based on the fiber optic elastic-optical effect, the optical time delay variation is converted into distributed strain data.

[0008] Furthermore, based on the fiber optic-elastic effect, the optical time delay variation is converted into distributed strain data, specifically as follows: ; in, For dynamic response; The refractive index; The elastic modulus of the optical fiber; The center frequency; This represents the pulse width.

[0009] Furthermore, the governing equations describing the dynamic behavior of the conductor are as follows: ; in, Mass per unit length of the conductor; For the conductor in position time lateral displacement; The spatial coordinates are along the arc length of the conductor; The damping coefficient; For bending stiffness; It is static tension; The external load per unit length of conductor; This is the loss function.

[0010] Furthermore, a hard constraint is used to force the neural network model output to satisfy the displacement boundary conditions at the suspension point of the line. Specifically, this is achieved by embedding a distance function in the output layer architecture of the neural network model, so that the network at the suspension point... and At that point, the displacement of the conductor must be 0, that is... and ;in This refers to the conductor span length.

[0011] Furthermore, during the training of the neural network model embedded with physical constraints using the input dataset, an adaptive weighting algorithm based on gradient magnitude is employed to dynamically adjust the balance between the data fitting loss term and the physical law loss term.

[0012] Furthermore, the implicit physical parameters include at least one of the following: effective conductor tension, equivalent elastic modulus, and equivalent icing load.

[0013] Furthermore, the method also includes analyzing the time-frequency characteristics and spatial distribution patterns of distributed strain data, and using a classification algorithm to identify at least one event type among conductor galloping, wind vibration, ice shedding impact, and external construction disturbance.

[0014] Secondly, this invention also proposes a distributed condition monitoring and reconfiguration system for the operation and maintenance protection of overhead transmission lines, comprising: Data acquisition unit: used to inject probe light signals into the optical fiber of overhead transmission line, acquire and demodulate the back Rayleigh scattering signals at each point along the optical fiber, and obtain the distributed strain data of the optical fiber. Data processing unit: Based on the feature points corresponding to the line structure in the distributed strain data, the spatial coordinates of the optical fiber are mapped to the geographical coordinates of the transmission line, and the geographical coordinates, spatial coordinates and distributed strain data are all normalized to form an input dataset. Model reconstruction unit: used to construct a neural network model with normalized geographic coordinates as input and conductor displacement as output; the control equations describing the dynamic behavior of the conductor are embedded as physical constraints into the loss function of the neural network model, and hard constraints are used to force the output of the neural network model to meet the displacement boundary conditions at the suspension point of the line. The neural network model with embedded physical constraints is trained using the input dataset to obtain the trained state reconstruction model. State inversion unit: used to input the spatiotemporal coordinates to be measured into the trained state reconstruction model to obtain the continuous displacement field and stress field of the transmission line over the entire span; and to invert the implicit physical parameters of the conductor by solving the inverse problem.

[0015] The effects described in the invention are merely those of the embodiments, and not all the effects of the invention. One of the above technical solutions has the following advantages or beneficial effects: This invention proposes a distributed state monitoring and reconstruction method for the operation and maintenance protection of overhead transmission lines. The method includes the following steps: injecting probe light signals into the optical fibers of the overhead transmission line to acquire and demodulate the backscattered Rayleigh signals at various points along the fiber to obtain distributed strain data of the optical fiber; mapping the spatial coordinates of the optical fiber to the geographical coordinates of the transmission line based on the feature points corresponding to the line structure in the distributed strain data, and normalizing the geographical coordinates, spatial coordinates, and distributed strain data to form an input dataset; constructing a neural network model with the normalized geographical coordinates as input and the conductor displacement as output; embedding the control equations describing the conductor's dynamic behavior as physical constraints into the loss function of the neural network model, and using hard constraints to force the neural network model output to satisfy the displacement boundary conditions at the line suspension points; training the neural network model with embedded physical constraints using the input dataset to obtain a trained state reconstruction model; inputting the spatiotemporal coordinates to be measured into the trained state reconstruction model to obtain the continuous displacement field and stress field of the transmission line over the entire span; and solving the inverse problem to inversely derive the implicit physical parameters of the conductor. Based on this method, a distributed condition monitoring and reconfiguration system for the operation and maintenance of overhead transmission lines is also proposed. This invention utilizes existing OPGW optical fibers in transmission lines to construct a long-distance continuous distributed strain and vibration sensing system. This system can directly measure minute deformations and dynamic stress changes along the conductor, overcoming the limitations of traditional visual monitoring which can only perceive external environmental targets, and achieving deep perception of the conductor's structural state.

[0016] This invention employs an ultra-narrow linewidth coherent light source and linear chirped pulse modulation technology, combined with coherent detection and cross-correlation time delay estimation methods, to achieve highly sensitive detection of minute strain changes. It can cover transmission lines of tens to hundreds of kilometers, significantly outperforming point sensors and camera monitoring in terms of spatial resolution and ranging range.

[0017] This invention is based on the principle of fiber optic distributed measurement, does not rely on visible light imaging, and is not sensitive to complex environmental conditions such as rain, fog, snow, and night, thus avoiding the perception blind spots and misjudgments caused by visual systems due to occlusion and changes in lighting.

[0018] This invention constructs a physical information neural network, embedding the conductor dynamics control equations, boundary conditions, and initial conditions into the algorithm model to achieve continuous reconstruction of the conductor's displacement, stress, and dynamic state across the entire field. This solves the key problems of traditional data-driven models lacking physical consistency and interpretability.

[0019] This invention can not only reconstruct the conductor deformation field, but also solve the implicit physical parameters such as conductor tension, equivalent elastic modulus and icing weight through inverse problem solving, providing a quantitative basis for conductor fatigue damage diagnosis and life assessment, and forming a closed-loop intelligent monitoring system from perception to diagnosis.

[0020] This invention uses a combination of analysis of vibration spectrum characteristics and strain space correlation patterns, along with machine learning algorithms, to intelligently classify events such as galloping, light wind vibration, ice shedding impact, and third-party construction disturbances. Based on a physical model, it predicts future motion trends, enabling early warning and risk classification. Attached Figure Description

[0021] Figure 1 This is a flowchart of the distributed state monitoring and reconfiguration method for operation and maintenance protection of overhead transmission lines proposed in Embodiment 1 of the present invention; Figure 2 This is an architecture diagram of the distributed state monitoring and reconfiguration method for operation and maintenance protection of overhead transmission lines proposed in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the physical neural network proposed in Embodiment 1 of the present invention; Figure 4 This is a schematic diagram of the distributed status monitoring and reconfiguration system for operation and maintenance protection of overhead transmission lines proposed in Embodiment 2 of the present invention. Detailed Implementation

[0022] To clearly illustrate the technical features of this solution, the invention will be described in detail below through specific embodiments and in conjunction with the accompanying drawings. The following disclosure provides many different embodiments or examples for implementing different structures of the invention. To simplify the disclosure of the invention, components and arrangements of specific examples are described below. Furthermore, reference numerals and / or letters may be repeated in different examples. This repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed. It should be noted that the components illustrated in the drawings are not necessarily drawn to scale. Descriptions of well-known components, processing techniques, and processes are omitted in this invention to avoid unnecessarily limiting the invention.

[0023] Example 1 Embodiment 1 of this invention proposes a distributed condition monitoring and reconstruction method for the operation and maintenance protection of overhead transmission lines. It solves the technical problem that existing visual monitoring schemes for transmission lines cannot directly, accurately, and in all weather conditions perceive minute deformations and stresses of the conductor body, and achieve intelligent reconstruction and in-depth diagnosis of the entire field condition based on physical laws.

[0024] Figure 1 This is a flowchart of the distributed state monitoring and reconfiguration method for operation and maintenance protection of overhead transmission lines proposed in Embodiment 1 of the present invention; Figure 2 This is an architecture diagram of the distributed state monitoring and reconfiguration method for operation and maintenance protection of overhead transmission lines proposed in Embodiment 1 of the present invention; combined with Figure 1 and Figure 2 The implementation process of Embodiment 1 of the present invention will be described in detail.

[0025] In step S1, a probe light signal is injected into the optical fiber of the overhead transmission line to acquire and demodulate the backscattered Rayleigh signals at various points along the optical fiber, thereby obtaining the distributed strain data of the optical fiber; the detailed process is as follows: A linearly swept-frequency chirped pulse optical signal is injected into the optical fiber composite overhead ground wire; the backscattered Rayleigh signals at various points along the optical fiber are received and demodulated to obtain an interferometric intermediate frequency signal reflecting the strain change of the optical fiber; the intermediate frequency signal is processed to obtain the optical time delay variation distributed along the optical fiber; based on the optical fiber elastic-optic effect, the optical time delay variation is converted into distributed strain data.

[0026] This invention uses an external cavity laser as the system source to generate a highly coherent continuous wave optical signal with a linewidth of less than 100Hz, so as to ensure that the phase noise is at an extremely low level within a monitoring range of 100km.

[0027] A periodic linear ramp voltage is generated by a signal generator to directly modulate the drive current of the external cavity laser, causing the laser output frequency to drift linearly over time. Subsequently, the continuous chirped light is shaped into a narrow pulse sequence with a high extinction ratio by a semiconductor optical amplifier to suppress in-band coherent noise.

[0028] The modulated swept-frequency chirped pulses are amplified by an optical fiber amplifier and filtered by a bandpass filter before being injected into the existing OPGW optical fiber of the overhead line through an optical circulator, transforming the entire power cable into a continuously distributed sensor array along the entire line.

[0029] Backscattered Rayleigh light transmitted through the optical fiber is coherently mixed with the local oscillator light from the laser at the receiving end. An interferometric intermediate frequency signal is extracted using a balanced photodetector, and common-mode rejection technology is employed to improve the signal-to-noise ratio of the system during long-distance detection. The continuously acquired optical trajectories are processed to extract the time delay caused by external physical disturbances. Based on the linear mapping characteristics of chirped pulses, the optical time delay variation is converted into distributed strain data, specifically: ; in, For dynamic response; The refractive index; The elastic modulus of the optical fiber; The center frequency; This represents the pulse width.

[0030] In step S2, based on the feature points in the distributed strain data that correspond to the line structure, the spatial coordinates of the optical fiber are mapped to the geographical coordinates of the transmission line, and the geographical coordinates, spatial coordinates, and distributed strain data are all normalized to form an input dataset. The algorithm identifies abrupt changes or specific waveform features at the tower suspension points in the distributed strain data. These points are used as anchor points to establish a mapping relationship from the fiber length coordinates to the spatial coordinates of the transmission line along the conductor. The mapped geographic coordinates, spatial coordinates, and distributed strain data are then normalized and scaled to a uniform numerical range [-1,1] to form a normalized input dataset, providing a stable numerical basis for the PINN algorithm.

[0031] In step S3, a neural network model is constructed with normalized geographic coordinates as input and conductor displacement as output; the control equations describing the dynamic behavior of the conductor are embedded as physical constraints into the loss function of the neural network model, and hard constraints are used to force the output of the neural network model to satisfy the displacement boundary conditions at the suspension point of the line. The neural network model with embedded physical constraints is trained using the input dataset to obtain the trained state reconstruction model. A multilayer perceptron is used as the base network to construct the PINN neural network. Spatiotemporal coordinates are used as inputs and the output is the predicted displacement of the conductor. The Sine activation function is used to improve the fitting ability of high-frequency vibration characteristics. Figure 3 This is a schematic diagram of the physical neural network proposed in Embodiment 1 of the present invention; The governing equations describing the dynamic behavior of the conductor are embedded as physical constraints into a neural network model, requiring the displacement field output by the neural network to be... The laws of mechanics must be strictly observed. In this scheme, the Euler-Bernoulli beam theory combined with tensioned string theory is primarily used to describe the vibration and galloping of the conductor. The algorithm uses automatic differentiation technology to calculate the derivative of the output with respect to spatiotemporal coordinates in real time, and constructs the following residual equation as part of the loss function. The specific formula is as follows: ; in, Mass per unit length of the conductor; For the conductor in position time lateral displacement; The spatial coordinates are along the arc length of the conductor; The damping coefficient; For bending stiffness; It is static tension; The external load per unit length of conductor; This is the loss function.

[0032] The formula takes into account inertial force, damping force, bending stiffness, and tension.

[0033] During training, the algorithm will force the residual to be... The mean square error tends to 0 across the entire domain, thus ensuring that the displacement curve generated by the model conforms to the laws of momentum conservation and energy balance.

[0034] Hard constraints are used to force the neural network model output to meet the displacement boundary conditions at the suspension point of the line. Specifically, this is achieved by embedding a distance function in the output layer architecture of the neural network model, ensuring that the network meets the displacement boundary conditions at the suspension point. and At that point, the displacement of the conductor must be 0, that is... and ;in This refers to the conductor span length.

[0035] In this invention, a distance function transformation is added to the output layer of the neural network, so that the network output is logically always equal to 0 at the boundary. This is more stable than the traditional method of adding penalty weights (soft constraints) and can prevent physically impossible drifts in long-distance inference.

[0036] Initial conditions and physical constraints requirements The initial displacement and initial velocity must conform to known static measurements (such as the initial sag distribution). This ensures that the dynamic evolution process has a correct logical starting point.

[0037] The initial displacement is expressed as: The initial velocity is expressed as: .

[0038] This invention, through physical constraints, allows the state reconstruction model to deduce the complete galloping pattern based on mechanical equations, even without any historical galloping samples, as long as the sensor detects even the slightest abnormal disturbance. It can filter out false alarms caused by environmental noise (such as vibrations from vehicles passing under fiber optic cables). Because signals generated by vehicle vibrations do not conform to the control equations for conductor motion, they are automatically excluded as physical noise due to the loss of physical terms. Using this constraint, the system can not only monitor displacement but also deduce the current tension or icing weight of the conductor through inverse problem solving, achieving in-depth diagnosis of structural health.

[0039] To achieve efficient training of the PINN network, the constructed physical consistency loss function includes: observation data term loss, partial differential equation residual loss, hard constraint boundary condition loss, and hierarchical gradient adaptive weight adjustment.

[0040] Loss of observed data items: Calculate the mean square error between the predicted displacement of the neural network at the sampling point and the measured strain sequence extracted by the hardware layer to ensure that the model output matches the real physical observation.

[0041] Partial differential equation residual loss: This relates to the conductor dynamics equations. Embedded loss function. Displacement calculated using automatic differentiation. Regarding time The second derivative (acceleration) and its relation to space The fourth derivative (bending stiffness contribution).

[0042] Among them, the residual loss term of the partial differential equation Approaching 0, specifically: ; Even in the blind zone where there is no sensor signal, the space of solutions can be constrained by physical laws to ensure the rationality of the prediction.

[0043] Hard constraint boundary condition loss: Apply hard physical constraints to the tower suspension points by embedding a distance function at the network architecture layer to force displacement. and This hard constraint at the architectural level eliminates the convergence instability problem caused by traditional weight penalties.

[0044] Hierarchical gradient adaptive weight adjustment: Introducing gradient-based adaptive weight coefficients The system dynamically adjusts the balance based on the magnitude of each loss term during training. For example, when the displacement prediction violates the conservation of momentum, the system automatically increases... The penalty weights ensure that the model always converges on a physically consistent trajectory.

[0045] In step S4, the spatiotemporal coordinates to be measured are input into the trained state reconstruction model to obtain the continuous displacement field and stress field of the transmission line over the entire span; and the implicit physical parameters of the conductor are obtained by solving the inverse problem.

[0046] As a continuous function, the PINN model can output the displacement field at any coordinate within the span, enabling a complete three-dimensional reconstruction of the conductor's galloping trajectory, sway amplitude, and static sag changes caused by icing, thus solving the problem that discrete sensors cannot detect the mid-span state.

[0047] By minimizing the physical residuals, the system can simultaneously invert implicit parameters such as the effective initial tension and equivalent elastic modulus of the conductor. If the inverted tension continues to decrease, it can be determined that the conductor is experiencing fatigue damage in the early stages of strand breakage or that there is a risk of slippage in the crimped pipe.

[0048] By combining the reconstructed frequency distribution (low-frequency galloping or high-frequency aerobatic vibration) with strain space correlation characteristics, K-means or LSTM algorithms are used to intelligently classify events. This can accurately distinguish between environmental wind loads, ice shedding impacts, and abnormal disturbances from third-party construction machinery, improving the accuracy of identification.

[0049] The system calculates the equivalent torque of the conductor under dynamic cyclic loads in real time and assesses the fatigue life of the line using the minimum linear cumulative damage theory. When the reconstructed physical displacement exceeds the safety and stability margin, the system will predict future movement trends through the physical model and trigger graded alarms in advance to guide maintenance personnel in carrying out targeted reinforcement.

[0050] Embodiment 1 of this invention proposes a distributed condition monitoring and reconstruction method for the operation and maintenance protection of overhead transmission lines. By reusing the existing OPGW optical fiber of the transmission line, a long-distance continuous distributed strain and vibration sensing system is constructed, which can directly measure the minute deformation and dynamic stress changes along the conductor. This breaks through the limitation of traditional visual monitoring that can only sense external environmental targets, and realizes in-depth perception of the structural state of the conductor itself.

[0051] To fully illustrate the implementation process of Embodiment 1 of the present invention, the data acquisition unit acquires strain data for the entire line at a sampling rate of 10Hz and a spatial resolution of 1m. The strain abrupt change characteristics at the two suspension points are identified, and coordinate mapping is completed. Data from a selected period is normalized and used to train the PINN model. After the model training converges, inputting real-time coordinates outputs a complete dynamic displacement cloud map of the conductor. The system inverts the current equivalent tension of the conductor to 45kN, a decrease of 8% from the initial value. Combined with the displacement amplitude, it is determined that there is a moderate risk of galloping, and the fatigue damage index is accumulating rapidly. The system immediately generates a level-two warning, prompting maintenance personnel to strengthen monitoring and prepare to take anti-galloping measures. Simultaneously, based on the vibration spectrum characteristics, the system classifies the event as conductor galloping, ruling out the possibility of construction interference.

[0052] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0053] Example 2 Based on the distributed state monitoring and reconstruction method for overhead transmission line operation and maintenance protection proposed in Embodiment 1 of this invention, Embodiment 2 of this invention also proposes a distributed state monitoring and reconstruction system for overhead transmission line operation and maintenance protection. Figure 4 This is a schematic diagram of a distributed condition monitoring and reconfiguration system for the operation and maintenance protection of overhead transmission lines, as proposed in Embodiment 2 of the present invention. The system includes: Data acquisition unit: used to inject probe light signals into the optical fiber of overhead transmission line, acquire and demodulate the back Rayleigh scattering signals at each point along the optical fiber, and obtain the distributed strain data of the optical fiber. Data processing unit: Based on the feature points corresponding to the line structure in the distributed strain data, the spatial coordinates of the optical fiber are mapped to the geographical coordinates of the transmission line, and the geographical coordinates, spatial coordinates and distributed strain data are all normalized to form an input dataset. Model reconstruction unit: used to construct a neural network model with normalized geographic coordinates as input and conductor displacement as output; the control equations describing the dynamic behavior of the conductor are embedded as physical constraints into the loss function of the neural network model, and hard constraints are used to force the output of the neural network model to meet the displacement boundary conditions at the suspension point of the line. The neural network model with embedded physical constraints is trained using the input dataset to obtain the trained state reconstruction model. State inversion unit: used to input the spatiotemporal coordinates to be measured into the trained state reconstruction model to obtain the continuous displacement field and stress field of the transmission line over the entire span; and to invert the implicit physical parameters of the conductor by solving the inverse problem.

[0054] The data acquisition unit includes an ultra-narrow linewidth laser source, modulator, amplifier, circulator, balanced detector, and signal processing module.

[0055] The distributed status monitoring and reconstruction system for operation and maintenance protection of overhead transmission lines provided in Embodiment 2 of this application achieves long-distance, distributed, and high-precision sensing of conductor status by reusing the existing optical fiber composite overhead ground wire (OPGW) of the line as the sensing medium. Through intelligent algorithms embedded with physical laws, it achieves accurate reconstruction and health assessment of the physical status of the entire line, thereby improving the depth, accuracy, and reliability of transmission line status sensing.

[0056] The description of the relevant parts of the distributed state monitoring and reconstructing system for operation and maintenance protection of overhead transmission lines provided in Embodiment 2 of this application can be found in the detailed description of the corresponding parts in the distributed state monitoring and reconstructing method for operation and maintenance protection of overhead transmission lines provided in Embodiment 1 of this application, and will not be repeated here.

[0057] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that the elements inherent in a process, method, article, or apparatus that includes a list of elements are included. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. Additionally, portions of the technical solutions provided in the embodiments of this application that are consistent with the implementation principles of corresponding technical solutions in the prior art have not been described in detail to avoid excessive elaboration.

[0058] While specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art can make other modifications or variations based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A distributed condition monitoring and reconfiguration method for operation and maintenance protection of overhead transmission lines, characterized in that, Includes the following steps: By injecting probe light signals into the optical fibers of overhead power transmission lines, acquiring and demodulating the back Rayleigh scattering signals at various points along the optical fiber, distributed strain data of the optical fiber can be obtained. Based on the feature points corresponding to the line structure in the distributed strain data, the spatial coordinates of the optical fiber are mapped to the geographical coordinates of the transmission line, and the geographical coordinates, spatial coordinates and distributed strain data are all normalized to form the input dataset. A neural network model is constructed with normalized geographic coordinates as input and conductor displacement as output. The control equations describing the dynamic behavior of the conductor are embedded as physical constraints into the loss function of the neural network model. Hard constraints are used to force the output of the neural network model to satisfy the displacement boundary conditions at the suspension point of the line. The neural network model with embedded physical constraints is trained using the input dataset to obtain the trained state reconstruction model. By inputting the spatiotemporal coordinates to be measured into the trained state reconstruction model, the continuous displacement field and stress field of the transmission line over the entire span are obtained. The implicit physical parameters of the conductor are then deduced by solving the inverse problem.

2. The method according to claim 1, characterized in that, The method also includes assessing the structural health of the line based on the reconstructed displacement field, stress field, and inverted physical parameters. Specifically, based on the reconstructed displacement field and stress field, combined with the material's SN curve and linear cumulative damage theory, the fatigue damage index of the conductor is calculated and the remaining fatigue life is predicted.

3. The method according to claim 1, characterized in that, A probe light signal is injected into the optical fiber composite overhead ground wire of the overhead transmission line. The back Rayleigh scattering signals at various points along the optical fiber are acquired and demodulated to obtain the distributed strain data of the optical fiber composite overhead ground wire, specifically: Injecting a linearly swept-frequency chirped pulse optical signal into the fiber optic composite overhead ground wire; The backscattered Rayleigh signals at various points along the optical fiber are received and demodulated to obtain an interferometric intermediate frequency signal reflecting the strain change of the optical fiber. The intermediate frequency signal is processed to obtain the optical time delay variation distributed along the optical fiber; Based on the fiber optic elastic-optical effect, the optical time delay variation is converted into distributed strain data.

4. The method according to claim 3, characterized in that, Based on the fiber optic-elastic effect, the optical time delay variation is converted into distributed strain data, specifically as follows: ； in, For dynamic response; The refractive index; The elastic modulus of the optical fiber; The center frequency; This represents the pulse width.

5. The method according to claim 1, characterized in that, The governing equations describing the dynamic behavior of the conductor are as follows: ； in, Mass per unit length of the conductor; For the conductor in position time lateral displacement; The spatial coordinates are along the arc length of the conductor; The damping coefficient; For bending stiffness; It is static tension; The external load per unit length of conductor; This is the loss function.

6. The method according to claim 1, characterized in that, Hard constraints are used to force the neural network model output to meet the displacement boundary conditions at the suspension point of the line. Specifically, this is achieved by embedding a distance function in the output layer architecture of the neural network model, ensuring that the network meets the displacement boundary conditions at the suspension point. and At that point, the displacement of the conductor must be 0, that is... and ;in This refers to the conductor span length.

7. The method according to claim 1, characterized in that, During the training of the neural network model with embedded physical constraints using the input dataset, an adaptive weighting algorithm based on gradient magnitude is used to dynamically adjust the balance between the data fitting loss term and the physical law loss term.

8. The method according to claim 1, characterized in that, The implicit physical parameters include at least one of the following: effective conductor tension, equivalent elastic modulus, and equivalent icing load.

9. The method according to claim 2, characterized in that, The method also includes analyzing the time-frequency characteristics and spatial distribution patterns of distributed strain data, and using classification algorithms to identify at least one event type among conductor galloping, wind vibration, ice shedding impact, and external construction disturbance.

10. A distributed condition monitoring and reconfiguration system for the operation and maintenance protection of overhead transmission lines, characterized in that, include: Data acquisition unit: used to inject probe light signals into the optical fiber of overhead transmission line, acquire and demodulate the back Rayleigh scattering signals at each point along the optical fiber, and obtain the distributed strain data of the optical fiber. Data processing unit: Based on the feature points corresponding to the line structure in the distributed strain data, the spatial coordinates of the optical fiber are mapped to the geographical coordinates of the transmission line, and the geographical coordinates, spatial coordinates and distributed strain data are all normalized to form an input dataset. Model reconstruction unit: used to construct a neural network model with normalized geographic coordinates as input and conductor displacement as output; the control equations describing the dynamic behavior of the conductor are embedded as physical constraints into the loss function of the neural network model, and hard constraints are used to force the output of the neural network model to meet the displacement boundary conditions at the suspension point of the line. The neural network model with embedded physical constraints is trained using the input dataset to obtain the trained state reconstruction model. State inversion unit: used to input the spatiotemporal coordinates to be measured into the trained state reconstruction model to obtain the continuous displacement field and stress field of the transmission line over the entire span. The implicit physical parameters of the conductor are then deduced by solving the inverse problem.

Images