Power transmission and transformation hoisting adaptive safety distance calculation method and system

Abstract

The application discloses a power transmission and transformation hoisting adaptive safety distance calculation method and system, and relates to the technical field of power engineering construction safety. The method comprises the following steps: acquiring multi-source sensing data containing real-time meteorological data, electric field intensity, voltage grade and the pose of hoisted objects, and preprocessing; extracting environmental feature vectors and electrical feature vectors for multi-modal fusion; calculating an insulation coordination correction coefficient and inputting a physics-data fusion prediction model to obtain a dynamic safety distance threshold value, and generating a first safety distance strategy; acquiring hoisting working condition data to calculate a working condition risk coefficient and secondarily optimize to obtain an optimization score; and correcting the dynamic threshold value according to the optimization score to obtain a second safety distance strategy containing upper and lower limits. The application realizes adaptive dynamic adjustment of the safety distance according to the electrical environment and the hoisting working condition, and improves the safety control precision of power transmission and transformation hoisting operations.

Description

Technical Field

[0001] This invention relates to the field of power engineering construction safety technology, and in particular to an adaptive safety distance calculation method and system for power transmission and transformation hoisting. Background Technology

[0002] Power transmission and transformation hoisting operations have long faced challenges from complex weather and working conditions. Existing safety management systems mainly rely on fixed geometric thresholds or human experience. Although multi-dimensional sensing and digital twin technologies have been introduced, safety distance thresholds are usually set statically, without considering the coupling relationship between electrical insulation parameters and real-time weather conditions, nor are they dynamically corrected based on working conditions such as hoisting weight and boom angle. This leads to frequent false alarms and missed alarms under severe weather or extreme working conditions, making it difficult to balance operational safety and construction efficiency.

[0003] Furthermore, existing pure physics engine models are unable to compensate for prediction errors caused by environmental disturbances, and pure data-driven models lack physical interpretability; both have limitations when applied alone. At the same time, existing solutions have not established a hierarchical collaborative optimization mechanism between electrical insulation constraints and mechanical operating condition risks. The calculation of safe distances is disconnected from the actual hoisting conditions, and the safety margin cannot be adaptively adjusted according to the real-time status, which restricts the improvement of the level of intelligent safety management and control of power transmission and transformation hoisting operations. Summary of the Invention

[0004] The purpose of this invention is to provide an adaptive safety distance calculation method and system for power transmission and transformation hoisting. By integrating real-time meteorological parameters with the electrical insulation coordination relationship, the dynamic safety distance threshold is calculated, and secondary optimization is performed based on the risk coefficient of hoisting conditions. This solves the technical problems of static rigidity of safety distance threshold and lack of coordination between electrical constraints and mechanical conditions in the existing technology, and realizes the adaptive dynamic adjustment of safety distance according to the environment and working conditions.

[0005] To achieve the above objectives, this invention provides a method for calculating adaptive safety distances during power transmission and transformation hoisting, comprising the following steps: S1. Acquire multi-source sensing data from the hoisting operation site. The multi-source sensing data includes real-time meteorological data, electric field strength data, voltage level data, and hoisting position and posture data. Perform time synchronization, outlier removal, and normalization processing on the multi-source sensing data to construct a standardized environmental dataset. S2. Extract environmental feature vectors and electrical feature vectors based on standardized environmental datasets, perform multimodal fusion processing on the environmental feature vectors and electrical feature vectors, and construct fused feature vectors; S3. Calculate the insulation coordination correction coefficient based on the fused feature vector and the preset insulation coordination model. Input the insulation coordination correction coefficient and voltage level data into the pre-built physical-data fusion adaptive safety distance prediction model to obtain the dynamic safety distance threshold and generate the first safety distance strategy. S4. Obtain the current lifting condition data, including lifting weight, boom angle and wind speed disturbance. Calculate the condition risk coefficient based on the lifting condition data. Perform secondary optimization on the first safety distance strategy based on the condition risk coefficient to obtain the optimization score. S5. Based on the optimization score, the dynamic safety distance threshold is corrected according to the working conditions to obtain the second safety distance strategy, and the second safety distance strategy is dynamically adjusted according to the real-time working conditions.

[0006] Preferably, in S1, time synchronization is based on the BeiDou time output by the BeiDou real-time dynamic differential positioning module, and the timestamps of the multi-source sensing data are aligned. Outlier removal uses the 3σ criterion to remove sampled values ​​that exceed three times the standard deviation of the mean; normalization uses Min-Max normalization to map multi-source sensing data to a unified dimension interval.

[0007] Preferably, in S2, multimodal fusion processing is performed on the environmental feature vector and the electrical feature vector to construct the fused feature vector, including: We employ a weighted fusion method based on an attention mechanism, and use a learnable weight matrix to perform linear transformations on the environmental feature vector and the electrical feature vector respectively, and calculate the fusion weights based on the attention vector. Calculate attention weights: ; Calculate the fused feature vector: ; in, For environmental feature vectors, For electrical feature vectors, To fuse feature vectors, For attention weights, For attention vectors, , All are learnable weight matrices. This is for splicing operations.

[0008] Preferably, in S3, the preset insulation coordination model corrects the standard insulation distance based on real-time meteorological parameters, and the formula for calculating the insulation coordination correction coefficient is: ; in, This is the insulation coordination correction factor. , , , These are real-time temperature, real-time humidity, real-time air pressure, and real-time wind speed, respectively. , , , These are the standard temperature reference value, standard humidity reference value, standard air pressure reference value, and standard wind speed reference value, respectively. α, β, γ, and δ are all weighting coefficients.

[0009] Preferably, in S3, the physical-data fusion adaptive safety distance prediction model includes a physical insulation coordination sub-model and a data-driven compensation sub-model; The physical insulation coordination sub-model calculates the basic safety distance based on the standard insulation coordination relationship. The data-driven compensation sub-model uses the residual between the historical predicted distance and the actual safety distance as the training target to dynamically compensate for the output deviation of the physical insulation coordination sub-model and outputs the dynamic safety distance threshold.

[0010] Preferably, in S4, the operational risk coefficient is quantified based on the combined impact of lifting weight, boom angle, and wind speed disturbance on lifting stability. The calculation formula is as follows: ; Where R is the risk factor for the working condition, M is the lifting weight, and θ is the boom angle. This represents the wind speed disturbance. , , These are the reference values ​​for the baseline lifting weight, the baseline boom angle, and the baseline wind speed disturbance. , , All are pre-configured sensitivity indices; The optimization score is calculated based on the coupling relationship between the working condition risk coefficient and the dynamic safety distance threshold, and is used to characterize the adjustment range of the safety distance under the current working condition.

[0011] Preferably, in S5, the condition correction of the dynamic safety distance threshold based on the optimization score includes: The optimized score is coupled with the dynamic safety distance threshold to obtain the upper and lower limits of the adaptive safety distance range. When the optimized score exceeds the preset score threshold, the upper limit is expanded and the lower limit is tightened to enhance the safety margin. When the optimized score is lower than the preset score threshold, the adaptive safety distance range is contracted to improve work efficiency.

[0012] Preferably, an adaptive safety distance calculation system for power transmission and transformation hoisting includes: The data acquisition module is used to acquire multi-source sensing data from the hoisting operation site. The multi-source sensing data includes real-time meteorological data, electric field strength data, voltage level data, and hoisting position and posture data. The module performs time synchronization, outlier removal, and normalization on the multi-source sensing data to construct a standardized environmental dataset. The feature fusion module is used to extract environmental feature vectors and electrical feature vectors based on a standardized environmental dataset, perform multimodal fusion processing on the environmental feature vectors and electrical feature vectors, and construct a fused feature vector. The distance prediction module is used to calculate the insulation coordination correction coefficient based on the fused feature vector and the preset insulation coordination model. The insulation coordination correction coefficient and voltage level data are input into the pre-built physical-data fusion adaptive safe distance prediction model to obtain the dynamic safe distance threshold and generate the first safe distance strategy. The working condition optimization module is used to acquire the current lifting working condition data, including lifting weight, boom angle and wind speed disturbance. The working condition risk coefficient is calculated based on the lifting working condition data, and the first safety distance strategy is optimized a second time based on the working condition risk coefficient to obtain the optimization score. The strategy output module is used to correct the dynamic safety distance threshold based on the optimization score to obtain the second safety distance strategy, and to dynamically adjust the second safety distance strategy according to the real-time operating conditions.

[0013] The advantages and beneficial effects of this invention compared to the prior art are: 1. This invention calculates the insulation coordination correction coefficient by using real-time meteorological parameters and a preset insulation coordination model, so that the dynamic safety distance threshold can be adaptively adjusted according to environmental meteorological conditions, thereby improving the accuracy of electrical safety distance calculation under complex meteorological conditions.

[0014] 2. This invention employs a physical-data fusion prediction model, using a physical insulation coordination sub-model to provide interpretable basic distances and a data-driven compensation sub-model to correct environmental disturbance residuals, thus balancing the interpretability and accuracy of the prediction results.

[0015] 3. This invention performs secondary optimization of the first safety distance strategy by using the risk coefficient of hoisting conditions, and establishes a hierarchical coordination mechanism between hard constraints of electrical insulation and soft constraints of mechanical conditions, so as to achieve a dynamic balance between safety margin and operation efficiency.

[0016] 4. The present invention outputs an adaptive safety distance range including upper and lower limits, and dynamically adjusts it according to real-time operating conditions, breaking through the limitations of traditional fixed thresholds or single thresholds, and enhancing the flexibility and reliability of safety management.

[0017] 5. This invention constructs a fusion feature vector through multimodal fusion processing, comprehensively encoding environmental meteorological attributes and electrical insulation attributes, providing multidimensional perception support for safe distance prediction.

[0018] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0019] Figure 1This is a flowchart of an adaptive safety distance calculation method for power transmission and transformation hoisting in an embodiment of the present invention; Figure 2 This is a structural diagram of an adaptive safety distance calculation system for power transmission and transformation hoisting according to an embodiment of the present invention.

[0020] Figure 3 This is a schematic diagram comparing the safe distance strategy of this invention with that of traditional methods; Figure 4 This is a schematic diagram comparing the false alarm rate and missed alarm rate of an embodiment of the present invention with those of a traditional method. Detailed Implementation

[0021] In the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product is in use. They are used only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," and "connect" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0022] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0023] like Figure 1 As shown, this invention provides a method for calculating adaptive safety distances during power transmission and transformation hoisting, comprising the following steps: S1. Acquire multi-source sensing data from the hoisting operation site. The multi-source sensing data includes real-time meteorological data, electric field strength data, voltage level data, and hoisting position and posture data. Perform time synchronization, outlier removal, and normalization processing on the multi-source sensing data to construct a standardized environmental dataset. S2. Extract environmental feature vectors and electrical feature vectors based on standardized environmental datasets, perform multimodal fusion processing on the environmental feature vectors and electrical feature vectors, and construct fused feature vectors; S3. Calculate the insulation coordination correction coefficient based on the fused feature vector and the preset insulation coordination model. Input the insulation coordination correction coefficient and voltage level data into the pre-built physical-data fusion adaptive safety distance prediction model to obtain the dynamic safety distance threshold and generate the first safety distance strategy. S4. Obtain the current lifting condition data, including lifting weight, boom angle and wind speed disturbance. Calculate the condition risk coefficient based on the lifting condition data. Perform secondary optimization on the first safety distance strategy based on the condition risk coefficient to obtain the optimization score. S5. Based on the optimization score, the dynamic safety distance threshold is corrected according to the working conditions to obtain the second safety distance strategy, and the second safety distance strategy is dynamically adjusted according to the real-time working conditions.

[0024] In some embodiments, multi-source sensing data is collected by meteorological sensors, electric field strength sensors, voltage transformers, and BeiDou positioning terminals deployed at the hoisting operation site, covering three types of heterogeneous information: environmental meteorology, electrical environment, and spatial pose of the hoisted object. Time synchronization, outlier removal, and normalization are used to eliminate differences in time reference, dimensional scale, and sampling quality of the heterogeneous data, constructing a standardized environmental dataset with a unified format. Environmental feature vectors represent meteorological attributes such as temperature, humidity, air pressure, and wind speed, while electrical feature vectors represent electrical attributes such as electric field strength and voltage level. After multimodal fusion, the two form a fused feature vector that simultaneously possesses environmental sensing and electrical discrimination capabilities. A preset insulation coordination model, based on insulation coordination theory in high-voltage engineering, quantifies the influence of real-time meteorological parameters on air insulation strength as insulation coordination correction coefficients. These coefficients, along with voltage level data, are input into a physical-data fusion adaptive safe distance prediction model. The model outputs a dynamic safe distance threshold and generates a first safe distance strategy, which reflects the baseline safe distance that allows close proximity between the hoisted object and energized equipment under the current electrical environment conditions. Current lifting operation data includes lifting weight, boom angle, and wind speed disturbance, used to quantify the mechanical stability risks of the lifting operation. The operation risk coefficient quantifies the combined impact of the above mechanical factors on the probability of load swaying and collision, and based on this, the first safety distance strategy is optimized a second time to obtain an optimization score. Finally, the dynamic safety distance threshold is adaptively corrected according to the operation conditions based on the optimization score to obtain the second safety distance strategy. This strategy is output in the form of an adaptive safety distance range including an upper limit and a lower limit. The upper limit serves as a warning trigger line, and the lower limit serves as a danger warning line, which is dynamically adjusted according to real-time operation conditions to achieve hierarchical coordination between electrical insulation constraints and mechanical operation condition constraints.

[0025] Preferably, in S1, time synchronization is based on the BeiDou time output by the BeiDou real-time dynamic differential positioning module, and the timestamps of the multi-source sensing data are aligned. Outlier removal uses the 3σ criterion to remove sampled values ​​that exceed three times the standard deviation of the mean; normalization uses Min-Max normalization to map multi-source sensing data to a unified dimension interval.

[0026] In some embodiments, time synchronization is based on the BeiDou time output by the BeiDou real-time dynamic differential positioning module. BeiDou time is a high-precision unified time reference provided by the BeiDou satellite navigation system, with timing accuracy down to the nanosecond level. By aligning the timestamps of data streams from various sensors, meteorological, electrical, and pose data are ensured to be correlated in a unified time coordinate system, avoiding distortion in subsequent fusion calculations due to sampling timing misalignment. Outlier removal adopts the 3σ criterion, which is based on the normal distribution assumption. Data with a deviation from the mean exceeding three times the standard deviation is judged as outliers and removed, thereby filtering out outliers caused by sensor momentary failures, electromagnetic interference, or communication packet loss. Normalization processing adopts Min-Max normalization, a transformation operation that linearly maps the original data to a preset numerical range (such as [0,1]), used to eliminate the scale interference of dimensional differences between different physical quantities (such as temperature in degrees Celsius, electric field strength in kilovolts per meter, and voltage level in kilovolts) on feature fusion and model calculation, so that features of each dimension have equal numerical weights in subsequent neural networks and physical models.

[0027] Preferably, in S2, multimodal fusion processing is performed on the environmental feature vector and the electrical feature vector to construct the fused feature vector, including: We employ a weighted fusion method based on an attention mechanism, and use a learnable weight matrix to perform linear transformations on the environmental feature vector and the electrical feature vector respectively, and calculate the fusion weights based on the attention vector. Calculate attention weights: ; Calculate the fused feature vector: ; in, For environmental feature vectors, For electrical feature vectors, To fuse feature vectors, For attention weights, For attention vectors, , All are learnable weight matrices. This is for splicing operations.

[0028] In some embodiments, the weighted fusion based on the attention mechanism originates from the simulation of human visual attention. After linearly transforming the environmental feature vector and the electrical feature vector using a learnable weight matrix, the attention vector calculates the correlation weight between them. This allows the model to automatically determine the relative importance of environmental meteorological attributes and electrical insulation attributes in the current scene and allocate the fusion ratio accordingly. The learnable weight matrix is ​​a parameter matrix iteratively optimized through backpropagation during neural network training, used to map input features to a feature space suitable for attention calculation. The concatenation operation connects the two feature vectors end-to-end along the channel dimension, forming a higher-dimensional joint representation to enhance the attention mechanism's ability to perceive cross-modal feature interactions. The LeakyReLU activation function introduces a small slope on the negative half-axis to avoid permanent neuron deactivation; the softmax function normalizes the attention score to a probability distribution, ensuring that the sum of all weight coefficients is 1. The fused feature vector generated by the above fusion process simultaneously encodes environmental meteorological attributes and electrical insulation attributes, providing a unified and complete feature input for subsequent calculation of insulation matching correction coefficients.

[0029] Preferably, in S3, the preset insulation coordination model corrects the standard insulation distance based on real-time meteorological parameters, and the formula for calculating the insulation coordination correction coefficient is: ; in, This is the insulation coordination correction factor. , , , These are real-time temperature, real-time humidity, real-time air pressure, and real-time wind speed, respectively. , , , These are the standard temperature reference value, standard humidity reference value, standard air pressure reference value, and standard wind speed reference value, respectively. α, β, γ, and δ are all weighting coefficients.

[0030] In some embodiments, insulation coordination is a technical principle in power systems that coordinates the configuration of insulation strength based on equipment insulation level, operating voltage, and environmental conditions to prevent insulation breakdown. The insulation strength of an air gap is not constant but is significantly affected by meteorological parameters such as temperature, humidity, air pressure, and wind speed: increased temperature leads to decreased air density, and increased humidity causes water molecules to adhere to the electrode surface, both reducing insulation strength; decreased air pressure also weakens the dielectric properties of air; wind speed affects discharge characteristics by altering the local flow field. A preset insulation coordination model uses standard meteorological reference values ​​(standard temperature, standard humidity, standard air pressure, and standard wind speed) as a benchmark, and calculates the insulation coordination correction coefficient K using a weighted summation method. This coefficient quantifies the degree of deviation of actual meteorological conditions from standard conditions. When K is greater than 1, it indicates that the current meteorological conditions are worse than standard conditions, the air insulation strength is reduced, and the safety distance needs to be increased; when K is less than 1, it indicates that the insulation conditions are better than standard conditions, and the safety distance can be appropriately reduced. The weighting coefficients α, β, γ, and δ reflect the relative contribution ratio of each meteorological parameter to the insulation strength, and are determined by historical insulation test data or on-site calibration.

[0031] Preferably, in S3, the physical-data fusion adaptive safety distance prediction model includes a physical insulation coordination sub-model and a data-driven compensation sub-model; The physical insulation coordination sub-model calculates the basic safety distance based on the standard insulation coordination relationship. The data-driven compensation sub-model uses the residual between the historical predicted distance and the actual safety distance as the training target to dynamically compensate for the output deviation of the physical insulation coordination sub-model and outputs the dynamic safety distance threshold.

[0032] In some embodiments, the physical insulation coordination sub-model adopts a dual-path coupling architecture, consisting of a physical insulation coordination sub-model and a data-driven compensation sub-model. The physical insulation coordination sub-model is based on insulation coordination theory in high-voltage engineering standards (such as GB / T311.1 or IEC60071-1), calculating the basic safety distance using analytical formulas based on voltage level and insulation coordination correction coefficients. This output has clear physical interpretability and can reflect the hard constraints of electrical insulation. However, the physical model is based on idealized assumptions and cannot fully encompass complex disturbances such as local microclimates, electromagnetic field distortions, and equipment surface conditions at actual hoisting sites, leading to a systematic deviation between the predicted and actual safety distances. The data-driven compensation sub-model uses the residual between historical predicted distances and measured safety distances (i.e., the difference between predicted and actual values) as its training objective. It employs machine learning networks (such as fully connected neural networks or gradient boosting trees) to learn nonlinear disturbance patterns not captured by the physical model and dynamically compensates for the output deviation of the physical insulation coordination sub-model. The two sub-models work together to output a dynamic safety distance threshold, which not only preserves the interpretability of the physical mechanism, but also improves the prediction accuracy under complex working conditions through data-driven methods.

[0033] Preferably, in S4, the operational risk coefficient is quantified based on the combined impact of lifting weight, boom angle, and wind speed disturbance on lifting stability. The calculation formula is as follows: ; Where R is the risk factor for the working condition, M is the lifting weight, and θ is the boom angle. This represents the wind speed disturbance. , , These are the reference values ​​for the baseline lifting weight, the baseline boom angle, and the baseline wind speed disturbance. , , All are pre-configured sensitivity indices; The optimization score is calculated based on the coupling relationship between the working condition risk coefficient and the dynamic safety distance threshold, and is used to characterize the adjustment range of the safety distance under the current working condition.

[0034] In some embodiments, during hoisting operations, the greater the hoisting weight, the higher the inertia of the hoisted object and the higher the rope tension, making displacement under sudden disturbances more difficult to control; the closer the boom angle is to the horizontal limit, the smaller the structural stability margin and the weaker the wind resistance; the wind speed disturbance directly determines the excitation intensity of the hoisted object's swing. The operational risk coefficient quantifies the combined impact of the above three factors on hoisting stability using a power function, where the sensitivity index... , , These factors reflect the nonlinear sensitivity of lifting weight, boom angle, and wind speed disturbance to risk contribution: when the sensitivity index is greater than 1, the risk increases superlinearly with this factor; when the sensitivity index is less than 1, the risk increases sublinearly. Reference values ​​for the baseline lifting weight, boom angle, and wind speed disturbance are used to make each physical quantity dimensionless, allowing mechanical parameters of different dimensions to be coupled and calculated in the same formula. The optimization score is calculated based on the coupling relationship between the operating condition risk coefficient and the dynamic safety distance threshold, and is used to characterize the additional safety margin requirement of the current mechanical operating condition relative to the electrical reference safety distance: the higher the operating condition risk coefficient, the larger the optimization score, indicating that a greater mechanical safety margin needs to be added on top of the electrical insulation threshold.

[0035] Preferably, in S5, the condition correction of the dynamic safety distance threshold based on the optimization score includes: The optimized score is coupled with the dynamic safety distance threshold to obtain the upper and lower limits of the adaptive safety distance range. When the optimized score exceeds the preset score threshold, the upper limit is expanded and the lower limit is tightened to enhance the safety margin. When the optimized score is lower than the preset score threshold, the adaptive safety distance range is contracted to improve work efficiency.

[0036] In some embodiments, the preset score threshold is a critical judgment value that distinguishes between high-risk and low-risk working conditions, and is calibrated by the system based on historical lifting accident statistics or engineering experience. The adaptive safety distance range is defined by an upper limit and a lower limit: the upper limit is the warning trigger line; when the predicted distance between the hoisted object and the energized equipment approaches the upper limit, the system triggers a warning to alert the operator; the lower limit is the danger warning line; when the predicted distance falls below the lower limit, the system triggers an alarm and suggests immediate avoidance measures. When the optimization score exceeds the preset score threshold, it indicates that the current lifting condition is high-risk. In this case, expanding the upper limit can push the warning trigger line to a further distance, achieving earlier warning; at the same time, tightening the lower limit (i.e., increasing the lower limit value) pushes the danger warning line to a further distance, providing the operator with a greater safety margin and reaction time. When the optimization score is below the preset score threshold, it indicates that the current working condition is stable. In this case, shrinking the adaptive safety distance range (i.e., reducing the distance between the upper and lower limits) allows the hoisted object to operate in a relatively closer range to the energized equipment while ensuring safety, reducing unnecessary downtime and thus improving construction efficiency.

[0037] Preferred, such as Figure 2 As shown, an adaptive safety distance calculation system for power transmission and transformation hoisting includes: The data acquisition module is used to acquire multi-source sensing data from the hoisting operation site. The multi-source sensing data includes real-time meteorological data, electric field strength data, voltage level data, and hoisting position and posture data. The module performs time synchronization, outlier removal, and normalization on the multi-source sensing data to construct a standardized environmental dataset. The feature fusion module is used to extract environmental feature vectors and electrical feature vectors based on a standardized environmental dataset, perform multimodal fusion processing on the environmental feature vectors and electrical feature vectors, and construct a fused feature vector. The distance prediction module is used to calculate the insulation coordination correction coefficient based on the fused feature vector and the preset insulation coordination model. The insulation coordination correction coefficient and voltage level data are input into the pre-built physical-data fusion adaptive safe distance prediction model to obtain the dynamic safe distance threshold and generate the first safe distance strategy. The working condition optimization module is used to acquire the current lifting working condition data, including lifting weight, boom angle and wind speed disturbance. The working condition risk coefficient is calculated based on the lifting working condition data, and the first safety distance strategy is optimized a second time based on the working condition risk coefficient to obtain the optimization score. The strategy output module is used to correct the dynamic safety distance threshold based on the optimization score to obtain the second safety distance strategy, and to dynamically adjust the second safety distance strategy according to the real-time operating conditions.

[0038] In some embodiments, the data acquisition module performs data acquisition and preprocessing in S1, the feature fusion module performs feature extraction and multimodal fusion in S2, the distance prediction module performs insulation coordination correction and dynamic threshold prediction in S3, the operating condition optimization module performs operating condition risk calculation and secondary optimization in S4, and the strategy output module performs operating condition correction and strategy output in S5. Each module can be deployed on an edge computing node or a cloud server, and data interaction is achieved through industrial Ethernet or a 5G private network. The data acquisition module is typically deployed at the edge of the hoisting site, directly interfacing with meteorological sensors, electric field sensors, and BeiDou positioning terminals to reduce data transmission latency; the distance prediction module and the operating condition optimization module can be deployed on the edge computing unit to meet real-time requirements; the strategy output module is linked with the on-site 3D visualization interface and audible and visual alarm devices to achieve real-time display and alarm output of the safe distance strategy.

[0039] The following verification is based on a specific embodiment.

[0040] At the construction site of a 500kV substation expansion project, a main transformer needed to be hoisted into place. The work area was adjacent to a 500kV GIS equipment that was already energized. The multi-source sensing system deployed on site included: a weather station (to collect temperature, humidity, air pressure, and wind speed in real time), an electric field strength sensor (to monitor the electric field around the GIS equipment), a voltage transformer (to confirm the operating voltage level), and a Beidou real-time dynamic differential positioning terminal (to monitor the position and attitude of the hoisted object).

[0041] S1 Execution Process: The collected real-time meteorological data includes a temperature of 36℃, humidity of 88%, air pressure of 101.0 kPa, wind speed of 6 m / s, electric field strength of 11.5 kV / m, and voltage level of 500 kV. The position and attitude of the suspended object are output by the BeiDou positioning module with centimeter-level accuracy. The above data is timestamped using BeiDou time as a reference. Two sets of abnormal sampling values ​​caused by instantaneous electromagnetic interference from the sensors are removed using the 3σ criterion. After Min-Max normalization, a standardized environmental dataset is constructed.

[0042] S2 execution process: Extract a 256-dimensional environmental feature vector (encoding temperature, humidity, pressure, and wind attributes) and a 128-dimensional electrical feature vector (encoding electric field and voltage attributes) from the standardized environmental dataset. Multimodal fusion is performed through a two-layer graph attention network, and a 384-dimensional fused feature vector is generated after attention weight calculation.

[0043] S3 Execution Process: The meteorological parameters from the fused feature vector are input into the preset insulation coordination model. Standard reference values ​​are taken: T0=20℃, H0=50%, P0=101.3kPa, v0=2m / s, with weighting coefficients α=0.3, β=0.4, γ=0.2, and δ=0.1. The insulation coordination correction coefficient K=1.22 is calculated. The physical insulation coordination sub-model calculates a basic safety distance of 5.0m based on the 500kV voltage level and standard insulation coordination relationships. The data-driven compensation sub-model compensates the physical model output by 0.3m based on historical residual training results, ultimately outputting a dynamic safety distance threshold of 6.1m, generating the first safety distance strategy.

[0044] S4 execution process: The current lifting condition data is as follows: lifting weight 45t, boom angle 60°, wind speed disturbance 2.5m / s. Taking the reference values ​​M0=30t, θ0=45°, v_w0=1m / s, and sensitivity indices λ1=1.2, λ2=0.8, λ3=1.5, the calculated condition risk coefficient R=1.15. Based on the coupling relationship between the condition risk coefficient and the dynamic safety distance threshold of 6.1m, the optimization score is calculated to be 0.82.

[0045] S5 Execution Process: The preset score threshold is 0.75. Since the optimized score of 0.82 exceeds the preset threshold, it is determined to be a high-risk condition. An adaptive correction is applied to the dynamic safety distance threshold of 6.1m: the upper limit is expanded to 7.0m (early warning line), and the lower limit is tightened to 5.5m (danger warning line), generating a second safety distance strategy with an adaptive safety distance range of [5.5m, 7.0m]. The system monitors the distance between the suspended load and the GIS equipment in real time. When the distance approaches 7.0m, a level one warning is triggered; when the distance drops to 5.5m, a level two alarm is triggered, prompting immediate adjustment of the boom.

[0046] Comparison of quantitative results with traditional methods: such as Figures 3-4 As shown, in the same work scenario, the traditional fixed threshold control system of 5.5m failed to trigger an alarm when the load distance was 5.8m, resulting in a near-electricity hazard and a false alarm rate of 15%, because it did not consider the decrease in air insulation strength caused by high temperature and humidity. Furthermore, because it did not consider operational risks, it continued to operate with a conservative threshold of 5.5m under stable conditions, leading to unnecessary downtime and limited construction efficiency. After adopting the adaptive safety distance calculation method of this embodiment, the dynamic threshold increases to 6.1m according to weather conditions, and after operational condition correction, an adaptive range of [5.5m, 7.0m] is output. In actual operation, the false alarm rate is reduced to 2%, the false alarm rate is reduced to 4%, and the construction efficiency is improved by 20% compared to the fixed threshold scheme, achieving a dynamic balance between safety margin and operational efficiency.

[0047] In this application, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. In case of any inconsistency, the meaning set forth in this specification or derived from the content described herein shall prevail. Furthermore, the terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit the scope of this application.

[0048] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for calculating adaptive safety distance in power transmission and transformation hoisting, characterized in that, Includes the following steps: S1. Acquire multi-source sensing data from the hoisting operation site. The multi-source sensing data includes real-time meteorological data, electric field strength data, voltage level data, and hoisting position and posture data. Perform time synchronization, outlier removal, and normalization processing on the multi-source sensing data to construct a standardized environmental dataset. S2. Extract environmental feature vectors and electrical feature vectors based on standardized environmental datasets, perform multimodal fusion processing on the environmental feature vectors and electrical feature vectors, and construct fused feature vectors; S3. Calculate the insulation coordination correction coefficient based on the fused feature vector and the preset insulation coordination model. Input the insulation coordination correction coefficient and voltage level data into the pre-built physical-data fusion adaptive safety distance prediction model to obtain the dynamic safety distance threshold and generate the first safety distance strategy. S4. Obtain the current lifting condition data, including lifting weight, boom angle and wind speed disturbance. Calculate the condition risk coefficient based on the lifting condition data. Perform secondary optimization on the first safety distance strategy based on the condition risk coefficient to obtain the optimization score. S5. Based on the optimization score, the dynamic safety distance threshold is corrected according to the working conditions to obtain the second safety distance strategy, and the second safety distance strategy is dynamically adjusted according to the real-time working conditions.

2. The method for calculating adaptive safety distance in power transmission and transformation hoisting according to claim 1, characterized in that, In S1, time synchronization is based on the BeiDou time output by the BeiDou real-time dynamic differential positioning module, and the timestamps of the multi-source sensing data are aligned. Outlier removal uses the 3σ criterion, removing sampled values ​​that exceed three times the standard deviation of the mean; The normalization process uses Min-Max normalization to map multi-source sensing data to a unified dimension range.

3. The method for calculating adaptive safety distance in power transmission and transformation hoisting according to claim 1, characterized in that, In S2, multimodal fusion processing is performed on the environmental feature vector and the electrical feature vector to construct the fused feature vector, including: We employ a weighted fusion method based on an attention mechanism, and use a learnable weight matrix to perform linear transformations on the environmental feature vector and the electrical feature vector respectively, and calculate the fusion weights based on the attention vector. Calculate attention weights: ； Calculate the fused feature vector: ； in, For environmental feature vectors, For electrical feature vectors, To fuse feature vectors, For attention weights, For attention vectors, , All are learnable weight matrices. This is for splicing operations.

4. The method for calculating adaptive safety distance in power transmission and transformation hoisting according to claim 1, characterized in that, In S3, the preset insulation coordination model corrects the standard insulation distance based on real-time meteorological parameters. The formula for calculating the insulation coordination correction coefficient is as follows: ； in, This is the insulation coordination correction factor. , , , These are real-time temperature, real-time humidity, real-time air pressure, and real-time wind speed, respectively. , , , These are the standard temperature reference value, standard humidity reference value, standard air pressure reference value, and standard wind speed reference value, respectively. α, β, γ, and δ are all weighting coefficients.

5. The method for calculating adaptive safety distance in power transmission and transformation hoisting according to claim 4, characterized in that, In S3, the physical-data fusion adaptive safe distance prediction model includes a physical insulation coordination sub-model and a data-driven compensation sub-model. The physical insulation coordination sub-model calculates the basic safety distance based on the standard insulation coordination relationship. The data-driven compensation sub-model uses the residual between the historical predicted distance and the actual safety distance as the training target to dynamically compensate for the output deviation of the physical insulation coordination sub-model and outputs the dynamic safety distance threshold.

6. The method for calculating adaptive safety distance in power transmission and transformation hoisting according to claim 5, characterized in that, In S4, the risk factor for the operating condition is quantified based on the combined impact of lifting weight, boom angle, and wind speed disturbance on lifting stability. The calculation formula is as follows: ； Where R is the risk factor for the working condition, M is the lifting weight, and θ is the boom angle. This represents the wind speed disturbance. , , These are the reference values ​​for the baseline lifting weight, the baseline boom angle, and the baseline wind speed disturbance. , , All are pre-configured sensitivity indices; The optimization score is calculated based on the coupling relationship between the working condition risk coefficient and the dynamic safety distance threshold, and is used to characterize the adjustment range of the safety distance under the current working condition.

7. The method for calculating adaptive safety distance in power transmission and transformation hoisting according to claim 6, characterized in that, In S5, the working condition correction of the dynamic safety distance threshold based on the optimization score includes: The optimized score is coupled with the dynamic safety distance threshold to obtain the upper and lower limits of the adaptive safety distance range. When the optimized score exceeds the preset score threshold, the upper limit is expanded and the lower limit is tightened to enhance the safety margin. When the optimized score is lower than the preset score threshold, the adaptive safety distance range is contracted to improve work efficiency.

8. A power transmission and transformation hoisting adaptive safety distance calculation system, characterized in that, include: The data acquisition module is used to acquire multi-source sensing data from the hoisting operation site. The multi-source sensing data includes real-time meteorological data, electric field strength data, voltage level data, and hoisting position and posture data. The module performs time synchronization, outlier removal, and normalization on the multi-source sensing data to construct a standardized environmental dataset. The feature fusion module is used to extract environmental feature vectors and electrical feature vectors based on a standardized environmental dataset, perform multimodal fusion processing on the environmental feature vectors and electrical feature vectors, and construct a fused feature vector. The distance prediction module is used to calculate the insulation coordination correction coefficient based on the fused feature vector and the preset insulation coordination model. The insulation coordination correction coefficient and voltage level data are input into the pre-built physical-data fusion adaptive safe distance prediction model to obtain the dynamic safe distance threshold and generate the first safe distance strategy. The working condition optimization module is used to acquire the current lifting working condition data, including lifting weight, boom angle and wind speed disturbance. The working condition risk coefficient is calculated based on the lifting working condition data, and the first safety distance strategy is optimized a second time based on the working condition risk coefficient to obtain the optimization score. The strategy output module is used to correct the dynamic safety distance threshold based on the optimization score to obtain the second safety distance strategy, and to dynamically adjust the second safety distance strategy according to the real-time operating conditions.

Images