Integrated detection method and system for preventing icing and melting ice of power transmission line
By employing a multi-dimensional data acquisition and fusion strategy and optimizing the random forest algorithm, the de-icing parameters are dynamically adjusted, solving the problems of long response time and high energy consumption in the de-icing technology for transmission lines. This achieves efficient and precise automated control, adapts to complex meteorological conditions, and improves the reliability of power grid supply and the safety of power lines.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- XINGAN ELECTRIC POWER CO OF STATE GRID EAST INNER MONGOLIA ELECTRIC POWER CO LTD
- Filing Date
- 2025-12-23
- Publication Date
- 2026-04-28
AI Technical Summary
Existing anti-icing and de-icing technologies for transmission lines lack real-time linkage mechanisms, have limited detection dimensions, weak anti-interference capabilities, and lack adaptive adjustment of de-icing strategies, resulting in long response times, high energy consumption, and difficulty in adapting to complex weather conditions.
By employing a multi-dimensional data acquisition and fusion strategy, combined with an optimized random forest algorithm, the de-icing parameters are dynamically adjusted to achieve fully automated control of the entire process, including real-time monitoring of parameters such as ice thickness, density, and line tension, as well as adaptive adjustment of de-icing power.
It has achieved full automation of the transmission line anti-icing and de-icing process, shortened response time, improved the accuracy of icing identification and anti-interference ability, reduced energy consumption, adapted to complex weather conditions, and improved the reliability of power grid supply and line safety.
Smart Images
Figure CN121939293A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power transmission line safety protection technology, specifically to an integrated detection method and system for preventing icing and melting on power transmission lines. Background Technology
[0002] The safe and stable operation of transmission lines is directly related to the reliability of power supply from the power grid. Icing is a major disaster that frequently occurs on transmission lines in winter. As power grid construction continues to extend to high-altitude and frigid regions, the icing patterns are becoming increasingly complex, such as mixed frost and wet snow icing. This places more stringent demands on the accuracy, real-time performance, and energy efficiency of anti-icing and de-icing technologies.
[0003] In existing technologies, icing detection for transmission lines largely relies on monitoring single physical quantities such as ice thickness and ambient temperature, while de-icing technology operates with fixed parameters like constant power heating and timed de-icing. The two lack an efficient real-time linkage mechanism. Although some technologies have attempted to achieve a preliminary integration of detection and de-icing, insurmountable technical problems remain, including: a single detection dimension, failing to fully consider the direct impact of key parameters such as ice density and line tension on de-icing effectiveness; simple signal processing algorithms with weak anti-interference capabilities in complex environments; a lack of adaptive adjustment capabilities in de-icing strategies, making it difficult to balance energy consumption and de-icing efficiency; and the absence of a closed-loop optimization mechanism, preventing dynamic iterative optimization of strategies based on different icing scenarios. These problems make existing technologies ill-suited for the refined protection needs under complex meteorological conditions, severely hindering the development of anti-icing and de-icing technologies for transmission lines.
[0004] Furthermore, existing technologies in the field of anti-icing and de-icing of transmission lines have structural defects. On the one hand, the detection module and the de-icing module are usually independent of each other, and the detection data needs to be manually analyzed and judged before the de-icing operation can be started. The response time from detection to execution is long, which can easily lead to continuous accumulation of ice and cause line faults. On the other hand, existing detection technologies mostly rely on a single sensor or a simple combination of parameters, and do not fully correlate the relationship between ice density, line tension and ice state. At present, fixed power and fixed duration heating modes are generally used, and the operating parameters are not dynamically adjusted according to the actual state such as ice thickness, density and type.
[0005] In summary, current anti-icing and de-icing detection technologies cannot meet the needs of refined power grid operation and maintenance, high-reliability power supply, low-energy operation, and applicability to complex meteorological conditions. There is an urgent need to develop an integrated detection method for anti-icing and de-icing of transmission lines. Summary of the Invention
[0006] This invention aims to provide an integrated detection method and system for anti-icing and de-icing of transmission lines, realizing full-process automation from icing monitoring to de-icing execution, and is applicable to anti-icing and de-icing operations of high-voltage / ultra-high-voltage transmission lines under complex meteorological conditions.
[0007] To address the aforementioned technical issues, the present invention employs the following specific solution: an integrated detection method for ice melting and anti-icing of transmission lines, comprising: collecting multi-dimensional data of the transmission line, including ice thickness, ice density, line tension, and environmental parameters; using a weighted average fusion strategy to integrate the multi-dimensional data to obtain a fused feature vector; inputting the fused feature vector into an optimized random forest algorithm to output the ice level and ice type; adaptively executing ice melting based on the ice level and ice type, with the ice melting parameters dynamically adjusted based on a thickness-density-power mapping relationship library, and dynamically adjusting the ice melting power and duration; monitoring changes in ice thickness in real time during the ice melting process, and stopping ice melting when the ice thickness drops below a safe threshold.
[0008] As a further optimization of the above technical solution, the weighting coefficients for integrating multi-dimensional data using a weighted average fusion strategy are determined by the entropy weight method.
[0009] As a further optimization of the above technical solution, the optimized random forest algorithm includes 50 decision trees, a maximum tree depth of 10 layers, and a minimum number of sample splits of 5. The output icing levels include: no icing: thickness < 2 mm; light icing: 2 mm ≤ thickness < 8 mm; moderate icing: 8 mm ≤ thickness < 15 mm; heavy icing: thickness ≥ 15 mm. The algorithm outputs icing types including rime, hoarfrost, and mixed rime.
[0010] As a further optimization of the above technical solution, the ice-melting parameters are dynamically adjusted based on a thickness-density-power mapping library, specifically for: light icing and / or a density of 0.8-1.0 g / cm³. 3 At that time, the power is 500-800W / m, and the duration is equal to the thickness × 3 minutes / mm; moderate icing and / or density is 1.0-1.1g / cm³. 3 Time: Power 800-1200W / m, duration = thickness × 4 minutes / mm; heavy icing and / or density 1.1-1.2g / cm³ 3 Time: Power 1200-1500W / m, Duration = Thickness × 5 minutes / mm.
[0011] As a further optimization of the above technical solution, when performing adaptive de-icing, segmented de-icing is adopted, with adjacent segments starting at intervals, and the surface temperature of the transmission line is monitored in real time during the de-icing process.
[0012] As a further optimization of the above technical solution, after the ice melting stops, the ice melting power, segmentation order and ice melting rate, energy consumption and transmission line temperature changes are recorded. The decision tree parameters of the random forest algorithm and the weight coefficients of the weighted average fusion strategy are optimized by the gradient descent algorithm, and the thickness-density-power mapping relationship library is updated.
[0013] As a further optimization of the above technical solution, the safety threshold is an ice thickness of less than 2 mm.
[0014] An integrated detection system for ice removal and de-icing of transmission lines includes: a multi-dimensional data acquisition module for collecting multi-dimensional data of the transmission line, including ice thickness, ice density, line tension, and environmental parameters; an intelligent decision-making module for integrating multi-dimensional data using a weighted average fusion strategy to obtain a fused feature vector, inputting the fused feature vector into an optimized random forest algorithm, and outputting the ice level and ice type; and an adaptive de-icing module for adaptively performing de-icing based on the ice level and ice type, dynamically adjusting the de-icing power and duration based on a thickness-density-power mapping relationship library, monitoring changes in ice thickness in real time during the de-icing process, and stopping de-icing when the ice thickness drops below a safe threshold.
[0015] As a further optimization of the above technical solution, the ice thickness data is collected by an ice thickness sensor, which is installed on the windward side of the conductor; the ice density data is collected by an ice density sensor, which is installed in pairs with the ice thickness sensor; the line tension data is collected by a line tension sensor, which is installed at the conductor suspension point on the crossarm of the tower; and the environmental parameter data is collected by an environmental sensor, which integrates temperature, humidity, wind speed, and visibility parameters and is installed in the unobstructed area at the top of the tower.
[0016] As a further optimization of the above technical solution, the ice-melting element of the adaptive ice-melting module adopts a segmented PTC electric heating element with a single segment length of 1-3m and a maximum surface temperature of ≤60℃. This ice-melting element is wrapped around the outside of the conductor.
[0017] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0018] This invention is the first to link the multi-dimensional data acquisition module of transmission lines with the ice melting execution module and the intelligent decision-making module in real time, realizing full-process automation from ice monitoring to ice melting execution, significantly shortening the response time and solving the response lag problem caused by functional separation.
[0019] This invention integrates multi-dimensional data such as thickness, density, tension, temperature and humidity through a weighted average fusion strategy, and combines it with an optimized random forest algorithm to achieve accurate identification of icing conditions such as icing thickness, density and type, and significantly improves anti-interference ability.
[0020] This invention dynamically adjusts the melting power and duration based on two parameters: ice thickness and density. It constructs a thickness-density-power mapping library to achieve on-demand melting, reducing energy consumption while avoiding localized overheating damage to conductors. By collecting ice data in real time during the melting process, it utilizes a gradient descent algorithm to optimize the mapping relationship between sensor weight coefficients and melting parameters, enabling self-iterative strategy upgrades to adapt to icing scenarios in different regions and under different meteorological conditions. Attached Figure Description
[0021] Figure 1 This is a schematic diagram of the device composition of the present invention;
[0022] Figure 2 This is a process flow diagram of the present invention;
[0023] Reference numerals: 1. Ice thickness sensor, 2. Ice density sensor, 3. Line tension sensor, 4. Environmental sensor, 5. Intelligent decision-making module, 6. Ice melting element, 7. Hybrid power supply module, 8. Wireless communication module, 9. Transmission line, 10. Pole tower, 11. Battery management system. Detailed Implementation
[0024] The technical solution of the present invention will be further described in detail below with reference to specific embodiments. Parts not described or disclosed in detail in the following embodiments of the present invention should be understood as prior art known or should be known by those skilled in the art, such as ice thickness sensors, ice density sensors, line tension sensors, environmental sensors, etc.
[0025] This invention discloses an integrated detection method for preventing icing and melting of transmission lines, comprising the following steps:
[0026] Collect multi-dimensional data of transmission lines, including ice thickness, ice density, line tension, and environmental parameters;
[0027] A weighted average fusion strategy is adopted to integrate multi-dimensional data to obtain a fused feature vector. The weight coefficients of the weighted average fusion strategy are determined by the entropy weight method, where the weight of icing thickness is 0.4, the weight of icing density is 0.3, the weight of line tension is 0.2, and the weight of environmental parameters is 0.1.
[0028] The fused feature vectors are input into the optimized random forest algorithm, which outputs icing level and icing type. The algorithm then adaptively executes the icing melting procedure based on these levels and types. The icing melting parameters are dynamically adjusted based on a thickness-density-power mapping library, controlling the icing melting power and duration. Specifically, the optimized random forest algorithm consists of 50 decision trees, with a maximum tree depth of 10 layers and a minimum number of sample splits of 5. The training set contains 1000 sets of experimental data on icing thickness, density, and type in different icing scenarios. The validation set accuracy is ≥98%. The algorithm outputs icing levels as follows: no icing: thickness <2mm; light icing: 2mm ≤ thickness <8mm; moderate icing: 8mm ≤ thickness <15mm; heavy icing: thickness ≥15mm. The algorithm outputs icing types as follows: rime, hoarfrost, and mixed rime.
[0029] The ice-melting parameters are based on a thickness-density-power mapping library, and the ice-melting power and duration are dynamically adjusted as follows:
[0030] Lightly iced and / or with a density of 0.8-1.0 g / cm³ 3 At that time, the power is 500-800W / m, and the duration is equal to the thickness × 3 minutes / mm;
[0031] Moderate icing and / or density 1.0-1.1 g / cm³ 3 Time: Power 800-1200W / m, Duration = Thickness × 4 minutes / mm;
[0032] Heavy icing and / or density 1.1-1.2 g / cm³ 3 Time: Power 1200-1500W / m, Duration = Thickness × 5 minutes / mm.
[0033] When adaptively executing the de-icing procedure, segmented de-icing is adopted, with a 30-second interval between the start of adjacent segments. During the de-icing process, the surface temperature of the transmission line is monitored in real time to ensure that it is ≤60℃.
[0034] The ice-melting process monitors the ice thickness in real time. The ice-melting procedure stops when the ice thickness drops below a safety threshold of less than 2 mm.
[0035] After the ice melting process stops, the ice melting power, segmentation order and ice melting rate, energy consumption and transmission line temperature changes are recorded. The decision tree parameters of the random forest algorithm and the weight coefficients of the weighted average fusion strategy are optimized by the gradient descent algorithm, and the thickness-density-power mapping relationship library is updated to achieve self-iteration.
[0036] The process of optimizing the decision tree parameters of the random forest algorithm and the weight coefficients of the weighted average fusion strategy using gradient descent, and updating the thickness-density-power mapping relationship library, to achieve strategy self-iteration is as follows: First, a comprehensive loss function is constructed, which integrates the classification error of ice cover state recognition, the regression error of ice melting power prediction, and the deviation of ice melting energy consumption efficiency. Using this loss function as the optimization objective, the system uses gradient descent to iteratively optimize the number of decision trees, maximum depth, and node splitting threshold of the random forest classifier, as well as the weight coefficients of each data source in the multi-sensor weighted average fusion strategy, until the loss function converges, thereby obtaining the current optimal recognition model and fusion weights.
[0037] Subsequently, the system dynamically updates the thickness-density-power mapping knowledge base based on recently accumulated high-confidence ice-melting process data. This process uses ice thickness and density as input features and optimal ice-melting power as the output target, training a multinomial regression model using gradient descent to fit the nonlinear relationship between the three factors. After training, the system uses this updated regression model to replace the original mapping relationships in the knowledge base and discards old samples with excessive errors from historical data.
[0038] like Figure 1 As shown, the present invention also discloses an integrated detection system for preventing icing and melting ice on transmission lines, comprising:
[0039] The multi-dimensional data acquisition module is used to collect multi-dimensional data of transmission lines, including ice thickness, ice density, line tension, and environmental parameters.
[0040] The intelligent decision-making module is used to integrate multi-dimensional data to obtain a fused feature vector by adopting a weighted average fusion strategy. The fused feature vector is then input into an optimized random forest algorithm to output the icing level and icing type.
[0041] The adaptive de-icing module is used to adaptively perform de-icing according to the icing level and icing type. The de-icing parameters are based on the thickness-density-power mapping relationship library, and the de-icing power and duration are dynamically adjusted. During the de-icing process, the changes in icing thickness are monitored in real time. When the icing thickness drops below the safety threshold, the de-icing stops.
[0042] Specifically, the icing thickness data is collected by an icing thickness sensor 1. The icing thickness sensor 1 is a capacitive sensor, installed at a 30° angle to the conductor axis on the windward side of the conductor, and fixed with an insulating bracket to avoid electromagnetic interference. It should be noted that the windward side of the conductor refers to the surface of the transmission line conductor that first comes into contact with and accumulates supercooled water droplets or snowflakes in the air under the influence of the prevailing local wind or the dominant wind direction during the icing period. In other embodiments of the invention, the icing thickness sensor 1 can be replaced with a laser ranging sensor, suitable for ultra-high voltage transmission lines with higher detection accuracy requirements.
[0043] The ice density data is collected by ice density sensor 2, which innovatively adopts an ultrasonic transmission method and is installed in pairs with ice thickness sensor 1 at a distance of 5cm.
[0044] The line tension data is collected by the line tension sensor 3, which is a fiber optic tension sensor installed at the conductor suspension point on the crossarm of the tower to monitor the changes in line tension caused by icing in real time.
[0045] Environmental parameter data are collected by environmental sensor 4, which integrates parameters such as temperature, humidity, wind speed, visibility, etc., and is installed on the top of the tower in an unobstructed area.
[0046] The intelligent decision-making module 5 employs a weighted average fusion strategy to integrate multi-dimensional data to obtain a fused feature vector. This fused feature vector is then input into an optimized random forest algorithm to output the icing level and icing type. The hardware of the intelligent decision-making module 5 utilizes an STM32H7 series microcontroller and an FPGA (Field-Programmable Gate Array) to achieve parallel data processing and algorithm acceleration. The intelligent decision-making module 7's storage unit is configured with 16GBeMMC flash memory to store historical detection data, melting parameters, and optimized model parameters. In other embodiments of this invention, the machine learning algorithm can be replaced with a support vector machine or a neural network algorithm. In scenarios with limited sample data, the support vector machine exhibits better recognition accuracy.
[0047] The de-icing element 6 of the adaptive de-icing module uses a segmented PTC electric heating element, with a single segment length of 1-3m, a rated voltage of 220V, and a power adjustment range of 500-1500W / m. It has self-limiting temperature characteristics, with a maximum surface temperature ≤60℃ to prevent localized overheating. This de-icing element 6 is wrapped around the outside of the conductor and fixed with high-temperature resistant insulating tape. Alternatively, the de-icing element can be replaced with an electromagnetic induction de-icing element, suitable for special circuits where electric heating is not advisable, such as circuits near flammable or explosive areas.
[0048] In addition, this system also includes a hybrid power supply module 7 and a wireless communication module 8. The hybrid power supply module 7 includes a solar panel, an energy storage battery, and a backup lithium battery. The solar panel of the hybrid power supply module 7 is a monocrystalline silicon solar panel, which is installed above the crossarm of the tower. Its energy storage unit adopts a lithium iron phosphate battery pack + a backup lithium battery, and is equipped with a BMS battery management system 11 to realize charge and discharge protection and power balance.
[0049] Wireless communication module 8 is a dual-mode communication module that supports 4G / 5G and LoRa dual-mode communication. 4G / 5G is used for long-distance data transmission with unlimited communication distance; LoRa is used for short-distance networking between poles to achieve data backup and collaborative work.
[0050] Below, in conjunction with Figure 1 , 2 The process steps and technical principles of the integrated detection method and system for anti-icing and de-icing of transmission lines of the present invention are described below:
[0051] S1. Equipment Installation and Commissioning
[0052] S101. Fix each sensor in the above positions to ensure that the thickness and density sensor probes are in contact with the surface of the wires, and that the tension sensor is coaxially installed with the wire suspension point to avoid stress displacement.
[0053] S102. Send calibration instructions to the intelligent decision-making module through the wireless communication module, perform zero-point calibration on the sensor in the non-icing state, and test the communication stability between modules.
[0054] S2, Multi-dimensional Data Acquisition and Preprocessing
[0055] S201 The sensor module collects data at a fixed frequency, and the collected data is transmitted to the intelligent decision-making module in real time via the CAN bus.
[0056] S202. A 4th-order cutoff frequency 5Hz low-pass filter is used to remove high-frequency noise caused by environmental vibration and electromagnetic interference.
[0057] S203. Correct sensor system errors based on a linear calibration model. The linear calibration model is y = ax + b, where a is the calibration coefficient and b is the offset, which is obtained by fitting experimental data.
[0058] S204. Use the Lagrange interpolation algorithm to complete occasionally lost data points to ensure data integrity.
[0059] S3, Intelligent Icing Status Recognition and Early Warning
[0060] S301. A weighted average fusion strategy is adopted, and the weight coefficients are determined by the entropy weight method. Specifically, the weight coefficients are: thickness weight 0.4, density weight 0.3, tension weight 0.2, and environmental parameter weight 0.1. Multi-dimensional data are integrated to obtain a fused feature vector.
[0061] S302. Input the fused feature vector into the optimized random forest algorithm. The optimized random forest algorithm includes 50 decision trees, with a maximum tree depth of 10 layers and a minimum number of sample splits of 5. The training set contains 1000 sets of experimental data for different icing scenarios—thickness, density, and type. The validation set accuracy is ≥98%. The algorithm outputs the icing level and icing type. The icing levels include: no icing: thickness <2mm; light icing: 2mm ≤ thickness <8mm; moderate icing: 8mm ≤ thickness <15mm; heavy icing: thickness ≥15mm. The icing types include rime, hoarfrost, and mixed rime.
[0062] S303. When the icing level reaches light or above, the intelligent decision-making module sends an early warning message to the operation and maintenance center via 4G / 5G communication. The early warning message includes the icing location, level, density, and expected growth rate, and triggers the de-icing preparation command at the same time.
[0063] S4, Adaptive Ice Melting Execution
[0064] S401. Based on the "thickness-density-power" mapping library, dynamically adjust the ice-melting power and duration:
[0065] Lightly iced (density 0.8-1.0 g / cm³) 3 Power: 500-800W / m; Duration: Thickness × 3 minutes / mm;
[0066] Moderate icing (density 1.0-1.1 g / cm³) 3 Power: 800-1200W / m, Duration: Thickness × 4 minutes / mm;
[0067] Heavy icing (density 1.1-1.2 g / cm³) 3 Power 1200-1500W / m, duration = thickness × 5 minutes / mm.
[0068] S402 The ice-melting module operates in an interval start-up mode, with a 30-second interval between the start-ups of adjacent segments to avoid voltage fluctuations in the line caused by simultaneous heating; during the heating process, the surface temperature of the conductor is monitored in real time by a temperature sensor to ensure that it is ≤60℃.
[0069] S5, Ice Melting Effect Feedback and Strategy Optimization
[0070] S501 During the ice melting process, the ice thickness is monitored in real time. When the thickness drops below 2mm, the intelligent decision module issues a stop command and the ice melting module is powered off.
[0071] S502. Record the parameters and effects of this ice melting operation. The parameters include power, duration, and segmentation order. The effects include ice melting rate, energy consumption, and conductor temperature change. Optimize the decision tree parameters of the random forest algorithm and the weight coefficients of the multi-sensor fusion through the gradient descent algorithm. Update the thickness-density-power mapping relationship library to achieve strategy self-iteration.
[0072] The hybrid power supply management includes: prioritizing solar power supply, automatically switching to energy storage battery power supply when the output power of the solar panel decreases; activating backup lithium battery power supply when the energy storage battery voltage decreases; entering sleep mode when there is no icing; and automatically waking up to normal working mode after detecting icing.
[0073] It should be noted that, in this invention, integrating multi-dimensional data to obtain a fused feature vector refers to the technology of integrating, analyzing, and optimizing information collected by different types of sensors to obtain a more comprehensive and accurate state assessment result than that of a single sensor; adaptive ice melting control is a control strategy that dynamically adjusts the ice melting power, duration, and heating method based on real-time parameters such as ice thickness, density, and type; closed-loop feedback optimization is a mechanism that achieves self-iterative upgrading of the technical solution by real-time monitoring of the execution effect and reverse adjustment of decision parameters; energy consumption ratio is the ratio of the total electrical energy consumed in the ice melting process to the effective energy for melting the ice, with the effective energy calculated as ice mass × melting heat.
[0074] This invention significantly improves the accuracy of identifying ice thickness, density, and ice type by combining multi-sensor fusion with an optimized random forest algorithm, while reducing the false positive rate in complex environments such as rain, snow, and heavy fog. Its fully closed-loop architecture shortens the response time from detection to ice melting; the adaptive ice melting strategy effectively reduces the ice melting energy consumption ratio, achieving significant optimization of response speed and ice melting efficiency. The segmented heating and self-limiting temperature design avoids localized overheating of the conductors, keeping the conductor surface temperature below 60°C to ensure unaffected service life; it also reduces the incidence of ice-related failures, improves power grid reliability, and enhances the safety and stability of line operation. This invention possesses strong adaptability to extreme environments and power supply stability, supports fully automated operation without manual intervention, thereby reducing maintenance costs. Through a closed-loop feedback mechanism, the technical solution can automatically adapt to different regions such as high altitudes and plains, as well as different ice types such as rime, hoarfrost, and mixed rime, without requiring manual parameter adjustments; its self-optimization strategy enables full-scene adaptation.
[0075] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. An integrated detection method for anti-icing and de-icing of transmission lines, characterized in that, include: Collect multi-dimensional data of transmission lines, including ice thickness, ice density, line tension, and environmental parameters; A weighted average fusion strategy is used to integrate multi-dimensional data to obtain a fused feature vector; The fused feature vectors are input into the optimized random forest algorithm, which outputs the icing level and icing type. Ice melting is adaptively performed based on the icing level and icing type. The ice melting parameters are based on a thickness-density-power mapping relationship library, and the ice melting power and duration are dynamically adjusted. During the ice melting process, the change in ice thickness is monitored in real time, and ice melting stops when the ice thickness drops below the safety threshold.
2. The integrated detection method for anti-icing and de-icing of transmission lines according to claim 1, characterized in that, The weighting coefficients for integrating multi-dimensional data using a weighted average fusion strategy are determined by the entropy weighting method.
3. The integrated detection method for anti-icing and de-icing of transmission lines according to claim 1, characterized in that, The optimized random forest algorithm includes 50 decision trees, with a maximum tree depth of 10 layers and a minimum number of sample splits of 5. The output icing levels include: no icing (thickness < 2 mm); light icing (2 mm ≤ thickness < 8 mm); moderate icing (8 mm ≤ thickness < 15 mm); and heavy icing (thickness ≥ 15 mm). The algorithm outputs icing types including rime, hoarfrost, and mixed rime.
4. The integrated detection method for anti-icing and de-icing of transmission lines according to claim 3, characterized in that, The ice-melting parameters are based on a thickness-density-power mapping library, and the ice-melting power and duration are dynamically adjusted as follows: Lightly iced and / or with a density of 0.8-1.0 g / cm³ 3 At that time, the power is 500-800W / m, and the duration is equal to the thickness × 3 minutes / mm; moderate icing and / or density is 1.0-1.1g / cm³. 3 Time: Power 800-1200W / m, duration = thickness × 4 minutes / mm; heavy icing and / or density 1.1-1.2g / cm³ 3 Time: Power 1200-1500W / m, Duration = Thickness × 5 minutes / mm.
5. The integrated detection method for anti-icing and de-icing of transmission lines according to claim 4, characterized in that, When performing adaptive de-icing, segmented de-icing is adopted, with adjacent segments starting at intervals, and the surface temperature of the transmission line is monitored in real time during the de-icing process.
6. The integrated detection method for anti-icing and de-icing of transmission lines according to claim 5, characterized in that, After the ice melting stops, record the ice melting power, segmentation order and melting rate, energy consumption and transmission line temperature changes. Optimize the decision tree parameters of the random forest algorithm and the weight coefficients of the weighted average fusion strategy using the gradient descent algorithm, and update the thickness-density-power mapping relationship library.
7. The integrated detection method for anti-icing and de-icing of transmission lines according to claim 1, characterized in that, The safety threshold is an ice thickness of less than 2 mm.
8. An integrated detection system for preventing icing and melting ice on power transmission lines, characterized in that, include: The multi-dimensional data acquisition module is used to collect multi-dimensional data of transmission lines, including ice thickness, ice density, line tension, and environmental parameters. The intelligent decision-making module is used to integrate multi-dimensional data to obtain a fused feature vector by adopting a weighted average fusion strategy. The fused feature vector is then input into an optimized random forest algorithm to output the icing level and icing type. The adaptive de-icing module is used to adaptively perform de-icing according to the icing level and icing type. The de-icing parameters are based on the thickness-density-power mapping relationship library, and the de-icing power and duration are dynamically adjusted. During the de-icing process, the changes in icing thickness are monitored in real time. When the icing thickness drops below the safety threshold, the de-icing stops.
9. An integrated detection system for anti-icing and de-icing of transmission lines according to claim 8, characterized in that, The ice thickness data is collected by an ice thickness sensor, which is installed on the windward side of the conductor. The ice density data is collected through an ice density sensor, which is installed in pairs with an ice thickness sensor. The line tension data is collected by a line tension sensor, which is installed at the conductor suspension point on the crossarm of the tower. Environmental parameter data are collected through environmental sensors, which integrate temperature, humidity, wind speed, and visibility parameters and are installed on the unobstructed area at the top of the tower.
10. An integrated detection system for anti-icing and de-icing of transmission lines according to claim 8, characterized in that, The de-icing element of the adaptive de-icing module adopts a segmented PTC electric heating element with a single segment length of 1-3m and a maximum surface temperature of ≤60℃. The de-icing element is wrapped around the outside of the conductor.