A power distribution line multi-stage ice melting method, system, device and medium based on fusibility index and uncertain band detection

Abstract

The application discloses a power distribution line multi-stage ice melting method, system, device and medium based on fusibility index and uncertain band detection, and belongs to the technical field of power distribution line operation, and comprises the following steps: collecting topologies and operation parameters of a target line, establishing a fusibility index calculation model, and outputting a frozen parameter set; determining preset parameters based on historical samples; loading the frozen parameter set and the preset parameters, calculating the current fusibility index, and generating a stage preliminary judgment result based on the fusibility index; according to the stage preliminary judgment result, performing low-power detection to confirm the final control stage and issuing ice melting control instructions, and applying stability constraints during stage switching; during the ice melting process, safety monitoring is performed and the ice melting process is ended when the process termination condition is met. In the presence of measurement noise and model uncertainty, the system is prevented from frequent dithering and mis-switching near the threshold, and stable and reliable multi-stage ice melting control is realized.

Description

Technical Field

[0001] This invention relates to the field of power distribution line operation and ice damage control technology, specifically to a multi-stage de-icing method, system, equipment and medium for power distribution lines based on meltability index and uncertainty zone detection. Background Technology

[0002] In cold and damp regions, icing on overhead power lines can easily lead to increased conductor sag, insulator flashover, abnormal stress on fittings, and even tripping and line breakage accidents.

[0003] Currently, de-icing in engineering mainly relies on experience to set the current or duration for de-icing. While this method is simple and easy to implement, it is difficult to adapt to environmental changes and dynamic fluctuations in energy supply boundaries, easily leading to over-melting, under-melting, or excessive energy consumption. Furthermore, most existing methods focus solely on increasing heat, lacking quantitative characterization of heat dissipation conditions and precipitation phase burdens, and failing to establish reproducible and calculable entry and exit criteria, resulting in significant differences in de-icing effects across different lines and seasons. In terms of control strategies, decisions are often made using a single or a few fixed thresholds, without fully considering measurement noise and the uncertainty of the model itself. This causes the system to frequently fluctuate and switch near the threshold, resulting in insufficient operational stability and potentially triggering safety boundaries such as voltage drops and current ramp-ups.

[0004] Overall, existing de-icing methods generally have the following limitations: threshold setting relies on manual experience, lacks unified data standards and traceable records, and is difficult to reuse across lines and seasons; there is a lack of quantitative absorption mechanisms for uncertainties in the threshold neighborhood, leading to frequent oscillations in state judgment within the gray zone; there is no hierarchical management and rhythm control for stages such as preheating, main de-icing, and insulation, and there is also a lack of stabilization constraints such as minimum dwell time and hysteresis, making it difficult to balance de-icing success rate and energy consumption economy; in addition, before issuing power or current commands, there is often no coordinated verification and dynamic adjustment with multiple operational safety boundaries such as the upper limit of available power, energy storage status, voltage drop limit, and current ramp rate limit, increasing the risk of system exceeding limits. Summary of the Invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by this invention is: how to avoid frequent jittering and erroneous switching of the system near the threshold in the presence of measurement noise and model uncertainty, so as to achieve stable and reliable multi-stage ice melting control.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a multi-stage de-icing method for power distribution lines based on the solubility index and uncertainty band detection, comprising,

[0008] Collect the topology and operating parameters of the target line, establish a melting index calculation model, calibrate and freeze it using the minimum objective function, and output the frozen parameter set;

[0009] After freezing is completed, the entry threshold, exit threshold and uncertain bandwidth required for stage division are determined based on historical samples and used as preset parameters;

[0010] Load the frozen parameter set and preset parameters, calculate the current fusibility index based on the real-time running parameters, and generate the preliminary judgment result based on the comparison relationship between the fusibility index and the entry threshold, exit threshold and uncertain bandwidth;

[0011] Based on the initial assessment results, a low-power detection was performed to confirm the final control stage and issue an ice-melting control command, while applying stability constraints during stage switching.

[0012] During the de-icing process, safety monitoring is performed, and the de-icing process is terminated when the termination conditions are met.

[0013] As a preferred embodiment of the multi-stage de-icing method for distribution lines based on the meltability index and uncertainty zone detection described in this invention, the method involves: collecting the topology and operating parameters of the target line, establishing a meltability index calculation model, calibrating and freezing the line using a minimum objective function, and outputting a set of frozen parameters, including...

[0014] The topology and equipment parameters of the target line, historical icing and handling records, operational monitoring data, and short-term meteorological data are collected to construct the main features and constraint signal vectors.

[0015] The main features include equivalent heat input characteristics, flow heat dissipation intensity characteristics, and precipitation phase burden indication characteristics;

[0016] The main features are normalized within the training window and scaled to the range of {0, 1}.

[0017] Based on the principal features and constraint signal vectors, a fusibility index calculation model is constructed through weighted combination to calculate the fusibility index, and the corresponding uncertainty can be calculated based on the fusibility index.

[0018] Using historical records of whether ice has been completely melted as training labels, a minimum objective function is constructed based on the weight factors, intercepts, and weight factors of each element in the constraint signal vector corresponding to the main features. The optimal parameter set is then solved, and the solved optimal parameter set is frozen, outputting the frozen parameter set.

[0019] As a preferred embodiment of the multi-stage de-icing method for power distribution lines based on the meltability index and uncertainty band detection described in this invention, wherein: after freezing is completed, the entry threshold, exit threshold, and uncertainty bandwidth required for stage division are determined based on historical samples, and these are preset parameters including...

[0020] Based on the solubility index of the historical samples, and taking the successfully melted samples as a benchmark, the entry threshold and exit threshold are optimized and determined according to the distribution of the corresponding solubility index.

[0021] For the entry threshold and exit threshold, the corresponding threshold neighborhood is determined in the historical samples, and a unified uncertainty bandwidth is determined based on the uncertainty of the samples in the threshold neighborhood.

[0022] Set the entry threshold, exit threshold, and uncertain bandwidth to preset parameters.

[0023] The current preferred solution uses an offline parameter setting process. The entry threshold, exit threshold, and uncertain bandwidth are objectively determined by historical samples under a unified evaluation standard and written into the configuration. This makes the source of the thresholds explainable and recalculated. The solution also uses four indicators to comprehensively select the best option, taking into account success rate, energy consumption, and operation frequency.

[0024] As a preferred embodiment of the multi-stage de-icing method for power distribution lines based on the solubility index and uncertainty zone detection described in this invention, the preliminary judgment results of the generation stage include:

[0025] The meltable index calculation model is initialized by loading the frozen parameter set, and the current meltable index is calculated based on the real-time collected operating parameters.

[0026] Based on the mutually exclusive segment division rules, the calculated current meltability index is compared with the entry threshold, exit threshold and uncertain bandwidth in the preset parameters to generate the preliminary judgment result of the stage.

[0027] If the current fusibility index falls within the uncertain bandwidth corresponding to the entry or exit threshold, an uncertain band marker is added to the preliminary judgment result, and it is retained as a result to be confirmed.

[0028] As a preferred embodiment of the multi-stage de-icing method for power distribution lines based on the meltability index and uncertain zone detection described in this invention, the step of performing a low-power detection confirmation of the final control stage based on the initial stage judgment results includes:

[0029] When the initial judgment result is a result pending confirmation, execute the low-power detection confirmation process;

[0030] Perform low-power detection and collect feedback data according to the detection amplitude and detection duration;

[0031] The fusibility index calculation model is based on the feedback data, and the current fusibility index is recalculated.

[0032] The recalculated solubility index is compared again with the entry threshold, exit threshold, and uncertain bandwidth to confirm the final control stage.

[0033] As a preferred embodiment of the multi-stage de-icing method for power distribution lines based on the meltability index and uncertainty band detection described in this invention, the step of issuing de-icing control commands and applying stability constraints during stage switching includes:

[0034] Based on the confirmed final control stage, a target power command or current command is generated, and the command is checked and trimmed according to the operating boundary before it is issued.

[0035] The verified and trimmed instructions are issued. The phase switching must meet the minimum dwell conditions and hysteresis conditions. The rate of change of instructions during the switching process is subject to limit constraints.

[0036] As a preferred embodiment of the multi-stage de-icing method for power distribution lines based on the meltability index and uncertainty zone detection described in this invention, the step of performing safety monitoring and ending the current de-icing process when the process termination conditions are met includes:

[0037] During the execution of the target power command or current command, the actual current, voltage, temperature and power status parameters of the system are continuously monitored; when any parameter exceeds the safety boundary, it is determined to be an out-of-limit event.

[0038] Once an out-of-limit event occurs, the minimum residence and hysteresis conditions will be ignored, and degraded control will be implemented.

[0039] Degradation control, in order of severity of exceeding limits, includes: trimming the current command amplitude to within the safety boundary, forcibly switching the control phase to the heat preservation phase, and executing a shutdown;

[0040] When the amount of residual ice on the line is determined to be below the safety threshold based on real-time data, or the system still cannot resume safe operation after downgrade control, or the cumulative running time or energy consumption reaches the preset upper limit, the current de-icing process will be terminated.

[0041] This invention provides a multi-stage de-icing system for power distribution lines based on the meltability index and uncertainty zone detection.

[0042] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a multi-stage de-icing system for power distribution lines based on the meltability index and uncertainty zone detection, comprising: a model construction module, a partitioning module, a calculation module, a confirmation control stage module, and an output module;

[0043] The model building module collects the topology and operating parameters of the target line, establishes a melting index calculation model, uses a minimum objective function for calibration and freezing, and outputs a frozen parameter set.

[0044] The segmentation module, after freezing, determines the entry threshold, exit threshold, and uncertain bandwidth required for stage segmentation based on historical samples, as preset parameters.

[0045] The calculation module loads the frozen parameter set and preset parameters, calculates the current fusibility index based on the real-time running parameters, and generates a preliminary judgment result based on the comparison relationship between the fusibility index and the entry threshold, exit threshold and uncertain bandwidth.

[0046] The confirmation control phase module performs low-power detection to confirm the final control phase and issues a melting control command based on the initial phase judgment result, and applies stability constraints when switching phases.

[0047] The output module performs safety monitoring during the ice melting process and terminates the ice melting process when the process termination conditions are met.

[0048] The present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, characterized in that the processor executes the computer program to implement the steps of the multi-stage de-icing method for power distribution lines based on the meltability index and uncertainty band detection.

[0049] The present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps of the multi-stage de-icing method for power distribution lines based on a meltability index and uncertainty band detection.

[0050] The beneficial effects of this invention are as follows: This invention pre-determines the criterion source through data-driven offline determination and versioning, ensuring that the entry or exit threshold and uncertain bandwidth have a consistent, interpretable, and recalculated standard, eliminating the need for repeated trial and error during commissioning; on the online side, only fixed parameters are compared, and if the signal falls into the gray area, it is first detected with low power before judgment, combined with minimum dwell time and hysteresis, which can absorb measurement noise and short-term disturbances in the threshold neighborhood, eliminating unnecessary stage switching of voltage drop without sacrificing the fusion success rate; before the instruction is generated, it is uniformly checked and trimmed according to available power, SOC, allowable voltage drop, and ramp rate, reducing the risk of out-of-bounds and protection malfunctions; during offline optimization, success rate and action frequency are constrained simultaneously, reducing energy consumption per unit length and time to reach the target while meeting the constraints, achieving a balance between energy efficiency and stability; the entire process is traced by parameter version and timestamp, allowing for traceability and continuous optimization in case of deviations. Attached Figure Description

[0051] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 This is a flowchart illustrating a multi-stage de-icing method for power distribution lines based on a meltability index and uncertainty band detection, as provided in one embodiment of the present invention.

[0053] Figure 2 This is a flowchart illustrating the initial judgment result of the generation stage of a multi-stage de-icing method for power distribution lines based on the meltability index and uncertainty zone detection, provided as an embodiment of the present invention.

[0054] Figure 3 This is a flowchart illustrating the final control stage of a multi-stage de-icing method for power distribution lines based on a meltability index and uncertainty zone detection, as provided in one embodiment of the present invention. Detailed Implementation

[0055] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0056] Example 1, referring to Figure 1 As one embodiment of the present invention, this embodiment provides a multi-stage de-icing method for power distribution lines based on the solubility index and uncertainty band detection, comprising:

[0057] Existing power distribution line de-icing solutions generally suffer from three main problems: First, start-stop thresholds are mostly set based on experience, lacking a unified data standard and record keeping, making them neither interpretable nor recalculated; second, the judgment logic is mostly based on single threshold decisions, without explicitly handling measurement noise and model errors, making it prone to fluctuations near the threshold, leading to frequent switching, reduced energy efficiency, and even triggering safety boundaries such as voltage drop or current ramp-up; third, the control side lacks phased constraints and rhythm management, with preheating, main de-icing, and insulation not being layered, lacking both gray zone detection-re-judgment closed loops and minimum dwell and hysteresis mechanisms, resulting in insufficient operational stability and traceability.

[0058] The aforementioned problems directly lead to engineering dilemmas such as over-financing, under-financing, high frequency of operations, increased risk of crossing boundaries, and difficulty in relocation across lines and seasons.

[0059] S1. Collect the topology and operating parameters of the target line, establish a melting index calculation model, calibrate and freeze it using the minimum objective function, and output the frozen parameter set.

[0060] S2. After freezing is completed, determine the entry threshold, exit threshold and uncertain bandwidth required for stage division based on historical samples, as preset parameters.

[0061] S3. Load the frozen parameter set and preset parameters, calculate the current fusibility index based on the real-time running parameters, and generate the preliminary judgment result based on the comparison relationship between the fusibility index and the entry threshold, exit threshold and uncertain bandwidth.

[0062] S4. Based on the initial assessment results, perform low-power detection to confirm the final control stage and issue the ice-melting control command, and apply stability constraints during stage switching.

[0063] S5. During the de-icing process, implement safety monitoring and terminate the de-icing process when the process termination conditions are met.

[0064] The sources of the threshold and bandwidth in this invention are objective, explainable, and recalculated, significantly reducing the difficulty of online tuning; the gray area adopts a detection and re-judgment closed loop and superimposes minimum dwell time and hysteresis to suppress threshold neighborhood jitter, reduce unnecessary stage switching, and maintain or improve the fusing success rate.

[0065] Before the instructions are issued, they are uniformly checked and trimmed according to the power supply and operation boundaries to avoid exceeding the limits and minimizing false triggers; offline optimization comprehensively constrains the success rate and frequency of actions, reducing energy consumption per unit length and shortening the time to meet the standards while meeting the constraints; the entire process is recorded according to parameter version and timestamp, which can be traced and continuously optimized.

[0066] Example 2, an embodiment of the present invention, provides a multi-stage de-icing method for power distribution lines based on the meltability index and uncertainty band detection, based on the previous embodiment, including:

[0067] S1. Collect the topology and operating parameters of the target line, establish a melting index calculation model, calibrate and freeze it using the minimum objective function, and output the frozen parameter set.

[0068] S11. Collect the topology and equipment parameters of the target line, historical icing and handling records, operation monitoring data and short-term meteorological data, and construct the main features and constraint signal vectors.

[0069] The main features, including equivalent heat input characteristics, flow heat dissipation intensity characteristics, and precipitation phase burden indication characteristics, are expressed as follows:

[0070] ,

[0071] ,

[0072] in, This represents the equivalent heat input characteristics; For current; It is a length of Time window averaging operator, It is a function of resistance with respect to temperature. For the temperature of the conductor, Reference resistor, Temperature coefficient; These are the current temperature and the reference temperature, respectively.

[0073] It should be noted that the temperature coefficient is determined based on the selected conductor material, and its value comes from the corresponding national standard or manufacturer's technical specifications.

[0074] ,

[0075] in, The heat dissipation intensity characteristics are as follows; The convective heat transfer coefficient per unit length under calm conditions is given. This represents the increase in heat transfer coefficient for every 1 m / s increase in wind speed. Wind speed; Temperature;

[0076] ,

[0077] in, This is an indicator of precipitation phase burden; take 1 if it is raining or snowing, otherwise take 0.

[0078] It should be further explained that The convective heat transfer coefficient per unit length under calm conditions is generally taken as 3W / (m·℃) to 6W / (m·℃). The heat transfer coefficient increment caused by each 1 m / s increase in wind speed is generally taken as 0.6 W / (m·℃) / (m / s) to 1.2 W / (m·℃) (m / s).

[0079] The constraint signal vector is constructed based on five types of constraint signals, and its expression is as follows:

[0080] ,

[0081] in, This is the power boundary signal vector during runtime. These represent the upper bound average of available AC power within the window, the upper bound average of available energy storage power within the window, and the average state of charge within the window, respectively. These are the upper limits for allowable voltage drop and current ramp-up rate, respectively. This is the transpose of a vector.

[0082] S12, to Normalization is performed within the training window, scaling it to the range {0, 1}.

[0083] Based on the main features and constraint signal vectors, a fusibility index calculation model is constructed through weighted combination. The expression for the fusibility index calculation model is as follows:

[0084] ,

[0085] ,

[0086] ,

[0087] ,

[0088] ,

[0089] in, The solubility index, The intercept is... They are respectively Weighting factors For linear weighting of the boundary signal vector, For Sigmoid normalization calculation, For the set of weighting factors of the energy supply boundary signal vector, These are the weighting factors for the upper bound of available AC power in the current window, the upper bound of available energy storage power in the current window, the average state of charge in the current window, the upper limit of allowable voltage drop, and the upper limit of current ramp rate, respectively. The Lth weight factor in the set of weight factors for the energy boundary signal vector. The index variable is the set of weight factors for the energy boundary signal vector. The first power supply boundary signal vector is the first power supply boundary signal vector. One parameter, For the index variable of the energy boundary signal vector, .

[0090] It should be further explained that: The equivalent heat input feature weighting factor is set to a positive value; the larger the value, the easier it is for the ice to melt. The heat dissipation intensity characteristic is negative; the larger the value, the less likely the ice is to melt. The negative value is used to indicate the phase burden of precipitation. When freezing rain or snow occurs, the value should be lowered.

[0091] The initial fixed starting value is set as follows: , , , The fluctuation range for each optimization is set to ±0.5.

[0092] The default value range is ±1.0. Within a fixed trial operation evaluation cycle (such as a shift, a day or a week, depending on the site selection and fixation), record the impact of each boundary signal on the decision-making, focusing on whether it frequently becomes a constraint or triggers protection or limiting events.

[0093] If a boundary signal becomes the dominant restriction or protection trigger multiple times during the evaluation period, the absolute value of its weight can be increased from 1.0 to 1.5 without changing its sign; if it continues to exhibit the same characteristic in subsequent evaluation periods, it can be further increased to 2.0. Each evaluation period can only increase the weight by one level.

[0094] If a certain boundary signal remains in a non-dominant state and has a weak impact on decision-making for a long period during the evaluation period, the absolute value of its weight can be reduced from 1.0 to 0.5 without changing its sign. The weight can be reduced by a maximum of one level per evaluation period.

[0095] The absolute values ​​of all boundary signal weights should be controlled within the range of 0.5–2.0; adjustments are limited to amplitude and do not change the sign. Each weight should be adjusted only once per evaluation period to avoid frequent fluctuations. If there are no significant changes over two consecutive evaluation periods, the weights should remain unchanged to ensure strategy stability. When there are substantial changes in equipment capabilities, operating strategies, or environmental conditions, a new evaluation should be conducted to determine whether the weights need to be returned to their initial settings or readjusted according to the rules.

[0096] S13. The corresponding uncertainty can be calculated based on the solubility index, and the expression is:

[0097] ,

[0098] in, The uncertainty is...

[0099] Real-time data collection to determine whether the ice has been completely melted. ,in, For time variables, for Residual ice volume at time The threshold for determining net asset value is... The result is used to determine whether the ice has been completely melted; 1 indicates complete melting, and 0 indicates otherwise.

[0100] It should be further explained that To determine the melting threshold, the value specified in the industry standard is adopted, and the value used in this invention is 2mm.

[0101] S14. Using historical data records of whether ice has been completely melted as training labels, construct a minimum objective function based on the weight factors, intercepts, and weight factors of each element in the constraint signal vector corresponding to the main features, solve for the optimal parameter set, freeze the solved optimal parameter set, and output the frozen parameter set.

[0102] Based on intercept, Weighting factors and We construct a minimum objective function using the weighting factors, solve for the optimal parameter set, freeze the optimal dataset, and output the frozen parameter set. The expression is:

[0103] ,

[0104] in, Let A be the total number of samples used in training. The set of weighting factors.

[0105] It should be further explained that the optimal parameter set defines a construction scheme, including intercept, Weighting factors and Weighting factors.

[0106] S2. After freezing is completed, determine the entry threshold, exit threshold and uncertain bandwidth required for stage division based on historical samples, as preset parameters.

[0107] S21. Based on the solubility index of historical samples, and taking the successfully melted samples as a benchmark, optimize and determine the entry threshold and exit threshold according to the distribution of the corresponding solubility index.

[0108] Based on the solubility index of historical samples, and taking successfully melted samples as a benchmark, a unique entry threshold and exit threshold are determined from candidate threshold combinations through multi-objective optimization according to the statistical distribution of their solubility index.

[0109] Further, the thawability index of historical samples is calculated using the frozen parameter set, and samples that have been successfully thawed are selected accordingly. Based on the distribution of the thawability index of successful samples, the candidate value ranges for the entry threshold and exit threshold are defined respectively.

[0110] Candidate values ​​are paired to form candidate threshold combinations that conform to physical logic. For each candidate threshold combination, the comprehensive performance on multiple performance indicators is evaluated through offline simulation.

[0111] Based on the preset bottom line constraints and sorting rules, the optimal combination is selected from all candidate combinations, and its corresponding threshold is determined as the system's entry threshold and exit threshold.

[0112] Multiple performance indicators include four indicators: energy consumption per unit length, average time to reach the target, success rate, and average frequency of actions.

[0113] For example: the candidate value range can be set based on percentiles. The candidate range for entering the threshold can be set between the 10th percentile (P10) and the 50th percentile (P50) of the exponential distribution of successful samples; the candidate range for exiting the threshold can be set between the 40th percentile (P40) and the 90th percentile (P90). Overlap between the two ranges in the P40 to P50 interval is allowed to ensure the completeness of the optimization process.

[0114] Discretize with a fixed step size (e.g., 0.05) to generate specific candidate values.

[0115] If the candidate value set obtained after discretization within the threshold range is {0.15, 0.20, 0.25, 0.30, 0.35, 0.40, 0.45, 0.50}, and the candidate value set obtained after discretization outside the threshold range is {0.40, 0.45, 0.50, 0.55, 0.60, 0.65, 0.70, 0.75, 0.80, 0.85, 0.90}.

[0116] Then, each entry candidate value in the two sets is paired with each exit candidate value. This step must impose a physical logical constraint that the entry threshold candidate value is less than the exit threshold candidate value to eliminate all invalid combinations. The pairing (0.45, 0.50) is valid, while the pairings (0.50, 0.45) or (0.45, 0.40) will be eliminated, ensuring that each combination to be evaluated conforms to the basic logic of first entry, then exit.

[0117] For each valid candidate combination, offline simulation can be used to statistically analyze its energy consumption per unit length, average time to achieve the target, success rate, and average frequency of actions.

[0118] Preset baseline constraints may include: a success rate of no less than 95% and an average action frequency of no more than 2 times / hour. The sorting rules can be set as follows: prioritize the combination with the lowest energy consumption per unit length; if the energy consumption difference between the best and second-best combinations is within a 5% tolerance range, then compare the average time to achieve the target and select the combination with the shorter time; if the difference in achievement time also does not exceed 5%, then further compare the average action frequency and select the combination with fewer switches; if the above comparisons still cannot distinguish between them, then adopt a conservative safety principle and select the combination with the smallest entry threshold value as the final parameter.

[0119] S22. For the entry threshold and exit threshold, determine the corresponding threshold neighborhood in the historical samples, and determine a unified uncertainty bandwidth based on the uncertainty of the samples in the threshold neighborhood.

[0120] For each determined entry and exit threshold, a corresponding threshold neighborhood is identified in the historical samples. Based on the uncertainty statistical characteristics of the samples within each threshold neighborhood, a unified uncertainty bandwidth is determined through conversion and constraint processing.

[0121] The threshold neighborhood is a set of samples whose absolute value of the difference between the current solubility index and the entry threshold or exit threshold does not exceed the preset neighborhood width.

[0122] Furthermore, for both the entry and exit threshold neighborhoods, a preset high quantile is calculated for the uncertainty values ​​of all samples within each neighborhood. This high quantile value is then multiplied by a preset conversion factor to amplify it and provide a safety margin.

[0123] The calculation results are then compared with a preset lower limit constraint value, and the larger one is selected to obtain two candidate bandwidths based on the entry threshold and exit threshold, respectively. Finally, the larger value of the two candidate bandwidths is taken as the uncertain bandwidth uniformly adopted by the system.

[0124] The preset neighborhood width is an offline preset parameter that defines the range of the threshold neighborhood. A width that is too small may result in insufficient sample numbers within the neighborhood, affecting the stability of subsequent statistics; a width that is too large will introduce samples unrelated to the characteristics of the threshold point, diluting the representativeness of the neighborhood. Based on engineering practice, this width is typically set to 3% to 10% of the solubility index range, for example, 0.05. Alternatively, it can be determined by analyzing the short-term fluctuation characteristics of the solubility index in historical data and taking a typical value of its fluctuation range (such as 1.5 times the daily average fluctuation amplitude).

[0125] The pre-defined high quantile refers to a relatively high percentile selected within the uncertainty set of samples in the threshold neighborhood. The high quantile is chosen to represent a generally high level of uncertainty within the neighborhood, rather than an extreme outlier. Typically, the 80th percentile (P80) is selected as this high quantile value. Within the threshold neighborhood, all sample uncertainty values ​​are arranged in ascending order, and the value ranked at the 80th percentile is taken as the representative uncertainty for that neighborhood. The same operation is performed for samples leaving the threshold neighborhood.

[0126] The pre-conversion mentioned above refers to a multiplier coefficient that amplifies the aforementioned high quantile uncertainty. Its purpose is to provide an additional safety margin beyond the uncertainties exhibited by the data, to address dynamic disturbances and measurement noise that the model cannot fully cover. This coefficient is typically set to a value greater than 1, such as 1.2, based on engineering experience.

[0127] The lower bound constraint refers to the minimum bandwidth value set to ensure that the system has basic buffering capacity under all circumstances. This is to prevent the calculated bandwidth from being too narrow and losing its practical buffering significance when the data quality is extremely high and the uncertainty is generally very small. The lower bound is usually set based on the requirements for system stability and empirical judgment, for example, 0.03. Within the neighborhood of the threshold, the uncertainty values ​​of all samples are arranged in ascending order, and the value ranked at the 80th percentile is taken as the representative uncertainty of that neighborhood. The same operation is performed on the neighborhood of the exit threshold.

[0128] Example: Neighborhood width is 0.05, high quantile is P80, conversion factor is 1.2, and lower limit constraint is 0.03. Based on historical data, the P80 value of the uncertainty of samples entering the threshold neighborhood is 0.025, and the P80 value of the uncertainty of samples leaving the threshold neighborhood is 0.020.

[0129] Candidate bandwidth on the ingress side: max(0.025×1.2,0.03)=max(0.030,0.03)=0.030

[0130] Candidate bandwidth for exit side: max(0.020×1.2,0.03)=max(0.024,0.03)=0.030

[0131] Ultimately, the unified uncertain bandwidth is: max(0.030,0.030)=0.030.

[0132] S23. Set the entry threshold, exit threshold, and uncertain bandwidth to preset parameters.

[0133] This invention, based on the meltability index distribution of historical successful melt samples, employs a multi-objective optimization strategy to automatically determine entry and exit thresholds, and determines a unified uncertainty bandwidth based on the statistical characteristics of uncertainty within the threshold neighborhood. This method transforms the source of thresholds and bandwidth from subjective experience into objective data, ensuring the interpretability and reproducibility of parameter settings. The system obtains a robust operating baseline adaptable to changes in different lines and environments, fundamentally avoiding over-melting, under-melting, or frequent switching caused by improper threshold settings.

[0134] Reference Figure 2 S3. Load the frozen parameter set and preset parameters, calculate the current fusibility index based on the real-time running parameters, and generate the preliminary judgment result based on the comparison relationship between the fusibility index and the entry threshold, exit threshold and uncertain bandwidth.

[0135] S31. Load the frozen parameter set to initialize the fusibility index calculation model, and calculate the current fusibility index based on the real-time collected operating parameters.

[0136] Read the frozen parameter set and load it into the meltable index calculation model, run it in a frozen state, and read the preset parameters at the same time.

[0137] The preset parameters are the entry threshold, exit threshold, and uncertain bandwidth.

[0138] In each sampling period, the fusibility index calculation model is invoked to calculate the current fusibility index and uncertainty without updating the parameters.

[0139] S32. Based on the mutually exclusive segment division rules, the calculated current meltability index is compared with the entry threshold, exit threshold and uncertain bandwidth in the preset parameters to generate the preliminary judgment result of the stage.

[0140] Based on the mutually exclusive segment division rules, the current meltability index is compared with the entry threshold, exit threshold and uncertain bandwidth to generate preliminary judgment results for the stages of preheating, main melting or heat preservation or suspension.

[0141] It should be further explained that the rules for dividing mutually exclusive sections are:

[0142] If the current solubility index is less than the entry threshold minus the uncertain bandwidth, it is judged as preheating.

[0143] If the current solubility index is greater than or equal to the entry threshold minus the uncertain bandwidth, and at the same time less than or equal to the entry threshold plus the uncertain bandwidth, it is judged as the entry threshold gray zone, and an output uncertain band mark is added.

[0144] If the current meltability index is greater than the entry threshold plus the uncertain bandwidth, and at the same time less than the exit threshold minus the uncertain bandwidth, then it is judged as the main melting.

[0145] If the current solubility index is greater than or equal to the exit threshold minus the uncertain bandwidth, and at the same time less than or equal to the exit threshold plus the uncertain bandwidth, it is judged as the exit threshold gray zone, and an output uncertain band mark is added.

[0146] If the current meltability index is greater than the exit threshold plus the uncertain bandwidth, it is judged to be either kept warm or suspended.

[0147] S33. If the current fusibility index is within the uncertain bandwidth corresponding to the entry threshold or exit threshold, then an uncertain band mark is added to the preliminary judgment result and it is retained as a result to be confirmed.

[0148] When the current fusibility index is within the uncertain bandwidth corresponding to the entry or exit threshold, an uncertain band marker is added to the output, and the preliminary judgment result of the current stage is retained as a result to be confirmed.

[0149] Record the current results along with the timestamp and parameter version information.

[0150] The entry threshold, exit threshold, and bandwidth are calculated from historical samples under a unified standard, ensuring clear sources and recalculation, facilitating reuse across lines and seasons. The bandwidth is taken from the high uncertainty level of the threshold neighborhood, then converted and constrained by a fixed lower limit, covering common disturbances without amplification and reducing critical point misjudgments. Threshold optimization considers success rate, action frequency, energy consumption, and time to achieve the target, avoiding under- or over-fusion caused by pursuing only a single indicator, and reducing the workload of repeated tuning during online deployment. The five-segment division using the entry threshold, exit threshold, and gray zones on both sides provides better isolation of critical uncertainties than single-threshold start / stop. Direct judgment outside the threshold and delayed judgment in the threshold neighborhood can reduce frequent switching caused by measurement noise or short-term weather fluctuations.

[0151] It operates in a frozen state, making inferences only and not drifting with on-site data, ensuring consistency of standards; it uses fixed thresholds and bandwidth to determine mutually exclusive sections, with simple and clear logic, facilitating implementation and joint debugging; gray areas are only marked and not switched temporarily, leaving a confirmation channel for detection in the same period; it records timestamps and parameter versions, which helps with future review and problem localization.

[0152] Reference Figure 3 S4. Based on the initial judgment results, perform low-power detection to confirm the final control stage and issue the ice melting control command, and apply stability constraints when switching stages.

[0153] S41. When the initial judgment result is pending confirmation, execute the low-power detection confirmation process. Perform low-power detection and collect feedback data according to the preset detection amplitude and detection duration.

[0154] The preset detection amplitude refers to the power or current increment added to the current command when performing low-power detection, which is sufficient to trigger a quantifiable system response but will not impact system safety and stability. The detection amplitude can be set to 5% to 10% of the typical power command value when the system performs the main de-icing phase. If the typical current of the main de-icing is 100A, the detection amplitude can be determined to be 5A to 10A.

[0155] The detection duration is the length of time during which the above-mentioned low-power detection is performed and feedback data is continuously collected.

[0156] S42. Perform low-power detection and collect feedback data according to the detection amplitude and detection duration.

[0157] S43. Input the fusibility index calculation model based on the feedback data, and recalculate the current fusibility index.

[0158] S44. The recalculated solubility index is compared again with the entry threshold, exit threshold, and uncertain bandwidth to confirm the final control stage.

[0159] Based on the feedback data, only the input data of the current sampling period is updated. The current fusibility index and uncertainty are calculated under the frozen parameter set, and then compared with the entry threshold, exit threshold and uncertainty bandwidth again.

[0160] The control phase was confirmed based on the results of the second comparison.

[0161] When the bandwidth is still uncertain and the preset maximum number of probes has been reached, maintain the current stage, switch to heat preservation or pause, and output a mark that requires manual review, without triggering stage switching.

[0162] S45. Based on the confirmed final control stage, generate the target power command or current command, and verify and trim the command according to the operating boundary before issuing it.

[0163] S46. Issue the verified and trimmed instructions. The stage switch must meet the minimum dwell condition and hysteresis condition. The rate of change of instructions during the switch is subject to limit constraints.

[0164] Based on the confirmed control phase, a target power or current command is generated, and the amplitude and rate of change of the command are pre-trimmed before issuance.

[0165] Before issuing the instructions, the instructions are checked against the preset power supply and operation boundaries.

[0166] The operating boundaries include the upper limit of available power on the AC side, the upper limit of available power on the energy storage side, the upper and lower limits of the state of charge, the upper limit of allowable voltage drop, the upper limit of current ramp rate, and the upper limit of conductor temperature rise.

[0167] When any boundary is touched, the instruction is trimmed to the nearest feasible value and the reason for trimming is recorded.

[0168] Each stage of execution should meet the preset minimum dwell time, and stage switching requests will not be accepted until the minimum dwell time is reached.

[0169] The execution phase switch is performed only if the determination of the target phase remains valid within a preset number of consecutive sampling periods; if the determination becomes unstable during the hysteresis confirmation period, the original phase is maintained.

[0170] Phase switching is executed step-by-step according to the current ramp-up rate limit, and a new phase switching is not triggered before the switching process is completed. Each instruction generation and issuance, phase start and end time, whether the boundary is touched and the corresponding pruning situation, whether the switching is rejected due to minimum dwell or hysteresis, as well as parameter version information and timestamps are all fully recorded.

[0171] S5. During the de-icing process, implement safety monitoring and terminate the de-icing process when the process termination conditions are met.

[0172] S51. During the execution of the target power command or current command, continuously monitor the actual current, voltage, temperature and power status parameters of the system; when any parameter exceeds the safety boundary, it is determined as an over-limit event.

[0173] S52. Once an out-of-limit event occurs, the minimum residence and hysteresis conditions are ignored, and degraded control is executed.

[0174] S53, Degradation control, according to the severity of exceeding the limit, includes in the following order: trimming the current command amplitude to within the safety boundary, forcibly switching the control phase to the heat preservation phase, and executing a shutdown.

[0175] S54. When the amount of residual ice on the line is determined to be below the safety threshold based on real-time data, or the system still cannot be restored to safe operation after downgrade control, or the cumulative running time or energy consumption reaches the preset upper limit, the current de-icing process shall be terminated.

[0176] An event is considered to be out of bounds if any of the following conditions occur: temperature rise exceeds the preset upper limit, allowable voltage drop exceeds the preset upper limit, current ramp rate exceeds the preset upper limit, available power on the AC side is insufficient, available power on the energy storage side is insufficient, or state of charge exceeds the preset range.

[0177] It should be further noted that the preset upper limit for temperature rise is determined based on the national standard for the selected conductor type or the long-term allowable operating temperature provided by the manufacturer. The preset upper limit for allowable voltage drop is determined based on the voltage deviation limits specified in the national standard for power quality of the power distribution system. The preset upper limit for current ramp-up rate is determined based on the manufacturer's factory parameters.

[0178] Insufficient available power on the AC side means that the actual available power value on the AC side, as monitored or received by the system in real time, is lower than the safe upper limit threshold for available power on the AC side. Insufficient available power on the energy storage side means that the actual available power value of the energy storage device, as monitored or received by the system in real time, is lower than the safe upper limit threshold for available power on the energy storage side.

[0179] The state of charge (SOC) exceeding the preset range refers to the actual value of the current SOC of the energy storage device as monitored in real time by the system being higher than the preset upper limit of SOC safety or lower than the preset lower limit of SOC safety (the upper and lower limits are set according to the battery technical specifications, such as 20% to 90%).

[0180] When an over-limit is detected, the current power or current command is first checked and trimmed. If the over-limit still occurs within a consecutive preset sampling period, the stage is downgraded to heat preservation (the heat preservation stage refers to issuing a fixed low power or zero current command to the line that is much lower than the main de-icing power). If the over-limit is still not eliminated, a shutdown is executed.

[0181] For each instance of exceeding limits, instruction pruning, stage downgrading, or shutdown, the event type, the preset threshold and actual measured value of the running boundary that triggered the limit exceeding, the stage at which it occurred, the values ​​before and after instruction pruning, the time of occurrence, and parameter version information must be recorded.

[0182] The current de-icing process will be terminated if any of the following conditions are met: the preset de-icing endpoint criteria are reached, the cumulative energy consumption reaches the preset upper limit, the cumulative duration reaches the preset upper limit, the gray area low-power detection has reached the preset maximum number of times but still cannot confirm the stage, or the main de-icing cannot be maintained due to boundary clipping within a continuous preset sampling period.

[0183] The preset melting endpoint criterion to be explained is when the system predicts the residual ice thickness based on real-time data. If the ice level is not greater than the preset melting threshold, the line is considered to have melted completely. Meeting this condition means that the goal of this ice-melting operation has been achieved, and the process can be safely terminated.

[0184] The cumulative energy consumption reaching the preset upper limit refers to the total electrical energy consumed in all ice-melting stages since the start of this ice-melting process, reaching the preset maximum value. The current preset maximum value is set based on the budgeted cost of a single ice-melting operation, energy quota, or typical energy consumption per unit length based on historical data, aiming to prevent energy consumption from getting out of control due to abnormal situations.

[0185] The phrase "cumulative duration reaching the preset upper limit" refers to the total duration of the current ice-melting process reaching a pre-set maximum allowable duration. This preset maximum allowable duration is set based on the operation time window, power supply reliability requirements, or equipment continuous operation capability.

[0186] The "gray zone low-power probe reaching the preset maximum number of attempts still cannot confirm" stage refers to the maximum number of consecutive low-power probe attempts allowed within a single gray zone (entry or exit). This number (typically 3 to 5 attempts) is designed to provide the system with sufficient opportunities for status confirmation while preventing it from getting stuck in meaningless probe loops under persistently uncertain anomalies.

[0187] The inability to maintain main de-icing due to boundary clipping within a continuous preset sampling period refers to the situation where, within a continuous preset number of sampling periods (e.g., 5-10 periods), the power or current command issued is continuously clipped to below the minimum effective value required to maintain main de-icing due to reaching the operating boundary (e.g., insufficient power, excessive voltage drop), resulting in the actual output failing to meet the de-icing requirements.

[0188] When the process terminates, the running data of this round, the sequence of phase changes, the running boundary thresholds and uncertain bandwidth that triggered the over-limit event, the boundary occupancy and instruction pruning logs, and the list of over-limit and shutdown events are archived together.

[0189] Example 3 is an embodiment of the present invention. This embodiment provides a multi-stage de-icing system for power distribution lines based on the meltability index and uncertainty zone detection, including a model construction module, a partitioning module, a calculation module, a confirmation control stage module, and an output module.

[0190] The model building module collects the topology and operating parameters of the target line, establishes a melting index calculation model, calibrates and freezes it using the minimum objective function, and outputs the frozen parameter set.

[0191] The segmentation module, after freezing, determines the entry threshold, exit threshold, and uncertain bandwidth required for stage segmentation based on historical samples, as preset parameters.

[0192] The calculation module loads the frozen parameter set and preset parameters, calculates the current fusibility index based on the real-time running parameters, and generates the preliminary judgment result based on the comparison relationship between the fusibility index and the entry threshold, exit threshold and uncertain bandwidth.

[0193] The confirmation control phase module performs low-power detection to confirm the final control phase based on the initial phase judgment results and issues ice melting control commands, and applies stability constraints during phase switching.

[0194] The output module performs safety monitoring during the ice melting process and terminates the current ice melting process when the process termination conditions are met.

[0195] This embodiment also provides an electronic device applicable to a multi-stage de-icing method for power distribution lines based on a meltability index and uncertainty zone detection, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the multi-stage de-icing method for power distribution lines based on a meltability index and uncertainty zone detection as proposed in the above embodiment.

[0196] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements a multi-stage de-icing method for power distribution lines based on a meltability index and uncertainty band detection as proposed in the above embodiment.

[0197] The storage medium proposed in this embodiment and the method for multi-stage de-icing of power distribution lines based on the meltability index and uncertainty band detection proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0198] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0199] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. 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 be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A multi-stage ice-melting method for power distribution lines based on the index of fusibility and uncertain band detection, characterized in that: include, Collect the topology and operating parameters of the target line, establish a melting index calculation model, calibrate and freeze it using the minimum objective function, and output the frozen parameter set; After freezing is completed, the entry threshold, exit threshold and uncertain bandwidth required for stage division are determined based on historical samples and used as preset parameters; Load the frozen parameter set and preset parameters, calculate the current fusibility index based on the real-time running parameters, and generate the preliminary judgment result based on the comparison relationship between the fusibility index and the entry threshold, exit threshold and uncertain bandwidth; Based on the initial assessment results, a low-power detection was performed to confirm the final control stage and issue an ice-melting control command, while applying stability constraints during stage switching. During the de-icing process, safety monitoring is performed, and the de-icing process is terminated when the termination conditions are met.

2. A multi-stage ice-melting method for power distribution lines based on the index of fusibility and the detection of uncertain zones as claimed in claim 1, characterized in that: The topology and operating parameters of the target line are collected, a melting index calculation model is established, and a minimum objective function is used for calibration and freezing. The output frozen parameter set includes... The topology and equipment parameters of the target line, historical icing and handling records, operational monitoring data, and short-term meteorological data are collected to construct the main features and constraint signal vectors. The main features include equivalent heat input characteristics, flow heat dissipation intensity characteristics, and precipitation phase burden indication characteristics; The main features are normalized within the training window and scaled to the range of {0, 1}. Based on the principal features and constraint signal vectors, a fusibility index calculation model is constructed through weighted combination to calculate the fusibility index, and the corresponding uncertainty can be calculated based on the fusibility index. Using historical records of whether ice has been completely melted as training labels, a minimum objective function is constructed based on the weight factors, intercepts, and weight factors of each element in the constraint signal vector corresponding to the main features. The optimal parameter set is then solved, and the solved optimal parameter set is frozen, outputting the frozen parameter set.

3. A multi-stage ice-melting method for power distribution lines based on the index of fusibility and the detection of uncertain zones as claimed in claim 2, characterized in that: After the freezing process is completed, the entry threshold, exit threshold, and uncertain bandwidth required for phase division are determined based on historical samples, and these are included as preset parameters. Based on the solubility index of the historical samples, and taking the successfully melted samples as a benchmark, the entry threshold and exit threshold are optimized and determined according to the distribution of the corresponding solubility index. For the entry threshold and exit threshold, the corresponding threshold neighborhood is determined in the historical samples, and a unified uncertainty bandwidth is determined based on the uncertainty of the samples in the threshold neighborhood. Set the entry threshold, exit threshold, and uncertain bandwidth to preset parameters.

4. The multi-stage de-icing method for power distribution lines based on the solubility index and uncertainty zone detection as described in claim 3, characterized in that: The initial judgment results of the generation stage include, The meltable index calculation model is initialized by loading the frozen parameter set, and the current meltable index is calculated based on the real-time collected operating parameters. Based on the mutually exclusive segment division rules, the calculated current meltability index is compared with the entry threshold, exit threshold and uncertain bandwidth in the preset parameters to generate the preliminary judgment result of the stage. If the current fusibility index falls within the uncertain bandwidth corresponding to the entry or exit threshold, an uncertain band marker is added to the preliminary judgment result, and it is retained as a result to be confirmed.

5. A multi-stage ice-melting method for power distribution lines based on the index of fusibility and the detection of uncertain zones as claimed in claim 4, characterized in that: The step of performing low-power detection to confirm the final control stage based on the initial assessment results includes... When the initial judgment result is a result pending confirmation, execute the low-power detection confirmation process; Perform low-power detection and collect feedback data according to the detection amplitude and detection duration; The fusibility index calculation model is based on the feedback data, and the current fusibility index is recalculated. The recalculated solubility index is compared again with the entry threshold, exit threshold, and uncertain bandwidth to confirm the final control stage.

6. A multi-stage ice-melting method for power distribution lines based on the index of fusibility and the detection of uncertain zones as claimed in claim 5, characterized in that: The issuance of ice-melting control commands and the application of stability constraints during phase switching include, Based on the confirmed final control stage, a target power command or current command is generated, and the command is checked and trimmed according to the operating boundary before it is issued. The verified and trimmed instructions are issued. The phase switching must meet the minimum dwell conditions and hysteresis conditions. The rate of change of instructions during the switching process is subject to limit constraints.

7. The multi-stage de-icing method for power distribution lines based on the solubility index and uncertainty zone detection as described in claim 6, characterized in that: The process of performing safety monitoring during the ice-melting process and ending the ice-melting process when the process termination conditions are met includes, During the execution of the target power command or current command, the actual current, voltage, temperature and power status parameters of the system are continuously monitored; when any parameter exceeds the safety boundary, it is determined to be an out-of-limit event. Once an out-of-limit event occurs, the minimum residence and hysteresis conditions will be ignored, and degraded control will be implemented. Degradation control, in order of severity of exceeding limits, includes: trimming the current command amplitude to within the safety boundary, forcibly switching the control phase to the heat preservation phase, and executing a shutdown; When the amount of residual ice on the line is determined to be below the safety threshold based on real-time data, or the system still cannot resume safe operation after downgrade control, or the cumulative running time or energy consumption reaches the preset upper limit, the current de-icing process will be terminated.

8. A multi-stage de-icing system for power distribution lines based on the solubility index and uncertainty zone detection, employing the multi-stage de-icing method for power distribution lines based on the solubility index and uncertainty zone detection as described in any one of claims 1 to 7, characterized in that, include: The module includes a model building module, a partitioning module, a calculation module, a confirmation and control phase module, and an output module. The model building module collects the topology and operating parameters of the target line, establishes a melting index calculation model, uses a minimum objective function for calibration and freezing, and outputs a frozen parameter set. The segmentation module, after freezing, determines the entry threshold, exit threshold, and uncertain bandwidth required for stage segmentation based on historical samples, as preset parameters. The calculation module loads the frozen parameter set and preset parameters, calculates the current fusibility index based on the real-time running parameters, and generates a preliminary judgment result based on the comparison relationship between the fusibility index and the entry threshold, exit threshold and uncertain bandwidth. The confirmation control phase module performs low-power detection to confirm the final control phase and issues a melting control command based on the initial phase judgment result, and applies stability constraints when switching phases. The output module performs safety monitoring during the ice melting process and terminates the ice melting process when the process termination conditions are met. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. When the processor executes the computer program, it implements the steps of the multi-stage de-icing method for power distribution lines based on the meltability index and uncertainty band detection as described in any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the multi-stage de-icing method for power distribution lines based on the meltability index and uncertainty band detection as described in any one of claims 1 to 7.

Images