An energy self-management sag monitoring terminal power supply optimization method

Abstract

This invention discloses a power supply optimization method for an energy-self-managed sag monitoring terminal, specifically relating to the field of online monitoring of transmission lines and terminal power supply management. It addresses the problems of repeated protection actions and disordered task recovery when power supply fluctuations are significant in complex electromagnetic environments, making it difficult to simultaneously manage critical monitoring and transmission content. The method generates a power supply status identifier by acquiring the protection trigger flag output by the power protection circuit and the power supply status information output by the energy storage unit. It then divides the monitoring retention task set and the delayed recovery task set according to the recovery sequence table, determines the minimum evidence retention segment based on the sag monitoring results, and further determines the recovery conditions based on the continuous energy release interval, net releaseable energy after protection, energy consumption for abnormal evidence fidelity, and task recovery discretion coefficient. Finally, it updates the current operating level and organizes the transmission content, thereby achieving orderly recovery and tiered transmission of the sag monitoring terminal under limited power supply conditions.

Description

Technical Field

[0001] This invention relates to the field of online monitoring of transmission lines and terminal power supply management, and more specifically, to a method for optimizing the power supply of a sag monitoring terminal with energy self-management. Background Technology

[0002] In online monitoring technology for power transmission lines, sag changes are crucial information reflecting the line's operational status, especially in deployment scenarios where ultra-high voltage (UHV) lines are exposed to the outdoor environment for extended periods, such as along the line, near towers, or at observation points beneath towers. Monitoring terminals typically not only handle image acquisition but also image analysis, sag point identification, status assessment, and result transmission. Existing technical document CN105987665A, titled "An Early Warning Monitoring Device and Method for the Variation Range of Sag Points in UHV Transmission Lines," discloses a sag early warning monitoring scheme composed of a video image analysis module, a communication module, and a power management module. The video image analysis module detects and identifies curve points on the transmission line and calculates the sag point location. The communication module transmits images, videos, and early warning information to the monitoring center. The power management module provides power management for the video image analysis module and the communication module. This document further describes that the power management module mainly consists of a power protection circuit and a power management circuit. The power protection circuit primarily protects against lightning strikes and electromagnetic interference in the complex electromagnetic environment of UHV transmission lines, while the power management circuit manages the power supply to the entire unit through branch switches. It is evident that existing technologies have recognized that when sag monitoring terminals operate near high-voltage lines, in addition to performing monitoring and transmission functions, they also need to possess corresponding power protection and basic power supply management capabilities to adapt to the operational requirements of complex on-site environments.

[0003] However, the power supply design of the aforementioned existing technologies still primarily focuses on circuit protection and branch control, lacking further coordination for the terminal power supply restoration process after protection actions are triggered in complex electromagnetic environments. While existing solutions can reduce direct damage to the circuit body from external shocks to some extent, they lack a unified control mechanism that matches the energy state regarding the order in which terminal functional modules should be restored after protection is lifted, the extent to which monitoring and communication tasks should be maintained under limited power conditions, and which key functions should be prioritized with remaining energy. In situations with frequent electromagnetic disturbances around the line and fluctuating power supply status, the terminal is prone to repeated power supply constraints, repeated startups, repeated communication establishments, and indiscriminate restoration of high-load modules. This results in a significant waste of limited energy in non-core processes, making it difficult to guarantee the continuity of sag monitoring and affecting the timeliness and stability of early warning information. Especially during operational phases where complex electromagnetic environments are more pronounced and real-time monitoring of line risks is crucial, this lack of hierarchical and rhythmic power restoration can further amplify monitoring gaps and alarm delays. Therefore, existing technologies still need to introduce a power supply optimization mechanism based on energy state on the basis of power supply protection, so that the sag monitoring terminal can maintain orderly and sustainable operation before and after protection action.

[0004] To address the aforementioned problems, a technical solution is provided. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a power supply optimization method for an energy-self-managed sag monitoring terminal. This method generates a power supply status identifier by acquiring the protection trigger flag output by the power protection circuit and the power supply status information output by the energy storage unit. It then divides the monitoring retention task set and the delayed recovery task set according to the recovery sequence table, determines the minimum evidence retention segment based on the sag monitoring results, and further determines the recovery conditions based on the continuous energy release interval, the net releaseable energy after protection, the energy consumption for abnormal evidence fidelity, and the task recovery discretion coefficient. Finally, it updates the current operating level and organizes the transmission content, thereby achieving orderly recovery and graded transmission of the sag monitoring terminal under limited power supply conditions, thus solving the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: S1: Obtain the protection trigger flag output by the power protection circuit and the power supply status information output by the energy storage unit, generate the current power supply status identifier of the terminal, input the power supply status identifier into the preset recovery sequence table, divide the monitoring retention task set and the delayed recovery task set according to the recovery sequence table, and determine the current operating level corresponding to the monitoring retention task set. S2: Start the monitoring retention task set according to the current operation level, perform sag identification processing to obtain sag monitoring results, and save the reporting content corresponding to the sag monitoring results as pending information; S3: Determine the continuous energy release interval based on the protection trigger mark, combine the power supply status information and the monitoring retention task set to obtain the net energy that can be released after protection, determine the minimum evidence retention segment based on the sag monitoring results and obtain the energy consumption for abnormal evidence fidelity, and input the two into the pre-trained dual-channel recovery evaluation model to obtain the task recovery discretion coefficient. S4: Determine whether the power supply status indicator meets the restoration conditions based on the task restoration discretion coefficient. If it does, put the delayed restoration task set into the restoration sequence table to update the current operating level. Organize the information to be sent based on the updated current operating level to obtain the sending content. Output the current operating level, sag monitoring results and sending content to the communication module, and record the current operating level as the level reference information.

[0007] Furthermore, generating the current power supply status identifier of the terminal includes collecting protection trigger markers and power supply status information at a fixed sampling period within the observation time window, calculating the duration of the protection non-triggered state within the observation time window, calculating the duration of the power supply status information satisfying the power supply requirements of each item in the recovery sequence table, and then comparing the duration with the preset criteria of each item to determine the power supply status identifier.

[0008] Furthermore, each entry in the recovery sequence table corresponds to a unique monitoring reserved task set, delayed recovery task set, and current operating level. Each entry also records the minimum operating voltage and minimum sustaining power of the monitoring reserved task set. When no entry meets the criteria, the basic reserved entry in the recovery sequence table is read, and the content corresponding to the basic reserved entry is used as the power supply status identifier, monitoring reserved task set, and current operating level.

[0009] Furthermore, sag identification processing is performed, including enabling monitoring image acquisition tasks, monitoring image preprocessing tasks, conductor contour extraction tasks, and sag identification result caching tasks in the monitoring retention task set according to the current operation level. Monitoring images are acquired at the same sampling time as in step S1, and calibration segments and left and right support points are extracted from the monitoring images to establish length conversion relationships and support baselines.

[0010] Furthermore, the sag recognition process also includes performing median filtering, grayscale enhancement, and edge extraction on the monitoring image in sequence, filtering out non-conductor contours to obtain a conductor contour point set, fitting a conductor contour curve based on the conductor contour point set, and obtaining the actual sag value based on the vertical distance between the lowest point of the conductor contour curve and the support baseline. The difference between the actual sag value and the reference sag value forms the sag offset.

[0011] Furthermore, the continuous energy release interval is determined based on the protection trigger mark, including searching along the sampling time sequence for candidate time periods when the protection trigger mark changes from triggered to non-triggered and the output voltage is not lower than the minimum operating voltage of the monitoring retention task set, and using the candidate time period containing the sampling time corresponding to the sag monitoring result as the continuous energy release interval; the net releasable energy supply after protection is obtained by subtracting the energy consumption of monitoring retention maintenance from the energy supply of the energy release interval within the continuous energy release interval.

[0012] Furthermore, based on the sag monitoring results, the minimum evidence retention segment is determined, including matching each evidence template entry in the evidence template library according to the actual sag value, sag offset, and the change in sag offset between adjacent sampling times, determining the current abnormal state, and reading the field set corresponding to the current abnormal state to form the minimum evidence retention segment; when multiple evidence template entries meet the requirements at the same time, the evidence template entry with the fewest fields and ranked first in the evidence template library is selected.

[0013] Furthermore, the energy consumption for maintaining the fidelity of abnormal evidence is accumulated from the energy consumption for generating the minimum evidence retention fragment, the energy consumption for encapsulating the minimum evidence retention fragment, and the energy consumption for storing the minimum evidence retention fragment. The dual-channel recovery evaluation model receives the net releasable energy supply after protection and the energy consumption for maintaining the fidelity of abnormal evidence, and outputs the task recovery discretion coefficient. The task recovery discretion coefficient and the task recovery discretion coefficient threshold pre-written in the terminal parameter area are used together to determine the recovery conditions.

[0014] Furthermore, the delayed recovery task set is put into operation sequentially according to the recovery sequence table to update the current operating level. This includes reading the recovery order and incremental maintenance power of each task in the delayed recovery task set recorded in the recovery sequence table, determining the allowable recovery supply energy based on the task recovery discretion coefficient and the net releasable supply energy after protection, and, under the constraint that the cumulative recovery energy consumption is not higher than the allowable recovery supply energy, putting the tasks in the delayed recovery task set into operation sequentially according to the recovery order, thereby determining the updated current operating level.

[0015] Furthermore, the current operating level, sag monitoring results, and transmission content are output to the communication module. This includes constructing a transmission sequence according to the transmission field set corresponding to the current operating level, and performing segmented transmission and acknowledgment verification in the order of basic transmission segment priority over evidence transmission segment. After segmented transmission, the transmission status, the amount of data transmitted, and the amount of data to be transmitted are written back to the transmission information. At the same time, the current operating level, sampling time, power supply status identifier, and transmission status are written into the level reference information.

[0016] The technical effects and advantages of the power supply optimization method for an energy-self-managed sag monitoring terminal of the present invention are as follows: This invention starts with the power supply status indicator and organizes the power protection status, energy storage unit power supply status, monitoring task retention, abnormal evidence retention, and communication transmission into the same processing chain. This allows the sag monitoring terminal to maintain the monitoring task set even under conditions of repeated protection triggers and significant power supply fluctuations, continuously generating sag monitoring results, and conditionally activating the delayed recovery task set based on these results. Compared to processing methods that rely solely on circuit protection or static power supply management, this invention establishes recovery determination based on the joint constraints of continuous energy release intervals, net releaseable energy after protection, and energy consumption for abnormal evidence fidelity. This gives the power supply recovery process clear hierarchical boundaries and execution order, thus reducing the impact of disordered recovery, repeated startups, and unnecessary energy consumption on the continuous operation of the terminal.

[0017] This invention further integrates the identification of current abnormal states, the determination of minimum evidence retention segments, and the organization of transmitted content into a unified process. By using a task recovery discretion coefficient to constrain recovery conditions and transmission levels, the terminal prioritizes the retention of key sag status information and necessary evidence fields even under power constraints, and organizes output content according to the current operating level. Thus, the transmitted content remains consistent with the current power supply conditions, current abnormal states, and current operating level. This not only improves the integrity and traceability of remotely received information but also allows level reference information to participate in the generation of subsequent power supply status identifiers, thereby enhancing the terminal's operational stability, monitoring continuity, and the reliability of information output in complex electromagnetic environments. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating a power supply optimization method for an energy self-management sag monitoring terminal according to the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1 This invention provides a power supply optimization method for an energy-self-managed sag monitoring terminal, comprising: S1: Obtain the protection trigger flag output by the power protection circuit and the power supply status information output by the energy storage unit, generate the current power supply status identifier of the terminal, input the power supply status identifier into the preset recovery sequence table, divide the monitoring retention task set and the delayed recovery task set according to the recovery sequence table, and determine the current operating level corresponding to the monitoring retention task set.

[0021] S2: Start the monitoring retention task set according to the current operation level, perform sag identification processing to obtain sag monitoring results, and save the reporting content corresponding to the sag monitoring results as pending information.

[0022] S3: Determine the continuous energy release interval based on the protection trigger mark, combine the power supply status information and the monitoring retention task set to obtain the net energy that can be released after protection, determine the minimum evidence retention segment based on the sag monitoring results and obtain the energy consumption for abnormal evidence fidelity, and input the two into the pre-trained dual-channel recovery evaluation model to obtain the task recovery discretion coefficient.

[0023] S4: Determine whether the power supply status indicator meets the restoration conditions based on the task restoration discretion coefficient. If it does, put the delayed restoration task set into the restoration sequence table to update the current operating level. Organize the information to be sent based on the updated current operating level to obtain the sending content. Output the current operating level, sag monitoring results and sending content to the communication module, and record the current operating level as the level reference information.

[0024] This invention addresses the practical problems of power supply fluctuations, repeated protection triggering, and disordered task recovery in sag monitoring terminals under complex electromagnetic environments. Its core approach is not simply to increase power extraction capacity or energy storage capacity, but rather to first convert the power protection status and energy storage unit's power supply capacity into power supply status indicators. Then, based on a recovery sequence table, the terminal tasks are divided into a set of monitoring tasks that must be prioritized and a set of delayed recovery tasks that can be recovered later. This allows the terminal to first stably obtain sag monitoring results when power supply is limited. Based on this, the available energy reserve is calculated using continuous energy release intervals. Simultaneously, the current abnormal state is identified based on the sag monitoring results, and the minimum evidence retention segment is determined. The energy required for retaining abnormal evidence is then included in the recovery judgment. Finally, a recovery assessment model is used to determine whether and to what extent the delayed recovery task set is implemented, and the content is organized and sent according to the updated current operating level, and layered transmission is executed. Therefore, this invention integrates power management, monitoring and processing, evidence preservation, and communication transmission into the same processing chain. This prevents the terminal from blindly resuming all tasks after protection is lifted. Instead, it prioritizes monitoring, then assesses remaining capacity, performs restoration, and finally transmits the data. This reduces repeated startups and ineffective energy consumption when power supply is insufficient, while ensuring that critical sag states and necessary evidence are preserved first. This improves the terminal's continuous operation capability under disturbance conditions, the reliability of anomaly detection, and the effectiveness of link transmission.

[0025] When deploying sag monitoring terminals near ultra-high voltage lines, the triggering and deactivation of power supply protection circuits often occur intermittently with the output status of energy storage units. Observing the protection trigger mark alone or the power supply status information alone is insufficient to directly determine which tasks the terminal can maintain, which tasks should be retained, and which operating level it should remain at. Therefore, the key point of step S1 is to first converge the protection trigger mark and power supply status information into a readable power supply status identifier, and then use the recovery sequence table to transform the power supply constraints into task constraints, thereby limiting the boundaries of subsequent monitoring and recovery actions from the source.

[0026] S101: Establishment of observation time window and acquisition of raw information.

[0027] Step S1 first establishes an observation time window, which consists of a start time and an end time. Within the observation time window, the power protection circuit and the energy storage unit are synchronously sampled according to a fixed sampling period. The sampling content is limited to the protection trigger mark, output voltage, and output current. The protection trigger mark is used to indicate whether the power protection circuit is in a triggered or non-triggered state at the corresponding sampling time. The output voltage is used to indicate the voltage level that the energy storage unit can provide at the corresponding sampling time. The output current is used to indicate the current level that the energy storage unit can provide at the corresponding sampling time. After obtaining the above three types of information at the same sampling time, the output power at the corresponding sampling time is obtained by multiplying the output voltage and the output current. The output power is used to indicate the instantaneous power supply capability of the energy storage unit to the terminal at the corresponding sampling time. The sampling results together constitute the original basis of the power supply status information.

[0028] The aforementioned observation time window is set in accordance with the terminal recovery judgment rhythm. In one feasible manner, the observation time window starts at the first sampling moment after the power protection circuit changes from the triggered state to the non-triggered state, and ends at the sampling moment when the preset judgment time is reached. Thus, step S1 always focuses on the power supply recovery stage after the protection action is released. The fixed sampling period is determined according to the principle that the output changes of the energy storage unit can be continuously identified and the processor can record stably. This ensures that the changes between adjacent sampling moments are not missed due to sparse sampling, nor are meaningless repeated records introduced due to dense sampling. The aforementioned arrangement ensures that the protection trigger mark, output voltage, output current and output power all fall on the same sampling sequence.

[0029] S102: Construction of the restored sequence list and definition of criterion items.

[0030] After obtaining the original sampling results, a preset recovery sequence table is further invoked. The recovery sequence table consists of a criterion area and a mapping area. The criterion area is used to determine which type of power supply status identifier corresponds to the current power supply capability. The mapping area is used to directly provide the monitoring retention task set, the delayed recovery task set, and the current operating level after the power supply status identifier is determined. Each row entry in the recovery sequence table corresponds to a power supply status, and each row entry contains at least the minimum protection static stability ratio, the minimum graded power supply continuity ratio, the minimum operating voltage, the minimum maintenance power, the monitoring retention task set, the delayed recovery task set, and the current operating level. The aforementioned fields have a one-to-one correspondence. Any power supply status identifier can only be mapped to a unique row entry, thus maintaining a unique mapping between the preceding and following terms.

[0031] The minimum operating voltage and minimum sustaining power must have clearly defined sources. The minimum operating voltage is extracted from the minimum power supply voltage requirements of all tasks within the corresponding monitoring and retention task set. Specifically, it is determined by listing the minimum power supply voltage requirements of each task in the monitoring and retention task set, and then taking the highest one as the minimum operating voltage for that row. If any task in the monitoring and retention task set cannot maintain power supply, the monitoring and retention task set cannot be considered to be in sustainable operation. The minimum sustaining power is summarized from the sustaining power requirements of all tasks within the corresponding monitoring and retention task set. Specifically, it is determined by listing the sustaining power requirements of each task in the monitoring and retention task set under continuous operation, and then summing the sustaining power requirements sequentially to obtain the minimum sustaining power for that row. Once the monitoring and retention task set enters a continuous operation state, the energy storage unit should at least provide a power supply capacity equivalent to the sum of the aforementioned sustaining power requirements for an extended period.

[0032] The minimum static stability ratio of protection can be determined based on historical operation records. Specifically, several observation time windows in the same line environment where sag identification processing can be completed continuously are selected. The proportion of protection non-triggered states within the aforementioned observation time windows is statistically analyzed, and the minimum proportion that can support the continuous operation of the corresponding monitoring reserved task set is taken as the minimum static stability ratio of protection for that row of entries. The minimum graded power supply continuity ratio can be determined according to the continuous operation requirements of the corresponding monitoring reserved task set. Specifically, during the trial operation phase, the proportion of continuous power supply time with output voltage not lower than the minimum operating voltage and output power not lower than the minimum maintenance power is recorded as the proportion of the total observation time window. The minimum proportion that can complete all basic actions of the corresponding monitoring reserved task set is taken as the minimum graded power supply continuity ratio for that row of entries. The aforementioned determination process does not rely on abstract experience but is directly derived from the power supply boundaries of the monitoring reserved task set during field operation.

[0033] S103: Calculation of the proportion of static stability protection and the proportion of continuous power supply in the graded power supply.

[0034] After the restoration sequence table structure is clear, the power supply status indicator determination and calculation stage begins. First, the protection static stability ratio is calculated. Specifically, within the observation time window, the time periods when the protection trigger mark is in an inactive state are identified segment by segment. The durations of all inactive state periods are summed sequentially to obtain the cumulative inactive protection time. Then, the cumulative inactive protection time is divided by the total duration of the observation time window to obtain the protection static stability ratio. The protection static stability ratio reflects the proportion of time the power supply protection circuit remains in an inactive state within the observation time window. Its physical meaning is not whether the circuit is completely undisturbed, but rather whether the energy storage unit has a sufficiently continuous power supply environment after the protection action has been released.

[0035] Subsequently, the corresponding graded power supply duration ratio is calculated for each row of entries in the recovery sequence table. During the calculation, the minimum operating voltage and minimum sustaining power of that row of entries are used as criteria. Judgments are made at each sampling moment within the observation time window. Any sampling segment that simultaneously satisfies the following conditions—the protection trigger flag being in an untriggered state, the output voltage not lower than the minimum operating voltage of that row of entries, and the output power not lower than the minimum sustaining power of that row of entries—is included in the effective continuous power supply time of that row of entries. All effective continuous power supply times are sequentially accumulated to obtain the cumulative effective continuous power supply time of that row of entries. This cumulative effective continuous power supply time is then divided by the total duration of the observation time window to obtain the graded power supply duration ratio of that row of entries. The graded power supply duration ratio is not an isolated evaluation of voltage, current, and power; rather, the duration is accumulated only when the protection state, voltage boundary, and power boundary are simultaneously met. Therefore, the result is a power supply duration capability oriented towards a specific task level, rather than an abstract power supply capability detached from task requirements.

[0036] In one embodiment, the sag monitoring terminal is installed at the observation location under the tower. Intermittent electromagnetic disturbances are likely to occur near the line in the afternoon. The power protection circuit is triggered multiple times within a period of time and then recovers. The output voltage and output current of the energy storage unit fluctuate accordingly. If only the output voltage at a certain sampling moment is observed, it may be considered that the terminal has the conditions to recover to a higher task level. However, after including the aforementioned period in the observation time window, the protection static stability ratio may still be low, and the continuous ratio of graded power supply may only meet the basic row items and not the higher row items. Based on the aforementioned calculation results, step S1 will limit the power supply status indicator to the items that can support basic monitoring. Thus, step S2 will first retain the sag identification processing and not prematurely invest in subsequent high-load tasks. The on-site result is that the terminal first recovers the image acquisition and sag identification result buffer, and then waits for the power supply conditions to stabilize further before entering the next stage of processing.

[0037] S104: Power supply status indicator generation and task allocation output.

[0038] After obtaining the protection static stability ratio and the graded power supply continuity ratio corresponding to each row entry, available entries are screened row by row according to the row order of the recovery sequence table. The screening rules include two aspects: first, the protection static stability ratio is not lower than the minimum protection static stability ratio of the current row entry; second, the graded power supply continuity ratio of the current row entry is not lower than the minimum graded power supply continuity ratio of that row entry. Any row entry that simultaneously meets the above two conditions is included in the available entry set. If there are multiple row entries in the available entry set, the row entry with the highest operating capacity is selected as the entry corresponding to the power supply status identifier. The operating capacity is determined by the row order preset in the recovery sequence table, and the later the row entry is arranged, the wider the range of tasks it can support. If the available entry set is empty, the basic reserved entries preset in the recovery sequence table are taken as the entries corresponding to the power supply status identifier. The monitoring reserved task set corresponding to the basic reserved entries only contains the minimum task combination required to maintain sag identification processing. Therefore, even if the power supply capacity is at a weak level, step S1 can still output a definite result without any task boundary gaps.

[0039] Once the power supply status identifier is determined, the corresponding table entry is used as the sole mapping source. The monitoring reserved task set, delayed recovery task set, and current operating level in that table entry are read as the final output of step S1. The monitoring reserved task set is used to directly start and execute sag identification processing in step S2. The delayed recovery task set is used to sequentially start operation after recovery condition determination in step S3. The current operating level is used to limit the operating range in step S2 and limit the transmission range in step S4.

[0040] In one embodiment, the basic reserved entries in the recovery sequence table only retain the video acquisition, basic image preprocessing, and sag recognition result caches. The intermediate entries add image encapsulation preparation on the basis of the above, and the higher entries add extended storage and subsequent transmission preparation. When the protection action near the line has just been released, step S1 may first output the power supply status identifier corresponding to the basic reserved entries, and the terminal will only recover the sag recognition processing. As the protection non-triggered state continues to extend within the observation time window, the output voltage and output power gradually stabilize. In a subsequent judgment, step S1 may switch to the power supply status identifier corresponding to the intermediate entries, and the terminal will then have the basis to further recover subsequent tasks. The aforementioned field actions and results are all directly driven by the task division output of step S1, without the need for steps S2 to S4 to reinterpret the power supply boundary.

[0041] After step S1 is completed, the power supply status identifier, monitoring retention task set, delayed recovery task set and current operating level have formed a stable correspondence. The front-end terminal is no longer in a state of ambiguous power supply conditions and unclear task boundaries, but has obtained a clear task division basis for the current power supply background. The aforementioned processing enables the terminal to first determine the maintainable operating level in the initial stage after protection is released, and then decide on the scope of tasks that can be entered, so that the power supply constraints have an executable landing point.

[0042] After the monitoring task set and current operating level have been defined in step S1, the front-end terminal has a range of tasks that can be started, but it has not yet formed monitoring results that can be used to determine abnormal states and organize the content to be sent. Therefore, the focus of step S2 is not to expand the types of tasks, but to sequentially execute monitoring image acquisition, image processing and sag recognition within the established task boundaries, so as to truly implement the current operating level into calculable, storable and referable sag monitoring results and information to be sent, so that subsequent steps can be carried out around specific monitoring content rather than abstract task names.

[0043] S201: Enable the monitoring retention task set according to the current run level.

[0044] At the start of step S2, the power supply status identifier, monitoring reserved task set, and current operating level output from step S1 are read first. Using the monitoring reserved task set as the sole enabled scope, all tasks within the terminal are checked item by item. All tasks belonging to the monitoring reserved task set are enabled, while all tasks belonging to the delayed recovery task set are disabled, thus maintaining consistency between the task boundaries. Subsequently, the monitoring reserved task sets are executed sequentially according to the predefined execution order in the recovery sequence table. The execution order includes at least monitoring image acquisition, monitoring image preprocessing, conductor contour extraction, and sag recognition result caching. The reason for using a sequential execution order is that monitoring image acquisition provides raw data for subsequent image processing, monitoring image preprocessing provides identifiable edges for conductor contour extraction, and conductor contour extraction provides geometric data for sag recognition result caching. Only after the previous processing is completed can the input object required for the next processing be formed. Therefore, they cannot be substituted or interchanged arbitrarily.

[0045] The aforementioned activation actions correspond one-to-one with the current operating level. The more basic the monitoring and retention task set corresponding to the current operating level, the shorter the processing chain after activation. The more complete the monitoring and retention task set corresponding to the current operating level, the longer the processing chain after activation. However, regardless of the current operating level, step S2 only allows sag identification processing to be completed within the monitoring and retention task set. High-load processing actions in the delayed recovery task set are not put into operation in advance. The significance of this processing is that the sag monitoring results output by step S2 can be generated stably, while not crowding out the power supply margin required for subsequent recovery determination in step S3.

[0046] S202: Monitoring image formation and calibration benchmark establishment.

[0047] After the monitoring retention task set has been activated, the monitoring image acquisition task acquires monitoring images sequentially according to the sampling time within the observation time window of step S1. Each sampling time corresponds to a monitoring image frame, and each monitoring image frame establishes a unique correspondence with the corresponding sampling time. This allows the subsequent sag monitoring results, pending information, and protection trigger markers of step S1 to be connected along the same time sequence. In each monitoring image frame, the calibration segment is identified first, followed by the left support point and the right support point. The calibration segment is a fixed visible structure whose actual length has been determined during installation. The left support point and the right support point are the suspension positions of the two ends of the conductor in the image. The aforementioned three types of objects constitute the geometric reference for sag identification processing.

[0048] After the calibration segment is identified, the pixel length of the calibration segment in the current monitoring image is calculated. Specifically, the pixel positions of the two endpoints of the calibration segment are extracted first, and then the straight-line distance between the two endpoints is calculated. The actual length of the calibration segment recorded during the installation phase is divided by the pixel length to obtain the length conversion factor. The length conversion factor represents the actual length corresponding to each pixel in the current monitoring image. Thus, the pixel-level sag distance obtained later is converted into the actual sag value. After the left and right support points are identified, a support baseline is established by connecting the two points mentioned above. The support baseline serves as a reference line for the suspension positions of the two ends of the conductor and is used to calculate the vertical sag distance of the sag point relative to the two ends of the conductor later.

[0049] In one embodiment, the sag monitoring terminal is installed at the observation position under the tower. The image shows both the conductor and a section of hardware structure adjacent to the conductor and maintaining a fixed position. During installation, the actual length of the hardware structure is first measured and written into the terminal parameter area. When processing the current monitoring image in step S2, the hardware structure is first located in the image, and then the length conversion coefficient is calculated from the positions of the two ends of the hardware structure. At the same time, the suspension positions of the two ends of the conductor are identified as the left support point and the right support point. On-site, the same frame of monitoring image has both the actual length reference and the geometric reference of the two ends of the conductor, thus providing a unified reference for subsequent sag value calculation.

[0050] S203: Conductor contour extraction and sag point location.

[0051] After the length conversion coefficient and support baseline are established, step S2 performs image preprocessing on the monitoring image. The image preprocessing performs median filtering, grayscale enhancement and edge extraction in a fixed order. Median filtering is used to suppress isolated noise points, grayscale enhancement is used to increase the grayscale difference between the conductor and the background, and edge extraction is used to output the contour candidate set in the image. After the contour candidate set is formed, it is then filtered according to the continuity, extension direction and positional relationship of the conductor in the image. Non-conductor contours formed by tree branches, tower material edges and sky background are removed, and only the conductor contour point set consistent with the extension direction of the conductor is retained.

[0052] After the guideline point set is determined, a quadratic curve is fitted using the guideline point set as input. The fitting method adopts the principle of minimizing the absolute residual, specifically minimizing the sum of the absolute values ​​of the vertical differences between all guideline point sets and the fitted curve, thus obtaining a quadratic curve that can describe the sag shape of the guideline. After the quadratic curve is established, the lowest point of the quadratic curve is taken as the sag point. The horizontal position of the sag point in the image is determined by the ratio of the coefficients of the first and second terms of the quadratic curve, and the vertical position of the sag point in the image is obtained by substituting the aforementioned horizontal position into the quadratic curve. Subsequently, the horizontal position of the sag point is substituted into the support baseline to obtain the vertical position of the support baseline at the same horizontal position. The absolute difference between the vertical position of the support baseline and the vertical position of the sag point is then used as the pixel-level sag distance.

[0053] After the pixel-level sag distance is converted using a length conversion factor, the actual sag value corresponding to the current monitoring image is obtained. The conversion relationship is the length conversion factor multiplied by the pixel-level sag distance. The length conversion factor comes from the ratio between the actual length of the calibration segment and the pixel length of the calibration segment. During the installation phase, a reference sag value is also pre-recorded. The current actual sag value is subtracted from the reference sag value to obtain the sag offset. The sag offset represents the degree of deviation of the current conductor state from the installation calibration state. The subsequent step S3 can directly use the sag offset and the actual sag value to identify abnormal states without having to re-execute image recognition.

[0054] To ensure feasible boundary conditions for quadratic curve fitting, step S2 also sets absolute residuals and thresholds, which are derived from normal image samples during the installation and trial operation phase. Specifically, during the installation and trial operation phase, multiple frames of images with normal conductor morphology are acquired. Conductor contour extraction and quadratic curve fitting are then performed on these images. The sum of absolute residuals for each frame is calculated, and the maximum value of the sum of absolute residuals is written into the terminal parameter area as the absolute residual threshold. If, during the operation phase, the sum of absolute residuals exceeds the absolute residual threshold, it is determined that the current quadratic curve cannot reliably describe the conductor contour. In this case, conductor contour extraction is re-executed, or the catenary fitting method is switched to continue calculating sag points.

[0055] In one embodiment, when cloud shadows, swaying branches and leaves, and background reflections appear in the monitoring image, the outline of the conductor and non-conductor outlines often appear simultaneously after edge extraction. Step S2 first filters out stray outlines according to the continuous direction of the conductor in the picture, and then performs quadratic curve fitting on the remaining conductor outline point set. If the fitted conductor curve deviates significantly from the original outline points, the absolute residual sum will exceed the absolute residual sum threshold recorded during the installation and trial operation phase. At this time, step S2 does not directly output the sag value, but returns to the outline extraction stage for re-screening. The on-site result shows that the sag point positioning will not be directly distorted by the swaying of branches or light spot interference, and the sag monitoring results generated subsequently are also easier to reference in step S3.

[0056] S204: Construction of sag monitoring results.

[0057] After the actual sag value and sag offset corresponding to the current monitoring image are calculated, step S2 begins to construct the sag monitoring results. The sag monitoring results include at least the sampling time, the current operating level, the actual sag value, the sag offset, the lateral position of the sag point, the longitudinal position of the sag point, and the monitoring image frame identifier. The sampling time is used to align with the protection trigger mark in step S1. The current operating level is used to indicate under what task boundary the aforementioned sag monitoring results were generated. The actual sag value and sag offset are used for subsequent abnormal state identification. The lateral and longitudinal positions of the sag point are used for subsequent construction of the minimum evidence preservation fragment. The monitoring image frame identifier is used to reverse locate the original monitoring image in the information to be sent.

[0058] After the aforementioned sag monitoring results are generated, they are first written into the sag identification result cache area, and then transferred from the cache area to the pending information processing stage. Since step S3 needs to determine the minimum evidence retention fragment based on the sag monitoring results, step S2 does not delete any fields related to geometric position when constructing the sag monitoring results, nor does it remove the current running level from the results. The aforementioned processing allows subsequent steps to directly read the complete monitoring results without repeated calculations.

[0059] S205: Generation of pending information.

[0060] Step S2: After the sag monitoring results have been generated, the reported content is organized according to the field range allowed by the current operating level, and the organized reported content is written into the storage area to form pending information. There is a one-to-one correspondence between the pending information and the sag monitoring results. Specifically, each pending information consists of the current operating level, actual sag value, sag offset, lateral position of the sag point, longitudinal position of the sag point, and monitoring image frame identifier corresponding to the same sampling time. The current operating level indicates the power supply boundary when the pending information is generated, and the monitoring image frame identifier indicates which monitoring image frame the aforementioned pending information can be traced back to.

[0061] The information to be sent is only saved and not sent directly in step S2. The reason is that step S3 needs to further organize the information to be sent based on the continuous energy release interval, the net release energy after protection, the energy consumption for abnormal evidence fidelity, and the task recovery discretion coefficient, and finally form the sending content corresponding to the updated current operating level. Therefore, the information to be sent generated in step S2 must retain the information elements required for subsequent judgment, and cannot include the high-load content that can only be generated by the delayed recovery task set.

[0062] After step S2 is completed, the sag monitoring results and pending information have been stably generated from the monitoring retention task set. The front-end terminal obtains monitoring records that correspond one-to-one with the sampling time, monitoring image frame identifier and current operating level. The processing makes the sag changes no longer stay at the level of the original image, but is organized into a structured result that can be directly used for anomaly judgment and sending cropping, thereby improving the referenceability and traceability of the monitoring output.

[0063] After the sag monitoring results and pending information have been generated in step S2, whether the delayed recovery task set can be put into operation and to what extent it can be put into operation cannot be determined solely based on whether the protection trigger flag has been released. This is because the power supply segment after the protection is released may be short-lived, and evidence preservation may occupy additional power supply. Therefore, step S3 needs to put the continuous power release interval of the power supply side and the abnormal evidence requirements of the monitoring side into the same judgment chain. First, the available power supply is calculated, then the power supply required for evidence preservation is calculated, and finally, it is decided whether the recovery conditions are met.

[0064] S301: Continuous energy release range determined.

[0065] Step S3 first uses the lowest operating voltage of the monitoring retention task set in the corresponding table entries of the sampling time sequence, protection trigger mark, output voltage, and power supply status identifier from step S1. Taking the sampling time corresponding to the current sag monitoring result in step S2 as the current reference time, candidate time periods are searched segment by segment within all sampling sections. Any time period that satisfies the following conditions is recorded as a candidate time period: the protection trigger mark is in the triggered state at the previous sampling time, the protection trigger mark is in the non-triggered state at the current sampling time, and the output voltage of all sampling times is always not lower than the lowest operating voltage of the monitoring retention task set from the aforementioned transition time until the next time the protection trigger mark is in the triggered state. If multiple candidate time periods exist, the candidate time period covering the current reference time is selected first as the continuous energy release interval. If there is no candidate time period covering the current reference time, the candidate time period with the end time closest to the current reference time is selected as the continuous energy release interval. The purpose of this processing is to make the continuous energy release interval and the current sag monitoring result be in the same power supply stage, so that the subsequent power supply calculation and abnormal state judgment have a consistent time background.

[0066] Once the continuous energy release interval is determined, the start and end points of the continuous energy release interval are fixed. The energy supply of the subsequent energy release interval, the energy consumption for monitoring and maintenance, and the net energy supply after protection are all calculated using the aforementioned continuous energy release interval as the sole calculation range. The time boundary is no longer redefined, thus ensuring that the processing objects before and after remain consistent. All subsequent judgments in step S3 revolve around the same continuous energy release interval.

[0067] S302: Energy supply and monitoring / retention energy consumption within the energy release zone.

[0068] After the continuous energy release interval is determined, the energy supply within the energy release interval is calculated. Specifically, for each sampling segment between adjacent sampling times within the continuous energy release interval, the output power corresponding to the previous sampling time is read, and then the aforementioned output power is multiplied by the sampling interval duration corresponding to that sampling segment to obtain the segment energy supply within that sampling segment. Then, the energy supply of all segments within the continuous energy release interval is sequentially accumulated to obtain the energy supply of the energy release interval. The energy supply of the energy release interval represents the total energy actually released by the energy storage unit to the terminal within the continuous energy release interval, which is all the energy sources that can be called upon in subsequent recovery actions.

[0069] Step S3 then calculates the energy consumption for monitoring and retention. The energy consumption for monitoring and retention is primarily determined based on the sampling results of the power supply branches of the monitoring and retention task set. Specifically, the branch power of the power supply branches of the monitoring and retention task set is read within the continuous energy release interval. The branch power in each sampling segment is then multiplied by the corresponding sampling interval duration to obtain the segment maintenance energy consumption for that sampling segment. The maintenance energy consumption of all segments within the continuous energy release interval is then sequentially accumulated to obtain the monitoring and retention energy consumption. If there are missing measurements in the sampling results of the power supply branches of the monitoring and retention task set, the minimum maintenance power of the monitoring and retention task set corresponding to the power supply status identifier in the recovery sequence table is read. The minimum maintenance power of the monitoring and retention task set is then multiplied by the duration of the continuous energy release interval to obtain the monitoring and retention energy consumption under the alternative caliber. The physical meaning of the alternative caliber is that the minimum power supply requirement already written in the recovery sequence table approximates the basic maintenance energy consumption of the monitoring and retention task set within the continuous energy release interval.

[0070] After obtaining the energy supplied during the energy release interval and the energy consumed for monitoring and maintenance, the net releaseable energy after protection is obtained by subtracting the energy consumed for monitoring and maintenance from the energy supplied during the energy release interval. The net releaseable energy after protection represents how much energy the energy storage unit has remaining to support subsequent recovery actions within the continuous energy release interval, assuming that the monitoring and maintenance task set has been continuously operating. If the net releaseable energy after protection is negative, it means that the energy supply within the continuous energy release interval is insufficient to cover the maintenance needs of the monitoring and maintenance task set. In this case, even if the subsequent recovery decision continues, the delayed recovery task set will not be activated.

[0071] In one embodiment, the protection trigger flag changes from a triggered state to a non-triggered state for a period of time, while the monitoring image acquisition and sag recognition processing continue to run. Step S3 first identifies the time period as a continuous energy release interval, and then accumulates the actual energy output of the energy storage unit within the continuous energy release interval as the energy supply of the energy release interval. At the same time, the maintenance energy consumed by the four tasks of monitoring image acquisition, monitoring image preprocessing, conductor contour extraction, and sag recognition result caching within the same time period is accumulated as the monitoring retention maintenance energy consumption. If the former is only slightly higher than the latter, the net release energy after protection is less, and subsequent recovery actions will be significantly limited. The visible result on site is that the sag recognition processing continues to be maintained, while the high-load subsequent tasks are not immediately put into operation.

[0072] S303: Determination of current abnormal state and determination of minimum evidence retention fragment.

[0073] After obtaining the net salient energy after protection, the process moves to the monitoring side. The sag monitoring results output in step S2 are read, and the actual sag value, sag offset, and sag offset at the current sampling time are extracted. The sag offset at the current sampling time is subtracted from the sag offset at the previous sampling time, and then divided by the time difference between the two sampling times to obtain the sag offset change rate. The sag offset change rate indicates how quickly the conductor sag changes over time and is an important basis for identifying the current abnormal state. Subsequently, the actual sag value, sag offset, and sag offset change rate are used as index conditions to search for evidence template entries in the evidence template library one by one.

[0074] Each evidence template entry consists of anomaly state determination conditions and a minimum set of evidence retention fields. The anomaly state determination conditions define which sag monitoring results should be classified into the corresponding current anomaly state. The minimum set of evidence retention fields defines the minimum number of evidence fields that must be retained to ensure subsequent interpretation and backtracking under the corresponding current anomaly state. The fields must include at least the monitoring image frame identifier corresponding to the current sampling time, the actual sag value, the sag offset, the lateral position of the sag point, the longitudinal position of the sag point, and the current anomaly state identifier. Some evidence template entries also include the monitoring image frame identifier corresponding to the previous sampling time to represent the direction of change of the anomaly state. In step S3, when searching the evidence template library, any sag monitoring result that meets the abnormal state judgment condition of a certain evidence template entry is classified into the current abnormal state corresponding to that evidence template entry, and the minimum evidence retention fragment is determined by the minimum evidence retention fragment field set corresponding to that evidence template entry. If multiple evidence template entries meet the condition simultaneously, the number of fields is compared first, and the one with fewer fields is given priority. If the number of fields is the same, the one with the higher ranking is selected according to the preset priority order of the evidence template library. Thus, the minimum evidence retention fragment always comes from the unique evidence template entry that is matched in the evidence template library.

[0075] The boundaries for determining abnormal states in the evidence template library need to be predetermined. Specifically, this can be achieved by replaying historical samples from the installation and trial operation phase. During this phase, multiple sets of sag monitoring results under stable, gradually changing, and abruptly changing states are collected and labeled. Candidate boundaries are then substituted one by one into the evidence template entries for playback matching. The number of misjudgments for each set of candidate boundaries is counted, and the set with the fewest misjudgments is written into the corresponding evidence template entry. Through this process, the current boundaries for determining abnormal states do not rely on empirical guesses but rather originate from sag change samples that can be reproduced during the installation and trial operation phase.

[0076] S304: Energy consumption for obtaining abnormal evidence.

[0077] After determining the minimum evidence retention fragment, the energy consumption for maintaining the fidelity of anomalous evidence is calculated around this fragment. The energy consumption for maintaining the fidelity of anomalous evidence is obtained by sequentially adding the generation energy consumption, encapsulation energy consumption, and storage energy consumption. Generation energy consumption represents the energy consumed in extracting the minimum evidence retention fragment field set from the sag monitoring results and monitoring image frames, specifically obtained by multiplying the average power of the generation phase by the duration of the generation phase. Encapsulation energy consumption represents the energy consumed in organizing the minimum evidence retention fragment field set into a unified record format, specifically obtained by multiplying the average power of the encapsulation phase by the duration of the encapsulation phase. Storage energy consumption represents the energy consumed in writing the encapsulated record to the storage medium, specifically obtained by multiplying the average power of the storage phase by the duration of the storage phase. The sum of these three parts sequentially yields the energy consumption for maintaining the fidelity of anomalous evidence, representing the total energy required to completely retain the minimum evidence retention fragment under the current anomalous state.

[0078] All three power levels and three durations have clearly defined sources. The average power during the generation, packaging, and storage phases is obtained by the power management module sampling the corresponding power supply branches. The durations of the generation, packaging, and storage phases are obtained by subtracting the start and end timestamps of the corresponding processing phases. In other words, the energy consumption for anomaly evidence fidelity is not an abstract estimate, but is measured and accumulated segment by segment according to the complete processing process from extraction to writing of the minimum evidence fragment to the storage medium. Therefore, it has a clear engineering implementation path.

[0079] In one embodiment, the current abnormal state is identified as a continuous sag state by the evidence template library. The minimum evidence retention segment requires retaining the monitoring image frame identifiers of the current sampling time and the previous sampling time, the current actual sag value, the current sag offset, and the current sag point position. After extracting the aforementioned fields, the duration of the field extraction stage and the corresponding power supply branch power are recorded first. Then, the time when the aforementioned fields are encapsulated to form an evidence record and the corresponding power supply branch power are recorded. Finally, the time when the evidence record is written to the storage area and the corresponding power supply branch power are recorded. The energy consumption for abnormal evidence fidelity is obtained by summing the three results. The on-site result shows that the evidence corresponding to the current abnormal state can be retained without introducing redundant data.

[0080] S305: Dual-channel recovery assessment model reasoning and recovery condition determination.

[0081] After simultaneously obtaining the net releasable energy supply after protection and the energy consumption for maintaining the fidelity of anomalous evidence, the aforementioned two input quantities are used as inputs to the dual-channel recovery assessment model. The net releasable energy supply after protection represents the remaining energy available for recovery actions after the continuous operation of the monitoring and retention task set within the current continuous energy release interval. The energy consumption for maintaining the fidelity of anomalous evidence represents the total energy consumption required to generate, encapsulate, and store the minimum evidence retention fragment under the current anomalous state. Thus, the aforementioned two input quantities correspond to the energy supply margin constraint and the evidence retention constraint, respectively. The dual-channel recovery assessment model consists of a first multilayer sensor channel, a second multilayer sensor channel, and a fusion output channel. The first multilayer sensor channel receives the net releasable energy supply after protection, and the second multilayer sensor channel receives the energy consumption for maintaining the fidelity of anomalous evidence. The two channels output channel feature results respectively. These two sets of channel feature results are then concatenated in a preset order to form a fusion result, which is then input into the fusion output channel to obtain the task recovery discretion coefficient. The task recovery discretion coefficient represents the degree to which the delayed recovery task set has the conditions for recovery and the depth to which it can be put into operation under the combined effect of the current energy supply margin and the current anomalous evidence requirement.

[0082] The dual-channel recovery assessment model was constructed using offline training. To eliminate the impact of different dimensions and value ranges on the model's training stability, before entering the dual-channel recovery assessment model, the net releasable energy supply after protection and the energy consumption for anomalous evidence fidelity preservation were first subjected to interval normalization. Specifically, the minimum and maximum values ​​of their respective input quantities in the training samples were used as normalization boundaries to map the original inputs to the interval between 0 and 1. The normalized results were then input into the corresponding channels. In one embodiment, the first multilayer perceptron channel has two hidden layers, with 8 neurons in the first hidden layer and 4 neurons in the second hidden layer. The second multilayer perceptron channel is configured in the same way as the first multilayer perceptron channel. The fusion output channel has one hidden layer and one output layer, with 8 neurons in the hidden layer and 1 neuron in the output layer. The hidden layer uses a modified linear unit as the activation function, and the output layer uses a logistic function to limit the output to between 0 and 1. This ensures that the task recovery discretion coefficient has a clear continuous numerical range and facilitates subsequent comparison with the task recovery discretion coefficient threshold.

[0083] The training samples for the dual-channel recovery assessment model are derived from protection release segments collected during the installation and trial operation phase and the historical operation phase. Each training sample records at least the net releasable energy supplied after protection, the energy consumed for abnormal evidence fidelity, whether the protection trigger flag reverts to the trigger state after the delayed recovery task set is put into operation, whether the monitoring retention task set continues to run, and whether the minimum evidence retention segment is generated and stored. Among them, samples that simultaneously meet the conditions of the protection trigger flag not reverting to the trigger state, the monitoring retention task set continuing to run, and the minimum evidence retention segment being generated and stored are recorded as successful recovery samples, and the remaining samples are recorded as failed recovery samples. During training, successfully recovered samples are designated as the positive class, and unsuccessfully recovered samples are designated as the negative class, constructing a supervised training set. In one embodiment, all training samples are divided into training, validation, and test sets in a 7:2:1 ratio. The loss function is the binary cross-entropy loss function, and the Adam optimizer is used for parameter optimization. The initial learning rate is set to 0.001, the batch size is set to 32, and the maximum number of training epochs is set to 200. Training stops when the validation set loss no longer decreases within 10 consecutive epochs, and the model parameters are reverted to the lowest level of validation set loss. These model parameters are then written to the terminal parameter area for use in step S3. To reduce the impact of a significant imbalance in the number of samples between the two classes on the training results, the number of successfully recovered samples and unsuccessfully recovered samples can be balanced before training to keep the number of samples in each class within a predetermined ratio range.

[0084] To use the output of the dual-channel recovery assessment model for recovery condition determination, step S3 also requires pre-determining the task recovery discretion coefficient threshold. The task recovery discretion coefficient threshold is determined through historical sample playback. Specifically, all samples in the test set are first input into the trained dual-channel recovery assessment model to obtain the corresponding task recovery discretion coefficients. Then, the samples are sorted from low to high according to their task recovery discretion coefficients, and subsequent observation segments of each sample are played back sequentially. Samples that simultaneously meet the requirements of continuous operation of the monitoring and retention task set, the protection trigger flag not being switched back to the trigger state, and the minimum evidence retention segment being generated and stored are considered successfully recovered samples. The sample with the smallest task recovery discretion coefficient among all successfully recovered samples is taken as the task recovery discretion coefficient threshold. In one embodiment, if the smallest task recovery discretion coefficient among the successfully recovered samples after test set playback is 0.65, then 0.65 is written into the terminal parameter area as the task recovery discretion coefficient threshold.

[0085] The recovery conditions employ a dual-judgment system. The first judgment is that the task recovery discretion coefficient is not lower than the task recovery discretion coefficient threshold. The second judgment is that the remaining result after subtracting the energy consumed for abnormal evidence fidelity from the net releaseable energy after protection is not negative. When both judgments are met simultaneously, the power supply status indicator satisfies the recovery conditions corresponding to the recovery sequence table; when either judgment is not met, the power supply status indicator does not satisfy the recovery conditions corresponding to the recovery sequence table. The aforementioned dual-judgment system also incorporates the recovery suitability learned from the model and the actual energy supply constraints within the current continuous energy release interval, ensuring that the recovery conditions are consistent with both historical operating patterns and the current actual energy supply status. This provides a unified judgment criterion for the subsequent deployment of delayed recovery task sets.

[0086] S306: Delay recovery task set deployment, current run level update, and content organization.

[0087] After the power supply status indicator meets the restoration conditions, step S3 does not deploy the entire delayed restoration task set at once, but deploys them one by one according to the preset restoration order field in the restoration order table. The restoration order table pre-writes the incremental maintenance power and restoration order number for each task in the delayed restoration task set. Step S3 first multiplies the net releasable energy after protection by the task restoration discretion coefficient to obtain the allowable restoration energy. The aforementioned allowable restoration energy represents the upper limit of energy that can be used for restoration actions after being discretionated by the dual-channel restoration evaluation model based on the current energy reserve. Subsequently, using the restoration observation time as a unified observation scale, the incremental maintenance power of the delayed restoration task set is accumulated item by item from front to back according to the restoration order number, and then multiplied by the restoration observation time to obtain the cumulative restoration energy consumption for the corresponding number of items. Any task item whose cumulative restoration energy consumption is not higher than the allowable restoration energy consumption can be included in the deployment scope of this round.

[0088] The recovery observation duration also needs to have a clearly defined logic. The recovery observation duration can be determined during the historical sample playback stage. Specifically, for historical samples that have returned to the triggered state after protection has been lifted, the duration from the start of the delayed recovery task set to the point where the protection trigger mark returns to the triggered state is counted. The shortest duration that covers the vast majority of the re-triggered segments is written into the terminal parameter area as the recovery observation duration. Therefore, the aforementioned recovery observation duration is not arbitrarily given, but rather derived from the actual historical process of being disturbed again after protection has been lifted.

[0089] After determining the number of delayed recovery tasks available for deployment in this round, step S3 activates the corresponding tasks one by one according to the recovery sequence number, and writes the entry in the recovery sequence table corresponding to the aforementioned deployment depth as the updated current operating level. After the updated current operating level is formed, step S3 reads the set of sending fields corresponding to the updated current operating level in the recovery sequence table, and then trims the information to be sent based on the set of sending fields. All fields belonging to the set of sending fields are retained, and all fields not belonging to the set of sending fields are not included in the sending content. If the set of sending fields contains evidence fields, the minimum evidence retention fragment is written into the sending content. If the set of sending fields does not contain evidence fields, the sending content only retains the current operating level, sampling time, actual sag value, sag offset, and monitoring image frame identifier. Thus, the aforementioned sending content is constrained by both the updated current operating level and the minimum evidence retention fragment corresponding to the current abnormal state.

[0090] After step S3 is completed, the net releasable energy after protection, the current abnormal state, the minimum evidence retention fragment, the energy consumption for abnormal evidence fidelity, the task recovery discretion coefficient, the updated current operating level, and the sent content have formed a continuous correlation. The input of the delayed recovery task set no longer depends on a single state signal, but is based on the joint constraints of the energy supply margin and the evidence requirement. The aforementioned processing provides a clear discretionary basis for the recovery action and also ensures that the sent content is consistent with the current power supply conditions and the current abnormal state.

[0091] After the current operating level and the content to be sent have been determined in step S3, although the front-end terminal already has the sending object and sending boundary, the sending process itself still needs to be organized in combination with the link carrying capacity, the confirmation receipt result and the subsequent state reuse requirements. Otherwise, the sent content is easy to lose its order in the link fluctuation, and the current operating level cannot continue to be used in the next round of power supply status determination. Therefore, the focus of step S4 is to transform the sent content into a link sending action with order, state and write-back record.

[0092] S401: Determining the transmission boundary under the current run level constraints.

[0093] At the start of step S4, the current run level and the content to be sent output in step S3 are read first, and the set of sending fields that uniquely corresponds to the current run level is read from the recovery sequence table. The aforementioned set of sending fields consists of fixed field contents in the recovery sequence table, which are used to limit the range of data allowed to enter the communication module under the current run level. Step S4 then checks the sending content field by field. All fields that belong to the set of sending fields are retained in the sending sequence, and all fields that do not belong to the set of sending fields are removed from the sending sequence. Thus, the boundary of the aforementioned sending sequence is directly derived from the current run level, and is no longer temporarily determined by the communication module.

[0094] When the set of transmitted fields contains the minimum evidence retention fragment, step S4 divides the transmission sequence into a basic transmission segment and an evidence transmission segment. The basic transmission segment includes at least the sampling time, current operating level, actual sag value, sag offset, and monitoring image frame identifier. The evidence transmission segment includes all fields in the minimum evidence retention fragment except for the aforementioned basic fields. When the set of transmitted fields does not contain the minimum evidence retention fragment, the transmission sequence consists only of the basic transmission segment. The purpose of the aforementioned division is to ensure that the communication module prioritizes transmitting the most critical sag status when the link capacity is limited, and then transmits the evidence fields used for interpretation and backtracking, thereby maintaining consistency between the transmission order and the task priority of the sag monitoring terminal.

[0095] S402: Determine the length of the content fragment and calculate the number of fragments.

[0096] After the transmission sequence is constructed, step S4 counts the data volume of the basic transmission segment and the evidence transmission segment, and then sums them sequentially to obtain the total data volume of the transmission content. That is, first, the data volume occupied by all fields in the basic transmission segment is added together to obtain the basic transmission segment data volume; then, the data volume occupied by all fields in the evidence transmission segment is added together to obtain the evidence transmission segment data volume; finally, the aforementioned basic transmission segment data volume and evidence transmission segment data volume are added together to obtain the total data volume of the transmission content. If the set of transmission fields does not contain the minimum evidence retention fragment, the evidence transmission segment data volume is recorded as zero, and the total data volume of the transmission content is equal to the basic transmission segment data volume.

[0097] Subsequently, step S4 reads the net payload capacity of a single communication module chip. The net payload capacity of a single communication module chip is not an arbitrary value, but is obtained by subtracting the header length, chip sequence number length, total chip length, sampling time marker length, and cyclic redundancy check (CRC) code length from the total length that a single frame of the link protocol used by the communication module can carry. This result is written into the terminal parameter area during the installation and debugging phase. Step S4 divides the total amount of data to be sent by the net payload capacity of a single communication module chip and rounds up to obtain the number of fragments. Specifically, if the total amount of data to be sent is divisible by the net payload capacity of a single communication module chip, the number of fragments is equal to the integer value obtained by the division. If the total amount of data to be sent is not divisible by the net payload capacity of a single communication module chip, one more fragment is added to the integer value obtained by the division. This ensures that the number of fragments always satisfies the requirement that all sent content can be loaded into the fragment sequence. After the number of fragments is determined, step S4 then performs fragmentation on the transmission sequence in ascending order of fragment number. Each fragment is appended with a fragment number, total number of fragments, sampling time identifier and cyclic redundancy check code, so that the remote end can verify the fragments one by one.

[0098] S403: The basic transmission segment is sent first, and the evidence transmission segment is sent later.

[0099] After the fragment sequence is formed, step S4 does not allow the communication module to send the basic transmission segment and the evidence transmission segment simultaneously. Instead, it proceeds in the order of basic transmission segment priority followed by evidence transmission segment. Specifically, the communication module first reads the fragment sequence corresponding to the basic transmission segment, sends each fragment sequentially from smallest to largest according to its fragment number, and waits for the corresponding acknowledgment receipt after each fragment is sent. Only when all fragments of the basic transmission segment have completed acknowledgment receipt processing, and the transmission field set indeed contains the minimum evidence retention fragment, does the communication module continue to send the fragment sequence corresponding to the evidence transmission segment. The aforementioned sequential arrangement ensures that the basic sag status field always enters the link before the evidence field, which is beneficial for the remote end to receive the current line status first, and then receive the evidence details.

[0100] The confirmation waiting time needs to be predetermined. The determination of the confirmation waiting time can be achieved using link replay records from the installation and trial operation phase. Specifically, during the installation and trial operation phase, test fragment transmission is repeatedly executed, and the arrival time from transmission to receipt of the confirmation receipt for each test fragment is recorded. Then, the sample with the longest arrival time from the consecutive successful receipt samples is taken as the confirmation waiting time, and this confirmation waiting time is written into the terminal parameter area. During the operation phase, if a fragment receives a confirmation receipt within the confirmation waiting time, the fragment is recorded as a successful fragment. If a fragment does not receive a confirmation receipt within the confirmation waiting time, the fragment is recorded as an incomplete fragment, and the transmission of subsequent fragments in the current transmission sequence is immediately stopped. This stop action prevents the communication module from continuing to occupy the current link, leaving the untransmitted content to be processed at a later time.

[0101] In one embodiment, if the link near the tower experiences short-term fluctuations during strong winds, step S4 first splits the transmitted content into a basic transmission segment and an evidence transmission segment. Then, the communication module prioritizes sending the corresponding fragments of the basic transmission segment. Once the remote end receives the sampling time, current operating level, actual sag value, and sag offset, it can determine whether the line is in an abnormal state. Subsequently, the communication module continues to send the evidence transmission segment. If a fragment fails to return an acknowledgment receipt during the transmission of the basic transmission segment or the evidence transmission segment, step S4 immediately stops the transmission of subsequent fragments. The on-site result is that the successfully transmitted fragments remain valid, while the untransmitted fragments remain inside the terminal to wait for the next link recovery before being processed.

[0102] S404: Calculation of effective transmission rate of communication module and recording of transmission time.

[0103] Step S4 involves continuously recording the transmission status of the most recent consecutive successfully transmitted fragments during the communication module's transmission process. This data is used to calculate the effective transmission rate of the communication module. Specifically, the process involves first selecting the most recent consecutive successfully transmitted fragments as the statistical range, then sequentially accumulating the effective data volume of each successful fragment within this range to obtain the cumulative successful data volume. Simultaneously, the transmission time taken for each successful fragment within the statistical range from transmission to receipt of the acknowledgment is sequentially accumulated to obtain the cumulative transmission time. Finally, the cumulative successful data volume is divided by the cumulative transmission time to obtain the effective transmission rate of the communication module. The effective transmission rate of the communication module represents the data transmission capability of the communication module in the current link state, enabling it to actually complete the acknowledgment closure.

[0104] After obtaining the effective transmission rate of the communication module, the total amount of data to be sent is divided by the aforementioned effective transmission rate of the communication module to obtain the estimated transmission time for the current content. This estimated transmission time is then written into the transmission log. The estimated transmission time does not change the transmission content boundary determined in step S3, but rather serves as a link occupancy record for subsequent operation and maintenance analysis and as a rhythm reference for retransmission in step S4.

[0105] S405: Sending status write-back and pending information continuation boundary update.

[0106] After the communication module completes the transmission of the current segment, it writes back the transmission status based on the acknowledgment results. If all segments receive acknowledgment, the transmission status at the corresponding sampling time is recorded as completed transmission; if only some segments receive acknowledgment, the data volume of all successfully transmitted segments is first added up sequentially to obtain the amount of data transmitted, and then the amount of data transmitted is subtracted from the total amount of transmitted content to obtain the amount of data to be transmitted. Thus, the amount of data to be transmitted explicitly represents the part of the transmitted content at the current sampling time that has not yet been transmitted.

[0107] Subsequently, step S4 writes the sampling time, monitoring image frame identifier, transmission status, amount of data sent, and amount of data to be sent back to the pending information storage location. The sampling time and monitoring image frame identifier are used to establish a unique correspondence with the sag monitoring results in step S2. The transmission status indicates whether the aforementioned pending information has been successfully transmitted, the amount of data sent indicates the portion that has been confirmed as received by the remote end, and the amount of data to be sent indicates the portion that still needs to be transmitted. After processing, the pending information is no longer just the reported content saved in step S2, but a record of resumable transmission with link transmission progress information. When entering step S4 again in the next round, the communication module prioritizes reading pending information whose transmission status is incomplete and resumes transmission from the first fragment of the unfinished fragment, instead of re-transmitting from the first fragment.

[0108] S406: Level reference information generation and retrieval boundary solidification.

[0109] After the transmission status is written back, step S4 records the current operating level as level reference information. The level reference information includes at least the current operating level, sampling time, power supply status identifier, and transmission status. The current operating level indicates the task level the terminal is in after this round of recovery and transmission is completed; the sampling time indicates the time position at which the current operating level was formed; the power supply status identifier indicates the power supply background corresponding to the current operating level; and the transmission status indicates whether the transmission action corresponding to the current operating level has been completed. After the level reference information is written to the level reference area, it is not rewritten again by step S4, but is directly called in the next round of step S1.

[0110] To limit the effective scope of the level reference information, step S4 further determines the level reference retention duration. The determination of the level reference retention duration can be achieved using a historical state change record playback method. Specifically, during the installation and trial operation phase and the historical operation phase, the state maintenance duration between two adjacent power supply status indicator changes is continuously recorded. The state maintenance durations are then sorted from shortest to longest, and the state maintenance duration that covers the vast majority of state maintenance processes is selected and written into the terminal parameter area as the level reference retention duration. When generating a power supply status indicator again in step S1, the most recent level reference information is retrieved backward from the current time. If the time interval between the formation time of the level reference information and the current time does not exceed the level reference retention duration, the current operating level in the level reference information is used as the starting point for the recovery sequence table search. If it exceeds the level reference retention duration, the search restarts from the default starting point of the recovery sequence table. Through this process, the level reference information can both pass the previous round of recovery results to the next round of power supply status determination and prevent it from continuing to affect new judgments after a long period of time.

[0111] In one embodiment, after a link interruption, the sag monitoring terminal restores communication. In step S4, a transmission sequence is constructed based on the transmission field set corresponding to the current operating level. Then, the communication module transmits the basic transmission segment and the evidence transmission segment in segments. After receiving the acknowledgment, some segments write back the amount of data already transmitted and the amount of data to be transmitted to the transmission information. At the same time, the current operating level, sampling time, power supply status identifier, and transmission status are written into the level reference area. When the next round of step S1 restarts the generation of the power supply status identifier, if the time since the formation of the previous round of level reference information has not exceeded the level reference retention time, step S1 directly uses the current operating level recorded in the previous round as the starting point for the recovery sequence table search. The result on site is that the terminal will not start from the lowest level every time, but can continue to judge forward along the most recently established operating level.

[0112] After step S4 is completed, the content to be sent has entered the communication module according to the sending boundary corresponding to the current operating level, the sending status is written back to the information to be sent, and the current operating level is also fixed as the level reference information. The aforementioned processing enables the front-end terminal to not only send the monitoring results into the link in a hierarchical manner, but also to save the recovery level of this round as the basis for subsequent judgment, so that the sending action and the state memory have both traceability and reusability.

[0113] Specifically, the above are merely preferred embodiments of this application and are not intended to limit this application.

[0114] In the description of this specification, references to terms such as "an embodiment," "example," and "specific example" indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0115] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A method for optimizing the power supply of an energy-self-managed sag monitoring terminal, characterized in that, Including the following steps: S1: Obtain the protection trigger flag output by the power protection circuit and the power supply status information output by the energy storage unit, generate the current power supply status identifier of the terminal, input the power supply status identifier into the preset recovery sequence table, divide the monitoring retention task set and the delayed recovery task set according to the recovery sequence table, and determine the current operating level corresponding to the monitoring retention task set. S2: Start the monitoring retention task set according to the current operation level, perform sag identification processing to obtain sag monitoring results, and save the reporting content corresponding to the sag monitoring results as pending information; S3: Determine the continuous energy release interval based on the protection trigger mark, combine the power supply status information and the monitoring retention task set to obtain the net energy that can be released after protection, determine the minimum evidence retention segment based on the sag monitoring results and obtain the energy consumption for abnormal evidence fidelity, and input the two into the pre-trained dual-channel recovery evaluation model to obtain the task recovery discretion coefficient. S4: Determine whether the power supply status indicator meets the restoration conditions based on the task restoration discretion coefficient. If it does, put the delayed restoration task set into the restoration sequence table to update the current operating level. Organize the information to be sent based on the updated current operating level to obtain the sending content. Output the current operating level, sag monitoring results and sending content to the communication module, and record the current operating level as the level reference information.

2. The power supply optimization method for an energy-self-managed sag monitoring terminal according to claim 1, characterized in that, Step S1 includes: The process of generating the current power supply status identifier of the terminal includes collecting protection trigger markers and power supply status information at a fixed sampling period within the observation time window, calculating the duration of the protection non-triggered state within the observation time window, calculating the duration of the power supply status information satisfying the power supply requirements of each item in the recovery sequence table, and then comparing the duration with the preset criteria of each item to determine the power supply status identifier.

3. The power supply optimization method for an energy-self-managed sag monitoring terminal according to claim 2, characterized in that, Step S1 also includes: Each entry in the recovery sequence table corresponds to a unique monitoring reserved task set, delayed recovery task set, and current operating level. Each entry also records the minimum operating voltage and minimum maintenance power of the monitoring reserved task set. When no entry meets the criteria, the basic reserved entry in the recovery sequence table is read, and the content corresponding to the basic reserved entry is used as the power supply status identifier, monitoring reserved task set, and current operating level.

4. The power supply optimization method for an energy-self-managed sag monitoring terminal according to claim 1, characterized in that, Step S2 includes: Perform sag identification processing, including enabling monitoring image acquisition tasks, monitoring image preprocessing tasks, conductor contour extraction tasks, and sag identification result caching tasks in the monitoring retention task set according to the current operation level, acquiring monitoring images at the same sampling time as in step S1, and extracting calibration segments and left and right support points from the monitoring images to establish length conversion relationships and support baselines.

5. The power supply optimization method for an energy-self-managed sag monitoring terminal according to claim 4, characterized in that, Step S2 also includes: The sag recognition process also includes performing median filtering, grayscale enhancement, and edge extraction on the monitoring image in sequence, filtering out non-conductor contours to obtain a conductor contour point set, fitting a conductor contour curve based on the conductor contour point set, and obtaining the actual sag value based on the vertical distance between the lowest point of the conductor contour curve and the support baseline. The difference between the actual sag value and the reference sag value forms the sag offset.

6. The power supply optimization method for an energy-self-managed sag monitoring terminal according to claim 1, characterized in that, Step S3 includes: The continuous energy release interval is determined based on the protection trigger mark, including searching along the sampling time sequence for candidate time periods when the protection trigger mark changes from triggered to non-triggered and the output voltage is not lower than the minimum operating voltage of the monitoring retention task set, and using the candidate time periods containing the sampling time corresponding to the sag monitoring results as the continuous energy release interval; the net releasable energy supply after protection is obtained by subtracting the energy consumption of monitoring retention maintenance from the energy supply of the energy release interval within the continuous energy release interval.

7. The power supply optimization method for an energy-self-managed sag monitoring terminal according to claim 1, characterized in that, Step S3 also includes: The minimum evidence retention segment is determined based on the sag monitoring results. This includes matching each evidence template entry in the evidence template library with the actual sag value, sag offset, and changes in sag offset between adjacent sampling times to determine the current abnormal state, and reading the set of fields corresponding to the current abnormal state to form the minimum evidence retention segment. When multiple evidence template entries meet the requirements simultaneously, the evidence template entry with the fewest fields and the highest ranking in the evidence template library is selected.

8. The power supply optimization method for an energy-self-managed sag monitoring terminal according to claim 7, characterized in that, Step S3 also includes: The energy consumption for maintaining the fidelity of abnormal evidence is accumulated from the energy consumption for generating the minimum evidence retention fragment, the energy consumption for encapsulating the minimum evidence retention fragment, and the energy consumption for storing the minimum evidence retention fragment. The dual-channel recovery evaluation model receives the net releasable energy supply after protection and the energy consumption for maintaining the fidelity of abnormal evidence, and outputs the task recovery discretion coefficient. The task recovery discretion coefficient and the task recovery discretion coefficient threshold pre-written in the terminal parameter area are used together to determine the recovery conditions.

9. The power supply optimization method for an energy-self-managed sag monitoring terminal according to claim 1, characterized in that, Step S4 includes: The delayed recovery task set is put into operation sequentially according to the recovery sequence table to update the current operating level. This includes reading the recovery order and incremental maintenance power of each task in the delayed recovery task set recorded in the recovery sequence table, determining the allowable recovery supply energy based on the task recovery discretion coefficient and the net releasable supply energy after protection, and, under the constraint that the cumulative recovery energy consumption is not higher than the allowable recovery supply energy, putting the tasks in the delayed recovery task set into operation sequentially according to the recovery order to determine the updated current operating level.

10. The power supply optimization method for an energy-self-managed sag monitoring terminal according to claim 1, characterized in that, Step S4 also includes: The current operating level, sag monitoring results, and transmission content are output to the communication module. This includes constructing a transmission sequence according to the transmission field set corresponding to the current operating level, and performing segmented transmission and acknowledgment verification in the order of basic transmission segment priority over evidence transmission segment. After segmented transmission, the transmission status, the amount of data transmitted, and the amount of data to be transmitted are written back to the transmission information. At the same time, the current operating level, sampling time, power supply status identifier, and transmission status are written into the level reference information.

Images