Contactless electronic voltage regulating power transformer online control method

Abstract

The application discloses a contactless electronic voltage regulation power transformer online control method, relates to the power transformation regulation and control technical field, and comprises the following steps: collecting voltage, power flow, communication, tapping and health and the like in step one, calculating communication credibility level, group saturation level and short-term over-limit probability, generating non-oscillation proof, and forming a first type of context data package; in step two, a budget surplus is obtained based on the data package and a health state coefficient, a rising edge or a falling edge is judged, a target dead zone and a target residence are generated, an action priority label is given, a second type of context data package is formed; in step three, a curve token containing a continuous time length is issued according to the gate control, and two types of receipts are completed, or the tapping is implemented according to the target residence; in step four, the token and the receipt are recovered and solidified in sections, the health state coefficient and the budget are updated, and a fourth type of context data package is formed; the method can inhibit mutual pursuit, reduce tapping actions, and improve consistency and traceability.

Description

Technical Field

[0001] This invention relates to the field of power transformer regulation technology, specifically to an online control method for contactless electronic voltage regulating power transformers. Background Technology

[0002] With the increasing proportion of distributed photovoltaic and other power sources connected to distribution substations, the voltage at the substation terminals fluctuates rapidly and the bus voltage is prone to rise during midday and periods of alternating cloud cover. Many grid-connected inverters default to using voltage-reactive power and voltage-active power limiting curves for local support. Simultaneously, main transformers or distribution transformers are equipped with on-load tap changers or electronic voltage regulators to maintain voltage levels. Furthermore, conditions such as aggregated control, remote parameter transmission and feedback, and communication link delays and packet loss are common on-site, leading to prominent coupling characteristics of fast-acting inverters, slow-acting tap changers, and multi-source uncertain communication within the same feeder.

[0003] The existing coordination approach mainly involves: setting fixed daily upper limits, fixed dead zones, and minimum dwell times to reduce tapping operations; issuing curves to inverters in batches or manually adjusting them at fixed intervals; and introducing offline optimization for some systems, but still using static thresholds and experience-based tuning during operation.

[0004] This approach exposes multiple flaws and mechanisms in high-penetration scenarios: First, a mismatch between time scale and sensitivity. Local inverters change reactive power or limit power generation within seconds, while tap changers respond within tens of seconds to minutes. Without dynamic gating and boundary verification, they chase each other, resulting in: photovoltaic upscaling → inverter absorbing reactive power → voltage slightly decreasing → tap downscaling → voltage dropping again → inverter withdrawing reactive power → voltage rising again, repeating this cycle and forming a limit loop. Second, communication and equipment heterogeneity leads to policy fragmentation. Remote transmissions often have partial success, inconsistent latency, or missing feedback, causing different inverters to execute different slopes or limits, resulting in voltage stratification and local over-limits. Third, the lack of timely management and verifiable rollback for temporary parameters leads to long-term stagnation of enhancement curves and difficulty in timely correction of baseline deviations. Fourth, the health status of tap changers is not included in the control premise; frequent tapping accelerates contact wear and paper aging, increasing maintenance risks. Fifth, the lack of runtime stability domain verification means that once the parameter combination crosses the safety boundary, oscillations and over-limits will be amplified.

[0005] The combination of these problems leads to a decrease in the pass rate, an increase in the number of actions, an extension of the time limit for issuing actions, and a deterioration in the user experience. Furthermore, it makes it difficult to establish an auditable and traceable operational trajectory. Therefore, the current technical challenge is that in power distribution sites with high-penetration distributed power sources, uncertain communication quality, and significant differences in the response characteristics of inverters and tap changers, the existing coordination methods based on fixed thresholds and periodic issuance cannot identify and avoid triggering conditions for mutual chasing during operation. They also cannot simultaneously constrain and verify parameter boundaries in real time while considering equipment health, communication, and the consistency of group execution, resulting in frequent oscillations and exceeding limits, uncontrolled tap changer actions, and long-term baseline deviations. Summary of the Invention

[0006] (a) Technical problems to be solved

[0007] To address the shortcomings of existing technologies, this invention provides an online control method for contactless electronic voltage regulating power transformers. Based on data packets and health status coefficients, a budget margin is obtained; rising or falling edges are determined; a target dead zone and target dwell time are generated, and an action priority label is provided, forming a second type of context data packet. Curve tokens with durations are issued according to gating and two types of receipts are completed, or tapping is implemented based on target dwell time. Tokens and receipts are collected and solidified in segments, and the health status coefficients and budget are updated, forming a fourth type of context data packet. This method suppresses mutual chasing, reduces tapping actions, and improves consistency and traceability; it solves the technical problems described in the background art.

[0008] (II) Technical Solution

[0009] To achieve the above objectives, the present invention is implemented through the following technical solution: an online control method for contactless electronic voltage regulating power transformers, comprising: collecting three-phase and terminal voltages and power flow, inverter communication and curves, tap status and equipment health; calculating communication reliability level, population saturation level and short-term over-limit probability, generating non-oscillation proof, and forming a first type of context data packet;

[0010] Based on the first type of context data packet and device health, calculate the budget margin and determine whether it is a rising edge or a falling edge; generate the target dead zone and target dwell according to the coupling mapping, verify that it meets the non-oscillation proof, output the second type of context data packet and mark the action priority label;

[0011] Based on budget margin, short-term over-limit probability, communication reliability level, group saturation level, and non-oscillation proof gating; when the conditions are met, a curve token is issued to the inverter group and two types of receipts are completed; otherwise, the deconnection is implemented according to the target residency; a third type of context data packet is formed.

[0012] When the voltage is qualified or the token expires, it is recycled in segments in the order of reducing the reactive power slope, relaxing the active power limit, and restoring the normal curve. The non-oscillation certificate is checked at the beginning and end of each segment. The token, receipt and curve trajectory are fixed, the health status coefficient and daily and hourly budgets are updated, and a fourth type of context data packet is formed.

[0013] Furthermore, the communication reliability level is determined by a combination of the online ratio, the receipt success rate, and the high quantile of the delay. The group saturation level is the maximum value of the high quantile of the reactive power occupancy ratio and the high quantile of the active power limited generation ratio. The short-term over-limit probability is generated by mapping the voltage relative deviation, the photovoltaic output change rate, and the neighborhood similarity term, and is written into the first type of context data packet along with the non-oscillation proof.

[0014] Furthermore, the neighborhood similarity term is calculated based on the directed increment of the end voltage of adjacent feeders at the same station according to the kernel weight, and the statistical window is aligned with the scheduling cycle; for missing measurements, the previous valid value is used with a safety margin, unreliable measurement points are blacklisted and downweighted, and non-oscillation proof is recorded by marking as passed or failed.

[0015] Furthermore, the budget margin is obtained by scaling the daily budget and hourly budget with a health status coefficient, and the rising or falling edge is determined within the statistical window; based on this, the target dead zone and target dwell time are generated, upper and lower bounds are set respectively, and the non-oscillation proof is used for verification, and the second type of context data packet and action priority label are output.

[0016] Furthermore, the action priority label is determined by the budget margin and the short-term over-limit probability. When the budget margin is low and the short-term over-limit probability is high, it is marked as a priority inverter; otherwise, it is marked as an available tap. The target dwell time is taken to be longer with the falling edge and shorter with the rising edge, and is coupled with the feeder electrical time constant to maintain consistency in a single cycle.

[0017] Furthermore, a trigger score is formed by combining budget margin, short-term over-limit probability, communication credibility level and group saturation level and compared with the threshold, with non-oscillation proof as a necessary condition; when the condition is met, a curve token is generated, which includes curve number, version, duration, recycling strategy number and signature. Two types of receipts are included: reception receipt and behavior receipt.

[0018] Furthermore, when the received receipts are incomplete or the communication trust level is below the threshold, curve tokens are issued in batches according to the reachable subgroup selection ratio; when the issuance window ends and the receipts are still incomplete, new tokens are suspended, and the on-loaded tapping device performs single or multiple steps according to the target's residency, and the tapping action is recorded with the third type of context data packet.

[0019] Furthermore, the recovery trigger is jointly satisfied by the consistency measurement of receipts and the proportion of continuous qualified duration, and the non-oscillation proof is maintained; the segmented recovery order is to reduce the reactive power slope, relax the active power limit, and restore the normal curve. The first segment is aligned with the timestamps of the two types of receipts, and the last segment is reviewed for the short-term over-limit probability and the non-oscillation proof.

[0020] Furthermore, the curve token signature, the two types of receipt timing, and the curve number trajectory are solidified into audit evidence items, and an audit evidence strength index is generated and written into the fourth type of context data packet; abnormal items and equipment lists are listed separately for subsequent review and retrieval.

[0021] Furthermore, the tapping action accounting and operating temperature rise in the third type of context data packet are mapped to health status coefficient updates, and the daily and hourly budgets for the next cycle are scaled down proportionally; the updated health status coefficients and budgets are returned with the fourth type of context data packet.

[0022] (III) Beneficial Effects

[0023] This invention provides an online control method for contactless electronic voltage regulating power transformers, which has the following beneficial effects:

[0024] Based on sensing, prediction, and non-oscillation proof, measurements such as voltage and power flow, communication and curves, tap and health are integrated into a first-class context data packet, unifying the judgment criteria, eliminating unreliable communication and group saturation scenarios in advance, reducing invalid commands, and providing a prerequisite for subsequent mapping and gating.

[0025] The health status coefficient is set forward as a budget margin, and the target dead zone and target dwell are generated by linking the rising and falling edges respectively. These are then bound to the action priority label to form a trend-oriented threshold and waiting time configuration, reducing the frequent operation of the on-load tap changer and ensuring that the inverter intervention and tap changer actions have a sequential order and are consistent within the same cycle.

[0026] The parameter boundaries are established by the number, version and duration of the curve token, and the execution loop is formed by the double confirmation of the received receipt and the action receipt. The tokens are distributed in batches according to the proportion of reachable subgroups, and local maintenance is maintained when communication degrades. When the conditions are not met, the target dwell is switched to the on-load tapping device and the tapping action is written into the maintenance ledger.

[0027] A fixed-sequence segmented recovery method is adopted, first reducing the reactive power slope, then relaxing the active power limit, and finally restoring the normal curve. The non-oscillation proof and short-term over-limit probability are checked at the beginning and end of each segment to prevent pullback jumps and temporary curve remnants, so that the system can smoothly transition from the enhanced state to the normal configuration and maintain the continuity and consistency of the configuration.

[0028] The curve token signature, two types of receipt timing and curve number trajectory are embedded in the third and fourth type context data packets to generate audit evidence strength indicators, realize integrated evidence storage of control instructions, equipment behavior and configuration trajectory, support compliance inspection, fault tracing and cross-cycle comparison, and ensure that policies and records correspond. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the online control method for contactless electronic voltage regulating power transformers according to the present invention. Detailed Implementation

[0030] 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.

[0031] Please see Figure 1 This invention provides an online control method for contactless electronic voltage regulating power transformers, including:

[0032] Step 1: Based on unified data collection, and considering four dimensions—communication trustworthiness, group saturation, limit-breaking tendency, and non-oscillation proof—a communication trustworthiness level is generated through rule-based calculation and consistency verification. Group saturation level Short-term over-limit probability The first type of context data packet with stability certificate tags This serves as the only valid input for the second step.

[0033] When a high proportion of distributed power sources are connected, the quality of measurement, communication, and feedback directly determines whether subsequent curve distribution and tap scheduling will be effective on the same time scale. If the communication link is unstable or the population is already under high occupancy, even if curves are distributed, it will be difficult to obtain a consistent response. Therefore, establishing measurable, comparable, and transferable hierarchical quantities for communication reliability and population saturation is a prerequisite for forming reliable input.

[0034] Using the in-station edge controller as a unified entry point, the system collects online ratio, acknowledgment success rate, and 95th percentile latency data. It also collects reactive power occupancy ratio sequences and active power limiting ratio sequences in parallel, and obtains the communication reliability level through rule-based mapping. With group saturation level This data, along with snapshots of the bus and terminal voltages, is then incorporated into subsequent calculations. This logic prioritizes reachability and availability, ensuring that all subsequent decisions are based on feasibility.

[0035] First, using online ratio, receipt success rate, and high-quantile latency as independent variables, a monotonically saturated synthetic mapping is constructed to determine the communication reliability level. The gain relationship is positive for more online users, higher response times, and shorter latency. To eliminate the dimensional differences, an exponentially decaying kernel is used for latency, reducing the impact of large latency on performance. The impact is contained within a controllable range. Specifically:

[0036]

[0037] Where: Online ratio The percentage of online devices during the statistical period. Success rate of receipts : The percentage of parameters that have been received and started. , used to characterize the degree of command closure achievement;

[0038] Delay quantile The highest percentile of the time from issuance to confirmation. To suppress long-tail latency, within the same statistical window, the 95th percentile of the single round-trip latency sample from issuing the token to receiving the receipt is taken.

[0039] Weight : Non-negative and satisfy Used to balance the contributions of the three indicators; time scale parameter Positive numbers map second-level latency to the feature scale of a dimensionless decay kernel; shape parameters Not less than It is used to adjust the intensity of the penalty for long-tail latency.

[0040] In practice, the introduction of an exponentially decaying kernel prevents long-tail latency from being linearly amplified and disturbing the overall evaluation; the synthetic mapping maintains a consistent response to the three metrics, facilitating the subsequent setting of explicit thresholds; and the communication reliability level... Constrained This facilitates comparison and transmission across feeders and time periods.

[0041] Within the same statistical period, high-quantile representative values ​​are extracted from the reactive power occupancy ratio sample series and the active power limited generation ratio sample series, respectively. The population saturation level is obtained by conservatively combining the higher and lower quantile values. It can sensitively capture scenes where most devices are approaching their limits without relying on complex learning methods. Specifically:

[0042]

[0043] Where: Reactive power occupancy ratio sample sequence Each device The sequence of samples, the range of values ​​for a single sample ; Active power limited generation ratio sample sequence Limited availability per device The sequence of samples, the range of values ​​for a single sample This reflects the degree of triggering of voltage-active power limiting;

[0044] quantile function Returns the 95th percentile value of the input sample. Used to represent the high-occupancy layer in the group; the 95th percentile of the ordered statistic of the sample within the statistical window is taken; the sample comes from the reactive power occupancy ratio sample sequence of each inverter in the current period. Active power limited generation ratio sample sequence Maximum operator The larger of the two high quantiles is taken to form a consistent conservative representative value.

[0045] In practice, the high quantile, rather than the mean, represents the group occupancy status to prevent local overload from being diluted by average; the maximum operator raises the conservative boundary to ensure that alerting is triggered when either of the two types of constraints approaches its upper limit; group saturation level. Also bound by This facilitates communication trust level Compare and link within the same coordinate system.

[0046] Whether or not instantaneous voltage exceeds the limit is insufficient to support the setting of thresholds and priorities. A description of the tendency to exceed the limit in the short term must be provided within the same statistical period. At the same time, in order to avoid round-tripping actions caused by subsequent curves or tap changes, it is necessary to complete the proof and conservative treatment of whether oscillations may be induced within the same period.

[0047] First, a monotonically saturated over-limit tendency function is constructed using the relative voltage deviation, the photovoltaic power output change rate, and the neighborhood similarity term to obtain the short-term over-limit probability. Then, an inequality-based stability criterion is constructed using the slope of the voltage-reactive power curve, the voltage dead zone, and the minimum residence time. This results in a certificate indicating whether the test was passed or failed; if the test fails, the dead zone and dwell time are immediately pre-adjusted to ensure that potential oscillations are suppressed at the source.

[0048] Based on continuous measurements, a weighted term is defined for voltage relative deviation, photovoltaic output change rate, and neighborhood similarity. These three factors are linearly synthesized into a trend quantity, which is then compressed using logical mapping. This makes it compatible with communication trust levels. With group saturation level Comparison within the same domain. Specifically, this is expressed as:

[0049]

[0050] Where: relative voltage deviation : A dimensionless quantity obtained by dividing the difference between the terminal voltage and the upper limit voltage by the upper limit voltage; a positive value indicates an upward trend; photovoltaic output change rate. : The rate of change of active power output per unit time within the statistical period; a positive value represents the upward slope.

[0051] Neighborhood similarity weighting : A scalar value weighted by kernels, representing the voltage change patterns of neighboring times or neighboring transformer areas, used to incorporate collaborative information based on spatial or temporal proximity; weight : Real numbers, satisfying Logical mapping A monotonically increasing function that saturates at both ends, compressing the real number line to an interval. .

[0052] In use, the three independent variables are incorporated into the synthesis with clear physical meanings to avoid a black box approach; the logical mapping enables short-term out-of-limit probabilities. Communication Trustworthiness Level With group saturation level Same-domain representation facilitates gating, and also reduces the probability of short-term limit violations. The increase is intuitive and explainable, as the voltage rises, the output increases, and the neighborhood resonance strengthens.

[0053] Furthermore, considering that steeper voltage-reactive power curves, narrower dead zones, and shorter dwell times are more likely to trigger round-trip movements, these three factors are selected as core criteria. Combined with the electrical time constant of the feeder, a ratio-based index is formed to determine whether there is an oscillation risk within the same statistical period, and a stability exceedance limit is defined. Specifically, it can be expressed as:

[0054]

[0055]

[0056] Where: slope of the voltage-reactive power curve The response slope of reactive power change to voltage change is used to represent the response slope. A larger value indicates a more intense response; voltage dead zone width : Range of values The smaller the value, the more frequent the intervention; minimum dwell time The shortest interval between two adjacent actions. ;

[0057] Feeder electrical time constant : A scalar obtained by fitting the voltage recovery process under a step disturbance. During the initial commissioning phase or periodic verification, a small step disturbance is performed on the substation (such as switching from fixed reactive power to mild reactive power support), and the voltage recovery sequence of the bus or key nodes is recorded. Using the first-order exponential recovery model The least squares fit yields and in Leave a trace.

[0058] Voltage-Active Power Limiting Curve Slope The response slope of the active power limiting to voltage changes is expressed as the slope. Suppression coefficient Dimensionless weight, range of values This is used to characterize the linear correction of active power limiting to the overall oscillation risk; threshold constant. The upper bound after regularization is used to determine whether a pass or fail is achieved.

[0059] When in use, when the slope of the voltage-reactive power curve is... Larger voltage dead zone width Smaller, minimum dwell time When it is too short, Ascending and prone to exceeding limits, the system enters a conservative pre-adjustment state to avoid triggering round-trip movements; introducing... Link dwell time with feeder inertia to avoid mechanically applying fixed dwell time; consider the slope of the voltage-active power limiting curve. The linear correction brings the side effects of active power limiting into a unified criterion, making the overall approach more in line with engineering intuition.

[0060] As an example, a distribution area containing multiple photovoltaic (PV) units experiences frequent cloud cover switching in the afternoon. After reading the list of online devices, the on-site edge controller calculates the online ratio and acknowledgment success rate, and also tracks the arrival and confirmation times of commands from the previous time period to obtain... Approaching a high level. Simultaneously, the controller collects the current ratio of reactive power utilization to active power generation from each inverter, extracts the high quantile, and synthesizes it into a... This indicates that most devices still have room for adjustment.

[0061] Subsequently, the system synthesizes the trend rate by combining the rise slope of the terminal voltage and the rise slope of the photovoltaic active power, and obtains it through logical mapping. Too high. Next, calculate using the current curve slope, dead zone, dwell time, and feeder time constant. The result was within acceptable limits, and a certificate approval marker was generated. At the end of the cycle, the controller assembled the certificate marker. It also includes a snapshot of the bus and end voltages as the sole input for the next step.

[0062] When used, the tendency to exceed limits is fixed as an interval quantity. Stability is fixed as an inequality index. Both and and Parallel entry into the same data packet provides a clear basis for subsequent threshold and priority adjustments; when stability criteria... If the process fails, a conservative pre-adjustment of the dead zone and the stationary phase is immediately triggered to prevent the risk from being passed on to the next stage.

[0063] Step 2: Under the unified certificate mark Calculate the budget surplus under the given input conditions. And based on this, the target dead zone is generated. Target stationing Action priority label Furthermore, rising and falling edges are treated as two mutually exclusive cases, resulting in the output of the second type of context data packet. It provides a unique and verifiable basis for subsequent token issuance or call scheduling.

[0064] After multiple operations, contact wear and mechanical fatigue of on-load tap changers will significantly increase. If a fixed daily limit or fixed penalty factor is still applied, more reciprocating operations are often triggered during periods of high volatility. Therefore, it is necessary to directly incorporate the health status into the budget calculation for the same cycle, and convert the judgment of whether the device can handle one tap changer operation into a continuous budget margin within that cycle. Then The trend judgment results are uniformly transmitted to the generation logic of dead zones and dwell times.

[0065] The specific processing logic is as follows: Step a1, Read the certificate tag. And lock in the short-term probability of exceeding the limit. With the bus-end voltage snapshot, ensure consistency with the variables of the previous period; Step a2, read the health status coefficient. and the Japanese budget Hourly budget Daily remaining budget Hourly Remaining Budget Merge the results to obtain the budget surplus. Step a3: Based on the time sequence changes of the voltage snapshot, generate trend indicator variables (rising edge indicator and falling edge indicator) to lay the foundation for subsequent dual-edge mapping; Step a4: Form a triplet of budget margin-trend-stability label.

[0066] Within this cycle, the daily-hour budget and health status coefficient are merged into a single interval variable, which directly serves as the core scaling variable for dead zones and residency. Specifically, this is expressed as follows:

[0067]

[0068] Where: Budget surplus : Range of values , The smaller the value, the less likely it is to trigger a breakout; daily remaining budget : Range of values This indicates the number of taps still available today; daily budget. : Range of values This indicates the total number of runs allowed today; remaining hourly budget. : Range of values This indicates the number of taps currently available in the small fashion app;

[0069] Hourly budget : Range of values This indicates the total number of runs allowed in the current hour; health status coefficient. : Range of values The output is from a health assessment; the higher the value, the worse the health, and it is used to reduce the budget.

[0070] Among them, calculation at the beginning of the cycle And lock it as a common scaling variable for dead zone and dwell time within this period, while also including the budget margin. With interval quantity The second type of context data packet draft is stored to ensure traceability across stages.

[0071] When in use, lifespan / maintenance constraints are directly quantified into budget margins. To avoid the passive situation of fixed daily limits in highly volatile scenarios; the same variable Simultaneously drive dead zone and dwell time, ensuring that both adjust in the same direction and with proportional magnitude; using the health status coefficient The constraint of reducing the number of movements when the equipment is fatigued is implemented from the rule layer to the numerical layer.

[0072] Furthermore, within the same period, the directed difference of the end-voltage snapshot sequence is used as the trend indicator: if the latest snapshot shows a significant upward movement relative to the previous snapshot, the rising edge indicator is marked as true, and the falling edge indicator as false; conversely, the opposite is also true. Based on the trend indicator, budget margin is considered. Short-term over-limit probability Generate action priority tags When the budget surplus Low and short-term probability of exceeding limits When the priority is high, label the action priority. Marked as a priority inverter; when High or When low, Marked as available for dispatch. To avoid random conflicts arising from multi-source conditions, trend indicators are only used for directional mapping. The generation of and With the core as the core.

[0073] In use, the direction information of voltage rise / fall and the budget information of whether there is a margin of error are synthesized in layers to avoid left and right swinging caused by simple threshold triggering; action priority label Maintaining stability within this cycle provides certainty for subsequent token or branch selection; trend indicators only affect the mapping direction and do not directly trigger control, reducing the disturbance of the execution surface caused by misjudgments.

[0074] As an example, in the afternoon, the voltage at the end of the feeder showed a slow rise. The station controller read... Among them, the probability of short-term over-limit Too high, stability criterion quantity The test was successful. Subsequently, the controller retrieved the health status coefficient from the health assessment. Then, combine the remaining budget for the day and the hour to calculate the budget surplus. The controller compares the two end voltage snapshots, determines it to be a rising edge, and then generates an operation priority tag for the priority inverter. At the end of this cycle, the controller carries... Trend indicators and Proceed to the next steps.

[0075] Furthermore, in scenarios prone to mutual chasing, if the handling rules for voltage rise and fall are not differentiated, the dead zone and dwell time on both sides will decrease simultaneously, thus triggering round trips. Therefore, it is necessary to classify rising and falling edges into two mutually exclusive cases, and to handle the target dead zone separately within the same cycle. Staying with the target Perform directional scaling; at the same time, this scaling must be in line with the budget margin. Short-term over-limit probability Coupled in the same direction and satisfying the stability criterion at any time. The constraints. The specific processing method includes the following steps: Step a1, link the rising edge / falling edge indicator with the budget margin. Short-term over-limit probability Combine to generate the target dead zone Scaling factor; Step a2, generate target dwell time in the same combination. The scaling factor is adjusted, but a stronger delay characteristic is introduced in the falling edge direction, which naturally reduces the operating frequency; step a3, respectively... and Perform cross-boundary interception and Perform parallel checks to ensure that the stability inequality is satisfied at the generation end; step a4, ... , Action priority label Packed as a Type 2 context data packet .

[0076] During this period, based on the basic dead zone As a baseline, combined with rising / falling edge indication and budget margin Short-term over-limit probability The target dead zone is generated, specifically expressed as follows:

[0077]

[0078] Where: target dead zone : Used to determine whether a gear shift or curve intervention is triggered on the execution plane; basic dead zone : The value is determined on-site or recommended by specifications; rising edge indicator. : Values or The rising edge is Otherwise Falling edge indicator : Values or The falling edge is Otherwise ,and Mutually exclusive;

[0079] Directional coefficient : Range of values Determine the magnitude of the narrowing at the rising edge and the widening at the falling edge; budget the coupling coefficient. : Determine the budget surplus Amplification intensity of dead zone; probabilistic coupling coefficient : Range of values Determine the short-term probability of exceeding the limit. The amplification intensity of the dead zone.

[0080] Use direction correction amount:

[0081]

[0082] Unify the notation of probabilistic coupling to the direction:

[0083]

[0084]

[0085] When the rising edge ( )and High time, , Automatic tightening and ;

[0086] When the falling edge ( )and When the height is high, the width will automatically increase, and the direction will be consistent.

[0087] When rising edge indicates At that time, direction factor Promote the target dead zone Shrink; when the falling edge indicates At that time, direction factor Promote the target dead zone Relaxing the budget margin and increasing the probability result in a larger multiplicative factor, thus making the dead zone larger overall in tense-high-risk scenarios to suppress action frequency. Then, the target dead zone is... With stability criterion Parallel verification is performed, and if any boundary of inequality is found, the amplitude is limited.

[0088] In practice, direction, budget, and probability are multiplicatively linked, ensuring that changes in the dead zone exhibit symmetrical but opposite-directional responses in both types of trends; when the budget margin... Small, short-term probability of exceeding limits The system automatically raises the dead zone when the temperature is high, reducing the probability of crossing the dead zone and thus reducing round-trip movements; this is related to the stability criterion. Parallel verification ensures that the generation of dead zones naturally conforms to stability conditions, thus avoiding the transmission of instability to the execution surface.

[0089] During this period, based on basic residency As a baseline, combined with rising / falling edge indication and budget margin Short-term over-limit probability Generate target dwell time and introduce feeder electrical time constant. This creates a proportional hysteresis to avoid the inconsistency of an enlarged dead zone coupled with an excessively short dwell time. Specifically, this means:

[0090]

[0091] In the formula: target dwell time : Used to constrain the shortest interval between adjacent actions; basic dwell. : Provided by on-site adjustment or specification recommendations; rising edge indicator Falling edge indicator Same as the aforementioned definition and values; direction coefficient : Range of values Determine the magnitude of the shortening of the rising edge and the lengthening of the falling edge; budget the coupling coefficient. : Determine the budget surplus Strength of dwell time; probabilistic coupling coefficient : Range of values Determine the short-term probability of exceeding the limit. Tensile strength of the stationary position; feeder electrical time constant : Range of values This reflects the inertial characteristics of voltage recovery; regularity. : , used to avoid The proportion is too large when the value is extremely small.

[0092] When the rising edge indicates At that time, the target will remain stationed according to Slightly shorten the time to allow the inverter to engage first during the rise phase; when the falling edge indicates... At that time, the target will remain stationed according to Extending the timeframe creates a natural cooling window; the smaller the budget margin and the higher the probability, the longer the dwell time is extended; finally, the proportional item... Steady-state compression is applied to the entire system to prevent the fast electrical system from being slowed down by excessively long dwell times.

[0093] When the target dead zone Target stationing and Generate and pass stability criterion quantity After review, including the period starting point Encapsulated together with trend indicators as a type 2 context data packet This serves as the sole input for the next step, either token issuance or de-scheduling.

[0094] In use, the direction, budget, probability, and time constant are bound multiplicatively to ensure that the dwell adjustment and dead zone are coupled in the same direction without conflict; the proportional term suppresses the side effect of excessive stretching of the ultrafast system, allowing for lateral alignment between feeders; and it is linked with the generated action priority labels. The matching process ultimately favors the inverter during the upward phase and suppresses back-and-forth movements during the downward phase.

[0095] Step 3: Within the same period, use the budget surplus. Short-term over-limit probability Communication Trustworthiness Level Group saturation level With stability criterion Based on the established gating criteria, the priority inverter or the on-load tap changer is determined, and a duration-sensitive variable inverter is generated in the former path. The curved token completes the dual confirmation of receiving and action receipts; under the latter path, it follows the target dwell. Perform call splitting and lifetime accounting; if communication degrades or acknowledgments are incomplete, perform subgroup delivery and degradation coordination to ensure scenario consistency and security.

[0096] Among these factors, relying solely on a single threshold cannot simultaneously address accessibility, availability, tendency to exceed limits, and stability; therefore, budget margins need to be considered. Short-term over-limit probability Risk characterization and communication trust level With group saturation level The feasibility of these parameters is synthesized on the same scale, thus providing an explicit triggering score for the priority inverter. After triggering, the intensity and duration of the curve shape must be considered. The token is fixed together with the recycling strategy number, so that the behavior of the execution end corresponds one-to-one with the upstream criteria and is traceable.

[0097] First, at the start point of this cycle, read... and ,locking , , , and Based on this, the trigger score is calculated. When the rating reaches the threshold and If the display passes, the system enters the priority inverter path and generates a curve token; otherwise, it directly enters the on-load tap changer path. Token generation is based on the target dead zone. Staying with the target The curve enhancement coefficient is given to ensure that the curve and the upstream criterion are coupled in the same direction in terms of strength; the token carries the curve number, version, and duration. The strategy number and signature are used to ensure a closed loop of fields for subsequent receipts and recycling.

[0098] To simultaneously reflect the tendency to exceed limits, budget constraints, and the availability of communication while the population is not saturated, a monotonic saturation trigger score is constructed and mapped to an interval quantity, serving as the sole triggering criterion for priority inverters:

[0099]

[0100]

[0101] In the formula: trigger score : This is used to determine whether to enter the priority inverter path; See above; budget surplus : The smaller the value, the tighter the budget; the greater the probability of a short-term limit. : The larger the value, the stronger the tendency to limit; communication trust level : Value A higher value indicates a more reliable link; group saturation level : Value The larger the value, the closer the group is to the limit; stability criterion quantity : Range of values ,when It is considered passed when Considered as failing; weighting coefficient : Non-negative, and at least one of them is positive, used to balance the influence of the five components; Logical mapping A monotonically increasing saturating function that compresses real numbers to a certain value. This ensures that the scoring aligns with subsequent thresholds.

[0102] Among them, calculation After and threshold Compare; if and If the inverter is selected, it will enter the priority inverter path; otherwise, it will enter the on-load tap changer path.

[0103] When in use, the trigger score compresses the four dimensions of risk, availability, availability, and stability into a single quantity, avoiding the uncertainty of priority caused by multiple thresholds. When the budget is tight, the probability of exceeding the limit is high, communication is reliable, and the group is not saturated, the inverter will naturally be given priority. When the stability fails, the score is suppressed due to the increased penalty, and the tapping path is automatically transferred to avoid forcibly issuing curves under unfavorable conditions.

[0104] Furthermore, to ensure that the curve within the token is coupled in the same direction with the upstream criterion, a curve enhancement coefficient is introduced to scale the slopes of voltage-reactive and voltage-active power limiting proportionally. Curve enhancement does not rely on black-box inference but is derived from the target dead zone generated in the previous cycle. Target stationing Short-term over-limit probability Budget surplus Decide:

[0105]

[0106] Where: Curve strengthening coefficient : Value Used as a scaling factor for the curve slope; short-term overshoot probability Budget surplus Target dead zone Target stationing Same definition and values ​​as described above; basic dead zone : Value , serving as the baseline for curve scaling; base dwell : Value , serving as the baseline for curve scaling; strengthening weights Non-negative, control and right Linear magnification; scale index Non-negative, used to adjust the curve enhancement coefficient of dead zone reduction and residence lengthening. The power effect.

[0107] In calculating the curve strengthening coefficient Then, the slopes of the voltage-reactive power curve and the voltage-active power limiting curve are proportionally amplified to the new slopes; the curve number, version, and duration are then incorporated into the token body. , and the strategy number and signature, and together with The fields are hashed together to ensure the traceability of subsequent receipts and audits.

[0108] When using, curve enhancement and , Target dead zone , The four upstream quantities correspond one-to-one to avoid empirical strong-to-weak switching; when the dead zone shrinks or the dwell time lengthens, the curve is amplified exponentially, and the direction is consistent with the operating philosophy of the priority inverter in the upward phase; the token assembly uses fixed fields to express the curve, timeliness and recycling strategy, which is convenient for subsequent receipt verification and audit tracking.

[0109] After the token is issued, the execution plane needs to obtain confirmation from two levels: first, a receipt to prove that the token has been successfully received and loading has begun; second, a behavior receipt to prove that the device has switched to the curve number specified by the token and remains effective for the duration.

[0110] Meanwhile, considering the coexistence of communication and equipment differences, layered deployment is required within the reachable subgroups, and degradation coordination paths should be set for incomplete receipts or communication degradation; if it is ultimately impossible to complete at the reachable level, seamless handover to on-load tap changers must be performed within the same cycle, and target dwell must be strictly followed. .

[0111] First, collect the receipt and register the timestamp within the delivery window; then, collect the behavior receipt and register the curve number within the behavior window; for devices without receipts, determine the communication reliability level. With group saturation level The reachable subgroup is assessed, and deployment continues within the subgroup until the deployment window ends; if basic consensus is not achieved by the end of the deployment window, degenerate collaboration is initiated: the dead zone is relaxed. Extended stay And suspend new tokens; at the same time, arrange for on-load tap changers to remain stationary according to the target. Perform single or multiple steps to record lifetime accounting entries and backfill budget margins. The entire process generates a third-type context data packet. This provides direct chronological evidence for the next steps of recovery and auditing.

[0112] To avoid inconsistencies caused by a one-size-fits-all approach when communication is uncertain, deployments are based on communication reliability levels. With group saturation level and trigger rating Determine the selection ratio for reachable subgroups, and deploy them proportionally from top to bottom in the equipment list:

[0113]

[0114] Where: the proportion of reachable subgroups : Value This indicates the percentage of devices targeted for deployment within the current period; communication trust level. Group saturation level Trigger rating Same as the aforementioned definitions and values; upper limit threshold : Range of values , is used to limit the upper bound of score normalization.

[0115] Therefore, the equipment list is sorted by online status, receipt history, and latency quantile, and the selection ratio of the reachable subgroup is chosen. The proportion is prioritized for delivery to reachable subgroups; within the delivery window, if the received receipts and action receipts are gradually delivered, the subgroup can be iteratively expanded; if the subgroups are not fully delivered by the end of the window, the expansion is stopped and degradation collaboration is triggered.

[0116] When using, the reachable subgroup selection ratio Simultaneously affected and Limitations ensure that deployment is not blindly expanded when communication is poor or the target audience is already highly occupied; score normalization avoids unrealistic 100% targets in high-scoring scenarios, allowing for flexibility in the receipt window; tiered deployment reserves time and space for subsequent receipt closure and degradation coordination, reducing the risk of execution surface fragmentation.

[0117] To transform the two types of receipts into comparable closed-loop metrics and to provide a clear timing point for degradation switching, a receipt consistency metric is introduced, combining behavioral coverage and time decay into a single quantity: where behavioral receipts receive receipts:

[0118]

[0119] Where: Receipt Consistency Measurement : Value This indicates the proportion of valid action receipts relative to received receipts within the token's duration; the number of action receipts. : The system counts the number of devices reporting currently enabled curve numbers within the behavior window; it also counts the number of received receipts. : Count the number of devices that have reported receiving and started within the receiving window; time difference. : Value , represents the interval between the current time and the token issuance time; token duration. : This is the target time for the token to remain valid on the device.

[0120] When Receipt Consistency Measurement If the threshold is lower than the consistency threshold and the delivery window ends, immediately trigger degradation coordination: suspend new tokens and relax restrictions. ,extend And arranged with a load tapping device according to Execute gear shift; when the consistency measurement of the receipt is performed. Reaching the threshold and continuously reaching the recycling judgment time Then, the next step is to proceed with segmented recycling.

[0121] When using it, the combination of two types of receipts plus timeliness is compressed into a single quantity. This ensures a clear boundary between degradation and recycling; the time decay term avoids misjudging consistency when reception is early but behavior is late; and it is related to the token duration. Direct association ensures that tokens enter the recycling or degradation path upon expiration, without the need for manual intervention.

[0122] As an example, during a cloudy afternoon, the site controller reads Type II context packets within a scheduling cycle. Among them, the target dead zone Narrowed, target stationed Lengthened, action priority label The inverter is prioritized for display. The controller synchronously reads the first type of context data packet. Calculate the trigger score The score reaches the threshold, and the stability criterion is met. To pass, the controller generates a curve token containing the curve number, version, and duration. The screen displays the strategy number and indicates that the policy is pending distribution. Once distribution is complete, receipts will pop up one by one in the distribution window, followed by action receipts arriving in the action window. The status bar will then switch from a normal curve to an enhanced curve.

[0123] The controller, upon observing that some devices failed to return a response, selected a proportion according to the reachable subgroup. Expand to the next batch. After the delivery window closes, a small number of devices remain undelivered. The controller suspends new token issuance and adds these devices to the undelivered set. This is due to the consistency measurement of receipts. Once the threshold is reached, the controller keeps the token valid until the time for revocation is determined. If there is a brief communication jitter during the process, the controller will only deploy to the subgroup of the missing set, without affecting the already active devices. If in another cycle... If the threshold is not met, the controller will no longer expand the subgroup, but will instead widen the target dead zone. ,extend And instruct the on-load tap changer to proceed. Perform a gear shift; the lifetime accounting entry for this gear shift is recorded in the third-class context data packet. This will serve as the basis for the next stage of auditing.

[0124] By triggering ratings The multidimensional criteria are compressed into a single threshold, eliminating the priority uncertainty of multiple thresholds; the coefficient is strengthened by curves. Make the token curve strength and , , , Coupling in the same direction avoids empirical switching between strong and weak coupling; the proportion is selected through reachable subgroups. Layered deployment under communication and occupancy constraints reduces execution fragmentation; consistency is measured through receipts. Determine the point of degradation or recycling within the time frame to make the switching boundary between the priority inverter and the on-load tap changer clear.

[0125] Step 4: Under the combined effect of the token duration and the redemption judgment time, the segmented redemption curve and the stability criterion quantity for time-synchronized verification are determined. Solidify traceable evidence and correlate tapping actions with operating temperature to the health status coefficient. The impact is quantified as budget backfilling, generating a fourth type of context data packet. This enables closed-loop coupling of the execution, asset, and compliance modules within the same cycle.

[0126] If the token expires or the voltage stabilizes to meet the recycling judgment time... If the system immediately reverts to the normal curve, it is prone to oscillations when cloud cover reappears or load changes abruptly. Therefore, a segmented annealing mechanism is needed to ensure that the recovery intensity is determined by a combination of factors, including the consistency of the feedback, the proportion of stable duration, the tendency to exceed limits, and the stability criteria. This ensures that the recovery rhythm is consistent with the on-site conditions. Simultaneously, to ensure the continuity of the segmentation-time synchronization-review process, the recovery action must be aligned with the timestamps of both types of feedback to prevent premature degradation before the equipment has completed its behavior switch.

[0127] The specific processing method is as follows: Step a1, calculate the recycling trigger score, and uniformly consider the consistency measurement of receipts. Stable duration percentage and short-term overrun probability and stability criterion quantity Step a2: Given a passing score, generate a segmented recovery plan, proceeding in a fixed sequence: first reduce reactive power slope, then relax power restrictions, and finally switch back to normal. Time checks are performed at the beginning and end of each segment. Step a3: If a short-term over-limit probability is found at the beginning of any segment... Re-elevation or stability criterion quantity If the target is not met, immediately interrupt the recycling process and return to the mapping logic in step two to generate a new target dead zone. Staying with the target .

[0128] To ensure that the retrieval trigger has a unified and verifiable threshold, the consistency of receipts, the proportion of stable duration, the suppression of out-of-limit tendencies, and the stability certificate are combined into a range-based retrieval trigger score:

[0129]

[0130] Where: Recycling trigger score : Value Used to determine whether to enter the segmented recycling phase; Receipt consistency measurement : Value The constraint token has been stably effective within the subgroup; the percentage of stable duration. : Value This indicates the current continuous qualified duration relative to the recovery judgment time. The proportion; the probability of short-term over-limit : Value The larger the value, the stronger the tendency to limit; stability criterion quantity : Range of values , Considered passed; weight : Non-negative coefficients, used to balance the effects of the three terms; logical mapping : A monotonically saturated function that compresses the linear synthesis terms to .

[0131] When recycling triggers scoring Reaching the threshold and stability criterion quantity Enter segmented recycling; if recycling triggers scoring. Insufficient or stability criterion quantity The token remains in the maintenance state and the tap slow channel remains ready.

[0132] When in use, the recyclability is transformed from an experience-based judgment into a unified scoring system, eliminating subjective human biases. The scoring is simultaneously influenced by the consistency of evidence, the stable proportion, and the suppression of exceeding limits, making recycling decisions more closely aligned with the actual situation. Logical mapping saturates and suppresses extreme values, preventing a single factor from becoming overly dominant.

[0133] Assuming the trigger score is passed, a three-segment intensity coefficient sequence is set for the recovery process, ensuring that the reactive power slope, active power limit, and curve numbering are annealed segment by segment in a predetermined order and proportion. To avoid abrupt changes between segments, a power-law-based slow release is employed.

[0134]

[0135] Where: segmented strength coefficient : , used for the The recovery range of the segment, Reduce reactive slope, Relaxing the active power limit Switch back to the normal curve; recycle trigger score : Same as above; sustained-release coefficient : Control the first Segment release range; power exponent : Control the first The shape of the release curve varies; the larger the exponent, the slower the initial phase and the faster the subsequent phase.

[0136] Therefore, at the beginning of each segment, it is related to the third type of context data packet. Align the two types of receipt timestamps to confirm that the previous action has been implemented; according to the segmentation strength coefficient Scaling curve slope or limiting threshold, until segment end verification; if any segment end... Lift or If the descent fails, immediately stop the recovery and return to the second step to generate a new target dead zone. Staying with the target Action priority label It remains unchanged.

[0137] When in use, the power-law release allows the recovery to proceed gradually from shallow to deep, avoiding a large pullback at the very beginning of the first segment; the segment-beginning timing ensures a strict sequence of command-implementation-verification, reducing scene fragmentation; the segment-end verification ensures that the recovery process dynamically converges with the on-site state rather than a one-time jump.

[0138] As an example, after the afternoon's overcast skies cleared, the station controller observed that the terminal voltage remained within the acceptable range for one cycle, the display bar changed from yellow to green, and the stable duration marked on the time axis gradually approached the recovery judgment time. In the receipt panel, the behavioral receipt coverage remains high, and the receipt consistency metric... The system stabilizes in the bright area. The controller calculates the recycling trigger score. , reaching the threshold.

[0139] At the beginning of the first segment, the controller will adjust the reactive power slope according to the segmented strength coefficient. A slight downward adjustment resulted in the curve number remaining unchanged on the screen, but the slope value decreased; a review a few minutes later showed that the tendency to exceed limits had not increased. At the start of the second segment, the active power limit upper limit was set as follows: The restrictions were appropriately relaxed, and the power-limiting indicator on the inverter side status bar disappeared; the end-of-stage review passed. At the start of the third stage, the controller switched the curve number back to the normal curve, the status bar indicated that recycling was complete, and the action priority label for this cycle was locked. Unchanged, awaiting evaluation in the next cycle. If the cloud shadow reappears midway through the second phase, the probability of short-term overshooting increases. Lift, controller interrupt recovery, return to step two to generate a new target dead zone. Staying with the target No back-and-forth movement was observed at the scene.

[0140] After the segmented recycling and certificate verification are completed, the execution sequence, curve number trajectory, receipt timestamp, and token signature must be solidified as searchable evidence, and the operational facts of the tapping action, temperature rise, and recycling duration must be mapped to the health status coefficient. Daily-hour budget , On the update. Only by backfilling the marginal impact of asset degradation into the budget layer within the same period can the budget surplus for the next period be ensured. Only then can it possess explainability and traceability.

[0141] The specific processing method is as follows: Step a1, combine the token signature hash, the receipt timing hash, and the curve number trajectory hash within the same period to form an audit evidence strength index, which is used to quickly retrieve the consistency of the event-behavior-curve; Step a2, based on the third type of context data packet... The recording of tap changer operations and operating temperature rise quantifies the wear contribution of this cycle into a health status coefficient. And proportionally fill in the daily-hour budget for the next cycle; Step a3, encapsulate the above results together with the exception logs, over-limit minutes and limited sending minutes of the current cycle into a fourth type of context data packet. .

[0142] In order to quickly assess the integrity of the audit chain without disclosing the original message, token signature hashing is used. Receipt timing hash Hash with curve number trajectory After being synthesized into interval quantities using a unified scale:

[0143]

[0144] Where: Audit evidence strength index : Used to measure the consistency and integrity of the token, receipt, and curve; token signature hash : , representing the strength of the token signature after normalized hashing; the token signature verification result is taken as a binary value (verification passed is 0). Failed );

[0145] Receipt timing hash : The strength of the receipt time series after normalized hashing is defined as the normalized consistency score of the difference between the receipt time series summarized by device and the issuance time; curve number trajectory hashing. : This represents the intensity of the curve number change sequence after normalized hashing, defined as the proportion of time within the behavior window when the actual curve number of the device is equal to the target number of the token; abnormal deviation count. : This indicates the count of events such as misaligned receipts or inconsistent curve numbers.

[0146] Weight : Non-negative coefficients, measuring the contribution of the three types of hashes to the audit; weights Non-negative coefficients, used to adjust for the relative influence of evidence and bias; logical mapping. : A monotonically saturated function that compresses the linear synthesis terms to .

[0147] When audit evidence strength indicators When the audit threshold is reached, the tokens and receipts for that period are considered as a closed loop of evidence; if the audit evidence strength index... If the threshold is not met, the pending verification mark for that period is retained and added to the fourth type of context data packet. Record the list of equipment and time periods that need to be reviewed.

[0148] When used, the strength of evidence index The three dimensions of signature, time series, and trajectory are unified into interval quantities to facilitate batch screening; anomaly deviation counting is introduced. It can directly reflect key inconsistencies such as misalignment; logical mapping avoids the whole being misjudged as incomplete due to a single hash anomaly.

[0149] To bring forward the impact of switching operations and temperature rise on assets into the budget, a multiplicative updated health status coefficient is used, and the budget ceiling for the next period is lowered accordingly.

[0150]

[0151] Where: the updated health status coefficient : Value Used for budget scaling in the next cycle; current health status coefficient : Value This represents the initial health level for this cycle; the number of tapping actions. : Value This represents the cumulative number of taps in this cycle; the baseline number of actions. : Value , which serves as the reference scale in the lifetime model;

[0152] Maximum temperature rise : Value This is the highest winding or oil temperature observed during this cycle; reference temperature rise. : Value , serving as the reference scale in the lifespan model; wear weight : Value Used to quantify the impact of the number of movements on health; temperature rise weighting : Value This is used to quantify the health effects of temperature rise.

[0153] Calculate the updated health status coefficient Then, the daily budget for the next cycle will be... With hourly budget according to The contraction ratio is calculated, and the current cycle's tapping action entries, temperature rise trajectory, and contraction ratio are written into the fourth type of context data packet. This is for direct use in the next cycle.

[0154] When in use, the asset status is brought forward to the budget level by multiplicative updates and proportional contractions to avoid the contradiction of poor health but frequent operation in the next cycle; tap changes and temperature rise are included at the same time to avoid focusing only on mechanical wear and ignoring thermal stress; the update volume is traceable, which is conducive to horizontal comparison between different feeders.

[0155] When used, scoring is triggered by recycling. With segmented strength coefficient The recycling process is driven by the same criteria across three levels: whether to recycle, how quickly to recycle, and in what order; this is achieved through auditing the strength of evidence indicators. Token signatures, receipt timing, and curve trajectories are uniformly encapsulated into searchable evidence; updated through health status coefficients. With the budget ratio shrinking, asset wear and tear is immediately reflected in the next cycle's budget, forming a closed-loop coupling of execution, evidence, and assets.

[0156] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0157] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0158] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0159] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0160] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A contactless electronic voltage regulating power transformer online control method, characterized in that: include, Collect three-phase and terminal voltage and power flow, inverter communication and curves, tap status and equipment health; Calculate the communication credibility level, the group saturation level, and the short-term limit violation probability to generate a non-oscillating proof and form a first-type context data packet; Based on the first type of context data packet and device health, calculate the budget margin and determine whether it is a rising edge or a falling edge; generate the target dead zone and target dwell according to the coupling mapping, verify that it meets the non-oscillation proof, output the second type of context data packet and mark the action priority label; Based on budget margin, short-term over-limit probability, communication reliability level, group saturation level, and non-oscillation proof gating; when the conditions are met, a curve token is issued to the inverter group and two types of receipts are completed; otherwise, the deconnection is implemented according to the target residency; a third type of context data packet is formed. When the voltage is qualified or the token expires, it is recycled in segments in the order of reducing the reactive power slope, relaxing the active power limit, and restoring the normal curve, and the non-oscillation certificate is checked at the beginning and end of the segment; the token, receipt and curve trajectory are fixed, the health status coefficient and daily and hourly budgets are updated, and the fourth type of context data packet is formed. The communication reliability level is determined by a combination of online ratio, receipt success rate and high quantile of delay. The group saturation level is the maximum value of the high quantile of reactive power occupancy ratio and the high quantile of active power limited generation ratio. The short-term over-limit probability is generated by mapping the voltage relative deviation, photovoltaic power output change rate and neighborhood similarity term, and is written into the first type of context data packet along with the non-oscillation proof. The budget margin is obtained by scaling the daily budget and hourly budget with a health status coefficient, and the rising or falling edge is determined within the statistical window; based on this, the target dead zone and target dwell time are generated, and upper and lower bounds are set respectively and verified with non-oscillation proof, and the second type of context data packet and action priority label are output. The trigger score is composed of budget margin, short-term over-limit probability, communication credibility level and group saturation level and compared with the threshold, with non-oscillation proof as a necessary condition; when the condition is met, a curve token is generated. The curve token contains curve number, version, duration, recycling strategy number and signature. Two types of receipts are included: reception receipt and behavior receipt.

2. The online control method for a contactless electronic voltage regulating power transformer according to claim 1, characterized in that: The neighborhood similarity term is calculated based on the directed increment of the end voltage of adjacent feeders at the same station according to the kernel weight, and the statistical window is aligned with the scheduling cycle. For missing measurements, the previous valid value is used with an increased safety margin. Unreliable measurement points are blacklisted and downweighted. Non-oscillating proofs are recorded as passed or failed.

3. The online control method for a contactless electronic voltage regulating power transformer according to claim 2, characterized in that: Action priority labels are determined by the budget margin and short-term over-limit probability. When the budget margin is low and the short-term over-limit probability is high, it is marked as a priority inverter; otherwise, it is marked as an available tap. The target dwell time is taken to be longer with the falling edge and shorter with the rising edge, and is coupled with the feeder electrical time constant to maintain consistency in a single cycle.

4. The online control method for a contactless electronic voltage regulating power transformer according to claim 3, characterized in that: When the received receipts are incomplete or the communication trust level is below the threshold, curve tokens are issued in batches according to the reachable subgroup selection ratio; when the delivery window ends and the receipts are still incomplete, new tokens are suspended, and the on-loaded tapping device performs single or multiple steps according to the target's residence. The tapping action is recorded with the third type of context data packet.

5. The online control method for a contactless electronic voltage regulating power transformer according to claim 4, characterized in that: The recycling trigger is jointly satisfied by the consistency measurement of receipts and the proportion of continuous qualified duration, and the non-oscillation proof is maintained. The segmented recycling order is to reduce the reactive power slope, relax the active power limit, and restore the normal curve. The first segment is aligned with the timestamps of the two types of receipts, and the last segment is reviewed for the probability of short-term over-limit and the non-oscillation proof.

6. The online control method for a contactless electronic voltage regulating power transformer according to claim 5, characterized in that: Curved token signatures, two types of receipt timings, and curve number trajectories are solidified into audit evidence items, and audit evidence strength indicators are generated and written into the fourth type of context data package; abnormal items and equipment lists are listed separately for subsequent review and retrieval.

7. The online control method for a contactless electronic voltage regulating power transformer according to claim 6, characterized in that: The tapping action accounting and operating temperature rise in the third type of context data packet are mapped to health status coefficient updates, and the daily and hourly budgets for the next cycle are scaled down proportionally; the updated health status coefficients and budgets are returned with the fourth type of context data packet.

Images