A low-voltage cable branch box energy storage peak load shifting power saving method and system

Abstract

The application discloses a low-voltage cable branch box energy storage peak clipping valley filling power saving method and system, which calculates the cable current-carrying heat accumulation curve by analyzing the branch box operation data, identifies the truly continuous heat peak period, arranges the basic charging plan accordingly, and calculates the short-time overload power and allocates the charging power quota for the period when the heat accumulation is not saturated. Further, the dischargeable capacity is calculated according to the charging control instruction, the branch with the hidden heat source is identified by the temperature drop performance after stopping charging, and the discharge support order is determined to form the discharge priority sequence and the scheduling instruction. Further, the progressive rising trend of the internal resistance in the charging and discharging process of the energy storage unit is tracked, the life margin is estimated, the safety level is converted, and the protection limit signal is generated for the unit below the threshold level. Finally, the emergency reserved capacity is determined according to the actual lag performance of the limit removal, the arbitrage window is selected combined with the load fluctuation condition, the allocation quota is determined, and the charging and discharging instruction considering the load safety and economic benefit is output.

Description

Technical Field

[0001] This invention relates to the field of low-voltage power distribution and energy storage dispatching technology, and in particular to a method and system for peak shaving and valley filling energy saving using energy storage in low-voltage cable branch boxes. Background Technology

[0002] When energy storage is introduced into a low-voltage cable branch box to achieve peak shaving and valley filling, the amount of additional charging load that it can safely handle essentially depends on the actual heating state of the cable at that moment, rather than a fixed rated current value. Most existing solutions simply avoid charging based on experience-set time periods. Even if a short current spike quickly drops, it is treated as something to be avoided. However, the continuous heat accumulation that repeatedly approaches the same temperature level over multiple cycles, which is what really needs to be guarded against, may be ignored because the value of a single current is not particularly high. Neither of these solutions reflects the actual pressure that the cable itself bears.

[0003] When energy storage systems undertake discharge support tasks, they also face similar judgment dilemmas: if a circuit fails to cool down after charging, it may be due to the continued heavy load or other heat sources on the line. Existing solutions cannot distinguish between these two causes, so it is impossible to use limited discharge resources where they are truly needed. In addition, the increase in battery internal resistance is a slow cumulative process. Current safety management relies on fixed threshold trigger protection, which may fail to detect risks in the early stages of the increase and lacks tracking of the actual recovery speed of the equipment after the restriction is lifted. This means that the reserve of emergency backup power always remains a rough estimate based on experience. Summary of the Invention

[0004] This invention discloses a method and system for energy storage, peak shaving, and valley filling in low-voltage cable branch boxes. The aim is to dynamically assess the available load margin of the cable based on its actual power-on history, identify the additional heat sources that may exist behind the abnormal cooling circuit after charging is stopped, and arrange the priority of discharge support accordingly. The system also determines the safety level and limits the discharge intensity based on the gradual increase trend of the battery's internal resistance, and dynamically determines the emergency reserve power and arbitrage resources according to the actual lag performance after the restriction is lifted, so as to achieve dynamic optimization scheduling that takes into account both safety margin and economic benefits.

[0005] The first aspect of this invention proposes a method for energy storage, peak shaving, and valley filling in low-voltage cable branch boxes to save electricity, comprising: Obtain branch box operation data, calculate current-carrying heat accumulation based on the branch box operation data to form cable heat accumulation curve, and identify the flat-top load duration through the short-term peak segment of the cable heat accumulation curve to determine the continuous heat peak period; A basic charging plan is obtained by arranging peak avoidance according to the continuous hot peak period. The short-term overload power of the unsaturated heat accumulation section is measured based on the basic charging plan to obtain the charging power quota. The charging power quota and the cable heat accumulation curve are used to generate charging control instructions by reverse allocation according to the heat accumulation difference. Based on the charging control command, the state of charge is calculated to obtain the dischargeable capacity. Using the cooling hysteresis branch within the dischargeable capacity, the support order is arranged to obtain the discharge priority sequence. A discharge scheduling command is output for the discharge priority sequence by inserting a hysteresis cooling gap. The unit status data is obtained by collecting the operating status according to the discharge scheduling command, and the evolution of the internal resistance degradation curve is tracked based on the unit status data to form a safety margin rating. The protection action is determined and a protection limit signal is generated based on the safety margin rating. By utilizing the delay in releasing the protection limit signal, the emergency lower limit demand is statistically analyzed to obtain the emergency reserve capacity. Based on the emergency reserve capacity, the arbitrage allocation quota is determined to generate a comprehensive adjustment scheme. Peak shaving and valley filling optimization output charging and discharging commands are executed for the comprehensive adjustment scheme.

[0006] A second aspect of this invention provides a low-voltage cable branch box energy storage peak shaving and valley filling power saving system, comprising: The heat peak identification module is used to acquire branch box operation data, calculate current-carrying heat accumulation based on the branch box operation data to form a cable heat accumulation curve, and identify the duration of flat-top load through the short-term peak segment of the cable heat accumulation curve to determine the continuous heat peak period. The charging decision module is used to arrange peak avoidance according to the continuous hot peak period to obtain a basic charging plan, to determine the short-term overload power of the unsaturated heat accumulation section based on the basic charging plan to obtain the charging power quota, and to generate charging control instructions by reverse allocation of the charging power quota and the cable heat accumulation curve according to the heat accumulation difference. The discharge decision module is used to calculate the state of charge based on the charging control command to obtain the dischargeable capacity, use the cooling hysteresis branch within the dischargeable capacity to arrange the support order to obtain the discharge priority sequence, and insert a hysteresis cooling gap into the discharge priority sequence to output the discharge scheduling command. The safety detection module is used to collect the operating status data of the unit according to the discharge scheduling command, track the evolution of the internal resistance degradation curve based on the unit status data to form a safety margin rating, and determine the protection action based on the safety margin rating to generate a protection limit signal. The instruction output module is used to release the hysteresis of the protection limit signal, calculate the emergency lower limit demand to obtain the emergency reserve capacity, determine the arbitrage allocation quota based on the emergency reserve capacity to generate a comprehensive adjustment scheme, and execute peak shaving and valley filling optimization output charging and discharging instructions for the comprehensive adjustment scheme.

[0007] The beneficial effects of this invention are reflected in the following points: First, the current-carrying heat accumulation curve of the cable itself can distinguish between short-term current surges and truly continuous load pressure. The former is characterized by a rapid temperature rise followed by a drop, while the latter is characterized by a temperature rise repeatedly approaching the same level over multiple cycles and being difficult to completely dissipate. Based on this, this invention changes the basis for judging peak charging avoidance from a simple current threshold to tracking the actual trend of cable heat accumulation, and calculates the specific short-term overload power limit during the period when heat accumulation has not yet saturated, so that charging capacity that was originally idle due to excessive caution can be safely utilized. Second, this invention discovers that the cooling performance of a charging branch after charging stops can reflect whether there are additional heat sources other than charging. If the temperature of a branch does not drop for a long time after charging stops, it often means that there are other loads on the same line that are continuously generating heat, rather than just residual heat caused by this charging. Based on this discovery, this invention prioritizes branches with abnormal cooling when arranging energy storage discharge support, shifting the allocation of discharge resources from simply checking whether the capacity is sufficient to truly taking care of circuits with more urgent heat pressure. Finally, this invention integrates two factors that were previously handled separately—the gradual increase in battery internal resistance and the actual recovery lag after the restriction is lifted—into the same scheduling logic: the trend of internal resistance increase is used to calculate the safety level and determine whether to limit the discharge intensity of a certain cell, while the lag performance after the restriction is lifted is used to determine the emergency power that actually needs to be reserved. Together, they determine which capacity must be locked and which can be released to participate in arbitrage, so that safety margin and economic benefits are no longer the result of simple superposition after independent calculations, but are dynamically weighed under the same set of constraints. Attached Figure Description

[0008] Figure 1 This is a flowchart of a low-voltage cable branch box energy storage peak shaving and valley filling power saving method according to the present invention.

[0009] Figure 2 This is a schematic diagram illustrating the identification of short-term peak periods and continuous hot periods in this invention.

[0010] Figure 3 This is a structural block diagram of a low-voltage cable branch box energy storage peak shaving and valley filling power saving system according to the present invention.

[0011] Among them: 1-Cable heat accumulation curve; 2-Short-term peak section; 3-Steepness indicator line; 4-Temperature lag rise section; 5-Ping-top candidate section; 6-Historical cycle curve; 7-Steepness convergence point; 8-Baseline rise indicator line; 9-Continuous heat peak period. Detailed Implementation

[0012] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0013] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0014] The technical solutions of the embodiments of this application will be described below.

[0015] like Figure 1 As shown, this embodiment of the invention provides a method for energy storage, peak shaving, and valley filling in low-voltage cable branch boxes to save electricity, including the following steps S101-S105: Step S101: Obtain branch box operation data, calculate current-carrying heat accumulation based on branch box operation data to form cable heat accumulation curve, and identify the duration of flat-top load by the short-term peak segment of cable heat accumulation curve to determine the continuous heat peak period.

[0016] Specifically, the branch box operation data is acquired, and the current-carrying heat accumulation is calculated based on this data to form a cable heat accumulation curve. The branch box operation data collection includes instantaneous current values ​​from the current transformers of each branch circuit, temperature sensor readings, and ambient temperature readings within the branch box. After being timestamped using a unified clock source, the data is stored separately according to the circuit number on the same timeline; data from different circuits are not merged. Current signals are sampled at a second-level cycle, and temperature signals at a minute-level cycle. The data is temporarily stored locally at the acquisition terminal at the end of each circuit and then uploaded in batches to the branch box main control unit at a set cycle. The calculation of current-carrying heat accumulation uses a first-order thermal inertia model, and the cable temperature rise is calculated using the formula T(t) = T(t-1) + (I(t)). 2The calculation is based on the iterative formula: ×R-(T(t-1)-T_amb) / τ)×Δt, where I(t) is the loop current at time t, R is the resistance per unit length of the cable (this value has been converted to an equivalent heating coefficient matching the temperature rise rate, not the original value of pure resistance), T_amb is the ambient temperature, τ is the cable thermal time constant, Δt is the sampling interval, and T(t) is the estimated temperature rise at time t. The cable heat accumulation curve is formed by connecting the T(t) values ​​obtained by iterating through this formula at each time step in chronological order. R and τ are taken from the product technical parameter manual of this cable model, and different values ​​correspond to different cross-sectional areas and different sheath materials. The temperature sensor readings serve as a calibration reference during the calculation process. Every set calibration cycle, the estimated temperature rise is compared with the measured terminal temperature readings. When the difference exceeds the set tolerance, the value of R or τ is slightly corrected and the iteration continues. The correction result is reflected in the subsequent trend of the cable heat accumulation curve. The cable heat accumulation curve is generated independently for each loop, covering all data collection periods of the branch box operation. Each moment on the curve corresponds to a calculated temperature rise value. When a calibration correction occurs, the historical value before the correction is not recalculated. Instead, the calculation continues from that moment onwards using the new parameters. The correction record is retained along with the correction moment.

[0017] In some embodiments, the step of identifying the flat-top load duration and determining the continuous heat peak period by means of the short-term peak segment of the cable heat accumulation curve includes: identifying the rising edge steepness based on the short-term peak segment of the cable heat accumulation curve to obtain peak steepness characteristics; identifying the steepness convergence curve through the temperature lag rise section in the peak steepness characteristics to determine the flat-top candidate interval; verifying the period repeatability of the flat-top candidate interval to obtain the flat-top load duration; and determining the heat peak level and the continuous heat peak period based on the flat-top load duration.

[0018] The steepness of the rising edge is identified based on the short-term peak segment of the cable heat accumulation curve to obtain the peak steepness characteristics. For example... Figure 2As shown, the calculated temperature rise is a continuous segment that rises rapidly in a short period of time and then slows down. The data is taken from the cable heat accumulation curve 1 and is identified as the short-term peak segment 2. The rapid rise phase corresponds to the initial temperature rise response caused by the sudden increase in current. For example, when a certain circuit energy storage device starts high-power charging, the current jumps several times within a few seconds, and the corresponding calculated temperature rise also rises sharply. The steepness of the rise edge of the short-term peak segment 2 is calculated by the formula k=ΔT / Δt_r, where ΔT is the temperature rise change from the start of the segment to the moment when the temperature rise rate first slows down significantly, Δt_r is the corresponding time interval, and k is the steepness of the rise edge. The slope trend is marked by the steepness indicator line 3 along the climbing direction of the segment. The steeper the inclination angle of the indicator line, the higher the corresponding k value. The two have a direct correspondence. The more rapid the current surge, the steeper the temperature rise curve rises at the initial moment, and the higher the steepness k-value. This correspondence forms the basis for using peak steepness characteristics to distinguish differences in the thermal response of different circuits. Circuits with significantly higher steepness under the same surge magnitude often have relatively poor heat dissipation conditions. For example, circuits in branch boxes with dense cable laying and poor surrounding ventilation are more likely to exhibit this behavior. For short-term peak segments occurring in the same circuit at different times, the rising edge steepness is calculated separately for each occurrence, without being averaged. For example, if the same circuit has a peak segment in the morning and another in the evening, the two steepness values ​​are recorded separately and not merged due to close intervals. Peak steepness characteristics are a set of records formed by adding the rising edge steepness values ​​and steepness narrowing times corresponding to each short-term peak segment identified for the circuit in the order of identification. Subsequently, steepness values ​​are extracted from this set one by one according to the observation period. Periods in the curve where the temperature rise is gradual throughout and there is no rapid rise phase are not included in this set.

[0019] For example, the step of identifying the steepness convergence curve and determining the flat-top candidate interval by means of the segment with lag in temperature rise in the peak steepness feature includes: performing multi-cycle steepness repetition sampling to obtain a steepness repetition sequence based on the segment with lag in temperature rise in the peak steepness feature; calculating the convergence degree to form a steepness convergence curve by means of the segment with convergence accompanied by baseline rise in the steepness repetition sequence; verifying the duration based on the steepness convergence curve to obtain a candidate duration threshold; and calibrating the interval boundary according to the candidate duration threshold to determine the flat-top candidate interval.

[0020] Based on the temperature lag rise segment in the peak steepness feature, multi-cycle steepness repetition sampling is performed to obtain a steepness repetition sequence. Temperature lag rise segment 4 is the segment in the peak steepness feature where the temperature rise continues at a relatively slow rate after the rising edge steepness narrows. The response lag is caused by the thermal inertia of the cable: the larger the thermal time constant τ, the slower the temperature rise response, and the longer it takes to stabilize after the peak. Multi-cycle steepness repetition sampling is performed on the same circuit, identifying temperature lag rise segments one by one within each observation period marked by the peak steepness feature, extracting the rise rate of each segment, and then arranging them into a sequence according to time. A steepness repetition sequence is valid only if the temperature lag rise segment actually appears at least 3 times. The rise rate extracted here is not the same as the rising edge steepness k of the short-term peak segment; the former describes the slope of the slow rise stage, while the latter describes the slope of the abrupt jump stage. Although they have the same dimensions, they correspond to different response stages. If certain periods are missing or do not exhibit temperature lag rise segments, the corresponding positions are left blank and marked as "no data," without interpolation to avoid fictitious values ​​interfering with convergence trend judgment. The extraction range covers all identified segments since the loop's commissioning, not limited to recent periods; results from earlier segments are also retained and included in the sequence. When multiple temperature lag rise segments appear within a single observation period, the rise rate for each segment is extracted independently. Multiple results from the same period are included in the sampling points for that period, without secondary averaging; instead, they are arranged in the sequence according to the order in which they belong. The steepness repetition sequence arranges the rise rates extracted from the peak steepness characteristics periodically in chronological order, continuously extending with the increase in observation periods, forming the direct data source for subsequent convergence calculations.

[0021] The steep convergence curve is formed by calculating the convergence degree in the segments of a steep repetitive sequence where convergence is accompanied by a baseline rise. The steep convergence curve is plotted based on the periodic ramp rate values ​​of the steep repetitive sequence, noting whether each period falls into a convergence segment and whether it is accompanied by a baseline rise. The horizontal axis of the curve represents the period number, and the vertical axis represents the corresponding ramp rate value. The ramp rate values ​​of adjacent historical period curves (6) are taken from the steep repetitive sequence. The segments where the differences gradually decrease and the decreasing trend is stable and continuous are marked as steep convergence points (7) after the trend is confirmed to be stable in a certain period. If the absolute level of the ramp rate corresponding to this convergence segment is higher than that of earlier periods, it is further determined to be convergence accompanied by a baseline rise (8). This combination of "convergence but baseline rise" usually corresponds to scenarios where the load on the circuit gradually increases while the heat dissipation conditions do not improve simultaneously. Figure 2The baseline rise indicator line 8 marks the longitudinal spacing between adjacent periodic curves. The wider the spacing, the greater the corresponding baseline rise. Convergence is calculated by the reduction ratio of the differences between two adjacent sets, i.e., (the difference between the previous set and the difference in the current set) divided by the difference between the previous set. The closer the ratio is to 1, the more sufficient the convergence. This process of gradually approaching 1 with the increase of the observation period is itself a reference for judging whether the heat accumulation mode of the loop tends to be stable. The baseline rise amplitude is taken as the difference between the absolute level of the rise rate of the initial period and the final period of the convergence segment. When the rise amplitude is positive and the convergence reaches the set criterion, it is determined that the loop has a trend of heat accumulation that intensifies period by period. This is visually represented on the steep convergence curve as the overall curve slowly shifts upward with the period. In steep repetitive sequences, if the rise rate fluctuates randomly during each period and does not show a convergence trend, or if the convergence value fluctuates and does not show a stable trend, it is determined that the current load mode has not been fixed and is not suitable for the calibration of candidate duration thresholds.

[0022] Candidate duration thresholds are obtained by verifying the duration of the steep convergence curve. The number of cycles corresponding to the convergence segment is obtained from the steep convergence curve and counted one by one. The number of cycles is converted into actual duration by multiplying by the observation cycle interval of the circuit. For example, if the observation cycle interval of a circuit is one day and the convergence segment spans six cycles, the corresponding duration is converted to six days. The converted duration is compared with the set minimum duration criterion. Convergence segments that reach or exceed the criterion are judged to be eligible for candidate status. The criterion value refers to the lower limit of the empirical duration that is generally accepted in the thermal aging mechanism of this type of cable, which requires continuous heating for multiple cycles to produce a substantial effect. The calibration of this lower limit also refers to the convergence performance of the steep convergence curve in historical real thermal aging cases. If multiple convergence segments qualify as candidates within the same loop during the observation period, the duration of each segment is retained separately and not merged into a single value. This is because different convergence segments often correspond to different load scenarios for the loop, such as high-load periods on weekdays and low-load periods on holidays, each forming an independent candidate segment. The duration value corresponding to each segment is marked with its respective observation time range. The candidate duration threshold is based on the duration value corresponding to the shortest set of convergence segments that qualify as candidates. The start and end periods of this set of convergence segments are also retained. When multiple sets of candidate segments coexist, the shortest one is used. Selecting the shortest rather than the longest or average provides a conservative lower limit benchmark for interval boundary calibration, preventing overly broad definitions that might mistakenly include short-term fluctuations. The candidate duration threshold is determined independently for each loop. The observation period interval and shortest duration criterion may differ for different loops. Before comparing across loops, they must be uniformly converted to the same observation period interval benchmark.

[0023] The flat-top candidate interval is determined by calibrating the interval boundary based on the candidate duration threshold. The start and end periods of the convergence segment corresponding to the candidate duration threshold are mapped back to the actual time coordinates on the cable heat accumulation curve 1. The mapped start and end times are the preliminary values ​​of the boundary of the flat-top candidate interval 5. Taking a circuit whose convergence segment corresponding to the candidate duration threshold spans from day 3 to day 9 as an example, the preliminary value of the interval boundary after mapping is the actual time range from 0:00 on day 3 to the end of day 9 for that circuit. If the starting point of the interval boundary happens to fall at a position where the steepness of the rising edge of a short-term peak segment has not yet narrowed, the starting point is shifted backward to the moment when the steepness of that peak segment has completed narrowing. The shift range is usually on the order of several minutes to more than ten minutes, depending on the magnitude of the cable thermal time constant of that circuit. Its function is to ensure that the period that is still in the violent response stage is not included in the flat-top candidate interval. The endpoint is handled similarly to the starting point. If it falls within the buffer window before the start of the next short-term peak, it is moved forward accordingly. The forward shift of both the start and end points is recorded and retained for verification of the adjustment amount relative to the initial value of the interval boundary. For example, if the endpoint of a loop is moved forward by twelve minutes due to its proximity to the peak, this value is fully recorded in the adjustment record. The performance of the flat-top candidate intervals after the start and end times are determined in multiple historical cycles is left to be judged in the cycle repeatability verification process. If there are multiple candidate duration thresholds for the same loop corresponding to different convergence segments, each is mapped to generate an independent flat-top candidate interval, which is not merged with each other, and the start and end times are recorded separately. If the time ranges of two flat-top candidate intervals overlap after mapping, the overlapping part is retained in the longer interval group, and the overlapping part of the other group is removed to prevent the same time period from being counted repeatedly in two interval groups.

[0024] The flat-top load duration is obtained by verifying the repeatability of the flat-top candidate interval. The start and end times corresponding to the flat-top candidate interval are checked one by one in multiple observation cycles of the loop to see if there are similar intervals with similar time ranges. The similarity criterion is that the deviation between the start and end times does not exceed a set tolerance. The tolerance is adjusted accordingly with the length of the observation cycle interval of the loop; the longer the observation cycle interval, the coarser the granularity of the start and end times themselves, and the corresponding tolerance is relaxed. When similar intervals appear repeatedly in multiple consecutive observation cycles, and the number of repetitions reaches the set minimum number, the flat-top candidate interval is determined to have periodic repeatability. After the determination is passed, the interval duration is officially confirmed as the flat-top load duration. If similar intervals appear only once or twice occasionally and do not reach the minimum number of repetitions, they are determined to be occasional flat-top loads. The original record is retained, and this record will be counted together when new cycle data is added to the loop. The counting caliber is consistent with the first occurrence and is not restarted due to the length of the intermediate interval. The specific value of the flat-top load duration is based on the median of the durations of similar intervals that have repeatedly occurred. Unlike simply taking the original duration or the average, the median is less affected by occasional deviations in length or length in individual cycles and better reflects the typical performance of the candidate flat-top load interval across multiple cycles. If there are multiple sets of candidate flat-top load intervals that have passed the cycle repeatability verification for the same circuit, each set generates an independent flat-top load duration, which is not forcibly merged into a single value. Each set is labeled with the date characteristics of the occurrence of the corresponding similar intervals to distinguish different usage patterns such as weekdays and non-weekdays. For example, for a circuit, weekdays and weekends each form an independent set of flat-top load durations, and the two values ​​often differ significantly, with the weekday set usually being longer than the weekend set.

[0025] The peak heat level is determined based on the duration of the flat-top load to identify the continuous peak heat period. The duration of the flat-top load is compared one by one with the set peak heat level classification thresholds, which are divided into three levels: mild, moderate, and severe. In the mild level, charging continues as usual. As the level increases towards moderate and severe, it needs to be avoided in advance during the charging plan. The level boundary is calibrated based on the mechanism analysis of whether the heat accumulation of the cable model approaches the material's tolerance limit at the corresponding duration. The calibration process also refers to the heat resistance rating of the cable sheath material. For example, PVC-sheathed cables and cross-linked polyethylene-sheathed cables have different level boundaries due to their different heat resistance ratings. Determining the severe level also requires considering the cable's heat accumulation during that period. The calculated absolute temperature rise level of the curve reaches a fixed proportion of the tolerance limit. Periods with long flat-top load durations but low temperature rise levels are classified as medium-risk. Both conditions must be met for a period to be classified as heavy-risk. This prevents misjudging long-duration periods with low actual heat accumulation as high-risk. The specific value of this fixed proportion is determined by referring to the correspondence between temperature rise level and insulation life loss in the thermal aging test of this cable model. A higher proportion indicates a more significant actual consumption of insulation life. Different sheath materials correspond to different proportion values. Polyvinyl chloride sheaths are usually lower than cross-linked polyethylene sheaths and do not use a uniform standard. If there are multiple sets of flat-top load durations in the same circuit, each set falls into a different heat peak level. Each set generates a corresponding continuous heat peak period 9 independently. The set with the higher level is given priority in the peak avoidance scheduling. The continuous heat peak period 9 corresponding to the light-risk level is also included in the basic charging plan scheduling, only with a lower priority than the medium and heavy levels, and the corresponding power reduction is also significantly smaller.

[0026] Step S102: Based on the continuous peak heat period, a basic charging plan is obtained by arranging peak avoidance. Based on the basic charging plan, the short-term overload power of the unsaturated heat accumulation period is measured to obtain the charging power quota. The charging power quota and the cable heat accumulation curve are used to generate charging control instructions by reverse allocation according to the heat accumulation difference.

[0027] Specifically, a basic charging plan is derived by scheduling peak avoidance based on the continuous peak heat periods. Time intervals within the continuous peak heat periods, marked as moderate or severe, are designated as prohibited charging periods when scheduling the charging plan. No energy storage charging tasks are scheduled for any branch circuits during these periods, as the additional current generated by charging itself would further increase the already high level of cable heat accumulation. Time intervals within the continuous peak heat periods, marked as mild, are designated as restricted charging periods. Charging is permitted during these periods, but a lower power limit is set. The power limit is determined by a fixed reduction ratio of the rated charging power of the circuit. This reduction ratio is set with reference to the specific value of the peak load duration corresponding to this period; the closer the peak load duration is to the moderate threshold, the lower the corresponding power limit. The remaining time intervals not covered by the continuous peak heat periods are scheduled as regular charging periods. During these periods, charging tasks are scheduled normally according to the rated charging power of each branch circuit, without additional power reduction. If multiple continuous peak periods exist within the same circuit, each corresponding to a different flat-top mode, the basic charging plan is arranged according to each mode. Weekday and non-weekday charging plans are generated independently and not mixed. The system automatically matches the corresponding plan version based on the date attribute; for example, the weekday version is used from Monday to Friday, and the non-weekday version is automatically switched on weekends. When determining the time period classification for a specific moment in the short-term overload power measurement stage of the unsaturated heat accumulation phase, the system directly determines whether the moment falls within the prohibited, restricted, or regular time periods already defined in the basic charging plan, without re-judging based on the original gear level of the continuous peak period. This classification result, along with the corresponding time period boundary, is recorded as the basis for direct querying in the next stage.

[0028] In some embodiments, the step of determining the short-term overload power of the unsaturated thermal accumulation section based on the basic charging plan to obtain the charging power quota includes: calculating the remaining temperature rise margin of the unsaturated thermal accumulation section based on the basic charging plan to obtain the temperature rise margin margin; calculating the temperature difference narrowing based on the temperature rise margin margin to determine the available overload duration; calculating the overload power upper limit using the available overload duration to obtain the short-term overload power; and determining the charging power quota according to the branch capacity allocated by the short-term overload power.

[0029] For example, the step of calculating the remaining temperature rise margin of the unsaturated heat accumulation segment based on the basic charging plan to obtain the temperature rise margin includes: determining the remaining temperature rise margin based on the thermal limit difference of the unsaturated heat accumulation segment calculated based on the basic charging plan; obtaining a corrected heat capacity difference by correcting the fluctuation range according to the remaining temperature rise margin; constructing a temperature rise margin margin ratio based on the corrected heat capacity difference; and converting the margin using the temperature rise margin margin ratio to obtain the temperature rise margin.

[0030] The remaining temperature rise margin is determined based on the difference between the thermal limit and the unsaturated section of the basic charging plan. During the restricted charging period and the regular charging period of the basic charging plan, the temperature rise is compared with the tolerance limit of the cable model at each sampling point along the cable's thermal accumulation curve. The interval where the temperature rise has not yet reached the limit is defined as the unsaturated section of thermal accumulation. The remaining temperature rise margin is calculated point-by-point at each sampling point within this interval using the formula C_r = T_limit - T_now, where T_limit is the temperature rise tolerance limit of the cable model, T_now is the current thermal accumulation value, and C_r is the remaining temperature rise margin. If the difference between the current thermal accumulation value and the tolerance limit remains relatively small within the most recent sampling points but has not yet truly reached the limit, it is considered to be continuously approaching the limit. This is distinctly different from the behavior of a sudden temperature rise approaching the limit and then rapidly falling back during a short-term peak period. The former reflects the continuous thermal pressure applied by the load mode, while the latter reflects the transient thermal response of a single current surge. For periods where the current heat accumulation value is within the unsaturated range but significantly different from the tolerance limit, the remaining temperature rise margin is still calculated using the same formula, with C_r being correspondingly larger due to the larger difference. When the current heat accumulation value exactly reaches the tolerance limit, the calculated C_r value is zero. This zero value is passed along the subsequent conversion chain, causing the available overload duration to synchronously decrease to zero, without a separate independent charging prohibition judgment. The calculation of the remaining temperature rise margin only covers the portion marked as restricted or regular charging periods in the basic charging plan. Each circuit is calculated separately for each sampling time. No calculation is performed during prohibited charging periods, as calculating the margin during these periods is meaningless since no charging tasks are scheduled. The critical point between restricted and regular charging periods is directly defined according to the period classification results without additional transition processing.

[0031] The corrected heat capacity difference is obtained by correcting the fluctuation range based on the remaining temperature rise margin. The original calculated value of the remaining temperature rise margin is based on the estimated temperature rise at a single moment. However, the estimated temperature rise itself has a reasonable error range due to factors such as short-term fluctuations in ambient temperature and current reading noise. The fluctuation range correction deducts the margin corresponding to this error range to prevent numerical fluctuations within the error range from being mistaken for the actual usable heat capacity space. The width of the fluctuation range is taken as the upper envelope value of the historical deviation distribution between the estimated temperature rise and the measured temperature rise recorded in the recent calibration of the cable heat accumulation curve of this circuit. The larger the deviation, the wider the fluctuation range of the circuit. The deviation data comes from the same calibration records in the estimated current-carrying heat accumulation stage and is not repeatedly collected or recalculated. The statistical window of the upper envelope value takes a fixed number of calibration records from the most recent period. Earlier records exceeding this number are not included to ensure that the width of the fluctuation range can reflect the measured deviation level of the circuit in a timely manner, rather than in the distant past. The corrected heat capacity difference is calculated by subtracting the fluctuation range width from the remaining temperature rise margin. If the subtraction result is negative, the corrected heat capacity difference is treated as zero, indicating that there is actually no verifiable remaining temperature rise margin at that moment. In this case, no additional overload charging task should be arranged for this circuit. The fluctuation range width is not a fixed value and gradually narrows as the circuit calibration process continues. The more historical calibrations a circuit has, the smaller the fluctuation range width will be. For newly commissioned circuits with insufficient calibration history, the initial fluctuation range width is temporarily substituted with the empirical value of circuits of the same model that have accumulated sufficient calibration data. If the fluctuation range width data is missing at a certain sampling moment, the corrected heat capacity difference is temporarily treated as the original value of the remaining temperature rise margin. It will be corrected again after the data is supplemented. The correction calculation of the entire curve will not be interrupted due to the missing data of a single sampling point.

[0032] The temperature rise margin ratio is constructed based on the margin ratio calculated using the corrected heat capacity difference. The margin ratio is calculated as the ratio of the corrected heat capacity difference to the withstand limit T_limit. A larger ratio indicates a higher proportion of the remaining usable heat capacity relative to the overall withstand capability of that cable type at that moment. The ratio itself is not affected by the absolute value of the withstand limit. A proportional form is used instead of directly using the absolute value of the corrected heat capacity difference because the withstand limits of different cable types vary significantly, and absolute values ​​are not comparable across circuits or models. The proportional form normalizes these differences, allowing the margin performance of different cable types and cross-sectional areas to be compared on the same scale. For example, the absolute value of the corrected heat capacity difference between a 95 square millimeter cross-sectional area cable and a 16 square millimeter cross-sectional area cable may differ by several times, but after conversion to a ratio, they can be directly compared. The temperature rise margin ratio ranges from zero to one. A ratio close to zero indicates that the remaining temperature rise margin is almost exhausted, while a ratio close to one indicates that the cable's heat accumulation level is still far below its tolerance limit, meaning the circuit has significant additional overload charging space. This ratio is read in real-time using the same formula during both the sudden increase in current and the stable load phase, without distinguishing between different scenarios. Before and after a short-term peak caused by a sudden current surge in the same circuit, the temperature rise margin ratio will show a rapid decrease followed by a slow recovery. Cables with a larger thermal time constant will experience a slower temperature rise after the surge, and the ratio will take longer to return to its pre-surge level. The temperature rise margin ratio is calculated separately for each circuit without additional smoothing, preserving the true and rapid change in the ratio itself due to the sudden current surge. This ensures that cross-circuit comparisons reflect the actual heat accumulation surplus state of each circuit at that moment, rather than having the true fluctuations masked by smoothing.

[0033] The temperature rise margin is obtained by converting the margin ratio. When converting the temperature rise margin ratio back to its absolute value, the ratio is multiplied by the corresponding loop's tolerance limit T_limit to obtain the conversion result, which is the temperature rise margin. The tolerance limit T_limit used for conversion is consistent with the value used when calculating the remaining temperature rise margin, without re-looking up tables or making adjustments, ensuring that all values ​​use the same benchmark throughout. Theoretically, the conversion process should be consistent with the corrected heat capacity difference value. In practice, because the ratio calculation and conversion process are rounded independently, there may be a very small rounding deviation between the two, which is negligible in the temperature difference narrowing calculation stage. The method of using a ratio conversion and then converting back to the absolute value, rather than directly using the corrected heat capacity difference as the final result, is to allow the temperature rise margin ratio itself to be used as an independent indicator for cross-loop comparison of the heat accumulation surplus of each loop in the branch box. For example, if multiple loops consistently show a low temperature rise margin ratio, it indicates that the heat dissipation conditions of the branch box itself may need to be improved. Each circuit determines a specific value of its temperature rise margin at each sampling time. The unit of this value is consistent with that of the remaining temperature rise margin and the corrected heat capacity difference, all of which are physical units of temperature rise difference. The conversion relationship between the three values ​​remains fixed with the circuit's tolerance limit and does not change over time. A circuit with a temperature rise margin of zero or close to zero indicates that the circuit currently has virtually no room to carry additional charging loads. Such circuits will directly obtain a very short or even zero available overload time in the subsequent temperature difference narrowing calculation. At this time, it is not advisable to arrange any additional overload charging tasks. It is necessary to wait for the load to drop and the temperature rise margin to reappear before making an evaluation.

[0034] The usable overload duration is determined by calculating the temperature difference narrowing based on the temperature rise margin. Temperature difference narrowing refers to the process where, under the assumption that the charging branch is continuously charged with a certain overload current, the temperature difference corresponding to the temperature rise margin gradually decreases over time until it reaches zero. Physically, this means that the additional heat generated by the overload current gradually consumes the remaining heat capacity. The usable overload duration can be estimated using the simplified thermal balance relationship t_avail=C_m / (ΔI²×R), where C_m is the temperature rise margin, ΔI is the increment of the assumed overload current relative to the rated current, R is the equivalent heating coefficient per unit length of the cable, converted to match the temperature rise rate, and t_avail is the usable overload duration. Assuming the specific value of the overload current is based on the upper limit of the rated charging current of the energy storage charging equipment in that branch, different branches with different specifications of charging equipment correspond to different assumed overload current benchmarks, with branches with larger rated currents corresponding to higher benchmark values. If the upper limit of the rated charging current changes due to the replacement of charging equipment in that branch, such as when the same model of charging equipment is upgraded from a rated 100A to 150A, the assumed overload current increases accordingly. The assumed overload current is adjusted synchronously according to the parameters of the new equipment. The iterative calculation of temperature difference narrowing is performed in seconds, consistent with the step size used to calculate current-carrying heat accumulation, to prevent additional errors introduced by cross-conversion of the two sets of step sizes. Both share the same sampling time axis, and each iteration synchronously produces a new set of available overload duration values. The available overload duration calculated by the formula for a circuit with zero temperature rise margin is also zero, indicating that the circuit does not have any additional overload charging time window at that moment. The available overload duration of different branches is estimated independently and does not affect each other. The estimation results of other branches are not borrowed from each other. Even if the current heat capacity of an adjacent circuit is sufficient, it cannot be used to calculate the overload duration supporting this circuit.

[0035] The short-time overload power is obtained by calculating the upper limit of the overload power using the available overload duration. The short-time overload power is calculated using the formula P_over=P_rated×[1+(β_max-1)×min(t_avail / t_ref,1)], where P_rated is the rated charging power of the branch circuit, taken from the nameplate parameters of the energy storage charging equipment of the branch circuit, β_max is the maximum short-time overload multiple nominally stated by the manufacturer of the charging equipment model, t_ref is the reference sufficient duration for using full multiple when this value is reached, t_avail is the available overload duration, and P_over is the short-time overload power. When t_avail reaches or exceeds t_ref, the multiple is taken as full β_max. When t_avail approaches zero, the multiple approaches 1, that is, no additional overload is allowed. The reference allowance duration t_ref is set as a fixed multiple of the thermal time constant of this cable model. The larger the thermal time constant of the cable, the longer the corresponding t_ref, because the large cross-section cable itself has a wider tolerance window for short-term overload. When t_avail takes a value between t_ref and zero, the multiple interpolates linearly between 1 and β_max without segmented jump processing. The upper limit of the multiple does not exceed the short-term overload capacity upper limit β_max specified by the equipment manufacturer. Even if the multiple calculated by the formula on the cable side still has a margin, the power limit of the equipment itself is always an insurmountable hard constraint. This upper limit is directly capped in the formula by β_max, and there is no need to compare and take the lower value separately. If the replacement of the charging equipment in this branch causes changes in P_rated or β_max, both parameters are simultaneously replaced with the parameters on the nameplate of the new equipment and substituted into the formula, and the values ​​corresponding to the old equipment are not used. The short-time overload power is calculated for each circuit individually, and the available overload duration of different circuits is not substituted into each other. The short-time overload power of the same circuit changes at different times due to the different available overload durations. For circuits where the temperature rise margin and available overload duration approach zero, the short-time overload power calculated by the formula approaches the original value of the rated charging power, meaning that there is no longer any additional overload space. At this time, the circuit can only be charged normally at the rated power, and no additional overload charging tasks are arranged.

[0036] Charging power quotas are determined based on the branch capacity allocated to short-term overload power. When the short-term overload power of multiple branches within the same branch box is aggregated, it is compared with the available capacity margin of the upstream distribution capacity. If the aggregated value does not exceed the available margin, each branch's original short-term overload power value is directly used as its charging power quota, without further capacity allocation adjustments between branches. If the aggregated value exceeds the available capacity margin of the upstream distribution capacity, the charging power quota of each branch is allocated and converted based on the proportion of its short-term overload power to the aggregated value. After conversion, the sum of the branch quotas does not exceed the available margin. For newly commissioned branches, since there is no historical data on short-term overload power, their rated charging power is temporarily used in the conversion calculation. The normal conversion method will be switched after the first short-term overload power value is accumulated. When allocating and converting power, branches with a high heat peak level are given priority to retain their original short-term overload power quota, while the reduction is more significant and allocated to branches with lower heat peak levels. This reflects the principle that circuits with higher risk levels should be given priority in receiving surplus capacity. For example, if the total short-term overload power of three branches in a branch box exceeds the available capacity of the upper-level distribution network, the branch with a high heat peak level retains its original quota without reduction, while the reduction is concentrated on the other two branches with lower heat peak levels. The charging power quota value is dynamically adjusted according to the real-time changes in short-term overload power and the available capacity of the upper-level distribution network. The charging control command generation process, which uses the difference in heat accumulation, directly adopts the latest adjustment result and does not cache old values. The conversion ratio is calculated and recorded separately for each branch. For example, if a dispute arises during a scheduling event, this record can be used to restore the actual proportion of each branch at that time.

[0037] Charging control commands are generated by reverse allocation based on the difference in thermal accumulation using charging power quotas and cable thermal accumulation curves. Before converting the charging power quota into an executable power setpoint, each branch first reads the difference between the calculated temperature rise and the withstand limit on the cable thermal accumulation curve at that moment as the thermal accumulation difference for that circuit. The smaller the difference, the closer the current thermal accumulation of that circuit is to saturation. Based on the sum of the thermal accumulation differences of all circuits participating in this round of allocation within the same branch box, the power setpoint of each circuit is reduced inversely proportional to the ratio of the thermal accumulation difference of that circuit to the total difference. The circuit with the smallest difference has the largest reduction, while the circuit with the largest difference is issued the original charging power quota without reduction. This reverse allocation mechanism ensures that the circuit with the heaviest thermal accumulation pressure always receives the most conservative charging arrangement. When the reduced power setpoint is lower than the minimum adjustable power limit of the energy storage charging equipment in that branch, the power setpoint in the corresponding circuit's charging control command for this round is treated as zero. That is, charging of that branch is not arranged for this round until the thermal accumulation difference recovers and the reduced power exceeds the lower limit before issuing the command again. The power setting and effective duration for each branch are encapsulated and sent within the same charging control command message. These two cannot be separated and sent individually; otherwise, there is a risk of power and time limit information being out of sync during device execution. If the cable thermal accumulation curve continues to rise within the effective duration, causing the thermal accumulation difference to narrow to below the set tolerance, the charging control command immediately triggers a step-down power reduction without waiting for the effective duration to expire. The switching process is smooth and seamless, without any sudden power drop. The power adjustment step size is the minimum adjustable level calibrated by the energy storage charging equipment for that branch at the factory. The finer the level, the smaller the proportion of credit loss due to rounding during reverse distribution.

[0038] Step S103: Calculate the state of charge based on the charging control command to obtain the dischargeable capacity, use the cooling hysteresis branch within the dischargeable capacity to arrange the support order to obtain the discharge priority sequence, and insert the hysteresis cooling gap into the discharge priority sequence to output the discharge scheduling command.

[0039] Specifically, the dischargeable capacity is obtained by calculating the state of charge (SOC) based on the charging control command. The SOC is calculated iteratively using the formula SOC(t) = SOC(t-1) + (P_charge(t) × η_c - P_discharge(t) / η_d) × Δt_h / E_rated, where P_charge(t) is the actual charging power issued by the charging control command at time t, η_c is the charging efficiency, P_discharge(t) is the actual discharging power at that time, η_d is the discharging efficiency, E_rated is the rated capacity of the energy storage device, Δt_h is the iteration step size (in hours to match the dimensions of power and capacity), and SOC(t) is the SOC at time t. The power setting value recorded in the charging control command serves as the basis for the value of P_charge(t). During periods when no charging command is issued or the power setting value is zero, the corresponding P_charge(t) is treated as zero, and the SOC is only affected by the discharging term during those periods. The dischargeable capacity is determined by the difference between the current state of charge and the set lower charge limit. This difference is multiplied by the rated capacity of the energy storage device to convert it into actual usable power. The lower charge limit is set considering the cycle life protection requirements of the energy storage battery itself, and does not allow discharge to zero. For example, the lower charge limit for a certain battery model is 10%, not zero. Different chemical systems correspond to different lower charge limit values; the typical lower limits for lithium iron phosphate batteries and ternary lithium batteries differ. The specific values ​​refer to the protection parameters preset by the battery management system at the time of device delivery and do not change dynamically with subsequent scheduling. E_rated uses the rated capacity indicated on the energy storage unit's nameplate at the time of delivery and is not recalibrated with the increase in service life. The initial value of SOC(t) is based on the measured state of charge reading detected during the first power-on of the unit. Thereafter, it is iteratively calculated according to the formula at each moment, without relying on external input for recalibration. The initial reading is only recorded once after unit replacement or major overhaul. If multiple branch circuits within the same branch box are each equipped with independent energy storage units, the discharge capacity is calculated separately for each energy storage unit, and the state of charge and discharge capacity of different units are not shared.

[0040] In some embodiments, the step of arranging the support order to obtain a discharge priority sequence using the cooling hysteresis branches within the dischargeable capacity includes: analyzing the temperature drop curve after charging is stopped based on the cooling hysteresis branches within the dischargeable capacity to obtain the actual cooling rate; identifying hidden heat sources and generating hidden heating branches from the charging rate cooling misalignment segment of the actual cooling rate; assessing the support urgency based on the hidden heating branches to determine the support order; and arranging the power resource allocation through the support order to obtain a discharge priority sequence.

[0041] The actual cooling rate is obtained by analyzing the temperature drop curve after charging is stopped based on the cooling hysteresis branch within the discharge capacity. For branches whose discharge capacity exceeds a set threshold, the drop segment of the cable's thermal accumulation curve after the most recent charging task is compared with the theoretical drop rate calculated by the natural cooling term of the first-order thermal inertia model. Branches with a drop rate significantly slower than the theoretical expectation are marked as cooling hysteresis branches. The comparison window is set according to a fixed observation duration after the charging task ends. The observation duration is a fixed multiple of the thermal time constant of the cable model; cables with larger thermal time constants have longer observation windows to ensure the complete natural cooling process is covered within the window. If the discharge capacity of the branch falls below the threshold during the observation window, the current determination is postponed and recalculated in the next charging cycle. The temperature drop curve after charging is stopped for the cooling hysteresis branch is taken from the end of the charging task of that circuit until the estimated temperature rise drops back to the baseline level before charging. The theoretical drop rate is calculated according to the natural cooling term of the first-order thermal inertia model, without considering the theoretical expectation of additional heat sources. The actual cooling rate is taken as the average slope of the temperature rise value decrease segment in the temperature drop curve after charging is stopped. The slope calculation excludes local rebound segments caused by several short-term fluctuations in the curve, and only counts the data points corresponding to the overall downward trend. Local short-term rebounds are not included in the numerator and denominator of the average slope. The actual cooling rate of the same cooling hysteresis branch after different charging tasks is calculated independently. No averaging is performed between different results. Because the charging rate and ambient temperature during charging are different in different charging tasks, averaging will mask the actual cooling anomaly that occurred after a certain charging. This anomaly is kept separately in the original record of the corresponding charging cycle for traceability.

[0042] By identifying the hidden heat sources and generating branches in the charging rate-induced cooling misalignment segment based on the actual cooling rate, and by calculating the expected cooling rate based on the actual cooling rate of each branch after charging is stopped, the charging rate of the last charge is used as the ratio of the short-term overload power of the last charge to the rated charging power. The higher the charging rate, the more heat is accumulated during the last charge, and the greater the temperature difference between the cable and the environment at the moment of charging stop. Theoretically, the cooling should be faster. The mechanism is that the heat generation increases with the square of the charging current. The actual cooling rate is compared with the expected cooling rate segment by segment. The segment where the actual cooling rate is significantly slower than the expected rate is identified as a charging rate cooling misalignment segment. The ratio of the two is used as the judgment index. The segment with the ratio below the lower limit is marked as misalignment and included in the candidate for hidden heat generation branch. The lower limit is adaptively relaxed according to the standard deviation of the measured cooling rate of the cable model under multiple natural cooling processes in the absence of hidden heat sources. The slower cooling during the high temperature period in summer than during the low temperature period in winter in the same cable trench is a normal seasonal fluctuation. The larger the standard deviation, the more relaxed it is, so as to avoid misjudging the natural slowdown caused by the ambient temperature difference as a charging rate cooling misalignment. In reverse, identify latent heat sources on the marked misaligned sections. These latent heat sources originate from factors that continuously apply additional heat load to the loop cable beyond charging. For example, temporary construction loads that continue operating within the same cable tray after charging has ended inject additional heat into the cable through the connection point, or the increasing heat generation due to insulation loss after the cable has exceeded half its design life. These factors are not reflected in the charging control command records and can only be indirectly inferred through abnormal cooling in the latent heat-generating branch. Identifying the misaligned section corresponding to the latent heat source generates the latent heat-generating branch. If multiple misaligned sections appear successively on the same branch, each section is recorded independently and not combined in the statistics, as the latent heat-generating causes corresponding to different misaligned sections may be different.

[0043] The urgency of support for hidden heat-generating branches is used to determine the priority of support. The circuit corresponding to a hidden heat-generating branch has accumulated cable heat due to the combined heating contributions of charging and the hidden heat source. The urgency assessment determines whether this circuit needs priority support from energy storage discharge to share its electrical load and indirectly reduce heating pressure. After load sharing, the current borne by the circuit itself decreases, helping to alleviate the pressure caused by the continuous heating of the hidden heat source. The urgency of support is graded according to the severity of the deviation between the actual cooling rate of the hidden heat-generating branch and the theoretical expectation. The greater the deviation, the higher the urgency rating. For example, if the ratio of a branch is only half of the theoretical expectation, the corresponding rating is classified as the highest level. For conventional circuits not identified as latent heat-generating branches, the support urgency rating is "Conventional." The rating is not upgraded solely based on the circuit's discharge capacity; the urgency rating reflects only the latent heat generation itself. The adequacy of discharge capacity is considered in conjunction with the support order during the power resource allocation process. These two factors are deliberately handled separately during the rating phase. After each charging task, the newly generated actual cooling rate replaces the old value in the rating calculation. The rating itself does not retain any weighting traces from historical ratings. The support order is arranged from highest to lowest support urgency rating. Circuits with the same rating are ranked secondary based on the ratio of their actual cooling rate to their theoretically expected cooling rate, from lowest to highest. If a circuit's latent heat-generating branch record is determined to be eliminated (i.e., no misalignment occurs after multiple consecutive charging tasks), the circuit's support urgency rating is downgraded to "Conventional," and its position in the support order shifts to make way for other circuits whose latent heat generation continues.

[0044] The power resource allocation is based on the support priority ranking to obtain a discharge priority sequence. Circuits ranked higher than those in the support priority ranking receive priority in the allocation of discharge capacity. The allocation amount is determined by the formula Q=min(D_req,C_avail), where D_req is the required discharge amount calculated based on the support urgency rating of the circuit, C_avail is the current remaining allocable discharge capacity of the circuit, and Q is the actual allocation amount. The required discharge amount D_req is set according to the support urgency rating, with higher ratings corresponding to larger required discharge amounts. The specific value is a fixed proportion of the rated load level of the circuit. When a single energy storage unit supports the discharge of multiple branch circuits simultaneously, the discharge capacity is allocated sequentially according to the support order. The circuits ranked first are deducted first, and the circuits ranked last can only use the remaining portion after deduction. The deduction order is strictly based on the support order and is not adjusted due to circuit number or physical location. Taking an energy storage unit that supports three branch circuits simultaneously as an example, the circuit ranked first in the support order is deducted first. If the capacity is insufficient after deduction, the second and third ranked circuits are allocated in turn. If the remaining dischargeable capacity allocated to a circuit is insufficient to meet the circuit's required discharge amount D_req, the actual allocation for that circuit will be given in full based on the remaining amount, without forcibly making up the required discharge amount. The unmet amount will be simultaneously recorded in the corresponding gap mark in the discharge priority sequence for that circuit, and will be tried again after the support order is re-arranged in the next round. The gap mark also records the specific difference between the actual allocation amount and the required discharge amount in this round. This difference is the direct basis for judging whether the gap has been filled in the next round of calculation. The difference will not be permanently deducted from the required discharge amount due to the temporary shortage in this round. If the same circuit has gaps in multiple consecutive rounds, the gap mark in the discharge priority sequence will be accumulated in each round and will not be cleared and recalculated due to cross-round.

[0045] For discharge priority sequences, a hysteresis cooling gap is inserted to output discharge scheduling commands. The actual allocation quota of each circuit in the discharge priority sequence is converted into two set values: discharge power and discharge duration that the energy storage unit of that circuit can directly execute. The conversion method refers to the rated discharge power upper limit of the energy storage unit, and the converted discharge power does not exceed this upper limit. The power is rounded down to prevent the actual power delivered from exceeding the actual allocation quota of this round. After the discharge duration of each circuit is determined, based on the actual cooling hysteresis after the cable heat accumulation curve of that circuit stops charging, a hysteresis cooling gap is inserted between the end of its discharge and the start of the next circuit's discharge. The gap duration is a fixed proportion of the time required for the temperature rise to fall back to the pre-discharge baseline after the circuit stops charging, ensuring that the additional heat generated by the previous circuit's discharge is fully dissipated during the gap period, preventing heat accumulation in the cable from rising round by round due to continuous discharge. Discharge scheduling instructions are generated and issued independently for each circuit. The issued instructions include the discharge power setting, discharge duration, and the duration of the hysteresis cooling gap after the discharge. During the gap, the actual temperature drop of the cable can be checked on-site to see if it matches the expectation. If the temperature drop is found to be significantly slower than expected, the maintenance personnel will report this through the maintenance terminal and include it in the adjustment of the gap duration for the next round of discharge for that circuit. For circuits with a support urgency rating of "normal," no additional discharge support will be arranged in the discharge scheduling instructions. The energy storage unit of this type of circuit will remain in standby mode and will not be forced to participate in the allocation due to the overall discharge demand of the branch box. This reserves discharge capacity for the actual overload charging demand that may occur at any time in this circuit. During the standby period, the energy storage unit will simultaneously maintain a basic self-monitoring status.

[0046] Step S104: Collect the operating status data of the unit according to the discharge scheduling command, track the evolution of the internal resistance degradation curve based on the unit status data to form a safety margin rating, and determine the protection action based on the safety margin rating to generate a protection limit signal.

[0047] Specifically, the unit status data is acquired based on the discharge scheduling instructions. The operational status data of the energy storage unit is collected based on the terminal voltage and charging / discharging current of each individual battery cell. The timing of the data acquisition is strictly aligned with the discharge period specified in the discharge scheduling instructions. Data acquisition begins one sampling cycle before the start of discharge and continues until the terminal voltage stabilizes after the discharge ends. The discharge power setting and discharge duration given in the discharge scheduling instructions serve as the basis for defining the length of the acquisition window. The sampling density is correspondingly increased for discharge tasks with higher power setting values. Because the individual cell status changes more rapidly during high-current discharge, sparse sampling can easily miss instantaneous voltage drops, which are precisely the key sampling points for judging the internal resistance of individual cells in the unit status data. When the energy storage unit remains in standby mode and has not received any discharge support arrangements, the data acquisition continues according to the low-frequency base cycle. This is because the self-discharge performance of individual cells during standby is also a basis for judging the health status of the unit, and abnormal self-discharge often appears in the data earlier than internal resistance degradation. The accuracy of single-cell voltage acquisition needs to reach the millivolt level. This accuracy requirement stems from the extremely small magnitude of the slight increase in internal resistance. The reading noise at the microvolt level is sufficient to make the slight increase in internal resistance within the early loss range completely undetectable. The unit status data is doubly identified by the energy storage unit number and the single-cell number. Data from different cells within the same energy storage unit are stored independently. Since the consistency differences between cells are themselves an important clue to determine whether local degradation has occurred in the unit, if they are combined and stored as a unit-level average, the abnormal performance of individual cells can easily be smoothed out by the other normal cells and become undetectable.

[0048] In some embodiments, the step of tracking the evolution of the internal resistance degradation curve based on the cell state data to form a safety margin rating includes: generating an internal resistance degradation curve by analyzing the internal resistance change trajectory during the charge and discharge cycle based on the cell state data; calculating the internal resistance degradation rate of the continuously deviating and non-falling-back segment of the internal resistance degradation curve; estimating the lifetime based on the narrowing segment of the rate of increase of the internal resistance degradation rate to obtain a lifetime margin index; and converting the lifetime margin index into a safety level to form a safety margin rating.

[0049] An internal resistance degradation curve is generated by analyzing the internal resistance change trajectory during charge-discharge cycles based on cell state data. The voltage change ΔU and current step amplitude ΔI_d before and after each current step are extracted from the sampling sequence of cell state data. The internal resistance is estimated successively using the formula R_int = ΔU ÷ ΔI_d, where ΔI_d is the actual measured current step at the same sampling moment on the discharge side, distinct from the assumed overload current increment ΔI, and R_int is the estimated internal resistance value corresponding to the step. The estimated internal resistance is compared cycle by cycle along the complete sampling sequence of cell state data. A slight increase in internal resistance, where the increase is not yet at the existing alarm threshold, falls within the allowable range of the normal wear curve for this battery model. This slow increase is precisely the part most easily missed by conventional threshold alarms; only continuous tracking across cycles can reveal its cumulative trend. For each charge-discharge cycle, the median of the estimated internal resistance values ​​corresponding to all current steps within that cycle is taken as the cycle representative value. The median is more effective than a simple average in resisting interference from individual steps that deviate significantly from the true internal resistance due to instantaneous current fluctuations or sampling timing deviations. Sampling points marked as pending verification in the cell status data are temporarily suspended in this internal resistance estimation. These sampling points typically originate from momentary disturbances in the acquisition circuit, including situations such as voltage drops causing instantaneous disturbances and brief clock synchronization failures. Arranging the representative internal resistance values ​​of each cycle according to the charge / discharge cycle number generates the internal resistance degradation curve. The curve shows a step-like decrease at the cycle where the battery pack has undergone equalization maintenance; this decrease is a normal recovery resulting from maintenance and is marked as a maintenance node to distinguish it from a true degradation reversal. Each cell within the same energy storage unit generates its own internal resistance degradation curve. Significant divergence in the curve trends between cells usually indicates that localized, consistent degradation has occurred within that unit.

[0050] Calculate the rate of internal resistance degradation in the rising segment of the internal resistance degradation curve that deviates continuously and does not decrease with each cycle. Compare the representative internal resistance value with the previous baseline level cycle by cycle along the internal resistance degradation curve. A continuous deviation is defined as a value that is consistently higher than the baseline for multiple cycles, with the deviation increasing cycle by cycle rather than fluctuating around a certain level. Distinguish this from the pseudo-rise caused by fluctuations in charge / discharge depth. Cycles with larger charge / discharge depths will naturally push up the estimated internal resistance value, but the representative internal resistance value will also fall after the depth decreases. This type of rise does not signify degradation and is represented on the internal resistance degradation curve as a waveform that fluctuates synchronously with the charge / discharge depth, clearly different from the unidirectional upward trend of true degradation. The key criterion for distinguishing between the two is that the rising segment begins at the cycle where the representative internal resistance value first continuously deviates from the baseline, and ends at the latest cycle accumulated on the internal resistance degradation curve. If a step down is encountered during this period, marked as a maintenance node, the rising segment is truncated at that node, and a new rising segment is calculated after maintenance. The internal resistance degradation rate is calculated by performing a linear regression of the representative internal resistance value within the rising segment against the cycle number, with the unit being the internal resistance increment per cycle. The goodness of fit of the regression is calculated simultaneously. Rising segments with excessively low goodness of fit are considered to have unclear trends, and rate conclusions are temporarily withheld. The goodness of fit threshold is taken as the lower quartile of the goodness of fit distribution of rising segments with confirmed real degradation in the history of this battery model. Rising segments below this threshold are mostly not true unidirectional degradation trends. The minimum number of sample points required for linear regression is linked to the minimum number of cycles; insufficient sample points are also marked as under-observed. For individual cells whose rising segment spans fewer than the minimum number of cycles set, the internal resistance degradation rate is marked as under-observed, and this cell is skipped in this round of rate estimation until the sample size is sufficient before being included.

[0051] Lifetime margin indicators are estimated based on the narrowing rate of internal resistance degradation. The narrowing rate of internal resistance degradation, after a period of rapid increase, is defined as the period when the rate of increase slows down significantly. In the battery aging mechanism, this corresponds to the period when by-products at the electrode interface accumulate to a certain thickness and then stabilize, leading to a slowdown in the increase of interface impedance. Lifetime estimates given at this stage are closer to the actual degradation rate of the individual cell. The lifetime margin indicator is estimated by substituting the internal resistance degradation rate *v* within the narrowing rate of increase into the formula *L=(R_th-R_now)÷v*, where *R_th* is the internal resistance threshold for determining the end of life for this battery model, *R_now* is the current representative internal resistance value, and *L* is the number of charge-discharge cycles that can continue operating at the current rate. *R_th* is taken from the internal resistance value corresponding to the lifetime termination criterion provided by the battery manufacturer for this model; if no manufacturer data is available, it is taken from the internal resistance level corresponding to the capacity decay to a set percentage during actual cycle life testing of this battery model. For cells where the growth rate has not yet narrowed and the internal resistance degradation rate is still in a rapid growth phase, the current instantaneous rate is used to estimate the lifespan margin indices. The resulting estimates are conservative and should be revised upwards once the cell enters a stable phase. If the internal resistance degradation rate rises significantly again after the growth rate narrows, it indicates that the battery has entered an accelerated degradation phase. The lifespan margin indices are immediately recalculated based on the new rate, not the optimistic estimates from the stable phase. The remaining number of cycles in the accelerated degradation phase is often much less than the estimated value from the stable phase. When there are differences in the lifespan margin indices of individual cells within the same energy storage unit, the overall lifespan margin of the unit is determined by the cell with the smallest value. In a series structure, if any cell reaches the end of its lifespan first, it will limit the availability of the entire unit; even if the remaining cells have sufficient margins, they cannot continue to perform discharge tasks independently.

[0052] A safety margin rating is formed by converting the lifespan margin index into a safety level. The remaining charge / discharge cycles calculated from the lifespan margin index are divided by the total number of cycles expected within the design life of the energy storage unit to obtain the remaining lifespan percentage. The safety level is divided into three categories based on this percentage: Ample, Watchlist, and Critical. The specific percentage values ​​for the Ample and Critical thresholds refer to the health grading recommendations given in the battery model's design document. Different battery models may have different threshold values ​​due to differences in their chemical systems. Units with a remaining lifespan percentage higher than the set Ample threshold are rated as Ample. These units can undertake all discharge support tasks as usual without any additional restrictions at the scheduling level, and the safety margin rating at this level does not affect normal discharge arrangements. Units with a percentage between the Ample and Critical thresholds are rated as Watchlist. These units can still participate in discharge support, but the depth of discharge per cycle needs to be correspondingly reduced. The closer the percentage is to the Critical threshold, the more significant the reduction; for those that have just fallen below the Ample threshold, the reduction is relatively slight. Cells with a percentage below the critical threshold are classified as critical. These cells have limited remaining usable lifespan, and continuing to handle high-intensity discharges will accelerate their demise. This is because cells with already high internal resistance generate more heat under high current, and the thermal effect, in turn, accelerates interfacial side reactions. Cells with insufficiently observed cells are classified as "concerned" in terms of lifespan margin. Their safety margin rating is conservative rather than ample, because underestimating the risk when data is insufficient is far more costly than overestimating it. Overestimation might at most reduce the discharge load on the cell, while underestimation could allow a cell nearing the end of its lifespan to continue operating at full capacity. Once the observation data is complete, the cell's rating is reassessed according to normal rules, and the conservative approach used during the period of insufficient observation is no longer applied.

[0053] Protection actions are determined based on the safety margin rating, generating protection limitation signals. For energy storage units with a safety margin rating of critical, protection action is triggered. Specifically, the maximum discharge power and maximum depth of discharge of the unit are limited. The limited values ​​are proportionally reduced according to the unit's current remaining lifetime percentage. The reduction is initially slight when the percentage just falls into the critical range, but becomes increasingly severe as the percentage continues to decline, until only basic emergency support can be maintained. The specific reduction ratio is determined when the unit is first rated as critical and is recalculated accordingly as the remaining lifetime percentage continues to decrease, without using the initial reduction ratio from the time of critical rating. For units with a safety margin rating of concern, protection action is triggered only when their lifetime margin indicators show a significant decline compared to the previous period. Units simply in the concern range with stable indicators are not triggered, preventing concern-range units from being overly restricted and losing their inherent discharge support capabilities. The protection limit signal also provides the expected release conditions for this limit. These conditions typically arise when the unit's lifespan margin index recovers to above the critical threshold after equalization maintenance, or when testing confirms that the relevant cells have been replaced. Before these conditions are met, even if the remaining lifespan percentage slightly increases, the limit remains unchanged to prevent frequent switching of the limit state due to repeated conditions. If only a few cells within the same energy storage unit trigger the protection judgment, the protection limit signal is still issued for the entire unit. Because the cells within the unit operate in series, the current flowing through each cell is the same value, and the limit can only apply to the entire current path. If the power setting value of an ongoing discharge task exceeds the limit value, it will be immediately adjusted downwards to the limit value.

[0054] Step S105: Use the protection limit signal to release the hysteresis, calculate the emergency lower limit demand to obtain the emergency reserve capacity, determine the arbitrage allocation quota based on the emergency reserve capacity to generate a comprehensive adjustment plan, and execute peak shaving and valley filling optimization output charging and discharging commands for the comprehensive adjustment plan.

[0055] In some embodiments, the step of using the limitation release hysteresis of the protection limiting signal to calculate the emergency lower limit demand and obtain the emergency reserve capacity includes: aggregating the recurrence interval based on the limitation release hysteresis of the protection limiting signal to form a hysteresis duration distribution; calculating the risk margin by measuring the deviation of the hysteresis duration distribution from the expected peak hysteresis to generate an emergency risk margin; calculating the lower limit capacity based on the emergency risk margin to determine the emergency lower limit demand; and calculating the reserve coefficient based on the emergency lower limit demand to obtain the emergency reserve capacity.

[0056] The delay duration distribution is formed by aggregating the delay and recurrence interval based on the release delay of protection and restriction signals. The release delay and recurrence interval are extracted event-by-event from the historical records of protection and restriction signals. After aggregation, the data is divided into several intervals based on the delay duration. The number of restriction release events falling within each interval is counted to form the delay duration distribution. The restriction release delay is the lag time between the actual fulfillment of the release conditions and the actual release of the restriction after the protection and restriction signal is issued. The sources of lag include waiting for equalization maintenance scheduling, the time consumed in the detection and confirmation process, and the response time of manual confirmation. The recurrence interval is the time elapsed between the previous restriction release and the next issuance of the protection and restriction signal for the same energy storage unit. A shorter interval indicates that the problem reappears quickly after the restriction is released, and the previous state was not truly restored. The recurrence interval and restriction release delay together characterize the overall degree to which the unit is trapped by protection and restriction. The interval boundaries are jointly set based on the typical cycle of equalization maintenance scheduling and the conventional time consumed in the detection and confirmation process. The boundaries are not equally spaced; the shorter delay side is more finely divided. Changes in the number of events within this interval are most instructive for process improvement. When the same unit triggers protection restriction signals multiple times, the delay duration and recurrence interval for each instance are statistically analyzed independently. The causes and lifting paths for each restriction may be completely different; one instance may be prolonged due to scheduling delays, while another may be prolonged due to repeated delays in the detection process. If the event duration in a certain interval is significantly higher than that in adjacent intervals, it indicates that there is a common bottleneck in the restriction lifting process at that time level. This bottleneck itself is the practical basis for the necessity of reserving emergency capacity. The process bottleneck is difficult to eliminate in the short term, and leeway can only be left in capacity arrangements.

[0057] By analyzing the deviation of the hysteresis duration distribution from the expected peak hysteresis, the risk margin is calculated to generate the emergency risk margin. The interval with the highest concentration of events in the hysteresis duration distribution is identified as the peak hysteresis. The corresponding hysteresis duration, t_peak, is taken and compared with the estimated processing time when the release conditions were set. A significant deviation from the estimate indicates a significant discrepancy between the actual processing time and the design expectation. The emergency risk margin is calculated using the formula M_e = t_peak × p_rec, where t_peak is the peak hysteresis duration, p_rec is the probability of the protection limiting signal recurring within the same hysteresis duration, and M_e is the emergency risk margin in hours, multiplied by P_base to obtain the corresponding energy quantity. Different energy storage units within the same branch box are calculated unit by unit. Since the hysteresis duration distributions of different units are often different, cross-unit merging or averaging is not performed. p_rec is calculated as the ratio of the number of times the recurrence interval of the unit was shorter than t_peak to the total number of limit release events in the unit's history. The higher the ratio, the greater the possibility that the unit will trigger protection again within the window period before the limit is completely released, and the emergency risk margin is significantly amplified, reflecting its actual state of repeated limitation. For units where the peak hysteresis does not deviate significantly from the expectation and the hysteresis duration is concentrated near the estimated processing time, the emergency risk margin is equal to t_peak itself and is no longer amplified by p_rec. Therefore, the limit release process of such units is operating normally, and the release time is basically consistent with the design expectation. The larger the emergency risk margin value, the more likely the unit will not be able to provide normal discharge support for a long time during the protection limitation period. The power demand during this window period must be covered by other units or reserved capacity. If there is hidden heat that needs to be supported at this time, there will be a risk of no available capacity to respond.

[0058] The emergency lower limit demand is determined based on the emergency risk margin calculation. The emergency lower limit demand is the minimum energy storage capacity that must be reserved during the discharge support window during the coverage protection restriction period. It is estimated using the formula Q_min=M_e×P_base, where M_e is the duration corresponding to the emergency risk margin, P_base is the basic discharge support power level of the branch box during that period, and Q_min is the emergency lower limit demand, which is the product of the window duration and the required support power during that period. The basic discharge support power level is based on the median power of the actual discharge support during the history of this branch box due to reasons such as latent heating. If the peak value is reserved instead, the reserved capacity will be idle for most of the time. If the average value is reserved instead, it will be easy to be in short supply during periods of high demand. The median value is between the two and is closer to the actual support intensity of the branch box on a daily basis. The basic discharge support power level of different branch boxes is statistically analyzed separately and the same empirical value is not applied because the number of circuits, load composition and frequency of latent heating of each branch box often vary greatly. Newly commissioned branch boxes do not yet have their own historical discharge records. Initially, the empirical value of the branch box with accumulated data of the same model is used. After the data of the branch box has accumulated to a set amount, it is switched to the actual statistical value. The switching process is completed in one go and there is no weighted transition between the old and new values. When multiple energy storage units in the same branch box are simultaneously under protection and restriction, the emergency risk margins of each unit are substituted into the formula, and the calculated emergency lower limit demand is added together. Since the gap needs to be filled by fewer available units when multiple units are simultaneously absent, the requirement for reserved capacity is naturally superimposed. It cannot be estimated based on the most severe unit. The contribution of each unit is marked separately, and the reserved share occupied by each unit is recorded. They are not combined into a single total that cannot be divided.

[0059] The emergency reserve capacity is obtained by calculating the reserve coefficient based on the emergency minimum demand. The reserve coefficient is a multiple greater than 1, used to add a safety margin above the emergency minimum demand. The coefficient value is determined by referencing the historical distribution of the actual discharge support demand exceeding the estimated emergency minimum demand for that branch box. This distribution is adjusted in small steps according to the actual performance of the latest period at a fixed period (e.g., quarterly), and the coefficient will not jump significantly due to a single abnormal fluctuation. The emergency reserve capacity is obtained by multiplying the emergency minimum demand by the reserve coefficient. This value is the bottom line of energy storage capacity that cannot be occupied by arbitrage arrangements. Any arbitrage arrangements can only operate in the surplus above this bottom line. The reserve coefficient is not fixed. For branch boxes that have never experienced actual demand exceeding the emergency minimum demand, the coefficient gradually converges to 1. Conversely, for branch boxes that have repeatedly exceeded the emergency minimum demand, the coefficient is adjusted upward accordingly, so that the reserve level reflects the actual fluctuation characteristics of the branch box itself, rather than applying an empirical multiple that is the same for all branch boxes. At a certain moment, if all energy storage units are in normal operation and no protection or limitation signals are active, the emergency minimum demand is zero, and the emergency reserve capacity is also zero. In this case, all dischargeable capacity can be considered for arbitrage. The emergency reserve capacity is numerically constrained by the total capacity of the branch box's energy storage equipment. If the estimated capacity exceeds the total capacity, it indicates that the energy storage configuration of the branch box itself is insufficient to cover its support needs during the protection limitation period. This situation needs to be addressed at the capacity configuration level, not the scheduling level. Even the most sophisticated scheduling algorithm cannot conjure up non-existent energy storage capacity out of thin air. In this case, capacity should be expanded rather than further compressing the arbitrage space.

[0060] In some embodiments, the step of determining the arbitrage allocation quota based on the emergency reserved capacity and generating a comprehensive adjustment scheme includes: determining the proportion of releaseable capacity based on the proportion of periods of fluctuating narrowing in the emergency reserved capacity analysis; using the proportion of releaseable capacity to screen arbitrage windows to obtain arbitrageable capacity; using the portion of the arbitrageable capacity where the profit is higher than the lower limit to determine the allocation quota to form an arbitrage allocation quota; and rearranging the allocation of charging and discharging resources based on the arbitrage allocation quota to generate a comprehensive adjustment scheme.

[0061] The proportion of releaseable capacity is determined based on the percentage of periods of fluctuating load narrowing in the emergency reserve capacity analysis. Periods of fluctuating load narrowing are defined as times when the short-term fluctuations in the load curves of each circuit in the branch box are lower than the set narrowing criterion. During these periods, the load trend is stable and predictable, and the probability of sudden discharge demand is relatively low. Therefore, the necessity of locking up a large portion of the emergency reserve capacity decreases accordingly. The percentage of periods of fluctuating load narrowing is calculated as the ratio of the duration of these periods to the total duration of the statistical period. A higher percentage indicates a more stable overall load pattern for the branch box, and a larger portion of the emergency reserve capacity can be temporarily released for arbitrage. However, this release is a conditional and temporary arrangement; it will be withdrawn once the load pattern returns to drastic fluctuations. The release capacity ratio is obtained by multiplying the proportion of periods with narrowing fluctuations by a set release limit. The release limit exists to ensure that even if the load is stable throughout, a portion of the emergency reserve capacity is always retained and not released. This is because the triggering of protection limits is not necessarily related to whether the load is stable, and the degradation of battery internal resistance accumulates with charge-discharge cycles, regardless of the shape of the current load curve. The specific value of the release limit is set with reference to the proportion of the longest continuous period during which protection limit signals have never appeared simultaneously in the history of this branch box to the statistical period. The shorter this longest continuous period, the more significant the possibility of multiple cells triggering limits simultaneously cannot be ignored, and the release limit is set more conservatively accordingly. For branch boxes with a very low proportion of periods with narrowing fluctuations and drastic load fluctuations throughout, the release capacity ratio approaches zero, and the emergency reserve capacity is basically fully locked in this case. For branch boxes with proportions between the two extremes, the release capacity ratio changes linearly with the proportion value, without additional grading processing.

[0062] Arbitrage opportunities are identified by using the release capacity ratio to screen arbitrage windows. The total discharge capacity and locked emergency reserve capacity of the branch box are existing values ​​determined in previous steps. The core of arbitrage capacity lies in further releasing the reserved portion using the release capacity ratio: the portion outside the reserve, i.e., the surplus after subtracting the emergency reserve capacity from the total discharge capacity, is fully included in the arbitrage capacity; the emergency reserve capacity within the reserve is only included according to the share corresponding to the release capacity ratio. The higher the release capacity ratio, the larger the share released from the reserve. When the ratio is zero, no reserve is released, and the arbitrage capacity is only equal to the surplus outside the reserve. An arbitrage window refers to a time interval where there is a significant price difference in time-of-use electricity prices, falling within the period of narrowing price fluctuations. Both conditions must be met simultaneously. Periods with only price differences but drastic load fluctuations are not considered arbitrage windows, as real discharge support demand may arise at any time during such periods. Once arbitrage activities fill the capacity, there will be no way to respond to real demand. The selection of arbitrage windows also requires verification of whether the time period overlaps with the prohibited charging periods in the basic charging plan. Overlapping portions are removed from the arbitrage window because the constraint of cable heat accumulation takes precedence over arbitrage profits; even a high release capacity ratio cannot exceed this constraint. If the same time interval falls within the prohibited charging periods of multiple branch lines, the entire interval is removed from the arbitrage window as soon as any one branch line is deemed prohibited, without any partial retention due to other branch lines still being in regular charging periods. Arbitrage capacity is not pre-allocated among arbitrage windows. The actual amount used in a particular window is determined jointly by the price difference level of that window and the allocation quota determination process. Pre-allocation would actually prevent the window with the largest price difference from receiving sufficient quota. The specific allocation ratio is determined uniformly in the next stage when determining the allocation quota.

[0063] Arbitrage allocation quotas are formed by determining the portion of the arbitrage capacity where the profit exceeds the lower limit of the loss. The determination process unfolds in three steps within the total arbitrage capacity: First, arbitrage profits and lower limit losses are calculated and screened for each arbitrage window. Arbitrage profits are calculated by multiplying the charge / discharge price difference for that window by the planned electricity consumption. The lower limit of the loss is the battery life loss cost incurred by the energy storage equipment to complete one charge / discharge cycle (calculated based on the actual cycle life of the battery model under the current operating temperature range and the equipment purchase cost, not the ideal lifespan claimed by the manufacturer). Window where the arbitrage profit is not higher than the lower limit of the loss is directly eliminated even if there is a price difference and does not enter the subsequent allocation, because the arbitrage income of this type is insufficient to cover battery losses, and it is actually operating at a loss. Secondly, the selected windows are sorted from highest to lowest based on their yield (arbitrage profit per unit of electricity). Arbitrage capacity is allocated sequentially to the windows ranked higher, with each window's required electricity fully met before moving on to the next, until the quota is exhausted. If a window's required electricity exceeds its remaining arbitrage capacity, only the remaining amount is allocated; the shortfall is not allocated in this round, and the quota for subsequent ranked windows is not overdrawn due to the window's high yield. Thirdly, the allocation results are reduced based on each energy storage unit's safety margin rating: units rated as "watchful" or "critical" receive a corresponding reduction in their quota; if multiple energy storage units share a window, each unit's arbitrage allocation is further distributed based on its own safety margin rating and discharge capacity, with units with ample ratings and capacity receiving a larger share. After these three steps, the arbitrage allocation quota is given as a specific electricity value for each window and each unit. This fine-grained quota format is the necessary input for verifying timeline conflicts during the rearrangement of charging and discharging resource allocation.

[0064] A comprehensive adjustment plan is generated by rearranging charging and discharging resource allocation based on arbitrage allocation quotas. The charging and discharging arrangements for each window determined by the arbitrage allocation quota must be combined with the original charging time slots in the basic charging plan. Both must be checked against the same timeline for conflicts, ensuring the granularity of the check matches the time slot divisions of the basic charging plan to prevent missed conflicts at boundaries due to different granularities. If a charging action in an arbitrage arrangement falls within a prohibited charging period of the basic charging plan, the arbitrage arrangement is cancelled, and the corresponding quota is carried over to the adjacent available window in the arbitrage allocation quota. If the adjacent window is also unavailable, this portion of the quota is forfeited in this round. If the discharge action arranged for arbitrage overlaps with the implicit heat-generating support discharge already scheduled in the discharge dispatch instruction, the power requirements of both shall be added together to check whether they exceed the maximum discharge power of the unit. If they do, the support discharge shall take priority, and the arbitrage discharge shall be reduced accordingly, because the support discharge is directly related to the mitigation of cable heat accumulation, while the arbitrage discharge is only an economic optimization. If the power requirements of both do not exceed the upper limit, they shall be executed simultaneously according to their respective arrangements, without being forced to choose one due to overlapping time periods. The power values ​​of both shall be recorded separately in the integrated regulation scheme, and the total value shall still meet the maximum discharge power constraint of the unit. The integrated regulation scheme provides the charging and discharging actions and power values ​​of each energy storage unit for each time period on the time axis. The same unit shall not have charging and discharging arrangements in both directions at the same time. Such conflicts usually arise from arbitrage charging and support discharge colliding at the same time. The solution is to cancel the arbitrage charging at that time. The verification shall be carried out one by one for each energy storage unit, and the time axis of each unit shall be verified separately. Conflicts found in the verification of one unit shall not affect the integrated regulation scheme content already scheduled for other units.

[0065] The integrated regulation scheme executes peak shaving and valley filling optimization output charging and discharging commands. The optimization objective is to minimize the peak-to-valley difference of the daily load curve on the branch box side while maximizing arbitrage profits, all while satisfying the constraints already defined in the integrated regulation scheme. When there is a trade-off between the two objectives, minimizing the peak-to-valley difference takes priority, as peak shaving and valley filling are the primary functions of this system. The optimization process makes small adjustments around the power values ​​given in the integrated regulation scheme for each time period. The adjustment range is limited by both the available capacity and the upper limit of equipment power for that period, without overturning the scheme's established action direction. This is because the scheme itself has already undergone multiple checks on accumulated constraints and protection limits; overturning it would mean that all these checks are invalidated. The peak-to-valley difference is measured by the difference between the maximum and minimum values ​​of the daily load curve of the branch box. This difference is recalculated after each iteration of the optimization until the difference no longer narrows significantly or the maximum number of iterations is reached. Setting an iteration limit ensures that the optimization yields results within a limited time; scheduling scenarios do not allow for indefinite waiting for the optimal solution. The charging and discharging commands provide the charging and discharging power and start and end times for each energy storage unit for each time period based on the optimization results. Before issuing the command, it is checked again to see if there is a conflict with the power limit in the protection limit signal. In case of a conflict, the protection limit shall prevail. The power exceeding the limit in the optimization results is reduced on the spot. The priority of equipment safety is always higher than the profit indicator, and the optimization will not be reversed due to the decrease in arbitrage profit after reduction. If the actual state of charge of a unit deviates from the calculated trajectory during the optimization process by more than the tolerance, the charging and discharging commands for the remaining time period are re-optimized. This is because the deviation means that the available capacity for the remaining time period no longer matches the assumption at the time of the initial optimization.

[0066] To execute the methods described in the above embodiments and achieve the corresponding functions and technical effects, this application also provides the following system. See [link to system]. Figure 3 , Figure 3 This paper illustrates a structural block diagram of a low-voltage cable branch box energy storage peak shaving and valley filling power saving system 300 provided in an embodiment of this application, including: The heat peak identification module 301 is used to acquire branch box operation data, calculate current-carrying heat accumulation based on the branch box operation data to form a cable heat accumulation curve, and identify the duration of flat-top load through the short-term peak segment of the cable heat accumulation curve to determine the continuous heat peak period. The charging decision module 302 is used to arrange peak avoidance according to the continuous hot peak period to obtain a basic charging plan, measure the short-term overload power of the unsaturated heat accumulation section based on the basic charging plan to obtain the charging power quota, and generate charging control instructions by reverse allocation of the charging power quota and the cable heat accumulation curve according to the heat accumulation difference. The discharge decision module 303 is used to calculate the state of charge based on the charging control command to obtain the dischargeable capacity, arrange the support order to obtain the discharge priority sequence using the cooling hysteresis branch within the dischargeable capacity, and output the discharge scheduling command by inserting the hysteresis cooling gap for the discharge priority sequence. The safety detection module 304 is used to collect the operating status data of the unit according to the discharge scheduling command, track the evolution of the internal resistance degradation curve based on the unit status data to form a safety margin rating, and determine the protection action based on the safety margin rating to generate a protection limit signal. The instruction output module 305 is used to release the hysteresis of the protection limit signal, calculate the emergency lower limit demand to obtain the emergency reserve capacity, determine the arbitrage allocation quota based on the emergency reserve capacity to generate a comprehensive adjustment scheme, and execute peak shaving and valley filling optimization output charging and discharging instructions for the comprehensive adjustment scheme.

[0067] The aforementioned low-voltage cable branch box energy storage peak shaving and valley filling power saving system 300 can implement one of the low-voltage cable branch box energy storage peak shaving and valley filling power saving methods described in the above-described method embodiments. The options in the above method embodiments are also applicable to this embodiment, and will not be detailed here. The remaining contents of this application embodiment can be referred to the contents of the above method embodiments, and will not be repeated in this embodiment.

[0068] The purpose of the above embodiments is to reproduce and derive the technical solution of the present invention by way of example, and to fully describe the technical solution, purpose and effect of the present invention. The purpose is to enable the public to have a more thorough and comprehensive understanding of the disclosure of the present invention, and not to limit the scope of protection of the present invention.

Claims

1. A method for energy storage, peak shaving, and valley filling in low-voltage cable branch boxes to save electricity, characterized in that... include: Obtain branch box operation data, calculate current-carrying heat accumulation based on the branch box operation data to form cable heat accumulation curve, and identify the flat-top load duration through the short-term peak segment of the cable heat accumulation curve to determine the continuous heat peak period; A basic charging plan is obtained by arranging peak avoidance according to the continuous hot peak period. The short-term overload power of the unsaturated heat accumulation section is measured based on the basic charging plan to obtain the charging power quota. The charging power quota and the cable heat accumulation curve are used to generate charging control instructions by reverse allocation according to the heat accumulation difference. Based on the charging control command, the state of charge is calculated to obtain the dischargeable capacity. Using the cooling hysteresis branch within the dischargeable capacity, the support order is arranged to obtain the discharge priority sequence. A discharge scheduling command is output for the discharge priority sequence by inserting a hysteresis cooling gap. The unit status data is obtained by collecting the operating status according to the discharge scheduling command, and the evolution of the internal resistance degradation curve is tracked based on the unit status data to form a safety margin rating. The protection action is determined and a protection limit signal is generated based on the safety margin rating. By utilizing the delay in releasing the protection limit signal, the emergency lower limit demand is statistically analyzed to obtain the emergency reserve capacity. Based on the emergency reserve capacity, the arbitrage allocation quota is determined to generate a comprehensive adjustment scheme. Peak shaving and valley filling optimization output charging and discharging commands are executed for the comprehensive adjustment scheme.

2. The method according to claim 1, characterized in that, The step of identifying the duration of the flat-top load and determining the sustained heat peak period by analyzing the short-term peak segment of the cable heat accumulation curve includes: The rise edge steepness is identified based on the short-time peak segment of the cable heat accumulation curve to obtain the peak steepness characteristics; The candidate interval for a flat top is determined by identifying the steepness convergence curve in the section where the temperature rises with lag in the peak steepness feature. The flat-top load duration is obtained based on the repeatability of the verification cycle of the flat-top candidate interval. The duration of the flat-top load is used to determine the heat peak level and the duration of the continuous heat peak.

3. The method according to claim 1, characterized in that, The step of obtaining the charging power quota by measuring the short-term overload power in the unsaturated heat accumulation period based on the basic charging plan includes: The remaining temperature rise margin in the unsaturated heat accumulation section is calculated based on the basic charging plan to obtain the temperature rise margin margin. Based on the aforementioned temperature rise margin, the available overload duration is determined by calculating the narrowing of the temperature difference. The short-term overload power can be obtained by calculating the upper limit of the overload power using the available overload duration. The charging power quota is determined according to the branch capacity of the short-time overload power allocation.

4. The method according to claim 1, characterized in that, The step of arranging the support order to obtain the discharge priority sequence using the cooling hysteresis branches within the discharge capacity includes: The actual cooling rate is obtained by analyzing the temperature drop curve after charging is stopped based on the cooling hysteresis branch within the discharge capacity. From the actual cooling rate of the charging rate cooling misalignment segment, identify the hidden heat source and generate a hidden heating branch. The urgency of supporting the hidden heat-generating branches is assessed to determine the support order. The discharge priority sequence is obtained by allocating power resources through the support sequence coding.

5. The method according to claim 1, characterized in that, The process of tracing the evolution of the internal resistance degradation curve based on the unit state data to form a safety margin rating includes: Based on the unit state data, the internal resistance change trajectory during the charge and discharge cycle is analyzed to generate an internal resistance degradation curve. Calculate the rate of internal resistance degradation of the rising segment where the internal resistance degradation curve continues to deviate and does not fall back with the cycle. The lifetime margin index is obtained by estimating the lifetime based on the narrowing segment of the rate of increase of the internal resistance degradation. The aforementioned life margin index is used to convert safety levels to form a safety margin rating.

6. The method according to claim 1, characterized in that, The process of using the protection limit signal to release the hysteresis and calculating the emergency lower limit demand to obtain the emergency reserve capacity includes: Based on the protection restriction signal, the hysteresis collection recurrence interval is used to form a hysteresis duration distribution; By measuring the deviation of the hysteresis duration distribution from the expected peak hysteresis, the risk margin is used to generate an emergency risk margin. Based on the aforementioned emergency risk margin, the lower limit capacity is calculated to determine the emergency lower limit demand. The emergency reserve capacity is obtained by calculating the reserve coefficient based on the emergency lower limit demand.

7. The method according to claim 1, characterized in that, The process of determining the arbitrage allocation quota based on the emergency reserve capacity and generating a comprehensive adjustment plan includes: The proportion of releaseable capacity is determined based on the percentage of periods when fluctuations narrow during the analysis of the aforementioned emergency reserve capacity. The arbitrage window is obtained by using the releasable capacity ratio; The portion of the arbitrage capacity where the profit exceeds the lower limit is allocated to form an arbitrage allocation quota; A comprehensive adjustment scheme is generated by rearranging the charging and discharging resource allocation based on the arbitrage allocation quota.

8. The method according to claim 2, characterized in that, The step of identifying the steepness convergence curve and determining the flat-top candidate interval by using the temperature lag rise segment in the peak steepness feature includes: Based on the temperature lag rise segment in the peak steepness feature, multi-cycle steepness repetition sampling is performed to obtain a steepness repetition sequence; A steep convergence curve is formed by calculating the convergence degree in the segments of the steep repetition sequence where convergence is accompanied by baseline rise. Based on the steep convergence curve, the duration of verification is used to obtain the candidate duration threshold; The flat-top candidate interval is determined based on the candidate duration threshold calibration interval boundary.

9. The method according to claim 3, characterized in that, The process of calculating the remaining temperature rise margin in the unsaturated heat accumulation section based on the basic charging plan to obtain the temperature rise margin includes: The remaining temperature rise margin is determined by calculating the thermal limit difference of the unsaturated section of thermal accumulation based on the basic charging plan. The corrected heat capacity difference is obtained by correcting the fluctuation range based on the remaining temperature rise margin. The temperature rise margin ratio is constructed based on the margin ratio calculated according to the corrected heat capacity difference. The temperature rise margin is obtained by converting the margin ratio using the aforementioned temperature rise margin margin.

10. A low-voltage cable branch box energy storage peak shaving and valley filling power saving system, characterized in that, include: The heat peak identification module is used to acquire branch box operation data, calculate current-carrying heat accumulation based on the branch box operation data to form a cable heat accumulation curve, and identify the duration of flat-top load through the short-term peak segment of the cable heat accumulation curve to determine the continuous heat peak period. The charging decision module is used to arrange peak avoidance according to the continuous hot peak period to obtain a basic charging plan, to determine the short-term overload power of the unsaturated heat accumulation section based on the basic charging plan to obtain the charging power quota, and to generate charging control instructions by reverse allocation of the charging power quota and the cable heat accumulation curve according to the heat accumulation difference. The discharge decision module is used to calculate the state of charge based on the charging control command to obtain the dischargeable capacity, use the cooling hysteresis branch within the dischargeable capacity to arrange the support order to obtain the discharge priority sequence, and insert a hysteresis cooling gap into the discharge priority sequence to output the discharge scheduling command. The safety detection module is used to collect the operating status data of the unit according to the discharge scheduling command, track the evolution of the internal resistance degradation curve based on the unit status data to form a safety margin rating, and determine the protection action based on the safety margin rating to generate a protection limit signal. The instruction output module is used to release the hysteresis of the protection limit signal, calculate the emergency lower limit demand to obtain the emergency reserve capacity, determine the arbitrage allocation quota based on the emergency reserve capacity to generate a comprehensive adjustment scheme, and execute peak shaving and valley filling optimization output charging and discharging instructions for the comprehensive adjustment scheme.

Images