Multi-parameter-based small generator grid connection management method and device

Abstract

The application discloses a small generator grid-connected management method and device based on multiple parameters, relates to the technical field of power grid connection control, and comprises the following steps: obtaining comprehensive energy operation data of a generator set, generating high-precision energy data, performing multi-scale island identification and dynamic evolution prediction on the high-precision energy data, and forming a multi-layer risk situation sequence; based on the multi-layer risk situation sequence, implementing adaptive power distribution and multi-data grid connection strategy collaborative optimization, generating real-time coordinated control instructions; performing heterogeneous resource mapping and multi-granularity scheduling arrangement on the real-time coordinated control instructions, and generating a hierarchical control execution sequence; performing multi-channel power distribution according to the hierarchical control execution sequence, and performing power dynamic compensation; through the collaborative optimization and multi-granularity scheduling arrangement technology driven by the multi-layer risk situation, the problems of extensive power distribution, lagging risk response and insufficient multi-source collaboration in traditional grid connection control are solved.

Description

Technical Field

[0001] This invention relates to the field of power grid connection control technology, and more specifically, to a method and apparatus for grid connection management of small generators based on multiple parameters. Background Technology

[0002] With the large-scale grid connection of distributed energy resources and small generator sets, the grid-connected operation environment exhibits complex characteristics of multi-source heterogeneity and dynamic coupling. The power output status of generator units is not only affected by load fluctuations, grid topology, and changes in operating conditions, but also closely related to grid disturbance propagation, local islanding formation, and system stability evolution. In actual operation, there are obvious temporal correlations and power interaction effects among generator units, and the risk status under different operating stages exhibits continuous evolution and cross-scale expansion characteristics. Therefore, how to achieve dynamic fusion and correlation analysis of multi-source information under multi-parameter operating data conditions, construct a risk situation sequence that can continuously characterize the islanding evolution process, and achieve coordinated optimization of power allocation, grid-connected control, and resource scheduling based on the identification of risk evolution trends, thereby forming a stable and efficient closed-loop control mechanism for grid-connected operation, has become an important technical problem that urgently needs to be solved in the field of small generator grid-connected management.

[0003] For example, the invention patent with publication number CN120782250A discloses a method and system for early warning of operational risks based on a power grid information system, relating to the field of power grid operational risk assessment technology. This method constructs a risk perception graph that integrates node and edge features to achieve equipment risk characterization and future risk score prediction; it combines clustering algorithms to identify high-risk areas and diffusion paths, and triggers hierarchical response strategies based on a rule base. This invention achieves real-time early warning and intelligent control of power grid operational risks.

[0004] For example, the invention patent with publication number CN120604535B discloses a dynamic coordination control method and system for microgrids used in distributed energy, relating to the field of microgrid energy management technology. Step one involves real-time acquisition of multi-source microgrid data, performing edge computing and data preprocessing operations, and early judgment of islanding modes. Step two involves aggregating the structure and status of each microgrid, dynamically constructing and updating the microgrid topology map, analyzing changes in the microgrid topology structure and health status, and assessing the comprehensive risk of microgrid nodes. Step three involves calculating the actual allocable power for each type of load, performing autonomous control and flexible mode switching. Step four involves, based on the actual allocable power for each type of load, real-time evaluation of the microgrid's collaborative energy dispatch capability, role identification, and implementation of optimization strategies. This solves the problem of delayed emergency dispatch response due to sudden main grid failures caused by extreme events as the scale of distributed energy and microgrid access continues to expand.

[0005] The above-disclosed technical solutions have at least the following technical problems: Traditional small generator grid connection management methods suffer from insufficient perception of multi-parameter operating status and lack of risk adaptation and collaborative optimization capabilities in power allocation strategies. This results in extensive power scheduling, lagging dynamic response, and difficulty in ensuring system stability during grid connection, making it difficult to meet the requirements of efficient, safe, and continuous operation in multi-source grid connection environments.

[0006] To address the above problems, this invention proposes a solution. Summary of the Invention

[0007] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a method and apparatus for grid-connected management of small generators based on multiple parameters. Through multi-level risk situation-driven collaborative optimization and multi-granularity scheduling and orchestration technology, the method and apparatus solve the problems of coarse power distribution, delayed risk response, and insufficient multi-source coordination in traditional grid-connected control.

[0008] To achieve the above objectives, the present invention provides the following technical solution: A method and apparatus for grid-connected management of small generators based on multiple parameters includes: acquiring comprehensive energy operation data of the generator set, generating high-precision energy data, and performing multi-scale island identification and dynamic evolution prediction on the high-precision energy data to form a multi-layer risk situation sequence; based on the multi-layer risk situation sequence, implementing adaptive power allocation and multi-data grid connection strategy collaborative optimization to generate real-time coordinated control commands; performing heterogeneous resource mapping and multi-granularity scheduling and orchestration on the real-time coordinated control commands to generate a hierarchical control execution sequence; and performing multi-channel power splitting and dynamic power compensation according to the hierarchical control execution sequence.

[0009] In a preferred technical solution, the acquisition of comprehensive energy operation data of generator sets, the generation of high-precision energy data, and the multi-scale island identification and dynamic evolution prediction of the high-precision energy data to form a multi-layer risk situation sequence are specifically as follows: Small generator set operation data are collected to form a raw energy operation data set; the raw energy operation data undergoes full-domain time-scale reconstruction and cross-channel constraint processing to form a standardized energy data sequence; the standardized energy data sequence undergoes multi-time-scale nested reconstruction to generate multi-scale coupled energy data; the multi-scale coupled energy data undergoes continuous electrical state expansion to form a continuous state chain; multi-window collaborative island identification is performed based on the high-precision energy state expression to generate an island identification result set; the island identification result set undergoes temporal evolution correlation analysis to generate an island dynamic evolution trajectory; and the island dynamic evolution trajectory is constructed using multiple risk situation hierarchical structures to form a multi-layer risk situation sequence.

[0010] In a preferred technical solution, the step of performing multi-window collaborative island identification based on high-precision energy state representation to generate an island identification result set is as follows: extracting data from each channel in the high-precision energy state representation to form a candidate path set; calculating the evolution deviation of adjacent channel island candidate paths in the candidate path set, and marking and eliminating them to obtain potential island paths; accumulating and counting the potential island paths within a continuous time window to obtain the main island evolution path and low-frequency island candidate paths; adjusting the abrupt changes of each path based on the main island evolution path and low-frequency island candidate paths; performing multi-path fusion on the adjusted island path set to generate a final island identification result set; and forming a dynamic trajectory by tracking the continuity and change trend of the final island identification result set along the time axis.

[0011] In a preferred technical solution, the step of implementing adaptive power allocation and multi-data grid connection strategy collaborative optimization based on a multi-layer risk situation sequence to generate real-time coordinated control commands specifically involves: merging the risk levels, evolution trends, and islanding evolution paths of each layer in the multi-layer risk situation sequence to form an initial risk mapping sequence; generating preliminary power allocation candidate schemes by performing risk-priority allocation and redundant buffer collaborative scheduling on the initial risk mapping sequence; implementing multi-data grid connection collaborative optimization on the preliminary power allocation candidate schemes, and adjusting the power output order of each power generation unit through cross-channel dependency constraints to form a collaborative optimization result; performing multi-granularity load matching on the collaborative optimization result, and adjusting the power... The output order and timing are dynamically adjusted to generate a power execution sequence. The consistency of power output order and timing within the power execution sequence is maintained, and the power execution sequence is mapped to a real-time control command sequence. Adaptive redundancy adjustment is implemented on the real-time control command sequence, and potential conflicting commands are automatically corrected based on multiple command paths and islanding risk levels to form a corrected real-time control command sequence. Multi-dimensional performance optimization is performed on the corrected real-time control command sequence, and the optimal control command path is obtained through joint evaluation of power smoothness, load matching, and islanding suppression effects. The optimal real-time control command sequence is continuously monitored, and its execution status is compared with the risk situation.

[0012] In a preferred technical solution, the multi-dimensional performance optimization of the modified real-time control command sequence, and the obtaining of the optimal control command path through joint evaluation of power smoothness, load matching degree, and islanding suppression effect, are as follows: A multi-dimensional performance sequence is generated based on the modified real-time control command sequence; based on the multi-dimensional performance sequence, parallel calculations are performed on each path in terms of power smoothness, load matching degree, and islanding suppression dimensions to form a parallel performance evaluation table; the parallel performance evaluation table is used to predict and adjust each candidate path, and the adjusted performance data is updated back to the parallel performance evaluation table; the Pareto optimal path in the parallel performance evaluation table is selected to form a preliminary optimized execution sequence; risk-priority scheduling is implemented on the preliminary optimized execution sequence to generate an optimized execution path; minor discontinuities and potential conflicts in the optimized candidate paths are corrected to form the final execution path; the final execution path is mapped to a real-time control command sequence, and risk priority and execution order markers are retained in the commands.

[0013] In a preferred technical solution, the heterogeneous resource mapping and multi-granularity scheduling of real-time coordinated control commands to generate a hierarchical control execution sequence is performed as follows: The real-time coordinated control command sequence is uniformly marked to generate a marked command feature set; the marked command feature set is initially mapped to available power generation units and grid-connected interface resources to generate multiple candidate mapping path chains; the multiple candidate mapping path chains are hierarchically segmented to generate multi-scale scheduling unit partitioning results; based on the multi-scale scheduling unit partitioning results, the output order of high-risk units and key node units is prioritized, while the order of low-risk units is adjusted and time-staggered to form a dynamically optimized scheduling arrangement structure; the dynamically optimized scheduling arrangement structure is intelligently matched and mapped to the resources in the power generation units to generate a preliminary hierarchical control path set; multi-granularity dynamic scheduling is implemented on the preliminary hierarchical control path set to generate a final optimized hierarchical control sequence; high-risk units and key load nodes in the final hierarchical control execution sequence are marked with risk priority, while auxiliary output paths are retained for low-risk units to form a full-time hierarchical control execution sequence.

[0014] In a preferred technical solution, the step of implementing multi-granularity dynamic scheduling on the preliminary hierarchical control path set to generate the final optimized hierarchical control sequence is as follows: Event extraction is performed on the preliminary hierarchical control path set to generate event-driven control units; based on the event-driven control units, the control units are clustered, and the power output within each cluster is analyzed for dependencies to generate dynamic resource allocation grouping results; for the event nodes within each group in the dynamic resource allocation grouping results, short-term power trend and islanding risk evolution predictions are performed to generate segmented dynamic control schemes, and the power output order within each segment is fine-tuned to form a segmented optimized sequence; the segmented optimized sequence is integrated according to resource clusters, and the cross-cluster power output order is compressed and optimized to generate resource compressed scheduling paths; multi-granularity time window scheduling is performed on the compressed optimized path set to generate multi-granularity time control schemes; each segmented optimization unit in the multi-granularity dynamic execution sequence is merged and rearranged according to the event triggering order to generate an integrated hierarchical control path; each execution path in the final hierarchical control sequence is marked with resource binding to form a full-time hierarchical control execution sequence.

[0015] In a preferred technical solution, the step of performing multi-channel power splitting and dynamic power compensation according to the hierarchical control execution sequence is as follows: Based on the hierarchical control execution sequence, the power generation channels and grid connection paths in each control unit are associated and marked to form a channel-associated execution structure; based on the channel-associated execution structure, the power output of each control unit is decomposed into channels to form a multi-channel power splitting sequence; by tracking the changing trends of load nodes in each channel of the multi-channel power splitting sequence, the power allocation ratio is adjusted in real time to form adaptive splitting execution data; based on the adaptive splitting execution sequence, the power output rhythm between channels is staggered and adjusted. The system sorts and reduces instantaneous superimposed power peaks to form a staggered collaborative power distribution sequence. Based on the staggered collaborative power distribution sequence, it continuously tracks the local deviations generated during the power output process of each channel, identifies time segments with insufficient and excessive power, and marks them with compensation tags in the corresponding channels to form a set of power tags to be compensated. According to the power sequence to be compensated, it performs cross-channel dynamic compensation on each channel to form a cross-channel collaborative compensation cluster. It performs trend adaptive correction on the cross-channel collaborative compensation cluster to generate a dynamic compensation execution path. According to the dynamic compensation execution sequence, it adjusts the final power output of each channel throughout the entire time period to form a multi-channel power distribution and dynamic compensation execution data stream.

[0016] In a preferred technical solution, the step of rearranging the power output rhythm of each channel according to the adaptive power splitting execution sequence and reducing the instantaneous superimposed power peaks to form a staggered collaborative power splitting sequence is as follows: Extracting the power rise interval, stable interval, and fall interval from the adaptive power splitting execution sequence to form a channel power rhythm feature sequence; based on the channel power rhythm feature sequence, identifying time-overlapping power rise intervals and marking the power superposition of multiple channels within the same time segment to form a peak superposition candidate domain; and dynamically rearranging the power output order of each channel within the peak superposition candidate domain. This process involves several steps: first, establishing a preliminary off-peak output layout; second, segmenting and mitigating power increases in channels with excessively rapid power rises, and pre-filling gaps in channels with gradual power changes to generate a rhythm coordination control structure; third, alternating high-load and low-load channels based on the rhythm coordination sequence to form an alternating collaborative scheduling flow; fourth, refining the time-shifted power output of low-risk channels in the alternating collaborative scheduling flow while prioritizing high-risk channels to generate a stable off-peak operation state; and fifth, unifying and integrating the power output sequence and time nodes of each channel in the stable off-peak sequence to form a off-peak collaborative diversion sequence.

[0017] In a preferred technical solution, a method and apparatus for grid-connected management of small generators based on multiple parameters includes an energy data construction module, a risk situation optimization construction module, a scheduling mapping and orchestration module, and a power shunting compensation control module. These modules are interconnected: the energy data construction module acquires comprehensive energy operation data of the generator set, generates high-precision energy data, and performs multi-scale island identification and dynamic evolution prediction on the high-precision energy data to form a multi-layered risk situation sequence; the risk situation optimization construction module implements adaptive power allocation and multi-data grid-connected strategy collaborative optimization based on the multi-layered risk situation sequence to generate real-time coordinated control commands; the scheduling mapping and orchestration module performs heterogeneous resource mapping and multi-granularity scheduling orchestration on the real-time coordinated control commands to generate a hierarchical control execution sequence; and the power shunting compensation control module executes multi-channel power shunting according to the hierarchical control execution sequence and performs dynamic power compensation.

[0018] The technical effects and advantages of this invention based on a multi-parameter small generator grid-connected management method and device are as follows: 1. This invention utilizes multi-parameter data fusion and multi-scale island identification technology to refine the original operating information and generate high-precision energy data after acquiring comprehensive energy operation data of generator sets. Simultaneously, by combining multi-scale island identification and dynamic evolution prediction, it continuously characterizes the island evolution path at different time scales, effectively enabling early identification and hierarchical representation of island risks, significantly improving the accuracy of risk perception and the ability to judge evolution trends. Furthermore, by constructing a multi-layered risk situation sequence and introducing evolution trend constraints for risk level division, it can clearly distinguish and dynamically track different risk stages, further enhancing the grid-connected system's responsiveness to complex operating conditions and overall operational stability.

[0019] 2. This invention utilizes adaptive power allocation and multi-data grid connection strategy collaborative optimization technology. Guided by a multi-layered risk situation sequence, it implements differentiated power scheduling for high-risk and low-risk areas. Through cross-data source collaborative optimization, it dynamically adjusts the power output sequence, effectively achieving coordination and unification of power allocation and grid connection strategies, significantly improving load matching and system balance during grid connection. Simultaneously, through heterogeneous resource mapping and multi-granularity scheduling orchestration technology, it maps real-time coordinated control commands to different power generation units and grid connection paths. Combined with multi-channel power diversion and dynamic compensation mechanisms, it performs peak-shifting adjustment and cross-channel compensation for power output, effectively reducing instantaneous power fluctuations and ensuring continuous power supply, further enhancing system stability and anti-disturbance capabilities. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating the multi-parameter-based grid-connected management method for small generators according to the present invention.

[0021] Figure 2 This is a schematic diagram of the device for the multi-parameter-based small generator grid-connected management method of the present invention. Technical solution

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Example 1, Figure 1 The present invention provides a multi-parameter-based grid-connected management method for small generators, comprising the following steps: S1: Acquire comprehensive energy operation data of generator sets, generate high-precision energy data, and perform multi-scale island identification and dynamic evolution prediction on the high-precision energy data to form a multi-layer risk situation sequence; In this embodiment, comprehensive energy operation data of the generator set is acquired, high-precision energy data is generated, and multi-scale island identification and dynamic evolution prediction are performed on the high-precision energy data to form a multi-layer risk situation sequence, as detailed below: The system collects voltage, current, frequency, phase, and power data during the operation of small generator sets, and simultaneously records switch status and load change data to form a raw energy operation data set. The raw energy operation data undergoes global time-scale reconstruction and cross-channel constraint processing, and adaptive adjustments are made to the data offsets of different sampling paths to form a standardized energy data sequence with temporal isomorphism. Specifically, the global time-scale reconstruction involves globally mapping the raw energy operation data from each channel based on a unified time reference, projecting measurement data with different sampling rates and start times onto the same unified time axis, and aligning with the signal through time interpolation. Redundant information from high-sampling-rate data is compressed and integrated, while low-sampling-rate data is padded, ensuring precise temporal alignment of all channel data to form a globally synchronized sequence. The cross-channel constraint processing involves adjusting the voltage, frequency, and other parameters collected from each channel of the generator set. A multi-dimensional constraint relationship sequence is established using phase and power data. The physical coupling characteristics and operational correlations between channels are mapped as constraints. Then, abnormal deviations, abrupt changes, or noise in each channel are detected in real time. Correction amounts are calculated using the constraint relationship sequence to ensure consistency in electrical coupling relationships across channels. The adaptive adjustment specifically involves: first, determining the deviation of each channel's data sampling point from a unified time reference using continuous time difference calculations; then, performing a timestamp remapping operation on sampling points with deviations exceeding a threshold, remapping the point to its corresponding standard time position; furthermore, generating a correction value by accumulating neighboring point values ​​of nearby synchronous sampling points according to time proportions, and combining this correction value with the original sampling value to form a... The corrected data points, to ensure continuous transient characteristics and consistency with other channels, are repeatedly processed for all offset points until the data of each channel achieves multi-channel dynamic alignment in time and amplitude. This alignment with the signal is achieved through time interpolation: First, a standard time axis is constructed using a unified time base. Based on the time interval between the original sampling time points of each channel and the standard time axis, the time positions requiring supplementation or resampling are determined. Then, high-sampling-rate channel data is sampled according to a time window. For low-sampling-rate or missing channel data, linear interpolation is used to generate intermediate data between adjacent known sampling points, covering the missing positions on the standard time axis. Finally, phase and amplitude alignment is performed on the signals of each channel. Ultimately, this achieves synchronous alignment and continuous representation of multi-channel data on a unified time axis. The specific calculation of the correction amount through the constraint relationship sequence is as follows: Based on the constraint relationship sequence, the measured values ​​of each channel at the current moment are substituted into the corresponding constraint sequence to determine the deviation between them and the constraint range. Then, based on the correlation degree of the associated channels in the constraint relationship and the closeness of their current state to the constraint target, the deviation amount is distributed among the associated channels, so that each associated channel participates in the deviation correction according to its correlation strength. Finally, when the correction direction of a certain channel is consistent with the change direction of its adjacent time points, the correction direction is retained. When there is an opposite direction or a sudden change, the magnitude of the correction amount of the channel is weakened or the direction is adjusted to generate the corresponding correction amount. A multi-timescale nested reconstruction is performed on a standardized energy data sequence to generate multi-scale coupled energy data with hierarchical correlation characteristics. Specifically, the multi-timescale nested reconstruction involves: first, dividing the standardized energy data sequence into short-term, medium-term, and long-term periods according to time length, and independently extracting data from each period to form data segments corresponding to the time scale; then, extracting continuously changing transient fluctuation features in the short-term period to form fine-grained change trajectories; integrating adjacent short-term data in the medium-term period to extract smooth evolution trends and form medium-scale evolution trajectories; summarizing the medium-scale trajectories in the long-term period to extract periodic change features and form long-term trend trajectories; then, embedding the short-term trajectories into the corresponding medium-term trajectories in chronological order, and further embedding the medium-term trajectories into the long-term trajectories; finally, maintaining the correspondence between the time scales during the nesting process, so that short-term changes can be mapped to medium- and long-term trends, thereby forming a multi-timescale nested structure with hierarchical correlation characteristics. The electrical state is continuously unfolded on multi-scale coupled energy data to form a continuous state chain covering the entire operation process, thereby obtaining a high-precision energy state expression with high resolution. Specifically, the continuous unfolding of the electrical state involves: first, generating a state increment sequence based on the state difference between adjacent time points, and accumulating it sequentially along the time axis to form a continuous change chain; then, performing adjacent channel synchronous mapping on each accumulated state point to introduce the electrical coupling relationship between channels into the continuous chain, ensuring the coordination and consistency of the state across multiple channels, and finally generating a continuous state chain covering the entire grid-connected operation cycle. Specifically, the continuous state chain involves: first, combining the voltage, current, frequency, phase, and power data of each channel at each time point according to a unified time axis to obtain the multi-channel state representation corresponding to each time point; then, sequentially connecting the states of adjacent time points in chronological order, maintaining the continuous correspondence between the state changes between each time point; next, recording the electrical changes between adjacent states; and finally, connecting the multi-channel states of each time point in the entire grid-connected operation cycle sequentially along the time axis to form a continuous state chain that reflects the gradual change process of the system from one time point to the next. Based on high-precision energy state representation, multi-window collaborative island identification is performed. By constructing state change sensing paths at different time scales and cross-verifying the multi-path identification results, local electrical isolation trends and potential island formation processes are identified, generating an island identification result set. Specifically, the cross-verification of multi-path identification results involves: first, extracting the corresponding change time points and their state change characteristics from the state change sensing paths constructed at different time scales and aligning them on a unified time axis; then, comparing the voltage, frequency, and power change directions and amplitudes of change points with similar or overlapping time positions at each time scale to determine whether they show a consistent change trend at multiple time scales; next, confirming the consistency of change points identified at multiple time scales and corresponding in time position and change characteristics, and determining them as valid change results; simultaneously, weakening change points that only appear at a single time scale but are not verified at other time scales; finally, summarizing and outputting the change results verified by multiple time scales, thereby achieving cross-verification between multi-path identification results. A temporal evolution correlation analysis is performed on the island identification result set to extract the gradual change characteristics during the island formation process (including: continuous offset data of electrical quantities, change rate evolution data, state persistence data, and multi-channel coupling change data). This generates a dynamic evolution trajectory of the islands reflecting the continuity of evolution. Specifically, the temporal evolution correlation analysis involves: first, converting the island result sequence at each time point into state vectors and arranging them chronologically to form an initial state chain; then, calculating the similarity and evolutionary deviation between state vectors at adjacent time points, identifying continuous similar state segments, and recognizing the continuation path of the island state; simultaneously, establishing a new state correlation with the previous evolution trajectory for time points where abrupt changes or new island formation occur, forming a complete evolution path sequence, and finally generating a dynamic evolution trajectory including initial disturbances, local island formation, and... The continuous trajectory of stable island evolution is calculated by determining the similarity and evolutionary deviation between state vectors at adjacent time points. Specifically, the state vectors at adjacent time points are aligned according to voltage, frequency, and power components, and the change amplitude and direction of each component between the two time points are extracted. When the change trends of each component are consistent and the change amplitudes are within a similar range, the time point pair is marked as a highly similar state. At the same time, the differences in the change amplitudes of each component are accumulated. When one or more components deviate or change direction abruptly, the corresponding time point is marked as an evolutionary deviation enhancement interval. Furthermore, the similarity results within a continuous time period are aggregated to form a continuous similar state segment, and the interval where the evolutionary deviation continues to expand is marked to distinguish between stable evolution processes and abrupt evolution processes. The dynamic evolution trajectory of isolated islands is constructed by multiple risk situation layers, and evolutionary trend constraints are introduced to classify the risk states into levels, forming a multi-layer risk situation sequence that includes initial disturbance state, transitional instability state and potential isolated state.

[0024] In this embodiment, multi-window collaborative island identification is performed based on high-precision energy state representation. By constructing state change sensing paths at different time scales and cross-verifying the multi-path identification results, local electrical isolation trends and potential island formation processes are identified, generating an island identification result set, as follows: The voltage, frequency, phase, and power state sequences of each channel in the high-precision energy state representation are extracted, and state abrupt change points are identified and candidate path sets are formed by using short and long time windows respectively. Calculate the evolution deviation of adjacent channel island candidate paths in the candidate path set, mark and remove paths that do not conform to channel coupling relationship, and retain potential island paths that are consistent across multiple channels; The potential island paths with consistent multi-channel characteristics are cumulatively counted within a continuous time window to obtain the main island evolution paths and low-frequency island candidate paths. Specifically, the cumulative counting to obtain the main island evolution paths and low-frequency island candidate paths involves: aligning the potential island paths with consistent multi-channel characteristics along the time axis, recording the occurrence of each path hourly within the continuous time window, counting the number of occurrences and the duration of continuous occurrence in each time segment, summarizing the cumulative occurrence frequency of each path throughout the continuous time window, and combining this with its continuous distribution on the time axis to identify paths with high occurrence frequency and continuous existence in multiple adjacent time segments, which are marked as main island evolution paths. At the same time, paths that only appear in local time segments, have low cumulative frequency, and lack continuity are marked and classified as low-frequency island candidate paths. Based on the main island evolution path and low-frequency island candidate path, the abrupt changes of each path are adaptively adjusted so that the path can accurately capture potential islands under different operating states. Specifically, the adaptive adjustment is as follows: extract the amplitude and duration of state abrupt changes in the main island evolution path and the low-frequency island candidate path in a continuous time interval. Determine the benchmark abrupt change interval based on the abrupt changes in the main island evolution path, and align and compare the low-frequency island candidate path with this benchmark. When the abrupt change amplitude of the low-frequency path is close to the benchmark abrupt change interval, its abrupt change segment is expanded to include adjacent time points to enhance continuity. When the abrupt change amplitude deviates from the benchmark interval, its abrupt change segment is compressed to remove isolated or short-term fluctuation points, thereby completing the adaptive and unified adjustment of the abrupt changes of each path. Multi-path fusion is performed on the adaptively adjusted set of isolated paths to generate a final set of isolated path identification results. Specifically, the multi-path fusion involves: aligning the adaptively adjusted isolated paths according to the time axis; matching the states at corresponding time points in different paths; merging path segments that overlap within the same time interval; retaining path intervals with more overlaps and longer durations as valid segments; and removing path segments that appear only in a single path and have a short duration. Subsequently, the retained valid path segments are connected in chronological order to form a continuous fused path. By tracing the continuity and changing trends of the final island identification results set along the time axis, a dynamic trajectory reflecting the island evolution process is formed.

[0025] In this embodiment, a multi-layered risk situation is constructed for the dynamic evolution trajectory of the isolated island, and evolutionary trend constraints are introduced to classify the risk states hierarchically, forming a multi-layered risk situation sequence that includes initial disturbance state, transitional instability state, and potential isolated state, as detailed below: The dynamic evolution trajectory of the isolated island is segmented in chronological order, and a segmented evolution sequence is generated by identifying intervals of state change continuity. During the segmentation process, the change amplitude and duration data of each segment are recorded synchronously. Specifically, the step of generating a segmented evolution sequence by identifying intervals of state change continuity involves: calculating the change difference of adjacent state points along the time axis, marking the time interval where the change difference remains within a continuous interval as the same state interval, and determining the interval boundary point when the change difference of adjacent time points changes abruptly or reverses direction. The overall trajectory is then divided based on the interval boundary point, and the state points in each continuous interval are combined in chronological order to form independent evolution segments. Each segment is numbered, and finally, a segmented evolution sequence arranged in chronological order is generated. Based on segmented evolutionary sequences, a cross-scale rate of change trajectory is formed by superimposing multiple window rates, enabling the joint expression of short-term mutations and long-term evolutionary trends within the same trajectory. Specifically, the multi-window rate superposition involves: distinguishing short-term, medium-term, and long-term windows on the segmented evolutionary sequence; calculating the rate of change for adjacent state points within each window at time intervals to obtain the rate sequence for the corresponding window; aligning the rate sequences of different windows along a unified time axis; furthermore, synchronously mapping the multiple window rates at the same time point; and superimposing them sequentially according to the time coverage corresponding to the window length, so that the short-window rate reflects transient changes, the medium- and long-term window rates reflect trend changes, and finally forming a cross-scale rate of change trajectory spanning the entire time domain. Under the constraint of the cross-scale rate of change trajectory, the state offsets of each segmented evolution sequence are progressively accumulated to generate a stage-sensitive cumulative offset sequence. Specifically, the progressive accumulation under the constraint involves: aligning the rate change trend of the corresponding time period in the cross-scale rate of change trajectory with the segmented evolution sequence segment by segment, and matching the state offset within each segment with its corresponding rate change direction. Then, when the rate of change of a certain segment shows a continuous increasing or stable upward trend, the state offsets within that segment are progressively accumulated in a forward direction according to time sequence, so that the offsets gradually increase with the evolution process. When the rate of change weakens or the direction reverses, the offset accumulation process of the segment is truncated to avoid excessive amplification of the overall accumulation result by abnormal mutations. At the same time, the cumulative offset of the previous segment is used as the initial reference for the next segment, thereby realizing the orderly evolution of the state offsets under the guidance of the cross-scale rate of change trajectory. Based on the cross-scale change rate trajectory and cumulative offset sequence, a state evolution determination space is constructed, and the distribution positions of each time period in the determination space are clustered and identified. Regions with small change amplitude, gentle rate and low cumulative offset are divided into initial perturbation states. Specifically, the construction of the state evolution determination space is as follows: the cross-scale change rate trajectory and cumulative offset sequence are used as two-dimensional coordinate axes, the horizontal axis represents the change rate of the state within the time period, and the vertical axis represents the cumulative offset of the corresponding time period. The evolution state of each time period is then mapped to a state point in the coordinate system, and the state points of adjacent time periods are connected in chronological order to form an evolution trajectory. Furthermore, by statistically analyzing the density of all state points, a state cluster region is generated. The distribution range of the trajectory in the determination space is used as the reference region for the risk level, thereby completing the construction of the state evolution determination space. Based on the initial perturbation dynamics, regions where the rate of change shows a continuous increasing trend and the cumulative offset grows synchronously are classified as transitional unstable states. Regions where the rate of change tends to stabilize but the cumulative offset remains at a high level and persists are classified as potential island states, thereby completing the multi-layer risk situation stratification construction of the dynamic evolution trajectory of islands.

[0026] S2, based on a multi-layered risk situation sequence, implements adaptive power allocation and multi-data grid connection strategy collaborative optimization to generate real-time coordinated control commands; In this embodiment, based on a multi-layered risk situation sequence, adaptive power allocation and multi-data grid connection strategy collaborative optimization are implemented to generate real-time coordinated control commands, as follows: The risk levels, evolution trends, and isolated evolution paths of each layer in the multi-layered risk situation sequence are merged to form an initial risk mapping sequence; By performing risk-priority allocation and redundant buffering coordinated scheduling on different risk areas in the initial risk mapping sequence, preliminary power allocation candidate schemes are generated. Specifically, the risk-priority allocation is as follows: based on the risk level and evolution trend of each time period in the initial risk mapping sequence, the time segments corresponding to high-risk areas and associated power generation units are given priority in allocating continuous and stable power output intervals during the power allocation process, so that high-risk areas are given priority protection during critical periods. At the same time, the power output sequence within high-risk areas is arranged to ensure its output continuity and stability. Specifically, the redundant buffering coordinated scheduling is as follows: the power generation units corresponding to low-risk areas are distributed in time sequence to form an adjustable power redundancy interval outside the high-risk areas. The power output of low-risk areas is arranged in a filling manner according to the overall power demand, so that it can undertake dynamic compensation and load buffering without interfering with the power output of high-risk areas, thereby forming a power allocation structure in which high-risk areas are given priority protection and low-risk areas are flexibly adjusted. Multi-data grid-connected collaborative optimization is implemented on the preliminary power allocation candidate schemes, and the power output sequence of each power generation unit is adjusted through cross-channel dependency constraints to form a collaborative optimization result. Specifically, the multi-data grid-connected collaborative optimization involves: based on the real-time power output, load changes, and risk status of each power generation unit, alternating scheduling of power output in high-risk and low-risk areas to ensure that high-risk areas receive priority stable power support, while low-risk areas provide auxiliary regulation, thereby achieving coordinated operation among the power generation units. During the execution process, the output sequence and power allocation are continuously iterated and adjusted in small increments to ensure that the overall power allocation simultaneously considers load demand and grid connection. To balance and control islanding risks, the power output sequence of each power generation unit is adjusted through cross-channel dependency constraints. Specifically, the power adjustment dependency sequence between units is determined based on the electrical coupling relationship and grid connection topology between each power generation unit. Then, within the same time window, the power output of each power generation unit is adjusted sequentially, with high-risk or critical node units outputting first and low-risk units adjusting their power subsequently. During the adjustment process, the power changes of each channel are monitored in real time. When a power conflict between units is detected or the safe operating range is exceeded, the output sequence is rearranged and the adjustment is repeated until the power output sequence of all power generation units satisfies the dependency constraints and takes into account grid stability. Multi-granularity load matching is performed on the collaborative optimization results, and the power output sequence and time nodes are dynamically adjusted to generate a power execution sequence. To maintain consistency between the power output order and time nodes in the power execution sequence, and to map the power execution sequence into a real-time control command sequence, specifically, the generated power execution sequence is divided into continuously executed power operation units according to time order. Each operation unit corresponds to the power output value of a single power generation unit within a specific time window. The power operation units are then converted into corresponding control commands, including power, output start / stop, and adjustment commands, and arranged in time order to form a continuous command stream. At the same time, priority execution flags corresponding to the islanding risk level are retained in the commands to ensure that the control commands are both continuous and meet risk management requirements in actual execution. Adaptive redundancy adjustment is implemented for the real-time control command sequence, and potential conflicting commands are automatically corrected based on multiple command paths and island risk levels to form a corrected real-time control command sequence. Specifically, the adaptive redundancy adjustment involves establishing redundant execution paths for each control command in the time sequence, and detecting potential conflicts or repeated execution risks by comparing multiple command paths with island risk levels in real time. When a potential conflict is detected, the priority execution path is automatically selected, and the power output order or execution time of the suboptimal path is adjusted. At the same time, redundant commands of low-risk or auxiliary units are delayed or distributed for execution to form dynamic redundancy compensation, so as to achieve safe and continuous execution of the overall command sequence under global risk control. The automatic correction of potential conflicting commands involves marking commands with overlapping power output times or conflicting sequences in the real-time control command sequence as conflicting commands. Then, the command priority is determined according to the island risk level. Commands of high-risk units are executed first, and commands of low-risk units are readjusted in terms of time nodes or output order without affecting the power output of priority units, so as to achieve automatic correction of conflicting commands and safe synchronous execution. Multidimensional performance optimization is performed on the modified real-time control command sequence, and the control command path with the best continuity and the lowest risk is selected by jointly evaluating power smoothness, load matching degree and islanding suppression effect. The optimized real-time control command sequence is continuously monitored, and the execution status is compared with the risk situation to maintain a closed loop of power distribution and grid connection stability throughout the entire period.

[0027] In this embodiment, multi-granularity load matching is performed on the collaborative optimization results, and the power output sequence and time nodes are dynamically adjusted to generate a power execution sequence, as detailed below: The output power of each power generation unit in the collaborative optimization results is coupled and matched with the load nodes in real time, and the power output order is arranged according to the load priority to generate a preliminary coupling matching sequence. Specifically, the real-time coupling matching is as follows: a one-to-one correspondence is established between the power of each load node and the real-time output capability of all power generation units. The loads are sorted according to their importance, and high-priority loads are matched with the most stable and fast-responding power generation units, while low-priority loads are matched with auxiliary units. At the same time, the output order is dynamically adjusted during the matching process to take into account instantaneous load fluctuations and overall power balance, so as to achieve continuous and collaborative matching between load demand and power generation capacity. The initial coupling matching sequence is mapped onto multiple time scales, and the output order and time nodes of each unit are adjusted by superposition to form a multi-time scale mapping sequence. Specifically, the superposition adjustment involves projecting the initial coupling matching sequence into independent execution trajectories according to three time scales: short-term transient, hourly evolution, and daily cycle trend. Then, the output order of the power generation units is arranged hierarchically at each time scale, with high-risk or critical node units being prioritized for output in the short-term trajectory. At the same time, the output order of the hourly and daily cycle trajectories is superimposed onto the short-term trajectory according to time periods, and the time nodes are adjusted to make the power output order of the multi-scale trajectories continuous and coordinated in time, forming a unified multi-time scale mapping sequence. Short-term redundant slots are reserved in the sequence for the power output of key load nodes in the multi-timescale mapping sequence, and the power of low-priority nodes is appropriately staggered or delayed to form a power execution sequence with dynamic redundancy. Based on the power execution sequence, the feasibility of the power output sequence and timing nodes is verified one by one according to the coupling relationship between the load and the power generation unit. The adjustment results are reflected in the entire sequence to form a stable and continuous execution sequence. Specifically, the feasibility of the power output sequence and timing nodes is verified one by one as follows: each output operation in the power execution sequence is traversed sequentially, the power capacity of the corresponding power generation unit is compared with the real-time demand of the load node, the overlap or conflict of the output timing nodes is checked, and the power response is evaluated to meet the load curve changes. For the discovered sequence conflicts or time deviations, the output sequence is adjusted or the timing nodes are fine-tuned to ensure that each operation meets the load demand while maintaining the continuity and smoothness of the entire sequence, thereby ensuring that the power output sequence is feasible and safe throughout the entire time period. The power output units that have completed the coupling mapping between load and power generation unit in the execution sequence are given priority marking. The power output units corresponding to low-risk loads are flexibly adjusted without affecting the output of high-priority loads to form the final power execution sequence. The priority marking is as follows: extract the output operations of the power generation units corresponding to the high-risk load nodes in the power execution sequence and mark them as high-priority execution units. At the same time, the output operations of the other corresponding low-risk loads are marked as low-priority execution units. The flexible adjustment is as follows: first, identify the time position and order of the power output operations corresponding to the low-risk nodes in the sequence. Then, according to the output time period of the high-priority nodes, shift the output operations of the low-risk nodes forward or backward on the time axis, or insert intervals in the gaps of discontinuous execution. At the same time, adjust the amplitude of the power output of the low-risk nodes to make them execute smoothly under the requirements of overall load matching and power continuity, thereby ensuring that the power output of the high-priority nodes is not disturbed and realizing dynamic optimization and stable execution of the entire sequence. By tracing the smoothness of power output and inter-node dependencies in the final power execution sequence along the time axis, it is ensured that the sequence can stably match the load and grid connection requirements throughout the entire time period, and output a continuously executable power operation unit sequence. The continuously executable power operation unit sequence is continuously compared with the collaborative optimization results to generate the final executable power control sequence.

[0028] In this embodiment, the real-time control command sequence is optimized in multiple dimensions, and the control command path with the best continuity and the lowest risk is selected by jointly evaluating power smoothness, load matching degree and islanding suppression effect, as follows: A multidimensional performance sequence is generated by sequentially arranging the power smoothness, load matching degree, and islanding suppression data of each instruction path in the modified real-time control instruction sequence. Based on the multidimensional performance sequence, parallel calculations are performed on each path in terms of power smoothness, load matching degree, and islanding suppression to form a parallel performance evaluation table. Specifically, the parallel calculation is as follows: for each candidate path in the multidimensional performance sequence, each power output point is scanned step by step along the time series in terms of power smoothness, load matching degree, and islanding suppression. The deviations from the ideal smooth curve, load matching degree, and islanding suppression are calculated, and the deviations are accumulated to form a path dimension score. At the same time, the same calculation is performed in parallel on all paths so that the performance of each path under the three-dimensional performance index is synchronously generated into a complete evaluation result. Using the parallel performance evaluation table, the current island risk level and short-term load change trend are used as constraints to predict and adjust each candidate path, optimize points with power mutations and load mismatches in advance, and update the adjusted performance data back to the parallel performance evaluation table. The Pareto optimal path that balances power smoothness, load matching, and islanding suppression in the updated parallel performance evaluation table is selected to form a preliminary optimized execution sequence. Specifically, the selection of the Pareto optimal path involves comparing each candidate path in the parallel performance evaluation table pairwise across the three dimensions of power smoothness, load matching, and islanding suppression. The path that is not surpassed in any dimension is retained as the Pareto optimal set. The remaining paths are then sorted according to their comprehensive performance across each dimension to form a preliminary optimized execution sequence that balances continuity, load matching accuracy, and islanding risk suppression. Risk-priority scheduling is implemented on the preliminary optimized execution sequence, prioritizing the execution of control instructions for high-risk units or critical load nodes, while low-risk units are appropriately time-shifted or ordered without affecting high-priority paths, thus generating optimized execution paths that balance continuity and security. Minor discontinuities or potential conflicts in the optimized candidate path are corrected to achieve final path stabilization and form a final execution path that can be directly mapped. Specifically, the correction of minor discontinuities or potential conflicts involves scanning each control instruction in the optimized execution path along the time axis, detecting abrupt changes and overlaps in the power output amplitude or time nodes of adjacent instructions, adjusting the power difference between adjacent time points through smooth transition interpolation for power abrupt changes to ensure continuous output for power abrupt changes, and rearranging conflicting instructions according to the island risk level for instruction time or sequence conflicts to ensure that high-risk units are executed first, while delaying or reversing the order of instructions for low-risk units to achieve a continuous, smooth, and conflict-free path throughout the entire time period. The final execution path, which has undergone multi-objective collaborative optimization, is mapped into a continuous and executable real-time control instruction sequence. Risk priority and execution order markers are retained in the instructions to ensure continuous power output and accurate load matching throughout the entire time period.

[0029] S3 performs heterogeneous resource mapping and multi-granularity scheduling and orchestration on real-time coordination and control commands to generate hierarchical control execution sequences; In this embodiment, heterogeneous resource mapping and multi-granularity scheduling orchestration are performed on real-time coordination and control commands to generate hierarchical control execution sequences, as detailed below: The power output, time nodes, and risk levels of each control unit in the real-time coordinated control command sequence are uniformly marked to generate a set of marked command features; A preliminary mapping is performed on the marked instruction feature set and available power generation units and grid connection interface resources. High-risk units are preferentially mapped to key resource nodes, and low-risk units are assigned to auxiliary nodes, generating multiple candidate mapping path chains. Specifically, the preliminary mapping is as follows: First, available power generation units and grid connection interface resources are identified according to the risk level and time node marked by the instruction. Then, according to the risk priority, high-risk unit instructions are matched to key resource nodes to enable them to obtain stable power output during critical periods. Low-risk unit instructions are matched to auxiliary or redundant resource nodes to ensure continuous input without interfering with high-risk units. During the mapping process, the resource and time position corresponding to each instruction are recorded, thereby generating multiple candidate mapping paths. Multiple candidate mapping path chains are hierarchically segmented to generate multi-scale scheduling unit partitioning results. Specifically, the hierarchical segmentation involves dividing each candidate mapping path into short-cycle, medium-cycle, and long-cycle windows according to different operational scales in time. The short-cycle window is used to capture instantaneous power fluctuations and rapid load changes, and the power output of consecutive short time periods in the path is divided into small units. The medium-cycle window is used to reflect hourly load trends, and consecutive short-cycle units are combined to form medium-cycle segments. The long-cycle window is used to cover daily or operational cycle trends, and the medium-cycle segments are summarized into long-cycle segments in chronological order. In addition, during the segmentation process, the instruction sequence, power value, and risk level within each time period are retained. Based on the multi-scale scheduling unit partitioning results, the output order of high-risk units and key node units is prioritized, while the order of low-risk units is adjusted or the time is staggered to form a dynamic optimized scheduling and orchestration structure. The power output, time nodes, and risk levels of each unit in the dynamic optimization scheduling and orchestration structure are intelligently matched and mapped with the resources in the power generation unit to generate a preliminary hierarchical control path set. Specifically, the intelligent matching and mapping is as follows: First, based on the power output, time nodes, and risk levels of the control commands, power generation units with sufficient output capacity, overlapping grid connection time periods, and matching risk levels with the commands are selected. Then, high-risk units are mapped to power generation units according to priority, and low-risk units are mapped to auxiliary nodes. During the mapping process, the matching order of commands and resources is dynamically adjusted so that each command can be continuously output on the resources, while ensuring that high-risk commands are executed first during critical time periods, thereby forming a preliminary hierarchical control path set. Multi-granularity dynamic scheduling is implemented on the preliminary hierarchical control path set. By reordering the paths and fine-tuning the time nodes within short, medium and long period windows, the final optimized hierarchical control sequence is generated. High-risk units and key load nodes in the final hierarchical control execution sequence are marked with risk priority, while auxiliary output paths are reserved for low-risk units, forming a continuous, efficient, and risk-controllable hierarchical control execution sequence for all time periods.

[0030] In this embodiment, multi-granularity dynamic scheduling is implemented on the initial hierarchical control path set. By reordering paths and fine-tuning time nodes within short-cycle, medium-cycle, and long-cycle windows, the final optimized hierarchical control sequence is generated, as detailed below: Events are extracted from the initial hierarchical control path set, and key power changes, load peaks and islanding risk nodes triggered by each instruction in the execution sequence are marked as event points, thereby generating event-driven control units; Based on the event-driven control unit, the control unit is clustered. Each cluster contains high-risk and low-risk units that are related to each other within the same time period. The power output within each cluster is analyzed for dependency, and dynamic resource allocation grouping results are generated. For event nodes within each group in the dynamic resource allocation grouping results, short-term power trend and islanding risk evolution prediction are performed to generate a segmented dynamic control scheme. Within each segment, the power output order is fine-tuned to form a segmented optimization sequence. Specifically, the short-term power trend and islanding risk evolution prediction is performed as follows: First, the power output change trend of event nodes in each cluster over a recent period is extracted. The fluctuation pattern of short-term power increase or decrease is identified through the cumulative analysis of continuous power increase or decrease. At the same time, combined with the islanding dynamic evolution trajectory, it is determined whether the event nodes may be affected by the formation, continuation or dissipation of islands. Then, the power change trend of each event node is correlated with the trend of adjacent nodes to form a short-term power evolution chain. The potential power jumps or islanding risk accumulation points in the chain are predicted. Finally, the power output order and amplitude of the event nodes are locally adjusted according to the prediction results to smooth the short-term power trend and mitigate the islanding risk, thereby generating a segmented dynamic control scheme. The segmented optimization sequences are integrated by resource clustering, and the cross-cluster power output order is compressed and optimized to minimize the overall execution path's resource consumption throughout the entire time period, while prioritizing the execution of high-risk units. This generates a resource-compressed scheduling path. Specifically, the resource clustering integration involves: grouping units with the same resource attributes into the same cluster based on the power output units and grid connection interfaces corresponding to the power output units in each segmented optimization sequence; analyzing potential power overlaps or conflicts within the cluster and fine-tuning the output order locally to ensure continuous power output and minimized resource consumption within the cluster; marking units with cross-cluster dependencies and establishing a priority chain to ensure priority execution of high-risk units during critical periods; and finally... The optimized segmented sequences within each cluster are integrated according to the inter-cluster dependency order to form an execution sequence with optimal overall resource consumption, continuous clusters, and adjustable controllability, thereby generating a set of compressed optimization paths. The compression optimization specifically involves: first, analyzing the time overlap and dependency relationship of power output between different resource clusters to identify time periods with idle or redundant resources; then, appropriately shifting the power output of low-risk units or non-critical units forward or backward, while maintaining priority execution of high-risk units during critical time periods; next, locally merging cross-cluster power output sequences to compress continuous idle or low-load time periods, reducing idle gaps during inter-cluster switching, and finally forming a set of compressed optimization paths with minimal resource consumption, continuous power output, and priority for high-risk units. Multi-granularity time window scheduling is performed on the compressed optimization path set. Short-term, high-frequency event control is embedded in the short-cycle window, medium-term power regulation is embedded in the medium-cycle window, and long-cycle power trend is embedded in the long-cycle window to generate a multi-granularity time regulation scheme. The segmented optimization units in the multi-granularity dynamic execution sequence are merged and rearranged according to the event triggering order to generate an integrated hierarchical control path. Specifically, the fusion and rearrangement are as follows: First, the segmented optimization units are initially sorted according to the event triggering time, and the event nodes within the same time window are arranged according to priority. Then, the power output units within each cluster are matched with the cross-cluster dependency chain, and the order of units with intersections or potential conflicts is fine-tuned so that high-risk events are executed first. Next, the segmented sequences of short-term, medium-term and long-term are spliced ​​together according to time and resource continuity to form a continuous power output chain, while keeping the dependencies between each cluster unbroken, and finally generating an integrated hierarchical control path. Each execution path in the final hierarchical control sequence is marked with resource binding, and the event trigger point, power output order and risk level are recorded to form a hierarchical control execution sequence that can be executed directly at all times.

[0031] S4, executes multi-channel power splitting according to the hierarchical control execution sequence, and performs dynamic power compensation; In this embodiment, multi-channel power splitting is performed according to the hierarchical control execution sequence, and dynamic power compensation is performed, as follows: Based on the power output sequence, time node, and risk priority of each control unit in the hierarchical control execution sequence, the power generation channels and grid connection paths in each control unit are associated and marked to form a channel association execution structure; Based on the channel-associated execution structure, the power output of each control unit is decomposed into channels to form a multi-channel power distribution sequence. Specifically, the channel decomposition involves: according to the mapping relationship between each control unit and the power generation channel and grid connection path in the channel-associated execution structure, control units with the same grid connection path or consistent physical connection relationship are grouped into the same candidate channel. Then, according to the power output order and time node of the control unit, the power output in each candidate channel is time-expanded, and initial channel output segments are formed according to the execution order. Next, segments that are adjacent in time and have smooth power changes in the initial channel output segments are merged, and segments that overlap or conflict are split. Furthermore, according to the load capacity and risk priority of each channel, high-priority control units are kept in independent or priority channels, and low-priority control units are allocated to shared channels according to time segments. Finally, the adjusted power output segments in each channel are combined in time order to form a multi-channel power distribution sequence with clear channel affiliation and execution order. By tracking the changing trends of load nodes in each channel of the multi-channel power split sequence, the power allocation ratio is adjusted in real time to form adaptive split execution data. Based on the adaptive power splitting execution sequence, the power output rhythm between each channel is rearranged to stagger peaks and the instantaneous superimposed power peaks are reduced to form a staggered collaborative power splitting sequence. Based on the peak-shifting and coordinated power distribution sequence, local deviations generated during the power output of each channel are continuously tracked. Time segments with insufficient or excessive power are identified, and compensation markers are added to the corresponding channels, forming a set of power markers to be compensated. Specifically, the continuous tracking involves comparing the actual power output value of each channel in the peak-shifting and coordinated power distribution sequence with the target power within the corresponding time segment along the time axis, calculating the power deviation value at each time point, and forming a continuous deviation change curve in chronological order. Then, each time point is classified according to the positive or negative attribute of the power deviation value. When the deviation value is consistently negative, the corresponding time interval is marked as having insufficient power. In the power excess interval, when the deviation value is continuously positive, the corresponding time interval is marked as the power excess interval. At the same time, abnormal points with short-term fluctuations but not forming a continuous interval are filtered out. Then, the identified power deficiency intervals and power excess intervals are merged. Adjacent intervals with small intervals of the same type are spliced ​​together to form a continuous time segment. Further, the channel identifier, start and end time, deviation amplitude and change trend are recorded in each time segment, and a compensation mark is added to the time segment in the corresponding channel. Finally, the time segments marked in all channels are summarized in chronological order to form a set of power marks to be compensated that includes power deficiency and power excess. Based on the power sequence to be compensated, cross-channel dynamic compensation is performed on each channel to form a cross-channel collaborative compensation cluster. Specifically, the cross-channel dynamic compensation is as follows: First, based on the power insufficiency and redundancy time segments marked in the power sequence to be compensated, the power status of each channel in the corresponding time window is aligned and matched to determine the channel pairs that can be compensated. Then, the channels are sorted according to risk priority, and high-risk or critical load channels are given priority to be included in the compensation objects. Adjustable power is extracted from low-risk or redundant channels. Next, while maintaining the continuity of the original power output, the power output of redundant channels is locally reduced, and the reduced part is inserted into the insufficient interval of the target channel according to time segments. Finally, a power compensation cluster with multi-channel collaborative linkage and dynamic balance is formed. Trend adaptive correction is performed on the cross-channel collaborative compensation cluster to generate a dynamic compensation execution path. Specifically, the trend adaptive correction is as follows: First, the power changes before and after compensation in each channel of the cross-channel collaborative compensation sequence are continuously tracked to extract the trend of power increase or decrease and identify time segments with abrupt changes or discontinuities. Then, the trend of change is compared with the power evolution trends of adjacent time periods and adjacent channels, and compensation segments that deviate from the overall evolution direction are marked. Next, the power output order is refined and the compensation amount is distributed in segments. At the same time, the priority compensation of high-risk channels is not weakened during the correction process, and the compensation rhythm of low-risk channels is coordinated. Finally, the power changes of each channel are continuously connected on the time axis to form a smooth and consistent dynamic compensation execution sequence. Based on the dynamic compensation execution sequence, the final power output of each channel is adjusted throughout the entire time period, thereby forming a multi-channel power splitting and dynamic compensation execution data stream. Specifically, the adjustment throughout the entire time period involves: first, expanding the dynamic compensation execution sequence segment by segment along the time axis to continuously connect the power output of each channel in different time segments; then, based on the output trend of each channel throughout the entire cycle, refining the power output order and connection position to ensure a natural transition between adjacent time periods; next, while ensuring that high-risk channels are prioritized for output, the power output of low-risk channels is time-staggered or slightly adjusted; and simultaneously, the power distribution of each channel throughout the entire time period is coordinated as a whole, ultimately forming a continuous power output structure covering the entire operating cycle.

[0032] In this embodiment, based on the channel-associated execution sequence, the power output of each control unit is decomposed into channels to form a multi-channel power splitting sequence, as follows: The risk priority of each control unit in the channel-related execution sequence is hierarchically divided. Control units with high risk and critical load are marked as core units, and the remaining control units are marked as auxiliary units. An initial decomposition structure map is formed according to the risk level. Based on the initial decomposed structure map, the power output of the core unit is divided into independent channels. At the same time, the auxiliary units are combined according to time proximity and output continuity to generate a dynamically expandable channel decomposed configuration. The independent channel division is as follows: First, according to the power generation channel and grid connection path corresponding to the core unit in the initial decomposed structure map, the channel resources that can be independently carried are determined, and each core unit is assigned a dedicated channel identifier so that each core unit corresponds to a unique channel in the same time segment. Next, the power output of the core unit in the assigned channel is unfolded over time, and its power change in continuous time segments is statistically analyzed. For segments with time conflicts or power superposition risks, local misalignment adjustments are made. When a channel has an overload trend, the output time window of the core unit is finely adjusted without changing its channel affiliation to ensure the stable carrying capacity of the independent channel. Finally, each core unit is bound to its corresponding dedicated channel one by one to form a core channel division result that maintains independent operation and is not coupled with each other throughout the time period. The power load of each channel in the channel decomposition configuration is continuously tracked, and the power occupancy of each channel in different time segments is recorded, thereby forming the channel load evolution trajectory. Based on the channel load distribution sequence, regions where channel power load tends to be concentrated are identified. The power output of some auxiliary units within these channels is then migrated to channels with lower loads in time segments, forming a channel capacity balancing scheduling scheme. Specifically, the migration to channels with lower loads in time segments involves: dividing the power occupancy of each channel along a continuous time axis based on the channel load distribution sequence; identifying high-load time segments where power load exceeds a preset threshold; locating the set of auxiliary units participating in the output of these time segments within the corresponding channels; and then selecting auxiliary units from the set that have minimal impact on output continuity and do not involve core components. Candidate migration units with high load are identified, and their power output sub-segments within high load time segments are extracted. Next, target channels with larger load margins are selected, and a migration mapping relationship between source and target channels is established. Furthermore, the power output sub-segments of candidate migration units are separated from the original channels and inserted into the corresponding time positions of the target channels in the original time order. At the same time, the order of adjacent power segments within the target channel is adjusted. Finally, if there are new load imbalances or time conflicts, the time boundaries of the migration segments are redistributed to target channels until the load of each channel tends to be balanced and the overall output remains continuous and stable. Based on the channel capacity equalization allocation sequence, the time nodes of power mutation and risk surge are identified, and the relevant control units are temporarily redirected to the event response channel within the corresponding time segment to form an event-driven channel reconstruction flow; The event-driven channel reconstruction flow is time-integrated, and the event response channel is connected with the original channel structure to form a stable multi-channel splitting sequence. Specifically, the time-integration is as follows: First, the reconstruction segments in the event-driven channel reconstruction sequence are sorted according to the event trigger time, and the event response channel and the original channel segment within the same time window are aligned. Then, the power output sequence before and after the event occurs is connected, and the power output order is adjusted at the start and end positions of the event. Next, the power output segments in the event response channel are inserted into the corresponding channel positions in chronological order, and adjacent time segments are continuously spliced ​​together to finally form a multi-channel splitting operation framework that is time-continuous, structurally stable, and where each channel works in a coordinated manner. Based on the multi-channel power splitting operation framework, the final channel assignment of the power output of each control unit is determined, and the power output sequence within the channel is kept consistent with the time node, thereby generating a multi-channel power splitting sequence with risk isolation capability, load balancing capability and event response capability.

[0033] In this embodiment, based on the adaptive power splitting execution sequence, the power output rhythm between each channel is rearranged to stagger peaks and the instantaneous superimposed power peaks are reduced, forming a staggered collaborative power splitting sequence, as follows: Extract the power rise interval, stable interval and fall interval of each channel on the time axis in the adaptive power splitting execution sequence, and form a channel power rhythm feature sequence; Based on the channel power rhythm feature sequence, the power rise interval with time overlap is identified, and the power superposition of multiple channels in the same time segment is marked to form a peak superposition candidate domain. The power output order of each channel within the peak superposition candidate domain is dynamically rearranged to form a preliminary peak-shifting output layout. Specifically, the dynamic rearrangement involves: first, sorting the power rise start time and change rate of each channel within the peak superposition candidate interval, identifying the channel with faster power growth within the same period, and marking it as a priority channel; then, using the output timing of the priority channel as a benchmark, performing time offset processing on the power rise interval of the remaining channels, that is, slightly adjusting the start time of some channels forward or backward; next, maintaining the continuity of power output within each channel during the rearrangement process, synchronously delaying or compressing the subsequent time nodes of the adjusted channels, thereby forming a preliminary peak-shifting output layout that reduces instantaneous peak values. Based on the initial staggered output layout, channels with excessively rapid power increases are segmented and gradually released, while channels with gradual power changes are appropriately filled in the gaps in advance, generating a rhythm coordination control structure. Specifically, the segmented release involves: first, identifying channels whose power increase rate exceeds a preset threshold, and dividing their power increase interval into time slices, dividing the continuous power growth process into multiple sub-time periods, and allocating the power increment within each sub-time period to avoid concentrated power jumps in a short period of time; then, while maintaining the overall output trend, fine-tuning the start and end times of each sub-time period to make the power change present a stepped distribution on the time axis; subsequently, sequentially reorganizing the segmented sub-time periods, ultimately decomposing the originally continuous and rapidly increasing power output process into multiple smoothly connected segmented output units, thereby achieving gradual release and uniform control of the power increase process; Based on the rhythm coordination sequence, high-load channels and low-load channels are alternately arranged to form an alternating collaborative scheduling flow. The alternating arrangement is as follows: First, the load levels of each channel in the rhythm coordination sequence are sorted, and the channels are divided into a high-load channel set and a low-load channel set, and their corresponding time interval characteristics are marked. Then, according to the power output start time on the time axis, the high-load channels and low-load channels are time-aligned. Using the key output interval of the high-load channel as the anchor point, the output interval of the low-load channel is inserted in the time neighborhood before and after it, thus forming an alternating insertion relationship. Next, the high-load channels and low-load channels are mapped and matched one by one. Subsequently, the boundary clipping process is performed on the local overlapping areas. Finally, the power output intervals of the high-load channels and the power output intervals of the low-load channels are staggered on the time axis, thus forming a load-balanced and time-coordinated alternating collaborative scheduling flow. While prioritizing the output of high-risk channels, the power output of low-risk channels in the alternating coordinated scheduling flow is refined with time-shifted offsets to ensure a smooth transition of the overall power curve throughout the entire time period, generating a stable off-peak operation state. Specifically, the refined time-shifted offsets are as follows: First, power output segments of low-risk channels in the alternating coordinated sequence that are adjacent to or overlap with high-risk channels or high-load intervals on the time axis are identified and marked as intervals to be adjusted. Then, based on the power output timing of high-risk channels, the corresponding output segments of low-risk channels are time-shifted by appropriately advancing or delaying their start times. Subsequently, the offset relationships between multiple low-risk channels are coordinated to further disperse them in time distribution. Finally, while maintaining the stability of the overall execution order, the power output of low-risk channels is evenly distributed on the time axis, thereby achieving a smooth transition of the overall power curve. The power output sequence and time nodes of each channel in the stable peak shifting sequence are unified and integrated to ensure that the channel correlation is consistent with that in the original adaptive current shifting execution sequence, thereby forming a peak shifting collaborative current shifting sequence with peak shaving and valley filling capabilities.

[0034] Example 2, Figure 2 The present invention provides an apparatus for a multi-parameter-based grid-connected management method for small generators, comprising an energy data construction module, a risk situation optimization construction module, a scheduling mapping and orchestration module, and a diversion compensation control module, wherein the modules are interconnected. The energy data construction module is used to acquire comprehensive energy operation data of generator sets, generate high-precision energy data, and perform multi-scale island identification and dynamic evolution prediction on the high-precision energy data to form a multi-layer risk situation sequence. The risk situation optimization module is used to implement adaptive power allocation and multi-data grid connection strategy collaborative optimization based on multi-layer risk situation sequences, and generate real-time coordinated control commands. The scheduling mapping and orchestration module is used to perform heterogeneous resource mapping and multi-granularity scheduling and orchestration on real-time coordination and control commands, and generate hierarchical control execution sequences. The shunt compensation control module is used to perform multi-channel power shunt according to the hierarchical control execution sequence and to perform dynamic power compensation.

[0035] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0036] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0037] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0038] The above description is merely a specific technical solution of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0039] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for grid-connected management of small generators based on multiple parameters, characterized in that, include: Acquire comprehensive energy operation data of generator units, generate high-precision energy data, and perform multi-scale island identification and dynamic evolution prediction on the high-precision energy data to form a multi-layered risk situation sequence, including: Collect operating data from small generator sets to form a raw energy operation data set; The raw energy operation data is reconstructed using a global timescale and subjected to cross-channel constraint processing to form a standardized energy data sequence. A multi-timescale nested reconstruction is performed on the standardized energy data sequence to generate multi-scale coupled energy data. Perform continuous electrical state expansion on multi-scale coupled energy data to form a continuous state chain; Multi-window collaborative island identification is performed based on high-precision energy state representation, generating a set of island identification results; Temporal evolution correlation analysis is performed on the island identification result set to generate the dynamic evolution trajectory of the island; The dynamic evolution trajectory of isolated islands is constructed by multiple risk situation layers to form a multi-layer risk situation sequence; Based on a multi-layered risk situation sequence, adaptive power allocation and multi-data grid connection strategies are collaboratively optimized to generate real-time coordinated control commands, including: The risk levels, evolution trends, and isolated evolution paths of each layer in the multi-layered risk situation sequence are merged to form an initial risk mapping sequence; By performing risk-priority allocation and redundant buffer coordinated scheduling on the initial risk mapping sequence, preliminary power allocation candidate schemes are generated. Multi-data grid-connected collaborative optimization is performed on the preliminary power allocation candidate schemes, and the power output order of each power generation unit is adjusted through cross-channel dependency constraints to form a collaborative optimization result; Multi-granularity load matching is performed on the collaborative optimization results, and the power output sequence and time nodes are dynamically adjusted to generate a power execution sequence. Maintain consistency between the power output order and time nodes in the power execution sequence, and map the power execution sequence to a real-time control command sequence; Adaptive redundancy adjustment is implemented on the real-time control command sequence, and potential conflicting commands are automatically corrected based on multiple command paths and the risk level of the island, forming a corrected real-time control command sequence; Multidimensional performance optimization of the modified real-time control command sequence is performed, and the optimal control command path is obtained by jointly evaluating power smoothness, load matching degree and islanding suppression effect. Continuously monitor the optimal real-time control command sequence and compare the execution status with the risk situation; Heterogeneous resource mapping and multi-granularity scheduling orchestration are performed on real-time coordination and control commands to generate hierarchical control execution sequences; Multi-channel power splitting is performed according to the hierarchical control execution sequence, and dynamic power compensation is carried out.

2. The method for grid-connected management of small generators based on multiple parameters according to claim 1, characterized in that, The process of performing multi-window collaborative island identification based on high-precision energy state representation to generate an island identification result set is as follows: Extract data from each channel in the high-precision energy state representation to form a candidate path set; Calculate the evolution deviation of adjacent isolated candidate paths in the candidate path set, and mark and remove them to obtain potential isolated paths; The potential island paths are cumulatively counted within a continuous time window to obtain the main island evolution paths and low-frequency island candidate paths; Based on the main island evolution path and low-frequency island candidate paths, the mutations of each path are adjusted; Multipath fusion is performed on the adjusted set of isolated paths to generate the final set of isolated path identification results. A dynamic trajectory is formed by tracing the continuity and changing trends of the final island identification result set along the time axis.

3. The method for grid-connected management of small generators based on multiple parameters according to claim 1, characterized in that, The process involves multi-dimensional performance optimization of the modified real-time control command sequence, and by jointly evaluating power smoothness, load matching degree, and islanding suppression effect, the optimal control command path is obtained, as detailed below: A multi-dimensional performance sequence is generated based on the modified real-time control command sequence. Based on the multidimensional performance sequence, parallel calculations are performed on each path in terms of power smoothness, load matching degree and islanding suppression to form a parallel performance evaluation table. The parallel performance evaluation table is used to adjust the prediction for each candidate path, and the adjusted performance data is updated back to the parallel performance evaluation table. The Pareto optimal path in the parallel performance evaluation table is selected to form a preliminary optimized execution sequence; Risk-priority scheduling is implemented on the initially optimized execution sequence to generate optimized execution paths; Minor discontinuities and potential conflicts in the optimized candidate paths are corrected to form the final execution path; The final execution path is mapped to a sequence of real-time control instructions, and risk priority and execution order markers are retained in the instructions.

4. The method for grid-connected management of small generators based on multiple parameters according to claim 1, characterized in that, The process of mapping heterogeneous resources and orchestrating multi-granularity scheduling for real-time coordination and control commands to generate hierarchical control execution sequences is as follows: The real-time coordination and control command sequence is uniformly marked to generate a marked command feature set; A preliminary mapping is performed on the mark instruction feature set and available power generation units and grid connection interface resources to generate multiple candidate mapping path chains; Multiple candidate mapping path chains are hierarchically segmented to generate multi-scale scheduling unit partitioning results; Based on the multi-scale scheduling unit partitioning results, the output order of high-risk units and key node units is prioritized, while the order of low-risk units is adjusted and the time is staggered to form a dynamic optimized scheduling and orchestration structure. The dynamic optimization scheduling and orchestration structure is intelligently matched and mapped with the resources in the power generation unit to generate a preliminary hierarchical control path set. Multi-granularity dynamic scheduling is performed on the initial hierarchical control path set to generate the final optimized hierarchical control sequence; High-risk units and critical load nodes in the final hierarchical control execution sequence are marked with risk priority, while auxiliary output paths are reserved for low-risk units, forming a hierarchical control execution sequence for all time periods.

5. The method for grid-connected management of small generators based on multiple parameters according to claim 4, characterized in that, The process of performing multi-granularity dynamic scheduling on the initial hierarchical control path set to generate the final optimized hierarchical control sequence is as follows: Events are extracted from the initial hierarchical control path set to generate event-driven control units; Based on the event-driven control unit, the control unit is clustered, and the power output within each cluster is analyzed for dependency, generating dynamic resource allocation grouping results. For event nodes within each group in the dynamic resource allocation grouping results, short-term power trend and islanding risk evolution predictions are performed to generate segmented dynamic control schemes. Within each segment, the power output order is fine-tuned to form a segmented optimization sequence. The segmented optimization sequence is integrated according to resource clusters, and the cross-cluster power output order is compressed and optimized to generate a resource compression scheduling path; Perform multi-granularity time window scheduling on the compressed and optimized path set to generate a multi-granularity time control scheme; The segmented optimization units in the multi-granularity dynamic execution sequence are merged and rearranged according to the event triggering order to generate an integrated hierarchical control path. Each execution path in the final hierarchical control sequence is marked with resource binding to form a full-time hierarchical control execution sequence.

6. The method for grid-connected management of small generators based on multiple parameters according to claim 1, characterized in that, The process of performing multi-channel power splitting and dynamic power compensation according to the hierarchical control execution sequence is as follows: Based on the hierarchical control execution sequence, the power generation channels and grid connection paths in each control unit are associated and marked to form a channel association execution structure; Based on the channel-associated execution structure, the power output of each control unit is decomposed into channels to form a multi-channel power splitting sequence; By tracking the changing trends of load nodes in each channel of the multi-channel power split sequence, the power allocation ratio is adjusted in real time to form adaptive split execution data. Based on the adaptive power splitting execution sequence, the power output rhythm between each channel is rearranged to stagger peaks and the instantaneous superimposed power peaks are reduced to form a staggered collaborative power splitting sequence. Based on the peak-shifting and coordinated power distribution sequence, the local deviations generated during the power output process of each channel are continuously tracked, the time segments of insufficient and excessive power are identified, and compensation marks are marked in the corresponding channels to form a set of power marks to be compensated. Based on the power sequence to be compensated, cross-channel dynamic compensation is performed on each channel to form a cross-channel collaborative compensation cluster; Trend-adaptive correction is performed on cross-channel collaborative compensation clusters to generate dynamic compensation execution paths; Based on the dynamic compensation execution sequence, the final power output of each channel is adjusted throughout the entire time period, forming a multi-channel power splitting and dynamic compensation execution data stream.

7. The method for grid-connected management of small generators based on multiple parameters according to claim 6, characterized in that, The process involves rearranging the power output rhythm of each channel according to the adaptive power splitting execution sequence, and reducing the instantaneous superimposed power peaks to form a staggered collaborative power splitting sequence, as detailed below: Extract the power rise interval, stable interval, and fall interval from the adaptive power splitting execution sequence, and form a channel power rhythm feature sequence; Based on the channel power rhythm feature sequence, the power rise interval with time overlap is identified, and the power superposition of multiple channels in the same time segment is marked to form a peak superposition candidate domain. The power output order of each channel in the peak superposition candidate domain is dynamically rearranged to form a preliminary staggered peak output layout. Based on the initial staggered output layout, the channels with excessively rapid power increases are segmented and slowed down, and the gaps in channels with gradual power changes are filled in advance to generate a rhythm coordination control structure. Based on the rhythm coordination sequence, high-load channels and low-load channels are alternately arranged to form an alternating collaborative scheduling flow; While prioritizing the output of high-risk channels, the power output of low-risk channels in the alternating coordinated scheduling flow is refined by time-shifting to generate a stable peak-shifting operation state. The power output sequence and time nodes of each channel in the stable peak-shifting sequence are integrated to form a peak-shifting collaborative diversion sequence.

8. An apparatus using the multi-parameter-based grid-connected management method for small generators as described in any one of claims 1-7, characterized in that, It includes an energy data construction module, a risk situation optimization construction module, a scheduling mapping and orchestration module, and a diversion compensation control module, and the modules are interconnected: The energy data construction module is used to acquire comprehensive energy operation data of generator sets, generate high-precision energy data, and perform multi-scale island identification and dynamic evolution prediction on the high-precision energy data to form a multi-layer risk situation sequence. The risk situation optimization module is used to implement adaptive power allocation and multi-data grid connection strategy collaborative optimization based on multi-layer risk situation sequences, and generate real-time coordinated control commands. The scheduling mapping and orchestration module is used to perform heterogeneous resource mapping and multi-granularity scheduling and orchestration on real-time coordination and control commands, and generate hierarchical control execution sequences. The shunt compensation control module is used to perform multi-channel power shunt according to the hierarchical control execution sequence and to perform dynamic power compensation.

Images