Distributed power source hierarchical access optimization method considering inter-level output deviation transmission

Abstract

The application relates to the technical field of power system optimization control, and discloses a distributed power supply hierarchical access optimization method considering inter-level output deviation transmission. Operation data of single-line-layer distributed power supply is collected, and an initial deviation mode is obtained through fluctuation characteristic identification. The initial deviation mode is taken as input of a wiring unit layer, regional power imbalance and duration are predicted, the prediction results are taken as input of a transformer station layer, real-time converged power data are analyzed in association, and an amplitude amplification coefficient and a duration extension coefficient in an output deviation transmission process are determined. The prediction results are corrected according to transmission characteristic parameters, expected amplitude and expected duration after the deviation is transmitted to the transformer station layer are obtained, and the expected amplitude and the expected duration are compared with regulation capacity thresholds of regulation resources of all levels, and regulation tasks of different levels are distributed. The application realizes optimal matching of regulation resources and deviation characteristics by quantifying hierarchical transmission characteristics of the deviation.

Description

Technical Field

[0001] This invention relates to the field of power system optimization and control technology, and in particular to a distributed power source hierarchical access optimization method that considers the transmission of output deviations between levels. Background Technology

[0002] Large-scale integration of distributed power sources has become a core pathway to promote the clean and low-carbon transformation of energy. Its importance lies in its ability to absorb renewable energy locally, reduce long-distance transmission losses, and enhance the resilience of the power system. As the penetration rate of distributed power sources such as photovoltaics and distributed wind power in the distribution network continues to increase, the inherent randomness and intermittency of their output pose challenges to grid operation.

[0003] Currently, the control of distributed power sources is mostly based on independent planning at a single level or simple superposition, resulting in a lack of effective connection between different voltage levels. When distributed power sources are directly connected to a low-voltage single line, their output deviation only manifests as localized, small fluctuations. However, when a large number of power sources are simultaneously connected to the same wiring unit, these local deviations begin to superimpose, forming regional power imbalances. Further amplified at the substation level, the magnitude and duration of the deviations are significantly increased, evolving into a major pressure requiring additional peak and frequency regulation for the system. This bottom-up, step-by-step amplification process makes traditional methods that rely solely on the self-regulation capabilities of a single level ineffective.

[0004] Some solutions attempt to deploy automatic generation control or automatic voltage control systems at the substation level to smooth power fluctuations by centrally dispatching resources such as thermal power and energy storage; others focus on local fluctuation absorption on the low-voltage side through local inverter control. However, these solutions generally suffer from a fundamental flaw: while low-level fast resources such as inverters have fast response times, their small capacity often leads to their rapid depletion due to the need to cope with long-lasting, large fluctuations; conversely, high-level slow resources such as thermal power units, while having large capacity, have slow response times and often respond passively only after rapid fluctuations have occurred, resulting in delayed and costly regulation. This regulatory mismatch makes it difficult for the power grid to achieve efficient and precise balance control at the overall level when facing a high proportion of distributed power sources. Summary of the Invention

[0005] Therefore, the purpose of this invention is to overcome the problem of regulation mismatch between low-level fast resources and high-frequency small fluctuations, and between high-level slow resources and low-frequency large fluctuations. It provides a distributed power generation hierarchical access optimization method that considers the transmission of output deviations between levels. By quantifying the amplitude amplification factor and duration extension factor of the output deviation transmission between levels, and based on this, it achieves optimized matching where high-frequency small fluctuations are regulated by low-level fast resources and low-frequency large fluctuations are regulated by high-level slow resources. This adapts the characteristics of the regulating resources to the deviation transmission characteristics, improves regulation efficiency, reduces regulation costs, and enhances the grid's ability to accept a high proportion of distributed power generation.

[0006] To address the aforementioned technical problems, this invention provides a distributed power source hierarchical access optimization method considering the transmission of output deviations between levels, comprising: The operation data of each distributed power source under the single-line layer is collected, and the fluctuation characteristics of the operation data are identified to obtain the initial deviation pattern characterizing the power output fluctuation characteristics of the single-line layer. The initial deviation pattern is used as the input of the wiring unit layer to obtain the real-time power information of multiple single lines under this layer. Based on the comparison between the initial deviation pattern and the real-time power information of each single line, the regional power imbalance and duration of the wiring unit layer in the future time window are predicted. The prediction results of the wiring unit layer are used as the input of the substation layer to obtain the real-time converged power data of the substation layer; the prediction results and the real-time converged power data are correlated and analyzed to determine the transmission characteristic parameters in the process of the output deviation being transmitted from the wiring unit layer to the substation layer; the transmission characteristic parameters include the amplitude amplification factor and the duration extension factor; The prediction results are corrected based on the transmitted characteristic parameters to obtain the expected amplitude and expected duration of the deviation after it is transmitted to the substation level. The expected amplitude and expected duration are compared with the regulation capacity threshold of each level of regulation resources, and regulation tasks of different levels are allocated according to the matching principle that high-frequency small fluctuations are regulated by low-level resources and low-frequency large fluctuations are regulated by high-level resources.

[0007] Preferably, the fluctuation feature identification includes: extracting the peak period, valley period, fluctuation amplitude and fluctuation frequency of the output curve of each distributed power source; clustering the distributed power sources of the single-line layer based on the extracted features; grouping distributed power sources with similar fluctuation characteristics into the same type; and generating a corresponding initial deviation pattern for each type.

[0008] Preferably, predicting the regional power imbalance and duration within a future time window for the wiring unit layer includes: setting a prediction time window; comparing the real-time power information of each individual line with the corresponding initial deviation pattern point by point to calculate the output deviation time series of each individual line; superimposing the output deviation time series of all individual lines within the same wiring unit to obtain the net power imbalance time series of the wiring unit layer; extracting the maximum net power imbalance from the net power imbalance time series as the amplitude prediction result of the regional power imbalance; and extracting the continuous duration for which the net power imbalance exceeds a threshold value as the duration prediction result.

[0009] Preferably, determining the transmission characteristic parameters during the transmission of power deviation from the wiring unit layer to the substation layer includes: calculating the ratio of the actual power fluctuation amplitude of the substation layer to the predicted amplitude of the wiring unit layer at the same time section, and obtaining the amplitude amplification factor; calculating the ratio of the actual duration of the power fluctuation of the substation layer exceeding the threshold value to the predicted duration of the wiring unit layer at the same time section, and obtaining the duration extension factor.

[0010] Preferably, the prediction results are corrected based on the transmission characteristic parameters to obtain the expected amplitude and expected duration of the deviation after it is transmitted to the substation, including: multiplying the predicted amplitude value of the regional power imbalance predicted at the wiring unit level by the amplitude amplification factor to obtain the expected amplitude of the deviation after it is transmitted to the substation; and multiplying the predicted duration value of the regional power imbalance predicted at the line unit level by the duration extension factor to obtain the expected duration of the deviation after it is transmitted to the substation.

[0011] Preferably, the regulation capacity thresholds for each level of regulation resources include: the maximum adjustable capacity and the longest sustainable regulation time of single-line level regulation resources; the maximum adjustable capacity and the longest sustainable regulation time after regional aggregation at the wiring unit level; and the total regulation capacity, unit ramp rate, and longest sustainable regulation time of substation level callable units.

[0012] Preferably, the expected amplitude and expected duration are compared with the regulation capacity threshold of each level of regulation resources, including: if the expected amplitude is less than the maximum adjustable capacity of the single-line layer regulation resources and the expected duration is less than the longest sustainable regulation time of the single-line layer regulation resources, then it is determined that the single-line layer will perform the regulation task; if the expected amplitude exceeds the regulation capacity of the single-line layer but is less than the maximum adjustable capacity after regional aggregation of the wiring unit layer and the expected duration is less than the longest sustainable regulation time of the wiring unit layer, then it is determined that the wiring unit layer will perform the regulation task; if the expected amplitude exceeds the regulation capacity of the wiring unit layer or the expected duration exceeds the longest sustainable regulation time of the wiring unit layer, then it is determined that the substation layer will perform the regulation task.

[0013] Preferably, when it is determined that the regulation task is to be performed by a single-line layer, the regulation task is decomposed into sub-regulation tasks for each single line, and allocated according to the regulation capacity of each single line. This includes: extracting the fluctuation amplitude statistics and fluctuation frequency characteristics of each single line from the initial deviation mode, and determining the regulation response speed level corresponding to each single line based on the fluctuation frequency characteristics; obtaining the inverter operating status of each single line, calculating the current output deviation value of each single line based on the comparison result between the real-time power information and the initial deviation mode, and determining the available regulation capacity and available regulation time of each single line based on the inverter operating status; allocating the total regulation power amplitude of the regulation task proportionally according to the available regulation capacity of each single line to obtain the allocated power amplitude of each single line, and constraining the total regulation time of the regulation task according to the available regulation time of each single line to ensure that the allocated regulation time of each single line does not exceed its available regulation time; and correcting the allocation result according to the regulation response speed level of each single line, prioritizing the allocation of high-fluctuation-frequency regulation components to single lines with fast response speeds, and allocating low-fluctuation-frequency regulation components to single lines with slow response speeds to obtain the sub-regulation tasks of each single line.

[0014] Preferably, before decomposing the adjustment task into sub-adjustment tasks for each single line when determining that the adjustment task is performed by a single line layer, the method further includes: identifying abnormal single lines whose fluctuation amplitude exceeds a threshold based on the fluctuation amplitude statistics in the initial deviation mode of each single line; setting the adjustment response speed level of the abnormal single line to the highest priority; and giving priority to the abnormal single line to undertake the adjustment component matching its fluctuation amplitude when allocating sub-adjustment tasks.

[0015] Preferably, when determining that the adjustment task is to be performed by the wiring unit layer, the adjustment task is decomposed into coordinated adjustment tasks for each individual line based on the adjustment capacity threshold of each individual line in the region. This includes: calculating the current output deviation value and deviation change trend of each individual line based on the comparison result between the real-time power information and the initial deviation mode corresponding to each individual line; determining the available upper regulation capacity, available lower regulation capacity, and available adjustment time of each individual line based on the inverter's operating status; performing correlation analysis on the current output deviation value and deviation change trend of each individual line to identify individual lines in the output rising phase and individual lines in the output falling phase, and adjusting the adjustment task of the wiring unit layer according to the adjustment method. The process is decomposed to obtain an upward adjustment component that requires increased output and a downward adjustment component that requires decreased output. For the upward adjustment component, it is allocated proportionally according to the available upward adjustment capacity of each line, with priority given to lines in the output decreasing phase. For the downward adjustment component, it is allocated proportionally according to the available downward adjustment capacity of each line, with priority given to lines in the output increasing phase. The allocation results are constrained based on the available adjustment time of each line to ensure that the allocated adjustment time of each line does not exceed its available adjustment time, thus obtaining the coordinated adjustment task of each line. The coordinated adjustment task includes the adjustment direction, adjustment power amplitude, and adjustment duration.

[0016] Compared with the prior art, the above-described technical solution of the present invention has the following advantages: The distributed power generation hierarchical access optimization method of the present invention, which considers the transmission of output deviation between levels, quantifies the amplitude amplification factor and duration extension factor of the output deviation transmission between levels, and based on this, achieves optimized matching of high-frequency small fluctuations by rapid resource adjustment at lower levels and low-frequency large fluctuations by slow resource adjustment at higher levels. This makes the characteristics of adjustment resources adapt to the characteristics of deviation transmission, improves adjustment efficiency, reduces adjustment costs, and enhances the grid's ability to accept a high proportion of distributed power generation. Attached Figure Description

[0017] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein: Figure 1 This is a flowchart of a distributed power supply hierarchical access optimization method considering the transmission of power output deviation between levels in a preferred embodiment of the present invention; Figure 2 This is a flowchart illustrating the prediction of power deviation in the wiring unit layer in a preferred embodiment of the present invention; Figure 3 This is a flowchart illustrating the determination of deviation propagation characteristic parameters in a preferred embodiment of the present invention; Figure 4 This is a flowchart of the process of obtaining the expected value of the substation layer by correcting the prediction result in a preferred embodiment of the present invention; Figure 5 This is a flowchart of comparing the expected value with the threshold and determining the adjustment level in a preferred embodiment of the present invention; Figure 6 This is a flowchart of single-line layer task decomposition and allocation in a preferred embodiment of the present invention; Figure 7 This is a flowchart illustrating the task decomposition of the wiring unit layer coordination and adjustment in a preferred embodiment of the present invention. Detailed Implementation

[0018] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0019] In the actual operation of large-scale distributed generation connected to the distribution network, the following dilemmas are faced: As the output deviation of distributed generation is transmitted step by step from low-voltage single line to high-voltage substation, its amplitude and duration will be significantly amplified, resulting in a mismatch between regulation resources and deviation characteristics; Specifically: although low-level fast resources such as inverters have fast response speeds, they have small capacities and are often used to deal with large fluctuations that last for a long time, causing their capacity to be quickly exhausted; although high-level slow resources such as thermal power units have large capacities, they have slow response speeds and often respond passively when rapid fluctuations have already occurred, resulting in lagging regulation effects and high costs.

[0020] To address this technical problem, this invention provides a distributed power source hierarchical access optimization method that considers the transmission of output deviations between levels. The method involves collecting operational data from each distributed power source at the single-line level and identifying fluctuation characteristics to obtain an initial deviation pattern. This initial deviation pattern is then used as input to the wiring unit level, combined with real-time power information to predict the regional power imbalance and its duration. Further, the prediction results are used as input to the substation level, and the prediction results are correlated with real-time aggregated power data to determine the transmission characteristic parameters during the transmission of output deviations from the wiring unit level to the substation level, including amplitude amplification and duration extension coefficients. The prediction results are then corrected based on these transmission characteristic parameters to obtain the expected amplitude and duration of the deviation after transmission to the substation level. Finally, the expected amplitude and duration are compared with the regulation capacity thresholds of regulation resources at each level, and regulation tasks for different levels are allocated according to the matching principle that high-frequency small-amplitude fluctuations are regulated by lower-level resources, while low-frequency large-amplitude fluctuations are regulated by higher-level resources.

[0021] For ease of understanding, the following explains some key terms in this embodiment: A single-line layer refers to the physical area covered by a single low-voltage feeder line in a distribution network, from the low-voltage side of the distribution transformer to the user-side distributed power source access point. It is based on a single line as the basic unit.

[0022] Distributed power sources refer to various small-scale power generation devices connected to the power distribution network, such as photovoltaic, wind power, and small hydropower. They are characterized by being close to the load side and having certain random and intermittent power output characteristics.

[0023] The initial deviation model refers to a typical pattern representing the power output fluctuation characteristics of a single-line layer, obtained by identifying the fluctuation characteristics of the operating data of each distributed power source under a single-line layer. This model reflects the power output fluctuation pattern of distributed power sources under specific operating conditions.

[0024] The wiring unit layer refers to an intermediate level in the distribution network formed by multiple individual lines converging through feeders or distribution transformer clusters. It typically corresponds to 10kV or 20kV feeders, distribution transformer clusters, or distribution automation areas. This level manages several subordinate individual lines as a whole, and the multiple individual lines within its coverage area are electrically connected through the same busbar or the same distribution transformer to achieve power aggregation.

[0025] The substation level refers to a higher voltage level in a distribution network centered around a 35kV, 66kV, or 110kV substation. It typically includes a main transformer, busbars, switchgear, and dispatch automation systems. This level is connected to the upper-level power grid via the main transformer and supplies power to lower-level connection units through multiple feeders, making it a crucial hub connecting the transmission and distribution networks.

[0026] Transmission characteristic parameters refer to parameters used to quantify the changes in deviation characteristics during the transmission of output deviation from lower to higher levels. These parameters reflect the dynamic law of deviation transmission between levels.

[0027] The amplitude amplification factor is one of the transmission characteristic parameters, used to characterize the degree of amplification of the power imbalance amplitude when the output deviation is transmitted from one level to a higher level.

[0028] The duration extension factor is one of the transmission characteristic parameters, used to characterize the extent to which the power imbalance duration is extended when the output deviation is transmitted from one level to a higher level.

[0029] The regulation capacity threshold refers to the maximum regulation capacity and the longest sustainable regulation time that the regulation resources available at each level can provide to smooth out power fluctuations; this threshold is used to assess the ability of each level to cope with different types of deviations.

[0030] The following is combined Figure 1 The embodiments of the present invention will be described in detail below with reference to the specific implementation process.

[0031] The first step involves collecting operational data from each distributed power source within the single-line layer and identifying fluctuation characteristics in this data to obtain an initial deviation pattern characterizing the power output fluctuations of that single-line layer. Operational data is acquired by deploying sensors or data acquisition terminals at the access points of the distributed power sources, and the collected voltage, current, power, and other data are transmitted to the data processing center. Fluctuation characteristic identification employs basic statistical analysis methods, such as calculating the average, standard deviation, and maximum and minimum values ​​of historical power output data, thereby summarizing the typical fluctuation characteristics of the single-line layer distributed power source. The initial deviation pattern is defined as the sequence of differences between the actual power output and the average power output of the distributed power source within a target time period.

[0032] The second step involves deploying a regional coordination controller at the wiring unit layer, which is connected to the edge computing units of each subordinate single-line layer via a communication network. The initial deviation pattern obtained above is used as input to the wiring unit layer, while simultaneously acquiring real-time power information from multiple single lines under the wiring unit layer. Based on the comparison between the initial deviation pattern and real-time power information of each single line, the regional power imbalance and duration within a future time window are predicted for the wiring unit layer. The initial deviation pattern can be stored in the system database for use by the data processing module of the wiring unit layer. The regional coordination controller compares the real-time power information of each single line with the typical output curve in the corresponding initial deviation pattern point by point, calculating the output deviation time series of each single line. The comparison process involves subtracting the real-time power data of each single line from the corresponding initial deviation pattern point by point to obtain the real-time output deviation of each single line. The prediction of the regional power imbalance and its duration is achieved by algebraically superimposing the real-time output deviations of all single lines within the same wiring unit, extracting the maximum imbalance as the amplitude prediction from the superposition result, and extracting the continuous duration for which the imbalance exceeds a preset threshold as the duration prediction. In this way, the random fluctuations that were originally dispersed in the single-line layer are transformed into quantifiable predictable values ​​in the wiring unit layer.

[0033] The third step involves deploying a station-level monitoring host at the substation level. This host connects to the regional coordination controllers at each wiring unit level via the dispatch data network and also interfaces with the substation's monitoring and automation system. The station-level monitoring host receives the predicted amplitude and duration values ​​from the wiring unit level from the regional coordination controllers and obtains real-time aggregated power data from the monitoring and automation system, including active and reactive power on both the high-voltage and low-voltage sides of the main transformer. At the same time point, the station-level monitoring host compares the actual power fluctuation amplitude at the substation level with the predicted amplitude at the wiring unit level, calculates the ratio, and obtains the amplitude amplification factor. It also compares the actual duration of power fluctuations exceeding a certain threshold at the substation level with the predicted duration at the wiring unit level, calculates the ratio, and obtains the duration extension factor. To improve the accuracy of the coefficients, a sliding window method can be used for dynamic calculation. The amplitude amplification factor and duration extension factor are calculated separately within multiple consecutive time windows, forming a coefficient sequence. The median or average value of the coefficient sequence is then used as the typical amplitude amplification factor and typical duration extension factor under the current operating conditions.

[0034] The fourth step involves the station-level monitoring host multiplying the predicted amplitude value from the wiring unit layer by an amplitude amplification factor to obtain the expected amplitude after the deviation is transmitted to the substation; and multiplying the predicted duration value from the wiring unit layer by a duration extension factor to obtain the expected duration after the deviation is transmitted to the substation. These two expected values ​​reflect the final form of the deviation after it is transmitted to the substation level, enabling the system to predict the final impact of the deviation in advance. Then, the station-level monitoring host compares the expected amplitude and expected duration with predefined regulation capacity thresholds for each level of regulation resources. These thresholds include: the maximum adjustable capacity and longest sustainable regulation time of single-line level regulation resources; the maximum adjustable capacity and longest sustainable regulation time after regional aggregation at the wiring unit layer; and the total regulation capacity, unit ramp rate, and longest sustainable regulation time of the units that can be called upon at the substation level. The comparison process follows a hierarchical decision-making logic: if the expected amplitude is less than the maximum adjustable capacity of the single-line layer's adjustment resources and the expected duration is less than the longest sustainable adjustment time of the single-line layer's adjustment resources, then the adjustment task is determined to be performed by the single-line layer, corresponding to scenarios with high-frequency, small-amplitude fluctuations; if the expected amplitude exceeds the adjustment capacity of the single-line layer but is less than the maximum adjustable capacity after aggregation of the wiring unit layer's regions, and the expected duration is less than the longest sustainable adjustment time of the wiring unit layer, then the adjustment task is determined to be performed by the wiring unit layer, corresponding to scenarios with medium-frequency, medium-amplitude fluctuations; if the expected amplitude exceeds the adjustment capacity of the wiring unit layer or the expected duration exceeds the longest sustainable adjustment time of the wiring unit layer, then the adjustment task is determined to be performed by the substation layer, corresponding to scenarios with low-frequency, large-amplitude fluctuations. This hierarchical decision-making method ensures that the allocation result of the adjustment task accurately corresponds to the inherent characteristics of the deviation, avoiding resource waste or adjustment failure caused by mismatch.

[0035] The distributed power generation hierarchical access optimization method of the present invention, which considers the transmission of output deviation between levels, quantifies the amplitude amplification factor and duration extension factor of the output deviation transmission between levels, and based on this, achieves optimized matching of high-frequency small fluctuations by rapid resource adjustment at lower levels and low-frequency large fluctuations by slow resource adjustment at higher levels. This makes the characteristics of adjustment resources adapt to the characteristics of deviation transmission, improves adjustment efficiency, reduces adjustment costs, and enhances the grid's ability to accept a high proportion of distributed power generation.

[0036] In some embodiments described above in this application, fluctuation feature identification is proposed to obtain an initial deviation pattern characterizing the power output fluctuation characteristics of a single-line layer. In its implementation, distributed power sources in a single-line layer are numerous, geographically dispersed, and have varying power output characteristics. If a separate prediction model is established for each distributed power source, not only will the excessive data volume lead to an overburdened computational burden, but the highly random and uncertain output of a single power source will also render the prediction results statistically meaningless, failing to provide a stable and reliable benchmark for subsequent inter-layer propagation predictions.

[0037] In response, this application further proposes fluctuation feature identification including: extracting the peak period, valley period, fluctuation amplitude and fluctuation frequency of the output curve of each distributed power source; clustering the distributed power sources of the single-line layer based on the extracted features; grouping distributed power sources with similar fluctuation characteristics into the same type; and generating a corresponding initial deviation pattern for each type.

[0038] Specifically, the peak periods, valley periods, fluctuation amplitude, and fluctuation frequency of each distributed power source output curve are extracted to comprehensively capture the dynamic characteristics of the distributed power source output curve. The peak periods and valley periods reflect the typical high and low points of power output within a cycle, reflecting the periodicity. The fluctuation amplitude quantifies the severity of power output changes, reflecting its randomness. The fluctuation frequency describes the speed of power output changes, reflecting its intermittency.

[0039] Based on the extracted features, distributed power sources in the single-line layer are clustered and grouped, classifying distributed power sources with similar output fluctuation characteristics into one category. The aim is to form a power source set with internal consistency. By grouping similar power sources into one category, it is ensured that power sources within the same type have a high degree of similarity in output behavior. This allows the initial deviation patterns generated for this type to accurately represent the overall fluctuation characteristics of that type of power source, avoiding pattern distortion caused by excessive differences in power source characteristics.

[0040] For each type, a corresponding initial deviation pattern is generated. This initial deviation pattern will serve as a benchmark for the power output fluctuation characteristics of that type of power source in a single-line layer, and will be used for subsequent comparison with real-time power information to accurately predict the regional power imbalance.

[0041] When the output deviation of distributed power sources is transmitted from the single-line layer to the wiring unit layer, and then from the wiring unit layer to the substation layer, its amplitude and duration will be significantly amplified and prolonged. This invention further proposes a method for predicting the regional power imbalance and its duration at the wiring unit layer within a future time window, referring to... Figure 2 As shown, it includes: setting a prediction time window, comparing the real-time power information of each single line with the corresponding initial deviation mode point by point, and calculating the output deviation time series of each single line; superimposing the output deviation time series of all single lines in the same wiring unit to obtain the net power imbalance time series of the wiring unit layer; extracting the maximum net power imbalance from the net power imbalance time series as the amplitude prediction result of the regional power imbalance; and extracting the continuous duration of the net power imbalance exceeding the threshold value as the duration prediction result.

[0042] Specifically, the regional coordination controller compares the real-time power information of each individual line with the typical output curve in the corresponding initial deviation mode point by point, calculates the difference at each sampling time, and forms the output deviation time series of each individual line. The core of this comparison process is that the initial deviation mode serves as a benchmark for normal operation, so that any deviation from the benchmark in the real-time power information is quantified as an output deviation value, thereby realizing the transformation from raw data to deviation characteristics.

[0043] Subsequently, the regional coordination controller algebraically superimposes the time series of output deviations of all individual lines within the same wiring unit on the same time coordinate to obtain the time series of net power imbalance at the wiring unit level. The significance of this superposition operation lies in the fact that the output deviations of each individual line are not synchronized over time; simple algebraic summation cannot reflect the true convergence effect. However, by superimposing the time series point by point, the superposition effect of different individual line deviations over time can be accurately captured, obtaining a physically meaningful trajectory of the net power imbalance changing over time.

[0044] From the time series of net power imbalance, the regional coordination controller extracts the maximum net power imbalance as the amplitude prediction result of the regional power imbalance. This amplitude prediction result reflects the maximum power deficit or surplus that may occur at the wiring unit level within the prediction time window. Simultaneously, it extracts the continuous duration for which the absolute value of the net power imbalance exceeds a preset threshold as the duration prediction result. This duration prediction result reflects the possible length of time the power imbalance state may last. The threshold value is typically set to match the activation threshold of each level of regulation resources, for example, it may be set as the upper limit of the power range that a single-line inverter can autonomously adjust, so that deviations exceeding this threshold are considered abnormal states requiring further processing.

[0045] This scheme transforms the scattered fluctuations of a single-line layer into aggregated predictions of a wiring unit layer through time series overlay and feature extraction. This information convergence method from local to global means that the prediction results no longer depend on the accurate prediction of the output of a single power source, but rather utilize the superposition effect of multiple power sources in a statistical sense, thereby effectively improving the stability and reliability of the prediction.

[0046] Furthermore, referring to Figure 3 As shown, the transmission characteristic parameters of the output deviation are determined during the transmission process from the wiring unit layer to the substation layer.

[0047] In practice, the station-level monitoring host first performs time alignment to ensure that the predicted values ​​at the wiring unit level and the actual values ​​at the substation level correspond to the same time segment. Then, at the same time segment, the actual power fluctuation amplitude at the substation level is compared with the predicted amplitude at the wiring unit level, and the ratio is calculated to obtain the amplitude amplification factor. The technical meaning of this ratio is that it quantitatively reflects the degree to which the amplitude is amplified during the transmission of the deviation from the wiring unit level to the substation level. For example, if the amplitude amplification factor is 1.5, it means that a 100kW power deviation predicted at the wiring unit level will become a 150kW actual power fluctuation after being transmitted to the substation level.

[0048] Similarly, the actual duration of power fluctuations exceeding the same threshold value at the substation level is compared with the predicted duration at the wiring unit level, and the ratio of the two is calculated to obtain the duration extension coefficient. This ratio quantitatively reflects the degree to which the duration of the deviation is extended during transmission. It is particularly important to emphasize that the threshold value used in calculating the duration extension coefficient is strictly consistent with the threshold value used in extracting the duration prediction results, ensuring the comparability of the duration definitions at the two levels.

[0049] In practical applications, to improve the stability and accuracy of the coefficients, a sliding window method can be used for dynamic calculation. This involves calculating the amplitude amplification factor and duration extension factor within multiple consecutive time windows, forming a coefficient sequence, and then taking the median or average of the coefficient sequence as the typical amplitude amplification factor and typical duration extension factor under the current operating conditions. This dynamic calculation method allows the coefficients to adaptively follow changes in operating conditions; when weather conditions, load levels, or distributed power generation characteristics change, the coefficients will automatically update.

[0050] This embodiment effectively solves the problem of inaccurate predictions caused by the amplification of amplitude and duration of output deviations during inter-level transmission. By introducing amplitude amplification and duration extension coefficients, the ambiguous step-by-step amplification phenomenon is transformed into calculable and predictable quantitative parameters. Specifically, by calculating the ratio of the actual power fluctuation amplitude at the substation level to the predicted amplitude at the wiring unit level within the same time segment, the amplification factor of the deviation amplitude is quantified, avoiding errors introduced by time asynchrony. Simultaneously, by calculating the ratio of the actual duration of power fluctuations exceeding a threshold value at the substation level to the predicted duration at the wiring unit level within the same time segment, the degree to which the deviation duration is extended is accurately reflected. Furthermore, the threshold value is used to filter out minor fluctuations, improving the reliability and practicality of the coefficients. The determination of these transmission characteristic parameters ensures that the expected amplitude and duration of the deviation after transmission to the substation level more accurately reflect the actual situation.

[0051] Furthermore, referring to Figure 4 As shown, the prediction results are corrected based on the transmission characteristic parameters to obtain the expected amplitude and expected duration of the deviation after it is transmitted to the substation.

[0052] In practice, the station-level monitoring host multiplies the predicted amplitude value from the wiring unit layer by an amplitude amplification factor to obtain the expected amplitude after the deviation is transmitted to the substation; it also multiplies the predicted duration value from the wiring unit layer by a duration extension factor to obtain the expected duration after the deviation is transmitted to the substation. The principle behind this correction is that, assuming the deviation transmission pattern is stable within a short time window (meaning the currently calculated amplification and extension factors are also applicable within the prediction time window), multiplying the predicted future deviation by the current factors yields the expected form of the deviation after it is transmitted to the substation. For example, if the wiring unit layer predicts an 80kW regional power imbalance in the next 10 minutes, lasting 3 minutes, and the currently calculated amplitude amplification factor is 1.6 and the duration extension factor is 1.5, then the corrected result is an expected amplitude of 128kW at the substation level, lasting 4.5 minutes.

[0053] These two expected values ​​directly reflect the final impact of the deviation after it is transmitted to the substation level, enabling the system to predict the severity and time scale of the deviation in advance, providing a basis for subsequent resource matching decisions.

[0054] In practical applications, when the substation layer is affected by multiple wiring units at the same time, it is necessary to decompose the actual fluctuation of the substation layer into the contribution component of each wiring unit, and then use the ratio of this component to the predicted amplitude of the corresponding wiring unit layer as the amplitude amplification factor of that wiring unit.

[0055] Compared with the uncorrected prediction result, the corrected expected value provided by the present solution is closer to the actual situation, which effectively improves the prediction accuracy of the influence of the power unbalance amount at the substation layer. This enables the subsequent regulation task allocation to be carried out based on more reliable data, and avoids the problem of mismatched regulation resources caused by underestimating or overestimating the deviation. For example, if no correction is performed, the actual amplitude and duration at the substation layer may be underestimated, resulting in insufficient capacity or insufficient duration of the allocated low-level regulation resources, which requires high-level resources to intervene passively and increases the regulation cost; on the contrary, if the deviation is overestimated, high-level resources may be excessively invoked, resulting in resource waste. Through the present correction solution, the final influence of the deviation can be predicted in advance, which provides a decision-making basis for the matching principle that high-frequency small fluctuations are regulated by low-level resources and low-frequency large fluctuations are regulated by high-level resources, so as to realize more efficient and economical utilization of the overall power grid regulation resources.

[0056] In the above embodiment solution, the regulation capability threshold of each level of regulation resources is clearly defined.

[0057] In specific implementation, the substation-level monitoring host at the substation layer pre-stores the capability parameters of each layer of regulation resources. For the single-line layer regulation resources, they refer to the inverters supporting distributed power sources, and their maximum adjustable capacity is dynamically calculated according to the rated capacity and current output of the inverter. For example, when the current output of the inverter is 80% of the rated capacity, its maximum upward regulation capacity is 20% of the rated capacity, and its maximum downward regulation capacity is 80% of the rated capacity; its maximum sustainable regulation time is determined according to the heat capacity and heat dissipation conditions of the inverter, and is usually set to no more than 30 minutes of continuous regulation.

[0058] For the regulation resources after regional aggregation at the wiring unit layer, they refer to the set of all single-line layer inverters in the wiring unit. The maximum adjustable capacity after regional aggregation is the algebraic sum of the available regulation capacities of each single line, and the maximum sustainable regulation time takes the minimum sustainable regulation time among all single lines, so as to ensure the consistency of the regional regulation capability.

[0059] For the callable units at the substation layer, including thermal power units, gas turbines, pumped storage, etc., their total regulation capacity is the sum of the reserve capacities of all units. The unit ramp rate is determined according to the unit type. For example, the ramp rate of a gas turbine can reach 20% of the rated capacity per minute, the ramp rate of a thermal power unit is usually 2% to 5% of the rated capacity per minute, and the maximum sustainable regulation time is determined according to the fuel reserve or water head height of the unit.

[0060] Through the above technical solutions, the embodiments of the present invention clearly define the specific capacity thresholds of each level of regulation resources, thereby providing a basis for accurately matching regulation tasks with resource capabilities and effectively solving the problem of regulation mismatch. Specifically, by defining the maximum adjustable capacity and the longest sustainable regulation time of single-line layer regulation resources, it is ensured that the rapid response but limited capacity characteristics of this layer can be fully considered when allocating tasks, avoiding the allocation of fluctuations exceeding its processing capacity, thereby preventing resource overload and efficiency degradation. By defining the maximum adjustable capacity and the longest sustainable regulation time after regional aggregation at the wiring unit layer, the aggregation characteristics of this level can be used to handle medium-amplitude fluctuations, ensuring that regional coordination capabilities can be fully utilized when allocating tasks, improving the stability and reliability of regulation. At the same time, by defining the total regulation capacity, unit ramp rate, and longest sustainable regulation time of the callable units at the substation layer, especially by introducing the key feature of unit ramp rate, it is emphasized that this layer has the ability to handle large but slow fluctuations, ensuring that task allocation can match its slow response speed but large capacity characteristics, avoiding the lag and cost waste caused by passive response when rapid fluctuations occur.

[0061] Reference Figure 5 As shown, the expected amplitude and expected duration are compared with the regulation capacity thresholds of each level of regulation resources, and the matching principle determines which level should perform the regulation task.

[0062] In practice, the station-level monitoring host compares the expected amplitude and duration with the defined adjustment capability thresholds for each level. The comparison process follows a hierarchical decision-making logic: first, it determines whether the expected amplitude is less than the maximum adjustable capacity of the single-line layer adjustment resource, and whether the expected duration is less than the longest sustainable adjustment time of the single-line layer adjustment resource. If both conditions are met, it indicates that the deviation is a high-frequency, small-amplitude fluctuation, and the inverter of the single-line layer is fully capable of completing the adjustment within its own sustainable time. Therefore, it is determined that the adjustment task should be performed by the single-line layer.

[0063] If the expected amplitude exceeds the adjustment capacity of a single-line layer, the next layer of judgment is initiated: it is determined whether the expected amplitude is less than the maximum adjustable capacity after aggregation in the wiring unit layer area, and whether the expected duration is less than the longest sustainable adjustment time of the wiring unit layer. If both conditions are met simultaneously, it indicates that the deviation is a medium-frequency, medium-amplitude fluctuation, requiring coordinated adjustment of resources from multiple single lines within the aggregated wiring unit. Therefore, it is determined that the adjustment task will be performed by the wiring unit layer.

[0064] If the expected amplitude exceeds the adjustment capacity of the wiring unit layer, or the expected duration exceeds the maximum sustainable adjustment time of the wiring unit layer, it indicates that the deviation is a low-frequency large fluctuation that neither the single-line layer nor the wiring unit layer can handle independently. The substation layer's units must provide support, and therefore the substation layer is determined to perform the adjustment task.

[0065] This hierarchical judgment method ensures that the allocation results of adjustment tasks correspond precisely to the inherent characteristics of the deviation—high-frequency small fluctuations are handled by fast-response small-capacity resources, while low-frequency large fluctuations are handled by slow-response large-capacity resources, achieving the optimal match between resource characteristics and deviation characteristics.

[0066] Based on the above embodiments, a threshold for the adjustment capability of each layer of adjustment resources is proposed to provide a basis for resource allocation. However, in this process, there is a lack of clear rules to accurately match the appropriate layer of resources according to the magnitude and duration of the expected deviation, which leads to improper allocation of adjustment tasks. This may cause low-level resources to be exhausted quickly due to excessive fluctuations, or high-level resources to have a delayed effect due to rapid fluctuations in response, thereby causing resource waste and adjustment failure.

[0067] This invention, through quantitative comparison of the expected deviation amplitude and duration with the regulation capacity thresholds of each level of regulation resources, dynamically and intelligently matches the most suitable regulation level. Specifically, for high-frequency, small-amplitude fluctuations, single-line layer resources with fast response speeds are prioritized for regulation, fully utilizing their rapid response characteristics and avoiding unnecessary intervention and startup costs of higher-level resources. For medium-amplitude fluctuations, the wiring unit layer coordinates regulation through regional aggregation capabilities, effectively sharing the pressure on the single-line layer and preventing overload. For low-frequency, large-amplitude, or long-term fluctuations, the large-capacity resources of the substation layer provide backup regulation, ensuring the stability and reliability of the system when facing severe power imbalances and preventing lower-level resources from failing due to insufficient capacity or duration limitations. This feature-matching-priority regulation paradigm effectively improves the overall regulation efficiency and economy after distributed power sources are connected to the grid.

[0068] In the above-described embodiments of the present invention, it is proposed that a single-line layer be used to perform the regulation task to handle high-frequency small-amplitude fluctuations. However, in this process, how to reasonably decompose the regulation task to each single line and match its regulation capacity to avoid uneven resource allocation, response delay, or overload risks caused by ignoring the real-time status and characteristics of the single line, thereby affecting the overall regulation efficiency and system stability, remains a challenge. The single-line layers not only differ in available regulation capacity and regulation duration, but also in their inherent fluctuation frequency characteristics. If allocation is based solely on capacity, lines with slow response speeds may be assigned high-frequency regulation components, failing to track fluctuation changes in a timely manner.

[0069] To address this issue, the embodiments of the present invention further propose that, when determining that a regulation task is to be performed by a single-line layer, the regulation task is decomposed into sub-regulation tasks for each single line, and allocated according to the regulation capacity of each single line. (Refer to...) Figure 6 As shown, it specifically includes the following steps.

[0070] The regional coordination controller at the wiring unit level receives the judgment results of the adjustment tasks performed by the single-line layer, as well as the total adjustment power amplitude and total adjustment duration of the adjustment tasks from the station-level monitoring host.

[0071] The regional coordination controller first extracts the fluctuation amplitude statistics and fluctuation frequency characteristics from the initial deviation patterns of each individual line. Based on the fluctuation frequency characteristics, it determines the corresponding regulation response speed level for each individual line. For example, the individual line with high fluctuation frequency (such as the rapid fluctuation of photovoltaic power caused by cloud shading) is set to a high response speed level, and the individual line with low fluctuation frequency (such as the fluctuation of wind power caused by slow changes in wind speed) is set to a low response speed level.

[0072] Then, the regional coordination controller obtains the inverter operating status of each line, calculates the current output deviation value of each line based on the comparison results of real-time power information and initial deviation mode, determines the available regulation capacity of each line based on the real-time rated capacity and current output of the inverter, and determines the available regulation time of each line based on the thermal state of the inverter.

[0073] Next, the regional coordination controller allocates the total regulation power amplitude of the regulation task proportionally to the available regulation capacity of each individual line. For example, if the available regulation capacity of a certain individual line accounts for 20% of the total capacity, then 20% of the total regulation power amplitude is allocated. At the same time, the total regulation duration of the regulation task is compared with the available regulation duration of each individual line to ensure that the regulation duration allocated to each individual line does not exceed its available regulation duration. If the available regulation duration of a certain individual line is insufficient, the allocated power amplitude of that individual line is reduced accordingly, and the reduced part is redistributed to other individual lines.

[0074] Finally, the regional coordination controller corrects the allocation results based on the regulation response speed level of each line. It prioritizes allocating regulation components with high fluctuation frequency to lines with fast response speeds and allocates regulation components with low fluctuation frequency to lines with slow response speeds. The significance of this correction step is that high-frequency fluctuation components need to track changes quickly and must be handled by inverters with fast response speeds, otherwise regulation lag will occur; low-frequency fluctuation components change slowly and can be handled by inverters with slow response speeds, thus reserving fast resources for more needed high-frequency tasks.

[0075] After the above allocation and correction, the regional coordination controller obtains the sub-regulation tasks of each individual line, including regulating power amplitude and regulation duration.

[0076] Furthermore, when determining whether a single-line layer is responsible for performing the regulation task, a mechanism for prioritizing abnormal single lines has been added. In specific implementation, before the aforementioned allocation process, the regional coordination controller first identifies abnormal single lines whose fluctuation amplitude exceeds a preset threshold based on the statistical values ​​of the fluctuation amplitude in the initial deviation patterns of each single line. For example, if the initial deviation pattern of a single line shows its normal fluctuation amplitude is within ±5%, but current real-time data indicates its output fluctuation amplitude reaches ±20%, far exceeding the normal range, then this single line is identified as an abnormal single line.

[0077] The regional coordination controller sets the response speed of these abnormal lines to the highest priority and prioritizes them for regulation tasks when allocating sub-regulation tasks. The principle behind this is that lines experiencing abnormal fluctuations are both the bearers of regulation tasks and the sources of deviations. Allowing other normal lines to regulate the deviations of these abnormal lines would not only increase unnecessary communication and coordination costs but also waste the regulation resources of the abnormal lines themselves. By prioritizing abnormal lines for regulation tasks matching their own fluctuations, a local mitigation strategy of prioritizing regulation based on the source of deviation is implemented, effectively preventing the propagation and amplification of deviations to higher levels.

[0078] Furthermore, referring to Figure 7 As shown, it is further proposed that when determining that the adjustment task is performed by the wiring unit layer, the adjustment task is decomposed into the coordinated adjustment task of each single line in the area according to the adjustment capability threshold of each single line.

[0079] In practice, the regional coordination controller at the wiring unit level receives from the station-level monitoring host the judgment results of the adjustment tasks performed by the wiring unit level, as well as the total adjustment power amplitude and total adjustment duration of the adjustment tasks.

[0080] The regional coordination controller first calculates the current output deviation value and deviation trend of each line based on the comparison results between the real-time power information of each line and the corresponding initial deviation mode. The deviation trend is obtained by calculating the time derivative of the output deviation value. The positive derivative represents the output rising stage, and the negative derivative represents the output falling stage.

[0081] At the same time, the available up-regulation capacity, available down-regulation capacity, and available regulation duration of each single line are determined based on the inverter's operating status.

[0082] Then, the regional coordination controller analyzes the current output deviation value and deviation change trend of each line, identifies the lines in the output rising phase and the lines in the output falling phase, and decomposes the adjustment task of the wiring unit layer according to the adjustment direction to obtain the upward adjustment component that needs to increase output and the downward adjustment component that needs to reduce output.

[0083] Next, the regional coordination controller adopts a trend-following adjustment principle: for upward adjustment components, allocation is made proportionally according to the available upward adjustment capacity of each individual line, with priority given to lines in the output decline phase; for downward adjustment components, allocation is made proportionally according to the available downward adjustment capacity of each individual line, with priority given to lines in the output increase phase. The core purpose of this allocation principle is that when the output of an individual line naturally declines, the decline itself is a negative deviation, and upward adjustment is to compensate for this negative deviation. The two are aligned, forming a synergistic effect, with the adjustment action complementing rather than opposing the natural trend. When the output of an individual line naturally increases, the increase itself is a positive deviation, and downward adjustment is to offset this positive deviation, again forming a synergistic effect. Through this trend-following adjustment, some adjustment is covered by the natural trend itself, and the actual power amplitude that the inverter needs to actively adjust is less than the total task amplitude, thus achieving the adjustment task by leveraging external forces.

[0084] Finally, the regional coordination controller constrains the allocation results based on the available adjustment time of each individual line to ensure that the allocated adjustment time of each individual line does not exceed its available adjustment time, thereby obtaining the coordination adjustment task of each individual line. Each coordination adjustment task includes the adjustment direction (up or down), adjustment power amplitude, and adjustment duration.

[0085] The present invention, through obtaining the current output deviation value and deviation change trend of each line, and combining this with the inverter's operating status to determine its available upward / downward regulation capacity and regulation duration, can comprehensively grasp the real-time regulation potential of each line. Based on this, correlation analysis of the deviation value and change trend identifies lines in the output increase or decrease phase, ensuring that the decomposition and allocation of regulation tasks match the natural change trends of the lines. The regulation task is decomposed into upward and downward adjustment components, with upward adjustment components preferentially allocated to lines with decreasing output and downward adjustment components preferentially allocated to lines with increasing output. This effectively avoids the conflict between the regulation task and the natural trend of the lines, significantly reducing the regulation burden and improving regulation efficiency.

[0086] The following specific example will provide a more in-depth explanation of the above technical solution: Consider a power distribution network serving an industrial park with multiple distributed power sources, including photovoltaic power generation systems installed on the roofs of different factory buildings. These photovoltaic systems are connected to low-voltage single lines through their respective inverters. Multiple low-voltage single lines converge to form medium-voltage connection units, and finally the power of all connection units is converged to the park's main substation.

[0087] First, operational data for each distributed power source at the single-line layer is continuously collected. For example, for a photovoltaic system connected to single-line A, its output data per minute is recorded. Next, fluctuation characteristics are identified in this operational data. Specifically, the system extracts the peak periods, valley periods, fluctuation amplitude, and fluctuation frequency of the photovoltaic output curve at single-line A. By performing similar analysis on all distributed power sources at the single-line layer within the park, and clustering based on the extracted characteristics, photovoltaic systems with similar fluctuation characteristics are grouped into the same type. For example, all photovoltaic systems with the same orientation and similar shading conditions are grouped together. For each type, an initial deviation model characterizing its output fluctuation characteristics is generated. This model reflects the output fluctuation pattern of that type of photovoltaic system under typical weather conditions.

[0088] Subsequently, these initial deviation patterns are used as inputs to the wiring unit layer. The system acquires real-time power information of multiple lines (e.g., single line A, single line B, and single line C) under wiring unit X. To predict the regional power imbalance and duration of wiring unit X within a future time window, a prediction time window is set. Within this window, the real-time power information of single line A is compared point-by-point with its corresponding initial deviation pattern to calculate the output deviation time series of single line A. The same operation is performed on single lines B and C. Then, the output deviation time series of all lines within wiring unit X are superimposed to obtain the net power imbalance time series at the wiring unit X level. From this net power imbalance time series, the maximum net power imbalance is extracted as the amplitude prediction result of the regional power imbalance, and the continuous duration for which the net power imbalance exceeds a preset threshold value is extracted as the duration prediction result. Unlike traditional methods that only respond after the actual imbalance occurs, this method predicts the potential power imbalance at the wiring unit level in advance by aggregating and predicting the fluctuation patterns of single lines, thus gaining time for subsequent adjustments.

[0089] Next, the predicted results of the wiring unit X layer (i.e., the predicted amplitude and duration of regional power imbalance) are used as input for the substation layer. Simultaneously, real-time aggregated power data for the substation layer is acquired. By correlating the predicted results with the real-time aggregated power data, the transmission characteristic parameters of the output deviation from the wiring unit layer to the substation layer are determined. These parameters include the amplitude amplification factor and the duration extension factor. Specifically, the ratio of the actual power fluctuation amplitude of the substation layer to the predicted amplitude of the wiring unit layer at the same time segment is calculated to obtain the amplitude amplification factor. For example, if the 1 MW imbalance predicted by the wiring unit layer actually manifests as a 1.5 MW fluctuation at the substation layer, the amplitude amplification factor is 1.5. Similarly, the ratio of the actual duration of the power fluctuation exceeding the threshold value at the substation layer to the predicted duration of the wiring unit layer at the same time segment is calculated to obtain the duration extension factor. For example, if the 10-minute duration predicted by the wiring unit layer is actually extended to 15 minutes at the substation layer, the duration extension factor is 1.5. These transmission characteristic parameters quantify the effect of the distributed power output deviation being amplified at each level, providing a basis for subsequent precise adjustment.

[0090] Based on the determined transmission characteristic parameters, the system corrects the prediction results at the wiring unit level to obtain the expected amplitude and duration of the deviation after it is transmitted to the substation level. For example, if the amplitude of the regional power imbalance predicted by the wiring unit level is 2 MW and the duration is 20 minutes, and the amplitude amplification factor is 1.5 and the duration extension factor is 1.5, then the expected amplitude at the substation level after correction is 3 MW and the expected duration is 30 minutes.

[0091] Finally, the expected amplitude and duration are compared with the regulation capacity thresholds of each level of regulation resources. Based on the matching principle that high-frequency, small-amplitude fluctuations are regulated by lower-level resources and low-frequency, large-amplitude fluctuations by higher-level resources, regulation tasks are allocated to different levels. The regulation capacity thresholds of each level of regulation resources include: the maximum adjustable capacity and longest sustainable regulation time of the single-line layer inverter; the maximum adjustable capacity and longest sustainable regulation time after regional aggregation at the wiring unit layer; and the total regulation capacity, unit ramp rate, and longest sustainable regulation time of the substation layer's callable units.

[0092] The specific allocation logic is as follows: 1. If the expected amplitude (e.g., 0.4 MW) is less than the maximum adjustable capacity of single-line A's regulation resource (e.g., 0.5 MW), and the expected duration (e.g., 10 minutes) is less than the maximum sustainable regulation time of single-line A's regulation resource (e.g., 15 minutes), then it is determined that single-line A will perform the regulation task. In this case, the system further decomposes the regulation task into sub-regulation tasks for single-line A. For example, based on the fluctuation amplitude statistics in the initial deviation mode of single-line A, it identifies whether it is an abnormal single-line with fluctuation amplitude exceeding a threshold, and prioritizes the abnormal single-line to undertake the regulation component matching its fluctuation amplitude. Simultaneously, based on the regulation response speed level of single-line A, lines with faster response speeds are prioritized for allocation of regulation components with higher fluctuation frequencies. This refined allocation avoids the improper use and rapid depletion of low-level resources.

[0093] 2. If the expected amplitude (e.g., 1.5 MW) exceeds the regulation capacity of a single line A but is less than the maximum adjustable capacity of the area after aggregation in wiring unit X (e.g., 2 MW), and the expected duration (e.g., 25 minutes) is less than the longest sustainable regulation time of wiring unit X (e.g., 30 minutes), then it is determined that wiring unit X will perform the regulation task. In this case, the system will decompose the regulation task into coordinated regulation tasks for each line based on the regulation margin of each line within the area. For example, lines in the output increase phase and lines in the output decrease phase are identified, and the regulation task is decomposed into upward and downward regulation components according to the regulation direction, and preferentially assigned to the lines that need corresponding regulation. This coordinated regulation fully utilizes the complementarity of the lines within the wiring unit.

[0094] 3. If the expected amplitude (e.g., 4 MW) exceeds the regulation capacity of connection unit X, or the expected duration (e.g., 45 minutes) exceeds the maximum sustainable regulation time of connection unit X, then the regulation task shall be performed by the substation level.

[0095] By employing the aforementioned hierarchical access optimization method, this solution effectively addresses the issues of distributed power generation output deviation amplifying across different levels and the mismatch of traditional regulation resources. Through quantitative analysis and prediction of deviation propagation characteristics, it achieves precise matching and efficient utilization of regulation resources at different levels. This prevents lower-level fast resources from being prematurely exhausted due to large, long-term fluctuations, and also avoids higher-level slow resources from becoming inefficient due to response lags, thereby improving the overall balance control capability and operational economy of the power grid.

[0096] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A distributed power generation hierarchical access optimization method considering the transmission of output deviation between levels, characterized in that, include: The operation data of each distributed power source under the single-line layer is collected, and the fluctuation characteristics of the operation data are identified to obtain the initial deviation pattern characterizing the power output fluctuation characteristics of the single-line layer. The initial deviation pattern is used as the input to the wiring unit layer to obtain the real-time power information of multiple single lines under this layer. Based on the comparison between the initial deviation pattern and the real-time power information of each single line, the regional power imbalance and duration of the wiring unit layer within a future time window are predicted. Specifically, this includes: setting a prediction time window; comparing the real-time power information of each single line with the corresponding initial deviation pattern point by point to calculate the output deviation time series of each single line; superimposing the output deviation time series of all single lines within the same wiring unit to obtain the net power imbalance time series of the wiring unit layer; extracting the maximum net power imbalance from the net power imbalance time series as the amplitude prediction result of the regional power imbalance; and extracting the continuous duration for which the net power imbalance exceeds a threshold value as the duration prediction result. The prediction results of the wiring unit layer are used as the input of the substation layer to obtain the real-time converged power data of the substation layer; the prediction results and the real-time converged power data are correlated and analyzed to determine the transmission characteristic parameters in the process of the output deviation being transmitted from the wiring unit layer to the substation layer; the transmission characteristic parameters include the amplitude amplification factor and the duration extension factor; The prediction results are corrected based on the transmitted characteristic parameters to obtain the expected amplitude and expected duration of the deviation after it is transmitted to the substation level. The expected amplitude and expected duration are compared with the regulation capacity threshold of each level of regulation resources, and regulation tasks of different levels are allocated according to the matching principle that high-frequency small fluctuations are regulated by low-level resources and low-frequency large fluctuations are regulated by high-level resources. Determining the transmission characteristic parameters of the output deviation from the wiring unit layer to the substation layer includes: calculating the ratio of the actual power fluctuation amplitude of the substation layer to the predicted amplitude of the wiring unit layer at the same time section, and obtaining the amplitude amplification factor; calculating the ratio of the actual duration of the power fluctuation of the substation layer exceeding the threshold value to the predicted duration of the wiring unit layer at the same time section, and obtaining the duration extension factor.

2. The distributed power generation hierarchical access optimization method considering the transmission of output deviation between levels as described in claim 1, characterized in that, The fluctuation feature identification includes: extracting the peak period, valley period, fluctuation amplitude and fluctuation frequency of the output curve of each distributed power source; clustering the distributed power sources of the single-line layer based on the extracted features; grouping distributed power sources with similar fluctuation characteristics into the same type; and generating a corresponding initial deviation pattern for each type.

3. The distributed power generation hierarchical access optimization method considering the transmission of output deviation between levels as described in claim 1, characterized in that, Based on the transmission characteristic parameters, the prediction results are corrected to obtain the expected amplitude and expected duration of the deviation after it is transmitted to the substation, including: The predicted amplitude of the regional power imbalance predicted at the wiring unit level is multiplied by the amplitude amplification factor to obtain the expected amplitude of the deviation after it is transmitted to the substation. The predicted duration of the regional power imbalance predicted at the wiring unit level is multiplied by the duration extension factor to obtain the expected duration after the deviation is transmitted to the substation.

4. The distributed power generation hierarchical access optimization method considering the transmission of output deviation between levels as described in claim 1, characterized in that, The adjustment capacity thresholds for resources at each level include: The maximum adjustable capacity and longest sustainable adjustment time of a single-line layer regulation resource; The maximum adjustable capacity and longest sustainable adjustment time after the wiring unit layer area is aggregated; The substation level can access the total regulating capacity of the generating units, the unit ramp rate, and the longest sustainable regulating time.

5. The distributed power generation hierarchical access optimization method considering the transmission of output deviation between levels as described in claim 4, characterized in that, The expected amplitude and expected duration are compared with the regulation capacity thresholds of regulation resources at each level, including: If the expected amplitude is less than the maximum adjustable capacity of the single-line layer adjustment resource, and the expected duration is less than the longest sustainable adjustment time of the single-line layer adjustment resource, then it is determined that the adjustment task will be performed by the single-line layer. If the expected amplitude exceeds the adjustment capability of a single-line layer but is less than the maximum adjustable capacity after aggregation of the wiring unit layer area, and the expected duration is less than the longest sustainable adjustment time of the wiring unit layer, then it is determined that the adjustment task will be performed by the wiring unit layer. If the expected amplitude exceeds the regulation capacity of the wiring unit layer, or the expected duration exceeds the maximum sustainable regulation time of the wiring unit layer, then the regulation task shall be performed by the substation layer.

6. The distributed power generation hierarchical access optimization method considering the transmission of output deviation between levels as described in claim 5, characterized in that, When determining that a regulation task should be performed by a single-line layer, the regulation task is decomposed into sub-regulation tasks for each single line, and allocated according to the regulation capacity of each single line, including: The fluctuation amplitude statistics and fluctuation frequency characteristics of each individual line are extracted from the initial deviation mode, and the adjustment response speed level corresponding to each individual line is determined based on the fluctuation frequency characteristics. The inverter operating status of each line is obtained. Based on the comparison result between the real-time power information and the initial deviation mode, the current output deviation value of each line is calculated. Based on the inverter operating status, the available regulation capacity and available regulation time of each line are determined. The total adjustment power amplitude of the adjustment task is proportionally allocated according to the available adjustment capacity of each single line to obtain the allocated power amplitude of each single line. The total adjustment time of the adjustment task is constrained according to the available adjustment time of each single line to ensure that the allocated adjustment time of each single line does not exceed its available adjustment time. The allocation results are corrected based on the regulation response speed level of each individual line. Individual lines with fast response speeds are assigned regulation components with high fluctuation frequencies first, while individual lines with slow response speeds are assigned regulation components with low fluctuation frequencies, thus obtaining the sub-regulation tasks of each individual line.

7. The distributed power generation hierarchical access optimization method considering the transmission of output deviation between levels as described in claim 5 or 6, characterized in that, Before decomposing the adjustment task into sub-adjustment tasks for each single line when determining that the adjustment task is to be performed by a single line layer, the process further includes: identifying abnormal single lines whose fluctuation amplitude exceeds a threshold based on the fluctuation amplitude statistics in the initial deviation mode of each single line; setting the adjustment response speed level of the abnormal single line to the highest priority; and giving priority to the abnormal single line to undertake the adjustment component that matches its fluctuation amplitude when allocating sub-adjustment tasks.

8. The distributed power generation hierarchical access optimization method considering the transmission of output deviation between levels as described in claim 5, characterized in that, When determining that the adjustment task should be performed by the wiring unit layer, the adjustment task is decomposed into coordinated adjustment tasks for each individual line based on the adjustment capability threshold of each line within the area, including: Based on the comparison results between the real-time power information and the initial deviation mode corresponding to each individual line, the current output deviation value and deviation change trend of each individual line are calculated; the available upper regulation capacity, available lower regulation capacity and available regulation time of each individual line are determined according to the operating status of the inverter. Correlation analysis is performed on the current output deviation value and deviation change trend of each single line to identify the single line in the output rising stage and the single line in the output falling stage. The adjustment task of the wiring unit layer is decomposed according to the adjustment direction to obtain the upward adjustment component that needs to increase the output and the downward adjustment component that needs to reduce the output. For upward adjustment components, the allocation is made proportionally according to the available upward adjustment capacity of each line, with priority given to lines in the output decreasing phase; for downward adjustment components, the allocation is made proportionally according to the available downward adjustment capacity of each line, with priority given to lines in the output increasing phase. The allocation results are constrained based on the available adjustment time of each individual line to ensure that the allocated adjustment time of each individual line does not exceed its available adjustment time, thereby obtaining the coordinated adjustment task of each individual line; the coordinated adjustment task includes adjustment direction, adjustment power amplitude and adjustment duration.

Images