A method for source-load response identification and operation segmentation of cascade hydropower stations in response to changes in grid load and renewable energy.

Abstract

This invention belongs to the technical field of hydropower station operation characteristics. Specifically, it proposes a method for identifying the source-load response and segmenting the operation of cascade hydropower stations in response to changes in grid load and renewable energy. The method includes: acquiring and preprocessing historical data sequences of load, renewable energy, and cascade hydropower stations in the target grid area; constructing a net load sequence; aggregating each time series across multiple time scales; calculating coupling indices between the output of cascade hydropower stations and total load, net load, and renewable energy output, and constructing a multidimensional coupling index tensor; calculating the structural change intensity of the multidimensional coupling index tensor, setting thresholds to identify operation segment boundaries, and identifying candidate segment points; forming a set of operation segment boundaries; calculating the response characteristics within each operation segment and outputting explanatory labels. This invention can uniformly quantify the coupling relationship between the output of cascade hydropower stations and grid load and renewable energy output under different time scales, load definitions, spatial regions, and grid levels, and automatically identify operation stage boundaries.

Description

Technical Field

[0001] This invention belongs to the technical field of hydropower station operation characteristics, specifically, it relates to a method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards changes in grid load and new energy sources. Background Technology

[0002] With the large-scale integration of new energy sources and the continuous changes in load-side electricity consumption patterns, the source-load balance relationship of power grid operation is gradually shifting from the traditional load-dominated and conventional power source-following model to a complex form involving the coupling of multiple factors such as load, new energy sources, inter-regional power reception, and regulation resources. Cascade hydropower stations, possessing the capabilities of power supply, capacity support, peak shaving, frequency regulation, and inter-regional mutual assistance, are important flexible regulation resources in the new power system. Their output changes are intricately coupled with the total load, net load, and new energy output of the power grid.

[0003] In existing technologies, the analysis of the relationship between grid load, renewable energy changes, and hydropower output typically employs correlation coefficient calculations within a fixed time window, linear regression, empirical rule judgment, or scheduling optimization models. Some technical solutions focus on short-term joint optimization scheduling of hydropower, wind power, and solar power under conditions of uncertainty in renewable energy output, while others focus on multi-timescale coordinated scheduling, load forecasting, or optimization of cascade hydropower station generation plans. These methods can support power generation planning and renewable energy consumption to some extent, but their technical objectives are mostly concentrated on generating scheduling plans or improving forecast accuracy, with insufficient attention paid to the spatiotemporal effect analysis of the relationship between cascade hydropower stations and grid response under the influence of grid load and renewable energy changes, the identification of abrupt changes in the correlation structure, and the output of operational segment boundaries.

[0004] The existing technology has the following drawbacks: (1) Correlation analysis is often carried out using a single time scale or a fixed window, which makes it difficult to reflect the changes in the coupling strength between hydropower output and load and new energy sources at different scales such as hour, day, ten days, and month, resulting in difficulty in unifying short-term and medium- to long-term conclusions.

[0005] (2) The total load is often used as a single explanatory variable. The different effects of total load, net load and new energy output on hydropower regulation demand are not fully distinguished, making it difficult to determine whether the fluctuation of new energy has changed the actual response object of hydropower.

[0006] (3) The analysis focuses on the overall region or a single receiving end region, lacks a comparison mechanism between spatial regions and power grid levels, and is difficult to explain the differences in the guiding effect of load signals on the output of cascade hydropower in different regions.

[0007] (4) Traditional correlation analysis or linear regression can only output static correlation coefficients or average response relationships. It usually relies on manual experience to divide the operation stages such as dry season, flood season, and water storage period. It cannot automatically identify the boundary of the operation stage based on the structural changes in the coupling relationship between load, new energy and hydropower output. It is difficult to characterize the sudden change in response relationship caused by the combined effects of water inflow, water level, flood control, channel, and planning constraints.

[0008] (5) The analysis results are usually limited to the level of chart display, lacking segmented boundaries, explanatory labels and quantitative analysis outputs that can be directly used for scheduling review, setting planning boundaries and switching operating strategies. Summary of the Invention

[0009] The technical problem to be solved by this invention is to provide a method for identifying the source-load response and segmenting the operation of cascade hydropower stations in response to changes in grid load and new energy sources. This method can uniformly quantify the coupling relationship between the output of cascade hydropower stations and grid load and new energy output under different time scales, different load calibers, different spatial regions and different grid levels, and automatically identify the boundaries of the operation stages where the relevant relationships change significantly.

[0010] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, comprising the following steps: S1. Obtain historical load data of the target power grid area, historical actual output data of new energy sources, and historical output data of cascade hydropower stations, and preprocess the acquired data. S2. Construct a net load sequence based on the total load and renewable energy output of the target power grid area; S3. Aggregate the preprocessed time series at multiple time scales to form synchronous samples at different time scales. S4. Calculate the coupling index between the output of cascade hydropower stations and the total load, net load and new energy output, and perform standardized transformation on the coupling index to construct a multidimensional coupling index tensor. S5. Calculate the structural change intensity of the multidimensional coupling index tensor within the sliding window, and set an adaptive threshold to identify the running segment boundary. When the structural change intensity at a certain moment exceeds the adaptive threshold, that moment is marked as a candidate segment point. S6. Perform minimum segment length constraints, merge adjacent breakpoints, and verify business rules on the identified candidate segment points to form a set of running segment boundaries. Calculate the response features within each running segment and output interpretation labels.

[0011] In a preferred embodiment, step S1 involves preprocessing the acquired data, including timestamp alignment, outlier handling, missing value imputation, and standardization, and then normalizing each time series after processing.

[0012] In the preferred embodiment, the specific operation of constructing the net load sequence in step S2 is as follows: For any spatial region p and any time scale t The net load calculation formula is as follows: ; In the formula: For the region p In time scale t Down t Net load at any given time; This corresponds to the total load; To contribute to the development of new energy sources.

[0013] In a preferred embodiment, the time scale in step S3 includes hours, days, ten-day periods, months, or user-defined scales.

[0014] In a preferred embodiment, in step S3, for any sequence On a time scale t The aggregation results are as follows: ; In the formula, Ag τ (·) represents the time scale. t The aggregation operator below, The number of time scales.

[0015] In a preferred embodiment, step S4 uses the Pearson correlation coefficient or regression explanatory power. R 2 Or Spearman correlation coefficient r Mutual information can be used as a coupling indicator.

[0016] In the preferred embodiment, the regression explanatory power R 2 The calculation formula is: ; In the formula, For the first Real-time water and electricity values; For the first Real-time load value, net load, or renewable energy output; This is the average load value; The number of samples; For the sake of explanation; The above formula is used to calculate the regression explanatory power of the output of cascade hydropower stations with total load, net load and new energy output respectively.

[0017] In a preferred embodiment, the coupling index is standardized using the Fisher Z-transform.

[0018] In the preferred embodiment, the expression for constructing the multidimensional coupling index tensor is: ; In the formula: p For spatial regions; t Time scale; k To determine the appropriate range, the total load can be used. or net load Or new energy sources ; g For the power grid level; M is the standardized coupling index; Indicates a specific spatial region p Time scale t Analytical caliber k and power grid level g The coupling index under; This represents the standardized function.

[0019] In the preferred embodiment, in step S5, the window length is set to... W The drift intensity between adjacent windows is calculated as the structural change intensity of the multidimensional coupling index tensor by using statistical distance, mean difference, variance difference, maximum mean difference, KL divergence, or change point detection methods.

[0020] In the preferred embodiment, the triggering rule for the candidate segmentation point in step S5 is as follows: ; In the formula: I (·) is the indicator function; i t for t Time-adaptive threshold; for t Intensity of structural change at any given time; When Trigger t When =1, that moment is identified as a candidate segmentation point.

[0021] In the preferred embodiment, when the identified candidate segmentation points are checked for business rules in step S6, the business rules include water inflow conditions, water level constraints, flood control scheduling rules, water storage plans, and channel constraints.

[0022] In the preferred embodiment, in step S6, the set of running segment boundaries is formed as follows: ,in If the segmentation point is , then the th m Each running segment is represented as: ; In the formula, For the first m Each running segment.

[0023] In the preferred embodiment, the intra-segment response characteristics include the intra-segment average correlation coefficient, correlation trend slope, fluctuation stability, net load dominance, spatial variability, and new energy impact.

[0024] In a preferred embodiment, the interpretation labels are output based on the response characteristics within a segment. The interpretation labels include the net load-dominated segment, the total load-dominated segment, the renewable energy disturbance enhancement segment, the hydrological constraint decoupling segment, and the high water level precise following segment.

[0025] In a preferred embodiment, the correspondence between the explanation labels and response features is as follows: ①If the average coupling index between net load and the output of cascade hydropower stations is not less than the set strong coupling threshold, and the net load dominance is not less than the net load dominance threshold, the net load dominant section is output. ②If the average coupling index between the total load and the output of the cascade hydropower stations is not less than the set strong coupling threshold, and the net load dominance is not less than the net load dominance threshold, the total load dominance section is output. ③ If the impact of new energy is not less than the impact threshold of new energy, and the fluctuation stability is less than the stability threshold or the slope of the relevant trend is greater than the trend slope stability threshold, output the new energy disturbance enhancement segment. ④ If the maximum value of the average coupling index between net load and cascade hydropower station output, or the average coupling index between total load and cascade hydropower station output is less than the weak coupling threshold, or the fluctuation stability is less than the stability threshold, and the business rule verification shows that there is at least one of the following in this segment: change in inflow conditions, water level constraints, flood control scheduling rules, water storage plan, or channel constraints, output the hydrological constraint decoupling segment. ⑤ If the maximum value of the average coupling index between net load and cascade hydropower station output, and the average coupling index between total load and cascade hydropower station output is not less than the strong coupling threshold, and the fluctuation stability is not less than the stability threshold, and the relevant trend slope is less than or equal to the trend slope stability threshold, output the high water level precise following segment.

[0026] The present invention provides a method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, which has the following beneficial effects: 1. By performing unified time aggregation and normalization processing on load, new energy output and cascade hydropower output, the coupling relationship at different scales such as hour, day, ten days and month is made comparable, avoiding misjudgment caused by different sampling frequency and sample length.

[0027] 2. By setting up multiple metrics such as total load, net load, and renewable energy output, it is possible to identify whether the output of cascade hydropower stations mainly responds to changes in total load and net load, or exhibits compensation, avoidance, and weak coupling relationships with fluctuations in renewable energy output. This clarifies the actual regulation targets and response characteristics undertaken by hydropower after a high proportion of renewable energy is connected.

[0028] 3. By introducing spatial region and power grid level dimensions, and by performing synchronous calculations on different regions, combinations of regions and the overall power grid level, the system can output the ranking of regional differences, scale convergence or divergence characteristics, and the reasons for differences between the overall level and the local level, thus avoiding the need to guide overall scheduling based solely on conclusions from a single region.

[0029] 4. The invention employs correlation drift detection to identify operational segment boundaries. Traditional methods typically use fixed windows or manual division of dry seasons, flood seasons, and water storage periods. This invention, however, automatically determines candidate breakpoints based on the structural changes of the coupling index tensor itself. This is then combined with minimum segment length constraints, merging of adjacent breakpoints, and verification using business rules to form stable segments. This reduces reliance on manual experience in defining operational periods and can reflect abrupt changes in response relationships under the combined effects of changes in energy structure, load path, enhanced new energy fluctuations, and changes in hydrological constraints.

[0030] 5. Transform the segmented results into interpretable engineering outputs. Each operational segment not only provides the start and end times, but also simultaneously provides the average correlation, trend slope, stability score, net load dominance, spatial variability, and explanatory labels within the segment. This enables the analysis results to directly serve scheduling review, strategy switching, and planning boundary setting.

[0031] 6. The correlation metrics, drift detection methods, and threshold determination methods in this invention can all be replaced. The correlation metrics can be Pearson, Spearman, mutual information, maximum information coefficient, or regression explanatory power; drift detection can be mean difference, statistical distance, change point detection, or machine learning methods; the threshold can be an empirical threshold, quantile threshold, or adaptive threshold. These substitutions do not affect the core technical concept of this invention. Attached Figure Description

[0032] The accompanying drawings, which are provided to further illustrate the invention and constitute a part of this invention, do not constitute an undue limitation thereof. In the drawings: Figure 1 This is a flowchart of the present invention; Figure 2 The results show the comparative explanatory power of total load and net load on the output of cascade hydropower stations at different time scales; Figure 3 The results of operational phase identification of the relationship between power output of a hydropower station and grid load response. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention.

[0034] Example 1: This invention provides a method for source-load response identification and operation segmentation of cascade hydropower stations in response to changes in grid load and renewable energy. Based on the data of grid load at the receiving end of the hydropower station, renewable energy output, and output of the cascade hydropower stations, a multi-time-scale synchronous sample is constructed to calculate the multi-dimensional coupling index between total load, net load, renewable energy output, and cascade hydropower output. Furthermore, through correlation drift detection and segmentation identification methods, the segmentation boundaries, intra-segment response characteristics, and explanatory labels of different operating stages are output, thereby providing technical support for source-grid coordinated scheduling, regulation capacity evaluation, and operation strategy review of cascade hydropower stations.

[0035] This invention can uniformly quantify the coupling relationship between the output of cascade hydropower stations and the power grid load and the output of new energy sources under different time scales, different load calibers, different spatial regions and different power grid levels, and automatically identify the boundaries of the operation stage where the relevant relationship changes significantly.

[0036] like Figure 1 As shown, the specific steps include: (1) Data preparation Historical load data of the target power grid area, historical actual output data of new energy sources, and historical output data of cascade hydropower stations are obtained, and the obtained data are preprocessed.

[0037] (2) Preprocessing of the acquired data includes: timestamp alignment, outlier handling, missing value filling and standardization of load, new energy output and cascade hydropower output data, and normalization of each time series after processing.

[0038] For any original time series x ( t The Max-Min linear normalization process is used, and the calculation formula is as follows: ; In the formula: x i For the first i The raw data at any given moment; x max and x min These are the maximum and minimum values ​​of the time series, respectively. x i * For the first iData after time-normalization.

[0039] (3) Construct the net load sequence of the target power grid area The net load sequence is constructed based on the total load and renewable energy output of the target power grid area. The specific operation is as follows: For any spatial region p and any time scale t The net load calculation formula is as follows: ; In the formula: For the region p In time scale t Down t Net load at any given time; This corresponds to the total load; To contribute to the development of new energy sources.

[0040] (4) Perform multi-timescale aggregation on the original time series. The preprocessed time series are aggregated across multiple time scales to form synchronized samples at different time scales. Specifically, synchronized samples are generated at hourly, daily, ten-day, monthly, or user-defined scales.

[0041] For any sequence On a time scale t The aggregation results are as follows: ; In the formula, Ag τ (·) represents the time scale. t The aggregation operators can be determined based on business metrics such as average power and cumulative power consumption. Time scale t Next k The number of samples contained in each aggregation window.

[0042] If the average power value is used as the business metric, then: ; If the cumulative electricity consumption value is used as the business criterion, then: ; In the formula, This represents the time interval for the original data.

[0043] (5) Calculate the coupling index between the output of cascade hydropower stations and the total load, net load and new energy output.

[0044] Using Pearson correlation coefficient or regression explanatory power R 2 Or Spearman correlation coefficientr Mutual information can be used as a coupling indicator.

[0045] Taking the Pearson correlation coefficient as an example, the calculation formula is as follows: ; In the formula: X and Y are any two time series to be analyzed; r This is the correlation coefficient between the two.

[0046] The above formula is used to calculate the correlation coefficients between the output of the cascade hydropower stations and the total load. Correlation coefficient between output and net load of cascade hydropower stations Correlation coefficient between power output of cascade hydropower stations and power output of new energy sources .

[0047] Indicated on the time scale The power output sequence of the cascade hydropower stations below; Indicates a spatial region p and time scale t The total load sequence below; Indicates a spatial region p and time scale t The net load sequence below; Indicates a spatial region p and time scale t The following is a sequence of new energy power outputs.

[0048] The explanatory power of regression R 2 The calculation formula is: ; In the formula, For the first Real-time water and electricity values; For the first Real-time grid-related variable values, selectable from load value, net load, or renewable energy output; The mean of the selected power grid-related variables; The number of samples; For the sake of explanation; Similarly, the above formula is used to calculate the regression explanatory power of the output of cascade hydropower stations with total load, net load, and new energy output.

[0049] (6) Standardize the coupling index and construct a multidimensional coupling index tensor.

[0050] To enhance the comparability of relevant indicators under different sample lengths and time scales, the correlation coefficients are standardized. Specifically, the Fisher Z-transform is used to standardize the coupling indicators, and the calculation formula is as follows: ; In the formula: z( r () is the correlation coefficient r The Fisher Z-transform results, Other coupling metrics can be used instead.

[0051] The expression for constructing the multidimensional coupling index tensor is: ; In the formula: p For spatial regions; t Time scale; k To determine the appropriate range, the total load can be used. or net load Or new energy sources ; g For the power grid level; M is the standardized coupling index; Indicates a specific spatial region p Time scale t Analytical caliber k and power grid level g The coupling index under; This represents the standardized function.

[0052] The physical structure of the index is shown in Table 1.

[0053]

[0054] (7) Calculate the structural change intensity of the multidimensional coupling index tensor within the sliding window. Let the window length be W The drift intensity between adjacent windows is calculated as the structural change intensity of the multidimensional coupling index tensor using statistical distance, mean difference, variance difference, maximum mean difference, KL divergence, or change point detection methods. The calculation formula is as follows: ; In the formula: for t The intensity of structural change at time t; Dist(·) is the statistical distance function; and These are sets of multidimensional coupling indicators for the two windows, one before and one after.

[0055] (8) Set adaptive threshold to identify the segment boundary of the operation.

[0056] The threshold can be determined by using historical quantiles, mean plus standard deviation, exponentially weighted moving average, or manually set methods.

[0057] When the intensity of structural change exceeds an adaptive threshold at a certain moment, that moment is marked as a candidate segmentation point. The triggering rule for candidate segmentation points is as follows: ; In the formula: I (·) is the indicator function; i t for t Time-adaptive threshold; for t Intensity of structural change at any given time; When Trigger t When =1, that moment is identified as a candidate segmentation point.

[0058] (9) Perform minimum segment length constraints, merge adjacent breakpoints and verify business rules on the identified candidate segment points to form a set of running segment boundaries. Calculate the response features within each running segment and output the interpretation label.

[0059] Let the set of running segment boundaries be ,in If the segmentation point is , then the th m Each running segment is represented as: ; In the formula, For the first m Each running segment.

[0060] 1) Minimum segment length constraint: Let the set of candidate segmentation points be: ; Let the minimum segment length be This value can be determined based on the time scale of the analysis. For example, in daily-scale analysis, it can be set to no less than a certain number of days; in ten-day-scale analysis, it can be set to no less than one or two ten-day periods; it can also be set to no less than the sliding window length. W .

[0061] For any adjacent candidate points and ,like Therefore, no separate operational segment is formed, and the intensity of structural changes is preserved within this range. The largest candidate segment point is selected, and the remaining candidate segment points are eliminated.

[0062] 2) Merging adjacent breakpoints: Set the breakpoint merging window as For multiple candidate segmentation points, if their time positions are too close, i.e., satisfying the condition... If the breakpoints are considered as a cluster of candidate breakpoints caused by the same change in response relationship, then the moment with the greatest structural change intensity within the cluster is taken as the final candidate breakpoint.b .

[0063] 3) Business rule verification: The purpose of business rule verification is not to recalculate coupling indicators, but to determine whether candidate segmentation points are consistent with changes in hydropower operation boundaries or power grid constraints, so as to avoid incorrect segmentation caused by outliers, missing measurements or short-term noise.

[0064] For candidate segmentation points after minimum segment length constraints and merging of adjacent breakpoints, further business rule verification is performed. (Based on candidate segmentation points...) b Centered on this, construct front and back verification windows respectively. and The system compares whether there are significant changes in inflow conditions, water level constraints, flood control scheduling rules, water storage plans, and channel constraints within the preceding and following windows. If there are changes in the structure of coupled indicators and at least one change in business boundaries near a candidate segment point, the candidate segment point is retained; if the candidate segment point is caused only by outliers, missing values, or isolated fluctuations, and does not correspond to any change in business boundaries, the confidence level of the candidate segment point is reduced or it is removed.

[0065] Business rule verification specifically includes: ① Verification of inflow conditions: Compare the magnitude and trend of the average inflow rate before and after the candidate segment point to see if they have changed; ② Water level constraint verification: Determine whether there are water levels near the candidate segment points that are close to the control water level, flood limit water level, water storage target water level, or water level fluctuation constraint changes; ③ Flood control scheduling rule verification: Determine whether the candidate segment points are in the period of flood control scheduling rule switching, flood limit water level control, flood control reservoir capacity reservation, or flood control constraint change; ④ Water storage plan verification: Determine whether there are any water storage plans initiated, water storage targets adjusted, or water storage progress constraints changed near the candidate segment points; ⑤ Channel constraint verification: Determine whether there are changes in transmission channel capacity, adjustments to the transmission plan, or channel restrictions near the candidate segment point.

[0066] 4) Calculate the response characteristics within each running segment and output interpretation labels: The intra-segment response characteristics include the intra-segment average correlation coefficient, correlation trend slope, fluctuation stability, net load dominance, spatial variability, and the impact of new energy sources. The specific calculation methods are as follows: ① Average correlation coefficient within the segment For analysis caliber ,in The total load, net load, or renewable energy output can be used, and the average correlation coefficient within the segment can be calculated directly within that segment. ; In the formula, Time scale The power output sequence of the cascade hydropower stations below, This is the input sequence for the corresponding caliber.

[0067] ② Slope of the correlation trend The slope of the correlation trend is used to indicate whether the coupling relationship within the segment is strengthening, weakening, or remaining stable. Coupling indices within the segment can be analyzed. Regarding time t Perform a linear fit, with the slope as: ; In the formula, For the first m Analysis caliber within each segment The slope of the relevant trend; This represents the average time sequence number within this segment.

[0068] ③ Fluctuation stability Fluctuation stability is used to indicate whether the coupling relationship within a segment is stable, and is calculated based on the standard deviation of the intra-segment coupling index: ; in: ; In the formula, This represents the standard deviation of the coupling index fluctuation within the segment. The smaller this value, the more stable the response relationship within the segment. The closer it is to 1, the higher the stability.

[0069] ④ Net load advantage Net load advantage is used to compare the interpretive advantage of net load versus total load: ; In the formula, This indicates that the net load metric has a stronger explanatory power for hydropower output than the total load metric.

[0070] ⑤ Spatial Difference Spatial variability is used to represent the difference in the guiding effect of load signals on the output of cascade hydropower stations in different spatial regions. The calculation formula is as follows: ; In the formula, For the first m Spatial regions within each segment p In the analysis caliber The average coupling index under the given conditions. The larger the value, the more pronounced the spatial regional differences.

[0071] ⑥ Impact of New Energy The impact of renewable energy sources is used to represent the degree to which changes in renewable energy output affect the hydropower response. The calculation formula is: ; In the formula, For the first m The average coupling index between the output of new energy sources and the output of cascade hydropower stations within each segment is used to represent the intensity of the impact of new energy sources by taking the absolute value.

[0072] The net load advantage can also be used to characterize the change in the regulation of demand by new energy sources: ; When this value is large, it indicates that the relationship between the net load formed after deducting new energy sources and the total load has changed significantly in terms of the interpretation of hydropower output, and that the output of new energy sources has a significant impact on the demand for hydropower regulation.

[0073] Set thresholds for each explanatory label. These thresholds can be derived from historical sample quantiles, empirical thresholds, or user-defined thresholds. The thresholds are defined as follows: Net Load Advantage Threshold. Total load advantage threshold Threshold for the impact of new energy Stability threshold Spatial variability threshold Trend slope stability threshold This invention uses the 70th percentile of historical samples as the strong coupling threshold. 30% quantile as a weak coupling threshold .

[0074] Interpretation labels are output based on the response characteristics within a segment. These labels include those for the net load-dominated segment, the total load-dominated segment, the renewable energy disturbance-enhanced segment, the hydrological constraint decoupling segment, and the precise following segment. The correspondence between each interpretation label and the response characteristics is as follows: ①If and This indicates that the net load is strongly coupled with the hydropower output, and the net load specification is significantly better than the total load specification, thus outputting the "net load dominant section".

[0075] ②If and This indicates that the total load is strongly coupled with the hydropower output, and the explanatory power of the total load caliber is higher than that of the net load caliber, thus outputting the "total load dominant section".

[0076] ③If satisfied or At the same time satisfy or This indicates that changes in renewable energy output significantly disrupt the hydropower response relationship, leading to a widening difference in the explanatory power between the total load and net load metrics, resulting in the output of the "renewable energy disturbance enhancement segment". ④ If satisfied or Meanwhile, the business rule verification shows that there is at least one of the following in this segment: changes in inflow conditions, water level constraints, flood control scheduling rules, water storage plans, or channel constraints. This indicates that the hydropower output is no longer mainly explained by changes in load or net load, but is dominated by hydrological conditions or scheduling boundaries, and outputs "hydrological constraint decoupling segment".

[0077] ⑤ If satisfied and ,at the same time This means that under favorable water level conditions, strong regulation capacity, and stable source-load coupling relationship, cascade hydropower stations can reliably follow changes in the target power grid demand and output a "precise following segment".

[0078] (10) Generate engineering output based on the segmentation results. The output includes source-load coupling indices at each time scale, comparison of explanatory power of different load calibers, differences in different regions or power grid levels, segment boundaries, intra-segment response characteristics, explanatory labels, threshold trajectories, and optional strategy suggestions. The output can be used for scheduling review, operation strategy switching, regulation capacity evaluation, planning boundary setting, and source-grid coordination analysis.

[0079] Example 2: To verify the effectiveness of the proposed method for source-load response identification and operation segmentation of cascade hydropower stations, historical data on the target power grid area load, renewable energy output, and cascade hydropower station output were selected for case verification. The results are as follows: Figure 2 and Figure 3 As shown.

[0080] exist Figure 2 The results show that the total load of the target power grid area and the net load after deducting renewable energy output were used as input variables to calculate their explanatory power for the output of cascade hydropower stations at different time scales. It can be seen that as the time scale increases from short to long periods, the explanatory power of both total load and net load for the output of cascade hydropower stations increases. This indicates that the present invention can identify the time scale effect between hydropower output and grid demand, avoiding the problem of unstable correlation judgments caused by using only a single short-term scale analysis.

[0081] at the same time, Figure 2The explanatory power of net load on the output of cascade hydropower stations is generally better than that of total load, and this advantage is more pronounced over longer time scales. This result indicates that, under conditions of increased fluctuations in renewable energy output, using only the total load metric can easily mask the actual regulatory demand of hydropower responses. This invention, by simultaneously constructing both total load and net load metrics and comparing them across multiple time scales, can identify the impact of renewable energy changes on hydropower regulation positioning, and more accurately reflect the response relationship of cascade hydropower stations to the actual net demand of the target power grid area.

[0082] Furthermore, based on the changes in multidimensional coupling indices, the operational phases of the output response relationship of cascade hydropower stations were identified, and the results are as follows: Figure 3 As shown, the results indicate that the response relationship between cascade hydropower stations and grid load is not fixed throughout the year, but exhibits phased differences with changes in inflow, water level, flood control, water storage, and grid demand. Different operational segments can exhibit characteristics such as rigid base load or weak response during the dry season, marginal decrease or threshold response during the pre-flood drawdown period, strong following response during the peak period, and weakened response under multi-objective constraints during the end-of-flood water storage period. Therefore, this invention can automatically identify operational segment boundaries based on changes in the coupling index structure and assign physically meaningful explanatory labels to different segments, solving the problem that traditional fixed-window methods struggle to identify abrupt changes in response relationships.

[0083] In summary, the following effects can be achieved after the implementation of this invention: (1) It can uniformly quantify the coupling strength between the output of cascade hydropower stations and the total load, net load and new energy output of the target power grid area under different time scales, and solve the problems of unstable conclusions and difficulty in comparison between scales in traditional single time scale analysis. (2) It can simultaneously compare the explanatory power of total load and net load on the output of cascade hydropower stations, identify the impact of changes in new energy output on the actual regulation demand of hydropower, and solve the problem that the traditional total load caliber is difficult to reflect the changes in net demand under the background of new energy fluctuations. (3) It can identify the stage-by-stage abrupt changes in the source-load response relationship of cascade hydropower stations based on the structural changes of multidimensional coupling indicators, and output the operation segment boundary, intra-segment response characteristics and interpretation labels, thus solving the problem that traditional fixed window analysis is difficult to characterize changes in operation status. (4) It can transform the complex relationship between load, new energy and hydropower output into calculable, comparable and interpretable source-load coupling indicators and operation segmentation results, providing technical support for cascade hydropower stations to participate in the regulation of new power systems, grid demand response analysis, dispatch review and planning boundary setting.

[0084] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, characterized in that, Includes the following steps: S1. Obtain historical load data of the target power grid area, historical actual output data of new energy sources, and historical output data of cascade hydropower stations, and preprocess the acquired data. S2. Construct a net load sequence based on the total load and renewable energy output of the target power grid area; S3. Aggregate the preprocessed time series at multiple time scales to form synchronous samples at different time scales. S4. Calculate the coupling index between the output of cascade hydropower stations and the total load, net load and new energy output, and perform standardized transformation on the coupling index to construct a multidimensional coupling index tensor. S5. Calculate the structural change intensity of the multidimensional coupling index tensor within the sliding window, and set an adaptive threshold to identify the running segment boundary. When the structural change intensity at a certain moment exceeds the adaptive threshold, that moment is marked as a candidate segment point. S6. Perform minimum segment length constraints, merge adjacent breakpoints, and verify business rules on the identified candidate segment points to form a set of running segment boundaries. Calculate the response features within each running segment and output interpretation labels.

2. The method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, as described in claim 1, is characterized in that... In step S1, the preprocessing of the acquired data includes: timestamp alignment, outlier handling, missing value imputation, and standardization, and the normalization of each time series after processing.

3. The method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, as described in claim 1, is characterized in that... In step S2, the specific operation for constructing the net load sequence is as follows: For any spatial region p and any time scale τ The net load calculation formula is as follows: ； In the formula: For the region p In time scale τ Down t Net load at any given time; This corresponds to the total load; To contribute to the development of new energy sources.

4. The method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, as described in claim 1, is characterized in that... In step S3, the time scale includes hours, days, ten-day periods, months, or user-defined scales.

5. The method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, as described in claim 1, is characterized in that... In step S3, for any sequence On a time scale τ The aggregation results are as follows: ； In the formula, Agg τ (·) represents the time scale. τ The aggregation operator below, The number of time scales.

6. The method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, as described in claim 1, is characterized in that... In step S4, Pearson correlation coefficient or regression explanatory power is used. R 2 Or Spearman correlation coefficient r Mutual information can be used as a coupling indicator.

7. The method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, as described in claim 6, is characterized in that... The explanatory power of regression R 2 The calculation formula is: ； In the formula, For the first Real-time water and electricity values; For the first Real-time grid-related variable values, selectable from load value, net load, or renewable energy output; The mean of the selected power grid-related variables; The number of samples; For the sake of explanation; The above formula is used to calculate the regression explanatory power of the output of cascade hydropower stations with total load, net load and new energy output respectively.

8. The method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, as described in claim 6, is characterized in that... The coupling index is standardized using the Fisher Z-transform.

9. A method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, as described in claim 6, is characterized in that... The expression for constructing the multidimensional coupling index tensor is: ； In the formula: p For spatial regions; τ Time scale; κ To determine the appropriate range, the total load can be used. or net load Or new energy sources ; g For the power grid level; M is the standardized coupling index; Indicates a specific spatial region p Time scale τ Analytical caliber κ and power grid level g The coupling index under; This represents the standardized function.

10. The method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, as described in claim 1, is characterized in that... In step S5, the window length is set to... W The drift intensity between adjacent windows is calculated as the structural change intensity of the multidimensional coupling index tensor by using statistical distance, mean difference, variance difference, maximum mean difference, KL divergence, or change point detection methods.

11. The method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, as described in claim 1, is characterized in that... In step S5, the triggering rule for candidate segmentation points is as follows: ； In the formula: I (·) is the indicator function; θ t for t Time-adaptive threshold; for t Intensity of structural change at any given time; When Trigger t When =1, that moment is identified as a candidate segmentation point.

12. The method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, as described in claim 1, is characterized in that... In step S6, when the identified candidate segmentation points are checked against business rules, the business rules include water inflow conditions, water level constraints, flood control scheduling rules, water storage plans, and channel constraints.

13. The method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, as described in claim 1, is characterized in that... In step S6, the set of running segment boundaries is formed as follows: ,in If the segmentation point is , then the th m Each running segment is represented as: ； In the formula, For the first m Each running segment.

14. The method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, as described in claim 13, is characterized in that... The intra-segment response characteristics include the intra-segment average correlation coefficient, correlation trend slope, fluctuation stability, net load dominance, spatial variability, and new energy impact.

15. The method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, as described in claim 14, is characterized in that... The interpretation labels are output based on the response characteristics within the segment. The interpretation labels include the net load-dominated segment, the total load-dominated segment, the renewable energy disturbance enhancement segment, the hydrological constraint decoupling segment, and the high water level precise following segment.

16. The method for source-load response identification and operation segmentation of cascade hydropower stations oriented towards grid load and new energy changes, as described in claim 15, is characterized in that... The correspondence between the explanatory labels and response features is as follows: ①If the average coupling index between net load and the output of cascade hydropower stations is not less than the set strong coupling threshold, and the net load dominance is not less than the net load dominance threshold, the net load dominant section is output. ②If the average coupling index between the total load and the output of the cascade hydropower stations is not less than the set strong coupling threshold, and the net load dominance is not less than the net load dominance threshold, the total load dominance section is output. ③ If the impact of new energy is not less than the impact threshold of new energy, and the fluctuation stability is less than the stability threshold or the slope of the relevant trend is greater than the trend slope stability threshold, output the new energy disturbance enhancement segment. ④ If the maximum value of the average coupling index between net load and cascade hydropower station output, or the average coupling index between total load and cascade hydropower station output is less than the weak coupling threshold, or the fluctuation stability is less than the stability threshold, and the business rule verification shows that there is at least one of the following in this segment: change in inflow conditions, water level constraints, flood control scheduling rules, water storage plan, or channel constraints, output the hydrological constraint decoupling segment. ⑤ If the maximum value of the average coupling index between net load and cascade hydropower station output, and the average coupling index between total load and cascade hydropower station output is not less than the strong coupling threshold, and the fluctuation stability is not less than the stability threshold, and the relevant trend slope is less than or equal to the trend slope stability threshold, output the high water level precise following segment.

Images