Risk scheduling method for water-wind-solar complementary system

Abstract

The invention discloses a risk scheduling method for a water-wind-solar complementary system, and the method comprises the steps: collecting historical data and power grid topological parameters, accessing global and regional numerical weather forecast data, employing a coupling model, fusing meteorological grid data with a historical power station output sequence, and generating hourly reservoir incoming water amount and wind-solar power probability prediction results. According to the predicted time sequence and the generated scene, calculating the scene probability based on the generated scene; according to the prediction time sequence, using a quantification method to obtain a peak regulation risk quantification value; calculating power grid power flow distribution and critical clearing time according to the generation scene and the power grid topological parameters, and calculating a system stability margin; a multi-target optimization model is constructed according to a peak regulation risk quantized value after splitting and a system stability margin, the system stability margin is introduced as an optimization target, the overall stability of the system is improved, a meteorological-hydrological-output three-mode feature mapping method is provided, and meteorological feature extraction of a key grid region is enhanced through an attention mechanism.

Description

Technical Field

[0001] The present invention relates to the technical field of power system dispatching, and in particular to a risk dispatching method for a hydro-wind-solar complementary system. Background Art

[0002] With the transformation of the global energy structure and the deepening of sustainable development concepts, the proportion of renewable energy in power systems is increasing. The hydro-wind-solar hybrid system, which integrates hydropower, wind power, and photovoltaic power generation, has attracted widespread attention because it can fully utilize the advantages of different energy sources to achieve complementary and optimized energy allocation.

[0003] For example, Chinese patent application No. 202311468432.0 discloses a risk scheduling method, device, equipment and medium for a hydro-wind-solar complementary system. The present invention utilizes a historical output data set and a time series prediction method for prediction. On this basis, the multiple uncertainty factors of the hydro-wind-solar complementary system participating in the peak regulation of the power grid are analyzed, and the peak regulation risk is quantified using a risk quantification method based on the upper partial moment index. A hydro-wind-solar multi-energy complementary joint peak regulation model is constructed with the goal of minimizing the sum of the expected value of the peak-to-valley difference of the residual load and the peak regulation risk.

[0004] Among existing patented technologies, most hydro, wind, and photovoltaic power stations currently adopt an independent scheduling mode, resulting in the failure to fully utilize resource complementarity and posing challenges to system flexibility and reliability. Based on historical hydrological data and short-term weather forecasts, a deterministic scheduling model is used to formulate power generation plans, relying on the constraints of reservoir scheduling regulations, using single numerical weather forecast data, combined with statistical methods or machine learning models to predict power generation. The scheduling strategy aims to maximize absorption, and static weight allocation or simple priority strategies are used in local complementary scenarios, without considering the multi-time scale coupling characteristics. Summary of the Invention

[0005] This application provides a risk scheduling method for a hydro-wind-solar complementary system. By introducing the system stability margin as an optimization target, the anti-disturbance capability of the power grid is quantified, the overall stability of the system is improved, and the risk of system collapse caused by improper peak regulation is reduced.

[0006] This application provides a risk scheduling method for a hydro-wind-solar complementary system, including: S11, collecting historical data and grid topology parameters, predicting time series and generating scenarios based on the collected historical data, and calculating scenario probabilities based on the generated scenarios; S12, using a quantitative method to obtain a peak-shaving risk quantitative value based on the predicted time series, and splitting the peak-shaving risk quantitative value into risk event frequency and average risk intensity; S13, calculating the power flow distribution and critical removal time of the power grid according to the generated scenario and the power grid topology parameters, and calculating the system stability margin based on the power flow distribution and the critical removal time; S14, construct a multi-objective optimization model based on the peak-shaving risk quantification value after splitting and the system stability margin.

[0007] Preferably, the formula for calculating the scene probability is: ,in, represents the probability of scenario S, P(W) represents the probability of weather type related to scenario S, P(L) represents the probability of load demand condition related to scenario S, and P(M) represents the probability of wind and solar output condition related to scenario S.

[0008] Preferably, the system stability margin is calculated using the formula: ,in, is the system stability margin, is the pre-set critical resection time threshold, It indicates the longest fault clearing time during which the system can maintain stable operation under actual operating conditions.

[0009] Preferably, S201, acquiring load forecast data according to a forecast time series, identifying high-risk periods according to the load forecast data, generating a time accumulation scenario, and calculating a time accumulation risk value based on the time accumulation scenario; S202, generating a spatial propagation scenario according to the power grid topology parameters, and calculating a spatial propagation risk value based on the spatial propagation scenario; S203, calculating a composite risk value and a system stability margin correction value based on the time-accumulated risk value and the spatial propagation risk value; S204, dynamically adjusting the multi-objective optimization model.

[0010] Preferably, the steps for forming high-risk periods are: dividing the load forecast data into continuous time periods, and for each time period, calculating the peak-to-valley difference of the remaining load in the time period, that is, the difference between the maximum and minimum values ​​of the load in the time period, traversing all continuous time period combinations, and identifying time periods in which the peak-to-valley difference of the remaining load for multiple consecutive time periods exceeds the threshold. These time periods are high-risk periods.

[0011] Preferably, the time cumulative risk value is calculated using the formula: ,in, represents the part exceeding the threshold value during the tth period, i.e. >r, = -r; when When ≤r, =0, λ is the time decay factor, and T is the number of time periods the scene lasts.

[0012] Preferably, the formula for correcting the system stability margin is: ,in, is the correction value of the system stability margin, Time cumulative risk value, is the spatial propagation risk value, is the preset threshold, is the system stability margin, and k is the attenuation coefficient.

[0013] Preferably, the steps of grouping the power grid according to the time-accumulated risk value, the spatial propagation risk value and the correction value of the system stability margin are: S301, based on the time-accumulated risk value, the spatial propagation risk value, and the correction value of the system stability margin, feature clustering is performed on the power grid nodes to obtain time-accumulated high-incidence areas, spatial propagation high-sensitivity areas, and stability margin weak areas; S302, setting grouping rules, grouping according to the grouping rules, and establishing a topology model for each group; S303, constructing an optimization model within the formation based on the time-accumulated risk value, the spatial propagation risk value, and the correction value of the system stability margin; S304: Construct a coordination mechanism between the groups according to the divided groups.

[0014] Preferably, the nodes whose time cumulative risk values ​​exceed the time cumulative risk threshold are screened out, and the screened out nodes constitute the high-incidence area of ​​risk time accumulation; the nodes whose space propagation risk values ​​exceed the space propagation risk threshold are screened out, and the screened out nodes constitute the high-sensitivity area of ​​space propagation; the nodes whose correction values ​​of the system stability margin are less than the correction threshold of the system stability margin are screened out, and the screened out nodes constitute the weak area of ​​stability margin.

[0015] Preferably, the objective function of the optimization model is: ,in, is a cost-related item, which represents the total cost during the system operation process. N is the number of time periods involved in the optimization. is the load change rate in the ith period, reflecting the load fluctuation; is the electricity price in the ith period, F is the frequency of risk events, is the average risk intensity, Time cumulative risk value, is the spatial propagation risk value, is the system stability margin of group G, 、 、 、 and are all weight coefficients.

[0016] One or more technical solutions provided in this application have at least the following technical effects or advantages: the frequency of risk events, the average risk intensity, and the system stability margin jointly determine the stability and reliability of the power system. When the frequency of risk events and the average risk intensity are low, and the system stability margin is high, the system has a strong anti-disturbance capability, can cope with various emergencies, ensure the stability and reliability of power supply, reduce the probability of high-frequency, low-intensity risk events, and suppress low-frequency, high-intensity extreme risks. The system stability margin (SSM) is introduced as an optimization objective to quantify the anti-disturbance capability of the power grid, improve the overall stability of the system, reduce the risk of system collapse caused by improper peak regulation, and provide a scheduling solution that takes into account economy, safety, and stability. Through a multi-objective optimization model, a hydropower output curve, a wind and solar power absorption strategy, and an energy storage scheduling solution are obtained to achieve the dual goals of maximizing system stability and controlling the peak-to-valley difference of residual load. The peak regulation risk is decomposed into the frequency of risk events and the average risk intensity, providing a more comprehensive risk assessment perspective. By constructing a multi-objective optimization model, the expected value of the peak-to-valley difference of residual load, the frequency of risk events, the average risk intensity, and the system stability margin are comprehensively considered to provide a more comprehensive and scientific scheduling solution. The introduction of spatiotemporal risk dimensions makes the quantification method more scientific, and the revised value of the system stability margin is more in line with the actual risk. The hierarchical optimization is combined with the rolling time domain optimization to adapt to the dynamic changes of the system. The time accumulation risk blocking and space propagation risk suppression strategies reduce the impact of high-risk scenarios. Through spatiotemporal risk assessment and blocking strategies, the frequency and intensity of risk events are reduced. The revised system stability margin more accurately reflects the system stability, ensuring that the system stability margin is greater than or equal to 0.5, improving the system stability margin, and constructing a multi-objective optimization model for the power system that comprehensively considers the time accumulation effect and space propagation characteristics, improving system stability and reliability, and reducing operational risks. Through the zoning and grouping method, the power grid is managed and optimized in a refined manner to improve the stability and reliability of the power grid, reduce operational risks, and avoid cross-regional transmission. By dividing the risk accumulation high-incidence areas, spatial transmission high-sensitivity areas, and stability margin weak areas, and implementing intra-group optimization and inter-group coordination mechanisms, the operational risks of the power grid are effectively reduced. Through the dynamic sharing mechanism of stability margin, the complementarity of stability margin between groups is achieved, and the overall stability of the power grid is improved. Through the zoning and grouping and intra-group optimization models, the efficiency and accuracy of power grid optimization are improved. Node risks are quantified through dynamic risk indexes to avoid local accumulation of risks. Grouping boundary adjustments are triggered based on the dynamic risk index to optimize the spatiotemporal distribution of water, wind, and solar resources and enhance the dynamic adaptability of the system. Through grouping and reorganization and coordination of flexible resources, precise risk prevention and control of water, wind, and solar complementarity is achieved. Through gradient layered release and triggered release mechanisms, resource utilization is maximized. By actively releasing resources, the regulation pressure of high-risk nodes is reduced, and they are guided in reverse to a low-dependence-self-balancing state. Actively releasing resources is a forward-looking, preventive peak-shaving strategy, proactive prevention and control, which reduces the system's sensitivity to risks by calling resources in advance and avoids chain failures caused by passive responses. Its core lies in risk threshold triggering, dynamic resource allocation, and reverse operation logic. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of a risk scheduling method for a water-wind-solar complementary system according to the present invention; Figure 2 A schematic diagram of a process for dynamically adjusting a multi-objective optimization model according to the present invention; Figure 3 This is a flow chart of grouping power grids according to the time-accumulated risk value, the spatial propagation risk value, and the correction value of the system stability margin in the present invention; Figure 4 A schematic diagram of the process flow of introducing dynamic risk-driven grouping and reorganization in the present invention. DETAILED DESCRIPTION

[0018] To facilitate understanding of the present invention, the present application will be described more comprehensively below with reference to the relevant drawings; the drawings show preferred embodiments of the present invention, but the present invention can be implemented in many different forms and is not limited to the embodiments described herein; on the contrary, the purpose of providing these embodiments is to enable a more thorough and comprehensive understanding of the disclosed content of the present invention.

[0019] It should be noted that the terms “vertical”, “horizontal”, “up”, “down”, “left”, “right” and similar expressions used in this document are for illustrative purposes only and do not represent the only implementation method.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains; the terms used herein in the specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention; the term "and / or" used herein includes any and all combinations of one or more of the associated listed items.

[0021] Example 1: Figure 1 1 is a flow chart of a risk scheduling method for a hydro-wind-solar complementary system according to an embodiment of the present invention, including: S101, collecting historical output data and grid topology parameters; Specifically, the collected historical output data includes output data of hydropower, wind power, photovoltaic power, and load curve data of the power grid. The historical output data covers different seasons and weather conditions to ensure the comprehensiveness and representativeness of the data. The historical output data and load curve data are obtained from the data acquisition and monitoring control system of the local power company or power grid operator, and the collected historical output data is stored in a database. The power grid topology parameters include line capacity, node voltage limit, transformer parameters, generator parameters, etc. The real-time topology parameters of the power grid are obtained through the data acquisition and monitoring control system (SCADA). The SCADA system can monitor the operating status of the power grid in real time, including line current, voltage, power and other parameters. The collected topology parameters are organized and stored according to a unified format and naming rules. The collected topology parameters are verified to ensure their accuracy and reliability. The verification is carried out by comparing with known data, performing simulation calculations, etc. The collected historical output data and power grid topology parameters are cleaned, outliers and missing values ​​are removed, and the data is normalized or standardized.

[0022] S102, predicting time series and generating scenarios based on the collected historical output data, and calculating scenario probabilities based on the generated scenarios; Furthermore, from the collected historical data, we screen out the wind and solar output data (wind power, photovoltaic) and load demand data for prediction. According to business needs and data characteristics, we extract the features that affect wind and solar output and load demand, such as time (hour, week, month), weather type (sunny, cloudy, rainy, etc.), temperature, humidity, etc., and encode the extracted features, such as converting the weather type into a numerical feature (sunny = 1, cloudy = 2, rainy = 3, etc.). We train the LSTM model, divide the processed historical data into a training set and a test set, and build an LSTM network. The structure includes an input layer, a hidden layer, and an output layer. The LSTM model is trained using a training set. The network weights are adjusted through the backpropagation algorithm to minimize the prediction error. The LSTM model is built and trained using the Python Keras library. The trained model is evaluated using a test set, and the prediction error index is calculated. Based on the evaluation results, the model is tuned. The trained LSTM model is used to predict the wind and solar power output and load demand for the next 24 hours. The features of the future time point are used as input to the LSTM model, and the LSTM model outputs the prediction results.

[0023] Based on business needs and actual conditions, the type of generated scenario is determined. The generated scenarios include high wind and solar output + low load scenarios, low wind and solar output + high load scenarios, normal wind and solar output + normal load scenarios, and scenarios under extreme weather conditions. For each scenario type, the key random variables that affect wind and solar output and load demand are determined. Based on the collected historical output data and expert experience, the probability distribution of each random variable is obtained. A large number of random samples are generated based on the determined random variable distribution using the Monte Carlo simulation method. For each random variable, the corresponding random value is generated using its probability distribution. The generated random samples are combined to form multiple groups of scenarios. The scenario probability is the product of the probabilities of each random variable. The formula is: ,in, represents the probability of scenario S, P(W) represents the probability of weather type related to scenario S, P(L) represents the probability of load demand condition related to scenario S, and P(M) represents the probability of wind and solar output condition related to scenario S.

[0024] S103, using a quantitative method to obtain a peak load risk quantitative value based on historical output data and forecast time series, and splitting the peak load risk quantitative value into risk event frequency and average risk intensity; Specifically, based on historical output data and predicted time series, the peak-shaving risk of the hydro-wind-solar complementary system is quantified using the risk quantification method of the upper bias moment indicator to obtain a quantified peak-shaving risk value. The upper bias moment indicator is an indicator used to measure the risk level of a random variable exceeding a certain threshold. In the peak-shaving risk quantification, the upper bias moment indicator can be defined as the moment when the system output exceeds a preset load demand threshold. All generated scenarios are traversed to screen out scenarios in which the peak-to-valley difference of the residual load exceeds the preset threshold. The peak-to-valley difference of the residual load refers to the difference between the peak value and the valley value of the load in a scenario. The screened scenarios are statistically analyzed and the sum of the probabilities of the screened scenarios is calculated. The formula is: ,in, is the risk event frequency, which represents the total probability of the scenario where the peak-to-valley difference of the residual load exceeds the threshold. is the probability of the i-th scenario, is the set of all scenarios exceeding the threshold; for each scenario where the peak-to-valley difference of the residual load exceeds the threshold, its excess value is calculated, and the average risk intensity is calculated based on the excess value. The formula is: ,in, is the average risk intensity, which indicates the average degree of exceeding the threshold, is the peak-to-valley difference of the residual load in the i-th scenario, is the preset threshold, is the frequency of risk events, is the probability of the i-th scenario, is the set of all scenes exceeding the threshold.

[0025] S104, calculating the power flow distribution and critical removal time of the power grid according to the generated scenario and power grid topology parameters, and calculating the system stability margin based on the power flow distribution and critical removal time; Furthermore, for the generated scenario, node information, branch information, generator parameters and load data in the grid topology parameters are prepared. The node information is the number, name, type and connection relationship between all nodes. The branch information is the connection relationship of all branches in the grid, including the node number at both ends of the branch, branch impedance, branch admittance and branch capacity limit. The generator parameters include the nodes and generator model of the generator. The load data includes load nodes and load characteristics. The nodes in the grid are divided into PQ nodes (load nodes), PV nodes (generator nodes) and balance nodes. According to the requirements of the power system analysis and synthesis program (PSASP) flow calculation tool, an input file is written. The input file includes the grid topology, generator parameters, load data, etc. Run PSASP, load the input file, and start the flow calculation process. The flow calculation is solved using an iterative method until the convergence conditions are met, and the results of the flow calculation are obtained, including the voltage amplitude and phase angle of each node, the active power and reactive power flow of each branch, etc.

[0026] By establishing a fault model to simulate various fault conditions that occur in the power grid, the fault location is selected according to the actual operation of the power grid and the existing weak links, and parameters such as the fault type, fault occurrence time, and fault duration are set. The fault duration is usually set from the moment the fault occurs to the moment the fault is cleared. According to each simulated fault scenario, the critical clearing time of the system is calculated, and parameters such as the simulation time step and total simulation time are determined. Starting from the moment the fault occurs, the power grid is simulated in the time domain to simulate the dynamic process of the system under the action of the fault. During the simulation process, the changes in the system's state variables over time are recorded. By gradually adjusting the fault clearing time, the dynamic response of the system after the fault is cleared is observed. When the fault clearing time is less than a certain value, the system can recover to a stable operating state; when the fault clearing time is greater than the specific value, the system will lose stability. This specific value is the critical clearing time of the system.

[0027] The system stability margin is calculated based on the critical cut-off time. The formula is: ,in, is the system stability margin, It is a pre-set critical cut-off time threshold used to evaluate the system's anti-disturbance capability. The threshold is set based on factors such as the grid's operating requirements, safety standards, and historical experience. It indicates the maximum fault clearing time during which the system can maintain stable operation under actual operating conditions. This is also the critical clearing time of the system obtained through transient stability analysis. It reflects the stability of the system in the face of actual faults.

[0028] S105, constructing a multi-objective optimization model based on the risk event frequency, average risk intensity and system stability margin; Specifically, the objective function of the multi-objective optimization model is constructed based on the obtained risk event frequency, average risk intensity and system stability margin. The objective function is: ,in, is a cost-related item, which represents the total cost during the system operation process, N is the number of time periods involved in the optimization, is the load change rate in the ith period, reflecting the load fluctuation; is the electricity price in the ith period. This term measures the cost of the system due to load changes and electricity price factors in different periods by the sum of the product of the load change rate and the electricity price. F is the frequency of risk events. is the average risk intensity, is the system stability margin, 、 and Both are weight coefficients. The constraints of the objective function are: Hydropower output constraint: P hydro,min ≤ P hydro ≤ P hydro,max Meaning: Phydro represents the actual hydropower output, Phydro,min and P hydro,max are the lower and upper limits of hydropower output, respectively. Hydropower output is limited by multiple factors such as reservoir water level, water inflow, and unit capacity; Wind and solar power output constraint: 0 ≤ P wind,P solar ≤ predicted maximum. Meaning: P wind and P solar are the actual wind power and photovoltaic power outputs, respectively, and their values ​​range from 0 to their respective predicted maximums. Wind and solar power output is intermittent and uncertain, and its maximum value is usually estimated based on weather forecasts and equipment capacity. Grid safety constraints include voltage limits and frequency deviations. The voltage limit requires that the voltage fluctuation range of each node in the system is within ±5% of the rated voltage. Voltage is one of the important indicators of power quality. Excessive voltage fluctuations may affect the normal operation of electrical equipment and even cause equipment damage. The frequency deviation stipulates that the system frequency deviation shall not exceed ±0.2Hz. Frequency stability is a basic requirement for the safe operation of the power system. Excessive frequency deviation will affect the synchronous operation and power quality of the power system. Stability constraint: SSM ≥ 0.4 (threshold is adjustable). The system stability margin SSM must be greater than or equal to 0.4 (this threshold can be adjusted according to actual conditions). The stability margin is an important indicator for measuring the system's ability to resist disturbances. A higher stability margin means that the system can better maintain stability in the face of various disturbances.

[0029] The multi-objective particle swarm algorithm (MOPSO) is used to solve the objective function, initialize the particle swarm, and determine the variables that need to be optimized, including the hydropower output plan (i.e., the hydropower output value in each time period) and the wind and solar power consumption ratio (i.e., the ratio of wind power and photovoltaic power generation in the total power generation). A set of initial solutions are randomly generated within the value range of the variables as the initial position of the particle swarm. The objective function value and the degree of constraint violation of each particle are calculated. The variable value of each particle is substituted into the objective function to calculate the corresponding objective function value. It is checked whether each particle meets all the constraints. For particles that do not meet the constraints, the degree of constraint violation is calculated. The degree of constraint violation can be expressed by calculating the sum of the degrees of constraint violation. The particle position is updated and the non-dominated solution set (Pareto frontier) is screened. According to the rules of the algorithm, the position of each particle is updated. In MOPSO, it is usually updated according to the individual optimal position and the global optimal position of the particle; in the updated particle swarm, the non-dominated solution is screened to form a Pareto frontier. Non-dominated solutions are those that cannot be simultaneously dominated by other solutions on multiple objectives. After multiple iterations, a set of optimal solutions is selected from the final Pareto frontier as the optimization result. Based on the variable values ​​in the optimal solution, a hydropower output curve, a wind and solar power consumption strategy, and an energy storage scheduling plan are generated. The hydropower output curve describes the hydropower output at different time periods; the wind and solar power consumption strategy determines the consumption ratio of wind power and photovoltaic power; and the energy storage scheduling plan specifies the charging and discharging strategies of energy storage devices at different time periods to achieve comprehensive system optimization.

[0030] The technical solutions in the above-mentioned embodiments of the present application have at least the following technical effects or advantages: the frequency of risk events, the average risk intensity, and the system stability margin jointly determine the stability and reliability of the power system. When the frequency of risk events and the average risk intensity are low, and the system stability margin is high, the system has a strong anti-disturbance capability, can cope with various emergencies, ensure the stability and reliability of power supply, reduce the probability of high-frequency, low-intensity risk events, and suppress low-frequency, high-intensity extreme risks. The system stability margin (SSM) is introduced as an optimization objective to quantify the anti-disturbance capability of the power grid, improve the overall stability of the system, reduce the risk of system collapse caused by improper peak regulation, and provide a scheduling solution that takes into account economy, safety, and stability. Through a multi-objective optimization model, a hydropower output curve, a wind and solar power consumption strategy, and an energy storage scheduling solution are obtained to achieve the dual goals of maximizing system stability and controlling the peak-to-valley difference of the remaining load. The peak regulation risk is decomposed into the frequency of risk events and the average risk intensity, providing a more comprehensive risk assessment perspective. By constructing a multi-objective optimization model, the expected value of the peak-to-valley difference of the remaining load, the frequency of risk events, the average risk intensity, and the system stability margin are comprehensively considered to provide a more comprehensive and scientific scheduling solution.

[0031] Example 2: Based on Example 1, this example further expands the risk quantification dimension, introduces the risk time accumulation effect (continuous high-risk period) and spatial propagation characteristics (spatial chain effect of risk events), and reduces the probability of long-term risk events. Figure 2 shown.

[0032] S201, obtaining load forecast data, identifying high-risk periods based on the load forecast data, generating time accumulation scenarios, and calculating time accumulation risk values ​​based on the time accumulation scenarios; Furthermore, load forecast data for the power system in the future is obtained. The data should have a high time resolution, such as one data point every 15 minutes, to ensure that the dynamic changes of the load can be accurately captured. According to the operating experience and safety requirements of the power system, a threshold for the peak-to-valley difference of the residual load is set. The threshold should comprehensively consider factors such as the system's peak-shaving capability, spare capacity, and stable operation requirements. The load forecast data is divided into continuous time periods. The length of each time period can be set according to actual conditions. For each time period, the peak-to-valley difference of the residual load in the time period is calculated, that is, the difference between the maximum and minimum values ​​of the load in the time period. All continuous time period combinations are traversed to identify time periods in which the peak-to-valley difference of the residual load exceeds the threshold for multiple consecutive time periods. These time periods are high-risk time periods. For the identified high-risk time periods, different time cumulative risk scenarios are constructed, and the time cumulative risk value of each scenario is calculated. The formula is: ,in, represents the part exceeding the threshold value during the tth period, i.e. >r, = -r; when When ≤r, =0, λ is the time decay factor, which is 0.9, indicating that the recent risk accumulation weight is higher, and T is the number of periods the scenario lasts.

[0033] S202, generating a spatial propagation scenario according to the power grid topology parameters, and calculating a spatial propagation risk value based on the spatial propagation scenario; Specifically, according to the grid topology parameters of the power system, the connection relationship between each node and line in the network is analyzed, and a topological model of the grid is constructed. The relevant indicators of betweenness centrality are used to identify the key nodes and lines in the grid. The identified key nodes or lines are used as single point failure points to simulate their failure situations. A cascading failure propagation model is constructed according to the load rate and betweenness centrality of the lines. The line load rate reflects the ratio of the current load carried by the line to its rated capacity, which is an important indicator for judging whether the line is overloaded; the betweenness centrality reflects the hub degree of the line in the network, that is, the frequency and importance of the power transmission of the line in the network. Starting from the single point failure point, the established cascading failure propagation model is used to gradually simulate the fault propagation process. In each step of the simulation, it is judged whether a new line is overloaded according to the change of the line load rate. If the load of a certain line If the load rate exceeds its rated capacity, the line is considered overloaded and the simulation continues with it as a new fault point. The load rate of each line at different stages is recorded, including the load rate values ​​before the fault occurs, at the moment of fault occurrence, and at each stage during the fault propagation process. A spatial propagation scenario is generated. By traversing the shortest paths between all pairs of nodes in the power grid, the number of times each node or line appears on these shortest paths is counted, and then normalized to obtain the value of betweenness centrality. Using numerical simulation methods, the load of the node or line is slightly disturbed, and the changes in parameters such as voltage and power of other nodes and lines in the network are observed. Based on these changes, the load sensitivity is quantified. The betweenness centrality and load sensitivity are multiplied by the corresponding weight coefficients and then added to obtain a comprehensive risk index. The comprehensive risk index of each node and line is weighted summed to calculate the spatial propagation risk value.

[0034] S203, calculating the composite risk value and the system stability margin based on the time-accumulated risk value and the spatial propagation risk value, and making corrections; Furthermore, the composite risk value is calculated based on the time-accumulated risk value and the spatial propagation risk value. The formula is: ,in, is the composite risk value, Time cumulative risk value, is the spatial propagation risk value, α and β are weight coefficients, and the formula for correcting the system stability margin is: ,in, is the correction value of the system stability margin, Time cumulative risk value, is the spatial propagation risk value, It is a preset threshold value, which is set according to the actual situation and stability requirements of the power grid. is the system stability margin, k is the attenuation coefficient, and they are calibrated through actual data.

[0035] S204, dynamically adjusting the multi-objective optimization model; Specifically, by suppressing high-risk scenarios in time and space, and based on the risk status monitored in real time, if a high-risk period occurs, the output of hydropower will be increased in advance to reduce the peak-shaving pressure of other power generation methods and avoid the situation where the peak-to-valley difference of the residual load continuously exceeds the threshold. When it is detected that the load rate of the key line is close to the overload threshold, the discharge of the energy storage equipment is controlled in time to share the line load; during the low-power consumption period, the charging of the energy storage equipment is controlled to provide support for the peak period. Taking into account the characteristics and costs of various power generation methods, such as thermal power, hydropower, wind power, photovoltaics, etc., the output of each power generation method is reasonably allocated according to the load demand and the correction value of the system stability margin. At the same time, the load forecast data, wind and solar output forecast data and risk status information are updated every 30 minutes. Based on the updated data, the improved algorithm is re-run for multi-objective optimization solution to obtain a new hydropower output plan, wind and solar absorption strategy and energy storage scheduling plan.

[0036] The technical solutions in the above-mentioned embodiments of the present application have at least the following technical effects or advantages: the introduction of the spatiotemporal risk dimension makes the quantification method more scientific, the revised value of the system stability margin is more in line with the actual risk, the hierarchical optimization is combined with the rolling time domain optimization to adapt to the dynamic changes of the system, the time accumulation risk blocking and space propagation risk suppression strategies reduce the impact of high-risk scenarios, and the frequency and intensity of risk events are reduced through spatiotemporal risk assessment and blocking strategies. The revised system stability margin more accurately reflects the system stability, ensuring that the system stability margin is greater than or equal to 0.5, improving the system stability margin, and constructing a multi-objective optimization model of the power system that comprehensively considers the time accumulation effect and space propagation characteristics, thereby improving system stability and reliability and reducing operational risks.

[0037] Example 3: Based on Example 2, the power grid is divided into risk-stability zoning groups from a macro perspective, and the power grid is divided into several independent or weakly coupled sub-areas (groups). Risk self-balancing and stability coordination are achieved in each group, such as Figure 3 shown.

[0038] S301, based on the time-accumulated risk value, the spatial propagation risk value, and the correction value of the system stability margin, feature clustering is performed on the power grid nodes to obtain time-accumulated high-incidence areas, spatial propagation high-sensitivity areas, and stability margin weak areas; Furthermore, the fluctuation range of the time cumulative risk value, spatial propagation risk value and the correction value of the system stability margin is identified. Using power grid simulation software, different operating scenarios are simulated and the changing patterns of various indicators are observed. Based on the simulation results, the time cumulative risk threshold, spatial propagation risk threshold and correction threshold of the system stability margin are set to ensure the safe and stable operation of the power grid. The determined thresholds are cross-validated to evaluate the rationality and effectiveness of the thresholds. The nodes whose time cumulative risk values ​​exceed the time cumulative risk threshold are screened out. The screened out nodes constitute the high-incidence areas of risk time accumulation, indicating that these areas have a high frequency of risk events in the time dimension. Similarly, the nodes whose spatial propagation risk values ​​exceed the spatial propagation risk threshold are screened out. The screened out nodes constitute the high-sensitivity areas of spatial propagation, indicating that they have a greater impact on risk propagation in the spatial dimension. The nodes whose correction values ​​of the system stability margin are less than the correction threshold of the system stability margin are screened out. The screened out nodes constitute the weak stability margin areas, indicating that they are more likely to lose stability when the power grid is disturbed.

[0039] S302, setting grouping rules, grouping according to the grouping rules, and establishing a topology model for each group; Specifically, the risk isolation principle and the stability coordination principle are set. The risk isolation principle is to isolate high-risk nodes from low-risk nodes to avoid chain reaction of risks. The stability coordination principle is to place groups with high stability margins adjacent to weak groups. Groups with high stability margins provide support to weak groups, thereby improving the overall stability of the power grid and ensuring that the power grid can recover quickly when disturbed. According to the risk isolation principle and the stability coordination principle, the power grid is divided into several groups. Each group contains several nodes and branches to form relatively independent operating units. Nodes within the same group should have similar risk-stability characteristics to ensure that the nodes within the group are physically connected to avoid isolated areas.

[0040] A topological model is established for each marshaling to describe the connection relationship between nodes and branches within the marshaling. The topological model includes node attributes (such as risk level and stability margin), branch attributes (such as impedance and capacity), and information on interaction interfaces between marshalings. The marshaling topological model is applied to power grid planning, operation, and scheduling to provide decision support for power grid management. The model is used for simulation analysis to evaluate power grid performance and risk levels under different operating scenarios. The marshaling topological model is optimized and adjusted based on simulation results and actual operating data.

[0041] S303, constructing an optimization model within the formation based on the time-accumulated risk value, the spatial propagation risk value, and the correction value of the system stability margin; Furthermore, an optimization model within the group is constructed based on the time-accumulated risk value, the spatial propagation risk value, and the correction value of the system stability margin. The objective function of the optimization model is: ,in, is a cost-related item, which represents the total cost during the system operation process. N is the number of time periods involved in the optimization. is the load change rate in the ith period, reflecting the load fluctuation; is the electricity price in the ith period. This term measures the cost of the system due to load changes and electricity price factors in different periods by the sum of the product of the load change rate and the electricity price. F is the frequency of risk events. is the average risk intensity, Time cumulative risk value, is the spatial propagation risk value, is the system stability margin of group G, 、 、 、 and Both are weight coefficients. The constraints are: the time cumulative risk value and spatial propagation risk value of group G shall not exceed the set threshold, and the interaction power between group G and other groups shall be less than or equal to the maximum interaction power allowed between group G and other groups.

[0042] S304, building a coordination mechanism between the groups according to the divided groups; Specifically, the coordination mechanism between the marshaling units includes two major mechanisms: dynamic sharing of stability margin and blocking of risk propagation. For the dynamic sharing mechanism of stability margin, by setting a system stability margin threshold, when the system stability margin of the marshaling unit is less than the system stability margin threshold, the stability margin sharing mechanism is triggered. By identifying the flexible resources available in the adjacent marshaling units, when the stability margin of the marshaling unit is insufficient, a support request is sent to the adjacent marshaling unit to call on its spare capacity; for the risk propagation blocking mechanism, the risk indicators of each marshaling unit are monitored in real time, and high-risk marshaling units, that is, units whose risk indicators exceed the set threshold, are identified. At the boundary between the high-risk marshaling unit and the adjacent marshaling unit, a virtual impedance is set (such as by adjusting the parameters of the connecting line) to increase the difficulty of risk propagation. The size of the virtual impedance can be dynamically adjusted according to the amount of risk propagation. The greater the amount of risk propagation, the greater the virtual impedance.

[0043] The technical solutions in the above-mentioned embodiments of the present application have at least the following technical effects or advantages: through the zoning and grouping method, the power grid is finely managed and optimized, the stability and reliability of the power grid are improved, the operation risk is reduced, and cross-regional propagation is avoided. By dividing the high-incidence areas of risk time accumulation, the high-sensitivity areas of spatial propagation and the weak areas of stability margin, and implementing the intra-grouping optimization and inter-grouping coordination mechanism, the operation risk of the power grid is effectively reduced. Through the dynamic sharing mechanism of stability margin, the complementarity of stability margin between groups is achieved, and the overall stability of the power grid is improved. Through the zoning and grouping and intra-grouping optimization models, the efficiency and accuracy of power grid optimization are improved.

[0044] Example 4: Based on Example 3, a dynamic risk-driven marshaling reorganization mechanism is introduced to split high-risk nodes into adjacent low-risk marshalings; adjust the support direction of flexible resources between marshalings; avoid local accumulation of risks caused by fixed marshaling, improve the dynamic adaptability of the system, and actively release resource gradients to reversely disperse risks to low-risk levels to avoid excessive concentration of resources. Figure 4 shown.

[0045] S401, obtaining volatility, uncertainty and failure probability, and calculating a dynamic risk index based on volatility, uncertainty and failure probability; Furthermore, the volatility is used to quantify the error in wind and solar power output forecasts, reflecting the instability of renewable energy output, so that the error calculation method can obtain volatility. The uncertainty is used to quantify the volatility of load demand, reflecting the uncertainty of load forecasts. Uncertainty is obtained by calculating the deviation between actual load and predicted load. The failure probability is used to quantify the failure risk of equipment (such as lines and transformers). The failure probability is calculated based on historical failure data. The dynamic risk index is calculated based on the obtained volatility, uncertainty and failure probability. The formula is: ,in, is the dynamic risk index of node i at time t, It is a volatility indicator, reflecting the wind and solar power output forecast error. is an uncertainty indicator, reflecting the load demand fluctuation rate. is a failure probability indicator that reflects the risk of equipment failure. 、 and are weight coefficients, corresponding to the importance of volatility, uncertainty, and failure probability, respectively.

[0046] S402, reorganize the group according to the dynamic risk index; Specifically, when the DRI mean of the group is greater than or equal to the set dynamic risk index threshold, the group reorganization is triggered. The reorganization steps are as follows: by calculating the DRI value of each node in the group, sorting them from high to low according to the DRI value, selecting the 20% nodes with the highest DRI as high-risk nodes, and splitting the highest-risk nodes in the high-risk group to the adjacent low-risk group to reduce the overall risk, reduce the high-risk group's dependence on external hydropower resources, reduce the priority of external support, increase the priority of calling the internal energy storage (such as battery energy storage, pumped storage) of the high-risk group, give priority to using its own resources to deal with risks, and update the topology of the group and the target group according to the node splitting results. According to the connecting lines between the groups, recalculate its maximum interactive power limit.

[0047] S403, constructing a risk-controllable scheduling model based on the dynamic risk index within the formation; Furthermore, the dynamic risk index within each group is weighted to obtain the global risk value, and the objective function of the risk-controllable scheduling model is obtained based on the dynamic risk index within each group: ,in, is the global risk value, calculated based on the weighted DRI of each group. is the wind and solar curtailment rate, constraint condition: curtailment rate ≤ 5%, Complementarity is the complementary income of water, wind and solar. 、 and is the weight coefficient, which is adjusted according to the grid operation target. The constraint condition is that the risk level of each group does not exceed the threshold.

[0048] S404: Acquire real-time data to obtain a peak-shaving risk index. Determine resource release based on the peak-shaving risk index. If the peak-shaving risk index is greater than or equal to a preset threshold, trigger-based release is initiated. Otherwise, gradient-based tiered release is initiated. Specifically, according to the acquired real-time output data and real-time load data, the data is substituted into step S401 to calculate the real-time peak-shaving risk index, and the situation of resource release is judged according to the real-time peak-shaving risk index. If the real-time peak-shaving risk index is greater than or equal to the preset threshold, the trigger release is started. Conversely, if the real-time peak-shaving risk index is less than the preset threshold, the gradient layered release is started. The trigger release is to linearly calculate the release ratio according to the real-time peak-shaving risk index, and give priority to releasing the energy storage resources of the high-risk grouping to reduce local risks. The gradient layered release divides the risk into gradients and releases it in layers. According to the real-time peak-shaving risk index, the risk is divided into three gradients, namely low-risk gradient, medium-risk gradient and high-risk gradient. For the low-risk gradient, the peak-shaving risk index is between 0 (including 0) and 0.5 (excluding 0.5), and priority is given to the high-risk grouping. Adjust the output of conventional units, temporarily do not release energy storage resources, adjust the output of thermal power units or gas turbines, increase or decrease power generation, monitor grid frequency and voltage, and ensure system stability; for medium-risk gradients, the peak-shaving risk index is between 0.5 (inclusive) and 0.8 (excluding 0.8), release some energy storage resources, reduce hydropower reserve capacity, release 5%-10% of energy storage resources, give priority to releasing energy storage in low-risk groups, reduce hydropower reserve capacity by 5%-10%, and reduce adjustable hydropower output; for high-risk gradients, the peak-shaving risk index is between 0.8 (inclusive) and 1 (excluding 1), give priority to trigger-type release of energy storage, superimpose gradient-based stratification rules, allow temporary abandonment of wind power and photovoltaic power, release 10%-30% of energy storage resources, give priority to releasing energy storage in high-risk groups, reduce hydropower reserve capacity by 15%-20%, and further release hydropower peak-shaving capacity.

[0049] S405, dynamically adjusting the release ratio of resources based on the change of the peak load risk index to form a coordinated resource allocation; Furthermore, the peak-shaving risk index is recalculated every 15 minutes, and the proportion of resource release is dynamically adjusted according to the changes in the peak-shaving risk index. At the same time, the hydropower backup capacity is dynamically adjusted.

[0050] The technical solutions in the above-mentioned embodiments of the present application have at least the following technical effects or advantages: quantifying node risks through dynamic risk indexes to avoid local accumulation of risks, triggering grouping boundary adjustments based on dynamic risk indexes, optimizing the spatiotemporal distribution of water, wind, and solar resources, and improving the dynamic adaptability of the system, and realizing precise risk prevention and control of water, wind, and solar complementarity through grouping and reorganization and coordination with flexible resources. Through gradient layered release and triggered release mechanisms, resource utilization is maximized, and the regulation pressure of high-risk nodes is reduced by actively releasing resources, guiding them in reverse to a low-dependence-self-balancing state. Actively releasing resources is a forward-looking, preventive peak-shaving strategy, active prevention and control, which reduces the system's sensitivity to risks by calling resources in advance and avoids chain failures caused by passive responses. Its core lies in risk threshold triggering, dynamic resource allocation, and reverse operation logic.

[0051] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Various modifications and variations are readily apparent to those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A risk scheduling method for a water-wind-solar complementary system, characterized in that: include: S11, collecting historical data and grid topology parameters, predicting time series and generating scenarios based on the collected historical data, and calculating scenario probabilities based on the generated scenarios; S12, using a quantitative method to obtain a peak-shaving risk quantitative value based on the predicted time series, and splitting the peak-shaving risk quantitative value into risk event frequency and average risk intensity; S13, calculating the power flow distribution and critical removal time of the power grid according to the generated scenario and the power grid topology parameters, and calculating the system stability margin based on the power flow distribution and the critical removal time; S14, construct a multi-objective optimization model based on the peak-shaving risk quantification value after splitting and the system stability margin.

2. The risk scheduling method for a water-wind-solar complementary system according to claim 1, characterized in that: The formula for calculating the scenario probability is: ,in, represents the probability of scenario S, P(W) represents the probability of weather type related to scenario S, P(L) represents the probability of load demand condition related to scenario S, and P(M) represents the probability of wind and solar output condition related to scenario S.

3. The risk scheduling method for a water-wind-solar complementary system according to claim 1, characterized in that: Calculate the system stability margin using the formula: ,in, is the system stability margin, is the pre-set critical resection time threshold, It indicates the longest fault clearing time during which the system can maintain stable operation under actual operating conditions.

4. The risk scheduling method for a water-wind-solar complementary system according to claim 1, characterized in that: S201, acquiring load forecast data according to a forecast time series, identifying high-risk periods according to the load forecast data, generating a time accumulation scenario, and calculating a time accumulation risk value based on the time accumulation scenario; S202, generating a spatial propagation scenario according to the power grid topology parameters, and calculating a spatial propagation risk value based on the spatial propagation scenario; S203, calculating a composite risk value and a system stability margin correction value based on the time-accumulated risk value and the spatial propagation risk value; S204, dynamically adjusting the multi-objective optimization model.

5. The risk scheduling method for a water-wind-solar complementary system according to claim 4, characterized in that: The steps to form high-risk periods are as follows: divide the load forecast data into continuous time periods. For each time period, calculate the peak-to-valley difference of the remaining load in the time period, that is, the difference between the maximum and minimum load values ​​in the time period. Traverse all continuous time period combinations and identify the time periods in which the peak-to-valley difference of the remaining load exceeds the threshold for multiple consecutive time periods. These time periods are high-risk periods.

6. The risk scheduling method for a water-wind-solar complementary system according to claim 4, characterized in that: Calculate the time cumulative risk value, the formula is: ,in, represents the part exceeding the threshold value during the tth period, i.e. >r, = -r; when When ≤r, =0, λ is the time decay factor, and T is the number of time periods the scene lasts.

7. The risk scheduling method for a water-wind-solar complementary system according to claim 4, characterized in that: The formula for correcting the system stability margin is: ,in, is the correction value of the system stability margin, Time cumulative risk value, is the spatial propagation risk value, is the preset threshold, is the system stability margin, and k is the attenuation coefficient.

8. The risk scheduling method for a water-wind-solar complementary system according to claim 7, characterized in that: According to the time-accumulated risk value, spatial propagation risk value and the correction value of the system stability margin, the steps for grouping the power grid are as follows: S301, based on the time-accumulated risk value, the spatial propagation risk value, and the correction value of the system stability margin, feature clustering is performed on the power grid nodes to obtain time-accumulated high-incidence areas, spatial propagation high-sensitivity areas, and stability margin weak areas; S302, setting grouping rules, grouping according to the grouping rules, and establishing a topology model for each group; S303, constructing an optimization model within the formation based on the time-accumulated risk value, the spatial propagation risk value, and the correction value of the system stability margin; S304: Construct a coordination mechanism between the groups according to the divided groups.

9. The risk scheduling method for a water-wind-solar complementary system according to claim 8, characterized in that: Nodes whose time cumulative risk values ​​exceed the time cumulative risk threshold are screened out, and the screened out nodes constitute the high-incidence area of ​​risk time accumulation; nodes whose space propagation risk values ​​exceed the space propagation risk threshold are screened out, and the screened out nodes constitute the high-sensitivity area of ​​space propagation; nodes whose correction values ​​of the system stability margin are less than the correction threshold of the system stability margin are screened out, and the screened out nodes constitute the weak area of ​​stability margin.

10. The risk scheduling method for a hydro-wind-solar complementary system according to claim 8, characterized in that: The objective function of the optimization model is: ,in, is a cost-related item, which represents the total cost during the system operation process. N is the number of time periods involved in the optimization. is the load change rate in the ith period, reflecting the load fluctuation; is the electricity price in the ith period, F is the frequency of risk events, is the average risk intensity, Time cumulative risk value, is the spatial propagation risk value, is the system stability margin of group G, 、 、 、 and are all weight coefficients.

Images