A method and system for generating a wind and solar power output scene with adjustable operation risk, and a medium

Abstract

This invention discloses a method, system, and medium for generating wind and solar power output scenarios with adjustable operational risk, relating to the field of power system operation optimization technology with a high proportion of new energy sources. The method includes: inputting a collection of historical wind and solar power operation scenarios into a power system optimization scheduling model to obtain the system operation status and constructing a flexibility index; quantifying the flexibility index into a system operation risk index based on the analytic hierarchy process (AHP) and entropy weight method, and using the system operation risk index as a system operation risk label to form a historical scenario library with system operation risk labels; constructing an operation scenario generation model based on a conditional generative adversarial network (GAN), and training the operation scenario generation model using the system operation risk labels as conditional variables based on the historical scenario library; and generating wind and solar power output scenarios according to specified risk levels based on the trained wind and solar power output scenario generation model. This invention enables the generation of scenarios with different risk levels on demand.

Description

Technical Field

[0001] This invention relates to the field of power system operation optimization technology with a high proportion of new energy sources, specifically to a method and system for generating wind and solar power output scenarios with adjustable operation risk. Background Technology

[0002] Against the backdrop of the ongoing pursuit of "dual carbon" goals, the proportion of clean energy sources such as wind power and solar power in the global energy mix is ​​continuously increasing. With the rapid rise in installed capacity and grid connection rates of new energy sources, the inherent volatility and intermittency of wind and solar power pose greater uncertainties to the power system during operation and dispatch. Particularly under the trend of rapid growth in new energy sources, wind and solar power output, influenced by extreme weather, seasonal variations, and random fluctuations, may lead to increased reserve consumption, intensified unit ramp-up pressure, and power flow exceeding limits, among other operational risks, posing a significant challenge to the safe and stable operation of the power system.

[0003] To address the uncertainty of wind and solar power generation, existing scene generation methods often employ parameterized approaches based on probabilistic models. These methods assume that wind and solar power output follows a known distribution and then sample based on model parameters. However, real-world wind and solar power output often exhibits highly nonlinear, multimodal, and drastic fluctuations, making it difficult to accurately describe with a single distribution. This results in significant limitations of such methods in generating realistic, extreme, and complex scenes. With the development of artificial intelligence technology, nonparametric generation models based on deep learning are emerging as a promising approach. type This approach, applied to the construction of new energy scenarios, generates similar time-series scenarios by learning the distribution characteristics of historical data. While such methods can generate diverse power output trajectories, existing technologies typically focus on generating only single characteristics, such as high output, low output, or high volatility, making it difficult to comprehensively characterize the multidimensional risks that scenarios may trigger from the perspective of power system operation. Furthermore, existing generative models lack targeted control over "risk levels," failing to generate corresponding wind and solar power output scenarios based on the type or level of risk faced by the system. This limits the application value of the generated results in practical dispatch research and risk simulation.

[0004] In terms of operational risk assessment, existing technologies mostly use a single operational indicator as the basis for risk measurement, such as power limit exceedance, frequency or voltage deviation, or branch power flow load factor. While these methods can describe local operational states, they cannot comprehensively reflect the overall risk level under the combined effects of multiple types of operational boundaries in a complex power grid. Furthermore, single-indicator-based assessments are also insufficient to support the needs of risk-oriented scenario generation, and cannot effectively perform coordinated modeling of "risk rating - scenario generation" for wind and solar power output.

[0005] In summary, under the existing technological framework, there is a significant disconnect between the generation of wind and solar power output scenarios and the characterization of power system operational risks. This is particularly evident in the inability to generate corresponding wind and solar power output scenarios based on predetermined operational risk levels. Existing scenario generation methods typically only reproduce historical statistical characteristics or generate single-feature scenarios such as high or low output, lacking the ability to quantitatively characterize and control comprehensive system risks. They cannot achieve the function of "generating corresponding scenarios given a risk level," which makes it difficult for the generated results to accurately reflect the operational pressure that the power system may face under different risk conditions, thus limiting their application value in risk assessment, dispatch verification, and system resilience analysis.

[0006] In view of the above, this application is hereby submitted. Summary of the Invention

[0007] The technical problem this invention aims to solve is that existing scenario generation methods typically only reproduce historical statistical characteristics or generate single-feature scenarios such as high or low output, lacking the ability to quantitatively characterize and control the comprehensive risks of the system. They cannot achieve the function of "generating corresponding scenarios given a risk level," which makes it difficult for the generated results to accurately reflect the operational pressure that the power system may face under different risk conditions. The purpose of this invention is to provide a method and system for generating wind and solar power output scenarios with adjustable operational risk. This invention realizes the ability to characterize the system operational risk index based on scheduling results and generate scenarios with different risk levels on demand. It can accurately reflect the operational pressure that the power system may face under different risk conditions, providing an effective technical means for simulating high-risk operating states of the power system.

[0008] This invention is achieved through the following technical solution:

[0009] In a first aspect, the present invention provides a method for generating wind and solar power output scenarios with adjustable operational risks, the method comprising:

[0010] Historical renewable energy output data is obtained, and a set of historical wind and solar power operation scenarios is constructed based on the historical renewable energy output data. Each scenario in the set of historical wind and solar power operation scenarios is input into a pre-constructed power system optimization and dispatch model to obtain the system operation status.

[0011] Based on the system's operating status, a flexibility index is constructed; the flexibility index is quantified into a system operation risk index based on the analytic hierarchy process and the entropy weight method, and the system operation risk index is used as a system operation risk label to form a historical scenario library with system operation risk labels.

[0012] A scenario generation model based on conditional generative adversarial networks is constructed. Based on a historical scenario library, the scenario generation model is trained with system operation risk labels as conditional variables to obtain a scenario generation model with adjustable operation risk for wind and solar power output.

[0013] Based on the wind and solar power output scenario generation model with adjustable operational risk, wind power and solar power output scenarios are generated according to the specified risk level.

[0014] Furthermore, historical renewable energy output data is obtained, and a set of historical wind and solar power operation scenarios is constructed based on this data. Each scenario in the historical wind and solar power operation scenario set is then input into a pre-built power system optimization and dispatch model to obtain the system operating status, including:

[0015] Obtain historical renewable energy output data, including historical wind power output data and photovoltaic power output data;

[0016] The historical renewable energy output data is preprocessed to obtain the preprocessed historical renewable energy output data.

[0017] The hourly output sequence of the preprocessed historical renewable energy output data is divided by operating day and constructed into a set of historical wind and solar operating scenarios. The wind output curve, solar output curve and photovoltaic output curve of each operating day in the set of historical wind and solar operating scenarios together constitute an independent historical scenario.

[0018] Each scenario from the historical operation scenarios of wind and solar power is input into a pre-built power system optimization scheduling model on a daily basis. The power system optimization scheduling model is then solved to obtain the system operation status under the corresponding scenario.

[0019] Furthermore, the parameters of system operation status include unit output, unit start-up and shutdown status, standby dispatch, and line power flow level.

[0020] Furthermore, flexibility metrics include ramp-up margin, multiple redundancy margin, adjustable capacity margin, and line transmission margin.

[0021] The uphill margin is the total uphill capacity of the power system minus the uphill demand.

[0022] Multiple redundancy margins include the capacity of 5-minute maximum available spinning reserve, 15-minute maximum available spinning reserve, and maximum available operating reserve;

[0023] Adjustable capacity margin is the available reserve of thermal power units to increase or decrease output when the power system needs it.

[0024] Line transmission margin is an assessment of the remaining capacity of a transmission line to carry additional power flow without violating thermal stability and safety constraints.

[0025] Furthermore, based on the analytic hierarchy process (AHP) and entropy weight method, the flexibility index is quantified into a system operation risk index, including:

[0026] With the first The flexibility index is relative to the first The importance ratio of each flexibility indicator Construct an AHP judgment matrix using matrix elements;

[0027] Calculate the largest eigenvalue of the AHP judgment matrix and the weights of each flexibility index; and calculate the weighted largest eigenvalue based on the weights of each flexibility index.

[0028] Calculate the consistency ratio of the consistency index based on the weighted maximum eigenvalue; perform a consistency test based on the consistency ratio; if the consistency test is not met, adjust the AHP judgment matrix; if the consistency test is met, determine the weights of each flexibility index.

[0029] Based on the entropy weight method, the entropy weight of each flexibility index is calculated according to the degree of dispersion of each flexibility index, and the weight of each flexibility index is corrected according to the entropy weight to obtain the corrected weight.

[0030] Based on the revised weights and corresponding flexibility indicators, a system operation risk index is generated on a daily basis.

[0031] Furthermore, the formula for calculating the system operation risk index is as follows:

[0032] ;

[0033] ;

[0034] In the formula, The system operation risk index; For the first The adjusted weights for each flexibility indicator; For the first The value of a flexibility indicator; The number of flexibility indicators; For the first The weighting of each flexibility indicator; For the first Entropy weight of a flexibility index.

[0035] Furthermore, the scene generation model includes a generator and a discriminator, both of which are constructed from two neural networks;

[0036] The neural network expression for the generator is: ,in, This represents a sequence of wind power and solar power output scenarios generated by the generator; This represents Gaussian random noise; These are weight parameters; For the mapping function of the generator network;

[0037] The neural network expression for the discriminator is: ,in, This represents the probability that the input scene belongs to the set of real-world historical operating scenes; This indicates the input scenario, which can be a real historical scenario or a scenario generated by the generator. These are weight parameters; For the discriminator network, the mapping function is used.

[0038] A conditional variable, namely the system operation risk index, is introduced into the operation scenario generation model. The conditional label corresponding to the system operation risk index is simultaneously input into the generator and the discriminator. The generator generates the corresponding wind power and solar power output scenarios under given conditions, while the discriminator evaluates whether the generated wind power and solar power output scenarios are both realistic and consistent with the specified operation risk level.

[0039] Secondly, the present invention provides a wind and solar power output scene generation system with adjustable operational risk, the system comprising:

[0040] The acquisition unit is used to acquire historical renewable energy output data.

[0041] Historical scenario construction unit, used to construct a set of historical operation scenarios for wind and solar power based on historical renewable energy output data;

[0042] The scheduling model unit is used to input the historical operation scenarios of wind and solar power into a pre-built power system optimization scheduling model to obtain the system operation status.

[0043] The flexibility metric construction unit is used to construct flexibility metrics based on the system's operating status.

[0044] The risk quantification unit is used to quantify the flexibility index into the system operation risk index based on the analytic hierarchy process and the entropy weight method, and to use the system operation risk index as the system operation risk label to form a historical scenario library with system operation risk labels.

[0045] Conditional Generative Adversarial Network (GAN) building blocks are used to construct runtime scenario generation models based on conditional generative adversarial networks (GANs).

[0046] The scene generation model training unit is used to train the operation scene generation model based on the historical scene library, with the system operation risk label as a condition variable, to obtain the operation risk adjustable wind and solar power output scene generation model.

[0047] The operation scenario generation unit is used to generate wind power and solar power output scenarios according to a specified risk level based on the wind and solar power output scenario generation model with adjustable operation risk.

[0048] Furthermore, based on the analytic hierarchy process (AHP) and entropy weight method, the flexibility index is quantified into a system operation risk index, including:

[0049] With the first The flexibility index is relative to the first The importance ratio of each flexibility indicator Construct an AHP judgment matrix using matrix elements;

[0050] Calculate the largest eigenvalue of the AHP judgment matrix and the weights of each flexibility index; and calculate the weighted largest eigenvalue based on the weights of each flexibility index.

[0051] Calculate the consistency ratio of the consistency index based on the weighted maximum eigenvalue; perform a consistency test based on the consistency ratio; if the consistency test is not met, adjust the AHP judgment matrix; if the consistency test is met, determine the weights of each flexibility index.

[0052] Based on the entropy weight method, the entropy weight of each flexibility index is calculated according to the degree of dispersion of each flexibility index, and the weight of each flexibility index is corrected according to the entropy weight to obtain the corrected weight.

[0053] Based on the revised weights and corresponding flexibility indicators, a system operation risk index is generated on a daily basis.

[0054] Thirdly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for generating a wind and solar power output scenario with adjustable operational risk.

[0055] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0056] This invention discloses a method and system for generating wind and solar power output scenarios with adjustable operational risks. The method involves acquiring historical renewable energy output data, constructing a set of historical wind and solar power operation scenarios based on this data, inputting each scenario from the set into a pre-built power system optimization scheduling model to obtain the system operating status, constructing a flexibility index based on the system operating status, quantifying the flexibility index into a system operating risk index using a subjective-objective fusion method based on the analytic hierarchy process (AHP) and entropy weighting method, forming a risk quantification method capable of characterizing system operating pressure, and using the system operating risk index as a system operating risk label to form a historical scenario library with system operating risk labels. A conditional generative adversarial network (GAN)-based operation scenario generation model is constructed, and trained using the system operating risk labels as conditional variables based on the historical scenario library to obtain a wind and solar power output scenario generation model with adjustable operational risks. Finally, based on the wind and solar power output scenario generation model with adjustable operational risks, wind and solar power output scenarios are generated according to specified risk levels. This invention enables the characterization of system operation risk index based on scheduling results and the generation of scenarios with different risk levels on demand. It can accurately reflect the operating pressure that the power system may face under different risk conditions and provides an effective technical means for simulating the high-risk operating state of the power system. Attached Figure Description

[0057] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:

[0058] Figure 1 This is an overall flowchart of the method in Embodiment 1 of the present invention;

[0059] Figure 2 This is a flowchart illustrating the generation process of the wind and solar power output scenario with adjustable operational risk in Embodiment 1 of the present invention.

[0060] Figure 3 This is a schematic diagram illustrating the weight calculation relationship between the analytic hierarchy process (AHP) and the entropy weight method in Embodiment 1 of the present invention.

[0061] Figure 4 This is a schematic diagram of the runtime scenario generation model based on conditional generative adversarial networks in Embodiment 1 of the present invention;

[0062] Figure 5 This is a graph of the loss function during the training process in Embodiment 1 of the present invention;

[0063] Figure 6 This is a discriminator score curve during the training process in Embodiment 1 of the present invention;

[0064] Figure 7 This is a comparison diagram of the wind power output generation scenario distribution in Embodiment 1 of the present invention;

[0065] Figure 8 This is a comparison diagram of the photovoltaic power generation scenario distribution in Embodiment 1 of the present invention;

[0066] Figure 9 This is a block diagram of the overall structure of the system in Embodiment 2 of the present invention. Detailed Implementation

[0067] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0068] Existing scenario generation methods can usually only reproduce historical statistical characteristics or generate single-feature scenarios such as high or low output. They lack the ability to quantitatively characterize and control the comprehensive risks of the system and cannot achieve the function of "generating corresponding scenarios given a risk level". This makes it difficult for the generated results to accurately reflect the operating pressure that the power system may face under different risk conditions.

[0069] Therefore, it is necessary to propose a method that can quantify scenario risks based on historical operating data and generate controllable risk level wind and solar power output scenarios through a condition generation model, so as to solve the shortcomings of existing technologies in risk identification and high-risk scenario generation, and improve the power system risk simulation and decision support capabilities.

[0070] Example 1

[0071] like Figure 1 and Figure 2 As shown, Figure 1 This is an overall flowchart of the method in Embodiment 1 of the present invention; Figure 2 This is a flowchart illustrating the generation process of a wind and solar power output scenario with adjustable operational risk in Embodiment 1 of the present invention; the present invention provides a method for generating a wind and solar power output scenario with adjustable operational risk, the method comprising:

[0072] Step 1: Obtain historical renewable energy output data, construct a set of historical wind and solar power operation scenarios based on the historical renewable energy output data, and input each scenario in the set of historical wind and solar power operation scenarios into the pre-constructed power system optimization and dispatch model to obtain the system operation status;

[0073] In this embodiment, step 1 specifically includes:

[0074] Step 11: Obtain historical renewable energy output data, which includes historical wind power output data and photovoltaic power output data for power systems with a high proportion of renewable energy.

[0075] Step 12: Preprocess the historical renewable energy output data to obtain preprocessed historical renewable energy output data. The preprocessing includes: identifying and removing abnormal data points caused by factors such as metering anomalies, communication loss, and equipment shutdown; and completing the missing output data. The completion method is based on the unit's technical parameters, sunshine duration, wind speed statistical characteristics, or interpolation methods between adjacent time periods.

[0076] Step 13: Divide the hourly output sequence of the preprocessed historical renewable energy output data into operating days and construct a set of historical operating scenarios for wind and solar power. The wind output curve, solar output curve and photovoltaic output curve of each operating day in the historical operating scenario set together constitute an independent historical scenario, reflecting the volatility of renewable energy.

[0077] Step 14: Input each scenario from the historical wind and solar power operation scenario set into the pre-built power system optimization scheduling model daily, solve the power system optimization scheduling model, and obtain the system operation status under the corresponding scenario. The parameters of the system operation status include unit output, unit start-up and shutdown status, reserve dispatch, and line power flow level, etc.

[0078] In this embodiment, a power system optimization scheduling model is constructed, which includes power balance constraints, system reserve constraints, unit output limits, unit ramping constraints, and unit start-up and shutdown time constraints. This model simulates the feasible operating states of the system under different wind and solar power output scenarios and provides operational data support for risk assessment and scenario labeling.

[0079] Power system optimal dispatch models include:

[0080] The objective function of the power system optimal scheduling model is:

[0081] (1)

[0082] In the formula: Indicates system operating costs. This indicates the penalty cost for wind and solar power curtailment;

[0083] System operating costs for:

[0084] (2)

[0085] In the formula: function Indicates the unit At any moment The power generation cost function is assumed to be a piecewise function. , and They represent thermal power units The no-load cost, start-up cost, and downtime cost are calculated using variables. Indicates thermal power unit At any moment The start / stop status, among which This indicates that the unit is in operation; otherwise, it indicates that it is in shutdown status. and The binary decision variables represent the thermal power units respectively. At any moment The start-up and stop status. (Using...) The number of thermal power units is expressed by using This indicates the total number of time periods within the scheduling time range.

[0086] Wind and solar power curtailment penalty costs for:

[0087] (3)

[0088] (4)

[0089] In equation (3): the penalty cost consists of the penalty costs for wind power and photovoltaic curtailment, denoted as follows: and Its definition is shown in equation (4). and These represent the corresponding penalty cost coefficients. (Using...) and These represent the number of wind farms and photovoltaic power plants, respectively. In equation (4): and They represent time intervals respectively. Given the available wind power and available photovoltaic power; and These represent the total amount of wind power and solar power curtailment over the entire dispatch period, respectively. and They represent the times at time 1 and 2 respectively. The wind power and solar power output dispatch values.

[0090] The system operation constraints are shown in equations (5)-(23):

[0091] Dispatchable output range of wind farms and photovoltaic power plants:

[0092] Constraints (5) and (6) limit the dispatchable output range of wind farms and photovoltaic power plants, and their dispatchable output can only be less than or equal to the given available output value:

[0093] (5)

[0094] (6)

[0095] Thermal power unit constraints:

[0096] The piecewise linearization constraints for the output of thermal power units are shown in equations (7)-(8), where Indicates thermal power unit At any moment No. The effort of the segment; Indicates the unit The output range of each segment; Indicates the number of segments. Segmentation method and cost function The segmentation method is consistent, and the cost function is a monotonically increasing function, thus the cost can be conveniently expressed in a linear form. The output constraint of the thermal power unit (9) limits the unit. At any moment The power output, of which and They represent the generating units. The upper and lower limits of output;

[0097] (7)

[0098] (8)

[0099] (9)

[0100] Constraints (10)–(11) represent thermal power units Climbing restrictions, among which and These represent the downhill and uphill ramp rates of the generator unit, respectively. This represents the 30-minute ramp time on an hourly scheduling timescale. (Binary variable) and They represent the generating units. At any moment The shutdown and startup status;

[0101] (10)

[0102] (11)

[0103] Constraints (12)-(14) are described in a typical "three-variable (3-bin)" form to describe the unit start-up and shutdown logic:

[0104] (12)

[0105] (13)

[0106] (14)

[0107] Constraints (15) and (16) limit thermal power units The minimum power-on / power-off time, of which and These represent the minimum start-up time and minimum downtime of the unit, respectively:

[0108] (15)

[0109] (16)

[0110] Constraints (17) and (18) limit spinning reserve and operating reserve capacity. and They represent the generating units. At any moment Provided spinning reserve and operating reserve capacity. Spinning reserve ( ) is only available when the unit is running, while the standby ( ) is only available when the unit is shut down. Equations (17) and (18) guarantee that both are non-negative. and These represent 5 minutes and 15 minutes of climbing time, respectively.

[0111] (17)

[0112] (18)

[0113] System power balance constraints:

[0114] Constraint (19) ensures power balance at the system level, where Indicates load At any moment Requirements:

[0115] (19)

[0116] Network trend constraints:

[0117] Constraints (20) and (21) jointly limit the line power flow. For simplicity, the phase shift factor is not defined at the bus level, but the phase shift factors corresponding to the generator units, wind farms, photovoltaic power plants, and loads are directly used. , , and . Indicates the line Maximum transmission capacity;

[0118] (20)

[0119] ; (twenty one)

[0120] System backup constraints:

[0121] Constraints (22) and (23) specify the system's requirements for spinning reserve and operational reserve, respectively. and These represent the system-level spinning reserve and operational reserve requirements, respectively:

[0122] ; (twenty two)

[0123] ; (twenty three)

[0124] Step 2: Based on the system's operating status, construct flexibility indicators; quantify the flexibility indicators into a system operation risk index based on the analytic hierarchy process and entropy weight method, and use the system operation risk index as a system operation risk label to form a historical scenario library with system operation risk labels;

[0125] In this embodiment, the flexibility indicators include ramp-up margin, multiple backup margin, adjustable capacity margin, and line transmission margin.

[0126] (1) Uphill margin

[0127] Based on the scheduling results, i.e. the system operating status, equation (24) calculates the net load of the system. Equation (25) calculates the time intervals between two adjacent time periods ( and The ramping demand between ) is calculated using equation (26). Based on this, the total ramping demand of the system is calculated using equation (27), which is defined as the total ramping capacity of the power system minus the ramping demand.

[0128] ; (twenty four)

[0129] (25)

[0130] (26)

[0131] (27)

[0132] In the formula, Indicates the system at time 10:00 Net load; No. The load at time Load demand; Indicates the first Wind farms at all times Wind power output; Indicates the first A photovoltaic power plant at time Photovoltaic power output; Indicates from time At the time Net load ramp-up; Indicates time The need for upward climbing; Indicates the system at time 10:00 Uphill margin; This indicates the unit's ramp-up rate; and This indicates the 15-minute and 30-minute ramp times on an hourly scheduling timescale. Indicates the unit At any moment The device is in its startup state. Indicates thermal power unit At any moment The start / stop status; Indicates the unit The lower limit of output.

[0133] (2) Multiple reserve margins

[0134] Multiple redundancy margins include the capacity of 5-minute maximum available spinning reserve, 15-minute maximum available spinning reserve, and maximum available operational reserve; these metrics are used to measure the system's ability to quickly increase output and ensure system reliability in the event of a sudden power imbalance.

[0135] (28)

[0136] (29)

[0137] (30)

[0138] In the formula, Indicates the system at time 10:00 The maximum usable rotation time is 5 minutes; Indicates the system at time 10:00 The maximum available rotation time is 15 minutes; Indicates the system at time 10:00 Maximum available operational standby; and They represent the generating units. The upper and lower limits of output; and This indicates the 5-minute and 15-minute ramp times on an hourly scheduling timescale. Indicates thermal power unit At any moment contribution; Indicates thermal power unit At any moment The start / stop status.

[0139] (3) Adjustable capacity margin

[0140] Adjustable capacity margin represents the available margin for a thermal power unit to increase or decrease its output when the power system requires it; its upward and downward adjustable capacity are calculated by equations (31)-(32) respectively:

[0141] (31)

[0142] In the formula, Indicates the system at time 10:00 The lower adjustment capacity margin; Indicates the system at time 10:00 The upper adjustment capacity margin; Indicates thermal power unit At any moment contribution; and They represent the generating units. The upper and lower limits of output.

[0143] (4) Line transmission margin

[0144] Line transmission margin is used to assess the remaining capacity of a transmission line to carry additional power flow without violating thermal stability and safety constraints. Equation (32) calculates the line power flow based on the dispatch results; Equation (33) calculates the relative transmission margin. , represents the smallest relative boundary among all lines; Equation (34) calculates the absolute transmission margin. , represents the smallest absolute boundary among all lines;

[0145] (32)

[0146] (33)

[0147] (34)

[0148] In the formula, Indicates the line At any moment The absolute value of the line power flow represents the magnitude of the bidirectional power flow. Indicates the system at time 10:00 The relative line transmission margin; Indicates the system at time 10:00 The relative line transmission margin; , , and These represent the phase shift factors corresponding to the generating unit, wind farm, photovoltaic power station, and load, respectively. and They represent the times at time 1 and 2 respectively. The dispatch values ​​for wind power and solar power output; Indicates load At any moment The demand.

[0149] All the above flexibility indicators are based on an hourly time scale. Since the above flexibility indicators are all positive indicators (i.e., the larger the indicator value, the better the system performance), the minimum value within 24 hours can be taken to integrate the hourly indicators into a single indicator. Then, the flexibility indicators are normalized as shown in equation (35), where and Representing indicators The maximum and minimum values, Index for indicators:

[0150] (35)

[0151] This invention comprehensively considers the above four types of indicators. By measuring the system's adjustable space under different operating conditions, it characterizes the flexibility level of the power system under the condition of new energy fluctuations. This makes the calculation of the system operation risk index more comprehensive in subsequent risk quantification and provides a basic evaluation framework for subsequent operation risk quantification.

[0152] In this embodiment, as Figure 3 As shown, Figure 3 This is a schematic diagram illustrating the weight calculation relationship based on the Analytic Hierarchy Process (AHP) and the Entropy Weight Method (EWM) in Embodiment 1 of the present invention; the flexibility index is quantified into a system operation risk index based on the Analytic Hierarchy Process (AHP) and the Entropy Weight Method (EWM), including:

[0153] An AHP judgment matrix was constructed to assess the relative importance of each flexibility indicator. Pairwise comparisons were performed using a scale of 1 to 9, where 1 represents the lowest importance and 9 represents the highest importance, with higher scores indicating greater importance. The resulting judgment matrix is ​​shown in equation (36), where the AHP judgment matrix... for Matrix, containing One indicator, Indicates the first The flexibility index is relative to the first The importance ratio of each indicator. Represents the AHP judgment matrix Maximum eigenvalue The corresponding feature vector, Indicates the first The weights of each flexibility index are determined by normalizing the feature vectors, as shown in Equation (37).

[0154] (36)

[0155] (37)

[0156] In obtaining Then, the weighted maximum eigenvalue can be calculated using equation (38). .based on The consistency index CI can be further calculated, and the consistency ratio CR can be calculated by combining it with the random consistency index RI (obtained from the standard RI table). A consistency test is then performed based on the consistency ratio; if the conditions are met... If the consistency check is passed, then the AHP judgment matrix is ​​valid; otherwise, the consistency check is required. Make adjustments;

[0157] (38)

[0158] (39)

[0159] (40)

[0160] After determining the weights of the flexibility indicators, the entropy weights of each indicator are calculated based on the information entropy method according to the dispersion of each flexibility indicator, and the weights are then adjusted.

[0161] First, the AHP judgment matrix is ​​determined by equation (41). Normalization is performed to obtain a standardized matrix that reflects the relative performance levels of each flexibility index. .in, Still matrix, Indicates the first The flexibility index is relative to the first The standardized values ​​of the importance ratios of each flexibility index. Based on the information entropy theory, the entropy value and entropy weight of each flexibility index are calculated by equations (42) and (43), respectively, where and They represent the first Information entropy and entropy weight of each indicator;

[0162] , (41)

[0163] (42)

[0164] (43)

[0165] Subsequently, the comprehensive weights can be calculated using equation (44), that is, based on the corrected weights and the corresponding flexibility indicators, a system operation risk index is generated per operating day. System operation risk index It is obtained from equation (45);

[0166] (44)

[0167] (45)

[0168] In the formula, The system operation risk index; For the first The adjusted weights for each flexibility indicator; For the first The value of a flexibility indicator; The number of flexibility indicators; For the first The weighting of each flexibility indicator; For the first Entropy weight of a flexibility index.

[0169] The above technical solution calculates flexibility indicators based on pre-built flexibility indicators and system operating status. Based on each flexibility indicator, the Analytic Hierarchy Process (AHP) and Entropy Weight Method (EWM) are used to integrate the flexibility indicators, thereby constructing a unified system operation risk index for this historical wind and solar power output scenario.

[0170] In this embodiment, the system operation risk index is used as the system operation risk label to form a historical scenario library with system operation risk labels, specifically:

[0171] The system operation risk index is normalized and mapped to historical wind and solar power output scenarios, so that different historical wind and solar power output scenarios correspond one-to-one with their operation risk levels, forming a set of historical wind and solar power output scenarios containing system operation risk labels, i.e., a historical scenario library with system operation risk labels.

[0172] Step 3: Construct a scenario generation model based on conditional generative adversarial networks. Based on the historical scenario library, train the scenario generation model with system operation risk labels as conditional variables to obtain a scenario generation model with adjustable operation risk for wind and solar power output.

[0173] In this embodiment, the scene generation model includes a generator and a discriminator, such as... Figure 4As shown, both the generator and discriminator are constructed from two neural networks, with corresponding weight parameters denoted as follows: , .

[0174] The neural network of the generator can be compactly represented by equation (46), where, This represents a sequence of wind power and solar power output scenarios generated by the generator; This represents Gaussian random noise; These are weight parameters; For the mapping function of the generator network;

[0175] (46)

[0176] The discriminator's neural network can be expressed as equation (47), where, This represents the probability that the input scene belongs to the set of real-world historical operating scenes; This indicates the input scenario, which can be a real historical scenario or a scenario generated by the generator. These are weight parameters; For the discriminator network, the mapping function is used.

[0177] (47)

[0178] To enable the model to learn features under specific conditions, a conditional variable, namely the system operation risk index, is introduced into the operational scenario generation model. The conditional label corresponding to the power system operation risk index, obtained by fusing the proposed flexibility index using the AHP-EWM method, is then used. Simultaneously input generator and discriminator The generator generates corresponding wind and solar power output scenarios under given conditions, while the discriminator evaluates whether the generated wind and solar power output scenarios are both realistic and consistent with the specified operational risk level.

[0179] The loss functions of the generator and discriminator are shown in equations (48)–(49), where This represents the expectation operator. By simultaneously training the generator and discriminator, which form an adversarial relationship, the game objective function is... As shown in equation (50). During training, the generator and discriminator continuously update their respective network parameters to generate scenes. The difference between the generated and real-world scenarios gradually decreases, while the discriminator's ability to distinguish between real and generated scenarios continuously improves. By minimizing the deviation in statistical characteristics between generated and real samples, the network gradually learns the probability distribution mapping relationship of wind power and photovoltaic output time series under different risk conditions, and finally the two networks reach a Nash equilibrium state.

[0180] (48)

[0181] (49)

[0182] (50)

[0183] In the formula, and These represent the losses of the generator and the discriminator, respectively. and These represent real-world and generated scene samples, respectively. Represents a condition variable; Represents the expected value of the distribution; express From the distribution of generated data The expected value obtained from sampling; express From the distribution of generated data The expected value obtained from sampling; This represents the objective function of the adversarial game between the generator and the discriminator.

[0184] To further improve the performance of the conditional generation network, a gradient penalty term is introduced to prevent gradient explosion and gradient vanishing problems, thereby enhancing training stability and convergence speed, enabling the model to learn the complex distribution patterns of wind and solar power output scenarios under given risk conditions. The improved objective function after adding the gradient penalty term is shown in equation (51):

[0185] (51)

[0186] In the formula, Represents the gradient penalty coefficient. This represents a sample obtained by random interpolation along the line connecting the real sample and the generated sample.

[0187] After training, any given system operation risk index and random noise are input into the conditional generation network, which can automatically generate wind power and solar power output scenarios corresponding to that risk condition. The generated scenarios can characterize the typical temporal characteristics and fluctuation patterns of wind and solar resources under the target risk level, thereby realizing the generation of diversified, controllable, and operationally risk-adjustable wind power and solar power output scenarios for high-proportion renewable energy power systems.

[0188] like Figure 5 and Figure 6 As shown, Figure 5 This is a graph showing the loss function during the training process; Figure 6 This is a graph showing the discriminator score during the training process.

[0189] Step 4: Based on the wind and solar power output scenario generation model with adjustable operational risk, generate wind and solar power output scenarios according to the specified risk level.

[0190] In this embodiment, based on the wind and solar power output scenario generation model with adjustable operational risk obtained in step 3, the system operation risk label is explicitly introduced into the generator as a condition constraint to construct a "risk perception" scenario generation mechanism. This enables the accurate generation of wind and solar power output scenarios according to the specified risk level, so that the generation results maintain statistical authenticity while having risk controllability and adjustability, providing a brand-new risk-driven scenario generation tool for the operation analysis of high-proportion new energy systems.

[0191] like Figure 7 and Figure 8 As shown, Figure 7 A comparison map of the distribution of wind power output scenarios; Figure 8 A comparison map showing the distribution of photovoltaic power generation scenarios. Figure 7 and Figure 8 This indicates a high degree of overlap in distribution between the real-world and generated scenarios. The KDE curve, used to characterize a continuous probability distribution, shows strong consistency between the KDE curves of the two scenarios. The results demonstrate that, after training, the generator can effectively capture the statistical characteristics of real-world scenarios and generate wind power output scenarios with high realism.

[0192] The above technical solution generates corresponding new energy output scenarios based on any specified system operation risk coefficient. By inputting risk condition vectors and random noise into the conditional generative adversarial network, it realizes the generation of wind and solar power output sequences at different operation risk levels. The generated scenarios can reflect the typical fluctuation characteristics under the target risk level and the pressure that may be caused to the system operation.

[0193] This invention constructs an optimized power system dispatch model and a flexibility index system including ramp-up margin, multiple reserve margin, adjustable capacity margin, and line transmission margin. It employs the analytic hierarchy process (AHP) and entropy weighting method to fuse subjective and objective weights, forming a risk quantification method capable of characterizing system operational pressure. Based on historical wind and solar power output data, a scenario set is constructed. Each scenario is input into the constructed optimized dispatch model to obtain the system operating status, and the operational risk coefficient is calculated using the flexibility index system, thus forming a historical scenario library with risk labels. On this basis, a conditional generative adversarial network (GAN) with gradient penalty terms is constructed. The operational risk label is used as a conditional variable to train the model, establishing a risk-adjustable generation model capable of generating wind and solar power output scenarios based on a target risk coefficient. This invention achieves the ability to characterize operational risks based on dispatch results and generate scenarios with different risk levels on demand, providing an effective technical means for simulating high-risk operating states of power systems.

[0194] Example 2

[0195] like Figure 9 As shown, the difference between this embodiment and Embodiment 1 is that this embodiment provides a wind and solar power output scenario generation system with adjustable operational risk, which corresponds one-to-one with the wind and solar power output scenario generation method with adjustable operational risk in Embodiment 1; the system includes:

[0196] The acquisition unit is used to acquire historical renewable energy output data.

[0197] Historical scenario construction unit, used to construct a set of historical operation scenarios for wind and solar power based on historical renewable energy output data;

[0198] The scheduling model unit is used to input the historical operation scenarios of wind and solar power into a pre-built power system optimization scheduling model to obtain the system operation status.

[0199] The flexibility metric construction unit is used to construct flexibility metrics based on the system's operating status.

[0200] The risk quantification unit is used to quantify the flexibility index into the system operation risk index based on the analytic hierarchy process and the entropy weight method, and to use the system operation risk index as the system operation risk label to form a historical scenario library with system operation risk labels.

[0201] Conditional Generative Adversarial Network (GAN) building blocks are used to construct runtime scenario generation models based on conditional generative adversarial networks (GANs).

[0202] The scene generation model training unit is used to train the operation scene generation model based on the historical scene library, with the system operation risk label as a condition variable, to obtain the operation risk adjustable wind and solar power output scene generation model.

[0203] The operation scenario generation unit is used to generate wind power and solar power output scenarios according to a specified risk level based on the wind and solar power output scenario generation model with adjustable operation risk.

[0204] As a further implementation, the flexibility index is quantified into a system operation risk index based on the analytic hierarchy process (AHP) and entropy weight method, including:

[0205] With the first The flexibility index is relative to the first The importance ratio of each flexibility indicator Construct an AHP judgment matrix using matrix elements;

[0206] Calculate the largest eigenvalue of the AHP judgment matrix and the weights of each flexibility index; and calculate the weighted largest eigenvalue based on the weights of each flexibility index.

[0207] Calculate the consistency ratio of the consistency index based on the weighted maximum eigenvalue; perform a consistency test based on the consistency ratio; if the consistency test is not met, adjust the AHP judgment matrix; if the consistency test is met, determine the weights of each flexibility index.

[0208] Based on the entropy weight method, the entropy weight of each flexibility index is calculated according to the degree of dispersion of each flexibility index, and the weight of each flexibility index is corrected according to the entropy weight to obtain the corrected weight.

[0209] Based on the revised weights and corresponding flexibility indicators, a system operation risk index is generated on a daily basis.

[0210] The execution process of each unit can be carried out according to the steps of the method for generating wind and solar power output scenarios with adjustable operation risk in Embodiment 1, and will not be described in detail in this embodiment.

[0211] Meanwhile, the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned method for generating wind and solar power output scenarios with adjustable operational risks.

[0212] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0213] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0214] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0215] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0216] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., 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 generating wind and solar power output scenarios with adjustable operational risks, characterized in that, The method includes: Historical renewable energy output data is obtained, and a set of historical wind and solar power operation scenarios is constructed based on the historical renewable energy output data. Each scenario in the set of historical wind and solar power operation scenarios is input into a pre-constructed power system optimization and scheduling model to obtain the system operation status. Based on the system's operating status, a flexibility index is constructed; the flexibility index is quantified into a system operating risk index based on the analytic hierarchy process and the entropy weight method, and the system operating risk index is used as a system operating risk label to form a historical scenario library with system operating risk labels. A scenario generation model based on conditional generative adversarial networks is constructed. The scenario generation model is trained using the system operation risk label as a condition variable based on the historical scenario library, resulting in a scenario generation model with adjustable operation risk for wind and solar power output. Based on the aforementioned wind and solar power output scenario generation model with adjustable operational risk, wind and solar power output scenarios are generated according to a specified risk level.

2. The method for generating wind and solar power output scenarios with adjustable operational risks according to claim 1, characterized in that, Historical renewable energy output data is obtained, and a set of historical wind and solar power operation scenarios is constructed based on this data. Each scenario in the historical wind and solar power operation scenario set is then input into a pre-constructed power system optimization and dispatch model to obtain the system operating status, including: Obtain historical renewable energy output data, including historical wind power output data and photovoltaic power output data; The historical renewable energy output data is preprocessed to obtain preprocessed historical renewable energy output data; The hourly output sequence of the preprocessed historical renewable energy output data is divided by operating day and constructed into a set of historical wind and solar operating scenarios; the wind output curve, solar output curve and photovoltaic output curve of each operating day in the set of historical wind and solar operating scenarios together constitute an independent historical scenario; Each scenario in the historical wind and solar operation scenario set is input into a pre-constructed power system optimization scheduling model on a daily basis. The power system optimization scheduling model is then solved to obtain the system operation status under the corresponding scenario.

3. The method for generating wind and solar power output scenarios with adjustable operational risks according to claim 2, characterized in that, The parameters of the system's operating status include unit output, unit start-up and shutdown status, standby dispatch, and line power flow level.

4. The method for generating wind and solar power output scenarios with adjustable operational risks according to claim 1, characterized in that, The flexibility metrics include ramp-up margin, multiple redundancy margin, adjustable capacity margin, and line transmission margin. The uphill margin is the total uphill capacity of the power system minus the uphill demand. The multiple redundancy margins include the capacity of 5-minute maximum available spinning reserve, 15-minute maximum available spinning reserve, and maximum available operational reserve. The adjustable capacity margin is the available capacity of a thermal power unit to increase or decrease its output when the power system requires it. The line transmission margin is an assessment of the remaining capacity of a transmission line to carry additional power flow without violating thermal stability and safety constraints.

5. The method for generating wind and solar power output scenarios with adjustable operational risks according to claim 1, characterized in that, The flexibility index is quantified into a system operation risk index based on the analytic hierarchy process (AHP) and entropy weight method, including: With the first The flexibility index is relative to the first The importance ratio of each flexibility indicator Construct an AHP judgment matrix using matrix elements; Calculate the maximum eigenvalue of the AHP judgment matrix and the weights of each flexibility index; and calculate the weighted maximum eigenvalue based on the weights of each flexibility index. Based on the weighted maximum eigenvalue, calculate the consistency ratio of the consistency index; perform a consistency check based on the consistency ratio; if the consistency check is not met, adjust the AHP judgment matrix; if the consistency check is met, determine the weight of each of the flexibility indices. Based on the entropy weight method, the entropy weight of each flexibility index is calculated according to the degree of dispersion of each flexibility index, and the weight of each flexibility index is corrected according to the entropy weight to obtain the corrected weight. Based on the revised weights and corresponding flexibility indicators, a system operation risk index is generated on a daily basis.

6. The method for generating wind and solar power output scenarios with adjustable operational risks according to claim 5, characterized in that, The formula for calculating the system operation risk index is as follows: ； ； In the formula, The system operation risk index; For the first The adjusted weights for each flexibility indicator; For the first The value of a flexibility indicator; The number of flexibility indicators; For the first The weighting of each flexibility indicator; For the first Entropy weight of a flexibility index.

7. The method for generating wind and solar power output scenarios with adjustable operational risks according to claim 1, characterized in that, The running scenario generation model includes a generator and a discriminator, both of which are constructed from two neural networks; The neural network expression for the generator is: ,in, This represents a sequence of wind power and solar power output scenarios generated by the generator; This represents Gaussian random noise; These are weight parameters; For the mapping function of the generator network; The neural network expression for the discriminator is: ,in, This represents the probability that the input scene belongs to the set of real-world historical operating scenes; This indicates the input scenario, which can be a real historical scenario or a scenario generated by the generator. These are weight parameters; For the discriminator network, the mapping function is used. A conditional variable, namely the system operation risk index, is introduced into the operation scenario generation model. The conditional label corresponding to the system operation risk index is simultaneously input into the generator and the discriminator. The generator generates corresponding wind power and photovoltaic output scenarios under given conditions, while the discriminator evaluates whether the generated wind power and photovoltaic output scenarios are both realistic and consistent with the specified operation risk level.

8. A wind and solar power output scenario generation system with adjustable operational risk, characterized in that, The system includes: The acquisition unit is used to acquire historical renewable energy output data. The historical scenario construction unit is used to construct a set of historical wind and solar power operation scenarios based on the historical new energy output data. The scheduling model unit is used to input the historical operation scenarios of wind and solar power into a pre-built power system optimization scheduling model to obtain the system operation status. A flexibility index construction unit is used to construct flexibility indices based on the system's operating status. The risk quantification unit is used to quantify the flexibility index into a system operation risk index based on the analytic hierarchy process and the entropy weight method, and to use the system operation risk index as a system operation risk label to form a historical scenario library with system operation risk labels. Conditional Generative Adversarial Network (GAN) building blocks are used to construct runtime scenario generation models based on conditional generative adversarial networks (GANs). The scene generation model training unit is used to train the operation scene generation model based on the historical scene library and with the system operation risk label as a condition variable, to obtain an operation risk adjustable wind and solar power output scene generation model. The operation scenario generation unit is used to generate wind power and photovoltaic output scenarios according to a specified risk level based on the operation risk adjustable wind and solar power output scenario generation model.

9. The wind and solar power output scenario generation system with adjustable operational risk according to claim 8, characterized in that, The flexibility index is quantified into a system operation risk index based on the analytic hierarchy process (AHP) and entropy weight method, including: With the first The flexibility index is relative to the first The importance ratio of each flexibility indicator Construct an AHP judgment matrix using matrix elements; Calculate the maximum eigenvalue of the AHP judgment matrix and the weights of each flexibility index; and calculate the weighted maximum eigenvalue based on the weights of each flexibility index. Based on the weighted maximum eigenvalue, calculate the consistency ratio of the consistency index; perform a consistency check based on the consistency ratio; if the consistency check is not met, adjust the AHP judgment matrix; if the consistency check is met, determine the weight of each of the flexibility indices. Based on the entropy weight method, the entropy weight of each flexibility index is calculated according to the degree of dispersion of each flexibility index, and the weight of each flexibility index is corrected according to the entropy weight to obtain the corrected weight. Based on the revised weights and corresponding flexibility indicators, a system operation risk index is generated on a daily basis.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements a method for generating wind and solar power output scenarios with adjustable operational risks as described in any one of claims 1 to 7.

Images