Dynamic aggregation method of virtual power plant considering reliability and flexibility of new energy

Abstract

This invention belongs to the field of smart grid technology, specifically relating to a dynamic aggregation method for virtual power plants that considers the reliability and flexibility of new energy sources. The steps include: collecting output data from wind farms and photovoltaic power plants, as well as load data from the distribution-side power system under the jurisdiction of the virtual power plant, and performing data preprocessing; generating time-series operation scenarios in batches using Monte Carlo simulation; evaluating the current performance of the virtual power plant based on the sequential Monte Carlo method; constructing a dynamic aggregation model based on the evaluation results, with the objective function of minimizing the expected power shortage; solving the dynamic aggregation model using the IFLRO algorithm, obtaining the optimal combination scheme of wind farms and photovoltaic power plants through iterative calculation, and outputting the optimal dynamic aggregation scheme; optimizing and adjusting the dynamic aggregation scheme based on verification and actual operation results. This invention can effectively reduce the risk of power outages, reduce wind and solar curtailment, and improve the power supply stability and flexible adjustment capabilities of virtual power plants.

Description

Technical Field

[0001] This invention belongs to the field of smart grid technology, specifically relating to a method for dynamic aggregation of virtual power plants that takes into account the reliability and flexibility of new energy sources. Background Technology

[0002] With the acceleration of energy transition and the advancement of new energy technologies, virtual power plants (VPPs) are becoming increasingly important as a key innovation in modern power systems. By integrating distributed energy resources such as wind power, solar power, and energy storage, VPPs form a flexible and efficient energy aggregator, aiming to improve the reliability and operational efficiency of the power grid system. However, the intermittency and uncertainty of new energy sources pose significant challenges to the stable operation of VPPs. How to accurately assess the reliability and flexibility of VPPs and achieve dynamic optimization aggregation has become a core issue in ensuring their effective operation.

[0003] Traditional virtual power plant assessment and aggregation technologies have several limitations. First, existing methods often focus on evaluating a single energy type or a single indicator, making it difficult to comprehensively assess virtual power plants that integrate multiple types of renewable energy. Second, the volatility of renewable energy output causes frequent power fluctuations and insufficient ramp-up rates in virtual power plants, which traditional methods struggle to address, severely hindering the accurate assessment and effective control of virtual power plant operation. Third, current aggregation methods are ill-suited to the complex and ever-changing scenarios of virtual power plants, failing to fully tap the potential of renewable energy and support optimized decision-making. These problems significantly impede the full realization of the effectiveness of virtual power plants. Summary of the Invention

[0004] In view of the shortcomings of the prior art, the purpose of this invention is to provide a dynamic aggregation method for virtual power plants that considers the reliability and flexibility of new energy sources. It relies on sequential Monte Carlo to complete multi-dimensional quantitative assessment of reliability and flexibility, constructs an aggregation model with the goal of minimizing the expected value of insufficient power, and optimizes the combination of wind and solar resources through the IFLRO algorithm. This can effectively reduce the risk of power outages, reduce wind and solar curtailment, and improve the power supply stability and flexible adjustment capability of virtual power plants.

[0005] To achieve the above objectives, this invention provides a method for dynamic aggregation of virtual power plants that considers the reliability and flexibility of new energy sources, comprising the following steps: S1. Collect power output data from wind farms and photovoltaic power plants, as well as load data from the distribution-side power system under the jurisdiction of the virtual power plant, and perform data preprocessing. S2. Using the Monte Carlo simulation method, the virtual power plant is simulated and analyzed under all-time operating conditions. Combined with random changes in weather and equipment failure probability, the random fluctuations of wind power and photovoltaic output are characterized, and time-series operation scenarios are generated in batches. S3. Based on the simulation analysis results, calculate the reliability and flexibility indicators, and evaluate the current performance of the virtual power plant based on the sequential Monte Carlo method. S4. Based on the evaluation results, construct a dynamic aggregation model with the objective function of minimizing the expected value of insufficient power, and set its constraints. S5. Improve the judgment conditions for light reflection and refraction in the meshless ray optimization algorithm to form the improved meshless ray optimization algorithm IFLRO. Solve the dynamic aggregation model based on the IFLRO algorithm, obtain the optimal combination scheme of wind farm and photovoltaic power station through iterative calculation, and output the optimal dynamic aggregation scheme. S6. Validate the dynamic aggregation scheme using reliability and flexibility indicators, and optimize and adjust the dynamic aggregation scheme based on the validation results. S7. Apply the optimized and adjusted dynamic aggregation scheme to the actual operation of the virtual power plant, perform real-time monitoring and adjustment, update the dynamic aggregation model parameters based on the regularly collected real-time data, and re-validate the current dynamic aggregation scheme using reliability and flexibility indicators, continuously optimizing and adjusting the dynamic aggregation scheme.

[0006] As a preferred embodiment of the present invention, in S1, historical and real-time output data of each wind farm and photovoltaic power station included in the aggregation scope of the virtual power plant are collected, and measured and predicted data of various user loads in the power distribution side of the virtual power plant are obtained simultaneously, and supplementary information including wind farm parameters, photovoltaic power station parameters, geographical location and meteorological environment data is added. The collected data is preprocessed by cleaning out abnormal noise and repairing missing or erroneous data, followed by data smoothing and standardization to unify the data format and units, forming a well-organized dataset.

[0007] As a preferred embodiment of the present invention, in S3, the reliability index includes the probability of insufficient power time, the expected value of insufficient power time, and the expected value of insufficient power; the flexibility index includes the probability of insufficient upward generation capacity, the expected value of insufficient upward generation capacity, the probability of insufficient downward generation capacity, the expected value of insufficient downward generation capacity, the probability of insufficient upward ramp rate, the expected value of insufficient upward ramp rate, the probability of insufficient downward ramp rate, the expected value of insufficient downward ramp rate, the probability of insufficient upward flexibility, the expected value of insufficient upward flexibility, the probability of insufficient downward flexibility, and the expected value of insufficient downward flexibility.

[0008] As a preferred embodiment of the present invention, in step S3, the current performance of the virtual power plant is evaluated based on the sequential Monte Carlo method, and the process is as follows: S3.1 Input the output data of wind farms and photovoltaic power stations; S3.2 Simulate the annual output sequence of each wind farm and photovoltaic power station; S3.3 Utilize the failure probability of wind and solar components in the virtual power plant to correct the annual output sequence of each wind farm and photovoltaic power station; S3.4 Input the annual load data of the power distribution system under the jurisdiction of the virtual power plant, calculate the reliability index of each wind farm and photovoltaic power station, and comprehensively compile the reliability index of the virtual power plant. S3.5 Determine whether the set lifespan M1 of the virtual power plant reliability assessment has been reached. If yes, output the virtual power plant reliability index; otherwise, return to S3.3. S3.6 Calculate the flexibility index of each wind farm and photovoltaic power station at this time and compile the virtual power plant flexibility index. S3.7 Determine whether the set lifespan M2 for the virtual power plant flexibility assessment has been reached. If yes, output the virtual power plant flexibility index; otherwise, return to S3.3. S3.8 Output virtual power plant reliability indicators and virtual power plant flexibility indicators.

[0009] As a preferred embodiment of the present invention, in S4, the objective function of the dynamic aggregation model is: ; In the formula, For dynamic aggregation scheme The corresponding objective function value; This is the power shortage expectation function, used to calculate the power shortage expectation value of a virtual power plant under a given dynamic aggregation scheme; A 0-1 variable representing whether the u-th wind farm or photovoltaic power station participates in the aggregation. Indicates participation in aggregation, M indicates non-participation; M represents the total number of wind farms and photovoltaic power plants that can participate in aggregation. This represents the equivalent reliable capacity of the u-th wind farm or photovoltaic power station, calculated based on reliability assessment. Represents dynamic aggregation scheme The equivalent reliable capacity of the virtual power plant; The constraints of the dynamic aggregation model include power system power balance constraints, wind power output unit operation constraints, photovoltaic power output unit operation constraints, and wind and solar curtailment rate constraints.

[0010] As a preferred embodiment of the present invention, in S5, the IFLRO algorithm specifically maps each dynamic aggregation scheme to a point in the search space. Each dimension corresponds to a decision variable for a wind farm or a solar power station, starting from the current point. Depart, in the direction propagation reference step size Arrive at the new point Then, with a step size of 0.1 Generate test points in each dimension Calculate the EENS value of the dynamic aggregation scheme corresponding to each test point, and use it as the fitness value of that test point. This fitness value corresponds to the virtual speed of light propagation. The optimal direction of the probe point is determined based on the fitness value, the propagation direction is updated, and then the continuous coordinate components are converted into 0-1 decision variables by threshold judgment. The iteration is repeated until the termination condition is met, and the optimal dynamic aggregation scheme is output, including the selection of wind farms and photovoltaic power plants participating in the aggregation and suggestions for energy storage configuration.

[0011] As a preferred embodiment of the present invention, the solution steps of the dynamic aggregation model in S5 are as follows: S5.1 Input the output data of wind farms and photovoltaic power plants, as well as the load data of the power distribution system under the jurisdiction of the virtual power plant, and perform data preprocessing, and input the constraints of the dynamic aggregation model; S5.2. Set the initial solution point, initial propagation direction vector, reference step size, and iteration termination condition for the IFLRO algorithm; S5.3. The optimization objective is to minimize the expected value of insufficient power corresponding to the dynamic aggregation scheme, and this is used as the fitness function of the algorithm. S5.4. Use the IFLRO algorithm to iteratively solve the dynamic aggregation model and output the optimal combination scheme of wind farm and photovoltaic power station and energy storage configuration suggestions.

[0012] The beneficial effects of this invention are: This invention significantly improves the power supply reliability and flexible control capabilities of virtual power plants by constructing a comprehensive reliability and flexibility evaluation index system and employing an improved FLRO algorithm to optimize the dynamic aggregation scheme. In terms of reliability, this invention can accurately quantify power outage risks, providing a scientific basis for the optimized configuration of energy storage devices, effectively reducing power outages, and ensuring the continuity and stability of power supply.

[0013] In terms of flexibility, this invention focuses on ramp-up capability and spinning reserve, establishing a complete flexibility index system to enable virtual power plants to adapt to load changes more efficiently and improve power system operating efficiency. Simultaneously, by optimizing the combination of wind farms and photovoltaic power plants through a dynamic aggregation method, the impact of renewable energy output fluctuations on power system stability is reduced, renewable energy absorption is increased, and wind and solar curtailment is reduced.

[0014] Based on reliability and flexibility assessments, this invention achieves rational resource allocation, avoids resource waste, and improves resource utilization efficiency. Overall, this invention not only improves the operational efficiency and economic benefits of virtual power plants but also promotes the widespread application of new energy sources and facilitates the sustainable development of the power system. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the principle of this invention. Detailed Implementation

[0016] The embodiments of the present invention will be further described below with reference to the accompanying drawings: Example 1: As Figure 1 As shown, the virtual power plant dynamic aggregation method considering the reliability and flexibility of new energy sources includes the following steps: S1. Collect power output data from wind farms and photovoltaic power plants, as well as load data from the distribution-side power system under the jurisdiction of the virtual power plant, and perform data preprocessing. S2. Using the Monte Carlo simulation method, the virtual power plant is simulated and analyzed under all-time operating conditions. Combined with random changes in weather and equipment failure probability, the random fluctuations of wind power and photovoltaic output are characterized, and time-series operation scenarios are generated in batches. S3. Based on the simulation analysis results, calculate the reliability and flexibility indicators, and evaluate the current performance of the virtual power plant based on the sequential Monte Carlo method. S4. Based on the evaluation results, construct a dynamic aggregation model with the objective function of minimizing the expected value of insufficient power, and set its constraints. S5. Improve the judgment conditions for light reflection and refraction in the meshless ray optimization algorithm to form the improved meshless ray optimization algorithm IFLRO. Solve the dynamic aggregation model based on the IFLRO algorithm, obtain the optimal combination scheme of wind farm and photovoltaic power station through iterative calculation, and output the optimal dynamic aggregation scheme. S6. Validate the dynamic aggregation scheme using reliability and flexibility indicators, and optimize and adjust the dynamic aggregation scheme based on the validation results. S7. Apply the optimized and adjusted dynamic aggregation scheme to the actual operation of the virtual power plant, perform real-time monitoring and adjustment, update the dynamic aggregation model parameters based on the regularly collected real-time data, and re-validate the current dynamic aggregation scheme using reliability and flexibility indicators, continuously optimizing and adjusting the dynamic aggregation scheme.

[0017] Light Ray Optimization (LRO) is an intelligent optimization algorithm based on the mechanisms of light propagation, reflection, and refraction. Light Ray Optimization Based on Grid Free Method (FLRO) is a gridless search method built upon LRO. FLRO relies on the physical mechanisms of light propagation, reflection, and refraction to achieve optimization, and abandons the traditional grid partitioning method, adopting a gridless search strategy to perform global optimization, effectively improving search flexibility and solution efficiency.

[0018] The IFLRO algorithm is based on the FLRO algorithm and incorporates the multidimensional 0-1 decision variable characteristics of the virtual power plant dynamic aggregation model. It improves the judgment conditions for light reflection and refraction to enhance the algorithm's adaptability and solution efficiency for wind farm and photovoltaic power plant combination optimization problems.

[0019] In S1, historical and real-time output data of each wind farm and photovoltaic power station included in the aggregation scope of the virtual power plant are collected. Simultaneously, measured and predicted data of various user loads in the power distribution side of the virtual power plant are obtained, and supplementary information including wind farm parameters, photovoltaic power station parameters, geographical location, and meteorological environment data is added. The collected data is preprocessed by cleaning out abnormal noise and repairing missing or erroneous data, followed by data smoothing and standardization to unify the data format and units, forming a well-organized dataset.

[0020] During preprocessing, data cleaning operations are first performed on the collected power output, user load, and auxiliary information of the power station (wind farm or photovoltaic power station). The discriminant method identifies and removes abnormal noise data that exceeds the normal fluctuation range. Linear interpolation is used to fill in missing data caused by time period gaps or acquisition failures. Then, a moving average filtering algorithm is used to smooth the time series data and eliminate instantaneous random jump interference. Finally, the min-max standardization method is used to normalize and scale various output, load, and meteorological parameters of different dimensions, unifying the storage format, time resolution, and physical dimensions of all data. This results in a standardized and regularized dataset that is free of missing data, anomalies, has a uniform scale, and is time-series continuous.

[0021] In S2, the Monte Carlo simulation method is used to conduct full-time operation simulation analysis of the virtual power plant. First, historical meteorological time series data, equipment failure rate and repair time parameters of each wind farm and photovoltaic power station are imported. Based on the random sampling mechanism, the meteorological random variation law of sunlight and wind speed and the random failure and shutdown events of the units are characterized. The actual installed capacity on the nameplate of the station is linked at the same time. Real-time output sequence of new energy sources affected by meteorological fluctuations and equipment shutdowns is generated for each time period. The time series operation scenarios covering the entire annual operation cycle and containing multiple sets of differentiated output characteristics are output in batches, providing complete basic operating condition sample data for subsequent reliability index calculation, equivalent reliable capacity conversion and dynamic aggregation optimization model solution.

[0022] In S3, reliability metrics include the probability of insufficient power time, the expected value of insufficient power time, and the expected value of insufficient power supply; flexibility metrics include the probability of insufficient upward generation capacity, the expected value of insufficient upward generation capacity, the probability of insufficient downward generation capacity, the expected value of insufficient downward generation capacity, the probability of insufficient upward ramp rate, the expected value of insufficient upward ramp rate, the probability of insufficient downward ramp rate, the expected value of insufficient downward ramp rate, the probability of insufficient upward flexibility, the expected value of insufficient upward flexibility, the probability of insufficient downward flexibility, and the expected value of insufficient downward flexibility.

[0023] Assessing the reliability and flexibility of virtual power plants is crucial for ensuring their stable operation and efficient dispatch. With the large-scale integration of new energy sources, virtual power plants integrate distributed energy resources such as wind power, solar power, and energy storage to form a flexible and efficient energy conglomerate. However, the intermittency and uncertainty of new energy sources pose significant challenges to the stable operation of virtual power plants. Therefore, accurate assessment of the reliability and flexibility of virtual power plants is essential.

[0024] Reliability assessment primarily focuses on the ability of virtual power plants to meet load demands. A series of metrics are used to quantify the likelihood and severity of power system outages, including the Probability of Time Without Power (LOLP), Expected Time Without Power (LOLE), and Expected Energy Without Power (EENS). These metrics reflect the outage risk of the virtual power plant from different perspectives; for example, LOLP measures the probability of a power shortage, while EENS comprehensively expresses information such as the number of outages, duration, and power output.

[0025] Flexibility assessment focuses on the virtual power plant's ability to adapt to and regulate load changes. This involves the characteristics of net load fluctuations, as well as indicators such as the power system's ramp-up capability and spinning reserve in response to these fluctuations. For example, indicators such as the probability of insufficient generation capacity during upward and downward movements and the probability of insufficient ramp-up rates can comprehensively evaluate the flexibility performance of the virtual power plant under different operating conditions.

[0026] To accurately assess the reliability and flexibility of virtual power plants, evaluation techniques such as Monte Carlo simulation are commonly employed. Monte Carlo simulation, by simulating numerous operating scenarios, effectively captures the randomness and uncertainty of renewable energy output, thus providing a scientific basis for the operation and scheduling of virtual power plants. Through reliability assessment, a reasonable capacity configuration scheme for energy storage batteries can be formulated, ensuring that the virtual power plant can meet load demands while possessing good ramp-up capabilities and sufficient spinning reserve to adapt to steeper ramp-up events and load descents. This evaluation method not only improves the power supply reliability of virtual power plants but also enhances their flexible control capabilities, laying a solid foundation for their widespread application in modern power systems.

[0027] In reliability metrics, the probability of power shortage time is calculated as follows: ; In the formula, LOLP is the probability of insufficient power time; z is the system outage capacity status number; The system is in a state of outage capacity. The probability of; The system is in a state of outage capacity. The ratio of the duration of the outage to the total duration of the current state; when the unit capacity does not meet the load requirements, this indicator can determine the probability of power outage time in the power system, but does not consider the magnitude of the outage.

[0028] The expected value of power shortage time is calculated as follows: ; In the formula, LOLE represents the expected duration of power shortage; m represents the number of time periods in a year; Let i be the number of days in the i-th time period; This represents the peak load on day j within the i-th time period; X represents the system's installed capacity during the i-th time period; X represents the system's downtime capacity. This represents the probability that the outage capacity on day j within the i-th time period is greater than or equal to the reserve capacity. This indicator can determine the probability that the outage capacity of the power system is greater than or equal to the reserve capacity.

[0029] The calculation method for the expected value of insufficient battery power is as follows: ; In the formula, EENS represents the expected value of insufficient power. The probability that the shutdown capacity is greater than or equal to X in the Kth hour of the jth day within the i-th time period; Let K be the hourly load on day j within time period i. This indicator represents the expected reduction in power supply to users due to forced unit shutdowns, and comprehensively expresses the number of outages, average duration, and average outage power.

[0030] In the flexibility index, the probability of insufficient upward generation capacity is calculated as follows: ; In the formula, Let t be the probability that the upward power generation capacity is insufficient at time t; The system has an available capacity for upscaling at time t; , respectively at t and Net load at any given time; , These represent the maximum and actual output of the unit at time t, respectively. For time intervals; This represents the probability of the event occurring.

[0031] The calculation method for insufficient upward power generation capacity is as follows: ; In the formula, The expected value of insufficient upward power generation capacity at time t; The insufficient power generation capacity at time t; The probability of insufficient downstream power generation capacity is calculated as follows: ; In the formula, Let t be the probability of insufficient power generation capacity at time t. This represents the minimum output of the unit at time t; The calculation method for the expected value of insufficient downstream power generation capacity is as follows: ; In the formula: The expected value of insufficient power generation capacity at time t.

[0032] The probability of insufficient uphill climbing speed is calculated as follows: ; In the formula, Let t be the probability that the upward climbing rate is insufficient at time t; This is the unit's maximum uphill ramp rate; Let be the rate of upward ramp of the net load at time t; The calculation method for insufficient upward climbing rate is as follows: ; In the formula, The upward climbing rate at time t is less than the expected value; The insufficient climbing rate at time t; The probability of insufficient downhill climbing speed is calculated as follows: ; In the formula, Let t be the probability that the downward climbing rate is insufficient at time t; This is the unit's maximum downhill ramp rate; The calculation method for the downward climbing rate being insufficient to meet the expected value is as follows: ; In the formula, The downward climbing rate at time t is less than the expected value.

[0033] The probability of insufficient upward flexibility is calculated as follows: ; In the formula, The probability of insufficient upward flexibility at time t; This is a function to find the maximum value. The calculation method for insufficient upward flexibility is as follows: ; In the formula, The expected value of insufficient upward flexibility at time t; The probability of insufficient downward flexibility is calculated as follows: ; In the formula, Let t be the probability of insufficient downward flexibility. The calculation method for the expected value of insufficient downward flexibility is as follows: ; In the formula, The expected value for downward flexibility at time t.

[0034] The current performance of the virtual power plant is evaluated using the sequential Monte Carlo method, and the process is as follows: S3.1 Input the output data of wind farms and photovoltaic power stations; S3.2 Simulate the annual output sequence of each wind farm and photovoltaic power station; S3.3 Utilize the failure probability of wind and solar components in the virtual power plant to correct the annual output sequence of each wind farm and photovoltaic power station; the failure probability of wind and solar components corresponds to the equipment failure probability in S2. Wind and solar components include wind turbines in wind farms and photovoltaic modules / inverters in photovoltaic power stations. Failure types include partial component failure and complete shutdown of the entire unit; based on the failure probability corresponding to each wind and solar component, random sampling is used to generate the shutdown status of each time period throughout the year. The theoretical meteorological output of the station in the corresponding time period is deducted according to the rated capacity of the shutdown unit to complete the failure correction of the annual time-series output sequence, and the actual effective output sequence that takes into account both random meteorological fluctuations and random shutdown of power generation components is obtained. S3.4 Input the annual load data of the power distribution system under the jurisdiction of the virtual power plant, calculate the reliability indicators of each wind farm and photovoltaic power station, and comprehensively compile the reliability indicators of the virtual power plant: ; In the formula, , , These represent the overall power shortage time probability, the expected value of the overall power shortage time, and the expected value of the overall power shortage, respectively; For the first In the simulation, the first One simulated state; , , The system is in state respectively. The corresponding probability function for power shortage time, the expected function for power shortage time, and the expected function for power shortage; , , The first The probability of insufficient power time, the expected value of insufficient power time, and the expected value of insufficient power in the next simulation are: the probability of insufficient power time for wind farms or photovoltaic power plants. Number of Monte Carlo simulations; This is a simulated number of hours; For the status of wind and solar components in the virtual power plant The continuous runtime; S3.5 Determine whether the set lifespan M1 of the virtual power plant reliability assessment has been reached. If yes, output the virtual power plant reliability index; otherwise, return to S3.3. S3.6 Calculate the flexibility index of each wind farm and photovoltaic power station at this time, and statistically analyze the flexibility index of the virtual power plant: ; In the formula, , , , , , , , These represent the probability of insufficient upward generation capacity, the expected value of insufficient upward generation capacity, the probability of insufficient downward generation capacity, the expected value of insufficient downward generation capacity, the probability of insufficient upward ramp rate, the expected value of insufficient upward ramp rate, the probability of insufficient downward ramp rate, the expected value of insufficient downward ramp rate, the probability of insufficient upward flexibility, the expected value of insufficient upward flexibility, the probability of insufficient downward flexibility, and the expected value of insufficient downward flexibility for the virtual power plant, respectively. , , , , , , , They represent the first The following values ​​are considered in the simulation: probability of insufficient upward power generation capacity, expected value of insufficient upward power generation capacity, probability of insufficient downward power generation capacity, expected value of insufficient downward power generation capacity, probability of insufficient upward ramp rate, expected value of insufficient upward ramp rate, probability of insufficient downward ramp rate, expected value of insufficient downward ramp rate, probability of insufficient upward flexibility, expected value of insufficient upward flexibility, probability of insufficient downward flexibility, and expected value of insufficient downward flexibility. S3.7 Determine whether the set lifespan M2 for the virtual power plant flexibility assessment has been reached. If yes, output the virtual power plant flexibility index; otherwise, return to S3.3. S3.8 Output virtual power plant reliability indicators and virtual power plant flexibility indicators.

[0035] In power systems, reliability and flexibility are key factors in ensuring stable operation and efficient dispatch. Reliability and flexibility assessments, through the establishment of a series of indicator systems, provide comprehensive quantitative tools for power system performance evaluation. However, assessments alone cannot directly address the power supply reliability issues caused by the uncertainty of renewable energy output. Therefore, this embodiment proposes a dynamic aggregation model, aiming to combine the results of reliability and flexibility assessments to optimize and aggregate renewable energy resources within a virtual power plant, thereby improving the overall reliability and flexibility of the power system.

[0036] The dynamic aggregation model aims to minimize key indicators (such as the expected amount of insufficient power generation) in the reliability assessment model, incorporating reliability constraints into the aggregation decision-making process to improve the overall power supply reliability of the virtual power plant. Simultaneously, it introduces flexibility assessment indicators (such as the probability of insufficient upward / downward generation capacity and the probability of insufficient ramp rate) as constraints to ensure that the aggregated virtual power plant possesses sufficient peak-shaving and ramp-up adjustment capabilities. The dynamic aggregation model can dynamically adjust the aggregation combination and power allocation strategies of wind farms and photovoltaic power plants based on load demand, renewable energy output characteristics, and power plant reliability status at different times. It transforms the quantitative results of the assessment model into implementable dispatch and control strategies, achieving efficient and reliable optimization management of virtual power plant resources.

[0037] In S4, the objective function of the dynamic aggregation model is: ; In the formula, For dynamic aggregation scheme ( The objective function value corresponding to the aggregated decision variable vector; This is the power shortage expectation function, used to calculate the power shortage expectation value of a virtual power plant under a given dynamic aggregation scheme; A 0-1 variable representing whether the u-th wind farm or photovoltaic power station participates in the aggregation. Indicates participation in aggregation, M indicates non-participation; M represents the total number of wind farms and photovoltaic power plants that can participate in aggregation. This represents the equivalent reliable capacity of the u-th wind farm or photovoltaic power station, calculated based on reliability assessment. Represents dynamic aggregation scheme The equivalent reliable capacity of the virtual power plant; The actual installed capacity of a wind farm or photovoltaic power station is its nameplate capacity, which can be directly obtained from the station's equipment ledger or grid connection registration information. Based on this actual installed capacity, and combined with the reliability indicators obtained from the sequential Monte Carlo reliability assessment, a reliability reduction factor is calculated, and the equivalent reliable capacity is obtained to characterize the station's long-term stable power supply capability after considering the randomness of output and equipment failures. For example, the reliability reduction factor is calculated based on EENS: ; In the formula, Represents the reliability conversion factor for the u-th wind farm or photovoltaic power station; Represents the EENS of the u-th wind farm or photovoltaic power station; This represents the minimum amount of electricity generated that falls short of the expected value among all participating wind farms or photovoltaic power plants. It is the EENS value corresponding to the power plant with the highest power supply reliability among all plants, and is used as a benchmark for calculation. The equivalent reliable capacity is calculated as: Equivalent Reliable Capacity = Actual Installed Capacity × Reliability Conversion Factor.

[0038] The constraints of the dynamic aggregation model include power system power balance constraints, wind power output unit operation constraints, photovoltaic power output unit operation constraints, and wind and solar curtailment rate constraints.

[0039] The power balance constraint of a power system is expressed as: ; In the formula, Let t be the load power at time t; , These are the predicted output power values ​​of the h-th wind farm and photovoltaic power station at time t, respectively. , 10% of the predicted output deviation for the h-th wind farm and photovoltaic power station at time t, respectively; H is the number of wind farms and photovoltaic power stations after aggregation; The operating constraints of the wind power output unit are expressed as follows: ; In the formula, , , These represent the output power of the wind power output unit at time t after aggregation, and its upper and lower limits, respectively. , , These represent the ramp rate and upper and lower limits of the wind power output unit at time t after aggregation; The operating constraints of the photovoltaic power output unit are expressed as follows: ; In the formula, , , These represent the output power of the photovoltaic power unit at time t after aggregation, and its upper and lower limits, respectively. , , These represent the ramp rate and upper and lower limits of the photovoltaic power output unit at time t after polymerization; The constraint on wind and solar curtailment rates is expressed as: ; ; In the formula, , , These represent the current, maximum, and minimum values ​​of the virtual power plant's wind and solar curtailment rate after aggregation at time t; , These represent the amount of wind curtailment at the h-th wind farm and the amount of solar curtailment at the h-th photovoltaic power station after aggregation at time t, respectively.

[0040] In S5, the IFLRO algorithm specifically maps each dynamic aggregation scheme to a point in the search space. Each dimension of this point corresponds to a decision variable for a wind farm or a photovoltaic power station, starting from the current point. Depart, in the direction propagation reference step size Arrive at the new point Then, with a step size of 0.1 Generate test points in each dimension Calculate the EENS value of the dynamic aggregation scheme corresponding to each test point, and use it as the fitness value of that test point. This fitness value corresponds to the virtual speed of light propagation. The optimal direction of the probe point is determined based on the fitness value, the propagation direction is updated, and then the continuous coordinate components are converted into 0-1 decision variables by threshold judgment. The iteration is repeated until the termination condition is met, and the optimal dynamic aggregation scheme is output, including the selection of wind farms and photovoltaic power plants participating in the aggregation and suggestions for energy storage configuration.

[0041] The complete implementation steps of the IFLRO algorithm are as follows: Step 1: Solution Space Mapping and Initialization: Map each feasible dynamic aggregation scheme in the virtual power plant dynamic aggregation model to a candidate solution point in the multidimensional search space. Where k represents the current iteration number, Each coordinate component corresponds to an aggregated decision variable (0-1 variable) for a wind farm or photovoltaic power station, and an initial solution point is set. Initial propagation direction vector and the reference step size for light propagation ; Step 2, Ray propagation and trial point generation: In the k-th iteration, from the current solution point... Depart, in the direction propagation reference step size Reaching a new solution point The calculation formula is: ; In the formula, the direction vector , This represents the c-th directional component in the first stage during the k-th iteration. This represents the supplementary component of the first-stage direction vector at the k-th iteration, where C is the number of dimensions; by Starting from the direction of exploration With a step size of 0.1 Generate C test points in each dimension Used to test the next point The direction of dissemination This represents the c-th trial point: ; Step 3: For each trial point Based on the corresponding dynamic aggregation scheme, the objective function value of the dynamic aggregation model, i.e., the expected value of insufficient power EENS, is calculated. This value is used as the fitness value of the test point and also as the virtual velocity of light propagation. The virtual velocity of the c-th test point is denoted as... ; In the k-th iteration, the first stage corresponds to the starting point of the current iteration stage. The second stage corresponds to the main propagation point of the current stage. The third stage corresponds to the testing point of the current stage. ; Step 4: Determining and updating the direction of reflection / refraction: Adjust the virtual velocity of each test point. Virtual velocity corresponding to the dimension of the previous stage Comparison: If the number of trial points that meet the condition is g > C / 2, then it is determined that the light ray is refracted: take the direction corresponding to the minimum virtual velocity among all trial points as the refraction direction, and update the direction vector components: ; ; In the formula, , These represent the c-th directional components of the second and third stages, respectively, during the k-th iteration; , These are the supplementary components of the direction vectors in the second and third stages during the k-th iteration; sign is the sign function. Otherwise, determine if light reflection has occurred: determine the direction corresponding to the minimum virtual velocity, reflect the light, and update the direction vector components: ; ; Finally, the updated propagation direction vector is obtained. ; Step 5: 0-1 conversion of decision variables: based on the updated trial points Extract its coordinate components , As a test point The corresponding continuous coordinate components are the u-th wind farm or photovoltaic power station; since the aggregation decision variables in the dynamic aggregation model are 0-1 variables, a threshold is used to determine... Transform into aggregate decision variables in the objective function, i.e. : ; The average of the absolute values ​​of all coordinate components is used as the threshold for 0-1 conversion.

[0042] Repeat steps two through five until one of the following termination conditions is met: The number of iterations is greater than 500; In two consecutive iterations, the virtual velocity difference of the optimal trial point satisfies ; After the iteration terminates, the current optimal combination of 0-1 decision variables is output, which is the optimal solution of the dynamic aggregation model (the optimal dynamic aggregation scheme).

[0043] The solution steps for the dynamic aggregation model are as follows: S5.1 Input the output data of wind farms and photovoltaic power plants, as well as the load data of the power distribution system under the jurisdiction of the virtual power plant, and perform data preprocessing, and input the constraints of the dynamic aggregation model; S5.2. Set the initial solution point, initial propagation direction vector, reference step size, and iteration termination condition for the IFLRO algorithm; S5.3. The optimization objective is to minimize the expected value of insufficient power corresponding to the dynamic aggregation scheme, and this is used as the fitness function of the algorithm. S5.4. Use the IFLRO algorithm to iteratively solve the dynamic aggregation model and output the optimal combination scheme of wind farm and photovoltaic power station and energy storage configuration suggestions.

[0044] In S6, if the verification results show that the overall power shortage expectation value of the virtual power plant is too high, the power shortage time probability and expected time exceed the preset threshold, or the flexibility index indicates insufficient adjustment margin, then the participation ratio of low reliability stations in the aggregation combination is adjusted according to the equivalent reliable capacity and reliability conversion coefficient of each station, the aggregation weight of high reliability stations is supplemented, and the adjustment resource allocation strategy of the dynamic aggregation scheme is optimized. If necessary, the optimization solution process of S4-S5 is re-executed to update the station aggregation decision variables until the verification results meet the preset reliability and flexibility constraints, forming the final dynamic aggregation scheme.

[0045] In S7, the dynamic aggregation scheme verified and optimized in S6 is applied to the actual operation of the virtual power plant. Real-time operational information such as output data, equipment status, and load demand of each station is collected through a real-time monitoring system. Key parameters such as the equivalent reliable capacity and reliability conversion factor of the dynamic aggregation model are updated regularly. Based on this, the reliability and flexibility indicators of the scheme are re-evaluated using the sequential Monte Carlo method. The power supply reliability and regulation margin of the current aggregation scheme are compared with preset thresholds to determine whether they meet the actual operation requirements. If the reliability or flexibility indicators fail to meet the standards, the station participation status and output allocation strategy of the aggregation combination are adjusted according to the station output characteristics and regulation capabilities fed back by real-time operational data. If necessary, the optimization solution process is re-executed to update the aggregation decision variables, continuously improving the adaptability and reliability of the dynamic aggregation scheme in actual operation scenarios.

[0046] Example 2: A virtual power plant dynamic aggregation device considering the reliability and flexibility of new energy sources, comprising: One or more processors; Memory, used to store one or more computer programs; When one or more programs are executed by one or more processors, the one or more processors execute the method in Example 1.

[0047] Example 3: A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, cause the processor to perform the method in Example 1.

[0048] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can make equivalent substitutions or modifications based on the technical solution and concept of the present invention within the scope of the technology disclosed in the present invention, and such modifications should also be considered to fall within the scope of protection of the present invention.

Claims

1. A virtual power plant dynamic aggregation method considering new energy reliability and flexibility, characterized in that, Includes the following steps: S1. Collect power output data from wind farms and photovoltaic power plants, as well as load data from the distribution-side power system under the jurisdiction of the virtual power plant, and perform data preprocessing. S2. Using the Monte Carlo simulation method, the virtual power plant is simulated and analyzed under all-time operating conditions. Combined with random changes in weather and equipment failure probability, the random fluctuations of wind power and photovoltaic output are characterized, and time-series operation scenarios are generated in batches. S3. Based on the simulation analysis results, calculate the reliability and flexibility indicators, and evaluate the current performance of the virtual power plant based on the sequential Monte Carlo method. S4. Based on the evaluation results, construct a dynamic aggregation model with the objective function of minimizing the expected value of insufficient power, and set its constraints. S5. Improve the judgment conditions for light reflection and refraction in the meshless ray optimization algorithm to form the improved meshless ray optimization algorithm IFLRO. Solve the dynamic aggregation model based on the IFLRO algorithm, obtain the optimal combination scheme of wind farm and photovoltaic power station through iterative calculation, and output the optimal dynamic aggregation scheme. S6. Verify the dynamic aggregation scheme using reliability and flexibility indicators, and optimize and adjust the dynamic aggregation scheme based on the verification results. S7. Apply the optimized and adjusted dynamic aggregation scheme to the actual operation of the virtual power plant, perform real-time monitoring and adjustment, update the dynamic aggregation model parameters based on the regularly collected real-time data, and re-validate the current dynamic aggregation scheme using reliability and flexibility indicators, continuously optimizing and adjusting the dynamic aggregation scheme. 2.The virtual power plant dynamic aggregation method considering new energy reliability and flexibility according to claim 1, wherein, In S1, historical and real-time output data of each wind farm and photovoltaic power station included in the aggregation scope of the virtual power plant are collected, and measured and predicted data of various user loads in the power distribution side of the virtual power plant are obtained simultaneously. Auxiliary information including wind farm parameters, photovoltaic power station parameters, geographical location and meteorological environment data is also supplemented. The collected data is preprocessed by cleaning out abnormal noise and repairing missing or erroneous data, followed by data smoothing and standardization to unify the data format and units, forming a well-organized dataset. 3.The virtual power plant dynamic aggregation method considering new energy reliability and flexibility of claim 1, wherein, In S3, the reliability indicators include the probability of insufficient power time, the expected value of insufficient power time, and the expected value of insufficient power; the flexibility indicators include the probability of insufficient upward generation capacity, the expected value of insufficient upward generation capacity, the probability of insufficient downward generation capacity, the expected value of insufficient downward generation capacity, the probability of insufficient upward ramp rate, the expected value of insufficient upward ramp rate, the probability of insufficient downward ramp rate, the expected value of insufficient downward ramp rate, the probability of insufficient upward flexibility, the expected value of insufficient upward flexibility, the probability of insufficient downward flexibility, and the expected value of insufficient downward flexibility. 4.The virtual power plant dynamic aggregation method considering new energy reliability and flexibility according to claim 3, wherein, In S3, the current performance of the virtual power plant is evaluated based on the sequential Monte Carlo method. The process is as follows: S3.1 Input the output data of wind farms and photovoltaic power stations; S3.2 Simulate the annual output sequence of each wind farm and photovoltaic power station; S3.3 Utilize the failure probability of wind and solar components in the virtual power plant to correct the annual output sequence of each wind farm and photovoltaic power station; S3.4 Input the annual load data of the power distribution system under the jurisdiction of the virtual power plant, calculate the reliability index of each wind farm and photovoltaic power station, and comprehensively compile the reliability index of the virtual power plant. S3.5 Determine whether the set lifespan M1 of the virtual power plant reliability assessment has been reached. If yes, output the virtual power plant reliability index; otherwise, return to S3.

3. S3.6 Calculate the flexibility index of each wind farm and photovoltaic power station at this time and compile the virtual power plant flexibility index. S3.7 Determine whether the set time limit M2 for the virtual power plant flexibility assessment has been reached. If yes, output the virtual power plant flexibility index; otherwise, return to S3.

3. S3.8 Output virtual power plant reliability indicators and virtual power plant flexibility indicators. 5.The virtual power plant dynamic aggregation method considering new energy reliability and flexibility of claim 1, wherein, In S4, the objective function of the dynamic aggregation model is: ； In the formula, For dynamic aggregation scheme The corresponding objective function value; This is the power shortage expectation function, used to calculate the power shortage expectation value of a virtual power plant under a given dynamic aggregation scheme; A 0-1 variable representing whether the u-th wind farm or photovoltaic power station participates in the aggregation. Indicates participation in aggregation, M indicates non-participation; M represents the total number of wind farms and photovoltaic power plants that can participate in aggregation. This represents the equivalent reliable capacity of the u-th wind farm or photovoltaic power station, calculated based on reliability assessment. Represents a dynamic aggregation scheme The equivalent reliable capacity of the virtual power plant; The constraints of the dynamic aggregation model include power system power balance constraints, wind power output unit operation constraints, photovoltaic power output unit operation constraints, and wind and solar curtailment rate constraints.

6. The method for dynamic aggregation of virtual power plants considering the reliability and flexibility of new energy sources according to claim 5, characterized in that, In S5, the IFLRO algorithm specifically maps each dynamic aggregation scheme to a point in the search space. Each dimension corresponds to a decision variable for a wind farm or a solar power station, starting from the current point. Depart, along the direction propagation reference step size Arrive at the new point Then, with a step size of 0.1 Generate test points in each dimension Calculate the EENS value of the dynamic aggregation scheme corresponding to each test point, and use it as the fitness value of that test point. This fitness value corresponds to the virtual speed of light propagation. The optimal direction of the probe point is determined based on the fitness value, the propagation direction is updated, and then the continuous coordinate components are converted into 0-1 decision variables by threshold judgment. The iteration is repeated until the termination condition is met, and the optimal dynamic aggregation scheme is output, including the selection of wind farms and photovoltaic power plants participating in the aggregation and suggestions for energy storage configuration.

7. The method for dynamic aggregation of virtual power plants considering the reliability and flexibility of new energy sources according to claim 6, characterized in that, In S5, the solution steps for the dynamic aggregation model are as follows: S5.1 Input the output data of wind farms and photovoltaic power plants, as well as the load data of the power distribution system under the jurisdiction of the virtual power plant, and perform data preprocessing, and input the constraints of the dynamic aggregation model; S5.

2. Set the initial solution point, initial propagation direction vector, reference step size, and iteration termination condition for the IFLRO algorithm; S5.

3. The optimization objective is to minimize the expected value of insufficient power corresponding to the dynamic aggregation scheme, and this is used as the fitness function of the algorithm. S5.

4. Use the IFLRO algorithm to iteratively solve the dynamic aggregation model and output the optimal combination scheme of wind farm and photovoltaic power station and energy storage configuration suggestions.

Images