Response reliability evaluation method and system of virtual power plant and related product

Abstract

The invention discloses a response reliability evaluation method and system for a virtual power plant and a related product, and belongs to the technical field of virtual power plant evaluation optimization. Stochastic state simulation is carried out based on a Monte Carlo method, a stochastic state sequence is generated, uncertain factors such as wind and light output fluctuation and energy storage state randomness are quantified into the stochastic state sequence capable of being analyzed, and the limitation of deterministic hypothesis is broken through; through energy storage dynamic division and real-time updating of a division result, and in combination with a design of calculating a reliability index by a random state sequence and an energy storage execution model, a virtual power plant dynamic reconstruction demand caused by wind and light output fluctuation and user demand change can be responded; and through dynamic clustering and attenuation quantification, the operation adaptability of the system under the uncertain working condition is ensured, the finally calculated reliability index can comprehensively reflect the real response capability of the virtual power plant under the uncertain environment, and an accurate evaluation basis is provided for improving the adaptive capability of the virtual power plant under the uncertain environment.

Description

Technical Field

[0001] This invention relates to the field of virtual power plant evaluation and optimization technology, specifically to a method, system, and related products for evaluating the response reliability of a virtual power plant. Background Technology

[0002] Virtual power plants (VPPs), as an important platform integrating distributed power sources, controllable loads, and energy storage systems, can leverage resource flexibility at low cost, optimize grid dispatch, and effectively mitigate the uncertainties brought about by the integration of new energy sources. Through VPP reliability assessment, weaknesses can be identified, enabling more effective measures to improve power supply reliability. Poor VPP reliability exacerbates the crowding-out effect of conventional energy on renewable energy, leading to a decrease in the grid's confidence in renewable energy supply capacity and severe wind and solar curtailment. Energy storage systems (ESS), with their rapid charging and discharging capabilities and frequency regulation response, provide short-term backup and energy balance support for the grid, thereby improving the overall system reliability. Therefore, research on dynamic clustering and capacity decay of energy storage systems is of great significance for VPP reliability assessment.

[0003] In energy storage modeling research, the energy storage system is considered part of a virtual power plant, and the proposed energy storage degradation index is crucial for the reliability assessment of the virtual power plant. Existing research has made significant achievements in the diversified resource clustering of virtual power plants, but it generally neglects the impact of dynamic clustering of energy storage systems on the reliability of virtual power plants. Furthermore, energy storage modeling assumes a constant energy storage capacity, ignoring the reality that the capacity of the energy storage system (ESS) gradually decays during multiple charge-discharge cycles. In addition, while various reliability indices are used in the assessment process, they are all based on deterministic assumptions and fail to consider the risk factors that depend on decision-making.

[0004] Therefore, how to consider the dynamic reconfiguration requirements under the fluctuation of wind and solar power output and changes in user demand, as well as the capacity decay characteristics of energy storage systems due to cyclic aging in actual operation, and at the same time improve the adaptability of virtual power plants in uncertain environments, has become an urgent technical problem to be solved. Summary of the Invention

[0005] The purpose of this invention is to provide a method, system, and related products for evaluating the response reliability of virtual power plants, so as to overcome the above-mentioned defects and deficiencies in the prior art, provide a technical solution for virtual power plant reliability evaluation that is more in line with actual operating scenarios and takes into account the influence of multiple factors, and provide support for improving the operational stability of virtual power plants and the renewable energy consumption capacity.

[0006] The present invention solves the above-mentioned technical problems through the following technical solution: This invention provides a method for evaluating the response reliability of a virtual power plant, comprising the following steps: S1. Based on Monte Carlo simulation, a random state sequence is generated for the virtual power plant, which includes the output of new energy sources and the state of the energy storage system. S2. Obtain the energy storage capacity and user response intention of each energy storage system, and dynamically divide the energy storage system into high-frequency frequency regulation energy storage system and low-frequency energy support energy storage system using a distributed dynamic clustering algorithm; construct an energy storage system execution model, which executes tasks according to the dynamic division results of the energy storage system. If the energy storage system is a high-frequency frequency regulation energy storage system, then the energy storage system executes the frequency and voltage regulation task; when the energy storage system is a low-frequency energy support energy storage system, then the energy storage system executes the energy support task; introduce a dynamic available capacity index into the energy storage system execution model to quantify the energy storage capacity decay, and update the dynamic division results of the energy storage system according to the real-time value of the dynamic available capacity index; S3. Based on the random state sequence, the dynamic partitioning results of the energy storage system, and the execution model of the energy storage system, calculate the response reliability index of the virtual power plant.

[0007] A further improvement of this invention lies in obtaining the energy storage capacity and user response intention characteristics of each energy storage system, and dynamically dividing the energy storage system into high-frequency frequency regulation energy storage systems and low-frequency energy support energy storage systems using a distributed dynamic clustering algorithm. Specifically, this includes the following steps: For each energy storage system, a feature vector containing normalized energy storage capacity and user response intention is constructed; Each energy storage system updates its local estimate of the cluster center iteratively based on a distributed average update formula through neighborhood communication. Each energy storage system determines whether it belongs to a high-frequency frequency regulation energy storage system or a low-frequency energy support energy storage system based on the updated cluster center and weighted Euclidean distance.

[0008] A further improvement of this invention is that the dynamic available capacity indicator is specifically as follows:

[0009] in, This represents the k-th time state in the Monte Carlo state simulation. for Dynamic available capacity metrics; The ESS charge / discharge energy for state k; ( ) represents the natural exponential function; This represents the initial rated capacity of the energy storage system. This is the capacity decay coefficient; The cycle depth decay index; The duration of the Monte Carlo state.

[0010] A further improvement of this invention lies in calculating the response reliability index of the virtual power plant based on the random state sequence, the dynamic partitioning result of the energy storage system, and the execution model of the energy storage system. This specifically includes the following steps: Construct a two-level optimization model with the upper-level objective of minimizing the total operating cost of the virtual power plant and the clustering error, and the lower-level objective of maximizing the system response reliability. In the upper-level optimization of the two-level optimization model, a risk value constraint is introduced. If the risk value meets the preset confidence level, proceed to the lower-level optimization; otherwise, return to step S1. The random state sequence, the dynamic partitioning result of the energy storage system, and the execution model of the energy storage system are input into the two-layer optimization model for solution, and the response reliability index of the virtual power plant is calculated. Determine whether the preset number of random state simulations has been reached. If yes, output the current response reliability index; otherwise, return to step S1.

[0011] A further improvement of this invention is that the Value at Risk constraint is specifically as follows:

[0012] in, Risk value constraint; Risk threshold; β To preset the credit level; This refers to the energy not expected to be supplied. A set of risk values; It represents probability.

[0013] A further improvement of the present invention is that the response reliability index is specifically:

[0014] in, In response to reliability indicators; This refers to the energy not expected to be supplied. This refers to the total energy that can theoretically be supplied throughout the entire time period.

[0015] The present invention also provides a response reliability assessment system for a virtual power plant, comprising: The first module is used to perform stochastic state simulation of virtual power plants based on Monte Carlo simulation, generating a stochastic state sequence that includes the output of new energy sources and the state of energy storage systems. The second module is used to obtain the energy storage capacity and user response intention of each energy storage system, and dynamically divide the energy storage system into high-frequency frequency regulation energy storage system and low-frequency energy support energy storage system using a distributed dynamic clustering algorithm; construct an energy storage system execution model, which executes tasks according to the dynamic division results of the energy storage system. If the energy storage system is a high-frequency frequency regulation energy storage system, it is instructed to perform frequency and voltage regulation tasks; if the energy storage system is a low-frequency energy support energy storage system, it is instructed to perform energy support tasks; a dynamic available capacity index is introduced into the energy storage system execution model to quantify energy storage capacity decay, and the dynamic division results of the energy storage system are updated according to the real-time value of the dynamic available capacity index. The third module is used to calculate the response reliability index of the virtual power plant based on the random state sequence, the dynamic partitioning result of the energy storage system, and the execution model of the energy storage system.

[0016] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the response reliability assessment method for the virtual power plant described above.

[0017] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the response reliability assessment method for the virtual power plant described above.

[0018] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the response reliability assessment method for the virtual power plant described above.

[0019] Compared with the prior art, the positive and progressive effects of the present invention are as follows: The virtual power plant response reliability assessment method provided by this invention is based on Monte Carlo simulation to generate a random state sequence containing the output of new energy sources and the state of the energy storage system. It quantifies uncertainties such as wind and solar power output fluctuations and the randomness of energy storage states into an analyzable random state sequence, breaking the limitations of deterministic assumptions. Through dynamic partitioning of energy storage and real-time updating of the partitioning results, combined with the design of reliability index calculation using random state sequences and energy storage execution models, a full-process adaptation mechanism for uncertainties is formed. This mechanism can respond to the dynamic reconfiguration needs of the virtual power plant caused by wind and solar power output fluctuations and changes in user demand. It also ensures the system's operational adaptability under uncertain operating conditions through dynamic clustering and attenuation quantification. The final calculated reliability index can comprehensively reflect the real response capability of the virtual power plant under uncertain environments, providing an accurate assessment basis for improving the adaptability of the virtual power plant under uncertain environments. This overcomes the shortcomings of existing technologies, such as one-sided assessment and inability to support decision-making under complex operating conditions. Attached Figure Description

[0020] The accompanying drawings are provided to further understand the invention and constitute a part of this invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0021] Figure 1 This is a flowchart illustrating the response reliability assessment method for a virtual power plant according to the present invention. Figure 2 The results of ESS clustering; Figure 3 This is an indicator of the dynamic available capacity of energy storage. Figure 4 A diagram showing the relationship between DAI and EENS; Figure 5 Output diagram of wind turbine unit; Figure 6 Output diagram of photovoltaic unit; Figure 7 The system voltage waveforms are shown in two scenarios. Figure 8 The diagram shows the system frequency waveforms for two different scenarios. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] In the description of this invention, it should be understood that the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0024] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0025] It should be understood that although terms such as first, second, third, etc., may be used in the embodiments of the present invention to describe the preset range, these preset ranges should not be limited to these terms. These terms are only used to distinguish the preset ranges from one another. For example, without departing from the scope of the embodiments of the present invention, the first preset range may also be referred to as the second preset range, and similarly, the second preset range may also be referred to as the first preset range.

[0026] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0027] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. This is an explanation of the present invention and not a limitation thereof.

[0028] This invention provides a method for evaluating the response reliability of a virtual power plant, comprising the following steps: S1. Based on Monte Carlo simulation, a random state sequence is generated for the virtual power plant, which includes the output of new energy sources and the state of the energy storage system. S2. Obtain the energy storage capacity and user response intention of each energy storage system, and dynamically divide the energy storage system into high-frequency frequency regulation energy storage system and low-frequency energy support energy storage system using a distributed dynamic clustering algorithm; construct an energy storage system execution model, which executes tasks according to the dynamic division results of the energy storage system. If the energy storage system is a high-frequency frequency regulation energy storage system, then the energy storage system executes the frequency and voltage regulation task; when the energy storage system is a low-frequency energy support energy storage system, then the energy storage system executes the energy support task; introduce a dynamic available capacity index into the energy storage system execution model to quantify the energy storage capacity decay, and update the dynamic division results of the energy storage system according to the real-time value of the dynamic available capacity index; S3. Based on the random state sequence, the dynamic partitioning results of the energy storage system, and the execution model of the energy storage system, calculate the response reliability index of the virtual power plant.

[0029] This solution uses the Monte Carlo method to simulate stochastic states, generating a stochastic state sequence that includes the output of new energy sources and the state of the energy storage system. It quantifies uncertainties such as fluctuations in wind and solar power output and the randomness of energy storage states into an analyzable stochastic state sequence, breaking the limitations of deterministic assumptions. Simultaneously, the dynamic partitioning of energy storage in step S2 and the real-time updating of the partitioning results, and the design of calculating reliability indicators by combining the stochastic state sequence with the energy storage execution model in step S3, form a full-process adaptation mechanism for uncertainties. This mechanism can respond to the dynamic reconfiguration needs of the virtual power plant caused by fluctuations in wind and solar power output and changes in user demand. Furthermore, dynamic clustering and attenuation quantification ensure the system's operational adaptability under uncertain conditions. The final calculated reliability indicators comprehensively reflect the true response capability of the virtual power plant under uncertain environments, providing an accurate evaluation basis for improving the virtual power plant's adaptability in uncertain environments. This overcomes the shortcomings of existing technologies, such as one-sided evaluation and inability to support decision-making under complex operating conditions.

[0030] By combining a distributed dynamic clustering algorithm, this design dynamically classifies energy storage systems into high-frequency frequency regulation energy storage systems and low-frequency energy support energy storage systems based on energy storage capacity and user response intentions, achieving accurate dynamic classification of energy storage resources. Simultaneously, the energy storage system execution model assigns specific tasks based on the classification results: high-frequency frequency regulation energy storage systems focus on frequency and voltage regulation, while low-frequency energy support energy storage systems undertake energy support tasks. This ensures a high degree of matching between the performance characteristics of various energy storage systems and the grid's task requirements, avoiding reliability improvement bottlenecks caused by the misuse or mismatch of energy storage resources. Furthermore, the solution updates the energy storage classification results in real time based on dynamic available capacity indicators, ensuring that the clustering results dynamically adjust with changes in energy storage status. This completely overcomes the problem of disconnect between assessment and actual operation caused by static or non-targeted clustering in existing technologies, providing a precise foundation for energy storage resource allocation in reliability assessment.

[0031] By introducing a dynamic available capacity index to quantify energy storage capacity decay and updating the dynamic energy storage partitioning results based on the real-time value of this index, the design incorporates energy storage capacity decay characteristics into the modeling and evaluation system, breaking the idealized assumptions of existing technologies. The dynamic available capacity index can accurately capture the capacity changes of the energy storage system during the cyclic aging process, ensuring that the operating state of the energy storage system execution model is highly consistent with the actual operating conditions. This ensures that subsequent reliability index calculations are based on the actual available energy storage capacity, effectively avoiding the evaluation distortion problem caused by the assumption of constant capacity, and solving the deficiency that existing evaluation results cannot guide actual operation and maintenance decisions.

[0032] The method of this invention can improve the accuracy of virtual power plant response reliability assessment, thereby providing a scientific basis for optimizing energy storage configuration and improving the overall reliability of virtual power plants. Precise dynamic clustering and attenuation quantification enable energy storage systems to fully leverage their rapid charging and discharging, frequency regulation response, and energy balance support capabilities. This effectively mitigates fluctuations in renewable energy output, alleviates uncertainties caused by renewable energy integration, enhances the grid's confidence in renewable energy supply capacity, reduces wind and solar curtailment, weakens the crowding-out effect of conventional energy, and helps virtual power plants more efficiently integrate distributed power sources, controllable loads, and energy storage systems, fully exploiting resource flexibility and optimizing grid dispatch efficiency. This method combines systematicity and operability. The application of Monte Carlo random simulation, distributed dynamic clustering algorithms, and the quantification method of dynamic available capacity indicators are all adaptable to actual engineering operation data and can be directly applied to the actual operation and maintenance assessment of virtual power plants. The response reliability index calculated in step S3 can accurately identify the weak links in the reliability of the virtual power plant, providing clear guidance for taking targeted measures (such as energy storage system operation and maintenance optimization, task allocation adjustment, capacity supplementation planning, etc.) to improve power supply reliability. It solves the problem that the existing technical evaluation methods are out of touch with engineering practice and cannot be transformed into actual operation and maintenance solutions, and has great engineering application value and promotion significance.

[0033] For details, see Figure 1 A method for assessing the response reliability of a virtual power plant, specifically including the following steps: S1. Initialize parameters and generate Monte Carlo states, including equipment failure and recovery time, wind speed, irradiance, etc. Calculate the output of new energy sources such as wind power and photovoltaics, and the status of equipment based on the generated state sequence.

[0034] (1) The Monte Carlo method is used to simulate the stochastic state of the virtual power plant. For each time state k The failure and recovery times of the equipment are constructed by generating uniformly distributed random numbers. For the first... i For each piece of equipment, the formulas for its time to failure (TTF) and time to repair (TTR) are defined as follows:

[0035] In the formula: For the first i Failure rate of the equipment; For repair rate; and All are random numbers that follow a uniform distribution on [0,1].

[0036] Define the status indicator function as follows:

[0037] In Monte Carlo simulations, through a large number of state samples , NSS Calculate the effective output of each device under each state.

[0038] (2) Wind turbine model Wind speed in natural environments can be represented by a two-parameter Weibull distribution with high goodness of fit, and its probability density function is as follows:

[0039] In the formula: Representing state k wind speed; and Let be the shape parameter and scale parameter of the Weibull distribution, respectively. i Typhoon generator in status k The following output is:

[0040] In the formula: , , These are the cut-in, rated, and cut-out wind velocities of the fan, respectively. For the first i The rated power of the typhoon turbine generator set. Based on the equipment reliability model, the wind turbine is introduced... Correct its actual output.

[0041] (3) Photovoltaic generator model This invention uses the beta distribution to predict solar radiation, thereby indirectly describing photovoltaic power output. The beta distribution is a continuous probability distribution defined on [0,1], and its probability density function is specifically expressed as follows:

[0042] In the formula: For state k Solar irradiance, W / m²; For gamma function; This represents the maximum solar irradiance. , Let be the shape parameter of the beta distribution. Then the th j The photovoltaic generator is in operation. k Lower output power Represented as:

[0043] In the formula: and These are the initial and rated irradiance of the photovoltaic module, respectively. For the firstj The rated power of the photovoltaic generator.

[0044] At the same time, in conjunction with the equipment reliability model, the photovoltaic generator is also introduced... Correct its actual output.

[0045] S2. A distributed dynamic clustering algorithm is proposed to construct an energy storage system. This algorithm primarily considers the energy storage capacity of each energy storage system (ESS) and user response intentions, dynamically dividing the system into two groups based on feature vectors: 1) High-frequency frequency regulation energy storage system (ESSFR): Primarily responsible for rapid frequency and voltage regulation. In power systems, frequency and voltage stability are crucial for ensuring the normal operation of power equipment. When power fluctuations occur in the grid, ESSFR can respond quickly, regulating the grid's frequency and voltage through rapid charging and discharging to maintain them within a stable range. This type of energy storage system typically requires high response speed and power density, capable of providing or absorbing large amounts of energy in a short time; 2) Low-frequency energy support energy storage system (ESSBS): Primarily provides system energy support. Unlike ESSFR, ESSBS is mainly used to address long-term energy supply and demand imbalances. When renewable energy generation is insufficient, ESSBS can release stored energy to provide continuous energy support to the grid; conversely, when renewable energy generation is excessive, ESSBS can absorb excess energy, avoiding energy waste. This type of energy storage system typically requires large energy storage capacity and long charging and discharging times.

[0046] The following are the basic principles of clustering algorithms.

[0047] 1) Let the first i The feature vectors of each ESS are:

[0048] In the formula: Indicates the normalized energy storage capacity; This indicates the user's willingness to respond or participate, such as their expected economic compensation price. User willingness to respond plays a crucial role in the dispatch of energy storage systems, as different users may have varying degrees of enthusiasm for participating in energy storage dispatch, which can affect the actual operational performance of the energy storage system.

[0049] 2) Each ESS updates the value of the first ESS using the following distributed average update formula through neighborhood communication. k Local estimation of class group centers:

[0050] In the formula: For the firsti A set of neighbors of an ESS; For communication weights; For convergence gain; This is the preset number of clusters.

[0051] Through this distributed average update formula, each ESS can update its estimate of the group center based on information from its neighbors, thereby gradually achieving the goal of clustering. This distributed approach avoids the communication burden and single point of failure problems associated with centralized control, improving the reliability and flexibility of the algorithm.

[0052] 3) Subsequently, the membership relationship of each ESS is determined using a weighted Euclidean distance, expressed by the formula:

[0053] In the formula: This is the feature weight matrix, and the weights can be adaptively adjusted based on feedback.

[0054] To ensure that the low-frequency energy support ESS has sufficient total energy capacity, an adaptive learning mechanism is designed for the weight matrix. Perform online adjustments. If feedback indicates that the total ESS capacity in the current cluster is lower than the expected demand, appropriately increase the weights related to energy storage capacity. Conversely, the energy density decreases. This adjustment causes the cluster centers to move, prompting more high-capacity ESSs to join the corresponding groups, thus achieving energy balance and optimized configuration within the groups.

[0055] S3. Under the guidance of clustering results, various energy storage systems perform frequency and voltage regulation or energy support tasks respectively, and achieve coordinated operation through distributed droop control and state of charge balancing mechanism.

[0056] (1) The distributed droop control model is as follows:

[0057] In the formula, For the first i The output frequency of each energy storage unit; For the first i The output voltage of each energy storage unit; , These are reference values ​​for frequency and voltage, respectively. , The first i The active and reactive power output of each energy storage unit; , The first i The frequency and voltage drop coefficients of each energy storage unit.

[0058] The distributed frequency recovery control model is as follows:

[0059] In the formula, This refers to the frequency control gain coefficient. The system's rated frequency; The guiding coefficient for pinning nodes.

[0060] The distributed voltage recovery control model is as follows:

[0061] In the formula, This refers to the voltage control gain coefficient. This is the system's rated frequency.

[0062] The SOC equalization control model is as follows:

[0063] In the formula, For the first i The state-of-charge variables of each energy storage unit; A , B The state matrix of the system; The gain is controlled by the SOC balance.

[0064] S4. The energy storage system operation model includes charge / discharge state transition and energy storage update.

[0065] (1) Update ESS charging and discharging status; (2) Calculate the number of loops and depth, and update the dynamic available capacity index.

[0066] Energy storage capacity decay is a cumulative process, with the core driving factors being the frequency and depth of charge-discharge cycles. Regarding cycle frequency, frequent charge-discharge accelerates the loss of active materials, thus exacerbating capacity decay. As for cycle depth, deep charge-discharge (near full charge or empty discharge) causes far more damage to the battery than shallow charge-discharge. Therefore, partial charge-discharge cycles need to be weighted according to actual throughput. To quantify the capacity decay effect of virtual energy storage during long-term operation, the dynamic available capacity index (DACI) is defined as the normalized ratio of the current available capacity to the initial capacity, and the impact of the number of charge-discharge cycles on capacity decay is introduced:

[0067] In the formula: This represents the k-th time state in the Monte Carlo state simulation. for Dynamic available capacity metrics; The ESS charge / discharge energy for state k; ( ) represents the natural exponential function; The initial rated capacity of the energy storage system is MWh; The capacity decay coefficient represents the proportion of capacity loss caused by each charge-discharge cycle. The cycle depth decay index reflects the accelerated decay effect of deep charge and discharge. The duration of the Monte Carlo state.

[0068] DCI normalizes the capacity status to the [0,1] interval. When DCI=1, it indicates no capacity degradation, the energy storage system is in its initial ideal state, and its response capability is strongest. When DCI=0, it indicates severe capacity degradation, the energy storage system is close to failure, and it cannot effectively participate in regulation. The closer the DCI value is to 1, the less severe the degradation of available energy storage capacity and the higher the reliability.

[0069] Update the energy storage capacity at each Monte Carlo state transition:

[0070] In the formula, For state k Next m The capacity of the ESS unit.

[0071] The capacity degradation of energy storage units is mainly reflected in the decrease in their usable capacity, leading to an overall decrease in the DCI value. When the DCI falls below a certain critical threshold, the clustering module reclassifies energy storage units originally belonging to the high-frequency modulation category (ESSFR) as low-frequency energy support category (ESSBS). This clustering migration effectively avoids additional losses caused by performance-degraded energy storage participating in frequent regulation.

[0072] To reflect the attenuation effect, the present invention designs the following clustering assignment adjustment function:

[0073] In the formula: The DAI classification threshold is set in accordance with the provisions of the national standard GB / T 36547-2018. It is 0.91.

[0074] (3) Adjust available capacity; (4) Update the ESS output status.

[0075] In state k Next, the m Energy state of Taiwan ESS The updated formula is:

[0076] In the formula: For state k The charging and discharging energy of the ESS is calculated based on the remaining energy in the virtual power plant and the energy storage charging and discharging constraints. The virtual power plant output is set to only account for a certain value of the load, i.e., the virtual power plant output ratio X. Specifically, there are two cases: 1. When the virtual power plant has surplus renewable energy, i.e., when the output of photovoltaic and wind power exceeds the load, the ESS is in a charging state. The calculation formula is as follows:

[0077] 2. When there is a load gap in the virtual power plant, the ESS discharge compensation formula is as follows:

[0078] In the formula: Representing state k The surplus output of wind or solar power; Representing state k The remaining or insufficient output of conventional generators; The duration of the state; , These represent the charging and discharging efficiencies of the energy storage system, respectively.

[0079] Meanwhile, the charge / discharge rate of the ESS is constrained by the maximum rate of change:

[0080] In the formula: and The first m Maximum and minimum energy capacity of Taiwan ESS; For the first m The time required for a Taiwan ESS to go from fully charged to empty.

[0081] S5. Considering the dual impact of dynamic clustering of energy storage systems on the reliability and economy of virtual power plants, a two-layer optimization model is adopted to construct the objective function. The risk value constraint ensures that the risk of power outage in extreme scenarios is within a controllable range, while the two-layer optimization structure further achieves a comprehensive trade-off between operating costs and reliability. The combination of the two enables the model to both defend against low-probability extreme risks and maintain an economical and efficient operating strategy in conventional scenarios.

[0082] (1) Upper-level optimization 1) Upper-level optimization objective: Minimize operating costs and clustering error. Considering the operating costs of the virtual power plant and the clustering results to achieve the upper-level optimization objective, balancing economic efficiency and the rationality of the clustering structure, this is specifically expressed as:

[0083] In the formula, This indicates that the energy storage devices classified as belonging to the low-frequency energy support cluster have a capacity lower than a set threshold. The number of devices; This represents the clustering error penalty coefficient, reflecting the degree of importance the operation and maintenance strategy places on cluster matching accuracy. Total operating cost of the virtual power plant. It consists of the power generation cost of wind turbines, the power generation cost of photovoltaic units, and the operating cost of energy storage systems, and is defined as:

[0084] In the formula: Cost of generating electricity from wind turbines; Cost of electricity generated by photovoltaic units; The operating cost of an ESS covers charge and discharge efficiency losses, maintenance costs, and additional expenses due to capacity decay.

[0085] 2) Calculate reliability indices based on optimization results. The outputs of wind power, photovoltaic power, and ESS are categorized by status. k After integration, the total output of the virtual power plant This can be expressed as:

[0086] In the formula: For the first i The typhoon generator is in operation. k Output power at the following levels; For the first j The photovoltaic generator is in operation. k Output power at the following levels; For the first m Taiwan's energy storage system is in status k The charging and discharging power is as follows: WTG is the number of wind turbines; PV is the number of photovoltaic generators; ESS is the number of energy storage systems.

[0087] Define the expected energy not served (EENS) and loss of load expectation (LOLE) indicators to quantify system reliability, and statistically analyze the load gap of virtual power plants under various conditions.

[0088]

[0089]

[0090]

[0091] In the formula: Provide system output margin; This is the system's total output; The load required by the system; This is a simulated number of years.

[0092] 3) Calculate the VaR risk value and check whether the confidence level is met. If it is met, proceed to the next level of optimization; otherwise, return to step S1.

[0093] Value at risk (VaR) is used to quantify the maximum potential loss from unsupplied energy by a virtual power plant at a given confidence level. Its mathematical definition is:

[0094] In the formula: A set of risk values; Represents probability; β As a confidence level, it is usually taken as 0.95, which means that the unsupplied energy does not exceed the VaR threshold with a 95% probability. The risk threshold, expressed in MWh, represents the confidence level. β The maximum unsupplied energy that is acceptable.

[0095] To ensure that the risks of the virtual power plant are controllable under extreme conditions, a value-at-risk constraint is incorporated:

[0096] In the formula: The maximum permissible risk value set for decision-makers, in units of same.

[0097] (2) Lower-level optimization 1) Objective: Maximize system response reliability. The response reliability index EIR is introduced as a lower-level optimization objective to evaluate the system's energy security capability under supply and demand disturbances. This objective function design ensures both the long-term stable operation reliability of the system and reflects the role of dynamic clustering in improving energy storage utilization efficiency and operational economy. Its definition is as follows:

[0098] EIR measures the overall system's ability to meet user electricity demand under various operating conditions, reflecting the actual support effect of clustering results on reliability. Further, an improved response reliability index (EIR) based on virtual power plants is introduced, defined as follows:

[0099] In the formula: EENS is the expected unsupplied energy, representing the cumulative amount of unsupplied power due to insufficient output of new energy sources and limitations in energy storage regulation capacity during the simulation period; TDE is the total energy that can theoretically be supplied throughout the entire period.

[0100] 2) Record EENS, LOLE, and EIR; 3) Determine if the simulation count has been reached. If so, output the reliability index, cost, and risk value of the result; otherwise, continue generating Monte Carlo states.

[0101] Example 1 The simulation used 100 states, each lasting 1 hour. New energy output was generated based on random wind speed and irradiance. After ESS charging and discharging regulation, the total system output and reserve capacity were statistically analyzed. The system load was fixed at 50MW, the system rated voltage was 230V, and the system rated frequency was 50Hz.

[0102] Five energy storage systems are configured. Energy Storage Systems (ESS) dynamically allocate charging and discharging capacity based on the remaining renewable energy. Each ESS has a charging / discharging duration of 0.5 hours, a charging efficiency of 0.95, and a discharging efficiency of 0.9. In this case, the virtual power plant output ratio X is set to 0.2; any excess is used for ESS charging, and any shortfall is compensated by ESS discharging. During operation, the energy storage system incurs both charging / discharging efficiency losses and daily maintenance expenses, as well as capacity degradation costs due to cyclic use. Therefore, the unit cost of energy storage is broken down into two parts: a cyclic operating cost of 10 yuan / MWh and a capacity degradation cost of 56 yuan / MWh, totaling 66 yuan / MWh.

[0103] Using a distributed dynamic clustering algorithm, the five energy storage systems were divided into two categories based on their current state of charge and response intentions. Figure 2As can be seen, the horizontal axis represents the normalized State of Charge (SOC) of the energy storage system, i.e., the ratio of the current energy storage state to the maximum capacity, and the vertical axis represents the user's willingness to respond. The results show that ESS1 and ESS3 are high-frequency frequency regulation energy storage systems, while ESS2, ESS4, and ESS5 are low-frequency energy support energy storage systems. When the normalized SOC is low and the user's willingness to respond is low, it indicates that the energy storage can release limited energy and has insufficient response capability to frequency fluctuations, as seen in ESS4 and ESS5. When the normalized SOC is high but the user's willingness to respond is low, it indicates that the energy storage can release sufficient energy, but will be unwilling to respond to frequency fluctuations, as seen in ESS2. When the normalized SOC is low but the user's willingness to respond is high, it indicates that the energy storage can release insufficient energy, but will respond actively to frequency fluctuations. Classifying these three types of ESSs as ESSFR would lead to large frequency deviations and reduced reliability of the virtual power plant; therefore, they can only be classified as ESSBS for providing system energy support. The normalized SOC and user response willingness of ESS3 in the figure are both at a moderate level, proving that it has a certain response capability to system frequency fluctuations. This can be attributed to EESFR's rapid adjustment of system frequency and voltage, thereby improving the overall real-time response capability and power supply reliability of the virtual power plant. Through this grouping, the virtual power plant can flexibly dispatch various energy storage systems under different time scales and operating conditions, which can not only respond to wind and solar power output fluctuations in a timely manner, but also maximize the utilization of energy storage resources.

[0104] During long-term operation, the actual usable capacity of energy storage devices decreases nonlinearly with the number of charge-discharge cycles due to the loss of active materials. The randomness of wind and solar power output, combined with the capacity decay of energy storage, creates a complex uncertainty. When the DAI (Dynamic Available Capacity Index) is low, the buffering capacity of the energy storage system decreases, making it unable to effectively absorb excess wind and solar energy or compensate for power output gaps, thus exacerbating the risk of supply-demand imbalance. To address this, this invention proposes a dynamic available capacity index; simulation results are available in [reference needed]. Figure 3 , Figure 3 The figure shows the distribution of the dynamic available capacity index of the energy storage system at the end of the simulation. The horizontal axis represents the energy storage system number, and the vertical axis represents the DCI value. The results show that the average DCI value of the five energy storage systems is 0.92, indicating that the capacity decay caused by long-term charge-discharge cycles has reduced the overall available capacity by 8%. The closer the DCI value is to 1, the lower the degree of energy storage capacity decay and the more effective energy the system can utilize. Conversely, a lower DCI value means that the energy storage regulation capability is significantly weakened due to capacity decay.

[0105] See Figure 4The model showed a high goodness of fit (R²=0.964), indicating a strong correlation and verifying the significant negative correlation between DCI and EENS. Therefore, energy storage capacity degradation is the core cause of decreased reliability in virtual power plants. By introducing a dynamic clustering algorithm and the DCI index, the model can accurately quantify the degradation effect, providing a basis for differentiated scheduling of energy storage systems (such as prioritizing the use of ESS with higher DCI for high-frequency regulation) and lifecycle management. This effectively balances reliability and economy in scenarios with increased uncertainty in wind and solar power.

[0106] In response to the inherent intermittent nature of photovoltaic and wind power generation, this invention integrates a risk value assessment system. Under the premise of ensuring that the probability of the system meeting load demand is not lower than the confidence level, it achieves proactive defense against extreme scenario risks in the operation of virtual power plants, and achieves a dynamic balance between power supply security and economy.

[0107] In the simulation results, when the confidence level is set to β When the coefficient of performance (COP) is 0.95, the calculated value at risk (VaR) is 3 MWh. This means that with a 95% probability, the system's unsupplied energy will not exceed 3 MWh in extreme cases, effectively controlling the potential power shortage risk within an acceptable range. This result fully demonstrates the virtual power plant's ability to cope with fluctuations in renewable energy output and the dynamic degradation effect of the energy storage system under the constraint of decision-dependent risk. It also proves that the model has high resilience in the face of various extreme situations.

[0108] Figure 5 The output power of five wind turbines under different conditions was demonstrated. The output of wind turbines is affected by the dual uncertainties of random wind speed and equipment failure.

[0109] Figure 6 The demonstration showcased the output power of five photovoltaic (PV) units under different conditions, highlighting the significant environmental constraints on PV power generation. Given such substantial uncertainties in wind and solar power generation, relying solely on traditional deterministic models in virtual power plant scheduling and resource allocation often fails to capture operational risks in extreme scenarios. By setting a confidence level, the project ensures that, with a probability of up to 95% or higher, the risk of power outages due to renewable energy shortages or equipment failures remains manageable. This allows the virtual power plant to achieve supply-demand balance and stable operation even under extreme uncertainty.

[0110] The model achieved a supply-demand balance constraint satisfaction rate of 98.7%, indicating that the system can achieve accurate power matching and load compensation in most simulation scenarios. Combined with distributed dynamic clustering algorithms and capacity decay, these results not only validate the scientific validity of the model's theoretical construction but also provide strong data support for optimizing the economic benefits and operational safety of virtual power plants.

[0111] 1.1 Reliability Indicator Results and Scenario Comparison Analysis To verify the impact of dynamic clustering and capacity decay of energy storage on system voltage and frequency, this invention sets up two scenarios as follows: Scenario A: Dynamic clustering and capacity decay of energy storage are not considered; Scenario B: Consider dynamic clustering and capacity decay of energy storage.

[0112] Solving the above two scenarios yields the following results, see [link / reference] Figure 7 and Figure 8 , Figure 7 The simulation demonstrates system voltage fluctuations under two scenarios. In Scenario A, the system bus voltage exhibits significant fluctuations throughout the simulation. This indicates that during fluctuations in renewable energy output or failures in conventional generating units, the lack of a dynamic clustering strategy hinders efficient coordinated scheduling of the energy storage system, preventing timely power replenishment or reduction and resulting in poor voltage control. Furthermore, because the model does not consider capacity decay, energy storage units are assumed to maintain their rated capacity, neglecting the decrease in actual usable capacity after multiple charge-discharge cycles. This may lead to more severe voltage fluctuations in actual operation. In contrast, Scenario B significantly improves voltage stability by performing hierarchical clustering scheduling of energy storage devices in real time based on normalized state of charge (SOC) and user response intentions. The introduction of a capacity decay index allows the dispatch center to more accurately assess the actual usable capacity of energy storage. By rotating or modifying the scheduling strategy for decaying energy storage units, the voltage curve oscillates within a narrow range, effectively controlling the fluctuation amplitude within a small interval.

[0113] Figure 8 The simulation demonstrates the system frequency fluctuations under two scenarios. In Scenario A, the system frequency exhibits significant dispersion and fluctuations between 49.8Hz and 50.2Hz. This indicates that when renewable energy output fluctuates significantly, the energy storage system without dynamic clustering lacks a targeted response, leading to frequent frequency overshoot or undershoot. This phenomenon is reflected in the simulation data as multiple large deviations from the target frequency. In contrast, in Scenario B, the introduction of dynamic clustering and a capacity decay model allows ESSFR and ESSBS to perform hierarchical collaborative scheduling, enabling the reasonable allocation of frequency regulation tasks in real-time. This stabilizes the system frequency within a narrower range, such as 49.9Hz to 50.1Hz, with a significant reduction in extreme value deviations. Data analysis shows that this scheduling strategy, through fine-grained management of available energy storage capacity, effectively reduces unsupplied energy and the expected gap duration under extreme conditions when facing various stochastic scenarios.

[0114] This invention utilizes the energy storage capacity of energy storage systems (ESS) and user response intentions to classify multiple ESS units into two types—one adaptable to high-frequency regulation and the other to low-frequency energy support—through a distributed dynamic clustering algorithm, thereby achieving flexible resource reconfiguration. Improvements are made to reliability indicators driven by energy storage degradation: a dynamic available capacity indicator is introduced, quantifying the long-term availability of energy storage based on the capacity degradation effect during charge-discharge cycles, thus improving the accuracy of reliability assessment. Furthermore, a probabilistic model of wind and solar power output and energy storage is constructed using Monte Carlo simulation, and a virtual power plant reliability assessment model considering decision-dependent risks is proposed, effectively quantifying the risk of supply-demand imbalance under extreme conditions.

[0115] Based on the same inventive concept, this application provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of a response reliability assessment method for a virtual power plant. The memory may include main memory, such as high-speed random access memory, or it may also include non-volatile memory, such as at least one disk storage device. The processor, network interface, and memory are interconnected via an internal bus, which may be an industry-standard architecture bus, a peripheral component interconnection standard bus, an extended industry-standard architecture bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. The memory stores the program; specifically, the program may include program code, which includes computer operation instructions. The memory may include main memory and non-volatile memory, and provides instructions and data to the processor.

[0116] Based on the same inventive concept, embodiments of this application provide a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the steps of a response reliability assessment method for a virtual power plant. Specifically, the computer-readable storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. The volatile memory may include RAM (Random Access Memory) and / or cache memory, etc. The non-volatile memory may include ROM (Read-Only Memory), hard disk, flash memory, optical disk, magnetic disk, etc.

[0117] Based on the same inventive concept, this application provides a computer program product, which includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions that, when executed by a computer device, cause the computer device to perform the steps of the above-described virtual power plant response reliability assessment method.

[0118] Those skilled in the art will understand that embodiments of the present invention can be provided as methods or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention 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 (Compact Disc Read-Only Memory), optical storage, etc.) containing computer-usable program code.

[0119] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as 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 apparatus or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0120] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer device 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.

[0121] These computer program instructions may also be loaded onto a computer device or other programmable data processing equipment to cause a series of operational steps to be performed on the computer device or other programmable equipment to produce a process implemented by the computer device, thereby providing instructions that execute on the computer device 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.

[0122] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0123] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for evaluating the response reliability of a virtual power plant, characterized in that, Includes the following steps: S1. Based on Monte Carlo simulation, a random state sequence is generated for the virtual power plant, which includes the output of new energy sources and the state of the energy storage system. S2. Obtain the energy storage capacity and user response intention of each energy storage system, and dynamically divide the energy storage system into high-frequency frequency regulation energy storage system and low-frequency energy support energy storage system using a distributed dynamic clustering algorithm; construct an energy storage system execution model, which executes tasks according to the dynamic division results of the energy storage system. If the energy storage system is a high-frequency frequency regulation energy storage system, then the energy storage system executes the frequency and voltage regulation task; when the energy storage system is a low-frequency energy support energy storage system, then the energy storage system executes the energy support task. A dynamic available capacity index is introduced into the energy storage system execution model to quantify the energy storage capacity decay, and the dynamic partitioning results of the energy storage system are updated based on the real-time value of the dynamic available capacity index. S3. Based on the random state sequence, the dynamic partitioning results of the energy storage system, and the execution model of the energy storage system, calculate the response reliability index of the virtual power plant.

2. The response reliability assessment method for a virtual power plant according to claim 1, characterized in that, The energy storage capacity and user response intention characteristics of each energy storage system are obtained. Combined with a distributed dynamic clustering algorithm, the energy storage systems are dynamically divided into high-frequency frequency regulation energy storage systems and low-frequency energy support energy storage systems. Includes the following steps: For each energy storage system, a feature vector containing normalized energy storage capacity and user response intention is constructed; Each energy storage system updates its local estimate of the cluster center iteratively based on a distributed average update formula through neighborhood communication. Each energy storage system determines whether it belongs to a high-frequency frequency regulation energy storage system or a low-frequency energy support energy storage system based on the updated cluster center and weighted Euclidean distance.

3. The response reliability assessment method for a virtual power plant according to claim 1, characterized in that, The dynamic available capacity metrics are as follows: in, This represents the k-th time state in the Monte Carlo state simulation. for Dynamic available capacity metrics; The ESS charge / discharge energy for state k; ( ) represents the natural exponential function; This represents the initial rated capacity of the energy storage system. This is the capacity decay coefficient; The cycle depth decay index; The duration of the Monte Carlo state.

4. The response reliability assessment method for a virtual power plant according to claim 1, characterized in that, Based on the random state sequence, the dynamic partitioning results of the energy storage system, and the execution model of the energy storage system, the response reliability index of the virtual power plant is calculated, specifically including the following steps: Construct a two-level optimization model with the upper-level objective of minimizing the total operating cost of the virtual power plant and the clustering error, and the lower-level objective of maximizing the system response reliability. In the upper-level optimization of the two-level optimization model, a risk value constraint is introduced. If the risk value meets the preset confidence level, proceed to the lower-level optimization; otherwise, return to step S1. The random state sequence, the dynamic partitioning result of the energy storage system, and the execution model of the energy storage system are input into the two-layer optimization model for solution, and the response reliability index of the virtual power plant is calculated. Determine whether the preset number of random state simulations has been reached. If yes, output the current response reliability index; otherwise, return to step S1.

5. The response reliability assessment method for a virtual power plant according to claim 4, characterized in that, The specific value at risk constraint is as follows: in, Risk value constraint; Risk threshold; β To preset the credit level; This refers to the energy not expected to be supplied. A set of risk values; It represents probability.

6. The response reliability assessment method for a virtual power plant according to claim 1, characterized in that, The specific response reliability index is as follows: in, In response to reliability indicators; This refers to the energy not expected to be supplied. This refers to the total energy that can theoretically be supplied throughout the entire time period.

7. A response reliability assessment system for a virtual power plant, characterized in that, include: The first module is used to perform stochastic state simulation of virtual power plants based on Monte Carlo simulation, generating a stochastic state sequence that includes the output of new energy sources and the state of energy storage systems. The second module is used to obtain the energy storage capacity and user response intention of each energy storage system, and dynamically divide the energy storage system into high-frequency frequency regulation energy storage system and low-frequency energy support energy storage system using a distributed dynamic clustering algorithm; it constructs an energy storage system execution model, which executes tasks according to the dynamic division results of the energy storage system. If the energy storage system is a high-frequency frequency regulation energy storage system, it is instructed to perform frequency and voltage regulation tasks; if the energy storage system is a low-frequency energy support energy storage system, it is instructed to perform energy support tasks. A dynamic available capacity index is introduced into the energy storage system execution model to quantify the energy storage capacity decay, and the dynamic partitioning results of the energy storage system are updated based on the real-time value of the dynamic available capacity index. The third module is used to calculate the response reliability index of the virtual power plant based on the random state sequence, the dynamic partitioning result of the energy storage system, and the execution model of the energy storage system.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the response reliability assessment method for the virtual power plant according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the response reliability assessment method for the virtual power plant according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the response reliability assessment method for the virtual power plant as described in any one of claims 1 to 6.

Images