Multi-cluster virtual power plant market transaction income distribution method, system and device based on improved Shapley value method and medium

Abstract

The invention discloses a multi-cluster virtual power plant market transaction income distribution method, system and device based on an improved Shapley value method and a medium, and belongs to the technical field of power systems and energy management, and the method comprises the steps: analyzing the probability distribution of historical electricity price data, constructing a representative scene set, solving a multi-virtual power plant day-ahead bidding model on this basis, and obtaining a multi-cluster virtual power plant market transaction income distribution model; and enumerating all internal cooperation combinations and calculating the income of each combination, carrying out initial income distribution by adopting a traditional Shapley value method, and finally introducing multiple contribution factors to carry out cooperation alliance income optimization distribution. According to the invention, a plurality of VPPs are aggregated to form a cluster so as to enhance the overall market bargaining capability and adjustment flexibility, and an internal energy sharing mechanism is established to realize resource complementation and cooperative operation; the uncertainty of the electricity price is processed, and a multi-scene day-ahead bidding model is constructed to ensure the strategy robustness; multiple factors such as the energy contribution degree and the profit contribution degree are innovatively introduced in the profit distribution link to improve a traditional Shapley value method, and fair and reasonable profit distribution is achieved.

Description

Technical Field

[0001] This invention relates to the field of power system and energy management technology, specifically to a method, system, equipment, and medium for distributing market transaction revenues of multi-cluster virtual power plants based on an improved Shapley value method. Background Technology

[0002] With the continuous integration of distributed energy resources (such as photovoltaics, wind power, energy storage, and adjustable loads), traditional centralized power systems are gradually evolving towards distributed, flexible, and intelligent systems. In this process, Virtual Power Plants (VPPs), as an important form of aggregating various distributed resources and participating in the electricity market, are receiving increasing attention. VPPs use information technology to uniformly schedule and optimize the management of their subordinate distributed power sources, energy storage systems, and controllable loads, thereby improving resource utilization efficiency and enhancing the system's economy and reliability.

[0003] Currently, research on VPP market trading mechanisms mainly focuses on the operation and optimization of a single VPP as an independent market participant, with the primary goal of maximizing its own economic benefits. However, facing a complex environment characterized by large fluctuations in electricity market prices, strong volatility in distributed resources, and intense competition among participants, a single VPP faces challenges in practical operation, including insufficient market competitiveness, limited regulatory capacity, and difficulty in internalizing volatility risks. Meanwhile, with the development of multi-VPP systems, how to achieve group collaboration while ensuring individual independence has become a key challenge in current research and practice.

[0004] Furthermore, in the context of multi-VPP collaborative operation, the issue of revenue distribution is particularly critical. The traditional Shapley value method, due to its theoretical completeness and fairness, is often used for revenue sharing in multi-participant collaborations. However, in actual electricity market trading scenarios, VPPs not only exhibit differences in marginal contributions but also face complex factors such as differences in energy sharing behavior, regulatory contributions, and risk-bearing. Simply applying the original Shapley model often fails to accurately reflect the actual roles of each party in the collaboration, potentially leading to unfair distribution and weakening the stability and incentive effect of the collaboration.

[0005] In summary, there is an urgent need to establish a transaction model for multi-VPP cluster collaborative participation in the market, and to design a more reasonable revenue distribution mechanism based on actual operating characteristics, so as to improve overall operating efficiency, promote resource sharing, and enhance system fairness and stability. Summary of the Invention

[0006] In view of the above-mentioned problems, the present invention is proposed.

[0007] Therefore, the technical problem solved by this invention is that current research and applications on VPP participation in electricity market transactions mainly focus on a single VPP participating in market bidding and dispatch as an independent entity. Although this can improve the integration capability and economic efficiency of distributed energy to a certain extent, it still has the following shortcomings in practical applications:

[0008] (1) Limited market competitiveness: The scale and regulation capacity of a single VPP are relatively limited. When facing large-scale market players (such as traditional power supply companies or aggregators), it lacks effective bargaining power and flexible scheduling methods, making it difficult to obtain considerable market benefits.

[0009] (2) Lack of coordination mechanism: Multiple VPPs usually operate independently, lacking energy interaction and coordination optimization mechanisms, and cannot realize the synergistic benefits of cluster operation. Resource complementarity and system redundancy capabilities are difficult to exploit.

[0010] (3) Single revenue distribution method: Even if there is a preliminary multi-VPP cooperation model, the traditional revenue distribution is mostly based on the original Shapley value method of marginal contribution, which fails to consider factors such as the degree of energy sharing and profit elasticity, which may cause the distribution results to be unbalanced and affect the sustainability and enthusiasm of cluster cooperation.

[0011] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for distributing market transaction revenues of multi-cluster virtual power plants based on the improved Shapley value method, comprising,

[0012] Historical market electricity price data is acquired and its probability distribution is analyzed to construct an electricity price scenario set and representative samples are built. Based on the electricity price scenario set, the day-ahead bidding model of multiple virtual power plants under multiple scenarios is solved. All combinations within the multi-virtual power plant cluster are listed and the revenue under each combination is calculated. The traditional Shapley value of each virtual power plant is calculated based on the revenue under each combination. Based on the traditional Shapley value, the revenue distribution of the cooperative alliance with multiple contribution factors is optimized.

[0013] As a preferred embodiment of the multi-cluster virtual power plant market transaction revenue distribution method based on the improved Shapley value method described in this invention, the probability distribution analysis includes:

[0014] Quantify the statistical characteristics of electricity prices based on historical data to capture the inherent uncertainty of historical data.

[0015] The virtual electricity price resource type is determined based on the inherent uncertainty of historical data, and the stochastic pattern of electricity prices is described.

[0016] An initial scenario is generated based on the virtual electricity price resource type.

[0017] As a preferred embodiment of the multi-cluster virtual power plant market transaction revenue distribution method based on the improved Shapley value method described in this invention, the generation of the initial scenario includes:

[0018] The probability distribution of electricity prices is divided into equal-length intervals to generate probability intervals.

[0019] Randomly select within a certain probability range.

[0020] Random variables are sampled using a probability distribution function to obtain the scenario value of electricity price.

[0021] Different scenario samples are obtained by repeatedly sampling randomly and sampling random variables.

[0022] As a preferred embodiment of the multi-cluster virtual power plant market transaction revenue allocation method based on the improved Shapley value method described in this invention, the representative sample construction includes:

[0023] Scenes are selected from scene samples using random sampling as initial cluster centers.

[0024] Calculate the distance from each scene to the cluster center.

[0025] Repeatedly assign each scene to the cluster center with the smallest Euclidean distance, calculate the mean of each class, and use it as the new cluster center until the cluster centers no longer change.

[0026] The probability estimation method based on sample frequency calculates the proportion of the original samples represented by each cluster center and uses it as the weight coefficient for representative scenarios.

[0027] As a preferred embodiment of the multi-cluster virtual power plant market transaction revenue allocation method based on the improved Shapley value method described in this invention, the step of solving the multi-virtual power plant day-ahead bidding model under multiple scenarios includes,

[0028] A bidding model for multiple VPPs in the day-ahead market is developed based on virtual electricity pricing resource types, denoted as follows:

[0029]

[0030] Among them, C DA，i Let λ represent the day-ahead total market revenue of the i-th virtual power plant. DA This indicates that the current electricity price forecast is uncertain. (P) DA，i S represents the electricity traded by the i-th virtual power plant. DA,i G DA,i Let C represent the day-ahead electricity sales and day-ahead electricity purchases of the i-th virtual power plant, respectively. CG，i Let C represent the operating cost of the distributed generator set of the i-th virtual power plant.IL，i Let C represent the IL user call cost for the i-th virtual power plant. TL，i Let C represent the cost of calling the i-th virtual power plant TL user. ES，i Let represent the energy storage operating cost of the i-th virtual power plant.

[0031] By adding constraints based on uncertainty, a multi-virtual power plant market bidding model is constructed, which considers the impact of market electricity price uncertainty. This model is represented as follows:

[0032]

[0033] Where k represents the scenario, C DA，i，k π represents the day-ahead market revenue of the i-th virtual power plant in the k-th scenario. k Represented as λ DA，k Let P represent the day-ahead forecast electricity price for the k-th scenario. DA，i，k Let C represent the traded electricity volume of the i-th virtual power plant in the k-th scenario. CG，i，k Let C represent the operating cost of the distributed generator unit of the i-th virtual power plant in the k-th scenario. IL，i，k Let C represent the IL user call cost for the i-th virtual power plant in the k-th scenario. TL，i，k Let C represent the TL user call cost for the i-th virtual power plant in the k-th scenario. ES，i，k Let S represent the energy storage operating cost of the i-th virtual power plant in the k-th scenario. DA,i，k G DA,i，k Let these represent the daily electricity sales and daily electricity purchases of the i-th virtual power plant in the k-th scenario, respectively.

[0034] The beneficial effects of this preferred technical solution are that, by constructing a multi-scenario bidding model that considers the uncertainty of market electricity prices, the present invention effectively improves the decision-making robustness and economy of virtual power plant clusters in the day-ahead market. This model represents electricity price fluctuations as multiple probabilistic scenarios and comprehensively optimizes the total revenue and cost structure under each scenario in the objective function, enabling the bidding strategy to cover various potential market states and reducing the revenue risk caused by electricity price prediction deviations.

[0035] As a preferred embodiment of the multi-cluster virtual power plant market transaction revenue distribution method based on the improved Shapley value method described in this invention, the calculation of the traditional Shapley value of each virtual power plant includes:

[0036] The number of virtual power plant users is defined, meaning the total number of combinations within a multi-virtual power plant cluster is 2. N -1, where each member's reward is represented as,

[0037]

[0038] Among them, Ri Let R(U) represent the revenue of the i-th virtual power plant, U represent the set of virtual power plants, R(U) represent the overall value of the combination U, and R(U-{i}) represent the value of the combination U excluding member vpp. i The subsequent benefits This represents the probability of a combination occurring.

[0039] The beneficial effect of this preferred technical solution is that by introducing the traditional Shapley value to quantify and allocate the cooperative benefits of the virtual power plant alliance, the present invention effectively ensures the fairness and rationality of the benefit distribution.

[0040] As a preferred embodiment of the multi-cluster virtual power plant market transaction revenue distribution method based on the improved Shapley value method described in this invention, the optimization of revenue distribution through a cooperative alliance with multiple contribution factors includes:

[0041] Calculate the energy and profit contributions of each virtual power plant, and then perform a secondary distribution of revenue using the improved Shapley value method.

[0042] The energy contribution is expressed as,

[0043]

[0044] Where, θ i P represents the energy contribution. i-j This represents the power exchanged between the i-th virtual power plant and the j-th virtual power plant.

[0045] The profit contribution rate is expressed as,

[0046]

[0047] Where, ω i R(U-{i}) represents the profit contribution rate, and R(U-{i}) represents the revenue of portfolio U after removing the i-th virtual power plant of member.

[0048] The revenue in the Shapley value method is redistributed secondaryly based on energy contribution and profit contribution, with the secondary distribution coefficient β. i Defined as:

[0049]

[0050] Where η1 and η2 are contribution weight coefficients, satisfying η1+η2=1.

[0051] After considering the secondary allocation factor, the revenue of each VPP member is:

[0052] R′ i =R i +R(N)β i

[0053] Among them, R′ i R(N) represents the revenue after secondary distribution, and R(N) represents the revenue from the participation of all employees in the market.

[0054] The beneficial effects of this preferred technical solution are as follows: By introducing dual indicators of energy contribution and profit contribution to the traditional Shapley value method for weighted optimization, this invention achieves a more refined and fairer distribution of revenue among multi-virtual power plant alliances. It not only considers the physical contribution of each member in energy interaction but also comprehensively reflects their marginal profit-creating capabilities in different cooperation combinations, overcoming the limitations of traditional methods that rely solely on marginal revenue for allocation. By constructing a secondary allocation coefficient based on multi-factor weighting, the final revenue distribution result simultaneously reflects the actual contributions of virtual power plants in both energy output and economic value, enhancing the internal collaborative efficiency and member cooperation stability of the alliance, and providing an effective profit distribution mechanism for multi-entity collaborative operation in complex market environments.

[0055] This invention provides a system for allocating market transaction revenues for multi-cluster virtual power plants based on an improved Shapley value method.

[0056] To address the aforementioned technical problems, this invention provides the following technical solution: a system for allocating market transaction revenues of multi-cluster virtual power plants based on an improved Shapley value method, comprising: a data acquisition and scenario generation module, a day-ahead bidding optimization module, a revenue calculation module, a traditional Shapley value calculation module, and a multi-contribution factor optimization allocation module.

[0057] The data acquisition and scenario generation module acquires historical data on market electricity prices and analyzes the probability distribution to construct an electricity price scenario set and build representative samples.

[0058] The day-ahead bidding optimization module solves the day-ahead bidding model for multiple virtual power plants under multiple scenarios based on the electricity price scenario set.

[0059] The revenue calculation module lists all combinations within the multi-virtual power plant cluster and calculates the revenue for each combination.

[0060] The conventional Shapley value calculation module calculates the conventional Shapley value of each virtual power plant based on the revenue under each combination.

[0061] The multi-contribution factor optimization allocation module optimizes the distribution of cooperative alliance benefits based on the traditional Shapley value.

[0062] The present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, characterized in that the processor executes the computer program to implement the steps of the multi-cluster virtual power plant market transaction revenue distribution method based on the improved Shapley value method.

[0063] The present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps of the method for distributing market transaction revenues of multi-cluster virtual power plants based on the improved Shapley value method.

[0064] The beneficial effects of this invention are as follows: This invention enables multiple VPPs to participate in the electricity market in a cluster form. Through resource aggregation and coordinated scheduling, it enhances the overall regulation capability and market bargaining power, and overcomes the shortcomings of individual VPPs in terms of scale, flexibility and market adaptability.

[0065] By establishing an energy-sharing mechanism among VPPs, resource complementarity and collaborative operation can be achieved, effectively mitigating volatility risks and improving the cluster's operational efficiency and market returns.

[0066] This invention uses scenario analysis to model the randomness of market electricity prices, improves the adaptability of bidding strategies to market fluctuations, and ensures that the optimization results are robust and feasible under multiple price scenarios.

[0067] Based on the traditional Shapley value method, this method comprehensively considers the energy contribution and profit contribution of VPPs to achieve a fairer and more reasonable distribution of benefits, thereby enhancing the cooperation enthusiasm and long-term stability of all participants in the cluster. Attached Figure Description

[0068] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0069] Figure 1 The overall flowchart of a method for allocating market transaction revenues of multi-cluster virtual power plants based on an improved Shapley value method, provided as an embodiment of the present invention, is shown below.

[0070] Figure 2 This is an overall framework diagram of a multi-virtual power plant market transaction revenue distribution system based on the improved Shapley value method, provided as an embodiment of the present invention. Detailed Implementation

[0071] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0072] Example 1, referring to Figure 1 This is one embodiment of the present invention, which provides a method for allocating market transaction revenues for multi-cluster virtual power plants based on the improved Shapley value method, including:

[0073] S1. Obtain historical data on market electricity prices and analyze the probability distribution to construct a set of electricity price scenarios and build representative samples;

[0074] S2. Based on the electricity price scenario set, solve the day-ahead bidding model of multiple virtual power plants under multiple scenarios;

[0075] S3. List all combinations within the multi-virtual power plant cluster and calculate the revenue for each combination;

[0076] S4. Calculate the traditional Shapley value of each virtual power plant based on the revenue under each combination;

[0077] S5. Based on the traditional Shapley value, optimize the distribution of cooperative alliance benefits through multiple contribution factors.

[0078] This invention provides a robust foundation for decision-making by constructing a representative set of scenarios that accurately reflects the uncertainty of electricity prices. Based on this, multi-scenario collaborative optimization is carried out to maximize the global returns of the day-ahead bidding strategy for virtual power plant clusters. Finally, by introducing a cooperative game allocation model with multiple contribution factors, the invention ensures the total returns of the alliance while achieving accurate quantification and fair allocation of the contributions of individual members, thereby effectively incentivizing members to continue participating in cooperation and ensuring long-term stability.

[0079] Example 2 is an embodiment of the present invention. Based on the previous embodiment, it provides a method for allocating market transaction revenue of multi-cluster virtual power plants based on the improved Shapley value method, including:

[0080] In this embodiment, the representative sample construction in S1 involves scene reduction and weighting. Through mathematical optimization, a small number of statistically representative core scenes are selected from a massive amount of original scenes, and each retained scene is assigned a corresponding probability weight. This significantly reduces the computational complexity of the subsequent optimization model while preserving the key features of the original probability distribution to the greatest extent possible. Specifically, scene reduction employs a probability distance-based clustering algorithm, iteratively optimizing and eliminating redundant scenes. The weighting operation calculates the weight coefficient based on the sample proportion of the original scene cluster represented by each retained scene, ensuring that the reduced scene set maintains a high degree of consistency with the initial scene set in terms of probabilistic statistical characteristics.

[0081] In one alternative implementation, representative sample construction can employ a scene reduction algorithm based on probabilistic distance. A distance matrix is ​​constructed by calculating the probabilistic distances between scenes in the initial scene set. Subsequently, through iterative forward selection or backward reduction, the most representative subset of scenes is selected with the optimization objective of minimizing the overall probabilistic distance between the reduced scene set and the initial scene set. Finally, based on the selection results, each retained scene is assigned a weight based on the sum of probabilities of the original scenes it represents.

[0082] In another alternative implementation, the construction of representative samples can also be achieved by applying a clustering analysis algorithm to divide all initial scenes into a preset number of categories; then, the cluster center of each category is calculated, and the center point is used as the representative scene of that category; finally, the proportion of the number of original scenes contained in each category to the total number of scenes is used as the weight coefficient of its corresponding representative scene, thereby completing the reduction and weighting of the scene set.

[0083] Furthermore, the analysis of the probability distribution in S1 includes the following steps A1-A3:

[0084] A1. Quantify the statistical characteristics of electricity prices based on historical data to capture the inherent uncertainty of historical data;

[0085] A2. Determine the virtual electricity price resource type based on the inherent uncertainty of historical data, and describe the stochastic pattern of electricity prices;

[0086] Furthermore, the types of virtual electricity pricing resources include photovoltaic, wind power, distributed generator sets, interruptible loads, transferable loads, and energy storage.

[0087] (1) Photovoltaics:

[0088] Photovoltaic power generation is a renewable energy technology that directly converts solar radiation into electrical energy using solar cells. It boasts advantages such as being environmentally friendly, having low operating costs, and high reliability. Its output power is significantly affected by environmental factors such as light intensity and temperature. The output power model is shown below:

[0089]

[0090] Among them, P PV Photovoltaic output power; P STC G STC T STC These represent the maximum output power, light intensity, and ambient temperature under standard test conditions, respectively. G and T represent the light intensity and temperature under the current environment, and k is the temperature coefficient, which is usually taken as -0.45.

[0091] (2) Wind power:

[0092] Wind power generation uses wind energy to drive a wind turbine, which then converts mechanical energy into electrical energy through a generator. The functional relationship between wind power generation and wind speed is shown below:

[0093]

[0094] k2=-k1v ci (4)

[0095] Among them, P WT P is the output power of the fan. r It is the rated power of the wind turbine, v ci v co v r v and v represent the cut-in wind speed, cut-out wind speed, rated wind speed, and actual wind speed of the wind turbine, respectively. k1 and k2 are constants for calculating wind power generation, which are determined based on the rated power and wind speed of the wind turbine.

[0096] (3) Distributed generator sets:

[0097] Controllable power generation equipment includes diesel generators and small coal-fired units, featuring rapid start-up and load adjustability. The power generation cost of distributed generator sets is:

[0098] C CG =aP CG +b(5)

[0099] Among them, C CG P CG The operating cost and actual output are given by , where a and b represent the operating cost coefficients of the unit, respectively.

[0100] The constraints to be satisfied include unit output constraints and ramp-up constraints:

[0101]

[0102] in, These represent the upper and lower limits of the unit's output. These represent the upper and lower limits of the unit's ramp rate, P and P respectively. CG (t) represents the unit output at time t, P CG (t-1) represents the unit output at time t-1.

[0103] (4) Interruptible load (IL):

[0104] Interruptible loads refer to load resources that can be temporarily reduced or disconnected under specific operating conditions, and scheduling rights are usually obtained through user subscriptions.

[0105] C IL =P IL c IL(8)

[0106]

[0107] Among them, C IL Indicates the cost of IL user calls, c IL P represents the unit compensation cost for IL user load interruption, respectively. IL This indicates the amount of load interruption for IL users. These represent the maximum interruption limit for the load.

[0108] (5) Transferable load (TL):

[0109] Transferable loads refer to loads that can be flexibly adjusted in terms of time or space, such as industrial electricity, refrigeration systems, and centralized air conditioning, which are load types with dispatch flexibility. By effectively managing transferable loads, the peak-to-valley difference in electricity supply can be reduced, the economy and stability of the power system can be improved, and excessive load during peak periods can be avoided.

[0110] C TL =(P TL,up +P TL,down )c TL (10)

[0111]

[0112] Among them, C TL Indicates the call cost between IL users and TL users, c TL These represent the unit compensation cost for load transfer of TL users; P TL,up P TL,down This indicates the load increase and decrease for TL users, and the increase and decrease should be the same throughout the entire scheduling cycle. This indicates the maximum upper limit of load transfer.

[0113] (6) Energy storage:

[0114] Energy storage systems are crucial regulation resources within a VPP (Virtual Power Supply), possessing bidirectional energy regulation capabilities. They can charge during off-peak hours and discharge during peak hours, participating in tasks such as peak-valley arbitrage, system backup, and frequency regulation. The operating costs and constraints of energy storage devices are shown below.

[0115] C ES =(P ES,c +P ES,d )c ES (14)

[0116] 0≤P ES,d ≤P ES,max (15)

[0117] 0≤P ES,c≤P ES,max (16)

[0118] E(t+1)=E(t)+P ES,c (t)η ES -P ES,d (t) / η ES (17)

[0119] E min ≤E≤E max (18)

[0120] E(0)=E(T)(19)

[0121] Among them, C ES For energy storage operating costs, P ES,c P ES,d These represent the energy storage charging and discharging power, c ES P represents the unit operating cost of energy storage. ES,max E represents the upper limit of energy storage charging and discharging, and E represents the energy storage capacity. min E max η represents the upper and lower limits of energy storage capacity, respectively. ES This refers to the energy storage charging and discharging efficiency.

[0122] A3. Generate the initial scenario based on the virtual electricity price resource type.

[0123] In this application's implementation, capturing the inherent uncertainty of historical data in A1 involves using probability distribution fitting and random scenario generation techniques to quantify the randomness, temporal correlation, spatial correlation, and complex statistical characteristics of electricity price fluctuations, such as their peaks and heavy tails, into a series of discrete but probabilistically controllable future possible states. Specifically, firstly, by conducting rigorous statistical tests and feature analysis on historical electricity price data, its inherent fluctuation patterns and dependency structures are identified. Subsequently, based on the analysis results, a probabilistic model capable of simultaneously capturing the marginal distribution and joint dependency structure of electricity price fluctuations is selected or constructed. Finally, through random sampling and simulation of this model, a large-scale initial scenario set that comprehensively and accurately reflects historical uncertainty patterns is generated.

[0124] In one alternative implementation, capturing the inherent uncertainty of historical data can be achieved using parameter estimation. First, a theoretical probability distribution model matching the statistical characteristics of the historical data is selected. Then, the parameters of the distribution model are calculated using methods such as maximum likelihood estimation or moment estimation. Finally, sampling is performed using the fitted parameterized distribution model to generate a large number of initial random scenarios, thereby capturing the implicit random patterns in the historical data in a mathematical form.

[0125] In another alternative implementation, capturing the inherent uncertainty of historical data can be achieved using nonparametric estimation methods. Specifically, instead of pre-setting a specific distribution, scenarios can be generated directly by resampling or smoothing based on the empirical distribution characteristics of historical data. For example, kernel density estimation techniques can be applied to approximate the true probability density curve of historical electricity prices through a smoothed kernel function; or bootstrap sampling can be used to construct multiple bootstrap sample sets to simulate the uncertainty of electricity prices by randomly resampling historical samples with replacement, thereby avoiding biases caused by model misspecification.

[0126] Common methods for generating scenarios include subjective methods, statistical methods, and simulation methods. Subjective methods rely on the decision-maker's experience and judgment, offering flexible modeling but being susceptible to subjective bias and lacking stability. Statistical methods extract typical scenarios through cluster analysis of historical data, but require a high quantity and quality of sample data. Simulation methods sample uncertain variables based on known probability distributions, improving modeling efficiency while ensuring data representativeness, and are therefore widely used in engineering practice.

[0127] Given that market electricity prices can be considered as random variables with known or estimable probability distributions, this paper employs Latin Hypercube Sampling (LHS) to generate a set of scenarios for electricity price uncertainty. This method stratifies the probability distribution intervals of the variable and randomly selects sample points from each stratum. This ensures uniform coverage of the entire distribution space while significantly improving sampling efficiency and reducing the required sample size, thus more accurately reflecting the overall distribution characteristics of electricity prices.

[0128] Common methods for generating scenarios include subjective methods, statistical methods, and simulation methods. Subjective methods rely on the decision-maker's experience and judgment, offering flexible modeling but being susceptible to subjective bias and lacking stability. Statistical methods extract typical scenarios through cluster analysis of historical data, but require a high quantity and quality of sample data. Simulation methods sample uncertain variables based on known probability distributions, improving modeling efficiency while ensuring data representativeness, and are therefore widely used in engineering practice.

[0129] Furthermore, generating the initial scene in step A3 includes the following steps A31-A34:

[0130] A31. Divide the probability distribution of electricity prices into equal-length intervals to generate probability intervals;

[0131] Furthermore, the probability distribution of electricity prices is divided into N equal parts, with each interval having a length of 1 / N, represented as [0, 1 / N], [1 / N, 2 / N]...

[0132] A32. Randomly select within a certain probability interval;

[0133] Furthermore, a number n is randomly selected within the i-th probability interval. i The extraction method is Where r is a random number on [0,1], and N represents the number of intervals in A31;

[0134] A33. Random variable sampling is performed using a probability distribution function to obtain the scenario value of electricity price;

[0135] The scenario value of electricity price is s i s i =F -1 (n i ).

[0136] A34. Repeated random sampling and random variable sampling are used to obtain different scenario samples.

[0137] Furthermore, repeat the above steps N times to obtain N sets of scene samples.

[0138] In this embodiment, the random variable sampling in step A33 involves an inverse transformation. First, the cumulative probability distribution function of the electricity price is obtained by fitting historical data. Then, a random number uniformly distributed in the interval [0,1] is generated. Finally, this random number is substituted as a probability value into the inverse function of the cumulative distribution function to solve for the corresponding sampled electricity price value. This method ensures that the generated random sample strictly follows the fitted probability distribution.

[0139] In one alternative implementation, random variable sampling can begin by selecting an easily sampled reference distribution and determining a constant such that the ratio of the density function of the reference distribution to that of the target probability distribution is always less than this constant. Then, random samples are generated from the reference distribution, and a uniformly random number is generated. If this uniformly random number is less than the ratio of the density function of the target distribution to the constant multiple of the density function of the reference distribution, the sample is accepted; otherwise, it is rejected and resampling is performed. This method is suitable for sampling complex distributions where the inverse function is difficult to calculate directly.

[0140] In another alternative implementation, random variable sampling can also be achieved by constructing a Markov chain with a stationary distribution equal to the target probability distribution, and after several state transitions, drawing samples from this Markov chain as approximate samples of the target distribution. The commonly used Metropolis-Hastings algorithm generates candidate samples through a proposal distribution and accepts the candidate state with a specific probability, thereby achieving asymptotic sampling of complex high-dimensional probability distributions.

[0141] Since the large number of scenarios generated by LHS can lead to excessively high dimensionality and significant computational overhead in model solving, it is necessary to reduce the original scenario set to improve computational efficiency while maintaining representativeness. Therefore, this paper employs the K-means clustering algorithm to perform cluster analysis on the electricity price scenarios sampled from LHS, grouping similar scenarios into the same category and using the cluster center value to represent the typical scenario of that category. By setting an appropriate number of clusters, the number of scenarios is compressed while preserving as much uncertainty as possible, thereby reducing the computational complexity of subsequent model solving.

[0142] Furthermore, the construction of representative samples in S1 includes the following steps B1-B4:

[0143] B1. Select scenes from the scene samples as initial cluster centers through random sampling.

[0144] Furthermore, by using a random sampling method, K scenes are selected from the N sets of scene samples generated by LHS as initial cluster centers;

[0145] B2. Calculate the distance from each scene to the cluster center;

[0146] Furthermore, the distance from each scene to the cluster center is calculated based on Euclidean distance, where Euclidean distance is defined as follows:

[0147]

[0148] Where X and Y are both M-dimensional vectors, x m y m Let m be the m-th term of an M-dimensional vector.

[0149] B3. Repeatedly assign each scene to the cluster center with the smallest Euclidean distance, calculate the mean of each class, and use it as the new cluster center until the cluster centers no longer change.

[0150] Furthermore, each scene is assigned to the cluster center with the smallest Euclidean distance, and the mean of each cluster is calculated and used as the new cluster center.

[0151] B4. A probability estimation method based on sample frequency is used to calculate the proportion of the original samples represented by each cluster center and use it as a weight coefficient for representative scenarios.

[0152] After scene reduction is completed, each retained typical scene needs to be assigned a corresponding occurrence probability weight π. kThis paper employs a probability estimation method based on sample frequency: the proportion of the original samples represented by each cluster center is statistically analyzed and used as the weight coefficient for that representative scenario. This approach effectively reflects the probability structure of the original electricity price distribution within the reduced scenario set, enabling the optimization model to reasonably assess the expected returns and risks under various electricity price conditions.

[0153] Furthermore, step S2 involves solving the multi-virtual power plant day-ahead bidding model under multiple scenarios, including the following steps C1-C2:

[0154] C1. A bidding model for multiple VPPs in the day-ahead market is developed based on virtual electricity price resource types, expressed as follows:

[0155]

[0156] Among them, C DA，i Let λ represent the day-ahead total market revenue of the i-th virtual power plant. DA This indicates that the current electricity price forecast is uncertain. (P) DA，i S represents the electricity traded by the i-th virtual power plant. DA,i G DA,i Let C represent the day-ahead electricity sales and day-ahead electricity purchases of the i-th virtual power plant, respectively. CG，i Let C represent the operating cost of the distributed generator set of the i-th virtual power plant. IL，i Let C represent the IL user call cost for the i-th virtual power plant. TL，i Let C represent the cost of calling the i-th virtual power plant TL user. ES，i Let $i$ be the energy storage operating cost of the $i$-th virtual power plant.

[0157] C2. Based on the uncertainty, add constraints to construct a multi-virtual power plant market bidding model that considers the impact of market electricity price uncertainty, denoted as follows:

[0158]

[0159] Where k represents the scenario, C DA，i，k π represents the day-ahead market revenue of the i-th virtual power plant in the k-th scenario. k Represented as λ DA，k Let P represent the day-ahead forecast electricity price for the k-th scenario. DA，i，k Let C represent the traded electricity volume of the i-th virtual power plant in the k-th scenario. CG，i，k Let C represent the operating cost of the distributed generator unit of the i-th virtual power plant in the k-th scenario. IL，i，k Let C represent the IL user call cost for the i-th virtual power plant in the k-th scenario. TL，i，k Let C represent the TL user call cost for the i-th virtual power plant in the k-th scenario. ES，i，kLet S represent the energy storage operating cost of the i-th virtual power plant in the k-th scenario. DA,i，k G DA,i，k Let these represent the daily electricity sales and daily electricity purchases of the i-th virtual power plant in the k-th scenario, respectively.

[0160] Furthermore, the constraints, including formulas (1)-(19), also include power balance constraints, power transmission constraints, and market transaction constraints.

[0161] The power balance constraint is expressed as follows:

[0162]

[0163] Among them, P L,i P PV,i P WT,i P represents the load, the output of photovoltaic and wind power, respectively. i-j This represents the power exchanged between the i-th virtual power plant and the j-th virtual power plant.

[0164] Power transfer constraints are expressed as follows:

[0165] -P p2p,max ≤P i-j ≤P p2p,max

[0166] Among them, -P p2p,max ≤P i-j ≤P p2p,max This indicates the upper limit of electrical energy interaction in the virtual power plant.

[0167] Market transaction constraints are represented as

[0168] 0≤G DA,i ≤P trade,max

[0169] 0≤S DA,i ≤P trade,max

[0170] Among them, P trade,max This indicates the upper limit of the amount of electricity that a VPP can trade with the electricity market.

[0171] At the same time, VPPs should also meet the restrictions on buying and selling electricity in market transactions, that is, they cannot buy and sell electricity at the same time. In the following formula, t represents the moment in the trading cycle T.

[0172] S DA,i (t)G DA,i (t)=0

[0173] It is worth noting that in the current electricity market, the market electricity price λ DAElectricity prices inherently possess significant uncertainty. As a price signal jointly determined by market supply and demand and transaction mechanisms, electricity prices often exhibit volatility and unpredictability in actual operation. This uncertainty is particularly pronounced during the day-ahead phase, due to the time lag between quoted prices and actual execution. Treating electricity prices as a fixed value when formulating bidding strategies may lead to a mismatch between the quoted strategy and the actual transaction price, thereby affecting the expected returns of VPPs and even introducing economic risks.

[0174] To address the uncertainty of electricity prices, this paper introduces scenario analysis for modeling and optimization. This method constructs a set of representative electricity price scenarios and assigns probability weights to each scenario based on historical market data or forecasts, thus transforming the original stochastic problem into a solvable multi-scenario optimization problem. This approach demonstrates good adaptability in both theory and engineering practice: on the one hand, it avoids over-assumptions about the probability distribution of electricity prices, making the modeling closer to actual market behavior; on the other hand, by optimizing under multiple possible electricity price scenarios, it can significantly improve the robustness and fault tolerance of the trading strategy, giving the VPP (Virtual Power Utility) strong adaptability under different market trends.

[0175] Therefore, the scenario analysis method not only effectively captures the decision-making risks brought about by market price fluctuations, but also provides theoretical support and operational paths for VPPs to formulate more robust and flexible bidding strategies, thereby improving overall operational efficiency and market participation capabilities.

[0176] The constraints are the operational constraints, power balance constraints, power transmission constraints, and market transaction constraints of distributed resources in each scenario.

[0177] Thus, by combining multi-scenario analysis, this invention has completed the construction of a multi-VPP market bidding model that fully considers the impact of market electricity price uncertainty.

[0178] In the day-ahead bidding model of multiple VPPs, multiple VPPs are regarded as a unified cluster, jointly participating in electricity market transactions. Within the cluster, each VPP not only collaboratively formulates market bidding strategies but also achieves internal cooperative operation through energy sharing and coordinated dispatch. This joint participation model can significantly enhance the overall regulation capacity and market competitiveness, thereby effectively overcoming problems faced by individual VPPs in actual operation, such as insufficient economies of scale, limited regulation capacity, and difficulty in internalizing fluctuation risks, thus improving the overall operational efficiency of the cluster.

[0179] However, during multi-VPP collaborative operation, since energy sharing within the cluster typically does not require additional market transaction fees, some VPPs may incur additional adjustment costs or sacrifice some autonomous operating space. While this collaboration is beneficial for improving the overall market returns of the cluster, it can also lead to uneven individual returns. Some VPPs may be at a disadvantage in the collaboration; if they are not given reasonable compensation, the incentive mechanism for cooperation will be weakened, affecting the stability and sustainability of the cluster operation. Therefore, after the market transaction is completed, how to achieve a fair, reasonable, and incentivizing distribution of returns among the VPPs becomes a key issue in the design of multi-VPP collaborative mechanisms.

[0180] The Shapley Value method, a classic approach to distributing payoffs in cooperative games, is widely used in multi-participant cooperative scenarios. Based on the principle of marginal contribution, this method calculates the marginal payoff for each member in different cooperative combinations and uses the average of the marginal contributions across all possible permutations as their share of the payoff. Shapley value allocation possesses excellent properties such as symmetry, fairness, and efficiency, theoretically guaranteeing that each participant receives a corresponding reward based on their actual contribution. Therefore, it is considered an effective means of solving the problem of distributing payoffs in multi-party cooperative games.

[0181] Furthermore, in step S4, calculating the conventional Shapley value for each virtual power plant includes:

[0182] The number of virtual power plant users is defined, meaning the total number of combinations within a multi-virtual power plant cluster is 2. N -1, where each member's reward is represented as,

[0183]

[0184] Among them, R i Let R(U) represent the revenue of the i-th virtual power plant, U represent the set of virtual power plants, R(U) represent the overall value of the combination U, and R(U-{i}) represent the value of the combination U excluding member vpp. i The subsequent benefits This represents the probability of a combination occurring.

[0185] The traditional Shapley method measures individual contribution based on marginal revenue, which has limitations when applied to multi-VPP scenarios, failing to fully reflect energy sharing behavior and actual profit contribution among virtual power plants. Therefore, this paper introduces an improved mechanism based on the traditional Shapley value method to more accurately measure the multi-dimensional contribution of each VPP in cluster operation, thereby improving the fairness and incentive effectiveness of revenue distribution.

[0186] In the implementation of this application, the optimization of the distribution of cooperative alliance benefits through multiple contribution factors in S5 is to calculate the energy contribution and profit contribution of each VPP; and to perform secondary distribution of benefits according to the improved Shapley value method. This method comprehensively considers the differentiated contributions of members in terms of output and benefits, and optimizes the traditional Shapley value by introducing contribution weight factors, so that the distribution results are more in line with the value creation law in actual cooperation.

[0187] In one alternative implementation, the optimization of profit distribution in a cooperative alliance with multiple contribution factors can be achieved using a nucleolus-based solution. Specifically, with the goal of improving the overall stability of the alliance, a linear programming approach is used to find the distribution scheme that minimizes the maximum dissatisfaction among all sub-alliances.

[0188] In an alternative implementation, the optimization of revenue distribution in a cooperative alliance with multiple contribution factors can also be based on a weighted Shapley value allocation mechanism. This method defines a weight coefficient for each VPP member to reflect its comprehensive importance in terms of energy contribution, profit contribution, and other aspects.

[0189] Furthermore, in S5, the optimization of revenue distribution through a cooperative alliance with multiple contribution factors includes calculating the energy and profit contributions of each virtual power plant and performing a secondary revenue distribution based on the improved Shapley value method.

[0190] In the current phase, multiple Virtual Power Planes (VPPs) operate collaboratively through energy interaction, sharing an energy-sharing mechanism. During this process, VPPs with more frequent energy interactions and larger energy sharing volumes typically play a more significant role in balancing supply and demand within the cluster and improving overall operational efficiency, thus contributing more to the cluster's overall performance. Therefore, an energy contribution level is defined.

[0191] The energy contribution is expressed as,

[0192]

[0193] Where, θ i P represents the energy contribution. i-j This represents the power exchanged between the i-th virtual power plant and the j-th virtual power plant.

[0194] When allocating profits, in addition to the marginal contribution considered in the Shapley value method, the profit contribution of each VPP to the entire cluster should also be considered. Define the profit contribution ω of each VPP. i It depends on the ratio of the increase in returns before and after participation to the overall increase in VPP returns under various combinations.

[0195] The profit contribution rate is expressed as,

[0196]

[0197] Where, ω i R(U-{i}) represents the profit contribution, where R(U-{i}) represents the revenue of portfolio U after removing the i-th virtual power plant from member U.

[0198] The revenue in the Shapley value method is redistributed secondaryly based on energy contribution and profit contribution, with the secondary distribution coefficient β. i Defined as:

[0199]

[0200] Where η1 and η2 are contribution weight coefficients, satisfying η1+η2=1;

[0201] After considering the secondary allocation factor, the revenue of each VPP member is:

[0202] R′ i =R i +R(N)β i

[0203] Among them, R′ i R(N) represents the revenue after secondary distribution, and R(N) represents the revenue from the participation of all employees in the market.

[0204] Example 3, referring to Figure 2 This embodiment of the present invention provides a system for the distribution of market transaction revenues of multi-cluster virtual power plants based on the improved Shapley value method, including a data acquisition and scenario generation module, a day-ahead bidding optimization module, a revenue calculation module, a traditional Shapley value calculation module, and a multi-contribution factor optimization allocation module.

[0205] The data acquisition and scenario generation module acquires historical data on market electricity prices and analyzes the probability distribution to construct a set of electricity price scenarios and build representative samples.

[0206] The day-ahead bidding optimization module solves the day-ahead bidding model for multiple virtual power plants under multiple scenarios, based on the electricity price scenario set.

[0207] The revenue calculation module lists all combinations within the multi-virtual power plant cluster and calculates the revenue for each combination.

[0208] The traditional Shapley value calculation module calculates the traditional Shapley value of each virtual power plant based on the revenue under each combination.

[0209] The multi-contribution factor optimization allocation module optimizes the distribution of cooperative alliance benefits based on the traditional Shapley value.

[0210] This embodiment also provides an electronic device applicable to the distribution method of market transaction revenue for multi-cluster virtual power plants based on the improved Shapley value method, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the distribution method of market transaction revenue for multi-cluster virtual power plants based on the improved Shapley value method proposed in the above embodiment.

[0211] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the method for distributing market transaction revenues of multi-cluster virtual power plants based on the improved Shapley value method proposed in the above embodiments.

[0212] The storage medium proposed in this embodiment and the method for distributing market transaction revenues of multi-cluster virtual power plants based on the improved Shapley value method proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0213] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0214] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for allocating market transaction revenues for multi-cluster virtual power plants based on the improved Shapley value method, characterized in that: include, Historical data on market electricity prices were obtained and their probability distribution was analyzed to construct a set of electricity price scenarios and to build representative samples. Based on a set of electricity price scenarios, we solve a day-ahead bidding model for multiple virtual power plants under multiple scenarios; List all combinations within a multi-virtual power plant cluster and calculate the revenue for each combination; Calculate the traditional Shapley value for each virtual power plant based on the revenue under each combination; Based on the traditional Shapley value, the distribution of cooperative alliance benefits is optimized through multiple contribution factors.

2. The method for distributing market transaction revenues of multi-cluster virtual power plants based on the improved Shapley value method as described in claim 1, characterized in that: The analyzed probability distribution includes, Quantify the statistical characteristics of electricity prices based on historical data to capture the inherent uncertainty of historical data; Based on the inherent uncertainty of historical data, the types of virtual electricity price resources are determined, and the stochastic patterns of electricity prices are described. An initial scenario is generated based on the virtual electricity price resource type.

3. The method for distributing market transaction revenues of multi-cluster virtual power plants based on the improved Shapley value method as described in claim 2, characterized in that: The generation of the initial scene includes, Divide the probability distribution of electricity prices into equal-length intervals to generate probability intervals; Randomly selected within a certain probability interval; Random variables are sampled using a probability distribution function to obtain the scenario value of electricity price; Different scenario samples are obtained by repeatedly sampling randomly and sampling random variables.

4. The method for distributing market transaction revenues of multi-cluster virtual power plants based on the improved Shapley value method as described in claim 3, characterized in that: The construction of the representative samples includes, Scenes were selected as initial cluster centers from scene samples using random sampling. Calculate the distance from each scene to the cluster center; Repeatedly assign each scene to the cluster center with the smallest Euclidean distance, calculate the mean of each class, and use it as the new cluster center until the cluster centers no longer change; The probability estimation method based on sample frequency calculates the proportion of the original sample number represented by each cluster center and uses it as the weight coefficient for representative scenarios.

5. The method for distributing market transaction revenues of multi-cluster virtual power plants based on the improved Shapley value method as described in claim 4, characterized in that: The solution to the multi-scenario, multi-virtual power plant day-ahead bidding model includes, A bidding model for multiple VPPs in the day-ahead market is developed based on virtual electricity pricing resource types, denoted as follows: Among them, C DA，i Let λ represent the day-ahead total market revenue of the i-th virtual power plant. DA This indicates that the current electricity price forecast is uncertain. (P) DA，i S represents the electricity traded by the i-th virtual power plant. DA,i G DA,i Let C represent the day-ahead electricity sales and day-ahead electricity purchases of the i-th virtual power plant, respectively. CG，i Let C represent the operating cost of the distributed generator set of the i-th virtual power plant. IL，i Let C represent the IL user call cost for the i-th virtual power plant. TL，i Let C represent the cost of calling the i-th virtual power plant TL user. ES，i Let $i$ be the energy storage operating cost of the $i$-th virtual power plant. By adding constraints based on uncertainty, a multi-virtual power plant market bidding model is constructed, which considers the impact of market electricity price uncertainty. This model is represented as follows: Where k represents the scenario, C DA，i，k π represents the day-ahead market revenue of the i-th virtual power plant in the k-th scenario. k Represented as λ DA，k Let P represent the day-ahead forecast electricity price for the k-th scenario. DA，i，k Let C represent the traded electricity volume of the i-th virtual power plant in the k-th scenario. CG，i，k Let C represent the operating cost of the distributed generator unit of the i-th virtual power plant in the k-th scenario. IL，i，k Let C represent the IL user call cost for the i-th virtual power plant in the k-th scenario. TL，i，k Let C represent the TL user call cost for the i-th virtual power plant in the k-th scenario. ES，i，k Let S represent the energy storage operating cost of the i-th virtual power plant in the k-th scenario. DA,i，k G DA,i，k Let these represent the daily electricity sales and daily electricity purchases of the i-th virtual power plant in the k-th scenario, respectively.

6. The method for distributing market transaction revenues of multi-cluster virtual power plants based on the improved Shapley value method as described in claim 5, characterized in that: The calculation of the conventional Shapley value for each virtual power plant includes... The number of virtual power plant users is defined, meaning the total number of combinations within a multi-virtual power plant cluster is 2. N -1, where each member's reward is represented as, Among them, R i Let R(U) represent the revenue of the i-th virtual power plant, U represent the set of virtual power plants, R(U) represent the total revenue of the combination U, and R(U-{i}) represent the total revenue of the combination U excluding member vpp. i The subsequent benefits This represents the probability of a combination occurring.

7. The method for distributing market transaction revenues of multi-cluster virtual power plants based on the improved Shapley value method as described in claim 6, characterized in that: The optimization of profit distribution in cooperative alliances through multiple contribution factors includes, Calculate the energy and profit contributions of each virtual power plant, and then perform a secondary distribution of revenue using the improved Shapley value method. The energy contribution is expressed as, Where, θ i P represents the energy contribution. i-j This represents the power exchanged between the i-th virtual power plant and the j-th virtual power plant; The profit contribution rate is expressed as, Where, ω i R(U-{i}) represents the profit contribution, and R(U-{i}) represents the revenue of portfolio U after removing the i-th virtual power plant of member U. The revenue in the Shapley value method is redistributed secondaryly based on energy contribution and profit contribution, with the secondary distribution coefficient β. i Defined as: Where η1 and η2 are contribution weight coefficients, satisfying η1+η2=1; After considering the secondary allocation factor, the revenue of each VPP member is: R′ i =R i +R(N)β i Among them, R′ i R(N) represents the profit after secondary distribution, and R(N) represents the profit from full participation in the market. i This represents the revenue of the i-th virtual power plant calculated using the traditional Shapley value method.

8. A system for distributing market transaction revenues of multi-cluster virtual power plants based on an improved Shapley value method, wherein the system applies the method for distributing market transaction revenues of multi-cluster virtual power plants based on an improved Shapley value method as described in any one of claims 1 to 7, characterized in that... include: The module includes data acquisition and scenario generation, day-ahead bidding optimization, revenue calculation, traditional Shapley value calculation, and multi-contribution factor optimization allocation. The data acquisition and scenario generation module acquires historical data on market electricity prices and analyzes the probability distribution to construct an electricity price scenario set and build representative samples. The day-ahead bidding optimization module solves the day-ahead bidding model for multiple virtual power plants under multiple scenarios based on the electricity price scenario set; The revenue calculation module lists all combinations within the multi-virtual power plant cluster and calculates the revenue for each combination. The conventional Shapley value calculation module calculates the conventional Shapley value of each virtual power plant based on the revenue under each combination. The multi-contribution factor optimization allocation module optimizes the distribution of cooperative alliance benefits based on the traditional Shapley value.

9. 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 method for distributing market transaction revenues of multi-cluster virtual power plants based on the improved Shapley value method as described in any one of claims 1 to 7.

10. 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 method for distributing market transaction revenues of multi-cluster virtual power plants based on the improved Shapley value method as described in any one of claims 1 to 7.

Images