Electric energy and auxiliary service market capacity distribution decision-making method and system

Abstract

The invention relates to the technical field of power markets, and discloses an electric energy and auxiliary service market capacity distribution decision method and system, and the method comprises the following steps: obtaining the operation parameters and market prediction data of a power generation main body, and generating a group of future operation scenes reflecting the uncertainty according to the operation parameters and market prediction data; based on the scene set, establishing risk-adjusted benefits of each power generation main body in the electric energy and auxiliary service market for each power generation main body; through a multi-objective decision conversion method, the double-market income is constructed into a single-objective optimization function of each subject; and finally, constructing the single objective functions of all the subjects into a non-cooperative game model, and solving the Nash equilibrium of the non-cooperative game model to obtain an optimal capacity allocation scheme of each subject between the two markets. According to the method, a random scene reflecting new energy output correlation is constructed, and a conditional value-at-risk (CVaR) model is introduced to quantify a decision-making tail risk into cost, so that a decision-making scheme capable of keeping robust in a changeable market environment is obtained.

Description

Technical Field

[0001] This invention relates to the field of electricity market technology, specifically to a method and system for decision-making on the allocation of electricity and ancillary services market capacity. Background Technology

[0002] In modern power systems, the structure of the electricity market is becoming increasingly complex. In particular, with the large-scale integration of intermittent renewable energy sources such as wind and solar power, the demand for ancillary services for the stable operation of the power system is growing. This creates new opportunities for power generators, enabling them not only to sell electricity in the traditional energy market but also to generate additional revenue by providing ancillary services. Therefore, power generators urgently need a scientific decision-making method to rationally allocate their limited generation capacity between the energy market and the ancillary services market, thereby maximizing their overall economic benefits while meeting system security constraints.

[0003] Several methods already exist to assist power generation entities in decision-making. Some methods employ deterministic optimization models, using mathematical tools such as linear or nonlinear programming, to calculate the theoretically maximum revenue point for the power generation entity based on a given set of market price and load forecast data. These methods are fast and provide intuitive results. Other methods, when dealing with multi-market revenue problems, employ weighted summation techniques. This technique allows decision-makers to allocate weights to different revenue objectives based on their operational strategies and market judgments, thus ensuring that the optimization direction reflects the decision-maker's strategic intent and offers better controllability.

[0004] However, existing technologies still reveal some inherent technical flaws when applied to the current complex electricity market environment.

[0005] First, optimization methods that rely on a single deterministic forecast depend entirely on the accuracy of the forecast for their decision-making results. In real-world scenarios where renewable energy output fluctuates significantly, actual market prices and system loads often deviate from forecasts. This approach fails to effectively assess and manage the financial risks arising from such deviations, and is particularly unable to prevent low-probability, high-loss extreme market events, resulting in a lack of robustness in its decision-making schemes.

[0006] Secondly, the practice of manually setting weights to balance the returns of different markets introduces a strong degree of subjectivity. The selection of weights often relies on the decision-maker's experience, lacking objective data support, and making it difficult to guarantee that the selected weights truly reflect the intrinsic value correlation between different markets within a specific time period. When the market environment changes dynamically, fixed weights cannot adaptively adjust, leading to decision outcomes that deviate from the actual optimal point.

[0007] Furthermore, most existing optimization methods analyze individual power generators as isolated entities during modeling, neglecting the fact that the electricity market is a competitive environment comprised of multiple rational participants. In such methods, an entity's optimal decision is derived under the assumption that other competitors' strategies are fixed. However, in a real market, any decision by one party will trigger strategic responses from others. Decision-making methods that ignore market game dynamics lack strategic stability and are unlikely to achieve the desired results in real competitive interactions. Therefore, those skilled in the art propose a decision-making method and system for the allocation of electricity and ancillary services market capacity to address these problems. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention provides a decision-making method and system for the allocation of market capacity for electrical energy and ancillary services, which solves the problems of insufficient quantification of market uncertainty risks, subjective setting of multi-objective decision weights, and neglect of game behavior among market participants in existing technologies.

[0009] To achieve the above objectives, the present invention provides the following technical solution: The first aspect of this invention provides a method for decision-making on the allocation of market capacity for electrical energy and ancillary services, the method comprising: The system acquires the operating parameters of a preset number of power generation entities in the market, system load forecast data within a decision cycle, and new energy output forecast data. Based on the aforementioned data, it generates a set of future operating scenarios that can reflect the uncertainty of market prices and load. Based on this set of future operating scenarios, risk-adjusted returns in the electricity market and ancillary services market are established for each power generation entity in the market by taking into account risk costs. The risk-adjusted returns of the aforementioned electricity market and the risk-adjusted returns of the ancillary services market are combined using a multi-objective decision transformation method to construct a single-objective optimization function for the power generation entity. The single-objective optimization function of all power generation entities in the market is constructed into a non-cooperative game model, and the model is solved to obtain its Nash equilibrium. The Nash equilibrium is the optimal capacity allocation scheme for each power generation entity between the electricity market and the ancillary services market.

[0010] Preferably, the step of generating a set of future operating scenarios reflecting market uncertainty specifically includes: The system net load curve is calculated based on the system load forecast data and the new energy output forecast data. A multidimensional random scenario is generated around the net load curve of the system using the Latin hypercube sampling method. To ensure the statistical properties of this multidimensional random scenario, Cholesky decomposition is used to process the correlations between the new energy output data.

[0011] Preferably, the step of establishing the risk-adjusted return for each power generation entity specifically includes: Based on market forecast prices and the capacity allocation of the power generation entity, calculate the basic revenue of the power generation entity in the electricity market and the ancillary services market, and calculate its power generation cost based on its inherent cost coefficient. Based on the future operating scenarios of this group, the Conditional Value at Risk (CVaR) method is used to quantify and determine the risk cost of capacity allocation decisions for the power generation entity by calculating the average loss of decisions under adverse scenarios. The risk-adjusted return is obtained by subtracting the power generation cost and the risk cost from the basic return.

[0012] Preferably, the multi-objective decision transformation method employs principal component analysis, and this step specifically includes: The risk-adjusted returns of the electric energy market and the risk-adjusted returns of the ancillary services market are used as two target variables to construct a sample matrix and then standardize the matrix. Calculate the correlation coefficient matrix of the standardized sample matrix, and solve for the eigenvalues ​​and eigenvectors of the correlation coefficient matrix; The eigenvector corresponding to the largest eigenvalue is selected as the weight coefficient, and the two risk-adjusted returns mentioned above are weighted and summed to obtain the single-objective optimization function.

[0013] Preferably, the solution to the Nash equilibrium of the model is obtained using a genetic algorithm, and this step specifically includes: Encode the capacity allocation scheme of each power generation entity within a decision cycle as an individual, and initialize a strategy population consisting of multiple individuals for each power generation entity. Based on the single-objective optimization function, calculate the fitness of each individual in each generation of the strategy population under the strategy combination with other subjects; For each power generation entity's strategy population, perform selection, crossover, and mutation operations to generate a new generation of strategy populations, and repeat the fitness calculation and population generation process until the strategies of all power generation entities converge to the Nash equilibrium state.

[0014] Preferably, the step of determining the risk cost specifically includes: The total revenue function of the power generation entity is negatively defined as the decision loss function of the power generation entity. At a preset confidence level, a Value at Risk (VaR) threshold is determined using this decision loss function; Calculate the average loss that exceeds the VaR threshold across all future operating scenarios, and use this average as the risk cost.

[0015] Preferably, the steps for determining the risk-adjusted return specifically include: The risk-adjusted return of the power generation entity in the electricity market is obtained by subtracting its power generation cost and risk cost in the electricity market from its basic revenue in the electricity market. The risk-adjusted revenue of the ancillary services market is obtained by subtracting the power generation cost and risk cost of the power generation entity in the ancillary services market from its basic revenue in the ancillary services market.

[0016] Preferably, the single-objective optimization function Determined by the following formula: ; in: The serial number of the power generation entity; and These are the weighting coefficients determined by principal component analysis. and They are the main power generation entities Risk-adjusted returns in the electricity market and ancillary services market.

[0017] Preferably, the optimal capacity allocation scheme also needs to meet the operational constraints of the power generation entity, which include: at any time within the decision-making cycle, the sum of the capacity allocated to the electricity market and the capacity allocated to the ancillary services market is between the maximum and minimum technical output of the power generation entity.

[0018] A second aspect of the present invention provides a decision-making system for the allocation of market capacity for electricity and ancillary services, the system comprising: The scenario generation module is used to acquire the operating parameters of power generation entities in the market, system load forecast data, and new energy output forecast data, and generate a set of future operating scenarios that reflect market uncertainties based on the data. The revenue modeling module is used to establish risk-adjusted revenues for each power generation entity in the market in the electricity market and ancillary services market, based on the future operating scenario. The objective conversion module is used to construct a single-objective optimization function for the power generation entity by converting the risk-adjusted returns of the electricity market and the ancillary services market into a multi-objective decision conversion method. The equilibrium solution module is used to construct the single-objective optimization function of all power generation entities in the market into a non-cooperative game model, and solve the Nash equilibrium of the model to obtain the optimal capacity allocation scheme for each power generation entity between the power market and the ancillary services market.

[0019] This invention provides a method and system for decision-making regarding the allocation of market capacity for electrical energy and ancillary services. It offers the following advantages: 1. This invention constructs a stochastic scenario reflecting the correlation of new energy output and introduces a conditional value at risk (CVaR) model to quantify the tail risk of decision-making into cost, resulting in a robust decision-making scheme that can remain robust in a volatile market environment. Existing technologies mostly use deterministic prediction or simple sensitivity analysis. The decision-making schemes obtained from these methods are often vulnerable when the market experiences unexpected and drastic fluctuations because they do not fully assess extreme losses, resulting in a huge deviation between actual and expected returns. This invention solves the problem of huge deviation between actual and expected returns and improves the risk resistance and execution robustness of the decision-making scheme in an uncertain market environment.

[0020] 2. This invention employs Principal Component Analysis (PCA) to objectively determine the weight coefficients of the two optimization objectives, the electricity market and the ancillary services market, based on the inherent characteristics of the data. This avoids the biases and limitations caused by relying on expert experience or subjectively setting weights in existing technologies. The fixed weights in existing technologies cannot adapt to changes in the internal structure of the market, leading to rigid decision-making. This invention automatically seeks the optimal balance point between the two markets through a data-driven approach, making capacity allocation decisions more adaptive and objective.

[0021] 3. This invention abstracts the multi-entity competitive relationship in the market into a non-cooperative game model and applies a genetic algorithm to solve its Nash equilibrium, thereby obtaining an optimal allocation scheme with strategic stability. Compared with the existing technology's isolated optimization scheme that only targets a single entity, this invention fully considers the rational reactions of other participants in the market. The isolated optimization scheme of the existing technology ignores the mutual influence between market entities, and its optimal solution can be easily broken in a real competitive environment, lacking stability. The Nash equilibrium solution obtained by this invention is the optimal response strategy formulated for itself under the premise of predicting that all opponents will adopt the optimal strategy, which has higher strategic foresight and practical feasibility. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the system structure of the present invention.

[0023] Among them, 100 is the scene generation module; 200 is the revenue modeling module; 300 is the target transformation module; and 400 is the equilibrium solution module. Detailed Implementation

[0024] The technical solutions in 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, and 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.

[0025] See attached document Figure 2 , Figure 2 This is a schematic diagram of a power and ancillary services market capacity allocation decision system according to an embodiment of the present invention. The present invention provides a power and ancillary services market capacity allocation decision system, comprising: a scenario generation module 100, a revenue modeling module 200, an objective conversion module 300, and an equilibrium solution module 400.

[0026] The scenario generation module 100 is configured to acquire operating parameters of power generation entities in the market, system load forecast data, and renewable energy output forecast data. The scenario generation module 100 is further configured to generate a set of future operating scenarios reflecting market uncertainties based on the data. The scenario generation module 100 transmits the generated future operating scenarios to the revenue modeling module 200.

[0027] The revenue modeling module 200 is configured to receive future operating scenarios from the scenario generation module 100. The revenue modeling module 200 is further configured to establish risk-adjusted revenues for each power generation entity in the market, separately in the electricity market and the ancillary services market, based on the received future operating scenarios. The revenue modeling module 200 transmits the established risk-adjusted revenue data to the target conversion module 300.

[0028] The objective conversion module 300 is configured to receive risk-adjusted revenue data from the electricity market and ancillary service market from the revenue modeling module 200. The objective conversion module 300 is further configured to construct a single-objective optimization function for the power generation entity from the received risk-adjusted revenue data using a multi-objective decision conversion method. The objective conversion module 300 then transmits the constructed single-objective optimization function to the equilibrium solution module 400.

[0029] The equilibrium solution module 400 is configured to receive single-objective optimization functions from the objective transformation module 300. The equilibrium solution module 400 is further configured to construct a non-cooperative game model from the single-objective optimization functions of all power generation entities in the market, and solve this model to obtain its Nash equilibrium. This Nash equilibrium represents the optimal capacity allocation scheme for each power generation entity between the electricity market and the ancillary services market. The equilibrium solution module 400 outputs the obtained optimal capacity allocation scheme as the system output.

[0030] This system, through logical connections and data transfer between modules, realizes a complete decision-making process, from raw data processing, market scenario construction, individual profit and risk quantification, multi-objective optimization transformation, to the final game equilibrium solution. The uncertain scenarios generated by the scenario generation module 100 serve as input for the profit modeling module 200 in its risk assessment. The risk-adjusted profit calculated by the profit modeling module 200 forms the basis for multi-objective optimization by the objective transformation module 300. The single-objective optimization function generated by the objective transformation module 300 serves as the payoff function for the equilibrium solution module 400 in its game-theoretic solution. During the iterative solution process, the equilibrium solution module 400 periodically calls the profit modeling module 200 and the objective transformation module 300 to evaluate the fitness of the current strategy.

[0031] See attached document Figure 1 , Figure 1 This is a schematic flowchart of a method for allocating market capacity for electricity and ancillary services according to an embodiment of the present invention. The present invention provides a method for allocating market capacity for electricity and ancillary services, which may include the following steps: S100 acquires the operating parameters of power generation entities in the market, system load forecast data, and new energy output forecast data, and generates a set of future operating scenarios that reflect market uncertainties based on the data.

[0032] S200, based on future operating scenarios, establishes risk-adjusted returns for each power generation entity in the market in both the electricity market and the ancillary services market.

[0033] S300 constructs a single-objective optimization function for the power generation entity by converting the risk-adjusted returns of the electricity market and the ancillary services market through a multi-objective decision-making transformation method.

[0034] S400 constructs a non-cooperative game model by taking the single-objective optimization function of all power generation entities in the market and solving the Nash equilibrium of the model to obtain the optimal capacity allocation scheme for each power generation entity between the power market and the ancillary services market.

[0035] The following is a detailed description of the method and steps: S100: Acquire operating parameters of power generation entities in the market, system load forecast data, and renewable energy output forecast data, and generate a set of future operating scenarios reflecting market uncertainties based on the data. This step is used to provide basic data and environmental modeling for subsequent risk assessment and decision optimization.

[0036] First, obtain the necessary data, which includes: system load forecast data for a specific decision-making period. Photovoltaic power generation output forecast data Wind power output forecast data .in, It can indicate time, for example, in hours, within a 24-hour decision-making cycle. The range is from 1 to 24. Simultaneously, it acquires the operating parameters of all power generation entities in the market, including but not limited to the power generation cost coefficient of thermal power units, the levelized cost of electricity (LCOE) of new energy units, and the maximum and minimum technical output constraints of each unit.

[0037] Secondly, based on the acquired system load forecast data and renewable energy output forecast data, the net load of the system at each time point is calculated. Net load refers to the system load demand that still needs to be met by conventional controllable power sources (such as thermal power) after deducting renewable energy generation output. Net load The calculation formula is as follows: ; in: For a moment The system net load; For a moment The predicted total system load; For a moment The predicted output value of photovoltaic power generation; For a moment The predicted output value of wind power generation.

[0038] Finally, to fully simulate the uncertainty of the market environment, a set of scenarios reflecting possible future operating states is generated based on the net load curve. Specifically, the Latin Hypercube Sampling (LHS) method is used to perform multidimensional random sampling of the uncertainty in the net load data to generate multiple future operating scenarios that can cover the entire uncertainty space and have a uniform sample distribution. Furthermore, considering the potential statistical correlation between the output of different types of renewable energy (e.g., wind power and photovoltaic), this invention employs the Cholesky decomposition method to process the correlation of the sampled renewable energy output sequences, ensuring that the generated future operating scenarios not only have randomness but also maintain the true statistical dependence between renewable energy outputs. Through this method, a set of scenarios containing… A collection of independent and representative future operating scenarios ,in Indicates the first A scenario.

[0039] S200, based on the future operating scenario generated in step S100, establishes the risk-adjusted returns for each power generation entity in the market in both the electricity market and the ancillary services market. This step aims to construct a comprehensive individual decision-making objective that simultaneously reflects expected returns and potential risks.

[0040] This step first involves the power generation main body. Establish basic revenue and generation cost models for the electricity market and ancillary services market, respectively. Basic revenue in the electricity market... Compared to basic revenue in the ancillary services market It can be determined by the following formula: ; ; in: The serial number of the power generation entity; The time sequence number within the decision-making cycle; The total duration of the decision-making cycle; and They are time points Predicted clearing prices for the electricity market and ancillary services market; and These are the main power generation units At any moment The power generation capacity allocated to the two markets.

[0041] The power generation cost of a power generation entity is determined based on its type. For thermal power units, the power generation cost... Determined by a quadratic function: ; in: For the unit At any moment The total output is and sum; , , This is the inherent cost coefficient for the thermal power unit. For new energy units, the power generation cost is calculated using the levelized cost of electricity (LCOE).

[0042] Secondly, this step introduces a risk quantification model to assess decision-making risk. Due to the uncertainty of market prices and renewable energy output, any capacity allocation decision faces the risk of revenue fluctuations. This invention employs the Conditional Value at Risk (CVaR) method to quantify this risk. First, a loss function for the decision is defined. This refers to the future operational scenario of this decision. The negative value of the total return. Then, at the preset confidence level. Under these conditions, the Value at Risk (VaR) is determined. This value represents the risk level at which a company's capital gains or losses incurs. The maximum possible loss for a decision under certain probabilities. CVaR further calculates the average of all losses exceeding the VaR threshold; this value is determined as the risk cost of the decision. Conditional Value at Risk The calculation formula is as follows: ; in: Main generator Capacity allocation decision vector throughout the entire decision-making cycle; A future running scenario generated for step S100; For the scene The probability density of occurrence; To be at confidence level Value at risk.

[0043] Finally, the basic revenue, generation cost, and risk cost are integrated to obtain the risk-adjusted revenue. (Power generation entity) Risk-adjusted returns in the electricity market and ancillary services market and Determined by the following formula: ; ; in: and It is the electricity generation cost allocated between the two markets based on the total electricity generation cost; and This is the corresponding capacity allocation decision vector; and This is for two markets at different confidence levels. and The risk cost is calculated below.

[0044] In step S300, the risk-adjusted returns from the electricity market and the ancillary services market obtained in step S200 are used to construct a single-objective optimization function for the power generation entity through a multi-objective decision transformation method. This step aims to transform a bi-objective optimization problem into a single-objective optimization problem with a clear optimization direction, laying the foundation for subsequent solutions.

[0045] This invention employs Principal Component Analysis (PCA) as a multi-objective decision transformation method. The execution process of this method is as follows: First, a sample matrix is ​​constructed. This matrix consists of the risk-adjusted returns from the electricity market corresponding to multiple decision options. Risk-adjusted returns compared to the ancillary services market The sample matrix was then standardized using Z-scores to eliminate the influence of differences in dimensions and orders of magnitude between the two target variables.

[0046] Secondly, based on the standardized data, the correlation coefficient matrix between the two target variables is calculated. This matrix reflects the degree of linear correlation between the two risk-adjusted returns.

[0047] Next, the correlation coefficient matrix Perform eigenvalue decomposition and solve for its eigenvalues. With the corresponding feature vector Each eigenvector defines a principal component, and the magnitude of its corresponding eigenvalue indicates the degree of contribution of that principal component to the total variance.

[0048] Then, select the principal component with the largest contribution rate, i.e., the largest eigenvalue. The corresponding eigenvector The elements of this eigenvector objectively reflect the relative importance of the two target variables in determining the main direction of change. These elements are used as weighting coefficients, denoted as... and .

[0049] Finally, the two risk-adjusted returns are summed to construct the power generation entity. The final single-objective optimization function This function is the fitness function used in subsequent game-solving steps. Its specific form is as follows: ; in: The serial number of the power generation entity; and These are the weighting coefficients determined by principal component analysis. and They are the main power generation entities Risk-adjusted returns in the electricity market and ancillary services market.

[0050] In step S400, the single-objective optimization functions obtained in step S300 for all power generation entities in the market are constructed into a non-cooperative game model, and the Nash equilibrium of this model is solved to obtain the optimal capacity allocation scheme for each power generation entity between the electricity market and the ancillary services market. This step aims to find a stable optimal strategy combination by simulating the competitive behavior of the parties in the market.

[0051] In this non-cooperative game model, the market components are defined as follows: all N power generation entities in the market are participants in the game; each participant... strategy The capacity allocation scheme for it throughout the entire decision-making cycle, i.e., the decision vector. Each participant The payment function is the single-objective optimization function constructed in step S300. The value of this payment function depends not only on the participants. Its own strategy It also depends on the set of strategies of all other participants in the market. .

[0052] The goal of this step is to find a Nash equilibrium point in the game model. A strategy combination Defined as a Nash equilibrium, if for any participant Maintaining a strategy with all its opponents Under the premise that the strategy remains unchanged, the participant cannot unilaterally change its own strategy. To obtain a higher payment. This condition is determined by the following formula: ; in: The serial number of the power generation entity; For participants Strategies under Nash equilibrium; To exclude participants The set of strategies of all other participants in the Nash equilibrium state; For participants Any other strategy may be chosen; For participants The payoff function is a single-objective optimization function.

[0053] Because the solution space of this game model is high-dimensional and non-convex, this invention employs a genetic algorithm (GA) to iteratively solve for the Nash equilibrium. The solution process specifically includes: First, the strategy of each participant is encoded. A complete capacity allocation scheme (i.e., a vector containing capacity allocation values ​​at all times within the decision-making cycle) is encoded as a chromosome individual.

[0054] Secondly, an initial strategy population consisting of multiple chromosome individuals is randomly generated for each participant in the market.

[0055] Then, iterative evolution is performed. In each generation, a policy profile is derived from the population of each participant, and then applied according to the aforementioned payoff function. The fitness of each individual is calculated. Subsequently, selection, crossover, and mutation operations are performed independently on the population of each participant to generate a new generation of strategic populations with higher fitness.

[0056] Finally, convergence is determined. The iterative evolution process described above is repeated until all participants' strategy populations reach the preset convergence criteria, such as the strategies no longer changing significantly or reaching the maximum number of iterations. At this point, the strategy combination formed by the optimal individuals in each participant population is a Nash equilibrium solution for the game model.

[0057] The final output of this step is the Nash equilibrium solution. Specifically, for each power generation entity, the output is an optimal capacity allocation scheme. This plan details the generation capacity to be allocated to the electricity market and ancillary services market at each point in time within the future decision-making cycle. This plan represents the most stable and optimal decision achievable after comprehensively considering market uncertainty, the company's own risk appetite, and the rational decision-making behavior of all competitors.

[0058] To further illustrate the collaborative working process of the technical solution of this invention, a specific working scenario example will be used below.

[0059] This embodiment aims to determine the optimal capacity allocation scheme for a specific power generation entity, "thermal power unit 1," within a future 24-hour decision-making cycle in a regional power market that includes multiple power generation entities.

[0060] The market setup includes three thermal power units (thermal power unit 1, thermal power unit 2, and thermal power unit 3), one photovoltaic power station, and one wind farm. The input data for the decision-making system in this embodiment includes: the system's total load forecast curve for the next 24 hours, the output forecast curves for the photovoltaic power station and wind farm, the market price forecast curves for electricity and ancillary services, and the technical and economic parameters for all five power generation entities. Specifically, the technical and economic parameters include the cost coefficients a, b, and c of the thermal power units, their maximum and minimum technical output, and the levelized cost of electricity (LCOE) of the renewable energy power station.

[0061] First, the system executes step S100. Based on the input total load forecast data and the output forecast data of photovoltaic and wind power, the system calculates the system net load curve for the next 24 hours. Subsequently, the system uses Latin hypercube sampling and Cholesky decomposition methods to generate 1000 random and statistically correlated future operating scenarios around the net load curve. This set of scenarios constitutes the uncertainty environment basis for subsequent decision analysis.

[0062] Next, the system executes step S200. For all five power generation entities in the market, the system establishes a revenue model for each. Taking "Thermal Power Unit 1" as an example, given a capacity allocation strategy, the system calculates the expected revenue and cost of the strategy in each of the 1000 scenarios generated in the previous step. Simultaneously, setting a confidence level of, for example, 95%, the system uses the CVaR method to calculate the risk cost corresponding to the strategy. Finally, the system obtains the risk-adjusted revenue of "Thermal Power Unit 1" under this strategy in the electricity market and ancillary services market. This process is performed simultaneously for all five entities.

[0063] Next, the system executes step S300. Taking "Thermal Power Unit 1" as an example again, the system uses its two risk-adjusted revenue objectives in the electricity market and ancillary services market as variables, and applies Principal Component Analysis (PCA) to process them. Through calculation, the system obtains a set of objective weighting coefficients, for example, the weight of the electricity market is 0.61, and the weight of the ancillary services market is 0.39. Using these weights, the system weights and sums the two revenue objectives, thus constructing a unified single-objective optimization function for "Thermal Power Unit 1". This process is also applied to the other four entities, generating their respective single-objective optimization functions.

[0064] Finally, the system executes step S400. The system constructs a non-cooperative game model by combining the five agents and their respective single-objective optimization functions. The operating parameters of the genetic algorithm are set, such as a population size of 100 and a maximum number of iterations of 200. After the algorithm starts, it simulates the five agents engaging in multiple rounds of strategy adjustments and competition in the market. In each iteration, the merit of an agent's strategy depends on the value of its single-objective optimization function under the current strategies of the other four agents. After multiple generations of evolution, when the strategy choices of all agents tend to stabilize and no longer change significantly, the system determines that it has converged to a Nash equilibrium.

[0065] The final output of this embodiment is a detailed optimal capacity allocation scheme for "thermal power unit 1". This scheme is presented in the form of a data table or graph, precisely listing the capacity values ​​to be allocated to the electricity market and the ancillary services market for each hour within the next 24 hours. This scheme is a stable and optimal strategy obtained after taking into account market uncertainty, its own risk appetite, and fully anticipating the rational behavior of all competitors.

[0066] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for decision-making on market capacity allocation for electrical energy and ancillary services, characterized in that, Includes the following steps: The system acquires operating parameters of power generation entities in the market, system load forecast data, and new energy output forecast data, and generates a set of future operating scenarios that reflect market uncertainties based on the data. Based on the aforementioned future operating scenario, risk-adjusted returns are established for each power generation entity in the market in the electricity market and the ancillary services market. The risk-adjusted returns of the electricity market and ancillary services market are used to construct a single-objective optimization function for the power generation entity through a multi-objective decision transformation method. The single-objective optimization function of all power generation entities in the market is constructed into a non-cooperative game model, and the Nash equilibrium of the model is solved to obtain the optimal capacity allocation scheme of each power generation entity between the power market and the ancillary services market.

2. The method for allocating market capacity for electrical energy and ancillary services according to claim 1, characterized in that, The generation of a set of future operating scenarios reflecting market uncertainty includes: The system net load curve is calculated based on the system load forecast data and the new energy output forecast data. A multidimensional random scenario is generated around the net load curve of the system using the Latin hypercube sampling method. The correlation between the new energy output data is processed using Cholesky decomposition to ensure the statistical characteristics of the multidimensional random scenario.

3. The method for allocating market capacity for electrical energy and ancillary services according to claim 1, characterized in that, The aforementioned risk-adjusted returns for each power generation entity in the electricity market and ancillary services market include: Calculate the basic revenue and generation cost of the power generation entity in the electricity market and ancillary services market, respectively; Based on the future operating scenario, the Conditional Value at Risk (CVaR) method is used to quantify and determine the risk cost of the power generation entity making capacity allocation decisions. The risk-adjusted return is obtained by subtracting the power generation cost and risk cost from the basic return.

4. The method for allocating market capacity for electrical energy and ancillary services according to claim 3, characterized in that, The steps for determining the risk cost include: The negative value of the revenue function of the power generation entity is defined as the decision loss function; At a given confidence level, the Value at Risk (VaR) threshold is determined using the decision loss function; Calculate the average loss that exceeds the VaR threshold in all scenarios to obtain the risk cost, i.e., the conditional value of risk (CVaR).

5. The method for allocating market capacity for electrical energy and ancillary services according to claim 3, characterized in that, The steps for determining the risk-adjusted return include: The power generation entity's basic revenue in the electricity market is subtracted from its power generation cost and risk cost in the electricity market, respectively. The power generation entity's basic revenue in the ancillary services market is reduced by its power generation costs and risk costs in the ancillary services market.

6. The method for determining the market capacity allocation of electrical energy and ancillary services according to claim 1, characterized in that, The multi-objective decision transformation method employs principal component analysis, which includes: The risk-adjusted returns of the electric energy market and the ancillary services market are used as two objective variables to construct a sample matrix and then standardize it. Calculate the correlation coefficient matrix of the standardized sample matrix, and solve for its eigenvalues ​​and eigenvectors; The eigenvector corresponding to the largest eigenvalue is selected as the weight coefficient, and the two risk-adjusted returns are weighted and summed to obtain the single-objective optimization function.

7. The method for determining the market capacity allocation of electrical energy and ancillary services according to claim 6, characterized in that, The single-objective optimization function Determined by the following formula: ； in: The serial number of the power generation entity; and These are the weighting coefficients determined by principal component analysis. and They are the main power generation entities Risk-adjusted returns in the electricity market and ancillary services market.

8. The method for determining the market capacity allocation of electrical energy and ancillary services according to claim 1, characterized in that, The Nash equilibrium of the non-cooperative game model is solved using a genetic algorithm, including: The capacity allocation scheme of each power generation entity is encoded as an individual, and a strategy population is initialized for each power generation entity. The fitness of each individual in each generation of the strategy population is calculated based on the single-objective optimization function. For each power generation entity, perform selection, crossover, and mutation operations on its strategy population to generate a new generation of strategy populations until the strategies of all power generation entities converge to a Nash equilibrium.

9. The method for allocating market capacity for electrical energy and ancillary services according to claim 1, characterized in that, The optimal capacity allocation scheme must also meet the operational constraints of the power generation entity, which include: the sum of the capacity allocated to the electricity market and the capacity allocated to the ancillary services market at any given time must be between the maximum and minimum technical output of the power generation entity.

10. A decision-making system for the allocation of market capacity for electricity and ancillary services, applied to the decision-making method for the allocation of market capacity for electricity and ancillary services as described in any one of claims 1-9, characterized in that, include: The scenario generation module is used to acquire the operating parameters of power generation entities in the market, system load forecast data, and new energy output forecast data, and generate a set of future operating scenarios that reflect market uncertainties based on the data. The revenue modeling module is used to establish risk-adjusted revenues for each power generation entity in the market in the electricity market and ancillary services market, based on the future operating scenario. The objective conversion module is used to construct a single-objective optimization function for the power generation entity by converting the risk-adjusted returns of the electricity market and the ancillary services market into a multi-objective decision conversion method. The equilibrium solution module is used to construct the single-objective optimization function of all power generation entities in the market into a non-cooperative game model, and solve the Nash equilibrium of the model to obtain the optimal capacity allocation scheme for each power generation entity between the power market and the ancillary services market.

Images