A wind-solar-storage-collaborative power financial decision method

Abstract

The application discloses a kind of wind light storage collaborative power financial decision-making method, belong to wind light storage collaborative power financial decision-making method technical field, solve the problem that due to conventional enterprise has not carried out the parameter needed by wind light storage collaborative decision-making collection, when accessing large power grid carries out electricity price operation, it can not appear to evaluate or the problem that the evaluation accuracy exists influence. Including the following steps: obtain regional meteorological data, regional new energy output data and power market data by external data access module;Through enterprise local data acquisition module, real-time acquisition enterprise total load power, energy storage operating state and equipment parameters;In migration calculation engine, execute output and load decomposition calculation.The application fuses regional public data and enterprise parameters, generates wind light output estimate value and decomposes load characteristics using migration learning, constructs power market bidding strategy combined with multi-stage robust optimization, dynamically configures financial derivatives to hedge risk, solves the problem that electricity price evaluation is inaccurate due to missing monitoring data.

Description

Technical Field

[0001] This invention relates to the technical field of wind-solar-storage collaborative power financial decision-making methods, and in particular to a wind-solar-storage collaborative power financial decision-making method. Background Technology

[0002] Enterprise microgrids participate in the electricity spot market, where aggregators integrate distributed energy sources such as photovoltaics, wind power, and energy storage, along with controllable loads, to buy and sell electricity in a market-based manner. Accurate prediction of output and electricity prices is crucial for trading. A CNN-LSTM hybrid model is commonly used to capture the spatiotemporal characteristics of new energy sources, combined with Monte Carlo simulations to generate multi-scenario electricity price probability distributions. In the decision-making phase, a two-stage robust optimization and the MOPSO algorithm are employed to balance returns and risks while satisfying constraints such as energy storage and the power grid, formulating the optimal bidding strategy. Risk control quantifies losses using a CVaR model, hedging fluctuations with contract tools, and automatic operation via blockchain smart contracts when thresholds are triggered. This algorithmic approach to trading and risk control facilitates the efficient entry of distributed resources into the market.

[0003] Currently, the power finance decision-making engine for wind, solar and energy storage synergy relies on key parameters such as wind and solar power output, energy storage status, and load fluctuations to achieve accurate decision-making.

[0004] However, conventional enterprises do not collect these parameters in a targeted manner, resulting in a lack of basic data input for the engine. When connecting to the large power grid for electricity pricing operations, due to the lack of key parameters or insufficient accuracy, the decision engine cannot fully assess the output potential, cost-benefit, and market risks of wind, solar, and energy storage synergy. This affects the accuracy of assessments in electricity price forecasting, transaction declaration, and other aspects, which may lead to decision-making biases and reduce the effectiveness and reliability of power financial operations.

[0005] Therefore, a wind-solar-storage integrated power finance decision-making method is proposed to solve or alleviate the above problems. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of existing technologies by proposing a wind-solar-storage collaborative power financial decision-making method.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] A wind-solar-storage integrated power financial decision-making method includes the following steps:

[0009] Regional meteorological data, regional renewable energy output data, and electricity market data are obtained through an external data access module.

[0010] The enterprise's total load power, energy storage operation status and equipment parameters are collected in real time through the enterprise's local data acquisition module.

[0011] The power and load decomposition calculations are performed in the migration computing engine, and it outputs the photovoltaic power output estimate, wind power output estimate and adjustable load to the scene generator.

[0012] The wind-solar-load joint probability distribution is constructed in the scene generator and a reduced scene set is generated. The reduced scene set containing the wind-solar-load joint state and real-time electricity price disturbance is output to the robust optimization solver.

[0013] After solving the two-stage optimization problem in the robust optimization solver, the baseline scheduling strategy is output to the rolling correction module.

[0014] The rolling correction module dynamically corrects decision commands based on real-time monitoring data.

[0015] The risk measurement module calculates the dynamic conditional value of risk and adjusts the bidding strategy. The risk measurement module outputs the dynamic conditional value of risk to the hedging strategy generator.

[0016] Configure financial derivatives to hedge residual risk in the hedging strategy generator.

[0017] Preferably, the step of performing output and load decomposition calculations in the migration calculation engine specifically includes the following steps:

[0018] To perform photovoltaic power output migration calculations, the average regional photovoltaic power output is divided by the total installed photovoltaic capacity in the region to obtain the power output ratio per unit capacity. This ratio is then multiplied by the company's installed photovoltaic capacity, and then by the tilt angle correction factor, system efficiency factor, and temperature degradation compensation factor. The tilt angle correction factor is calculated based on the company's latitude, and the temperature degradation compensation factor is determined based on the difference between the real-time temperature and the standard test temperature.

[0019] To perform wind power output migration calculations, the regional wind speed data is corrected according to the hub height and surface roughness to obtain the enterprise's hub height wind speed; the corrected wind speed is divided by the wind turbine's rated wind speed, cubed, multiplied by the enterprise's wind power installed capacity, and then multiplied by the air density correction factor.

[0020] Adjustable load separation is performed, and cluster analysis is conducted on historical total load data to identify load curves under different production modes. The average load rate of each load mode cluster is calculated by subtracting the product of the base load and the load rate of that cluster from the current total load.

[0021] Preferably, the step of constructing a joint probability distribution of wind, solar, and load in the scene generator and generating a reduced scene set specifically includes the following steps:

[0022] The joint probability distribution of wind, solar and load is constructed based on the Gaussian connection function. The marginal distributions of each variable are transformed using the inverse function of the standard normal distribution. Then, a multivariate normal distribution is constructed using the correlation coefficient matrix.

[0023] The initial scene set was generated using Latin hypercube sampling technique.

[0024] The K-medoids clustering algorithm based on Mahalanobis distance reduces the scene size, calculates the covariance weighted distance between each scene, and selects representative scene center points to form a reduced scene set.

[0025] For each scenario, a real-time electricity price disturbance following a Johnson-SU distribution is superimposed.

[0026] Preferably, solving the two-stage optimization problem in the robust optimization solver specifically includes the following steps:

[0027] The first phase of day-ahead decision optimization minimizes the sum of day-ahead market electricity purchase costs and ancillary service revenues, while considering real-time adjustment costs under the worst-case scenario; where the uncertainty set is defined as the set of all scenarios whose covariance weighted distance from the benchmark scenario is less than the chi-square distribution critical value;

[0028] The second-stage real-time adjustment model minimizes the sum of real-time market deviation costs and penalty costs. The constraints include that the energy storage power adjustment amount does not exceed the real-time available power margin, the margin is the difference between the maximum energy storage power and the planned power, and the minimum of the available capacity divided by the time step.

[0029] Preferably, the step of dynamically correcting the decision instruction based on real-time monitoring data in the rolling correction module specifically includes the following steps:

[0030] The prediction error feedback correction multiplies the difference between the actual output and the predicted value in the current period by a dynamically updated autoregressive coefficient and adds it to the predicted value for the next period.

[0031] The decision instruction is smoothly adjusted to solve an optimization problem that minimizes the deviation between the new instruction and the original baseline strategy, and minimizes the change in energy storage action. The penalty weight for energy storage action is set according to the equipment loss characteristics.

[0032] Preferably, the step of calculating the dynamic conditional value of risk and adjusting the bidding strategy in the risk measurement module specifically includes the following steps:

[0033] Dynamic conditional value at risk calculation, on a reduced scenario set, calculates the average of the last 5 percent of the loss distribution;

[0034] The risk buffer adjustment for bid volume is to deduct the positive deviation between the risk value and the expected bid volume from the original bid volume, where the risk value is taken as the decimal of the loss distribution.

[0035] Preferably, configuring financial derivatives to hedge residual risk in the hedging strategy generator specifically includes the following steps:

[0036] Residual risk exposure quantification involves calculating the probability-weighted sum of the loss values ​​for the 5% scenario with the greatest potential loss.

[0037] Solving for the optimal hedging amount in CFDs involves finding the contract purchase amount that minimizes the deviation between the risk exposure and the maximum contract return.

[0038] Automated execution of hedging transactions through blockchain smart contracts.

[0039] The present invention has the following beneficial effects:

[0040] This invention solves the problem of inaccurate electricity price operation assessment caused by the lack of monitoring data in conventional enterprises by integrating regional public data with basic enterprise operating parameters, using transfer learning technology to generate missing wind and solar power output estimates and decompose load characteristics, combining multi-stage robust optimization to construct a risk-aware electricity market bidding strategy, and dynamically configuring financial derivatives to hedge residual risks. Attached Figure Description

[0041] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This is a structural block diagram of the present invention.

[0043] The diagram shows: 1. External data access module; 2. Enterprise local data acquisition module; 3. Migration computing engine; 4. Scenario generator; 5. Robust optimization solver; 6. Rolling correction module; 7. Risk measurement module; 8. Hedging strategy generator. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0045] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0046] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0047] In the description of this invention, it should be understood that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of this invention is in use, or the orientation or positional relationship commonly understood by those skilled in the art. They are only used to facilitate the description of this invention and to simplify the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0048] Furthermore, the terms "first," "second," and "third" are used only to distinguish descriptions and should not be interpreted as indicating or implying relative importance.

[0049] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0050] A wind-solar-storage integrated power finance decision-making method, such as Figure 1 As shown, it includes the following steps:

[0051] Regional meteorological data, regional renewable energy output data, and electricity market data are acquired through external data access module 1.

[0052] This step obtains regional meteorological data, regional new energy output data, and electricity market data through external data access module 1, directly replacing the real-time monitoring parameters that the enterprise lacks.

[0053] Regional meteorological data, including irradiance and wind speed, fills the monitoring gap for enterprises that have not installed meteorological stations. Regional renewable energy output data provides a capacity ratio reference to estimate the theoretical wind and solar output of enterprises. Electricity market data constructs a benchmark for electricity price assessment. The three work together to reconstruct the assessment parameter system required for enterprises' wind, solar and energy storage collaborative decision-making, compressing the prediction errors caused by data gaps and ensuring the assessability of electricity price operation strategies from the source.

[0054] The enterprise's total load power, energy storage operation status and equipment parameters are collected in real time through the enterprise's local data acquisition module 2.

[0055] This step uses the enterprise's local data acquisition module 2 to acquire total load power, energy storage operating status, and equipment nameplate parameters in real time. This minimizes monitoring costs and addresses the problem of missing key data. The total load curve replaces the missing sub-metering data to support load decomposition. The total energy storage power and SOC status fill the gap in battery pack-level monitoring. Equipment parameters, including capacity and efficiency, are anchored to physical boundary constraints, making wind and solar power output migration calculation and optimization control feasible. This reduces the error in identifying the load adjustable potential and improves the accuracy of energy storage status estimation, providing a highly reliable local data foundation for subsequent collaborative decision-making.

[0056] The power output and load decomposition calculations are performed in the migration computing engine 3, and it outputs the photovoltaic power output estimate, wind power output estimate and adjustable load to the scene generator 4.

[0057] To perform photovoltaic power output migration calculations, the average regional photovoltaic power output is divided by the total installed photovoltaic capacity in the region to obtain the power output ratio per unit capacity. This ratio is then multiplied by the company's installed photovoltaic capacity, and then by the tilt angle correction factor, system efficiency factor, and temperature degradation compensation factor. The tilt angle correction factor is calculated based on the company's latitude, and the temperature degradation compensation factor is determined based on the difference between the real-time temperature and the standard test temperature.

[0058] To perform wind power output migration calculations, the regional wind speed data is corrected according to the hub height and surface roughness to obtain the enterprise's hub height wind speed; the corrected wind speed is divided by the wind turbine's rated wind speed, cubed, multiplied by the enterprise's wind power installed capacity, and then multiplied by the air density correction factor.

[0059] Adjustable load separation is performed, and cluster analysis is conducted on historical total load data to identify load curves under different production modes. The average load rate of each load mode cluster is calculated by subtracting the product of the base load and the load rate of that cluster from the current total load.

[0060] In this step, the average regional photovoltaic output is divided by the total installed capacity of the regional photovoltaic system to obtain the output ratio per unit capacity through photovoltaic output migration calculation. This ratio is then multiplied by the enterprise's photovoltaic installed capacity and a tilt angle correction coefficient and a temperature attenuation compensation coefficient are introduced. The tilt angle correction coefficient is a spatial irradiance optimization factor calculated based on latitude, while the temperature attenuation compensation coefficient is a power correction term based on the difference between real-time temperature and standard test temperature. This addresses the blind spot in output assessment caused by the enterprise's failure to install string-level monitoring.

[0061] By calculating wind power output migration based on a logarithmic wind speed profile model of surface roughness and meteorological tower height, regional wind speed data is corrected for hub height and converted to the cube ratio of rated wind speed, with air density correction added to compensate for the power coefficient affected by altitude, thus overcoming the monitoring limitations of enterprises lacking anemometers. By separating adjustable loads and performing K-means cluster analysis on historical total loads, typical production mode clusters are identified and the load rate of each cluster is calculated. The current total load is subtracted from the product of the base load and the load rate, with the base load being the moving average of the minimum value of the cluster, solving the problem of quantifying adjustable potential when individual metering is not deployed.

[0062] This reduces the estimation error of photovoltaic power output, lowers the estimation error of wind power output, improves the accuracy of adjustable load identification, and reconstructs the complete parameter system required for wind-solar-storage collaborative decision-making under the condition of zero new monitoring equipment, thus eliminating some of the risk of assessment bias.

[0063] In the scenario generator 4, a joint probability distribution of wind, solar and load is constructed and a reduced scenario set is generated. The reduced scenario set containing the joint state of wind, solar and load and real-time electricity price disturbance is output to the robust optimization solver 5.

[0064] The joint probability distribution of wind, solar and load is constructed based on the Gaussian connection function. The marginal distributions of each variable are transformed using the inverse function of the standard normal distribution. Then, a multivariate normal distribution is constructed using the correlation coefficient matrix.

[0065] The initial scene set was generated using Latin hypercube sampling technique.

[0066] The K-medoids clustering algorithm based on Mahalanobis distance reduces the scene size, calculates the covariance weighted distance between each scene, and selects representative scene center points to form a reduced scene set.

[0067] For each scenario, a real-time electricity price perturbation following a Johnson-SU distribution is superimposed;

[0068] This step constructs a Gaussian Copula joint probability distribution through the scenario generator 4, integrates the photovoltaic / wind power output estimates and adjustable load output from the migration calculation engine 3, and solves the assessment distortion problem caused by enterprises' failure to collect wind-solar-load correlation data. Based on the correlation coefficient matrix of regional historical data, a wind-solar-load joint distribution model is established. Latin hypercube sampling is used to generate an initial scenario set, and then the scenario scale is compressed to 50 representative scenarios through Mahalanobis distance-weighted K-medoids clustering algorithm. The Johnson-SU distribution of electricity price disturbance is superimposed to simulate real-time market fluctuations, which improves the insufficient scenario coverage caused by data loss and significantly enhances the predictive ability of extreme events, such as sudden power drop, thereby improving the risk perception accuracy of subsequent optimization decisions and overcoming the problem of inaccurate assessment from a probabilistic perspective.

[0069] After solving the two-stage optimization problem in the robust optimization solver 5, the baseline scheduling strategy is output to the rolling correction module 6.

[0070] The first phase of day-ahead decision optimization minimizes the sum of day-ahead market electricity purchase costs and ancillary service revenues, while considering real-time adjustment costs under the worst-case scenario; where the uncertainty set is defined as the set of all scenarios whose covariance weighted distance from the benchmark scenario is less than the chi-square distribution critical value;

[0071] The second-stage real-time adjustment model minimizes the sum of real-time market deviation costs and penalty costs. The constraints include that the energy storage power adjustment amount does not exceed the real-time available power margin, the margin is the difference between the maximum energy storage power and the planned power, and the minimum of the available capacity divided by the time step.

[0072] This step involves establishing a decision-making model under ellipsoidal uncertainty set constraints using a two-stage robust optimization solver 5 to address the problem of strategy inaccuracies caused by the lack of real-time parameters for wind, solar and load. The first stage aims to minimize the day-ahead market electricity purchase cost and maximize ancillary service revenue. Under uncertainty set constraints, energy storage plans and market bidding are formulated. The uncertainty set covers all wind, solar and load fluctuation scenarios where the weighted distance of the covariance of the benchmark scenario is less than 95% of the chi-square distribution critical value.

[0073] The second phase involves constructing a min-max-min game model for the real-time market to simulate worst-case scenarios, such as sudden drops in wind speed and electricity price crashes. This model dynamically optimizes the adjustment of energy storage charging and discharging and the response of ancillary services. The constraints strictly limit the adjustment of energy storage power to not exceed the real-time availability margin. The model takes the minimum of the difference between the maximum power and the planned value, as well as the minimum of the capacity converted from the state of charge boundary. This transforms the decision-making bias risk caused by missing data into quantifiable worst-case scenario costs, enabling enterprises to maintain a certain rate of revenue achievement even under zero wind and solar power output monitoring conditions, and reducing losses caused by extreme weather.

[0074] In the rolling correction module 6, decision instructions are dynamically corrected based on real-time monitoring data;

[0075] The prediction error feedback correction multiplies the difference between the actual output and the predicted value in the current period by a dynamically updated autoregressive coefficient and adds it to the predicted value for the next period.

[0076] The decision instruction is smoothly adjusted to solve an optimization problem that minimizes the deviation between the new instruction and the original baseline strategy and minimizes the change in energy storage action. The penalty weight for energy storage action is set according to the equipment loss characteristics.

[0077] This step establishes a real-time feedback mechanism through the rolling correction module 6. To address the prediction deviation problem caused by the enterprise's failure to collect detailed data on wind and solar power output, the ARMA(1,1) model is used to update the autoregressive coefficients online based on the dynamic error between the actual total power output and the migration prediction value in the current period. Specifically, the historical error sequence fitting weights are used to add the error compensation amount to the prediction value in the next period, thereby achieving closed-loop correction of prediction deviation.

[0078] Simultaneously, a quadratic programming model is constructed to find the optimal solution that minimizes the Euclidean distance between the rolling command and the robust optimization benchmark strategy, and minimizes the change in energy storage power adjustment, while satisfying the boundary constraints of the energy storage charge state. The penalty weight is 0.02 × charge-discharge cycle cost. Through a dual dynamic calibration mechanism, the average error of migration prediction is further compressed, and the frequency of invalid energy storage actions is reduced. This enables enterprises to maintain high reliability of dispatch commands even in the absence of string-level monitoring, and completely eliminates the risk of strategy execution inaccuracy caused by the lack of real-time data.

[0079] The dynamic conditional risk value is calculated and the bidding strategy is adjusted in the risk measurement module 7. The risk measurement module 7 outputs the dynamic conditional risk value to the hedging strategy generator 8.

[0080] Dynamic conditional value at risk calculation, on a reduced scenario set, calculates the average of the last 5 percent of the loss distribution;

[0081] The risk buffer adjustment for bid volume is to deduct the positive deviation between the risk value and the expected bid volume from the original bid volume, where the risk value is taken as the tenths of the loss distribution.

[0082] This step dynamically calculates the conditional value of risk on the reduced scenario set through the risk measurement module 7. It addresses the issue of inaccurate tail risk quantification caused by enterprises' failure to collect real-time details of wind and solar power output and electricity prices. Based on the bidding strategy output by robust optimization and the 50 probability-weighted scenarios provided by the scenario generator 4, the Monte Carlo path simulation is used to calculate the average value of the tail 5% of the loss distribution, with a confidence level of 0.95, to capture the potential losses of extreme events, such as a sudden drop in wind and solar power output coupled with negative electricity prices.

[0083] Furthermore, a risk buffer adjustment for the bid volume is implemented, deducting most of the positive deviation between the quantile risk value and the expected bid volume from the original bid volume, thus constructing a two-layer protection structure of "core bid volume + dynamic buffer layer".

[0084] This reduces the tail risk exposure caused by missing data.

[0085] Configure financial derivatives to hedge residual risk in Hedging Strategy Generator 8;

[0086] Residual risk exposure quantification involves calculating the probability-weighted sum of the loss values ​​for the 5% scenario with the greatest potential loss.

[0087] Solving for the optimal hedging amount in CFDs involves finding the contract purchase amount that minimizes the deviation between the risk exposure and the maximum contract return.

[0088] Automated execution of hedging transactions through blockchain smart contracts.

[0089] The tail risk exposure output by the hedging strategy generator 8 and the risk measurement module 7 is the probability weighted sum of the scenario with the maximum loss of 5%. The optimal CFD hedging amount is then solved, which is the quadratic programming solution that minimizes the deviation between the exposure and the maximum return of the contract. The hedging transaction is automatically triggered by the blockchain smart contract to solve the problem of residual risk out of control caused by the failure to collect real-time carbon price and blocking cost data.

[0090] The cross-market correlation risk of wind and solar power output fluctuations and electricity price anomalies is transformed into fixed income of financial contracts. When the real-time electricity price falls below the threshold or the carbon price fluctuation exceeds the limit, the option can be exercised instantly. This further reduces the residual risk exposure caused by data gaps without human intervention, reduces the frequency of extreme losses, and improves the stability of income.

[0091] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A wind-solar-storage coordinated power financial decision-making method, characterized in that, Includes the following steps: Regional meteorological data, regional new energy output data and electricity market data are obtained through the external data access module (1); The enterprise's total load power, energy storage operation status and equipment parameters are collected in real time through the enterprise's local data acquisition module (2); The output and load decomposition calculations are performed in the migration calculation engine (3), and it outputs the photovoltaic output estimate, wind power output estimate and adjustable load to the scene generator (4); In the scene generator (4), the joint probability distribution of wind, solar and load is constructed and a reduced scene set is generated. The reduced scene set containing the joint state of wind, solar and load and real-time electricity price disturbance is output to the robust optimization solver (5). After solving the two-stage optimization problem in the robust optimization solver (5), the baseline scheduling strategy is output to the rolling correction module (6); The two-stage optimization problem is solved in the robust optimization solver (5), specifically including the following steps: the first stage is day-ahead decision optimization, which minimizes the sum of day-ahead market electricity purchase cost and ancillary service revenue, and considers the real-time adjustment cost under the worst scenario; where the uncertainty set is defined as the set of all scenarios whose covariance weighted distance from the benchmark scenario is less than the chi-square distribution critical value; The second-stage real-time adjustment model minimizes the sum of real-time market deviation costs and penalty costs. The constraints include that the energy storage power adjustment amount does not exceed the real-time available power margin, the margin is the difference between the maximum energy storage power and the planned power, and the minimum of the available capacity divided by the time step. In the rolling correction module (6), decision instructions are dynamically corrected based on real-time monitoring data; The dynamic conditional risk value is calculated and the bidding strategy is adjusted in the risk measurement module (7). The risk measurement module (7) outputs the dynamic conditional risk value to the hedging strategy generator (8). Configure financial derivatives to hedge residual risk in the hedging strategy generator (8).

2. The wind-solar-storage coordinated power financial decision-making method according to claim 1, characterized in that, The process of performing output and load decomposition calculations in the migration calculation engine (3) specifically includes the following steps: To perform photovoltaic power output migration calculations, the average regional photovoltaic power output is divided by the total installed photovoltaic capacity in the region to obtain the power output ratio per unit capacity. This ratio is then multiplied by the company's installed photovoltaic capacity, and then by the tilt angle correction factor, system efficiency factor, and temperature degradation compensation factor. The tilt angle correction factor is calculated based on the company's latitude, and the temperature degradation compensation factor is determined based on the difference between the real-time temperature and the standard test temperature. Wind power output migration calculations were performed, and regional wind speed data were corrected according to hub height and surface roughness to obtain the enterprise's hub height wind speed. Divide the corrected wind speed by the rated wind speed of the wind turbine, take the cube, multiply by the enterprise's wind power installed capacity, and then multiply by the air density correction factor. Adjustable load separation is performed, and cluster analysis is conducted on historical total load data to identify load curves under different production modes. The average load rate of each load mode cluster is calculated by subtracting the product of the base load and the load rate of that cluster from the current total load.

3. The wind-solar-storage coordinated power financial decision-making method according to claim 1, characterized in that, The process of constructing a joint probability distribution of wind, solar and load and generating a reduced scene set in the scene generator (4) specifically includes the following steps: The joint probability distribution of wind, solar and load is constructed based on the Gaussian connection function. The marginal distributions of each variable are transformed using the inverse function of the standard normal distribution. Then, a multivariate normal distribution is constructed using the correlation coefficient matrix. The initial scene set was generated using Latin hypercube sampling technique. The K-medoids clustering algorithm based on Mahalanobis distance reduces the scene size, calculates the covariance weighted distance between each scene, and selects representative scene center points to form a reduced scene set. For each scenario, a real-time electricity price disturbance following a Johnson-SU distribution is superimposed.

4. The wind-solar-storage coordinated power financial decision-making method according to claim 1, characterized in that, The step of dynamically correcting decision instructions based on real-time monitoring data in the rolling correction module (6) specifically includes the following steps: The prediction error feedback correction multiplies the difference between the actual output and the predicted value in the current period by a dynamically updated autoregressive coefficient and adds it to the predicted value for the next period. The decision instruction is smoothly adjusted to solve an optimization problem that minimizes the deviation between the new instruction and the original baseline strategy, and minimizes the change in energy storage action. The penalty weight for energy storage action is set according to the equipment loss characteristics.

5. The wind-solar-storage coordinated power financial decision-making method according to claim 1, characterized in that, The calculation of dynamic conditional value of risk and adjustment of bidding strategy in the risk measurement module (7) specifically includes the following steps: Dynamic conditional value at risk calculation, on a reduced scenario set, calculates the average of the last 5 percent of the loss distribution; The risk buffer adjustment for bid volume is to deduct the positive deviation between the risk value and the expected bid volume from the original bid volume, where the risk value is taken as the decimal of the loss distribution.

6. The wind-solar-storage coordinated power financial decision-making method according to claim 1, characterized in that, The configuration of financial derivatives to hedge residual risk in the hedging strategy generator (8) specifically includes the following steps: Residual risk exposure quantification involves calculating the probability-weighted sum of the loss values ​​for the 5% scenario with the greatest potential loss. Solving for the optimal hedging amount in CFDs involves finding the contract purchase amount that minimizes the deviation between the risk exposure and the maximum contract return. Automated execution of hedging transactions through blockchain smart contracts.

Images