Electric power spot market transaction decision-making method and system based on multi-factor dynamic analysis

Abstract

The invention discloses an electric power spot market transaction decision-making method and system based on multi-factor dynamic analysis, and the method comprises the steps: building a power generation power prediction model based on historical power generation data, meteorological data and time information, predicting the power generation power of a future time period based on the power generation power prediction model, and carrying out the uncertainty analysis of the power generation power; establishing a market price prediction model based on the multi-source data, performing prediction based on the market price prediction model to obtain a market price in a future time period, and performing uncertainty analysis of the market price; building a transaction decision optimization model by taking profit maximization as a target and considering power generation cost, market risk and new energy consumption based on the power generation power and the market price of a future time period; and solving the transaction decision optimization model by adopting an improved genetic algorithm to obtain an optimal transaction decision scheme. Scientific, accurate and efficient transaction decision support is provided for the new energy power station in the electric power spot market, and the economic benefits and market competitiveness of the new energy power station are improved.

Description

Technical Field

[0001] This invention relates to the field of electricity market trading technology, and in particular to a method and system for electricity spot market trading decisions based on multi-factor dynamic analysis. Background Technology

[0002] The statements in this section are merely background information relating to this disclosure and do not necessarily constitute prior art.

[0003] With the increasing global demand for clean energy, the proportion of renewable energy power plants in the power system is gradually increasing. As a key component of power market reform, the electricity spot market provides a more direct and flexible platform for renewable energy power plants to trade electricity. However, renewable energy power plants face numerous challenges in participating in electricity spot market transactions.

[0004] On the one hand, new energy power generation is characterized by significant intermittency, volatility, and uncertainty, and is greatly affected by natural factors such as sunlight and wind. This makes it extremely difficult to accurately predict power generation, which in turn affects the formulation of trading strategies. For example, photovoltaic power generation depends on solar irradiance, and weather changes such as cloudy or rainy weather can cause large fluctuations in power generation. Wind power generation depends on wind speed and stability, and the uncontrollable changes in wind speed increase the difficulty of power generation prediction.

[0005] On the other hand, electricity spot market prices fluctuate frequently and are complex, influenced by a combination of factors including electricity supply and demand, generation costs, grid operation status, policies and regulations, and the behavior of market participants. Currently, most new energy power plants lack comprehensive and dynamic analysis of these multiple factors when making trading decisions, and their trading models are too simplistic to accurately reflect market changes. Furthermore, traditional decision-making algorithms are computationally inefficient, struggling to process large amounts of data and generate optimal trading strategies in a short time, putting new energy power plants at a disadvantage in market competition and preventing them from fully realizing their potential. Summary of the Invention

[0006] To overcome the shortcomings of the existing technology, this invention provides a method and system for electricity spot market trading decisions based on multi-factor dynamic analysis. It aims to provide scientific, accurate, and efficient trading decision support for new energy power plants in the electricity spot market, thereby enhancing their economic benefits and market competitiveness. By constructing a multi-factor dynamic analysis model, designing a trading decision optimization model, applying intelligent optimization algorithms, and establishing a risk assessment and decision support mechanism, it provides new energy power plants with scientific and reasonable trading decision schemes, improving their economic benefits and risk resistance in the electricity spot market.

[0007] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions: In a first aspect, the present invention provides a method for electricity spot market trading decisions based on multi-factor dynamic analysis, including: Historical power generation data, meteorological data, and time information are acquired. A power generation prediction model is established based on the historical power generation data, meteorological data, and time information. The power generation power for future periods is predicted based on the power generation prediction model, and uncertainty analysis of power generation power is performed. Acquire multi-source data from the electricity spot market, establish a market price prediction model based on the multi-source data, predict market prices for future periods based on the market price prediction model, and conduct uncertainty analysis on market prices; Based on the power generation capacity and market price in the future period, and with the goal of maximizing profits, a transaction decision optimization model is constructed, taking into account power generation costs, market risks, and renewable energy consumption. An improved genetic algorithm is used to solve the trading decision optimization model to obtain the optimal trading decision scheme.

[0008] A further technical solution is that the power generation prediction model adopts a long short-term memory network and uses the Monte Carlo simulation method to analyze the uncertainty of power generation.

[0009] A further technical solution is that the market price prediction model adopts a hybrid model of convolutional neural networks and recurrent neural networks.

[0010] A further technical solution is proposed, in which the market risks include price fluctuation risk and power generation forecast error risk.

[0011] A further technical solution is that the objective function of the transaction decision optimization model is expressed as:

[0012] in, Represents the true market price. Indicates power generation capacity. Indicates the cost of electricity generation. express Price volatility risk value over a period of time. This indicates the risk value of the power generation prediction error. Indicates the actual consumption ratio. Indicates the target proportion of new energy consumption; , , These are the weighting coefficients.

[0013] A further technical solution is that the constraints of the transaction decision optimization model include power generation constraints, supply and demand balance constraints, and grid security constraints.

[0014] Further technical solutions, based on the optimal trading decision scheme, take into account the uncertainties of power generation and market prices, and use the Monte Carlo simulation method to conduct risk assessment to evaluate the risk level of the trading decision scheme.

[0015] Secondly, this invention provides a power spot market trading decision-making system based on multi-factor dynamic analysis, including: The power generation prediction module is configured to: acquire historical power generation data, meteorological data and time information; establish a power generation prediction model based on the historical power generation data, meteorological data and time information; predict the power generation for future periods based on the power generation prediction model; and perform uncertainty analysis on the power generation. The market price forecasting module is configured to: acquire multi-source data from the electricity spot market, establish a market price forecasting model based on the multi-source data, predict market prices for future periods based on the market price forecasting model, and perform uncertainty analysis on market prices. The optimization model construction module is configured to: construct a transaction decision optimization model based on the power generation and market price of the future period, with the goal of maximizing profits, and considering power generation costs, market risks and new energy consumption; The optimization decision module is configured to use an improved genetic algorithm to solve the transaction decision optimization model and obtain the optimal transaction decision scheme.

[0016] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the electricity spot market trading decision-making method based on multi-factor dynamic analysis as described in the first aspect.

[0017] Fourthly, the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in the electricity spot market trading decision-making method based on multi-factor dynamic analysis as described in the first aspect.

[0018] The above one or more technical solutions have the following beneficial effects: This invention provides scientific and reasonable trading decision-making schemes for new energy power plants by constructing a multi-factor dynamic analysis model (including a power generation prediction model and a market price prediction model), designing a trading decision optimization model, applying intelligent optimization algorithms, and establishing a risk assessment and decision support mechanism, thereby improving their economic benefits and risk resistance in the electricity spot market.

[0019] This invention comprehensively considers multiple influencing factors, including the uncertainty of new energy power generation, fluctuations in electricity spot market prices, power generation costs, grid operation constraints, and new energy consumption. Through dynamic analysis and modeling, it can more accurately grasp the laws of market changes, provide more comprehensive information support for trading decisions, and improve the scientific nature and accuracy of decision-making.

[0020] The transaction decision optimization model constructed in this invention comprehensively considers multiple objectives and solves them using an improved genetic algorithm. It can effectively handle complex optimization problems, quickly find the optimal transaction decision scheme, and improve the transaction efficiency and economic benefits of new energy power plants in the electricity spot market.

[0021] This invention uses methods such as probability distribution, Monte Carlo simulation, and CVaR to quantify risk, and combines them with adaptive genetic algorithm operations to make the optimization results more stable and resilient to market uncertainties, thus helping new energy power plants better cope with market fluctuations and protect their economic interests.

[0022] The decision support system constructed by this invention provides decision-makers with intuitive and comprehensive information and analytical tools. Through various analytical methods, it helps decision-makers to deeply understand the advantages and disadvantages of different decision-making schemes and improve the decision-making level and management capabilities of new energy power plants in the electricity spot market. Attached Figure Description

[0023] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0024] Figure 1 This is a flowchart of a power spot market transaction decision-making method based on multi-factor dynamic analysis, according to an embodiment of the present invention. Detailed Implementation

[0025] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0026] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0027] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0028] Example 1 like Figure 1 As shown in the figure, this embodiment discloses a method for electricity spot market trading decisions based on multi-factor dynamic analysis. The method includes the following steps: S1: Obtain historical power generation data, meteorological data, and time information; establish a power generation prediction model based on the historical power generation data, meteorological data, and time information; predict the power generation for future periods based on the power generation prediction model; and perform uncertainty analysis on the power generation. In this embodiment, for new energy power generation methods such as photovoltaic power generation and wind power generation, a power generation prediction model is established by comprehensively considering historical power generation data, meteorological data (such as solar irradiance, temperature, wind speed, wind direction, etc.), time information and geographical location information (optional), and using a Long Short-Term Memory (LSTM) network.

[0029] Taking photovoltaic power generation as an example, the input data includes historical solar irradiance series, temperature series, and time information. By using an LSTM network to learn the time series characteristics and complex nonlinear relationships in the data, the output is a predicted value of photovoltaic power generation for future periods. Simultaneously, considering the uncertainty of power generation, it is described by calculating the prediction interval and probability distribution. Using the Monte Carlo simulation method, the possible values ​​of power generation are generated through multiple simulations based on the power generation prediction model, and its probability distribution function is obtained. Assuming the photovoltaic power generation... The range of values ​​is within By simulating the frequency of different power values, the probability distribution can be approximated.

[0030] The process of building, training, and predicting power generation is as follows: (1) Model selection and training Taking photovoltaic power generation as an example, a Long Short-Term Memory (LSTM) network is selected for power generation prediction. LSTM networks can effectively handle long-term dependencies in time series data, making them suitable for predicting data with complex fluctuation characteristics, such as new energy power generation.

[0031] 1) Data partitioning: There are N samples in total. The preprocessed data is divided into training set, validation set and test set according to the ratio of 70%, 15% and 15% respectively.

[0032] 2) Model Setup: The LSTM network parameters were set as follows: 128 hidden layer neurons, learning rate of 0.001, batch size of 32, and 200 iterations. Historical solar irradiance sequences, temperature sequences, and time information (represented by timestamps) were used as input features and organized into a feature matrix according to a specific format. X Photovoltaic power generation in future time periods (e.g., 1 hour or 3 hours in advance, depending on the forecast duration). Y The predicted values ​​are used as output labels. The accuracy of the model is verified by comparing the predicted values ​​with historical power generation data.

[0033] 3) Train the LSTM model using the training set data, setting the learning rate to 0.001 and the batch size to 32. Take 32 samples from the training set each time to form a batch, and input them into the model to obtain the power generation prediction value. Calculate the predicted value Compared with the true value The loss is calculated using the mean squared error (MSE) loss function, as shown in the formula below. (1) in, Indicates the first The actual photovoltaic power generation value of each sample is the data obtained from actual measurement. The model represents the first The predicted photovoltaic power generation value for each sample is calculated by the LSTM model based on the input features.

[0034] 4) The model's weights and biases are continuously adjusted using the backpropagation algorithm to gradually reduce the prediction error on the training set. This process is repeated, with each 200 iterations constituting a training cycle. After each training cycle, the model's loss and prediction accuracy are calculated using the validation set data. If the loss on the validation set no longer decreases or the prediction accuracy no longer improves, the model's performance is considered to have reached a point where training is stopped. The resulting trained LSTM model, i.e., the power generation prediction model, is used to predict power generation on the test set data.

[0035] Considering the uncertainty of power generation, this invention utilizes the Monte Carlo simulation method for analysis. Based on a trained power generation prediction model, 1000 simulations are performed. In each simulation, a certain amount of random perturbation is introduced into the predicted value, taking into account the fluctuations in factors such as solar irradiance and temperature from historical data. It is assumed that the historical standard deviation of solar irradiance is... The historical standard deviation of temperature fluctuation is In the simulation, the perturbation values ​​of solar irradiance follow a mean of 0 and a standard deviation of . The temperature follows a normal distribution, and the perturbation values ​​follow a mean of 0 and a standard deviation of . The normal distribution is followed. Perturbed input data is fed into the power generation prediction model, resulting in 1000 different predicted power generation values. Statistical analysis is performed on these prediction values ​​to calculate the prediction interval and probability distribution. Assuming the photovoltaic power generation... The range of values ​​is within By statistically analyzing the frequency of occurrence of 1000 predicted values ​​within different power ranges, the probability distribution is approximated. For example, when calculating photovoltaic power generation within a given range... probability within the interval P The calculation formula is: (2) Among them, the molecule is the one that falls within the specified power range out of 1000 predicted values ​​generated by the Monte Carlo simulation. The number of predicted values ​​within a given range. Through extensive simulations, the probability is approximated by frequency, reflecting the likelihood of photovoltaic power generation occurring within this range, helping to assess the uncertainty distribution of power generation.

[0036] S2: Obtain multi-source data from the electricity spot market, establish a market price prediction model based on the multi-source data, predict the market price for future periods based on the market price prediction model, and conduct uncertainty analysis of the market price. This study collects multi-source data, including historical electricity spot market price data, electricity supply and demand data (installed capacity of various types of power generation, load demand changes, etc.), power generation cost data (fuel prices, equipment maintenance costs, etc.), power grid operation data (transmission line capacity, congestion status, etc.), and policy and regulatory information. A model combining Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) is used to extract and analyze features from this data, capturing the patterns and trends of market price changes. This leads to the establishment of an electricity spot market price forecasting model, which outputs predicted market prices for future periods. Similarly, considering price uncertainty, the standard deviation of price fluctuations can be determined by statistically analyzing historical price volatility. A confidence interval for the price is constructed by combining the predicted value, which is used to measure the degree of uncertainty in price prediction.

[0037] The process of building, training, and predicting market price forecasting models is as follows: (1) Multi-source data acquisition and integration Historical price data from the electricity spot market is comprehensively collected, accurate to the hourly level. Simultaneously, electricity supply and demand data for the same period are collected, including installed capacity of various types of power generation (thermal, hydro, wind, solar, etc.), load demand changes in different regions; power generation cost data, such as fuel prices and maintenance costs of various power generation equipment; grid operation data, such as transmission line capacity and congestion status; and relevant policy and regulatory information, such as subsidy policies and electricity price adjustment policies. This multi-source data is integrated and aligned chronologically to form a comprehensive dataset. Different types of data are normalized or standardized to ensure they have the same scale, facilitating subsequent model training and analysis. For electricity price data... The minimum-maximum normalization method is used, and the formula is: (3) in, It is the first in the original electricity price data One value, It is the normalized value. and These are the minimum and maximum values ​​in the original electricity price data, respectively. After this normalization process, the data can be mapped to... Intervals eliminate differences in the units and numerical ranges of different data features, making model training more effective.

[0038] (2) Model building and training A model combining Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) was constructed for market price prediction. CNNs were used to extract spatial features from the data and capture local correlations between different types of data; RNNs were used to process time-series features and learn price trends over time. The integrated data was divided into training, validation, and test sets at a ratio of 75%, 15%, and 10%, respectively. During model training, the CNN was configured with a 3×3 kernel size, 3 convolutional layers, and 2×2 pooling layers; the RNN used a GRU (Gated Recurrent Unit) with 64 hidden neurons, a learning rate of 0.0005, a batch size of 64, and 300 iterations. Multi-source data, including historical price data, electricity supply and demand data, power generation cost data, grid operation data, and policy and regulatory information, were used as input features, and the market price for different prediction periods (e.g., 1 hour, 6 hours, 1 day in advance) was used as the output label. The model is trained using training set data, evaluated using validation set data, and its root mean square error (RMSE) is monitored to measure the average error between the model's predictions and the true values. Training is stopped when the model's performance on the validation set no longer improves.

[0039] (4) in, To verify the size of the sample in the set, For the first The true market price of a sample For the model to the first The RMSE (Record Price Estimate) measures the predicted market price of a sample. It takes into account both the magnitude and direction of the error, and can intuitively reflect the degree of deviation between the model's predicted value and the actual value, helping to judge the model's predictive performance.

[0040] (3) Quantification of uncertainty Calculate the standard deviation of price volatility At that time, assuming the hourly price data for the past year is (Calculated as 8760 hours per year), price fluctuation value series ,in The formula for calculating the standard deviation of price volatility is: (5) in, This is the mean of the price fluctuation value sequence D. This formula quantifies the uncertainty of price fluctuations by calculating the dispersion of these fluctuations, thus providing a key parameter for constructing price confidence intervals. Price confidence intervals are used to quantify the dispersion (uncertainty) of price fluctuations; by quantifying the uncertainty of price fluctuations, a range of price distributions is obtained. The more prices fall within a certain range (95%), the higher the probability (95%) of the predicted price falling within that range. Therefore, a 95% confidence interval indicates a 95% probability that the predicted price will fall within that range.

[0041] S3: Based on the power generation and market price in the future period, construct a transaction decision optimization model with the goal of maximizing profits, taking into account power generation costs, market risks and new energy consumption; In this embodiment, the main objective is to maximize the profits of new energy power plants in the electricity spot market, while also considering factors such as power generation costs, market risks, and the absorption of new energy.

[0042] (1) The cost of power generation includes fixed costs (equipment investment depreciation, site rental fees, etc.) and variable costs (equipment maintenance fees, raw material consumption, etc.). Let the fixed cost be... Variable costs and power generation Related, the unit variable cost is Then the cost of electricity generation The calculation formula is: (6) in, For time period Power generation capacity, Total number of trading sessions. Power generation capacity. It can be the predicted power output of the power generation prediction model, from which the predicted power generation cost is obtained, or it can be the actual power generation, from which the actual power generation cost is obtained.

[0043] (2) Market risk is measured by price fluctuation risk and power generation forecast error risk.

[0044] 1) Price volatility risk is quantified using the Conditional Value at Risk (CVaR) indicator. For confidence level (e.g., 0.95). For time period The market price, For power generation, then the loss function for: (7) (8) in, Indicates taking The result is a negative value; if the result is positive, the value is 0.

[0045] The formula for calculating conditional value at risk is: (9) in, To be at confidence level Value at risk (VaR) is determined by adjusting the loss function. Sort the data and select the corresponding quantile values.

[0046] 2) The risk of power generation forecast error is measured by the variance of the forecast error, which is expressed as: (10) in, To predict power generation rate.

[0047] (3) New Energy Consumption: Considering the principle of prioritizing the consumption of new energy power plants, a constraint on the proportion of new energy consumption is introduced, and the target proportion of new energy consumption is set as follows: Actual consumption ratio The calculation formula is: (11) in, For time period Total load demand of the power system.

[0048] The objective function of the trading decision optimization model can be expressed as: (12) in, Represents the true market price. Indicates power generation capacity. Indicates the cost of electricity generation. express Price volatility risk value over a period of time. This indicates the risk value of the power generation prediction error. Indicates the actual consumption ratio. Indicates the target proportion of new energy consumption; , , The weighting coefficients are used to balance the relationships between different objectives, and their values ​​are determined based on factors such as the risk appetite of new energy power plants, the market environment, and policy requirements. The objective of the trading decision optimization model is to maximize the returns obtained in the actual market, i.e., market returns minus market risk losses. The actual return is the actual price multiplied by the actual power generation. Market risk gains and losses consist of price fluctuation risk and power generation prediction risk.

[0049] Constraints are set for the transaction decision optimization model, including power generation constraints, supply and demand balance constraints, and grid security constraints.

[0050] Power generation constraints: The power generation capacity of new energy power plants is limited by the installed capacity of equipment and technology. Minimum power generation capacity must be met and maximum power generation The limitations are: (13) Supply and demand balance constraints: In the electricity spot market, the power generation of new energy power plants must match market demand. Considering factors such as grid transmission losses, let the grid transmission efficiency be... Then we have: (14) in, Indicates power grid transmission efficiency. It is a time period The market demand for electricity.

[0051] Power grid safety constraints: Transmission capacity limitations of transmission lines must be considered to prevent line overload. This applies to transmission lines connected to renewable energy power plants. Its transmission power Must meet: (15) in, For the line Maximum transmission capacity.

[0052] S4: The improved genetic algorithm is used to solve the transaction decision optimization model to obtain the optimal solution.

[0053] The solution process for the trading decision optimization model is as follows: S401: Solution Algorithm Selection. An improved genetic algorithm (IGA) is used to solve the trading decision optimization model. A genetic algorithm (GA) is a stochastic search algorithm that simulates the natural evolutionary process, progressively searching for the optimal solution in the solution space through selection, crossover, and mutation operations. The improved genetic algorithm, based on the traditional GA, introduces adaptive crossover and mutation rates, as well as an elite retention strategy, to improve the algorithm's convergence speed and search accuracy.

[0054] S402: Encoding and Population Initialization. The trading decision variables (such as power generation and traded electricity volume in different time periods) are encoded using real numbers, with each individual representing a trading decision scheme. Assume the trading decision optimization model has... Given several decision variables, the chromosome vector for each individual is: (16) in, Indicates the first The individual in the first The values ​​that can be taken on the decision variables.

[0055] When initializing the population, a certain number of individuals are randomly generated within the feasible region of the decision variable to form the initial population. For the power generation decision variable, its value range is... If the initial value is between these ranges, then an initial value is randomly determined within that range.

[0056] S403: Calculate the fitness value: Decode the chromosome vector of each individual into the actual trading decision scheme, substitute it into the objective function of the trading decision optimization model, and calculate its fitness value. (i.e., objective function value) The fitness value reflects the overall performance of the transaction decision scheme represented by the individual in terms of maximizing profits, controlling risks, and meeting the needs of renewable energy consumption. The higher the value, the better the scheme.

[0057] S404: Adaptive Cross Rate Crossover is performed on the selected individuals. The adaptive crossover rate is adjusted based on the individual's fitness value. For individuals with fitness values ​​higher than the average fitness value, a lower crossover rate is used to protect the traits of superior individuals; for individuals with fitness values ​​lower than the average fitness value, a higher crossover rate is used to promote individual diversity and evolution. The crossover operation generates new individuals by randomly selecting two individuals and exchanging partial gene segments between them. Assuming individuals are selected... and If the intersection point is k, then the new individual generated after the intersection... and .

[0058] S405: Adaptive mutation rate Mutation operations are performed on individuals. The adaptive mutation rate is also adjusted based on the individual's fitness value; individuals with lower fitness values ​​have a higher mutation rate to increase population diversity and help the algorithm escape local optima. The mutation operation involves randomly selecting one or more dimensions of an individual's genes for mutation. Using uniform mutation, the mutated gene values for: (17) in, Is Random numbers between and They are the first The maximum and minimum values ​​of the decision variables.

[0059] S406: Set the maximum number of iterations (MaxIter) and the convergence accuracy of the objective function. This serves as a termination condition. The fitness value of the optimal solution in the population changes less than the maximum iterative time (MaxIter) after multiple iterations (e.g., m iterations). When the iteration stops, that is: (18) in, It is the fitness value of the population's optimal solution in the current iteration. It is the fitness value of the optimal solution in the population before m iterations.

[0060] When the termination condition is met, the algorithm stops iterating. At this point, the transaction decision scheme corresponding to the best individual in the population is the optimal solution of the transaction decision optimization model, including the optimal power generation and transaction volume for each time period.

[0061] In this embodiment, the method also includes the construction of a risk assessment and decision support system.

[0062] Based on the optimized trading decision-making scheme, and considering the uncertainties in renewable energy power generation and market prices, a Monte Carlo simulation method is used for risk assessment. According to the probability distribution functions of power generation and market prices, multiple simulations are generated to create different power generation and market price scenarios. Profit and risk indicators (such as CVaR and standard deviation) for renewable energy power plants are calculated under each scenario. Through statistical analysis of numerous simulation results, the distribution of profits and the statistical values ​​of risk indicators are obtained, thus assessing the risk level of the trading decision-making scheme.

[0063] A decision support system is constructed to present optimization results and risk assessment results to decision-makers in an intuitive manner, including information such as power generation arrangements, traded electricity volume, expected profits, and risk indicators for different time periods. Simultaneously, it provides comparative analysis of decision-making schemes under different market scenarios (such as varying electricity supply and demand conditions and price fluctuation ranges) to help decision-makers understand the impact of different decisions on the economic benefits and risks of the power plant. Sensitivity analysis is used to study the impact of key parameters (such as power generation costs, market price prediction model parameters, and risk preference weights) on optimization results and risk assessments, providing a reference for decision-makers to formulate reasonable trading strategies and adjust decision parameters. For example, analyzing the impact of changes in power generation costs on optimal power generation and profits shows that if power generation costs rise, the optimal power generation may be adjusted accordingly, and profits will also change. Decision-makers can formulate contingency strategies in advance based on these analytical results.

[0064] Example 2 This embodiment discloses a power spot market trading decision-making system based on multi-factor dynamic analysis, including: The power generation prediction module is configured to: acquire historical power generation data, meteorological data and time information; establish a power generation prediction model based on the historical power generation data, meteorological data and time information; predict the power generation for future periods based on the power generation prediction model; and perform uncertainty analysis on the power generation. The market price forecasting module is configured to: acquire multi-source data from the electricity spot market, establish a market price forecasting model based on the multi-source data, predict market prices for future periods based on the market price forecasting model, and perform uncertainty analysis on market prices. The optimization model construction module is configured to: construct a transaction decision optimization model based on the power generation and market price of the future period, with the goal of maximizing profits, and considering power generation costs, market risks and new energy consumption; The optimization decision module is configured to use an improved genetic algorithm to solve the transaction decision optimization model and obtain the optimal transaction decision scheme.

[0065] Example 3 The purpose of this embodiment is to provide a computing device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method of Embodiment 1.

[0066] Example 4 The purpose of this embodiment is to provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps of the method of Embodiment 1.

[0067] The steps and methods involved in the apparatuses of Embodiments 3 and 4 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0068] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0069] 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.

[0070] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A decision-making method for electricity spot market transactions based on multi-factor dynamic analysis, characterized in that, include: Historical power generation data, meteorological data, and time information are acquired. A power generation prediction model is established based on the historical power generation data, meteorological data, and time information. The power generation power for future periods is predicted based on the power generation prediction model, and uncertainty analysis of power generation power is performed. Acquire multi-source data from the electricity spot market, establish a market price prediction model based on the multi-source data, predict market prices for future periods based on the market price prediction model, and conduct uncertainty analysis on market prices; Based on the power generation capacity and market price in the future period, and with the goal of maximizing profits, a transaction decision optimization model is constructed, taking into account power generation costs, market risks, and renewable energy consumption. An improved genetic algorithm is used to solve the trading decision optimization model to obtain the optimal trading decision scheme.

2. The electricity spot market transaction decision-making method based on multi-factor dynamic analysis as described in claim 1, characterized in that, The power generation prediction model employs a long short-term memory network and utilizes Monte Carlo simulation to analyze the uncertainty of power generation.

3. The electricity spot market transaction decision-making method based on multi-factor dynamic analysis as described in claim 1, characterized in that, The market price prediction model employs a hybrid model combining convolutional neural networks and recurrent neural networks.

4. The electricity spot market transaction decision-making method based on multi-factor dynamic analysis as described in claim 1, characterized in that, The market risks mentioned include price volatility risk and power generation forecast error risk.

5. The electricity spot market transaction decision-making method based on multi-factor dynamic analysis as described in claim 1, characterized in that, The objective function of the transaction decision optimization model is expressed as: in, Represents the true market price. Indicates power generation capacity. Indicates the cost of electricity generation. express Price volatility risk value over a period of time. This indicates the risk value of the power generation prediction error. Indicates the actual consumption ratio. Indicates the target proportion of new energy consumption; , , These are the weighting coefficients.

6. The electricity spot market transaction decision-making method based on multi-factor dynamic analysis as described in claim 5, characterized in that, The constraints of the transaction decision optimization model include power generation constraints, supply and demand balance constraints, and grid security constraints.

7. The electricity spot market transaction decision-making method based on multi-factor dynamic analysis as described in claim 1, characterized in that, Furthermore, based on the optimal trading decision scheme, and taking into account the uncertainties of power generation and market prices, a Monte Carlo simulation method is used to conduct risk assessment and evaluate the risk level of the trading decision scheme.

8. A power spot market trading decision-making system based on multi-factor dynamic analysis, characterized in that: include: The power generation prediction module is configured to: acquire historical power generation data, meteorological data and time information; establish a power generation prediction model based on the historical power generation data, meteorological data and time information; predict the power generation for future periods based on the power generation prediction model; and perform uncertainty analysis on the power generation. The market price forecasting module is configured to: acquire multi-source data from the electricity spot market, establish a market price forecasting model based on the multi-source data, predict market prices for future periods based on the market price forecasting model, and perform uncertainty analysis on market prices. The optimization model construction module is configured to: construct a transaction decision optimization model based on the power generation and market price of the future period, with the goal of maximizing profits, and considering power generation costs, market risks and new energy consumption; The optimization decision module is configured to use an improved genetic algorithm to solve the transaction decision optimization model and obtain the optimal transaction decision scheme.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the electricity spot market trading decision-making method based on multi-factor dynamic analysis as described in any one of claims 1-7.

10. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the electricity spot market trading decision-making method based on multi-factor dynamic analysis as described in any one of claims 1-7.

Images