An Efficient Operation Method for a Comprehensive Green Hydrogen Metallurgy Energy System

Abstract

The present invention discloses an efficient operation method for a green hydrogen metallurgy integrated energy system. First, a probability interval of uncertain variables is formed by probability distribution fitting. Secondly, a distributionally robust optimization method is adopted to obtain an optimized operation scheme considering multiple operation objectives, and a continuous convex hull approximation strategy is applied to optimize the solution accuracy and calculation efficiency. In the context of coping with energy supply uncertainty, a large number of operation scenarios are generated to evaluate the energy efficiency, flexibility and risk of the system, and dynamic association rules are proposed to support tradable energy blocks and carbon blocks in multi-market transactions. Further analyze the time-series coupling characteristics of energy blocks and carbon blocks, study the bidding strategies and different operation modes, and explore an adaptive trading decision-making method based on multi-agent reinforcement learning. The purpose of the present invention is to improve the safety, economy and flexibility of the green hydrogen metallurgy integrated energy system, propose a multi-market trading support strategy, and provide theoretical and practical support for the efficient utilization of green energy and sustainable development.

Description

Technical Field

[0001] The present invention belongs to the technical field of energy system operation optimization, and particularly relates to an efficient operation method for a green hydrogen metallurgy integrated energy system. Background Art

[0002] With the global emphasis on low-carbon development, green hydrogen metallurgy technology, as an important green steel production method, is developing rapidly. The green hydrogen metallurgy integrated energy system combines renewable energy and hydrogen energy to significantly reduce carbon emissions in the steel production process, becoming an important way for energy structure transformation and carbon emission reduction. However, the high complexity and multi-energy-quality flow coupling characteristics of this system bring many technical challenges, especially in the flexible regulation and optimal operation of the system.

[0003] Currently, research on the green hydrogen metallurgy integrated energy system has established multi-time-scale optimization planning and scheduling models, focusing on the regulation characteristics and flexible regulation capabilities of various flexible adjustable resources in the system (such as grid connection, renewable energy power generation, electrolyzers, hydrogen storage devices, etc.). These models take economy as the main goal and balance the safety, economy, and flexibility of the system through multi-objective optimization methods. In addition, existing research has also explored the cross-link regulation characteristics in the multi-energy-quality flow transmission process, such as the time delay problem in the transmission and conversion processes, providing certain theoretical support for the coordinated optimal operation of the system.

[0004] Although the existing coordinated optimal operation technologies have improved the operation efficiency of the system to a certain extent, there are still significant deficiencies in the face of a complex dynamic coupling environment. First, the dynamic coupling relationship among system energy efficiency, flexibility, and risk has not been fully considered. Especially under the fluctuation of a high proportion of renewable energy, energy efficiency, flexibility, and risk restrict each other, and existing models cannot effectively balance these factors. Second, with the increase in tradable categories such as green hydrogen, green electricity, and carbon blocks, existing modeling technologies lack in-depth research on dealing with the time-series dynamic coupling of tradable energy blocks and carbon blocks and improving flexibility in the market environment. In addition, the optimal operation of the system under the dynamic correlation and multiple uncertainties in multi-market transactions has not been fully resolved. Therefore, the coordinated operation method for the green hydrogen metallurgy integrated energy system needs to be further improved to achieve the efficient operation and flexible control of the system. Summary of the Invention

[0005] Aiming at the above deficiencies in the prior art, the efficient operation method for the green hydrogen metallurgy integrated energy system provided by the present invention solves the problem of insufficient operation efficiency existing in the existing related methods when dealing with the green hydrogen metallurgy integrated energy system with dynamic coupling, multiple uncertainties, and complex constraints.

[0006] To achieve the above-mentioned invention objectives, the technical solution adopted by the present invention is as follows: An efficient operation method for a green hydrogen metallurgy integrated energy system, comprising the following steps:

[0007] S1. Based on the uncertainty variables in the green hydrogen metallurgy integrated energy system, construct a multi-objective optimization operation model for realizing stochastic production simulation based on the distributionally robust optimization method, and use the continuous convex hull approximation strategy to transform it into a linear model;

[0008] S2. Based on the uncertainty of energy supply in the green hydrogen metallurgy integrated energy system, generate a large number of operation scenarios through stochastic production simulation, and construct an oblique decision tree model based on multi-index sparse weights and introducing regularization according to the generated operation scenarios;

[0009] S3. Use the constructed oblique decision tree model to extract the association rules among energy efficiency, flexibility, and risk in the operation scenario data;

[0010] S4. Based on the extracted association rules among energy efficiency, flexibility, and risk and the transformed linear model, conduct sequential coupling analysis and confidence interval analysis on the energy block and carbon block in the green hydrogen metallurgy integrated energy in turn, and then quantitatively infer the coupling characteristics of the energy block and carbon block;

[0011] The energy block and carbon block respectively represent the physical energy transmission and carbon emission quota in the green hydrogen metallurgy integrated energy system;

[0012] S5. Analyze the optimal bidding strategies of the energy block and carbon block in different markets in multi-market transactions through a game theory model;

[0013] S6. Based on the optimal bidding strategies of the energy block and carbon block, construct an adaptive trading auxiliary decision-making model based on multi-agent reinforcement learning, and output the scheduling and trading strategies of the energy block and carbon block through it to realize the efficient operation of the system.

[0014] Further, the step S1 includes the following sub-steps:

[0015] S11. Conduct statistical analysis on the historical operation data in the green hydrogen metallurgy integrated energy system, fit the probability distribution of the uncertainty variables therein, and characterize the uncertainty variables as corresponding uncertainty sets based on the distance between different probability distributions; the uncertainty variables include power supply fluctuations and market price fluctuations;

[0016] S12. Based on the multiple uncertainty conditions of the uncertainty set, use the distributionally robust optimization method to construct a multi-objective optimization operation model for realizing stochastic production simulation;

[0017] S13. Transform the multi-objective optimization operation model with non-linear constraints into a linear model and solve it using the continuous convex hull approximation strategy to obtain a high-value optimal solution.

[0018] Further, in the step S12, the multi-objective optimization operation model for realizing stochastic production simulation is expressed as:

[0019]

[0020] In the formula, x is the decision variable of the system, ξ is the uncertainty variable, f(x, ξ) corresponds to the profit and loss function and optimization objective of the system, θ represents the robustness parameter, which is used to control the trade-off between the expectation and conservatism of the system, represents the expectation function based on , represents the uncertainty set, P ξ represents the probability distribution of the uncertainty variable ξ, represents the reference distribution fitted from the operation historical data;

[0021] The constraint condition of the multi-objective optimization operation model is the extended resource-task network of the green hydrogen metallurgy integrated energy system.

[0022] Further, the step S2 includes the following sub-steps:

[0023] S21. Regard the power supply fluctuation and market price fluctuation as the uncertainty variables of the operation scenario;

[0024] S22. Randomly generate the uncertainty variables included in different operation scenarios, and take maximizing the operation efficiency of the system and minimizing the risk as the optimization objective of the operation scenario, and then generate a large number of operation scenarios;

[0025] S23. Analyze the generated operation scenarios, extract the main characteristics affecting the key indicators of the system operation, and construct an oblique decision tree based on multi-index sparse weights;

[0026] S24. Regularize the sparse weights in the oblique decision tree to obtain the oblique decision tree model.

[0027] Further, in the step S22, the optimization objective of the operation scenario is expressed as:

[0028]

[0029] In the formula, x s represents the decision variable of the operation scenario s, f(x s , ξ s , η s ) represents the operation performance of the system under the operation scenario s, R(x s , ξs , η s ) represents the system risk, η s and ξ s respectively represent the power supply fluctuation and market price fluctuation in the operation scenario s. λ represents the weight coefficient for balancing system performance and risk, s ∈ S, and S represents the set of operation scenarios.

[0030] Furthermore, in the step S24, the objective function E(q) for sparse weight regularization in the oblique decision tree is expressed as:

[0031]

[0032] In the formula, represents the weight entropy, l1 and l2 represent the regularization parameters for controlling the degree of sparsity, |q| and ||q|| represent the sparse regularization terms, W i (q) is expressed as the weight function related to the sample i, H i (q) represents the index or function related to the sample i, N represents the number of samples, and i ∈ I represents the sample index.

[0033] Furthermore, in the step S4, the sequential coupling analysis of the energy block and carbon block is performed through a linear coupling relationship, which is expressed as:

[0034] C(t) = αE(t) + β + ε t

[0035] In the formula, (C(t)) represents the carbon block, (E(t)) represents the energy block, α and β respectively represent the coupling coefficients between energy and carbon emissions, and ε t represents the error term.

[0036] Furthermore, the step S5 is specifically as follows:

[0037] Describe the optimal bidding strategies of the participants in the market through a game theory model. When the bids of each market satisfy the Nash equilibrium condition, obtain the optimal bidding strategies of the energy block and carbon block in different markets by optimizing the utility function that characterizes the total revenue of the energy market and carbon market;

[0038] Among them, the Nash equilibrium condition is expressed as:

[0039]

[0040] Among them, U m (b m , b -m ) represents the utility function of market m under the bids b - m of other markets, and it is expressed as:

[0041]

[0042] where b m represents the price, and b -m represents other prices. E m (b m ) represents the expected utility related to the potential field m and b m . C m (b m ) represents the cost related to the market m and b m . p m represents the price of the market m.

[0043] Furthermore, in step S6, the adaptive trading assistant decision-making model learns and optimizes the trading strategies of energy blocks and carbon blocks by simulating the interaction behaviors of multiple agents in the market, so as to adapt to different market changes, and outputs the scheduling and trading strategies of energy blocks and carbon blocks;

[0044] The optimization objective of the adaptive trading assistant decision-making model is to maximize the cumulative return of each agent through the output scheduling and trading strategies of energy blocks and carbon blocks. The agents include power trading agents, carbon emission rights trading agents, and green hydrogen trading agents.

[0045] The beneficial effects of the present invention are as follows:

[0046] (1) The present invention provides an efficient operation method for a green hydrogen metallurgy integrated energy system. Aiming at the power fluctuations and market information uncertainties caused by renewable energy power generation in the green hydrogen metallurgy integrated energy system, a characterization method for the power supply and market uncertainty sets is studied and constructed, and the probability intervals of uncertain variables are formed through probability distribution fitting. Secondly, based on these uncertainties, a multi-objective optimization problem is studied. The distributionally robust optimization is adopted to balance the expectation and conservatism, and at the same time, the nonlinear constraints of the multi-energy-quality flow transmission network are solved. The solution accuracy and efficiency are improved through the continuous convex hull approximation strategy. Finally, a large number of operation scenarios are generated, the energy efficiency, flexibility, and risks of the system are evaluated, dynamic association rules are proposed to support the management of energy blocks and carbon blocks in multi-market trading, and an adaptive trading decision-making method based on multi-agent reinforcement learning is explored. The present invention aims to improve the safety, economy, and flexibility of the green hydrogen metallurgy integrated energy system, propose multi-market trading support strategies, and provide theoretical and practical support for the efficient utilization of green energy and sustainable development.

[0047] (2) The present invention proposes to construct the probability intervals of uncertain variables through probability distribution fitting in the green hydrogen metallurgy integrated energy system, and adopt the distributionally robust optimization method to balance the objectives of system safety, economy, and flexibility. At the same time, aiming at the nonlinear constraints of the multi-energy-quality flow transmission network, the continuous convex hull approximation strategy is innovatively applied to improve the optimization solution accuracy and calculation efficiency.

[0048] (3) The present invention innovatively proposes a method for extracting the association rules among energy efficiency, flexibility, and risk based on large-scale operation scenarios. By deeply mining the multi-dimensional operation data of the integrated energy system for green hydrogen metallurgy, the complex dynamic association relationships among energy efficiency, flexibility, and risk in different scenarios of the system are extracted. This method is based on data-driven analysis and uses technical tools such as decision trees and regression models to deeply reveal the interactions and influence paths among the system performance indicators. This innovation provides accurate decision-making support for the optimal operation and scheduling strategies of the system, effectively improving the overall efficiency and risk control ability of the system in the face of uncertainty and complex market environments.

[0049] (4) The present invention proposes an analysis method for the time-series coupling characteristics of energy blocks and carbon blocks in the integrated energy system for green hydrogen metallurgy, and designs an adaptive trading decision-making method through multi-agent reinforcement learning technology. This innovation provides decision-making support for the optimization of energy blocks and carbon blocks in multi-market trading, effectively improving the flexibility and control efficiency of the system in complex and changeable market environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a flowchart of the efficient operation method for the integrated energy system for green hydrogen metallurgy provided by the present invention.

[0051] Figure 2 It is a schematic block diagram of the stochastic production simulation considering multiple uncertainties of power supply and market provided by the present invention.

[0052] Figure 3 It is a schematic block diagram of the dynamic quantitative analysis of energy efficiency - flexibility - risk based on the oblique decision tree model provided by the present invention.

[0053] Figure 4 It is a schematic block diagram of the flexible and efficient operation based on multi-market trading of energy blocks and carbon blocks provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0054] The following describes the specific embodiments of the present invention to facilitate those skilled in the art of the present technology to understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those of ordinary skill in the art of the present technology, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions created using the concept of the present invention are within the scope of protection.

[0055] The embodiments of the present invention provide an efficient operation method for an integrated energy system for green hydrogen metallurgy, as Figure 1 shown, including the following steps:

[0056] S1. Based on the uncertain variables in the integrated green hydrogen metallurgy energy system, construct a multi-objective optimization operation model for stochastic production simulation using the distributionally robust optimization method, and transform it into a linear model using the continuous convex hull approximation strategy;

[0057] S2. Based on the uncertainty of energy supply in the integrated green hydrogen metallurgy energy system, generate a large number of operation scenarios through stochastic production simulation, and construct an oblique decision tree model based on multi-index sparse weights and introducing regularization according to the generated operation scenarios;

[0058] S3. Use the constructed oblique decision tree model to extract the association rules between energy efficiency, flexibility, and risk in the operation scenario data;

[0059] S4. Based on the extracted association rules between energy efficiency, flexibility, and risk and the transformed linear model, conduct sequential coupling analysis and confidence interval analysis on the energy block and carbon block in the integrated green hydrogen metallurgy energy, and then quantitatively infer the coupling characteristics of the energy block and carbon block;

[0060] The energy block and carbon block respectively represent the physical energy transmission and carbon emission quota in the integrated green hydrogen metallurgy energy system;

[0061] S5. Analyze the optimal bidding strategies of the energy block and carbon block in different markets in multi-market trading through a game theory model;

[0062] S6. Based on the optimal bidding strategies of the energy block and carbon block, construct an adaptive trading auxiliary decision-making model based on multi-agent reinforcement learning, and output the scheduling and trading strategies of the energy block and carbon block through it to achieve the efficient operation of the system.

[0063] Step S1 of the embodiment of the present invention includes the following sub-steps:

[0064] S11. Conduct statistical analysis on the historical operation data in the integrated green hydrogen metallurgy energy system, fit the probability distribution of the uncertain variables therein, and characterize the uncertain variables as corresponding uncertainty sets based on the distance between different probability distributions; the uncertain variables include power supply fluctuations and market price fluctuations;

[0065] S12. Under the multiple uncertainty conditions of the uncertainty set, use the distributionally robust optimization method to construct a multi-objective optimization operation model for stochastic production simulation;

[0066] S13. Use the continuous convex hull approximation strategy to transform the multi-objective optimization operation model with non-linear constraints into a linear model and solve it to obtain a high-value optimal solution.

[0067] In step S11 of this embodiment, during the operation of the green hydrogen metallurgy integrated energy system, the power supply fluctuations and market price changes have significant uncertainties. To effectively address these uncertainties, an uncertainty set characterization method based on probability distribution fitting is proposed in this embodiment.

[0068] Specifically, by statistically analyzing historical operation data, the uncertainties of power supply fluctuations (such as new energy power generation output fluctuations) and market price changes (such as power spot market price fluctuations) are respectively modeled. Using probability distribution fitting technology, the probability intervals of multiple power supply and market uncertainty variables are generated, which can effectively represent various possible operating states and provide a data basis for subsequent optimal scheduling.

[0069] In an example of this embodiment, when describing the uncertainties of power supply and market, first use the probability distribution fitting method to generate the uncertainty set. Set the power supply fluctuation (such as new energy power generation fluctuation) and market price fluctuation as random variables ξ, and its probability distribution is P ξ . Suppose there is historical data {ξ1, ξ2,......., ξ N}, the probability distributions of these variables can be estimated by kernel density estimation or other probability distribution fitting methods.

[0070] For the characterization of the uncertainty set, the Wasserstein distance is used to measure the distance between different probability distributions. The definition of the Wasserstein distance W is:

[0071]

[0072] Among them, P1 and P2 are two probability distributions respectively, and ∏(P1, P2) represents the joint distribution set of the two distributions. Through this distance, the probability distribution differences in different power supply and market environments can be described.

[0073] In this embodiment, in order to characterize the uncertainty set, the distribution set of the uncertainty variable is defined as:

[0074]

[0075] Among them, is the reference distribution fitted by historical data, and ρ is the Wasserstein radius, indicating the degree of uncertainty that the system can accept.

[0076] In step S12 of this embodiment, considering the impact of power supply and market uncertainties on system operation, the Distributionally Robust Optimization (DRO) method is introduced to optimize stochastic production simulation. In this process, robustness parameters are introduced through the optimization model to enable the system to balance the expectations and conservativeness of operation under different power supply and market environments. Specifically, the optimization objectives cover the safety, economy, and flexibility of system operation. In a stochastic environment, distributionally robust optimization can not only provide a stable and reliable production strategy for the green hydrogen metallurgy system but also minimize operation risks to the greatest extent.

[0077] Specifically, in this embodiment, under multiple uncertainty conditions, the Distributionally Robust Optimization (DRO) method is used. Considering the stochastic production optimization problem of the green hydrogen metallurgy system, the objective function can be expressed as the worst-case expectation under the uncertainty set:

[0078]

[0079] In the formula, x represents the decision variables of the system (such as power generation scheduling, market trading volume, etc.), ξ represents the uncertainty variables (such as power supply fluctuations, market prices, etc.), and f(x, ξ) is the profit and loss function of the system (including indicators such as economy, safety, and flexibility).

[0080] To further consider the safety and flexibility of system operation, the extended resource-task network of the green hydrogen metallurgy integrated energy system is introduced as a model constraint condition to enable the system to achieve an optimized effect under a certain robustness level, and the robustness parameter θ is introduced. The multi-objective optimization operation model for realizing stochastic production simulation is expressed as:

[0081]

[0082] In the formula, x is the decision variable of the system, ξ is the uncertainty variable, f(x, ξ) corresponds to the profit and loss function and optimization objective of the system, θ represents the robustness parameter, which is used to control the trade-off between expectations and conservativeness of the system, represents the expectation function based on of, represents the uncertainty set, P ξ represents the probability distribution of the uncertainty variable ξ, represents the reference distribution fitted by historical operation data;

[0083] In step S1 of the embodiment of the present invention, in view of the complexity and non-linear constraints of the multi-energy mass flow transmission network in the system, a continuous envelope approximation strategy is proposed in this embodiment. Since the non-linear problems in system operation are usually difficult to solve directly, the use of this approximation strategy can transform complex non-linear problems into linear problems, improving the accuracy and efficiency of the solution. Through the continuous convex hull technology, it is possible to quickly find a high-value optimal solution with limited computing resources, so as to improve the operation efficiency of the system in the face of multiple uncertainties.

[0084] Specifically, in an example of this embodiment, it is assumed that the objective function f(x) in the original problem is non-linear and difficult to solve directly, and the problem can be simplified by finding its convex hull approximation; assume that f(x) is a non-linear function, and its linear approximation at the point x k is:

[0085]

[0086] Through this linearization method, the original problem is approximated as a linear programming (LP) problem, so that a high-quality solution can be obtained with limited computing resources. In the optimization solution process, this strategy can be used to improve the computing efficiency and ensure that the accuracy of the solution meets the operation requirements of the system.

[0087] As Figure 2 shown, it shows the principle of the stochastic production simulation of the integrated green hydrogen metallurgy energy system considering multiple uncertainties of power supply and market in step S1 above.

[0088] Step S2 of the embodiment of the present invention includes the following sub-steps:

[0089] S21. Take the power supply fluctuation and the market price fluctuation as the uncertainty variables of the operation scenario;

[0090] S22. Randomly generate the uncertainty variables included in different operation scenarios, and take maximizing the operation efficiency of the system and minimizing the risk as the optimization goal of the operation scenario, and then generate a large number of operation scenarios;

[0091] S23. Analyze the generated operation scenarios, extract the main characteristics affecting the key indicators of system operation, and construct an oblique decision tree based on multi-index sparse weights;

[0092] S24. Regularize the sparse weights in the oblique decision tree to obtain the oblique decision tree model.

[0093] In step S22 of this embodiment, on the basis of the implementation process of the above step S1, in order to comprehensively evaluate the operating performance of the green hydrogen metallurgy integrated energy system under different asynchronous deterministic conditions, a large number of different operating scenarios are generated through stochastic production simulation. By simulating a large number of possible power supply and market environments, the energy efficiency, flexibility, and risk characteristics of the system under various operating conditions can be captured. These scenario data not only provide a solid foundation for subsequent optimization and decision-making, but also provide a comprehensive risk assessment and flexibility analysis tool for system operation.

[0094] Specifically, the power supply fluctuation ξ and the market price fluctuation η are set as uncertainty variables, and different scenario sets S are constructed. Each scenario s ∈ S contains variables (ξ s , η s ) generated by random sampling or Monte Carlo simulation. For each scenario, the optimization goal is to maximize the operating efficiency of the system while minimizing the risk; based on this, in step S22 of this embodiment, the optimization goal of the operating scenario is expressed as:

[0095]

[0096] In the formula, x s represents the decision variable of the operating scenario s, f(x s , ξ s , η s ) represents the operating performance of the system under the operating scenario s, R(x s , ξ s , η s ) represents the system risk, η s and ξ s respectively represent the power supply fluctuation and the market price fluctuation in the operating scenario s, λ represents the weight coefficient for balancing the system performance and risk, s ∈ S, and S represents the set of operating scenarios.

[0097] In step S23 of this embodiment, on the basis of a large number of operating scenarios, an oblique decision tree based on multi-index sparse weights is proposed to dynamically quantify and analyze the energy efficiency, flexibility, and risk of the system. By introducing sparse weights, this oblique decision tree extracts the main features affecting the key indicators of system operation, and at the same time controls the complexity of the model through regularization methods to avoid overfitting. The model can not only efficiently process multi-dimensional data, but also dynamically identify the internal relationships between various indicators during system operation, and provide high-precision quantitative analysis for multi-objective optimization problems under uncertain environments.

[0098] Specifically, the oblique decision tree is a decision tree based on linear segmentation, used to process multi-dimensional and continuous data. Let the linear segmentation function of the k-th node in the decision tree be:

[0099]

[0100] Among them, x is the input feature vector (such as indicators like energy efficiency, flexibility, risk, etc.), ω k is the sparse weight vector of the oblique decision tree at the k-th node, and b k is the bias term.

[0101] In step S24 of this embodiment, in order to ensure the sparsity of the model and avoid overfitting, regularization is introduced. The objective function E(q) for regularizing the sparse weights in the oblique decision tree is expressed as:

[0102]

[0103] In the formula, represents the weight entropy, l1 and l2 represent the regularization parameters for controlling the degree of sparsity, |q| and ||q|| represent the sparse regularization terms, W i (q) is expressed as the weight function related to sample i, H i (q) represents the index or function related to sample i, N represents the number of samples, and i ∈ I represents the sample index.

[0104] In this embodiment, through regularization processing, the model can automatically select the most important features, remove the variables that have less impact on the system operation, thereby avoiding the overfitting phenomenon of complex models.

[0105] In step S3 of the embodiment of the present invention, using the large-scale data of the operation scenario, a data-driven method for extracting the association rules of energy efficiency, flexibility, and risk is proposed. Through data mining technology, the complex dynamic relationship between energy efficiency, flexibility, and risk is extracted. This rule extraction process is based on technologies such as decision trees and regression analysis, and can deeply explore the correlation between various performance indicators in the system, and provide a reference basis for the optimal scheduling and operation strategy of the system. This method provides theoretical support for the efficient operation of the integrated energy system in the face of uncertainty by analyzing the energy efficiency performance and risk characteristics under different operation scenarios.

[0106] Specifically, in this embodiment, based on the operation scenario data, data mining technology is used to extract the association rules between energy efficiency, flexibility, and risk. Specifically, the above-mentioned oblique decision tree model is used to identify the correlation between variables. It is assumed that the correlation between the target variable y (such as energy efficiency) and the input feature x can be expressed by a regression model as:

[0107]

[0108] Among them, β iis the regression coefficient, and ε is the error term. Through decision tree analysis, significant relationships between energy efficiency and indicators such as flexibility and risk under different operating scenarios can be identified, providing a basis for system optimization and operation strategies.

[0109] In optimal scheduling, energy efficiency, flexibility, and risk are usually set as multi-objective optimization problems. A typical multi-objective optimization model is:

[0110] min f(E,F,R)=w1E+w2F+w3R(10)

[0111] Where w1, w2, and w3 are the weights of energy efficiency, flexibility, and risk respectively, representing their importance in system optimization.

[0112] By restricting the constraint conditions, it can be ensured that the optimization results meet the system requirements:

[0113] subject to:E≥E min F≥F min R≥R max (11)

[0114] Thus, the system optimization goals of maximizing energy efficiency, enhancing flexibility, and minimizing risk are achieved.

[0115] Such as Figure 3 shown, the above shows the principle of dynamic quantitative analysis of energy efficiency - flexibility - risk of the green hydrogen metallurgy integrated energy system based on the oblique decision tree in steps S2 and S3.

[0116] In step S4 of the embodiment of the present invention, for the energy block and carbon block in multi-market trading, a time-series coupling analysis method is proposed and quantitatively evaluated through confidence inference. The energy block and carbon block respectively represent the transmission of physical energy and the change of carbon emission allowances in the integrated energy system, and their trading in the market has a high degree of correlation. Through the coupling analysis model, the interaction between energy and carbon is studied, especially its transmission and trading modes in different time periods. In this embodiment, an inference method based on the confidence interval is also proposed to quantify the coupling characteristics of the energy block and carbon block under different market conditions, providing quantitative analysis results for system optimization.

[0117] Specifically, in step S4 of this embodiment, there is an interrelationship between energy and carbon emissions at different time periods t. The time-series coupling analysis of the energy block and carbon block is carried out in sequence through a linear coupling relationship, which is expressed as:

[0118] C(t)=αE(t)+β+ε t (12)

[0119] Wherein, (C(t)) represents a carbon block, (E(t)) represents an energy block, α and β respectively represent the coupling coefficients between energy and carbon emissions, and ε t represents an error term.

[0120] Furthermore, in order to evaluate the confidence level of this coupling relationship, a confidence interval is constructed. Assuming that energy and carbon emissions under a set of samples satisfy a normal distribution, the confidence interval can be expressed as:

[0121]

[0122] Wherein, is the estimated coupling coefficient, is the quantile of the standard normal distribution, σ α is the estimated standard error, and n is the number of samples.

[0123] In this embodiment, through confidence interval analysis, quantitative inference can be made on the coupling characteristics of the energy block and the carbon block.

[0124] In step S5 of the embodiment of the present invention, quantitative inference is made on the coupling characteristics of the energy block and the carbon block. By performing multi-market tendering and bidding coupling analysis, the bidding mechanism and market coupling characteristics between different markets are explored. Based on this, step S5 in this embodiment is specifically:

[0125] Describe the optimal bidding strategies of the participants in the market through a game theory model. When the bids of each market satisfy the Nash equilibrium condition, by optimizing the utility function that characterizes the total revenue of the energy market and the carbon market, the optimal bidding strategies of the energy block and the carbon block in different markets are obtained;

[0126] Among them, the Nash equilibrium condition is expressed as:

[0127]

[0128] Among them, U m (b m ,b -m ) represents the utility function of market m under the bids b-m of other markets, and it is expressed as:

[0129]

[0130] In the formula, b m represents the price, b -m represents other prices, E m (b m ) represents the expected utility related to potential field m and b m , and C m (b m ) represents the relationship with market m and b mThe related cost, p m represents the price of market m.

[0131] Through the above optimized utility function, the bidding strategies of energy blocks and carbon blocks in different markets can be obtained, realizing the coupling analysis of the energy and carbon markets.

[0132] In this embodiment, the above analysis not only helps to improve the market competitiveness of the integrated energy system, but also can effectively promote the coordinated development between the carbon market and the energy market, achieving a double improvement in economic and environmental benefits.

[0133] In step S6 of the embodiment of the present invention, the adaptive trading assistant decision-making model learns and optimizes the trading strategies of energy blocks and carbon blocks by simulating the interaction behavior of multiple agents in the market to adapt to different market changes, and outputs the scheduling and trading strategies of energy blocks and carbon blocks; the optimization goal of the adaptive trading assistant decision-making model is to maximize the cumulative return of each agent through the output scheduling and trading strategies of energy blocks and carbon blocks;

[0134] In this embodiment, the agents include power trading agents, carbon emission rights trading agents, and green hydrogen trading agents obtained based on the coupling time series analysis of carbon blocks and energy.

[0135] Specifically, in this embodiment, in order to improve the flexibility and adaptability of energy blocks and carbon blocks in market trading, the multi-agent reinforcement learning technology is introduced to design the adaptive trading assistant decision-making model; in this model, the interaction between multiple agents is continuously improved through reinforcement learning, enabling the system to adapt to different market changes and make the best decisions in a complex trading environment. Through continuous policy update and optimization, this method can achieve the efficient scheduling and trading of energy blocks and carbon blocks in a multi-market environment, enhancing the market response ability and profitability of the system.

[0136] Exemplarily, a multi-agent reinforcement learning (MARL) model is introduced. Suppose there are N agents, and the state of each agent i is s i (t), the action is a i (t), and the reward is γ i (t). The goal of reinforcement learning is to maximize the cumulative return R of each agent through the policy π i : i :

[0137]

[0138] Through continuous interactions among agents, the reinforcement learning system can adaptively adjust its strategy to optimize the market trading of energy blocks and carbon blocks. Ultimately, such a multi-agent system can achieve the best scheduling and trading decisions in a complex market environment.

[0139] As Figure 4 shown, it demonstrates the principle of the flexible and efficient operation of the integrated green hydrogen metallurgy energy system based on multi-market trading of energy blocks and carbon blocks in the above steps S4 to S6.

[0140] In the present invention, specific embodiments are used to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only for helping to understand the method of the present invention and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

[0141] Those of ordinary skill in the art will realize that the embodiments described herein are for helping readers understand the principle of the present invention, and it should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those of ordinary skill in the art can make various specific deformations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed in the present invention, and these deformations and combinations are still within the protection scope of the present invention.

Claims

1. An efficient operation method for a green hydrogen metallurgy integrated energy system, characterized in that It includes the following steps: S1. Based on the uncertain variables in the integrated green hydrogen metallurgy energy system, construct a multi-objective optimization operation model for stochastic production simulation using the distributionally robust optimization method, and transform it into a linear model using the continuous convex hull approximation strategy; S2. Based on the uncertainty of energy supply in the integrated green hydrogen metallurgy energy system, generate a large number of operation scenarios through stochastic production simulation, and construct an oblique decision tree model based on multi-index sparse weights and introducing regularization according to the generated operation scenarios; S3. Use the constructed oblique decision tree model to extract the association rules among energy efficiency, flexibility, and risk in the operation scenario data; S4. Based on the extracted association rules among energy efficiency, flexibility, and risk and the transformed linear model, conduct sequential time-series coupling analysis and confidence interval analysis on the energy block and carbon block in the integrated green hydrogen metallurgy energy, and then quantitatively infer the coupling characteristics of the energy block and carbon block; The energy block and carbon block respectively represent the physical energy transmission and carbon emission quota in the integrated green hydrogen metallurgy energy system; S5. Analyze the optimal bidding strategies of the energy block and carbon block in different markets in multi-market transactions through a game theory model; S6. Based on the optimal bidding strategies of the energy block and carbon block, construct an adaptive trading auxiliary decision-making model based on multi-agent reinforcement learning, and output the scheduling and trading strategies of the energy block and carbon block through it to achieve the efficient operation of the system; The step S1 includes the following sub-steps: S11. Conduct statistical analysis on the historical operation data in the integrated green hydrogen metallurgy energy system, fit the probability distribution of the uncertain variables therein, and characterize the uncertain variables as corresponding uncertainty sets based on the distance between different probability distributions; the uncertain variables include power supply fluctuations and market price fluctuations; S12. Based on the multiple uncertainties of the uncertainty set, use the distributionally robust optimization method to construct a multi-objective optimization operation model for stochastic production simulation; S13. Use the continuous convex hull approximation strategy to transform the multi-objective optimization operation model with non-linear constraints into a linear model and solve it to obtain a high-value optimal solution; In the step S12, the multi-objective optimization operation model for stochastic production simulation is expressed as: In the formula, is the decision variable of the system, is the uncertainty variable, is the profit and loss function of the system corresponding to the optimization objective, represents the robustness parameter, which is used to control the trade-off between the expectancy and conservatism of the system, represents based on the expectation function, represents the uncertainty set, represents the uncertainty variable the probability distribution of, represents the reference distribution fitted by fitting the operation historical data; The constraint condition of the multi-objective optimization operation model is the extended resource-task network of the integrated green hydrogen metallurgy energy system; The step S2 includes the following sub-steps: S21. Regard power supply fluctuations and market price fluctuations as the uncertain variables of the operation scenario; S22. Randomly generate the uncertain variables included in different operation scenarios, and take maximizing the operation efficiency of the system and minimizing the risk as the optimization objective of the operation scenario, and then generate a large number of operation scenarios; S23. Analyze the generated operation scenarios, extract the main characteristics affecting the key indicators of system operation, and construct an oblique decision tree based on multi-index sparse weights; S24. Regularize the sparse weights in the oblique decision tree to obtain the oblique decision tree model.

2. The efficient operation method of the integrated green hydrogen metallurgy energy system according to claim 1, characterized in that In the step S22, the optimization objective of the operation scenario is expressed as: In the formula, represents the decision variable of the operation scenario s, represents the operation performance of the system under the operation scenario , represents the system risk, and respectively represent the power supply fluctuation and the market price fluctuation in the operation scenario , represents the weight coefficient for balancing the system performance and risk, , represents the set of operation scenarios.

3. The efficient operation method of the integrated green hydrogen metallurgy energy system according to claim 1, characterized in that, In the step S24, the objective function of sparse weight regularization in the oblique decision tree is expressed as: In the formula, represents the weighted entropy, and represent the regularization parameter for controlling the degree of sparsity, and represent the sparse regularization term, is expressed as the weight function related to the sample i , represents the index or function related to the sample i ; represents the number of samples, represents the sample index.

4. The efficient operation method of the green hydrogen metallurgy integrated energy system according to claim 1, characterized in that, In the step S4, conduct sequential time-series coupling analysis on the energy block and carbon block through a linear coupling relationship, which is expressed as: In the formula, represents a carbon block, represents an energy block, and respectively represent the coupling coefficients between energy and carbon emissions, represents an error term.

5. The efficient operation method of the integrated green hydrogen metallurgy energy system according to claim 1, wherein The specific steps of step S5 are as follows: Describe the optimal bidding strategies of participants in the market through a game theory model. When the bids in each market satisfy the Nash equilibrium condition, obtain the optimal bidding strategies of energy blocks and carbon blocks in different markets by optimizing the utility function that represents the total revenue of the energy market and the carbon market. Among them, the Nash equilibrium condition is expressed as: Among them, represents the market The utility function under bidding in other markets is expressed as: The utility function under bidding in other markets is expressed as: In the formula, Indicates price, Indicates other prices, Represents the potential field m and The expected utility associated with Representation and Market and The associated costs, Indicates the market price.

6. The efficient operation method of the integrated green hydrogen metallurgy energy system according to claim 1, characterized in that In step S6, the adaptive trading assistant decision-making model learns and optimizes the trading strategies of energy blocks and carbon blocks by simulating the interactive behaviors of multiple agents in the market to adapt to different market changes, and outputs the scheduling and trading strategies of energy blocks and carbon blocks. The optimization objective of the adaptive trading assistant decision-making model is to maximize the cumulative return of each agent through the output scheduling and trading strategies of energy blocks and carbon blocks. The agents include electricity trading agents, carbon emission rights trading agents, and green hydrogen trading agents.

Images