A multi-energy coupling scheduling method for hydrogen-based steel industry park considering green steel constraint

Abstract

This invention discloses a multi-energy coupled scheduling method for hydrogen-based steel industrial parks that considers green steel constraints. First, an operational model of the park's equipment is constructed, including a refined operation model for vertical shaft furnaces, an electric arc furnace, and an electrolytic cell. The refined operation model for vertical shaft furnaces includes a process model and a thermal model. Then, based on the park's equipment operation model, an optimal scheduling model is constructed. Finally, the optimal scheduling problem is modeled as a constrained Markov decision process (CMM) considering green steel constraints, and the green steel certification constraint is modeled as a cost constraint within the CMM. A Lagrange constrained near-end strategy optimization algorithm is used to solve the CMM model considering green steel constraints, yielding executable actions, i.e., the scheduling results. This method achieves accurate and reliable solutions for multi-energy coupled scheduling in hydrogen-based steel industrial parks, providing strong support for the low-carbon economic operation and green steel certification compliance of these parks.

Description

Technical Field

[0001] This invention belongs to the field of integrated energy system optimization and scheduling technology, specifically a multi-energy coupling scheduling method for hydrogen-based steel industrial parks that takes into account green steel constraints. Background Technology

[0002] The steel industry is a high-energy-consuming and high-carbon-emission sector. Hydrogen-based direct reduction steelmaking, which uses hydrogen instead of coal as a reducing agent, has become a key path for the steel industry's low-carbon transformation. Hydrogen-based steel industrial parks integrate wind power, photovoltaics, electrolytic hydrogen production, thermochemical hydrogen production, hydrogen-based shaft furnaces, electric arc furnaces, and various types of energy storage (electricity, heat, and hydrogen) into a single system for scheduling. Meanwhile, the international market has imposed green steel certification requirements on low-carbon steel products, setting clear upper limits on carbon emission intensity per ton of steel. Therefore, optimizing the scheduling of hydrogen-based steel industrial parks under multiple uncertainties, ensuring both compliance with green steel certification standards and economic operation, is a pressing issue that needs to be addressed in this field.

[0003] However, the current optimized scheduling of hydrogen-based steel industrial parks faces the following two key technical challenges: (1) Inaccurate characterization of the operating mechanism of hydrogen-based shaft furnaces: Two closely coupled physical processes occur simultaneously inside hydrogen-based shaft furnaces: the chemical reaction of hydrogen reducing iron ore and heat transfer. Both hydrogen utilization rate and heat demand exhibit nonlinear and temperature-dependent characteristics. However, existing scheduling models mostly treat the effective utilization rate of hydrogen as a fixed constant and use a linear proportional coefficient to describe the heat demand of the shaft furnace, ignoring the nonlinear variation of utilization rate with hydrogen supply rate and furnace temperature, as well as the dependence of heat demand on furnace temperature. This leads to a systematic bias in the characterization of the actual hydrogen metallurgical operating mechanism by the scheduling models. At the same time, existing models lack explicit modeling of the distribution relationship between the regenerator waste heat in the shaft furnace auxiliary heating and the park's heating network, making it difficult to accurately characterize the heat coupling relationship between the shaft furnace and the park's heating network.

[0004] (2) Scheduling problems involving green steel constraints are difficult to solve under multiple uncertainties: The operation of the park faces multiple uncertainties such as wind and solar power output, real-time electricity prices, and the intensity of carbon emissions from the power grid. The scheduling problem is a sequential decision problem, and the green steel constraint is a cumulative ratio constraint over the entire cycle. Traditional mathematical programming methods are difficult to fully cope with multiple uncertainties, and the nonlinearity introduced by the refined modeling of hydrogen-based vertical furnaces will exacerbate the difficulty of convergence. The standard proximal policy optimization algorithm takes the maximization of cumulative reward as the sole optimization objective, but does not set a mechanism for handling constraints in the algorithm structure. Although the Lagrange constraint proximal policy optimization algorithm can handle cost constraints, its standard implementation requires the construction of an independent cost-value network, which increases training overhead. Moreover, the cost advantage estimation using the discounted cumulative form is incompatible with the non-discounted full-cycle cumulative evaluation logic of green steel constraints, which can easily lead to the attenuation of long-term constraint violation signals.

[0005] To address the aforementioned issues, this invention proposes a multi-energy coupling optimization scheduling method for hydrogen-based steel industrial parks that takes into account green steel constraints. By refining the modeling of the vertical shaft furnace process and thermal processes, and by subjecting the cost constraints of green steel constraints to sequential decision-making frameworks, the shortcomings of existing technologies are overcome. Summary of the Invention

[0006] To address the shortcomings of existing technologies, the technical problem this invention aims to solve is to provide a multi-energy coupling scheduling method for hydrogen-based steel industrial parks that takes into account green steel constraints. This method solves the technical problems that existing scheduling models do not accurately characterize the operating mechanism of hydrogen-based vertical shaft furnaces and that existing optimization methods are unable to solve scheduling models that take into account green steel constraints under multiple uncertainties.

[0007] The present invention solves the aforementioned technical problem by adopting the following technical solution: A multi-energy coupling scheduling method for hydrogen-based steel industrial parks considering green steel constraints, characterized by comprising the following steps: Step 1: Construct an operational model for the park's equipment, including a refined operational model for vertical shaft furnaces, an operational model for electric arc furnaces, and an operational model for electrolytic cells; The refined operation model of the vertical shaft furnace includes the vertical shaft furnace process model and the vertical shaft furnace thermodynamic model; the vertical shaft furnace process model is represented as follows: (6) (7) (5) (4) In the formula, for The production rate of sponge iron from the vertical shaft furnace during the period. The yield coefficient of sponge iron per unit of net hydrogen consumption in the reduction reaction. for Iron ore consumption rate over a period of time. This represents the iron ore consumption coefficient corresponding to the unit output of sponge iron. for Net hydrogen consumption of the reduction reaction during the period for Hydrogen recovery volume in tail gas during the period for Hydrogen utilization rate of vertical shaft furnace during the period for Total hydrogen supply from vertical shaft furnace during the period This represents the thermodynamic equilibrium upper limit for hydrogen utilization. The effective void volume of the reduction zone in a hydrogen-based vertical shaft furnace. The porosity of the material bed in a hydrogen-based vertical shaft furnace. The density of hydrogen gas under standard conditions. The nominal equivalent reaction characteristic time at the reference temperature. As the apparent activation energy, Let be the ideal gas constant. for Equivalent temperature of the reduction zone in the vertical shaft furnace during the time period For reference temperature; The thermal model of the vertical shaft furnace is represented as follows: (13) In the formula, for Equivalent temperature of the reduction zone in the vertical shaft furnace during the time period The equivalent thermal mass of the material in the reduction zone of the vertical furnace, The average specific heat capacity of the materials in the furnace. For scheduling step size, for The heat output power of the electric heating furnace during the time period for The auxiliary heating power of the time-period regenerator for the vertical furnace. for Total heat demand power of vertical shaft furnace during the time period; Step 2: Based on the park equipment operation model, construct an optimized scheduling model; The objective function for optimizing the scheduling model is as follows: (17) (18) In the formula, The total cost of the park during the scheduling period. for The overall cost of the park during the period for Cost of raw materials purchased during the period; for Electricity purchase cost during specific time periods for Carbon emission costs per period for Thermochemical hydrogen production costs over time for Equipment maintenance costs during specific time periods for The cost of penalties for wind and solar power curtailment during certain periods This represents the total number of time periods in the scheduling cycle. The constraints of the optimized scheduling model include vertical shaft furnace operation constraints, hydrogen production equipment operation constraints, new energy output constraints, energy bus balance constraints, energy storage unit operation constraints, steel production constraints, and green steel certification constraints; among them, the green steel certification constraints are: (36) In the formula, The carbon emission coefficient for 1 ton of crude steel. for Crude steel production rate over a period of time for Carbon emission intensity of electricity purchased from the power grid during a given period for Power purchased by the park during different time periods The carbon emission coefficient for the entire life cycle of purchased hydrogen. for The amount of hydrogen purchased during the period, The carbon emission coefficient for the entire life cycle of purchased iron ore. for Iron ore consumption rate over a period of time. The maximum carbon emission intensity per ton of steel allowed for green steel certification; Step 3: Model the optimization scheduling problem as a constrained Markov decision process that takes into account the green steel constraint; The state is defined as: (37) In the formula, for The state of the time period for Theoretical maximum available power of wind power during a given period for Theoretical maximum available power of photovoltaic power during a given period for Time-of-use electricity pricing For time period Basic electrical load of the park for The basic heat load of the park during the period for Hydrogen storage capacity of the time-limited hydrogen storage tank. for The basic hydrogen load of the park during the period for Time-of-use energy storage capacity for Time-of-use thermal energy storage for storing heat. for Input power of the electrolytic cell during the time period for Remaining steel production to be completed during the period; , up to Cumulative carbon emissions and cumulative steel production over a given period; Action is defined as: (38) In the formula, for Actions during a period of time, for Input electrical power of the electrolytic cell during the time period for Hydrogen production capacity of time-phase thermochemical hydrogen production units for Actual wind power utilization during the period for Actual photovoltaic power utilization during the period for Power consumption of the electric heating furnace during the period for Waste heat distribution ratio of the time-period regenerator for Time-of-use energy storage capacity, for Net power of thermal energy storage during the time period for Net hydrogen exchange volume during the hydrogen storage period; The reward function is: (42) In the formula, for Instant rewards for specific time periods; The cost function is: (43) (44) In the formula, for Dynamic green steel evaluation indicators for different time periods for The immediate cost of a given time period; The optimization objective is to find the optimal strategy that maximizes the expected cumulative discount reward and the expected cumulative discount cost constraints, which are as follows: (45) (46) In the formula, In strategy The expected cumulative discount reward In strategy The expected cumulative discount cost, For the expectation, To constrain tolerance, for Discount factor for the time period; Step 4: The Lagrange constrained near-end policy optimization algorithm is used to solve the constrained Markov decision process model that takes into account the green steel constraint to obtain the executable actions, i.e. the scheduling results.

[0008] Furthermore, the electric arc furnace operation model is represented as follows: (15) In the formula, for Power consumption of electric arc furnace during specific time periods The electric arc furnace power consumption coefficient is the unit of sponge iron. The electrolytic cell operating model is represented as follows: (16) In the formula, for Hydrogen production per time period of electrolyzer For the efficiency of the electrolytic cell, The lower heating value is the calorific value of hydrogen per unit volume.

[0009] Furthermore, the cost of the purchased raw materials is expressed as follows: (20) The cost of carbon emissions is expressed as: (twenty one) In the formula, The unit carbon price; The cost of thermochemical hydrogen production is expressed as: (twenty two) In the formula, The fixed costs of the thermochemical hydrogen production unit during the scheduling cycle, The variable cost coefficient per unit of hydrogen production; Equipment maintenance costs are expressed as follows: (twenty three) In the formula, For the collection of equipment in the park; For the first The unit operating volume maintenance cost coefficient for this type of equipment; for Time period Operating volume of this type of equipment; The penalty cost of curtailing wind and solar power is expressed as follows: (twenty four) In the formula, This is the wind curtailment penalty coefficient; This represents the penalty coefficient for discarded light.

[0010] Compared with the prior art, the beneficial effects of the present invention are: (1) Refined modeling of hydrogen-based vertical shaft furnace process and thermodynamic process: Based on gas residence time and Arrhenius reaction kinetics, an equivalent model of hydrogen effective utilization rate is established that simultaneously couples hydrogen supply rate and furnace temperature; the heat absorption of reduction reaction, heat consumption of iron ore heating and heat dissipation of furnace body are combined into a joint function of temperature and sponge iron production, and the equivalent temperature of vertical shaft furnace is introduced as a dynamic state variable, so that the process model and thermodynamic model form a closed loop coupling; the waste heat distribution ratio of regenerator is introduced as a continuous control variable, and the recovered heat is explicitly diverted into two paths: auxiliary vertical shaft furnace heating and connection to the park heat network. The heat flow of each path is directly determined by controllable variables.

[0011] (2) Lagrange constraint near-end policy optimization algorithm considering green steel constraint: The scheduling problem is constructed as a constrained Markov decision process, and the portion of the real-time carbon emission intensity per ton of steel that exceeds the certification limit is used as the immediate cost to realize the time-by-time monitoring of green steel constraint; the policy objective function is constructed as the difference between the reward side and the cost side pruning term, so that the constraint penalty term has a non-zero gradient with respect to the policy network parameters.

[0012] The refined modeling of hydrogen-based vertical furnaces and the explicit allocation method of waste heat from regenerators provide a system model foundation with high physical fidelity for optimized scheduling. Meanwhile, the Lagrange-constrained near-end strategy optimization algorithm, which takes into account green steel constraints, provides an efficient sequential decision-making solution framework based on this model. Together, they achieve accurate and reliable solutions for multi-energy coupled scheduling in hydrogen-based steel industrial parks, providing strong support for the low-carbon economic operation and green steel certification of hydrogen-based steel industrial parks. Attached Figure Description

[0013] Figure 1 This is a diagram of the architecture of a hydrogen-based steel industrial park; Figure 2 This is the overall flowchart. Detailed Implementation

[0014] Specific embodiments are given below with reference to the accompanying drawings. These specific embodiments are only used to describe the technical solution of the present invention in detail and are not intended to limit the scope of protection of this application.

[0015] like Figure 1As shown, the hydrogen-based steel industrial park integrates a power grid, wind power units, photovoltaic units, electrolytic hydrogen production units, thermochemical hydrogen production units, a purchased hydrogen supply channel, electric heating furnaces, hydrogen-based shaft furnaces, electric arc furnaces, regenerators, condenser purifiers, and three types of energy storage devices: electricity, heat, and hydrogen. Through electrolytic and thermochemical hydrogen production, electrical and thermal energy are converted into hydrogen energy to provide a reducing agent for the hydrogen-based shaft furnaces, achieving low-carbon steelmaking by replacing carbon with hydrogen. The dispatching system aims to optimize the coordinated operation of each unit to achieve low-carbon economic operation while meeting the park's electricity, heat, hydrogen, and steel product needs. The key challenge lies in ensuring that the final steel products meet green steel certification standards when relying on the power grid for supplemental electricity for electrolytic hydrogen production.

[0016] like Figure 2 As shown, this invention proposes a multi-energy coupling scheduling method for hydrogen-based steel industrial parks that takes into account green steel constraints, including the following steps: Step 1: Construct an operational model for the park's equipment, including a refined operational model for vertical shaft furnaces, an operational model for electric arc furnaces, and an operational model for electrolytic cells; 1.1 Refined Operation Model of Vertical Shaft Furnace The vertical shaft furnace is the core equipment for steelmaking in the industrial park. Its operation involves two closely coupled physical processes: the chemical reaction of hydrogen reduction of iron ore and heat transfer. Existing scheduling models often treat the effective utilization rate of hydrogen as a fixed constant and use a linear proportionality coefficient to describe the heat demand of the furnace. This ignores the nonlinear variation of hydrogen utilization rate with hydrogen supply rate and temperature, as well as the dependence of heat demand on furnace temperature. Consequently, the scheduling models deviate from the actual operating mechanism of hydrogen metallurgy. Therefore, this invention constructs a refined operation model for vertical shaft furnaces to characterize the actual operating mechanism of hydrogen-based vertical shaft furnaces. The refined operation model for hydrogen-based vertical shaft furnaces consists of two parts: a process model and a thermodynamic model. The process model is used to characterize the material transformation relationships inside the vertical shaft furnace, with the total hydrogen supply and the equivalent temperature of reducing gas as the core inputs. Key variables such as sponge iron production and iron ore consumption are derived through a nonlinear hydrogen utilization rate model. The thermodynamic model is used to characterize the heat balance and temperature changes inside the vertical shaft furnace, establishing the coupling relationship between the furnace's heat demand and the furnace temperature. The process model and the thermodynamic model form a closed-loop coupling through the equivalent temperature and the sponge iron production rate.

[0017] Vertical shaft furnace process model: (1) Construct a coupled model of hydrogen utilization rate of vertical furnace with hydrogen supply rate and temperature; During the operation of a hydrogen-based vertical shaft furnace, not all the supplied hydrogen (reducing gas) participates in the iron ore reduction reaction. A portion enters the reduction process as net hydrogen consumption, while the other portion is discharged as tail gas and recycled after condensation and purification. The ratio of these two components is determined by the effective utilization rate of hydrogen in the vertical shaft furnace. Numerical simulations and experimental studies of the hydrogen-based vertical shaft furnace show that hydrogen utilization is significantly affected by two factors. Firstly, as the hydrogen flow rate increases, its residence time in the furnace shortens, and it is discharged from the top of the furnace before fully reacting with the iron ore, thus decreasing the hydrogen utilization rate with increasing flow rate. Secondly, the furnace temperature determines the reduction reaction rate; as the temperature increases, the reaction rate increases exponentially, resulting in more complete hydrogen conversion within the same residence time, and thus increasing the hydrogen utilization rate with increasing temperature. Therefore, a nonlinear coupling model of the vertical shaft furnace hydrogen utilization rate with hydrogen supply rate and temperature is constructed in three steps. First, based on the fundamental principles of chemical reactor design, the average residence time of hydrogen in a hydrogen-based vertical shaft furnace is determined by the ratio of the effective void volume in the reduction zone to the hydrogen volumetric flow rate. After converting the hydrogen supply mass rate to the equivalent volumetric flow rate under standard conditions, the equivalent residence time of hydrogen in the vertical shaft furnace is expressed as: (1) In the formula, for The equivalent residence time of hydrogen in the vertical shaft furnace during a given period; The effective void volume of the reduction zone in a hydrogen-based vertical shaft furnace; The porosity of the material bed in a hydrogen-based vertical shaft furnace; This refers to the density of hydrogen gas under standard conditions. for Total hydrogen supply from the vertical shaft furnace during the specified time period; Next, the hydrogen reduction process of iron ore can be described at the microscale by a reduced-core model. This model indicates that the reduction reaction of iron ore particles proceeds from the surface inwards, and the reaction rate is jointly controlled by two mechanisms: interfacial chemical reaction and product layer diffusion. The corresponding chemical reaction rate constant and effective diffusion coefficient of the product layer both obey the Arrhenius law, i.e., they increase exponentially with increasing temperature. The equivalent reaction characteristic time comprehensively reflects the combined effect of the above two mechanisms and therefore varies with the equivalent temperature of the shaft furnace reduction zone. Using the equivalent Arrhenius form, a macroscopic reduction-order model of the temperature dependence of the equivalent reaction characteristic time is performed, unifying the temperature effects of the two rate-limiting mechanisms—chemical reaction and diffusion—into a single apparent activation energy parameter. Therefore, the equivalent reaction characteristic time of the shaft furnace is: (2) In the formula, for The equivalent temperature in the reduction region during this period The equivalent reaction characteristic time and equivalent temperature of the lower vertical furnace reflect the constraint of the furnace's thermal inertia on current heating and production decisions. The nominal equivalent reaction characteristic time at the reference temperature; The apparent activation energy reflects the temperature sensitivity of both the chemical reaction and product layer diffusion rate-limiting mechanisms. It is the ideal gas constant; For reference temperature, the median of the design operating temperature of the vertical shaft furnace is usually taken; when hour, This characterizes how high temperatures accelerate the reaction, thus shortening the equivalent reaction characteristic time; when hour, The characteristic of low temperature slowing down the reaction leads to a prolonged equivalent reaction characteristic time; when When this happens, the model degenerates into the constant case.

[0018] Finally, hydrogen utilization rate essentially reflects the proportion of hydrogen that participates in the effective reduction reaction during its residence time in the furnace; when the residence time is sufficient and the temperature is high, hydrogen reacts fully with iron ore, and the utilization rate approaches the upper limit of thermodynamic equilibrium; when the residence time is insufficient or the temperature is low, the reaction is incomplete, and the utilization rate decreases; a saturation-type function is used. The nonlinear relationship between hydrogen utilization rate and the ratio of residence time to reaction characteristic time is characterized. This functional form is often used in reaction engineering to describe the change in reaction conversion rate with the characteristic time ratio. It has the physical characteristics of linear growth at zero and approaching the upper limit of saturation far from zero, making it suitable as an equivalent functional form of utilization rate. Based on this, a vertical shaft furnace hydrogen utilization rate model that simultaneously couples hydrogen supply rate and temperature is constructed as follows: (3) Substituting the expressions for equivalent residence time and equivalent reaction characteristic time into the equation, we obtain the coupling model of hydrogen utilization rate of the vertical shaft furnace with hydrogen supply rate and temperature as follows: (4) In the formula, for Hydrogen utilization rate of vertical shaft furnace during a given period, i.e., the mass proportion of hydrogen supplied to participate in the reduction reaction; The upper limit of the thermodynamic equilibrium of hydrogen utilization rate is limited within the allowable operating temperature range of the vertical furnace. The equilibrium conversion rate corresponding to the midpoint of this temperature range is taken as a constant approximation, and the relative error introduced therefrom is acceptable in engineering.

[0019] The above-mentioned coupling model of hydrogen utilization rate of vertical shaft furnace with hydrogen supply rate and temperature has the following physical characteristics: when the hydrogen supply rate is low, the residence time is sufficient, the exponential term approaches zero, and the utilization rate approaches the upper limit of thermodynamic equilibrium; when the hydrogen supply rate is high, the residence time is shortened, the exponential term increases, and the utilization rate decreases significantly; when the temperature rises, the equivalent reaction characteristic time decreases, which is equivalent to a more complete reaction within the same residence time, and the utilization rate increases; the opposite is true when the temperature decreases. This model forms a complete closed-loop coupling relationship between the vertical shaft furnace process model and the thermodynamic model. The equivalent temperature of the vertical shaft furnace affects the equivalent reaction characteristic time through the Arrhenius relation, which in turn affects the hydrogen utilization rate and the sponge iron production rate. The sponge iron production rate, in turn, affects the vertical shaft furnace temperature in the next period through the total heat demand of the vertical shaft furnace.

[0020] (2) Model of material balance and sponge iron production in vertical shaft furnace Substituting the hydrogen utilization rate of the vertical shaft furnace into the material balance model of the vertical shaft furnace, the net hydrogen consumption of the reduction reaction and the amount of hydrogen recovered from the tail gas are obtained as follows: (5) In the formula, for Net hydrogen consumption of the reduction reaction during the time period; for The amount of hydrogen recovered from the tail gas during the period; when the hydrogen supply rate increases, although the total hydrogen supply increases, the hydrogen utilization rate decreases simultaneously, which causes the growth rate of net hydrogen consumption in the reduction reaction to gradually slow down, showing a marginal decrease characteristic.

[0021] In the vertical shaft furnace, the net hydrogen consumption of the reduction reaction reduces iron ore to sponge iron. The sponge iron production rate is determined by the stoichiometric relationship of the reduction reaction, and the iron ore consumption rate is directly proportional to the sponge iron production rate, as expressed respectively: (6) (7) In the formula, for The production rate of sponge iron in the vertical shaft furnace during the time period shows a non-linear increase and marginal decrease in the sponge iron production rate relative to the hydrogen supply rate, since the hydrogen utilization rate of the vertical shaft furnace is a non-linear decreasing function of the hydrogen supply rate. The yield coefficient of sponge iron corresponding to the net hydrogen consumption per unit of reduction reaction; for Iron ore consumption rate over a period of time; The iron ore consumption coefficient corresponding to the unit output of sponge iron; Vertical shaft furnace thermal model: Significant heat demand exists during vertical shaft furnace operation, primarily including heat absorption during reduction reaction, heat loss due to iron ore heating, and heat loss from the furnace body. The hydrogen reduction of iron ore is a strongly endothermic reaction; the heat carried by the hydrogen gas entering the furnace must simultaneously meet the heat absorption requirements of the reduction reaction and the iron ore heating requirements. Otherwise, the furnace temperature will continuously decrease, affecting reduction efficiency. All three heat consumption items are related to the vertical shaft furnace temperature. Existing scheduling models use a fixed proportional coefficient to represent the total heat demand of the vertical shaft furnace as a linear function of the production rate, which is insufficient to reflect the coupling characteristics with temperature. Therefore, the equivalent temperature of the reduction zone in the vertical shaft furnace is introduced as a dynamic state variable, and a lumped parameter method is used to establish a vertical shaft furnace thermal model. (1) Combine the heat absorption of the reduction reaction, the heat consumption of iron ore heating and the heat loss of the furnace body to construct a total heat demand model for the vertical furnace; (8) In the formula, for Total heat demand power of vertical shaft furnace during the time period; It is the basic heat production demand coefficient, reflecting the baseline heat endothermic demand of the reduction reaction corresponding to a unit output of sponge iron. The heat consumption coefficient for iron ore heating reflects the unit heat required for the iron ore to rise from the feed temperature to the reaction temperature inside the furnace. The heat dissipation coefficient reflects the intensity of heat loss from the vertical furnace to the outside world; The temperature at which iron ore enters the furnace; The ambient temperature.

[0022] (2) Vertical furnace heat source structure and regenerator waste heat distribution model The heat supply to the vertical shaft furnace within the park comes from two sources. One source is an electric heating furnace that converts electrical energy into high-temperature heat and sends it directly into the furnace. This heat is used only to heat the hydrogen entering the furnace and does not contribute to the park's basic heat load. The other source is a regenerator that recovers waste heat from the exhaust gas discharged from the top of the furnace. The waste heat recovered by the regenerator is physically divided into two outputs: one is returned to the furnace to assist in preheating the gas entering the furnace, and the other is connected to the park's heating network to meet the basic heat load. The thermal energy storage only exchanges heat with the park's heating network and does not directly supply heat to the vertical shaft furnace.

[0023] The total waste heat recovery power of the regenerator is determined by the exhaust gas flow rate and the furnace temperature. (9) In the formula, for Total waste heat recovery power of the time-period regenerator; To improve the heat exchange efficiency of the regenerator; The specific heat capacity of hydrogen at constant pressure; The waste heat distribution ratio of the regenerator represents the proportion of the recovered heat used to assist the vertical furnace. Therefore, the auxiliary heating power of the regenerator to the vertical furnace and its heating power to the park's heating network are respectively: (10) (11) In the formula, for The auxiliary heating power of the time-period regenerator to the vertical furnace; for The heating capacity of the time-period regenerator to the park's heating network; for The waste heat distribution ratio of the regenerator during each time period is dynamically determined by the dispatching decision based on the relative relationship between the vertical furnace heat demand and the park's heat load during each time period. When the vertical furnace heat demand is large, it can be appropriately increased. To reduce the burden on the electric heating furnace, the load is reduced when the park's heat load is high. Prioritize ensuring basic heat load.

[0024] (3) Dynamic equation of vertical furnace temperature The heat output of the electric heating furnace is: (12) In the formula, for Heat output power of electric heating furnace during a given period; The electrothermal conversion efficiency of the electric heating furnace; for Power consumption of electric heating furnace during specific time periods; Combining the heat output of the electric heating furnace with the auxiliary heating power of the regenerator for the vertical shaft furnace, based on the law of conservation of energy, the dynamic update equation for the equivalent temperature of the reduction zone of the vertical shaft furnace is: (13) In the formula, for Equivalent temperature of the reduction zone in the vertical shaft furnace during a specific time period; The equivalent thermal mass of the material in the reduction zone of the vertical furnace; The average specific heat capacity of the materials inside the furnace; For scheduling step size; The right side of the equation reflects the difference between the net heat input and the total heat demand on the vertical furnace side. When the net input is greater than the total heat demand, the furnace temperature rises, and vice versa.

[0025] 1.2 Electric Arc Furnace Operation Model The sponge iron produced in the vertical shaft furnace is smelted in the electric arc furnace to form steel products. The electric arc furnace is the core equipment in the process chain that converts sponge iron into steel products, and it is also one of the main power-consuming equipment in the industrial park. The yield of the electric arc furnace reflects the efficiency of the quality conversion from sponge iron to steel products, and satisfies the following formula: (14) In the formula, for Crude steel production rate over a period of time For electric arc furnace yield; The power consumption of an electric arc furnace is directly proportional to the amount of sponge iron it processes, therefore: (15) In the formula, for Power consumption of electric arc furnace during specific time periods; The electric arc furnace power consumption coefficient is the unit of sponge iron. 1.3 Electrolyzer Operation Model (16) In the formula, for Hydrogen production per time period in the electrolyzer; for Input electrical power of the electrolytic cell during the time period; For electrolytic cell efficiency; The lower heating value is the calorific value of hydrogen per unit volume.

[0026] Step 2: Based on the park equipment operation model, construct a multi-energy coupling optimization scheduling model for the hydrogen-based steel industrial park that takes into account green steel constraints; 2.1 The operating costs of the hydrogen-based steel industrial park originate from multiple sources, including the purchase of iron ore and hydrogen, the purchase of electricity, carbon emissions, thermochemical hydrogen production operations, equipment maintenance, and penalties for unused wind and solar resources. To balance economic efficiency and low carbon emissions, and with the goal of minimizing the overall cost within the scheduling cycle, the objective function of the multi-energy coupled optimization scheduling model for the hydrogen-based steel industrial park, taking into account green steel constraints, is as follows: (17) (18) In the formula, This represents the overall cost of the park during the scheduling period; for The overall cost of the park during the same period; for Cost of raw materials purchased during the period; for Electricity purchase cost during specific time periods; for Carbon emission costs over time; for Cost of thermochemical hydrogen production over time; for Equipment maintenance costs during specific time periods; for The penalty costs for curtailing wind and solar power during certain periods; This represents the total number of time periods in the scheduling cycle. The industrial park requires iron ore and purchased hydrogen during the production process. The cost of these purchased raw materials is expressed as follows: (19) In the formula, This refers to the unit price of iron ore; This refers to the unit price of purchased hydrogen. for The amount of hydrogen purchased during the specified period; When the solar and wind power output of the industrial park is insufficient to support the electricity consumption for hydrogen production and steelmaking, electricity needs to be purchased from the grid. The cost of purchasing electricity is expressed as follows: (20) In the formula, for Time-of-use electricity pricing; for Power purchased by the park during the specified time period; The park's operation involves direct emissions from technological processes, indirect emissions from electricity purchases, and upstream emissions related to purchased hydrogen and iron ore. The carbon emission cost is expressed as follows: (twenty one) In the formula, The unit carbon price; The carbon emission coefficient for 1 ton of crude steel; for Crude steel production rate over a given period; for Carbon emission intensity of electricity purchased from the power grid during a given time period; The carbon emission coefficient for the entire life cycle of purchased hydrogen; The carbon emission coefficient for the entire life cycle of purchased iron ore; The cost of thermochemical hydrogen production during the scheduling cycle consists of two parts: fixed operating costs and variable operating costs. Therefore, the cost of thermochemical hydrogen production is expressed as: (twenty two) In the formula, The fixed cost of the thermochemical hydrogen production unit during the scheduling cycle; The variable cost coefficient per unit of hydrogen production; for Hydrogen production capacity of a time-phase thermochemical hydrogen production unit; The various equipment in the park incur maintenance and operation management costs during operation. Therefore, the equipment operation and maintenance cost is expressed as follows: (twenty three) In the formula, For the collection of equipment in the park; For the first The unit operating volume maintenance cost coefficient for this type of equipment; for Time period The operating volume of equipment is calculated by power for electrical equipment and production rate for gas production equipment. To improve the utilization rate of wind and solar resources and curb wind and solar curtailment, a penalty cost for wind and solar curtailment is set to guide the system to prioritize the use of renewable energy within the park, as shown below: (twenty four) In the formula, This is the wind curtailment penalty coefficient; This is the penalty coefficient for discarded light; for The theoretical maximum available power of wind power during a given time period; for Actual wind power utilization during the time period; for Theoretical maximum available power of photovoltaics during a given time period; for Actual photovoltaic power utilization during the time period.

[0027] 2.2 The constraints of the optimized scheduling model include vertical furnace operation constraints, hydrogen production side equipment operation constraints, new energy output constraints, energy bus balance constraints, energy storage unit operation constraints, steel production constraints, and green steel certification constraints; (1) Operating constraints of vertical shaft furnace The internal temperature of the vertical furnace must be maintained within the effective reaction temperature range to ensure that the reduction reaction proceeds normally. If the temperature is below the lower limit, the reaction rate will be too slow, and if the temperature is above the upper limit, it may cause the pellets to agglomerate. The power consumption of the electric heating furnace and the waste heat distribution ratio of the regenerator must meet their respective capacity constraints. (25) In the formula, , These are the minimum and maximum hydrogen supply capacities allowed for the vertical shaft furnace, respectively. , These are the minimum and maximum permissible operating temperatures for the vertical shaft furnace, respectively. This refers to the maximum allowable input power of the electric heating furnace. (2) Operational constraints of hydrogen production equipment Electrolyzers are a crucial source of green hydrogen supply for the industrial park. Their hydrogen production capacity depends on the input electrical power and is limited by rated power and ramp rate. The operating constraints of electrolyzers are as follows: (26) In the formula, This represents the maximum input electrical power of the electrolytic cell; This represents the maximum ramp rate of the electrolytic cell; for Input power of the electrolytic cell during the time period; Thermochemical hydrogen production units, serving as supplementary hydrogen sources, are constrained by their maximum hydrogen production capacity. (27) In the formula, This refers to the maximum rated hydrogen production capacity of the thermochemical hydrogen production unit. (3) Constraints on new energy output The actual utilization power of wind and solar power is constrained by their predicted upper limit of output: (28) (4) Energy bus balance constraints To ensure the physical consistency of energy supply and consumption across different time periods, real-time balance constraints are established for three types of busbars: electrical energy, hydrogen energy, and thermal energy. On the electrical energy side, the power grid, wind power, photovoltaic power, and energy storage discharge together constitute the power supply source, while electrolytic cells, electric heating furnaces, electric arc furnaces, energy storage charging, and the park's basic electrical load constitute the electricity demand. Therefore, the balance constraints on the electrical energy side are as follows: (29) In the formula, , They are respectively Periodic energy storage charging and discharging power, for Power consumption of the electric heating furnace during the period for Power consumption of electric arc furnace during specific time periods For time period Basic electrical load of the park; On the hydrogen energy side, purchased hydrogen, hydrogen produced by electrolysis, hydrogen produced by thermochemical processes, hydrogen recovered from tail gas of hydrogen-based vertical shaft furnaces, and hydrogen released from hydrogen storage together constitute the hydrogen source. Hydrogen demand consists of hydrogen supply from hydrogen-based vertical shaft furnaces, basic hydrogen load in the industrial park, and hydrogen charging from hydrogen storage. Therefore, the balance constraint on the hydrogen energy side is: (30) In the formula, for Hydrogen recovery volume in tail gas during the period , They are respectively Net hydrogen storage and net hydrogen release during different time periods for Total hydrogen supply from vertical shaft furnace during the period for The basic hydrogen load of the park during the period; On the thermal energy side, the heat output of the electric heating furnace is entirely used for heating the reducing gas in the hydrogen-based vertical furnace, directly characterized by the temperature dynamic equation of the hydrogen-based vertical furnace, and does not enter the park's heat network balance. The heat source of the park's heat network is the heating power supplied to the network by the regenerator and the heat released by the thermal energy storage. The demand consists of the park's basic heat load and the heat stored in the thermal energy storage. Therefore, the balance constraint on the thermal energy side is: (31) In the formula, for The heating capacity of the time-period regenerator to the park's heating network. , They are respectively Net thermal storage capacity and net heat release capacity of time-limited thermal energy storage for The basic heat load of the park during the period; (5) Energy storage unit operation constraints Electric energy storage is used for peak shaving and valley filling and to improve the local consumption capacity of renewable energy. Its operational constraints are as follows: (32) In the formula, , They are respectively , Time-of-use energy storage capacity; , They are respectively Periodic energy storage charging and discharging power, This refers to the rated capacity of the electrical energy storage. and These are the charging efficiency and discharging efficiency of electrical energy storage, respectively. and These are the lower and upper limits of the state of charge of electrical energy storage, respectively. and These are the maximum allowable charging power and the maximum allowable discharging power of electric energy storage, respectively. , These are the initial and final charge of the energy storage during the scheduling cycle, respectively. Thermal energy storage is used to buffer fluctuations in the park's heat load. Considering the natural heat loss rate in its dynamic process, the operational constraints of thermal energy storage are: (33) In the formula, , They are respectively , Time-of-use thermal energy storage; This refers to the natural loss rate of thermal energy storage. This refers to the rated capacity of thermal energy storage. and These are the heat storage efficiency and heat release efficiency of thermal energy storage, respectively. and These represent the lower and upper limits of the thermal energy storage state, respectively. and They are respectively Maximum allowable thermal storage power and maximum allowable heat release power for time-limited thermal energy storage; , These are the initial and final heat storage capacity of thermal energy storage during the scheduling cycle, respectively. Hydrogen energy storage is used to store hydrogen during periods of low electricity prices or surplus wind and solar power, and release it during peak hydrogen consumption periods in vertical shaft furnaces, thereby solving the time scale mismatch problem between hydrogen production and consumption. The operational constraints of hydrogen energy storage are: (34) In the formula, , They are respectively , Hydrogen storage capacity of the hydrogen storage tank during a given time period; This refers to the natural loss rate of hydrogen energy storage. This refers to the rated capacity of the hydrogen storage tank. and These are the hydrogen storage efficiency and hydrogen release efficiency of hydrogen energy storage, respectively. and These represent the lower and upper limits of hydrogen storage capacity, respectively. and These refer to the maximum allowable hydrogen charging capacity and the maximum allowable hydrogen discharging capacity for hydrogen energy storage, respectively. , These represent the initial and final hydrogen storage capacity of the hydrogen storage tanks during the scheduling cycle.

[0028] (6) Steel production constraints To ensure the industrial park completes its production tasks within the scheduling cycle, the cumulative steel output must meet the planned requirements: (35) In the formula, This refers to the planned steel production within the scheduling cycle.

[0029] (7) Green Steel Certification Constraints While ensuring economic operation, the scheduling model also needs to ensure that steel products meet the requirements for green steel certification. Using carbon emission intensity per ton of steel as a constraint indicator, the average carbon emission level of the steel production process within the scheduling cycle is limited. Therefore, the green steel certification constraint is: (36) In the formula, The maximum carbon emission intensity per ton of steel allowed for green steel certification; The green steel certification constraint means that the average carbon emission level corresponding to the unit steel production within the scheduling cycle shall not exceed the given standard. This constraint works together with the carbon emission cost term in the objective function. The carbon emission cost reflects the economic loss signal under the carbon trading mechanism, while the green steel constraint reflects the certification access threshold requirements. The two have different mechanisms of action but work together.

[0030] Step 3: Construct a constrained Markov decision process model that takes into account green steel constraints; The optimized scheduling model established in the second step is characterized by strong nonlinearity, strong temporal coupling, and multiple constraints in its mathematical structure. Furthermore, the system needs to make decisions under dynamic conditions with multiple uncertainties, such as wind and solar power output, real-time electricity prices, and grid carbon emission intensity. Simultaneously, the green steel certification constraint is a cumulative ratio constraint over its entire lifecycle, and its satisfaction level dynamically evolves over time during the scheduling process. Therefore, the optimized scheduling problem is transformed into a sequential decision-making form that can be handled by deep reinforcement learning. To explicitly handle cumulative constraints within the sequential decision-making framework, this invention models the green steel certification constraint as a cost constraint in a constrained Markov decision process, rather than the soft penalty term of the objective function commonly used in traditional methods. The constrained Markov decision process model includes: 1) State space: The state represents all decision-related information of the system. Status of the time period Defined as: (37) In the formula, for Remaining steel production to be completed during the period; , up to Cumulative carbon emissions and cumulative steel production over a given period; 2) Action Space: An action represents the device control quantity taken by the agent in a given state. Actions during a period of time Defined as: (38) Power purchased by the park Instead of being treated as an independent action variable, it is derived from the internal environment based on the energy bus balance, in order to reduce the dimensionality of the action space and avoid redundant decisions. Therefore: (39) Based on the non-negative requirement for the purchased power capacity The non-negativity constraint should be satisfied; 3) State transition probability State transition probability Indicates the system from its current state Take action Then transition to the next state The probability of state transition is determined by both internal deterministic updates and external random disturbances. The internal deterministic updates are given by the system model established in the first step. The three types of energy storage states are updated by their respective dynamic constraint equations. The equivalent temperature of the vertical shaft furnace reduction zone is updated according to the temperature dynamic equation. The remaining steel production, cumulative steel production, and cumulative carbon emissions are recursively derived using the following formula: (40) (41) In the formula, for Remaining steel production to be completed during the period , up to Cumulative steel production and cumulative carbon emissions over the period. for Instantaneous carbon emission rate over a given period; State transition probability The external random disturbance component mainly consists of uncertain environmental quantities such as theoretical available wind power output, theoretical available photovoltaic power output, time-of-use electricity price, grid carbon emission intensity, and basic electricity, heat, and hydrogen loads in the industrial park. These variables are not directly controlled by the agent's actions and can be sampled based on historical data, prediction error models, or typical operating scenarios. Due to the randomness of external environmental quantities, under the same conditions... and the same action The system may then transition to different next states. .

[0031] 4) Reward Function: This refers to the immediate reward obtained by the agent after performing an action. To achieve the optimization goal of minimizing the overall cost of the park, the immediate reward is defined as the instantaneous negative value of the overall cost. Therefore: (42) In the formula, for Instant rewards for specific time periods; 5) Cost function To achieve time-by-time monitoring of green steel certification constraints within a constrained Markov decision process framework, the dynamic green steel evaluation index is defined as the real-time carbon emission intensity per ton of steel up to the current time period: (43) In the formula, for Dynamic green steel evaluation indicators for different time periods; when ( When (is a very small positive number), let To avoid the division-by-zero problem when the cumulative steel production is zero at the beginning of the scheduling process; Based on the dynamic green steel evaluation index, the immediate cost is defined as: (44) In the formula, for The immediate cost of a given time period; 6) Constraint tolerance and discount factor Constraint tolerance refers to the preset upper limit of the expected cumulative discount cost. In this implementation, the constraint tolerance is set to zero to drive the agent to learn an operating strategy that strictly adheres to the Green Steel certification standard; discount factor It is used to balance the relative importance of immediate rewards and future rewards, and to influence the contribution weight of constraint violations in each time period in the cumulative cost.

[0032] 7) Optimization Objective The goal of a constrained Markov decision process is to find the optimal strategy. To ensure that the strategy maximizes the expected cumulative discount reward while satisfying the expected cumulative discount cost constraint, we have: (45) (46) In the formula, In strategy The expected cumulative discount reward In strategy The expected cumulative discount cost, For the expectation, To constrain tolerance, for Discount factor for time period.

[0033] Step 4: Solve the constrained Markov decision process model using the Lagrange constrained proximal policy optimization algorithm to obtain the executable actions, i.e., the scheduling results; Constrained Markov Decision Process (CPM) models simultaneously involve maximizing cumulative discounted rewards and cost constraints. While the Proximal Policy Optimization (PPO) algorithm exhibits good stability in maximizing cumulative rewards, it struggles to directly handle the cost constraints of CPM models. Therefore, the Lagrange multiplier method is introduced and combined with PPO to obtain the Lagrange Constrained Proximal Policy Optimization (LPO). This algorithm, through alternating policy updates and Lagrange multiplier updates, transforms the problem of solving CPM models into an approximate solution of the following Lagrange dual problem: (47) In the formula, For Lagrange functions, For policy network parameters; The non-negative Lagrange multiplier serves as a penalty coefficient, dynamically balancing the relationship between reward maximization and constraints. When the constraints tend to be violated... Increase the penalty for the policy when the constraint is satisfied. Reduce the size to make the strategy focus more on reward maximization; The policy network of a constrained Markov decision process model takes the state as input and outputs the action probability distribution. The standard Lagrange constrained proximal policy optimization algorithm only updates the policy with a reward-side pruning objective, without addressing the cost constraint. To ensure that the cost constraint directly affects policy updates, this implementation adds a cost-side pruning objective to the reward-side pruning objective, constructing the policy objective function as the difference between the two. Therefore: (48) In the formula, Let the policy objective function be... The reward-side pruning objective is the same as the standard Lagrangian-constrained proximal policy optimization algorithm, which is obtained by taking the minimum value of the probability ratio of the new and old policies and the reward advantage after pruning. The cost-side pruning target is represented as: (49) In the formula, for The ratio of the probability of actions under the old and new strategies during a given time period. for The probability of actions under the new strategy during a given time period. for The probability of actions during a given time period under the old strategy; for The cost advantage of time period is used to measure action. The impact of the average action on the cumulative constraint cost relative to the current state; The clipping factor; A truncation operator that restricts the first parameter to a given interval; Since the cost advantage is explicitly multiplied by the action probability ratio, the cost-side pruning objective maintains a non-zero gradient with respect to the policy network parameters. Differential suppression based on the degree of constraint violation can be applied to each action through policy gradient updates.

[0034] Any aspects not covered in this invention are applicable to existing technologies.

Claims

1. A multi-energy coupling scheduling method for hydrogen-based steel industrial parks considering green steel constraints, characterized in that, Includes the following steps: Step 1: Construct an operational model for the park's equipment, including a refined operational model for vertical shaft furnaces, an operational model for electric arc furnaces, and an operational model for electrolytic cells; The refined operation model of the vertical shaft furnace includes the vertical shaft furnace process model and the vertical shaft furnace thermodynamic model; the vertical shaft furnace process model is represented as follows: （6） （7） （5） （4） In the formula, for The production rate of sponge iron from the vertical shaft furnace during the period. The yield coefficient of sponge iron per unit of net hydrogen consumption in the reduction reaction. for Iron ore consumption rate over a period of time. This represents the iron ore consumption coefficient corresponding to the unit output of sponge iron. for Net hydrogen consumption of the reduction reaction during the period for Hydrogen recovery volume in tail gas during the period for Hydrogen utilization rate of vertical shaft furnace during the period for Total hydrogen supply from vertical shaft furnace during the period This represents the thermodynamic equilibrium upper limit for hydrogen utilization. The effective void volume of the reduction zone in a hydrogen-based vertical shaft furnace. The porosity of the material bed in a hydrogen-based vertical shaft furnace. The density of hydrogen gas under standard conditions. The nominal equivalent reaction characteristic time at the reference temperature. As the apparent activation energy, Let be the ideal gas constant. for Equivalent temperature of the reduction zone in the vertical shaft furnace during the time period For reference temperature; The thermal model of the vertical shaft furnace is represented as follows: （13） In the formula, for Equivalent temperature of the reduction zone in the vertical shaft furnace during the time period The equivalent thermal mass of the material in the reduction zone of the vertical furnace, The average specific heat capacity of the materials in the furnace. For scheduling step size, for The heat output power of the electric heating furnace during the time period for The auxiliary heating power of the time-period regenerator for the vertical furnace. for Total heat demand power of vertical shaft furnace during the time period; Step 2: Based on the park equipment operation model, construct an optimized scheduling model; The objective function for optimizing the scheduling model is as follows: （17） （18） In the formula, The total cost of the park during the scheduling period. for The overall cost of the park during the period for Cost of raw materials purchased during the period; for Electricity purchase cost during specific time periods for Carbon emission costs per period for Thermochemical hydrogen production costs over time for Equipment maintenance costs during specific time periods for The cost of penalties for wind and solar power curtailment during certain periods This represents the total number of time periods in the scheduling cycle. The constraints of the optimized scheduling model include vertical shaft furnace operation constraints, hydrogen production equipment operation constraints, new energy output constraints, energy bus balance constraints, energy storage unit operation constraints, steel production constraints, and green steel certification constraints; among them, the green steel certification constraints are: （36） In the formula, The carbon emission coefficient for 1 ton of crude steel. for Crude steel production rate over a period of time for Carbon emission intensity of electricity purchased from the power grid during a given period for Power purchased by the park during different time periods The carbon emission coefficient for the entire life cycle of purchased hydrogen. for The amount of hydrogen purchased during the period, The carbon emission coefficient for the entire life cycle of purchased iron ore. for Iron ore consumption rate over a period of time. The maximum carbon emission intensity per ton of steel allowed for green steel certification; Step 3: Model the optimization scheduling problem as a constrained Markov decision process that takes into account the green steel constraint; The state is defined as: （37） In the formula, for The state of the time period for Theoretical maximum available power of wind power during a given period for Theoretical maximum available power of photovoltaic power during a given period for Time-of-use electricity pricing For time period Basic electrical load of the park for The basic heat load of the park during the period for Hydrogen storage capacity of the time-limited hydrogen storage tank. for The basic hydrogen load of the park during the period for Time-of-use energy storage capacity for Time-of-use thermal energy storage for storing heat. for Input power of the electrolytic cell during the time period for Remaining steel production to be completed during the period; , up to Cumulative carbon emissions and cumulative steel production over a given period; Action is defined as: （38） In the formula, for Actions during a period of time, for Input electrical power of the electrolytic cell during the time period for Hydrogen production capacity of time-phase thermochemical hydrogen production units for Actual wind power utilization during the period for Actual photovoltaic power utilization during the period for Power consumption of the electric heating furnace during the period for Waste heat distribution ratio of the time-period regenerator for Time-of-use energy storage capacity, for Net power of thermal energy storage during the time period for Net hydrogen exchange volume during the hydrogen storage period; The reward function is: （42） In the formula, for Instant rewards for specific time periods; The cost function is: （43） （44） In the formula, for Dynamic green steel evaluation indicators for different time periods for The immediate cost of a given time period; The optimization objective is to find the optimal strategy that maximizes the expected cumulative discount reward and the expected cumulative discount cost constraints, which are as follows: （45） （46） In the formula, In strategy The expected cumulative discount reward In strategy The expected cumulative discount cost, For the expectation, To constrain tolerance, for Discount factor for the time period; Step 4: The Lagrange constrained near-end policy optimization algorithm is used to solve the constrained Markov decision process model that takes into account the green steel constraint to obtain the executable actions, i.e. the scheduling results.

2. The multi-energy coupling scheduling method for hydrogen-based steel industrial parks considering green steel constraints as described in claim 1, characterized in that, The electric arc furnace operation model is represented as follows: （15） In the formula, for Power consumption of electric arc furnace during specific time periods The electric arc furnace power consumption coefficient is the unit of sponge iron. The electrolytic cell operating model is represented as follows: （16） In the formula, for Hydrogen production per time period of electrolyzer For the efficiency of the electrolytic cell, The lower heating value is the calorific value of hydrogen per unit volume.

3. The multi-energy coupling scheduling method for hydrogen-based steel industrial parks considering green steel constraints according to claim 1 or 2, characterized in that, The cost of the purchased raw materials is expressed as follows: （20） The cost of carbon emissions is expressed as: （21） In the formula, The unit carbon price; The cost of thermochemical hydrogen production is expressed as: （22） In the formula, The fixed costs of the thermochemical hydrogen production unit during the scheduling cycle, The variable cost coefficient per unit of hydrogen production; Equipment maintenance costs are expressed as follows: （23） In the formula, For the collection of equipment in the park; For the first The unit operating volume maintenance cost coefficient for this type of equipment; for Time period Operating volume of this type of equipment; The penalty cost of curtailing wind and solar power is expressed as follows: （24） In the formula, This is the wind curtailment penalty coefficient; This represents the penalty coefficient for discarded light.

Images