Hydrogen energy park scheduling method based on improved cloud drift optimization algorithm

Abstract

The invention discloses a hydrogen energy park scheduling method based on an improved cloud drift optimization algorithm, and relates to the technical field of intelligent development and scheduling of a comprehensive energy system, and the method comprises the following steps: building a target function with the minimum economic cost and carbon processing cost of the park comprehensive energy system as a target; and after wind and light power supply constraints, energy storage equipment constraints, power balance constraints, carbon sequestration capacity constraints, carbon capture constraints, gas turbine output and climbing constraints, gas boiler constraints, fuel cell constraints and power-to-gas constraints are set, an improved cloud drift optimization algorithm is used for solving, and the hydrogen energy park is scheduled. According to the invention, the bottleneck of insufficient regulation capability of a single energy network is broken.

Description

Technical Field

[0001] This invention relates to the field of intelligent development and scheduling technology for integrated energy systems, and more specifically, to a scheduling method for hydrogen energy parks based on an improved cloud drift optimization algorithm. Background Technology

[0002] Currently, in integrated energy systems, hydrogen energy utilization technologies often revolve around a single core function, failing to achieve deep and diversified coupling of hydrogen energy as a multi-energy carrier with the system. The main technological forms include: **Single Energy Storage Utilization:** This is currently the most mainstream single-use model. The system is equipped with an electrolyzer, hydrogen storage tank, and fuel cell, with its core function being to use hydrogen energy as a large-scale electricity storage medium. The specific process involves using surplus renewable energy from wind and solar power within the park to electrolyze water to produce hydrogen, converting the electricity into hydrogen energy for storage; during power shortages, the hydrogen fuel cell generates electricity, feeding it back to the grid. In this process, hydrogen energy primarily plays the role of a "giant battery." **Single Transportation Fuel Utilization:** This model focuses hydrogen energy utilization on the transportation sector. The park's energy system provides hydrogen refueling services for its internal fuel cell buses, logistics vehicles, or passenger cars. Hydrogen production facilities and refueling stations constitute relatively independent infrastructure, narrowing hydrogen energy's role from "energy carrier" to "transportation fuel," with weak interaction with other energy forms within the park, such as electricity, heat, and cooling. Single-use industrial feedstock / fuel: In industrial parks such as chemical and metallurgical plants, hydrogen energy may be used solely as a feedstock or high-temperature heat source in industrial processes. For example, it may be used as a syngas feedstock in chemical production or as a reducing agent to replace coke in steel plants. Although this achieves deep decarbonization, the hydrogen energy system is usually decoupled from the park's power and heating networks, failing to realize its synergistic value in balancing the power grid and providing multi-energy supply.

[0003] The coupling of carbon capture technology with multiple energy flows such as electricity, heat, and gas mainly relies on the existing "segment-by-segment adaptation" technology model, and has not yet formed a deeply integrated and collaborative architecture. Regarding power coupling, carbon capture equipment (such as amine absorption capture devices) mostly adopts a "constant load" operation design, using grid backup power or energy storage systems to cope with the volatility of renewable energy generation. For example, some power plants directly use electricity generated from photovoltaic and wind power for hydrogen production in electrolyzers, combined with... Methane synthesis achieves a preliminary "electricity-gas-carbon" conversion; in thermodynamic coupling, industrial waste heat or power plant exhaust steam is mainly used to provide heat energy for the solvent regeneration process of carbon capture, such as introducing waste heat from boiler tail gas into the amine liquid regeneration tower in coal-fired power plants to reduce additional heat source consumption; in the field of gas coupling, carbon treatment is indirectly linked through hydrogen blending technology in natural gas pipelines, combining green hydrogen production with... The captured "electricity-to-gas" products are mixed into the natural gas pipeline network, achieving a simple connection between the energy carrier and the carbon stream.

[0004] In terms of solution, heuristic algorithms are widely used in integrated energy dispatching in industrial parks due to their adaptability to complex problems. Among them, the Genetic Algorithm (GA) simulates the biological evolution process, iteratively optimizing the population through selection, crossover, and mutation, making it suitable for multi-objective dispatching scenarios. For example, in the dispatching of hydrogen-utilizing industrial parks, GA can use "economic efficiency - carbon emission reduction - equipment wear" as a multi-objective function, and treat electrolyzer output and hydrogen storage tank filling / discharging amounts as chromosome gene segments, selecting the Pareto optimal solution set through multiple generations of evolution. The Ant Colony Optimization (ACO) algorithm, based on the pheromone mechanism of ants foraging, is suitable for dispatching problems involving path optimization, such as the coordinated transmission dispatching of hydrogen pipelines and the power grid in industrial park energy networks. ACO guides the ants (i.e., the dispatching scheme) to select the optimal path through pheromone concentration, gradually constructing a globally optimal energy flow scheme.

[0005] The most critical drawback of relying solely on hydrogen energy is that it ignores hydrogen's essential nature as a multi-energy carrier. The electrolysis of water to produce hydrogen generates a large amount of high-grade heat energy. In a single-use power generation or transportation fuel model, these byproducts are typically directly emitted and discarded, resulting in significant energy and resource waste and failing to achieve the comprehensive energy efficiency improvement of "temperature matching and cascaded utilization." Furthermore, its system flexibility and resilience are limited. A hydrogen energy system used solely for energy storage or transportation has limited ability to cope with fluctuations in the supply and demand of multiple energy sources. When a park experiences peak heat demand while electrical load remains stable, a single energy storage-type hydrogen system cannot respond effectively. Unlike a combined heat and power (CHP) system, it cannot simultaneously meet both electricity and heat demands, thus weakening the multi-energy complementarity and operational resilience that a comprehensive energy system should possess.

[0006] The coupling patterns in current integrated energy systems have significant flaws, the core issue being the lack of deeply optimized synergistic mechanisms and technological support. From a dynamic response perspective, the operational inertia of carbon treatment technologies is mismatched with the volatility of multiple energy flows. For example, the start-up and shutdown response time of electrolyzers can take several minutes to tens of minutes, while wind power fluctuations can occur within seconds, leading to inefficient coordination and even grid load shocks. In thermodynamic coupling, there is a contradiction between the temperature and flow changes of industrial waste heat and the stable heat energy requirements for carbon capture solvent regeneration. When waste heat supply is insufficient, it is necessary to switch to fossil fuel supplementary combustion, which increases carbon emissions. From a system optimization perspective, existing technologies mostly focus on the local coupling of a single energy source and carbon treatment, lacking a system-wide energy-carbon flow synergistic optimization model. For example, in the "electricity-to-gas" process, only electricity consumption is considered, without fully integrating the load demand of the heating network to adjust methane production, resulting in reduced overall energy utilization efficiency.

[0007] Heuristic algorithms demonstrate strong adaptability to complex park scheduling problems, but they generally suffer from drawbacks such as unstable convergence performance, sensitivity to parameters, and a lack of rigorous theoretical guarantees of global optimality. When facing scheduling scenarios involving deep coupling of hydrogen energy and hydrogen blending with natural gas, and real-time coordination of multiple energy flows, a single heuristic algorithm often struggles to balance solution accuracy and efficiency. For example, the GA algorithm suffers from slow convergence, is prone to getting trapped in local optima in the later stages of evolution, and the settings of parameters (such as population size and crossover / mutation probability) significantly affect the solution results, requiring extensive trial and error adjustments. The ACO algorithm has low computational efficiency, the pheromone update rule has a significant impact on the solution speed, and when dealing with continuous decision variables (such as energy storage charging and discharging power), it requires discretization to reduce accuracy, making it difficult to meet the refined requirements of scheduling schemes. Summary of the Invention

[0008] To address the aforementioned issues, the present invention aims to provide a hydrogen energy park scheduling method based on an improved cloud drift optimization algorithm, which addresses the difficulty of existing technologies in meeting the refined requirements of scheduling schemes.

[0009] To achieve the aforementioned technical objectives, this application provides a hydrogen energy park scheduling method based on an improved cloud drift optimization algorithm. The hydrogen energy park comprises wind power, solar power, electricity-to-gas conversion, carbon capture, gas turbines, gas boilers, hydrogen fuel cells, and energy storage equipment for these three energy sources, including batteries, thermal storage tanks, and hydrogen storage tanks. The wind power, solar power, electricity-to-gas conversion, batteries, and hydrogen storage tanks together constitute the hydrogen refueling station within the park, providing hydrogen energy for hydrogen-powered vehicles within the park. The method includes the following steps: The objective function is constructed with the goal of minimizing the economic cost and carbon treatment cost of the integrated energy system in the park. After setting constraints including wind and solar power supply, energy storage equipment, power balance, carbon sequestration capacity limit, carbon capture, gas turbine output and ramping constraints, gas boiler constraints, fuel cell constraints, and electricity-to-gas constraints, the improved cloud drift optimization algorithm is used to solve the problem and schedule the hydrogen energy park.

[0010] Preferably, when constructing the objective function, the objective function is expressed as: ; In the formula, , , , These are carbon trading costs, carbon storage costs, natural gas purchase costs, and wind / solar curtailment costs; among them, Carbon trading costs: ; In the formula, This represents the total carbon dioxide emissions; Indicates carbon quota; Indicates the carbon trading price; Total carbon dioxide emissions The calculation formula is: ; In the formula, This represents the carbon emissions of the gas turbine at time t; This represents the carbon emissions of the gas-fired boiler at time t; The cost of carbon sequestration is: ; In the formula: For the unit quality of sealing The cost; This represents the power of the carbon capture device at time t; The annual cost of purchasing natural gas is: ; In the formula: For natural gas prices; These represent the natural gas consumption of the gas turbine and the gas boiler at time t, respectively. The penalty cost for curtailing wind and solar power is: ; In the formula: These represent the wind curtailment price and the solar curtailment price at time t, respectively. These are the predicted power output values ​​for wind power and solar power, respectively. These represent the actual wind power generation and photovoltaic power generation at time t, respectively.

[0011] Preferably, when setting constraints, the wind-solar power supply constraints consist of wind power constraints and solar power constraints, wherein the wind power constraints are: ; Photoelectric constraint is: ; In the formula: These represent the amount of wind and solar power curtailment at time t, respectively. These represent the predicted wind power and solar power output at time t, respectively. Let represent the power output of wind power and photovoltaic power at time t, respectively.

[0012] Preferably, when setting constraints, the power balance constraint is represented as follows: ; In the formula: This represents the electrical power consumed and output by the battery at time t, as well as the electrical load demand of the system. This represents the heat power absorbed and output by the thermal storage tank at time t, as well as the system's heat load demand. This represents the volume of natural gas purchased by the system at time t; This represents the volume of natural gas output from the methane reactor at time t; This represents the volume of natural gas consumed by the gas turbine and gas boiler at time t. This represents the hydrogen power absorbed and released by the hydrogen storage tank at time t, as well as the system's hydrogen load requirement.

[0013] Preferably, when setting constraints, the carbon sequestration capacity limit is expressed as follows: ; In the formula: They represent the values ​​at time t, respectively. The amount of storage and the system Maximum storage capacity; Carbon capture constraints are expressed as follows: CCS operating energy consumption Do not exceed maximum operating power : ; Gas turbine output and ramp constraint representation: ; In the formula: For the total output of the gas turbine; These represent the electrical power and thermal power output by the gas turbine at time t, respectively. These are the upper and lower limits of the electrical output of the gas turbine; These are the upper and lower limits of the thermal output of the gas turbine; The upper and lower limits of the gas turbine's ramp output; Constraints for gas-fired boilers are indicated as follows: ; In the formula, These are the upper and lower limits of the heat output of the gas-fired boiler; The upper and lower limits of the ramp output of the gas-fired boiler; Fuel cell constraints are expressed as follows: ; In the formula, This is the upper limit of fuel cell output; Electro-pneumatic constraint representation: ; In the formula, the operating energy consumption of P2G is... Within the scope ; These represent the upper and lower limits of P2H's ramp output.

[0014] Preferably, when obtaining the improved cloud drift optimization algorithm, the improved cloud drift optimization algorithm is obtained by setting an adaptive parameter control mechanism, a weight update mechanism, a position update strategy, a diverse maintenance mechanism, and convergence monitoring and reinforcement search based on the cloud drift optimization algorithm CDO.

[0015] Preferably, when setting the adaptive parameter control mechanism, the adaptive parameter control mechanism is set by nonlinear parameter decay and dynamic balance adjustment, so that the algorithm model achieves the best balance between global search and local refinement.

[0016] Preferably, when setting the weight update mechanism, a differentiated search strategy allocation is implemented based on individual performance differences. A hierarchical processing strategy is adopted to divide the population into elite individuals, average individuals, and lagging individuals: the top 30% of elite individuals are given a larger weight to enhance their development capabilities and accelerate convergence to the optimal solution; the middle 40% of average individuals are given a balanced strategy to take into account both deterministic updates and random perturbations; and the weight of the bottom 30% of lagging individuals is reduced to encourage exploratory search to maintain population diversity. Furthermore, the base weight wbase dynamically changes over time to achieve adaptive adjustment of the strategy.

[0017] Preferably, when setting the location update strategy, the location update strategy is set by incorporating multi-source information fusion and time decay mechanisms, so as to achieve a more comprehensive search coverage of the algorithm model.

[0018] Preferably, when setting up a diversity maintenance mechanism, the core is to actively maintain population diversity and prevent premature convergence. The diversity maintenance mechanism is achieved through dynamic subgroup division, subgroup information exchange, diversity monitoring, and active restart mechanism: it is used to automatically determine the number of subgroups based on the population distribution to achieve an adaptive population structure; it periodically exchanges information between different subgroups to promote the spread and recombination of superior patterns; it calculates the population diversity index in real time to quantify the exploration status, and automatically restarts some individuals to re-inject diversity when the diversity is below the threshold. At the same time, it achieves distributed coverage of the search space through the subgroup mechanism to improve the global search capability.

[0019] The present invention discloses the following technical effects: (1) Off-grid systems can fully utilize the dispersed renewable energy sources within the region, combining energy production and consumption locally, avoiding line losses during long-distance power transmission, and improving energy efficiency. For regions with abundant local energy resources, the system can maximize the absorption of local clean energy, reduce dependence on external fossil fuels, and achieve green and efficient energy utilization. In remote mountainous areas, islands, border outposts, and other areas where large power grids are difficult to extend their coverage, off-grid integrated energy systems can achieve energy self-sufficiency by relying on local wind and solar resources, small hydropower, etc. At the same time, the system is not affected by external factors such as main grid failures and natural disasters. Through energy storage devices and multi-energy complementary configurations, it can ensure continuous and stable power supply to critical loads, significantly improving the resilience and risk resistance of energy supply. Gas grid connection promotes a significant improvement in comprehensive energy efficiency. Natural gas generates electricity and heat in gas turbine units, and the high-temperature waste heat generated during power generation is used for heating, realizing "electricity-heat" cogeneration, with comprehensive energy efficiency far exceeding that of traditional coal-fired units; at the same time, the system can dynamically allocate the flow of natural gas between different equipment according to the diverse energy needs of users, such as electricity and heat.

[0020] (2) The economic and environmental benefits of the model were verified by comparing five differentiated scenarios. Its core advantage lies in its ability to comprehensively, accurately, and specifically explore the impact of key factors, providing solid support for model optimization: By controlling single variables (such as whether carbon capture is coupled, whether the methane reactor is activated, whether hydrogen is added, and differences in carbon dioxide sources), the core components (carbon capture, methane reactor) and key parameters (hydrogen addition, etc.) can be clearly identified. The source of the data has an independent impact on economic and environmental efficiency, avoiding ambiguity in conclusions caused by the confounding of multiple factors; at the same time, the scenarios cover "coupled and uncoupled carbon capture"; "hydrogen-doped and non-hydrogen-doped"; "methanation and non-methanation"; Typical operating conditions such as "self-sufficiency" and "external purchase" can verify the model's adaptability to different practical application scenarios and clarify the model's advantages and applicable scope. Furthermore, by comparing "with or without carbon capture"; "Emissions / Buying Difference" can quantify the environmental benefits of the model (such as carbon reduction effects), combined with cost differences under different component configurations (such as buying externally). (Cost, methane reactor operating cost), can accurately analyze the trade-off between economic efficiency and environmental protection, providing a basis for optimal solution decision-making.

[0021] (3) The CDO algorithm accurately locates the optimal solution through adaptive parameter control and other mechanisms; avoids premature convergence by relying on diversity preservation and random reset strategies; the parameters can be automatically adjusted to adapt to the scenario and parameter changes; it meets the real-time scheduling requirements by leveraging elite guidance and experience learning; it reduces the sensitivity of initial values, and can find a satisfactory solution even if the initial scheme is not good; it has convergence theoretical proof, and the application reliability is guaranteed; it integrates multiple strategies to handle multi-objective and multi-constraint relationships, and maintains population diversity through subgroup division and information exchange to avoid local optima; it triggers reinforcement search when convergence stalls, further improving the solution quality; at the same time, through dimension-independent operation and adaptive weights, it can effectively deal with high-dimensional scheduling problems and fully adapt to the demands of actual scenarios. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a schematic diagram of the method described in this invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0025] like Figure 1 As shown, this invention provides a hydrogen energy park scheduling technology based on an improved cloud drift optimization algorithm, specifically including the following: This invention constructs an off-grid integrated energy system for industrial parks, which includes wind power, solar power, electricity-to-gas conversion, carbon capture, gas turbines, gas boilers, hydrogen fuel cells, and energy storage devices for all three energy sources (batteries, thermal storage tanks, and hydrogen storage tanks). The wind power, solar power, electricity-to-gas conversion, batteries, and hydrogen storage tanks together constitute the hydrogen refueling station within the park, providing hydrogen energy for hydrogen-powered vehicles within the park.

[0026] 1. Define optimization goals: This invention constructs an objective function with the goal of minimizing the economic cost and carbon treatment cost of the integrated energy system in the industrial park: ; In the formula, , , , These are carbon trading costs, carbon storage costs, natural gas purchase costs, and wind / solar curtailment costs.

[0027] (1) Carbon trading costs: ; In the formula, This represents the total carbon dioxide emissions; Indicates carbon quota; This indicates the price of carbon trading.

[0028] Total carbon dioxide emissions The calculation formula is: ; In the formula, This represents the carbon emissions of the gas turbine at time t; This represents the carbon emissions of the gas-fired boiler at time t.

[0029] (2) The cost of carbon sequestration is: ; In the formula: For the unit quality of sealing The cost; This represents the power of the carbon capture device at time t; (3) The annual natural gas purchase cost is: ; In the formula: For natural gas prices; These represent the natural gas consumption of the gas turbine and the gas boiler at time t, respectively.

[0030] (4) The penalty cost for wind and solar power curtailment is: ; In the formula: These represent the wind curtailment price and the solar curtailment price at time t, respectively. These are the predicted power output values ​​for wind power and solar power, respectively. These represent the actual wind power generation and photovoltaic power generation at time t, respectively.

[0031] 2. Construct constraints: (1) Constraints of wind and solar power supply Wind power constraints: ; Photoelectric confinement: ; In the formula: These represent the amount of wind and solar power curtailment at time t, respectively. These represent the predicted wind power and solar power output at time t, respectively. Let represent the wind power and photovoltaic power output at time t, respectively; (2) Constraints of energy storage devices: To prevent energy storage devices from charging and discharging simultaneously, a unique identifier for the charging / discharging status must be set. Taking a hydrogen storage tank as an example, the following constraints are set; the same applies to thermal storage tanks and batteries, as detailed below: ; To extend the lifespan of energy storage devices, a State of Charge (SOC) constraint needs to be added, which is the state of charge of the hydrogen storage tank at time t. All must be in Within the specified range, and to prevent simultaneous charging and discharging of hydrogen storage tanks, charging and discharging status indicators must be installed. The value can be 0 or 1; at the same time, the state of charge of the hydrogen storage tank at time 0 must be consistent with the state of charge at the last time.

[0032] (3) Power balance constraints: To ensure supply and demand balance within the system and on the load side, the following constraints are imposed on the four energy flows (electricity, heat, gas, and hydrogen) within the system: ; In the formula: This represents the electrical power consumed and output by the battery at time t, as well as the electrical load demand of the system. This represents the heat power absorbed and output by the thermal storage tank at time t, as well as the system's heat load demand. This represents the volume of natural gas purchased by the system at time t; This represents the volume of natural gas output from the methane reactor at time t; This represents the volume of natural gas consumed by the gas turbine and gas boiler at time t. This represents the hydrogen absorption and release power of the hydrogen storage tank at time t, as well as the system's hydrogen load requirement. (4) Carbon sequestration capacity limitations ; In the formula: They represent the values ​​at time t, respectively. The amount of storage and the system Maximum storage capacity; (5) Carbon capture constraints: CCS operating energy consumption Do not exceed maximum operating power : ; (6) Gas turbine output and ramping constraints: ; In the formula: For the total output of the gas turbine; These represent the electrical power and thermal power output by the gas turbine at time t, respectively. These are the upper and lower limits of the electrical output of the gas turbine; These are the upper and lower limits of the thermal output of the gas turbine; The upper and lower limits of the gas turbine's ramp output; (7) Constraints on gas-fired boilers: ; In the formula, These are the upper and lower limits of the heat output of the gas-fired boiler; The upper and lower limits of the ramp output of the gas-fired boiler; (8) Constraints of fuel cells: ; In the formula, This is the upper limit of fuel cell output; (9) Electro-pneumatic confinement: ; In the formula, the operating energy consumption of P2G is... Within the scope ; These represent the upper and lower limits of P2H's ramp output.

[0033] 3. Improvements to the Cloud Drift Optimization Algorithm (CDO): This invention makes several improvements to the Cloud Drift Optimization (CDO) algorithm, and the specific improvement steps are as follows: (1) Adaptive parameter control mechanism: To achieve a smooth transition from global exploration to local development in the algorithm, an adaptive parameter control mechanism is designed to achieve the best balance between global search and local refinement. Its specific working mechanism includes nonlinear parameter decay and dynamic balance adjustment: nonlinear changes are achieved through parameters such as γ, β, and α to replace simple linear decay in order to conform to the natural law of the optimization process. In the early stage of the algorithm, the exploration ability is strengthened (larger a, b, z values), and in the later stage, the development ability is strengthened (smaller a, b, z values). At the same time, the parameters are automatically adjusted according to the iteration progress, without the need for manual setting of fixed values. This not only improves the adaptability of the algorithm, but also prevents premature convergence by maintaining appropriate exploration parameters.

[0034] ; In the above formula: β is the nonlinear decay exponent, which controls the curvature of the parameter decay; β is the exponential decay coefficient, which controls the decay rate of the exploration capability. The hyperbolic tangent adjustment coefficient controls the smoothness of the transition in the random reset probability. These represent the upper and lower bounds of the development parameters, respectively. Explore the initial maximum value of the parameter; Explore the initial maximum value of the parameters; t represents the current iteration number; T represents the maximum number of iterations; Represents control parameters during the development phase; Represents control parameters during the exploration phase; This represents the probability of a random reset.

[0035] (2) Weight update mechanism: The weight update mechanism focuses on allocating differentiated search strategies based on individual performance differences, and adopts a hierarchical processing strategy to divide the population into elite individuals ( ≤0.3), medium-sized individuals (0.3 < ≤0.7) and lagging individuals ( >0.7) Three levels: The top 30% of elite individuals are given higher weights to enhance their development capabilities and accelerate convergence to the optimal solution; a balancing strategy is adopted for the middle 40% of individuals, taking into account both deterministic updates and random perturbations; the bottom 30% of lagging individuals have lower weights to encourage exploratory search to maintain population diversity, and the base weights are... The strategy adapts dynamically over time, improving convergence while maintaining population diversity.

[0036] ; In the formula: This represents the current optimal fitness value; represents the fitness value of the i-th individual; S represents the fitness range; It is the numerical stability constant; represents the standardized rank of the i-th individual in the population; i is the index of the individual in the sorted population; N represents the population size and the total number of candidate solutions. The fundamental weight of elite individuals oscillates over time; This represents the basic weight of an average individual, which changes periodically. The base weight of lagging individuals, exhibiting exponential decay; It serves as a moderating factor for the weight of elite individuals; The variable is a random disturbance and follows a standard normal distribution N(0,1).

[0037] (3) Position update strategy: The position update strategy achieves more comprehensive search coverage through a combination of multiple strategies. The specific mechanism includes multi-source information fusion and time decay mechanism: 70% of the development stage adopts deterministic search guided by elite individuals to make full use of the current best information; 30% of the development stage introduces historical best solution information to avoid over-reliance on the current best solution and premature convergence; the exploration stage adopts a nonlinear random generation strategy to ensure extensive coverage of the search space, while combining multiple information sources such as the current best solution, historical experience, and random perturbation, and making the exploration intensity decay exponentially with time; thus achieving the effect of maintaining the convergence speed while significantly improving the probability of finding the global optimum.

[0038] ; Where: xij represents the position value of the i-th individual in the j-th dimension; The position value of the current optimal solution in the j-th dimension; The position values ​​of two randomly selected individuals A and B in the j-th dimension; The probability of dynamic development shifts, with an early stage leaning towards exploration and a later stage leaning towards development. Learning factors guided for elites; The learning rate based on historical experience; For random disturbance coefficients; Shape parameters for exploring diversity control the skewness of random distributions; represents a random number that follows a standard normal distribution; U represents a uniformly distributed random number. The value of the historical best solution in the j-th dimension; These are the upper and lower bounds of the j-th dimension search space, respectively.

[0039] (4) Diversity maintenance mechanism: The diversity maintenance mechanism is centered on actively maintaining population diversity and preventing premature convergence. It is achieved through dynamic subgroup division, subgroup information exchange, diversity monitoring and active restart mechanism: the number of subgroups is automatically determined according to the population distribution to achieve adaptive population structure; information is exchanged between different subgroups regularly to promote the spread and recombination of superior patterns; the population diversity index is calculated in real time to quantify the exploration status, and when the diversity is lower than the threshold, some individuals are automatically restarted to re-inject diversity. At the same time, the subgroup mechanism realizes distributed coverage of the search space and improves the global search capability.

[0040] ; In the formula: C represents the set of subgroups, which divides the population into k clusters; k is adaptive to the number of subgroups, with a minimum of 2 and a maximum of N / 5; Subgroup Locally optimal individuals within ρ subgroups; the stochastic intensity of information exchange between ρ subgroups; Information exchange occurs once every 10 generations.

[0041] ; In the formula, This represents an index of population diversity, and calculates the average of the standard deviations of each dimension. represents the standard deviation of all individual values ​​in the j-th dimension; dim is the problem dimension, i.e., the number of decision variables.

[0042] (5) Convergence monitoring and reinforcement search: Convergence monitoring and reinforcement search aim to detect convergence stagnation and trigger targeted improvement measures. Its mechanism includes stagnation detection, intelligent diagnosis and adaptive search: the convergence status is monitored by the improvement of the optimal solution over multiple generations. When stagnation is detected, it is automatically determined that reinforcement search is needed. A small-scale fine search is carried out around the current optimal solution to improve the quality of the solution. At the same time, some individuals are reinitialized in a guided manner to introduce new changes while maintaining good genes. The search range decreases over time to ensure the accuracy of the later search.

[0043] ; In the formula: Let represent the optimal solution at time t; This represents the optimal fitness value 50 generations ago. Strengthen the search radius, which decreases over time; This represents a random vector that follows a multidimensional standard normal distribution.

[0044] 4. Model Validation: To verify the economic and environmental feasibility of the model, the following five scenarios were established: Scenario 1 is a coupled system of power-to-gas conversion and carbon capture, with a fixed hydrogen blending ratio; Scenario 2 does not consider a methane reactor, and produces hydrogen for hydrogen-powered vehicles and for blending hydrogen into fuel gas, but not for methane production, with a fixed hydrogen blending ratio; Scenario 3 does not consider carbon capture, and emitted carbon dioxide is directly released into the atmosphere, while the carbon dioxide from methanation is purchased externally, with a fixed hydrogen blending ratio; Scenario 4 is a coupled system of power-to-gas conversion and carbon capture, without hydrogen blending.

[0045] The improved algorithm results show that the total cost of Scenario 2 is higher than that of Scenario 1. The core reason is that Scenario 2 eliminates the methane reactor, reducing the electricity required for hydrogen production, leading to a decrease in wind and solar power consumption and an increase in the cost of curtailment. Simultaneously, methane must be purchased entirely externally, increasing gas purchase costs and weakening the park's carbon cycle capability. Scenario 3 has a significantly higher total cost than Scenario 1, primarily because it lacks a carbon capture system, resulting in direct carbon dioxide emissions into the atmosphere. The park needs to purchase a large amount of additional carbon emission credits to meet compliance requirements. Scenario 1 includes hydrogen blending, while Scenario 4 does not, hence the higher total cost of Scenario 4. The core reason is that the addition of hydrogen blending directly increases the energy consumption of the power-to-gas conversion equipment, while simultaneously requiring increased output from the low-carbon emission gas turbine to meet the system's electricity demand. Furthermore, the gas turbine generates heat during power generation, which indirectly reduces the operating output of the gas boiler, thereby reducing the overall carbon emissions of the park. In conclusion, Scenario 1 can improve the economic efficiency and environmental friendliness of the park's operation.

[0046] 5. Algorithm Evaluation The improved cloud drift optimization algorithm was compared with the genetic algorithm, particle swarm optimization algorithm, and the unimproved cloud drift optimization algorithm using the controlled variable method in the following two categories of tests: 1) To evaluate the adaptability and robustness of the four optimization algorithms, this invention conducts sensitivity analyses for different population sizes (N=60, 100, 150, 200, 250, 300). To ensure fairness in the comparison, the analyses are performed under the same test function, changing only the single variable of population size. The core parameters of all algorithms are kept consistent, with a maximum number of iterations set to 1000. Specifically, the crossover rate of the genetic algorithm is set to 0.75 to support robust genetic recombination, and the mutation rate is set to 0.01 to maintain population diversity and avoid premature convergence; the maximum particle velocity of the particle swarm optimization algorithm is set to 5, and both the cognitive and social coefficients are set to 0.5.

[0047] 2) This invention selects four benchmark functions from the CEC2005 test suite, which is a widely recognized standard in the field of optimization algorithm evaluation. These benchmark functions can simulate various real-world challenges such as random disturbances, multimodal landscapes, flat regions, and complex constraints. The table below shows the four selected test functions. The specific implementation involves maintaining a consistent population size and conducting tests under the four selected test functions with different structures, as shown in Table 1. Table 1 Table 2 shows the results of the four algorithms after 1000 iterations, under the conditions of N=60, 100, 150, 200, 250, and 300: Table 2 As shown in Table 3, the following data was obtained from the tests conducted using the four test functions: Table 3 Clearly, under the improved method proposed in this invention, the algorithm has the characteristics of higher convergence accuracy, stronger global search capability, better adaptive capability, faster convergence speed, and stronger robustness.

[0048] This invention establishes an off-grid integrated energy system, whose core advantages lie in the autonomy and comprehensive energy efficiency of energy supply. It eliminates reliance on the public power grid, integrating renewable energy sources such as photovoltaics and wind power with energy storage systems to achieve multi-energy complementary supply. This effectively avoids the risks of energy interruptions caused by grid outages, power rationing, or faults, and allows for flexible allocation of different energy sources according to load characteristics. It provides continuous and stable comprehensive energy security for core production, living, and critical facilities on the user side, making it particularly suitable for scenarios with high requirements for energy reliability and diversity. It allows for the rapid construction of an energy supply system without large-scale grid deployment, significantly reducing upfront infrastructure investment and time costs. Furthermore, the system is primarily powered by clean energy, resulting in extremely low carbon emissions during energy production and conversion. In addition, the system capacity can be customized and flexibly expanded according to changes in user energy demand, and energy dispatch and management can be precisely controlled through an intelligent platform, further improving resource utilization efficiency and operational convenience.

[0049] The system described in this invention adopts an off-grid power grid and a gas grid-connected system architecture. This distributed energy system, independent of the main power grid and enabling autonomous energy production and allocation, exhibits unique advantages in energy supply flexibility and scenario adaptability by integrating local renewable energy, energy storage devices, and diverse energy-consuming facilities. The gas grid connection provides a flexible energy regulation platform for the integrated energy system. When electricity load fluctuates between peak and off-peak times or renewable energy output is unstable, fuel cells, gas boilers, gas turbines, and batteries can quickly start and stop to adjust output, smoothing out power system fluctuations. Simultaneously, through electricity-to-gas conversion, surplus electricity can be converted into hydrogen, and then through methanation, natural gas can be used to generate electricity and heat through gas turbine units, while also reducing carbon emissions. This achieves bidirectional "electricity-gas" energy conversion, significantly improving the system's ability to absorb new energy sources such as wind and solar power, and overcoming the bottleneck of insufficient regulation capacity of a single energy network.

[0050] The improved CDO algorithm demonstrates several significant advantages in solving the integrated energy system scheduling problem. Its most prominent improvement lies in the algorithm's convergence accuracy, which has increased from the original... Upgraded to This significant improvement in scale enables the algorithm to generate ultra-precise scheduling schemes. Economically, it accurately calculates the marginal cost of coordinating multiple energy sources (electricity, heat, and gas), maximizing economic benefits. Reliably, it strictly meets the operational constraints of various equipment, completely avoiding constraint violations caused by calculation errors. In terms of global search performance, the improved algorithm effectively overcomes the tendency of traditional optimization algorithms to get trapped in local optima by introducing a diversity preservation mechanism. Integrated energy system scheduling is essentially a multimodal optimization problem, with local optima existing in economic and environmental scheduling. The improved CDO algorithm can simultaneously discover multiple feasible schemes, including combined electricity and heat optimization, coordinated gas-electricity scheduling, multi-energy complementarity, and optimal charging and discharging strategies for energy storage, providing dispatchers with a wealth of decision-making options. The algorithm's adaptive adjustment capability makes it particularly suitable for handling uncertainties in integrated energy systems. When faced with load fluctuations, equipment failures, or random changes in new energy output, the adaptive parameter mechanism can quickly respond to changes in system state and re-optimize the scheduling scheme. This adaptive characteristic allows the algorithm to quickly converge to meet real-time requirements in short-term scheduling. In multi-energy coordinated optimization, the improved algorithm demonstrates superior ability to handle complex coupling relationships. It can precisely optimize the system's operating status and coordinate the interaction between various units. Through strategies such as energy cascade utilization optimization and minimizing wind and solar curtailment, the overall energy efficiency of the system is improved. Enhanced multi-objective optimization capabilities are another important improvement. The algorithm can simultaneously coordinate economic objectives (minimizing operating costs) and environmental objectives (minimizing carbon emissions), laying a solid foundation for multi-objective trade-off analysis.

[0051] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0052] In the description of this invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0053] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A hydrogen energy park scheduling method based on an improved cloud drift optimization algorithm, wherein the hydrogen energy park consists of wind power, photovoltaic power, electricity-to-gas conversion, carbon capture, gas turbines, gas boilers, hydrogen fuel cells, and energy storage equipment for the three energy sources, including batteries, thermal storage tanks, and hydrogen storage tanks; wherein the wind power, photovoltaic power, electricity-to-gas conversion, batteries, and hydrogen storage tanks together constitute the hydrogen refueling station within the park, providing hydrogen energy for hydrogen-powered vehicles in the park; characterized in that, The method includes the following steps: The objective function is constructed with the goal of minimizing the economic cost and carbon treatment cost of the integrated energy system in the park. After setting constraints including wind and solar power supply, energy storage equipment, power balance, carbon sequestration capacity limit, carbon capture, gas turbine output and ramping constraints, gas boiler constraints, fuel cell constraints, and electricity-to-gas constraints, the improved cloud drift optimization algorithm is used to solve the problem and schedule the hydrogen energy park.

2. The hydrogen energy park scheduling method based on an improved cloud drift optimization algorithm according to claim 1, characterized in that: When constructing the objective function, the objective function is expressed as follows: ； In the formula, , , , These are carbon trading costs, carbon storage costs, natural gas purchase costs, and wind / solar curtailment costs; among them, Carbon trading costs: ； In the formula, This represents the total carbon dioxide emissions; Indicates carbon quota; Indicates the carbon trading price; Total carbon dioxide emissions The calculation formula is: ； In the formula, This represents the carbon emissions of the gas turbine at time t; This represents the carbon emissions of the gas-fired boiler at time t; The cost of carbon sequestration is: ； In the formula: For the unit quality of sealing The cost; This represents the power of the carbon capture device at time t; The annual cost of purchasing natural gas is: ； In the formula: For natural gas prices; These represent the natural gas consumption of the gas turbine and the gas boiler at time t, respectively. The penalty costs for curtailing wind and solar power are: ； In the formula: These represent the wind curtailment price and the solar curtailment price at time t, respectively. These are the predicted power output values ​​for wind power and solar power, respectively. These represent the actual wind power generation and photovoltaic power generation at time t, respectively.

3. The hydrogen energy park scheduling method based on an improved cloud drift optimization algorithm according to claim 2, characterized in that: When setting constraints, the wind-solar power supply constraints consist of wind power constraints and solar power constraints, wherein the wind power constraints are: ； The photoelectric constraint is: ； In the formula: These represent the amount of wind and solar power curtailment at time t, respectively. These represent the predicted wind power and solar power output at time t, respectively. Let represent the power output of wind power and photovoltaic power at time t, respectively.

4. The hydrogen energy park scheduling method based on an improved cloud drift optimization algorithm according to claim 3, characterized in that: When setting constraints, the power balance constraint means: ； In the formula: This represents the electrical power consumed and output by the battery at time t, as well as the electrical load demand of the system. This represents the heat power absorbed and output by the thermal storage tank at time t, as well as the system's heat load demand. This represents the volume of natural gas purchased by the system at time t; This represents the volume of natural gas output from the methane reactor at time t; This represents the volume of natural gas consumed by the gas turbine and gas boiler at time t. This represents the hydrogen power absorbed and released by the hydrogen storage tank at time t, as well as the system's hydrogen load requirement.

5. The hydrogen energy park scheduling method based on an improved cloud drift optimization algorithm according to claim 4, characterized in that: When setting constraints, the carbon sequestration capacity limit is expressed as follows: ； In the formula: They represent the values ​​at time t, respectively. Storage volume and system Maximum storage capacity; The carbon capture constraint means: CCS operating energy consumption The maximum operating power must not be exceeded. : ； The gas turbine output and ramping constraint are expressed as follows: ； In the formula: For the total output of the gas turbine; These represent the electrical power and thermal power output by the gas turbine at time t, respectively. These are the upper and lower limits of the electrical output of the gas turbine; These are the upper and lower limits of the thermal output of the gas turbine; The upper and lower limits of the gas turbine's ramp output; The constraints of the gas-fired boiler are as follows: ； In the formula, These are the upper and lower limits of the heat output of the gas-fired boiler; The upper and lower limits of the ramp output of the gas-fired boiler; The fuel cell constraints are expressed as follows: ； In the formula, This is the upper limit of fuel cell output; The electro-pneumatic constraint means: ； In the formula, the operating energy consumption of P2G is... Within the scope ; These represent the upper and lower limits of P2H's ramp output.

6. The hydrogen energy park scheduling method based on an improved cloud drift optimization algorithm according to claim 5, characterized in that: When obtaining the improved cloud drift optimization algorithm, the cloud drift optimization algorithm CDO is improved by setting an adaptive parameter control mechanism, a weight update mechanism, a position update strategy, a diverse preservation mechanism, and convergence monitoring and reinforcement search.

7. The hydrogen energy park scheduling method based on an improved cloud drift optimization algorithm according to claim 6, characterized in that: When setting up the adaptive parameter control mechanism, the adaptive parameter control mechanism is set up by nonlinear parameter decay and dynamic balance adjustment, so that the algorithm model can achieve the best balance between global search and local refinement.

8. The hydrogen energy park scheduling method based on an improved cloud drift optimization algorithm according to claim 7, characterized in that: When setting up the weight update mechanism, a differentiated search strategy is allocated based on the individual performance differences. A hierarchical processing strategy is adopted to divide the population into elite individuals, medium individuals, and lagging individuals: the top 30% of elite individuals are given greater weight to enhance their development capabilities and accelerate convergence to the optimal solution; a balancing strategy is adopted for the middle 40% of medium individuals to take into account both deterministic updates and random perturbations. The weights of the bottom 30% of individuals are reduced to encourage exploratory searches to maintain population diversity, and the base weights are dynamically changed over time to achieve adaptive adjustment of the strategy.

9. A hydrogen energy park scheduling method based on an improved cloud drift optimization algorithm according to claim 8, characterized in that: When setting the location update strategy, the strategy is implemented by incorporating multi-source information fusion and time decay mechanisms to achieve more comprehensive search coverage of the algorithm model.

10. A hydrogen energy park scheduling method based on an improved cloud drift optimization algorithm according to claim 9, characterized in that: When setting up a diversity maintenance mechanism, the core is to actively maintain population diversity and prevent premature convergence. The diversity maintenance mechanism is achieved through dynamic subgroup division, subgroup information exchange, diversity monitoring and active restart mechanism: it is used to automatically determine the number of subgroups according to the population distribution to achieve adaptive population structure; it exchanges information between different subgroups regularly to promote the spread and recombination of superior patterns; it calculates the population diversity index in real time to quantify the exploration status, and automatically restarts some individuals to re-inject diversity when the diversity is lower than the threshold. At the same time, the subgroup mechanism realizes distributed coverage of the search space and improves the global search capability.

Images