An energy island wind-solar-hydrogen storage system scheduling method based on adaptive variable step size

Abstract

The application discloses a kind of energy island wind light hydrogen storage system scheduling method based on adaptive variable step length, comprising: obtaining wind power, photovoltaic output and load prediction value, combined with deviation prediction model, obtain wind power, photovoltaic output and load actual value;Multiple initial scenes are generated using Monte Carlo simulation;Multiple initial scenes are screened using synchronous back substitution scene reduction technology, and the probability of selecting typical scene and each typical scene that meets the overall distribution of random variable;Based on typical scene and the probability of each typical scene, the rolling optimization method of adaptive variable step length SMPC is used to solve the energy island wind light hydrogen storage system scheduling optimization model, and the energy island wind light hydrogen storage system scheduling result of current period is obtained.The application considers the uncertainty of wind and light, and simultaneously feedbacks the prediction deviation of prediction model, so as to effectively cope with the volatility and uncertainty of wind, light and load, improve the scheduling accuracy and speed.

Description

Technical Field

[0001] This invention belongs to the field of wind farm automatic control, specifically relating to a scheduling method for a wind-solar-hydrogen storage system in an energy island based on adaptive variable step size. Background Technology

[0002] Multi-energy complementarity of wind, solar, hydrogen, and storage leverages the significant short-, medium-, and long-term complementary characteristics of wind and solar energy, relies on the energy regulation capabilities of hydrogen and hydro storage, and coordinates with a considerable scale of wind and solar new energy power generation projects. Through capacity configuration, dispatching, and unit control, it achieves joint complementarity, thereby realizing the efficient utilization of clean energy sources such as hydropower, wind, and solar power. Therefore, research on the optimized dispatching and operation of large-scale offshore energy islands with wind, solar, hydrogen, and storage systems has great development potential. The key components of wind-solar complementary power generation coupled with hydrogen energy storage systems are technically relatively mature. By optimizing capacity configuration design and system control, system costs can be reduced and energy conversion efficiency improved, thus achieving efficient energy utilization. Therefore, research on the coordinated dispatching and operation of offshore energy islands with wind, solar, hydrogen, and storage systems is of great significance for promoting the sustainable development of the renewable energy industry.

[0003] Among numerous control methods, Model Predictive Control (MPC) stands out for its ability to handle delays, multivariables, and constrained conditions. Furthermore, its rolling optimization and feedback correction techniques enhance disturbance suppression capabilities, making it one of the most important methods in engineering control. The main principles of MPC include model prediction, rolling optimization, and feedback correction. MPC excels at handling uncertainties, nonlinearities, and large inertia in the problem set. Currently, MPC is widely used in the scheduling research of energy complementary systems.

[0004] However, stochastic model predictive control is currently mainly applied to cooling, heating and power complementary systems, and its advantages have not been fully utilized. It has not yet been applied to the field of multi-energy complementary power generation of offshore wind, solar and hydrogen storage. Furthermore, SMPC has problems such as the impact of disturbances at the end of the optimization cycle and constant optimization step size in the rolling optimization stage, which affect the prediction accuracy and computational efficiency.

[0005] Therefore, it is necessary to further optimize the SMPC rolling optimization process and study how to improve the prediction accuracy of the SMPC method in order to improve the stable operation of the offshore energy island wind, solar and hydrogen storage system. Summary of the Invention

[0006] Purpose of the invention: In view of the problems and deficiencies of the existing technology, the purpose of this invention is to provide a scheduling method for wind, solar and hydrogen storage systems in energy islands based on adaptive variable step size. By considering the uncertainty of wind and solar power, and simultaneously correcting the prediction deviation of the prediction model, this method can effectively cope with the fluctuation and uncertainty of wind, solar and load, and improve scheduling accuracy and speed.

[0007] Technical solution: To solve the above technical problems, the technical solution adopted by the present invention is as follows:

[0008] Firstly, an adaptive variable-step-size scheduling method for wind-solar-hydrogen-storage systems in energy islands is provided, including:

[0009] S1. Obtain the predicted values ​​of wind power and solar power output and load, and combine them with the deviation prediction model to obtain the actual values ​​of wind power and solar power output and load;

[0010] S2. Based on the actual values ​​of wind power, photovoltaic power output, and load, multiple initial scenarios are generated using Monte Carlo simulation.

[0011] S3. Use the synchronous back-substitution scene reduction technique to screen and reduce multiple initial scenes, and select typical scenes that satisfy the overall distribution of random variables and the probability of each typical scene.

[0012] S4. Based on typical scenarios and the probabilities of each typical scenario, the rolling optimization method of adaptive variable step size SMPC is used to solve the scheduling optimization model of the energy island wind-solar-hydrogen-storage system, and the scheduling result of the energy island wind-solar-hydrogen-storage system for the current period is obtained.

[0013] Secondly, the present invention provides a scheduling device for an energy island wind-solar-hydrogen storage system based on adaptive variable step size, including a processor and a storage medium;

[0014] The storage medium is used to store instructions;

[0015] The processor is configured to operate according to the instructions to execute the method according to the first aspect.

[0016] Thirdly, the present invention provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.

[0017] Beneficial Effects: Compared with existing technologies, this invention has the following advantages: Considering the volatility and uncertainty of renewable energy power generation, this invention processes prediction biases in wind speed and solar irradiance data to achieve uncertain prediction of wind and solar power output; it employs scenario reduction technology to eliminate calculations for numerous repetitive and extreme scenarios, balancing computational speed and model typicality; the proposed adaptive variable-step-size stochastic model predictive control (SMPC) method for wind-solar-hydrogen-storage system scheduling and control improves the problem of constant step size and built-in confidence within the rolling optimization cycle affecting computational accuracy and solution rate, thereby improving model prediction accuracy while ensuring solution efficiency and accuracy. This significantly enhances the absorption capacity of wind and solar energy and the utilization rate of energy storage power stations, thereby improving the grid's friendly operation capability. Through optimized scheduling and control, this technology is expected to effectively address the challenges of power system volatility and uncertainty while promoting the increase of renewable energy's share in the energy structure. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the Jensen wake in an embodiment of the present invention;

[0019] Figure 2 This is a schematic diagram of the floating photovoltaic power output considering mechanical vibration in an embodiment of the present invention;

[0020] Figure 3 This is a schematic diagram illustrating the adaptive stochastic model predictive control principle of an embodiment of the present invention;

[0021] Figure 4 This is a schematic diagram of the rolling optimization framework according to an embodiment of the present invention;

[0022] Figure 5 This is a diagram showing the scheduling results of a wind-solar-hydrogen storage system in a specific embodiment of the present invention;

[0023] Figure 6 This is a diagram showing the water network scheduling results of a wind-solar-hydrogen storage system in a specific embodiment of the present invention;

[0024] Figure 7 This is a diagram showing the hydrogen grid scheduling results of the wind-solar-hydrogen storage system in a specific embodiment of the present invention;

[0025] Figure 8 The figure shows the results of the system prediction step size Np changing in a specific embodiment of the present invention. Detailed Implementation

[0026] The present invention will be further illustrated below with reference to the accompanying drawings and specific examples. It should be understood that these examples are for illustrative purposes only and are not intended to limit the scope of the invention. After reading this invention, any modifications of the invention in various equivalent forms by those skilled in the art will fall within the scope defined by the appended claims.

[0027] In the description of this invention, the terms "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0028] Example 1

[0029] like Figure 1 As shown, a scheduling method for an energy island wind-solar-hydrogen storage system based on adaptive variable step size includes:

[0030] S1. Obtain the predicted values ​​of wind power and solar power output and load, and combine them with the deviation prediction model to obtain the actual values ​​of wind power and solar power output and load;

[0031] In this application, a deviation prediction model is generated based on the day-ahead forecast data and the day-ahead output and load data of the wind-solar-hydrogen-storage system established in this application. In some embodiments, the deviation prediction model includes:

[0032] The prediction error ΔP for wind power, photovoltaic output, and load follows a normal distribution N~(μ,σ) 2 ), where the mean μ is 0, and the variances of wind power, solar power output, and load are σ. 2 The formulas for predicting offsets are: 10, 10, and 5 respectively.

[0033]

[0034] Where f(ΔP) represents the predicted offset of wind power, photovoltaic power, and various types of loads.

[0035] S2. Based on the actual values ​​of wind power, photovoltaic power output, and load, multiple initial scenarios are generated using Monte Carlo simulation.

[0036] In this application, based on the probability density distribution of the prediction error of the parameters at each time point, the Monte Carlo model is used to sample the actual values ​​of wind power, photovoltaic power output and load to generate a large number of wind and solar power scenarios as initial scenarios; each scenario contains the error sequence prediction values ​​of wind power, photovoltaic power output and load in the future prediction time domain, and the initial probability of each scenario is 1 / N.

[0037] S3. Use the synchronous back-substitution scene reduction technique to screen and reduce multiple initial scenes, and select typical scenes that satisfy the overall distribution of random variables and the probability of each typical scene.

[0038] In some embodiments, step S3 specifically includes:

[0039] S31. The generated N initial scenarios are used as sample scenarios. Each sample scenario contains the error sequence prediction values ​​of wind power, photovoltaic power output and load in the future prediction time domain. The initial probability of each sample scenario is 1 / N.

[0040] S32. Perform Gaussian normalization on N sample scenes to eliminate the influence of parameter amplitude changes on the results;

[0041] S33. For each sample scenario x i Calculate x with other sample scenarios j The probability distance;

[0042]

[0043] In the formula, d ij Represents sample scenario x i and sample scenario x j The probability distance between them For sample scenario x i The probability, where N is the number of sample scenarios;

[0044] S34. Find the sample scenario x corresponding to the minimum value among the required probability distances. i Then delete it, but add the probability of that sample to the nearest sample scene x. j At this point, the updated sample scene x j probability for:

[0045]

[0046] In the formula, Sample scenarios x before and after the update are respectively. j The probability of;

[0047] S35. Repeat steps S32-S34 until the updated number of sample scenarios is reduced to the required number of typical scenarios Ns. Use the sample scenarios at this point as typical scenarios and calculate the probability of each typical scenario.

[0048] S4. Based on typical scenarios and the probabilities of each typical scenario, the rolling optimization method of adaptive variable step size SMPC is used to solve the scheduling optimization model of the energy island wind-solar-hydrogen-storage system, and the scheduling result of the energy island wind-solar-hydrogen-storage system for the current period is obtained.

[0049] In some embodiments, the scheduling optimization model of the energy island wind-solar-hydrogen-storage system is constructed based on the energy island wind-solar-hydrogen-storage system model, which includes: a floating photovoltaic power station scheduling model, a gas turbine power station model, a conversion-side equipment system model, and an energy storage station scheduling model.

[0050] (1) The dispatch model for floating wind farms includes:

[0051] The wake model of a floating wind farm is a wind farm model that considers the Jensen wake model, which is expressed as follows:

[0052]

[0053] In the formula, r0 is the radius of the wake after the generator, r d The radius of the wind turbine is represented by r, where a is the axial induction factor and r is the axial induction factor. x Let x be the wake radius at a distance x downstream of the generator unit, k be the expansion coefficient, and v be the radius of the wake. x v0 is the wake velocity at a distance x downstream of the unit, and v0 is the effective incoming wind velocity of the fan.

[0054] Downstream units may be located within the wake region of multiple upstream units. Based on the sum of squares of energy and wake superposition model, the effective incoming wind speed v at the downstream unit's rotor surface... j Represented as:

[0055]

[0056] In the formula, v i A is the effective incoming air velocity of the i-th unit. ij Let v be the overlap area from the wake of the i-th unit to the j-th unit. ij It is the wake velocity from the i-th unit to the j-th unit;

[0057] After considering the characteristics of the Jensen wake model, the relationship between wind farm output and wind speed is as follows:

[0058]

[0059] In the formula, E WT (t) represents the output of the wind farm at time t, n represents the number of turbines, ρ represents the air density, R represents the radius of the wind turbine blades, and C represents the output of the wind farm at time t. P The wind energy utilization coefficient of the wind turbine;

[0060] For floating wind turbines, combining the Jensen wake model and considering the impact of the six degrees of freedom motion of the floating wind turbine on the power output, the influence of the floating platform on the power characteristics of the wind turbine is mainly manifested in the deflection of the rotor surface relative to the direction of the incoming wind, including pitching and yaw; and the relative motion between the rotor surface and the incoming wind, including swaying and rolling.

[0061] This application only considers the deflection motion on the wind turbine surface. Considering only the two-degree-of-freedom floating platform motion causing the wind turbine surface to deflect, the actual wind-catching area of ​​the wind turbine is expressed as:

[0062] S′=M y M z S

[0063] When the pitch angle is α and the yaw angle is β, My is the transformation matrix when the wind turbine surface deviates from the Z-axis by an angle α; Mz is the transformation matrix when the wind turbine surface rotates around the Z-axis by an angle β, where S′ is the projected area of ​​the wind turbine surface in the direction of the incoming wind after the floating platform moves, and S is the wind turbine area.

[0064] In summary, the aerodynamic power model of a floating wind turbine that takes into account the wake is as follows:

[0065]

[0066] E WT (t) represents the power output of the wind farm at time t; n is the number of turbines, ρ is the air density, R is the radius of the wind turbine blades, and C is the power output of the wind farm at time t. P The wind turbine's wind energy utilization coefficient; the pitch angle is α, the yaw angle is β, and v. j Effective incoming air velocity at the turbine rotor surface;

[0067] (2) The floating photovoltaic power plant scheduling model includes:

[0068] Photovoltaic modules are sensitive to changes in light intensity. Floating power generation places photovoltaic modules on a floating water surface, and the mechanical vibrations caused by the undulations of the water surface can affect the temperature of the photovoltaic modules and the amount of solar radiation they receive.

[0069] Therefore, considering the output characteristics of floating photovoltaic modules under mechanical vibration conditions, a low-power photovoltaic panel is selected to simplify the actual large photovoltaic module. When the module vibrates mechanically, its parameters such as the direct sunlight area will also change periodically, causing its output characteristics to change.

[0070]

[0071] M represents the effective area of ​​the photovoltaic panel for illumination. The oscillation angle is 30° (here); f is the frequency of the mechanical oscillation (here 1.7 Hz); t is the time.

[0072] Figure 2This is a schematic diagram of the floating photovoltaic power output considering mechanical oscillation in this invention. Based on its power theory model, the factor of the periodic change in the effective area of ​​the photovoltaic panel under mechanical oscillation is added. That is, the output power generated under static conditions is multiplied by the periodically changing direct illumination area to obtain the power output relationship under mechanical oscillation.

[0073]

[0074] In the formula: E PV (t) represents the DC output power of the photovoltaic module at time t; G t (t) represents the actual radiated power; P pvr The rated power of the photovoltaic module under standard test conditions; η pv β is the power reduction factor for photovoltaic modules. T G represents the power temperature coefficient of a photovoltaic module. STC T STC For radiation intensity and battery temperature under standard test conditions, T C This refers to the actual battery temperature; it can be determined by the ambient temperature T. e The calculation yields the following result, as shown in Equation 2: T STC —Battery temperature under standard test conditions (25℃). Where:

[0075] T c =T e +KG t (t).

[0076] (3) The gas turbine power plant model includes:

[0077] Although wind power and solar power have many outstanding advantages, they also have strong uncertainties and intermittency. Therefore, this embodiment uses a gas turbine as a supplementary energy device, and the gas turbine output and scheduling model is as follows:

[0078]

[0079] In the formula, E GT (t) represents the electrical power output of the gas-fired power plant at time t, and η GT For power generation efficiency, q GT,fu LHV is the mass flow rate of fuel. fu The low calorific value of fuel; δ GT This refers to the start-up and shutdown status of the gas turbine. These are the maximum and minimum output of the gas turbine, respectively.

[0080] (4) The system model of the conversion side equipment includes: the load includes water load and hydrogen load, so the system must have equipment to produce fresh water and hydrogen. Here, the percolation method is used to produce fresh water and the electrolysis of water is used to produce hydrogen. The output and scheduling model of the desalination equipment is as follows:

[0081]

[0082] In the formula E SD (t) represents the power consumption of the de-diffusing device at time t, η SD1 and η SD2 These are the power consumption coefficient and desalination coefficient of the desalination equipment, respectively, V. SDOUT (t) is the output water volume of the desalination equipment, V SDIN (t) represents the influent flow rate of the desalination equipment. and These are the minimum and maximum water inflow rates for the desalination equipment. These are the minimum and maximum power consumption of the desalination equipment, δ SD It fades the start / stop status of the equipment;

[0083] The output and scheduling model for hydrogen production equipment is as follows:

[0084]

[0085] In the formula E HY (t) is the power consumption of the electrolysis equipment at time t, η HY1 and η HY2 These are the power consumption coefficient and electrolysis coefficient of the electrolysis equipment, V. HYOUT (t) is the outlet hydrogen volume of the electrolysis equipment, V HYIN (t) represents the water inlet flow rate of the electrolysis equipment. and These are the minimum and maximum water inlet volumes for the electrolysis equipment. These are the minimum and maximum power consumption of the electrolysis equipment, δ HY This refers to the start-up and shutdown status of the electrolysis equipment;

[0086] (5) Various energy storage station scheduling models include:

[0087] Combining the functions of batteries, water storage, and hydrogen tanks for peak shaving and valley filling, this wind-solar-hydrogen storage system incorporates such energy storage devices to improve system efficiency. For energy storage power stations, operational constraints include state of charge (SOC) constraints, state of charge / discharge (SDC) constraints, and charge / discharge power constraints. SOC constraints prevent deep charging and discharging, ensuring timely response and a longer lifespan for the energy storage system; furthermore, the batteries should maintain the same initial and final states after a 24-hour scheduling cycle. SDC constraints prevent batteries from charging and discharging simultaneously during operation.

[0088] Power constraints on energy storage power stations can prevent energy loss caused by heat generation during charging and discharging. The dispatching model for energy storage power stations is as follows:

[0089] In the formula S BA (t) represents the remaining battery charge at time t, S BA (t-1) represents the remaining battery charge at time t-1, E BA,ch (t) represents the electrical power stored in the battery at time t, E BA,dch (t) represents the electrical power released by the battery at time t, Δt is the scheduling interval, and η is the power released by the battery at time t. BA,ch For battery charging efficiency, η BA,dch For the battery's release efficiency; E represents the minimum and maximum battery capacity. BA,ch (t), E BA,dch (t) represent the charging and discharging power of the energy storage station during time period t, respectively;

[0090] These represent the minimum and maximum electrical power during the charging state, respectively. δ represents the minimum and maximum electrical power during discharge, respectively. BA,ch δ BA,dch These represent the charging and discharging states, respectively.

[0091] The freshwater storage tank scheduling model is as follows:

[0092]

[0093] In the formula S WA (t) represents the remaining water volume in the storage tank at time t, S WA (t-1) represents the remaining water volume in the storage tank at time t-1, V WA,ch (t) represents the amount of water stored in the storage tank at time t, V WA,dch (t) represents the amount of water released from the storage tank at time t, and η WA,ch η is the water storage efficiency of the water storage tank. WA,dch The release efficiency of the water storage tank; V represents the minimum and maximum water volume of the storage tank. WA,ch (t), V WA,dch (t) represent the amount of water filled and discharged from the storage tank during time period t, respectively; These represent the minimum and maximum water volume in the water storage state, respectively; δ represents the minimum and maximum water volume during the discharge state, respectively. WA,ch δ WA,dch These represent the water storage and water release states, respectively.

[0094] The scheduling model for hydrogen storage tanks is as follows:

[0095]

[0096] In the formula S HY (t) represents the remaining hydrogen quantity in the hydrogen storage tank at time t, S HY (t-1) represents the remaining hydrogen quantity in the hydrogen storage tank at time t-1, V HY,ch (t) represents the amount of hydrogen stored in the hydrogen storage tank at time t, V HY,dch (t) represents the amount of hydrogen released from the hydrogen storage tank at time t, and η HY,ch η represents the hydrogen storage efficiency of the hydrogen storage tank. HY,dch The release efficiency of the hydrogen storage tank; V represents the minimum and maximum hydrogen capacity of the hydrogen storage tank. HY,ch (t), V HY,dch (t) represent the amount of hydrogen charged and discharged from the hydrogen storage tank during time period t, respectively;

[0097] These represent the minimum and maximum hydrogen quantities in the hydrogen storage state, respectively. δ represents the minimum and maximum amount of hydrogen gas released, respectively. BA,ch δ BA,dch These represent the water storage and water release states, respectively.

[0098] The electricity-freshwater-hydrogen dispatch model in a wind-solar-hydrogen storage system refers to achieving a balance between the supply and demand of electricity, water, and hydrogen networks in a comprehensive energy system on an offshore energy island through optimized dispatch and energy management. Specifically: electricity balance involves the forecasting and dispatching of floating wind and solar power, as well as electricity storage and consumption. The electricity must not only meet the island's direct electricity demand but also provide power for hydrogen production via water electrolysis and seawater desalination. Freshwater balance involves the operation of seawater desalination equipment to meet freshwater load demand. The water production of the desalination unit needs to match the island's freshwater demand, and peak shaving and valley filling are implemented through water storage tanks. Hydrogen balance involves the operation of the water electrolysis hydrogen production unit, as well as hydrogen storage and use to meet hydrogen load demand during peak periods.

[0099] More specifically, the scheduling optimization model for the energy island wind-solar-hydrogen-storage system includes an objective function and constraints; the objective function F of the energy island wind-solar-hydrogen-storage system scheduling optimization model is expressed as:

[0100] F = max(F1 - F2)

[0101]

[0102] Where F1 is the objective function for the rolling optimization cycle revenue, F2 is the objective function for the unit operation penalty cost, s is the typical scenario after scenario reduction, and π s These are the probabilities of a typical scenario s, E(t) and Q(t). W (t), QG (t) represents the amount of electricity, fresh water, and hydrogen sold by the system under typical scenario s, and C Esell C Wsell C Gsell These are the system's electricity, water, and hydrogen energy prices, respectively; Δt is the dispatch interval; and λ is the energy level. gas , λ gt E GT (t), E HY (t), E SD (t) represents the gas price, gas turbine penalty coefficient, gas turbine output in typical scenario s, power consumption of electrolysis equipment, and power consumption of desalination equipment, respectively. N S N is the number of typical scenarios. p To optimize the cycle step size for rolling.

[0103] The constraints include electrical power balance constraints, water network balance constraints, and hydrogen network balance constraints.

[0104] The power balance constraint is expressed as follows:

[0105] E WT (t)+E PV (t)+E GT (t)+E BA,dch (t)

[0106] =E load (t)+E SF (t)+E HY (t)+E SD (t)+E BA,ch (t)

[0107] In the formula, E WT (t) represents the power output of the wind farm at time t, E PV (t) represents the output of the photovoltaic power station at time t, E GT (t) is the output of the gas turbine at time t, E load (t) is the load at time t, E SF (t) represents the electricity consumption of the marine ranch at time t, E HY (t) is the power consumption of the electrolysis equipment at time t, E SD (t) represents the power consumption of the de-emphasis device at time t, E BA,ch (t), E BA,dch (t) represent the charging and discharging power of the energy storage station during time period t, respectively;

[0108] The water network balance constraints are as follows:

[0109] V SDOUT (t)-V WA,ch (t)+V WA,dch (t)=VWA,load (t)

[0110] In the formula, V SDOUT (t) is the output water volume of the desalination equipment, V WA,ch (t), V WA,dch (t) represent the amount of water filled and discharged from the storage tank during time period t, respectively; V WA,load (t) represents the water load.

[0111] The equilibrium constraints of the hydrogen network are as follows:

[0112] V HYOUT (t)-V HY,ch (t)+V HY,dch (t)=V HY,load (t)

[0113] In the formula, V HYOUT (t) is the outlet hydrogen volume of the electrolysis equipment, V HY,ch (t), V HY,dch (t) represent the amount of hydrogen charged and discharged from the hydrogen storage tank during time period t, respectively. HY,load (t) represents the hydrogen load.

[0114] In integrated wind-solar-hydrogen-storage energy systems, wind and solar resources exhibit significant uncertainties. Model predictive control (MMDC) can better address the stochasticity of renewable energy output. Furthermore, in energy island integrated wind-solar-hydrogen-storage energy systems, stochastic model predictive control can effectively reduce the impact of uncertainties on scheduling results. The stochastic model predictive control scheduling strategy for the energy island integrated wind-solar-hydrogen-storage energy system proposed in this embodiment is as follows: Figure 3 As shown, this scheduling strategy mainly consists of four parts: a prediction model, scenario reduction, rolling optimization, and feedback correction. The prediction model is used to predict the outputs of both supply and demand sides in the prediction time domain. Based on the prediction results, a large number of initial scenarios are filtered using scenario reduction techniques. Then, based on the reduced scenarios, optimization is performed in the prediction time domain with the objective function of minimizing the expected operating cost of the system. The first variable of the resulting control sequence is then used to control the system. Finally, the deviation between the actual system output and the predicted output is fed back to correct the prediction model for the next step of rolling optimization.

[0115] To improve the robustness and prediction accuracy of the rolling optimization cycle, this embodiment innovatively proposes an adaptive control step size technique in the rolling optimization stage, based on SMPC. The purpose of this method is to reduce the prediction deviation between the scheduling command value and the actual output value of the wind-solar-hydrogen-storage system. By tracking the actual system load, the adaptive variable step size method is expected to adaptively adjust the step size of the rolling optimization cycle according to the scheduling error, thereby improving the flexibility of the optimization stage. The deviation reference coefficient F is obtained by comparing the deviation between the system output scheduling command and the actual output of two adjacent rolling cycles, as shown in the following formula. The objective of this method is to reduce the prediction deviation between the scheduling command value and the actual output value of various energy storage power stations, gas turbine power stations, and conversion equipment. The adaptive variable step size method is expected to adaptively adjust the step size of the rolling optimization cycle according to the scheduling error, thereby improving the flexibility of the optimization stage.

[0116] Furthermore, the rolling optimization method of adaptive variable step size SMPC is used to solve the scheduling optimization model of the energy island wind-solar-hydrogen-storage system, and the scheduling results of the energy island wind-solar-hydrogen-storage system for the current period are obtained, including:

[0117] S41. Obtain the actual operating data of the energy island wind-solar-hydrogen storage system after executing the previous cycle scheduling command;

[0118] S42. Based on the predicted and actual operating data of various energy storage power stations, gas turbine power stations, and conversion-side equipment in the previous cycle, adaptively determine the rolling optimization cycle step size N. P ;

[0119] In some embodiments, S42 specifically includes:

[0120] Based on the predicted and actual operating data of various energy storage power stations, gas turbine power stations, and conversion-side equipment in two adjacent rolling optimization cycles, the deviation reference coefficient F is obtained;

[0121]

[0122] In the formula, E WT (t) represents the power output of the wind farm at time t, E PV (t) represents the output of the photovoltaic power station at time t, E GT (t) is the output of the gas turbine at time t, E load (t) is the load at time t, E SF (t) represents the electricity consumption of the marine ranch at time t, E HY (t) is the power consumption of the electrolysis equipment at time t, E SD (t) represents the power consumption of the de-emphasis device at time t, E BA,ch (t), E BA,dch (t) represent the charging and discharging power of the energy storage station during time period t, respectively; VSDOUT (t) is the output water volume of the desalination equipment, V WA,ch (t), V WA,dch (t) represent the amount of water filled and discharged from the storage tank during time period t, respectively; V WA,load (t) represents the water load; V HYOUT (t) is the outlet hydrogen volume of the electrolysis equipment, V HY,ch (t), V HY,dch (t) represent the amount of hydrogen charged and discharged from the hydrogen storage tank during time period t, respectively. HY,load (t) represents the hydrogen load;

[0123] δ E (t), δ Vwa (t), δ Vhy (t) represents the deviation reference coefficients of the power grid system, water network system, and hydrogen network system at time t, respectively; δ(t) and δ(t-1) represent the deviation reference coefficients of the wind, solar, hydrogen, and storage systems of the energy island at times t and t-1, respectively.

[0124] Based on the deviation reference coefficient F, the rolling optimization cycle step size N is... P Adaptive changes can be made; when the scheduling prediction error decreases, i.e., F is smaller, the prediction step size N can be increased. P To achieve more global scheduling optimization; when the scheduling prediction error increases, i.e., when F is large, the prediction step size N is shortened. P This reduces prediction errors and achieves more accurate scheduling optimization. The specific criteria for determining the adaptive step size are as follows:

[0125]

[0126] S43, Based on the rolling optimization cycle step size N P The variable weighting method is used to distinguish the confidence levels of each cycle in the rolling optimization process;

[0127] This embodiment, based on SMPC, incorporates the concept of variable weight coefficients. In traditional SMPC scheduling strategies, the confidence level of the prediction result remains constant across all optimization cycles. As the optimization cycle step size increases, data at the end of the cycle affects the prediction result, causing the prediction accuracy to gradually decrease. Within the optimization cycle, the further away from the current time, the larger the error of the predicted result and the lower the reliability; however, shortening the optimization cycle length is not conducive to achieving global optimization and cannot comprehensively consider the operating state at future times. Therefore, it is necessary to reasonably set the optimization cycle length, by reducing the confidence level at the end of the optimization cycle and increasing the confidence level at the beginning of the cycle, to achieve accurate scheduling control at the current time. To reflect the difference in the confidence level of the prediction result at different times during the simulation, a variable weight method is introduced to distinguish the reliability of the results obtained from different prediction step sizes, that is, the weight coefficient gradually decreases as the current time becomes further away. The specific formula for the variable weight coefficient proposed in this embodiment is as follows:

[0128] f1 = diag(φ i-1 ), 0 < φ < 1, i = 1, ..., N p ,

[0129] In the formula, f1 represents the confidence level of different step lengths. As the step length increases, the confidence level decreases. φ is a constant between 0 and 1. In this embodiment, φ is 0.98. i represents different step lengths.

[0130] S44. Based on typical scenarios and the probabilities of each typical scenario, the scheduling optimization model of the energy island wind-solar-hydrogen-storage system is solved to obtain the scheduling result of the energy island wind-solar-hydrogen-storage system for this period as prediction data, and a scheduling instruction is generated and sent to the energy island wind-solar-hydrogen-storage system for execution.

[0131] The specific working principle of rolling optimization is that the system calculates the predicted input for the future time period based on the prediction model to obtain the optimal control result for this period. However, the system only outputs the result at the current sampling time, and the other results are used as the prediction input for the next round of optimization. During this process, the length of the prediction sequence remains unchanged.

[0132] Based on the current load forecast, the renewable energy supply side includes wind power, photovoltaic power, and [unspecified source]. The power supply method involves the complete consumption of wind and photovoltaic power, followed by gas turbine regulation, and finally compensation by batteries. Therefore, the rolling optimization cycle revenue objective function F1 (revenue from electricity sales, fresh water, and hydrogen production) has the highest value. Secondly, the system operation requires the consumption of gas and some electricity that could otherwise be sold (this electricity is used for seawater desalination and water electrolysis to produce hydrogen). Furthermore, frequent fluctuations in the power output of the gas turbine units are detrimental to the safe and stable operation of the units. Therefore, a unit operation penalty cost is introduced, resulting in objective function F2. Δt is the scheduling interval, which is 1 hour in this embodiment. The initial scenario in this embodiment is 100, while Ns and Np are taken as 10 and 8, respectively.

[0133] Feedback correction refers to the process of adjusting the system to more closely approximate the desired output by comparing the difference between the actual output and the expected output and feeding that difference back to the input. Wind-solar-hydrogen-storage system models consider uncertainties on both the supply and demand sides, requiring the addition of biases to the day-ahead forecast data to simulate the scheduling results of the wind-solar-hydrogen-storage system. By comparing and correcting the actual scheduling results with the forecast model through a feedback loop, the model's prediction error over the time scale is reduced. Increasing the normal distribution error by N(0,10) for wind speed and solar irradiance, and by N(0,5) for load, reduces the uncertainty of system forecasts and improves the model's prediction accuracy.

[0134] Simulation verification experiments: such as Figure 3 An adaptive variable-step stochastic model predictive control scheduling model for a wind-solar-hydrogen-storage system is established to verify and analyze the scheduling of offshore wind-solar-hydrogen-storage systems. The region has one floating photovoltaic power station, one floating wind farm, one seawater desalination and hydrogen production station, one marine ranch, and three energy storage power stations. Electricity is sold to the region and its coastal areas, with prices calculated based on local time-of-use pricing, as shown in Table 1. The wind farm has a rated installed capacity of 264MW (33*8MW), the photovoltaic power station has a rated capacity of 150MW, the gas turbine power station has a rated capacity of 300MW, and the energy storage power station has a capacity of 45MW / 150MW·h.

[0135] Table 1: Time-of-use electricity pricing table for a certain area

[0136]

[0137]

[0138] Table 2: Time-of-use Water Pricing Table for a Certain Area

[0139]

[0140] Table 3: Time-of-Use Hydrogen Price Table for a Certain Location

[0141]

[0142] 1) Construct a biased prediction model that considers the uncertainties of wind and solar energy:

[0143] Generate a deviation prediction model. The prediction errors ΔP for wind power, photovoltaic output, and load follow a normal distribution N~(μ,σ). 2 The mean μ is 0, and the variances of wind speed, light intensity, and various loads are σ. 2 The values ​​are 10, 10, and 5 respectively.

[0144] 2) Generate initial scenarios based on the deviation prediction model, and use scenario reduction technology to screen and reduce the initial scenarios to obtain typical scenarios; ① Generate 100 initial scenarios based on the prediction data of wind speed, light intensity, various loads and the deviation prediction model.

[0145] ② Perform Gaussian normalization on the 100 initial scenes obtained;

[0146] ③ Calculate the probabilistic distance between the sample scene and other scenes;

[0147] ④ Find the sample corresponding to the minimum probability distance and delete it. Then add the probability of that sample to the sample closest to it.

[0148] Sample location;

[0149] ⑤ Repeat steps ②-④ until 10 typical scenarios are obtained, and calculate the probability of each typical scenario.

[0150] 3) Design an adaptive variable step size stochastic model predictive controller:

[0151] ① Use adaptive control step size technology to change the rolling optimization step size N P ;

[0152] ② Use the variable weighting method to distinguish the confidence levels of each cycle in the rolling optimization process;

[0153] ③ The objective function optimizes the scheduling of the wind-solar-hydrogen-storage system. This optimization problem can be solved by calling the Gurobi solver.

[0154] Output the scheduling results for each cycle and generate scheduling instructions;

[0155] ④ Input the actual running data after executing the scheduling instruction into step ①, and compare the actual running data with the predicted data to achieve...

[0156] Feedback correction;

[0157] ⑤ Simulation solution:

[0158] Figure 4 This is a schematic diagram of the rolling optimization framework according to an embodiment of the present invention; Figure 5The scheduling results of the wind-solar-hydrogen storage system are obtained using the adaptive variable step size stochastic model predictive control method. Figure 6 This is a diagram showing the water network scheduling results of a wind-solar-hydrogen storage system in a specific embodiment of the present invention; Figure 7 This is a diagram showing the hydrogen grid scheduling results of the wind-solar-hydrogen storage system in a specific embodiment of the present invention; Figure 8 The figure shows the results of the system prediction step size Np changing in a specific embodiment of the present invention.

[0159] ⑥ Determination of evaluation indicators and comparison methods:

[0160] This invention uses root mean square deviation I RMSE The following evaluation metrics were proposed to assess the prediction performance of the proposed adaptive variable step size SMPC algorithm and to compare the prediction bias of different predictive control strategies:

[0161]

[0162] In addition, to demonstrate the effectiveness of the method proposed in this invention, the following methods are used as comparative methods:

[0163] 1. SMPC: Traditional stochastic model predictive control method;

[0164] 2. Variable Weight SMPC: A predictive control method based on variable weight stochastic models;

[0165] 3. Adaptive Variable Step Size SMPC: namely, the adaptive variable step size stochastic model predictive control method proposed in this application.

[0166] Table 4 presents a comparison of the comprehensive scheduling benefits of three different strategies, and Table 5 presents quantitative comparison indicators for different methods.

[0167] Table 4 Comparison of scheduling benefits under different strategies

[0168] Strategy Scheduling revenue (yuan) Traditional SMPC 1702780 Variable Weight SMPC 1806518 Adaptive Variable Step Size SMPC 1782269

[0169] Table 5. Quantitative Comparison Indicators of Different Methods

[0170] Strategy IRMSE Traditional SMPC 1.5436 Variable Weight SMPC 2.3623 Adaptive Variable Step Size SMPC 1.9634

[0171] The table shows that, quantitatively, the average relative deviations of different methods based on the proposed prediction deviation evaluation index are roughly the same, indicating that different methods can track scheduling instructions well.

[0172] Based on the comparison of indicators, the adaptive variable step size SMPC strategy and algorithm proposed in this application, compared with the traditional SMPC algorithm, increase the scheduling benefit by 0.85%, thus improving the scheduling economy to a certain extent. Compared with the variable weight SMPC algorithm, the scheduling cost is slightly reduced, but the reduction is only 0.009%. Its prediction error is reduced by 11.2% compared with the traditional SMPC and by 24.1% compared with the variable weight SMPC strategy. The adaptive variable step size can adjust the step size according to the prediction error, effectively solving the prediction error caused by the constant weight and step size problem in the rolling optimization cycle. The results show that the adaptive variable step size SMPC strategy proposed in this application can effectively reduce the prediction error between the scheduling commands and actual outputs of the wind-solar-hydrogen-storage system.

[0173] Example 2

[0174] Secondly, this embodiment provides a scheduling device for an energy island wind-solar-hydrogen storage system based on adaptive variable step size, including a processor and a storage medium;

[0175] The storage medium is used to store instructions;

[0176] The processor is configured to operate according to the instructions to execute the method according to Embodiment 1.

[0177] Example 3

[0178] Thirdly, this embodiment provides a storage medium on which a computer program is stored, which, when executed by a processor, implements the method described in embodiment 1.

[0179] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0180] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. 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... Figure 1One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0181] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0182] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0183] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A scheduling method for an energy island wind-solar-hydrogen storage system based on adaptive variable step size, characterized in that, include: S1. Obtain the predicted values ​​of wind power and solar power output and load, and combine them with the deviation prediction model to obtain the actual values ​​of wind power and solar power output and load; S2. Based on the actual values ​​of wind power, photovoltaic power output, and load, multiple initial scenarios are generated using Monte Carlo simulation. S3. Use the synchronous back-substitution scene reduction technique to screen and reduce multiple initial scenes, and select typical scenes that satisfy the overall distribution of random variables and the probability of each typical scene. S4. Based on typical scenarios and the probabilities of each typical scenario, the rolling optimization method of adaptive variable step size SMPC is used to solve the scheduling optimization model of the energy island wind-solar-hydrogen-storage system to obtain the scheduling result of the energy island wind-solar-hydrogen-storage system for the current period; wherein, the scheduling optimization model of the energy island wind-solar-hydrogen-storage system includes an objective function and constraints. The objective function F of the energy island wind-solar-hydrogen storage system scheduling optimization model is expressed as: ； Where F1 is the objective function for the revenue of the rolling optimization cycle, F2 is the objective function for the unit operation penalty cost, and s is the typical scenario after scenario reduction. It represents the probability of typical scenario s. , , This refers to the amount of electricity, fresh water, and hydrogen sold by the system under typical scenario s. , , These are the prices of the system's electricity, water, and hydrogen energy sources, respectively. It is the scheduling interval. , , , , These are, respectively, the gas price, the gas turbine penalty coefficient, the gas turbine output in typical scenario s, the power consumption of the electrolysis equipment, and the power consumption of the desalination equipment, N. S These are typical scenario numbers. To optimize the cycle step size for rolling optimization; The constraints include electrical power balance constraints, water network balance constraints, and hydrogen network balance constraints. The power balance constraint is expressed as follows: ； In the formula, It is the output of the wind farm at time t. It is the output of the photovoltaic power station at time t. It is the output of the gas turbine at time t. It is the load at time t. This represents the power consumption of the marine ranch at time t. It is the power consumption of the electrolysis equipment at time t. The power consumption of the device is reduced at time t. , These represent the charging and discharging power of the energy storage power station within the time period t; The water network balance constraints are as follows: ； In the formula, It refers to the water output of the desalination equipment. , These represent the amount of water filled and discharged from the storage tank during time period t, respectively. Represents water load; The equilibrium constraints of the hydrogen network are as follows: ； In the formula, This refers to the output hydrogen volume of the electrolysis equipment. , These represent the amount of hydrogen added and released from the hydrogen storage tank during time period t. This represents the hydrogen load.

2. The method according to claim 1, characterized in that, The deviation prediction model includes: The prediction error ∆P of wind power, photovoltaic output, and load follows a normal distribution N~(μ,σ) 2 ), where the mean μ is 0, and the variances of wind power, solar power output, and load are σ. 2 The formulas for predicting offsets are: 10, 10, and 5 respectively. ； in, This refers to the predicted offset for wind power, solar power, and various types of loads.

3. The method according to claim 1, characterized in that, The synchronous back-substitution scene reduction technique is used to screen and reduce multiple initial scenes, selecting typical scenes that satisfy the overall distribution of random variables and the probability of each typical scene, including: S31. The generated N initial scenarios are used as sample scenarios. Each sample scenario contains the error sequence prediction values ​​of wind power, photovoltaic power output and load in the future prediction time domain. The initial probability of each sample scenario is 1 / N. S32. Perform Gaussian normalization on N sample scenes to eliminate the influence of parameter amplitude changes on the results; S33. For each sample scenario x i Calculate x with other sample scenarios j The probability distance; ； In the formula, d ij Represents sample scenario x i and sample scenario x j The probability distance between them For sample scenario x i The probability, where N is the number of sample scenarios; S34, find the sample scene x corresponding to the minimum value in the sought probability distance i and delete it, but add the probability of this sample to the sample scene x j closest to it at this time, the probability of the updated sample scene x j is: ​ ； wherein , are the probabilities of the sample scenario x j before and after the update, respectively. S35. Repeat steps S32-S34 until the updated number of sample scenarios is reduced to the required number of typical scenarios Ns. Use the sample scenarios at this point as typical scenarios and calculate the probability of each typical scenario.

4. The method of claim 1, wherein, The scheduling optimization model of the energy island wind-solar-hydrogen-storage system is constructed based on the energy island wind-solar-hydrogen-storage system model, which includes: a floating photovoltaic power station scheduling model, a gas turbine power station model, a conversion-side equipment system model, and an energy storage station scheduling model. (1) The floating wind farm dispatch model includes: The wake model of a floating wind farm is a wind farm model that considers the Jensen wake model. The aerodynamic power model of a floating wind turbine considering the wake is as follows: ； P(t) represents the output of the wind farm at time t; n is the number of units, p is the air density, R is the radius of the fan blade, C P is the wind energy utilization coefficient of the fan; the pitch angle is , the yaw angle is , the effective incoming wind speed of the unit wind wheel surface; (2) The floating photovoltaic power station scheduling model includes: ； ； In the formula: is the output DC power of the photovoltaic module at time t; is the actual radiation power; is the rated power of the photovoltaic module under standard test conditions; is the power reduction coefficient of the photovoltaic module; is the power temperature coefficient of the photovoltaic module; , is the radiation intensity under standard test conditions, the battery temperature, is the actual battery temperature; M is the light effective area of the photovoltaic panel; is the oscillation angle; f is the frequency of mechanical oscillation; (3) The gas turbine power plant model includes: ； wherein is the electric power output by the gas power plant at time t, is the electric power output by the gas power plant at time t, is the mass flow of fuel, is the lower heating value of the fuel; is the start-stop state of the gas turbine, , are the maximum and minimum power output of the gas turbine, respectively; (4) The system model of the conversion side equipment includes: the load includes water load and hydrogen load, so there must be equipment in the system to produce fresh water and hydrogen. Here, the percolation method is used to produce fresh water and the electrolysis of water to produce hydrogen. The output and scheduling model of the desalination equipment is as follows: ； wherein is the power consumption of the desalination device at time t, and are the power consumption coefficient and the desalination coefficient of the desalination device, respectively, is the water output of the desalination device, is the water input of the desalination device, and are the minimum and maximum water inputs of the desalination device, respectively, , are the minimum and maximum power consumptions of the desalination device, respectively, is the start-stop state of the desalination device; The output and scheduling model for hydrogen production equipment is as follows: ； wherein is the power consumption of the electrolysis device at time t, and are the power consumption coefficient and the electrolysis coefficient of the electrolysis device, respectively, is the outlet hydrogen quantity of the electrolysis device, is the inlet water quantity of the electrolysis device, and are the minimum and maximum inlet water quantities of the electrolysis device, respectively, , are the minimum and maximum power consumption of the electrolysis device, respectively, is the start-stop state of the electrolysis device; (5) Various energy storage station scheduling models include: The energy storage power station scheduling model is as follows: ； In the formula Let be the remaining charge of the battery at time t. Let be the remaining charge of the battery at time t-1. Let be the electrical power stored in the battery at time t. Let be the electrical power released by the battery at time t. For scheduling intervals, For battery charging efficiency, For the battery's release efficiency; , These are the minimum and maximum battery capacity. , These represent the charging and discharging power of the energy storage power station within the time period t; , These represent the minimum and maximum electrical power during the charging state, respectively. , These represent the minimum and maximum electrical power during discharge, respectively. , These represent the charging and discharging states, respectively. The freshwater storage tank scheduling model is as follows: ； In the formula Let be the remaining water volume in the storage tank at time t. Let be the remaining water volume in the storage tank at time t-1. Let t be the amount of water stored in the storage tank. Let t be the amount of water released from the storage tank. For the water storage efficiency of the water storage tank, The release efficiency of the water storage tank; , These are the minimum and maximum water volumes of the storage tank; , These represent the amount of water filled and discharged from the storage tank during time period t, respectively. , These represent the minimum and maximum water volume in the water storage state, respectively; , These represent the minimum and maximum water volumes during the discharge process, respectively. , These represent the water storage and water release states, respectively. The scheduling model for hydrogen storage tanks is as follows: ； In the formula Let t represent the remaining hydrogen quantity in the hydrogen storage tank at time t. Let be the amount of hydrogen remaining in the hydrogen storage tank at time t-1. Let t be the amount of hydrogen stored in the hydrogen storage tank at time t. Let t be the amount of hydrogen released from the hydrogen storage tank. The hydrogen storage efficiency of the hydrogen storage tank. The release efficiency of the hydrogen storage tank; , These represent the minimum and maximum hydrogen capacity of the hydrogen storage tank. , These represent the amount of hydrogen added and released from the hydrogen storage tank during time period t, respectively. , These represent the minimum and maximum hydrogen quantities in the hydrogen storage state, respectively. , These represent the minimum and maximum amounts of hydrogen gas released, respectively. , These represent the water storage and water release states, respectively.

5. The scheduling method for an energy island wind-solar-hydrogen storage system based on adaptive variable step size according to claim 1, characterized in that, The rolling optimization method of adaptive variable step size SMPC is used to solve the scheduling optimization model of the energy island wind-solar-hydrogen-storage system, and the scheduling results of the energy island wind-solar-hydrogen-storage system for the current period are obtained, including: S41. Obtain the actual operating data of the energy island wind-solar-hydrogen storage system after executing the previous cycle scheduling command; S42. Based on the predicted and actual operating data of various energy storage power stations, gas turbine power stations, and conversion-side equipment in the previous cycle, adaptively determine the rolling optimization cycle step size N. P ; S43, Based on the rolling optimization cycle step size N P The variable weighting method is used to distinguish the confidence levels of each cycle in the rolling optimization process; S44. Based on typical scenarios and the probabilities of each typical scenario, the scheduling optimization model of the energy island wind-solar-hydrogen-storage system is solved to obtain the scheduling result of the energy island wind-solar-hydrogen-storage system for this period as prediction data, and a scheduling instruction is generated and sent to the energy island wind-solar-hydrogen-storage system for execution.

6. The scheduling method for an energy island wind-solar-hydrogen storage system based on adaptive variable step size according to claim 5, characterized in that... S42 includes: Based on the predicted and actual operating data of various energy storage power stations, gas turbine power stations, and conversion-side equipment in two adjacent rolling optimization cycles, the deviation reference coefficient F is obtained; ； In the formula, It is the output of the wind farm at time t. It is the output of the photovoltaic power station at time t. It is the output of the gas turbine at time t. It is the load at time t. This represents the power consumption of the marine ranch at time t. It is the power consumption of the electrolysis equipment at time t. The power consumption of the device is reduced at time t. , These represent the charging and discharging power of the energy storage power station within the time period t; It refers to the water output of the desalination equipment. , These represent the amount of water filled and discharged from the storage tank during time period t, respectively. Represents water load; This refers to the output hydrogen volume of the electrolysis equipment. , These represent the amount of hydrogen added and released from the hydrogen storage tank during time period t. Represents hydrogen load; These represent the deviation reference coefficients of the power grid system, water network system, and hydrogen network system at time t, respectively. These represent the deviation reference coefficients of the energy island wind-solar-hydrogen storage system at times t and t-1, respectively. Based on the deviation reference coefficient F, the rolling optimization cycle step size N is... P Adaptive changes are made; 。 7. The scheduling method for an energy island wind-solar-hydrogen storage system based on adaptive variable step size according to claim 5, characterized in that, S43 includes: ， In the formula, This represents the confidence level for different step sizes; as the step size increases, the confidence level decreases. It is a constant between 0 and 1, where i represents different time lengths.

8. A scheduling device for an energy island wind-solar-hydrogen storage system based on adaptive variable step size, characterized in that, Including processor and storage media; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the method according to any one of claims 1 to 7.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method described in any one of claims 1 to 7.

Images