Energy island wind-solar-hydrogen storage system scheduling method based on self-adaptive variable step size

Abstract

The invention discloses an energy island wind-solar-hydrogen storage system scheduling method based on adaptive variable step size, which comprises the following steps: acquiring wind power, photovoltaic output and load prediction values, and combining a deviation prediction model to obtain wind power, photovoltaic output and load actual values; utilizing Monte Carlo simulation to generate a plurality of initial scenes; screening and reducing the plurality of initial scenes by using a synchronous back substitution scene reduction technology, and selecting typical scenes meeting overall distribution of random variables and the probability of each typical scene; and based on the typical scenes and the probability of each typical scene, solving the scheduling optimization model of the energy island wind-light-hydrogen storage system by using a rolling optimization method of adaptive variable step size SMPC to obtain a scheduling result of the energy island wind-light-hydrogen storage system in the current period. According to the method, the wind and light uncertainty is considered, and the prediction deviation of the prediction model is fed back and corrected, so that the fluctuation and uncertainty of wind, light and load are effectively dealt with, and the scheduling precision and speed are improved.

Description

Technical Field

[0001] The present invention belongs to the field of wind farm automatic control, and in particular relates to an energy island wind-solar-hydrogen storage system scheduling method based on adaptive variable step size. Background Art

[0002] Wind, solar, hydrogen, and storage multi-energy complementarity utilizes the obvious short-term, medium-term, and long-term complementary characteristics of wind and solar energy, relies on the regulatory capabilities of hydrogen and water storage, and cooperates with large-scale wind and solar energy power generation projects to achieve joint complementarity through capacity configuration, scheduling operation, and unit control, thereby realizing the efficient utilization of clean energy such as hydropower, wind power, and solar power. Therefore, research on the optimized scheduling and operation of "wind, solar, hydrogen, and storage" on large-scale offshore energy islands has great development potential. The key components of wind and solar complementary power generation coupled with hydrogen energy storage systems are relatively mature in technology. Through capacity configuration design and system control optimization, system costs can be reduced and energy conversion efficiency can be improved, thereby achieving efficient energy utilization. Therefore, research on the coordinated scheduling and operation of "wind, solar, hydrogen, and storage" on offshore energy islands is of great significance to promoting the sustainable development of the renewable energy industry.

[0003] Among many control methods, Model Predictive Control (MPC) has the advantages of being able to handle delays, multiple variables, and constraints, and its rolling optimization and feedback correction ideas improve the ability to suppress disturbances, making it one of the most important methods in engineering control. The main ideas of Model Predictive Control (MPC) include model prediction, rolling optimization, and feedback correction. The advantage of MPC is that it can handle problems such as uncertainty, nonlinearity, and large inertia in the solution object. Currently, Model Predictive Control is widely used in the scheduling research of energy complementary systems.

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

[0005] Therefore, it is necessary to further optimize the SMPC rolling optimization link and study how to improve the prediction accuracy of the SMPC method to improve the smooth 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 shortcomings of the above-mentioned prior art, the purpose of the present invention is to provide an energy island wind-solar-hydrogen storage system scheduling method based on adaptive variable step size. By taking into account the uncertainty of wind and solar power, and at the same time feedback-correcting the prediction deviation of the prediction model, it can effectively deal with the volatility and uncertainty of wind, light and load, and improve the scheduling accuracy and speed.

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

[0008] In a first aspect, a method for scheduling an energy island wind-solar-hydrogen storage system based on an adaptive variable step size is provided, comprising:

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

[0010] S2. Generate multiple initial scenarios using Monte Carlo simulation based on wind power, photovoltaic output, and actual load values;

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

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

[0013] In a second aspect, the present invention provides an energy island wind-solar-hydrogen storage system scheduling device 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 perform the method according to the first aspect.

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

[0017] Beneficial effects: Compared with the existing technology, the present invention has the following advantages: the present invention takes into account the volatility and uncertainty of renewable energy power generation, performs prediction deviation processing on wind speed and light data, and realizes the uncertainty prediction of wind and solar output; uses scene reduction technology to eliminate the calculation of a large number of repeated scenes and extreme scenes, taking into account both the operation rate and the typicality of the model; the proposed adaptive variable step-size random model predictive control SMPC wind, solar and hydrogen storage system scheduling control method improves the problem that the constant step size and the constant built-in confidence in the rolling optimization cycle affect the calculation accuracy and solution rate, thereby improving the model prediction accuracy while ensuring the solution efficiency and accuracy, and can significantly improve the absorption capacity of wind and solar energy and the utilization rate of energy storage power stations, thereby improving the friendly operation capacity of the power grid. By optimizing scheduling control, this technology is expected to effectively respond to the volatility and uncertainty challenges of the power system while promoting the increase in the proportion of renewable energy in the energy structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 Schematic diagram of Jensen wake according to an embodiment of the present invention;

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

[0020] Figure 3 Schematic diagram of the adaptive stochastic model predictive control principle of an embodiment of the present invention;

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

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

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

[0025] Figure 8 This is a diagram showing the change in the system prediction step size Np in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0026] The present invention is further illustrated below with reference to the accompanying drawings and specific examples. It should be understood that these examples are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, modifications of various equivalent forms of the present invention made by those skilled in the art all fall within the scope defined by the claims attached to this application.

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

[0028] Example 1

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

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

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

[0032] The forecast error ΔP of wind power, photovoltaic output and load obeys the normal distribution N~(μ,σ 2 ), where the mean μ is 0, and the variance of wind power, photovoltaic output, and load is σ 2 They are 10, 10, and 5 respectively. The formula for predicting the offset is:

[0033]

[0034] Among them, f(ΔP) is the predicted offset of wind power, photovoltaic power and various loads.

[0035] S2. Generate multiple initial scenarios using Monte Carlo simulation based on wind power, photovoltaic output, and actual load values;

[0036] In this application, based on the probability density distribution of the prediction error of the parameters at each moment, the Monte Carlo model is used to sample the actual values ​​of wind power, photovoltaic output and load to generate a large number of wind and solar scenarios as initial scenarios; each scenario contains the error sequence prediction values ​​of wind power, photovoltaic 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 scenario reduction technique to screen multiple initial scenarios and select typical scenarios that satisfy the overall distribution of random variables and the probability of each typical scenario;

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

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

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

[0041] S33. For each sample scene x i , calculate and other sample scenes x j The probability distance;

[0042]

[0043] Where, d ij Represents a sample scene x i and sample scene x j The probability distance between For the sample scene x i The probability of , N is the number of sample scenes;

[0044] S34. Find the sample scene x corresponding to the minimum value in the required probability distance i And delete it, but add the probability of the sample to the sample scene x closest to it j At this time, the updated sample scene x j Probability for:

[0045]

[0046] Where, The sample scenes x before and after the update are j probability;

[0047] S35. Repeat steps S32-S34 until the number of updated sample scenes is reduced to the required number of typical scenes Ns. The sample scenes at this time are used as typical scenes, and the probability of each typical scene is calculated.

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

[0049] In some embodiments, the energy island wind-solar-hydrogen storage system scheduling optimization model is constructed based on the energy island wind-solar-hydrogen storage system model, and the energy island wind-solar-hydrogen storage system model 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 floating wind farm dispatch model includes:

[0051] The floating wind farm wake model is a wind farm model that takes the Jensen wake model into consideration. The Jensen wake model is expressed as follows:

[0052]

[0053] Where r0 is the wake radius behind the generator, r d represents the radius of the wind turbine, a is the axial induction factor, r x is the wake radius at x downstream of the unit, k is the expansion coefficient, v x is the tail wind speed at a distance x downstream from the unit, v0 is the effective incoming wind speed of the fan;

[0054] The downstream unit may be in the wake area of ​​multiple upstream units. According to the energy square and wake superposition model, the effective incoming wind speed v of the downstream unit's wind wheel surface is j Expressed as:

[0055]

[0056] Where, v i is the effective incoming wind speed of the i-th unit, A ij is the overlapping area from the tail flow of the i-th unit to the j-th unit, v ij is the wake wind speed from the i-th unit to the j-th unit;

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

[0058]

[0059] Where, E WT (t) is the output of the wind farm at time t, n is the number of units, ρ is the air density, R is the radius of the wind turbine blade, C P is the wind energy utilization coefficient of the wind turbine;

[0060] For floating wind turbines, combined with the Jensen wake model and the impact of the floating wind turbine's six-degree-of-freedom motion on power output, the impact of the floating platform on the wind turbine's power characteristics is mainly manifested in the deflection of the rotor surface relative to the incoming wind direction, including pitch and bow; the relative motion between the rotor surface and the incoming wind, including two types of surge and pitch;

[0061] In this application, only the deflection motion of the wind rotor surface is considered. Only the two degrees of freedom are considered. After the floating platform motion causes the wind rotor surface to deflect, the actual wind rotor wind-catching area is expressed as:

[0062] S′=M y M z S

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

[0064] In summary, the aerodynamic power model of a floating wind turbine considering the wake is:

[0065]

[0066] E WT (t) represents the output of the wind farm at time t; n is the number of units, ρ is the air density, R is the radius of the wind turbine blade, C P is the wind energy utilization coefficient of the wind turbine; the pitch angle is α, the bow angle is β, v j Effective incoming wind speed on the wind wheel surface of the unit;

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

[0068] Photovoltaic modules are sensitive to changes in light intensity. Floating power generation places photovoltaic modules on a floating water surface. The mechanical oscillations caused by the ups and downs of the water surface will affect the temperature of the photovoltaic modules and the received solar radiation illumination.

[0069] In view of this, considering the output characteristics of floating photovoltaic modules under mechanical vibration conditions, a micro-power photovoltaic panel is selected to simplify the actual large-scale photovoltaic modules. When the module mechanically oscillates, its parameters such as the direct illumination area will also change periodically, causing its output characteristics to change.

[0070]

[0071] M is the effective illumination area of ​​the photovoltaic panel; is the oscillation angle, here it is 30°; f is the frequency of mechanical oscillation, here it is 1.7HZ; t is the time;

[0072] Figure 2This is a schematic diagram of the floating photovoltaic output considering mechanical oscillations of the present invention. Based on its power theoretical model, the factor of the periodic change in the effective illuminated area of ​​the photovoltaic panel when the photovoltaic panel oscillates mechanically is added. That is, the output power generated under static conditions is multiplied by the periodically changing direct illuminated area, resulting in the output relationship under mechanical oscillation.

[0073]

[0074] Where: E PV (t) is the output DC power of the photovoltaic module at time t; G t (t) is the radiation power under actual conditions; P pvr is the rated power of the photovoltaic module under standard test conditions; η pv is the power reduction coefficient of the photovoltaic module; β T is the power temperature coefficient of the photovoltaic module; G STC 、T STC is the radiation intensity and battery temperature under standard test conditions, T C is the actual battery temperature; it can be determined by the ambient temperature T e Calculated as follows, T STC —Battery temperature under standard test conditions (25°C). Where:

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

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

[0077] Although wind power and photovoltaic power have many outstanding advantages, they are highly uncertain and intermittent. Therefore, this embodiment uses a gas turbine as the energy supplement equipment, where the gas turbine output and scheduling model is:

[0078]

[0079] Where, E GT (t) is the electric power output of the gas power plant at time t, η GT is the power generation efficiency, q GT,fu is the mass flow rate of fuel, LHV fu is the lower calorific value of the fuel; δ GT is the start and stop state of the gas turbine, are the maximum and minimum outputs of the gas turbine, respectively;

[0080] (4) The model of the conversion side equipment system includes: the load includes water load and hydrogen load, so there must be equipment to produce fresh water and hydrogen in the system. Here, the osmosis 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:

[0081]

[0082] Where E SD (t) is the power consumption of the fading equipment at time t, η SD1 and η SD2 are the power consumption coefficient and desalination coefficient of the desalination equipment, V SDOUT (t) is the water output of the desalination equipment, V SDIN (t) is the water inflow of the desalination equipment, and are the minimum and maximum water inflow of the desalination equipment, is the minimum and maximum power consumption of the desalination equipment, δ SD It is the start and stop status of the dilution equipment;

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

[0084]

[0085] Where E HY (t) is the power consumption of the electrolysis equipment at time t, η HY1 and η HY2 They 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) is the water inflow of the electrolysis equipment, and are the minimum and maximum water inflow of the electrolysis equipment, is the minimum and maximum power consumption of the electrolysis equipment, δ HY It is the start and stop status of the electrolysis equipment;

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

[0087] This wind-solar-hydrogen storage system incorporates energy storage devices to improve system efficiency, combining peak-shaving and valley-filling functions with batteries, water storage, and hydrogen tanks. Operational constraints for the energy storage station include state-of-charge (SOC), charge-discharge (CD) status, and charge-discharge power. SOC constraints prevent deep charge and discharge, ensuring timely response and a longer lifespan for the energy storage system. Furthermore, after a 24-hour scheduling cycle, the battery's initial and final states must be identical. Charge-discharge state constraints prevent simultaneous charging and discharging of the battery during operation.

[0088] The power constraint of the energy storage station can prevent the heat generated by charging and discharging, which will also cause power loss. The scheduling model of the energy storage station is:

[0089] Where S BA (t) is the remaining battery power at time t, S BA (t-1) is the remaining battery power at time t-1, E BA,ch (t) is the electric power stored in the battery at time t, E BA,dch (t) is the electric power released by the battery at time t, Δt is the scheduling interval, η BA,ch is the battery charging efficiency, η BA,dch is the release efficiency of the battery; is the minimum and maximum capacity of the battery; E BA,ch (t), E BA,dch (t) represents the charging and discharging power of the energy storage station in time period t;

[0090] Respectively represent the minimum and maximum electric power of the charging state; Respectively represent the minimum electric power and maximum electric power in the discharge state, δ BA,ch , δ BA,dch They are charging and discharging states respectively;

[0091] The fresh water tank scheduling model is:

[0092]

[0093] Where S WA (t) is the remaining water in the water tank at time t, S WA (t-1) is the remaining water in the water tank at time t-1, V WA,ch (t) is the amount of water stored in the water tank at time t, V WA,dch (t) is the amount of water released from the water storage tank at time t, η WA,ch is the water storage efficiency of the water tank, η WA,dch is the release efficiency of the water storage tank; is the minimum and maximum water volume of the water storage tank; V WA,ch (t), V WA,dch (t) represents the amount of water filled and discharged in the water storage tank during the time period t; Respectively represent the minimum and maximum water volume of the water storage state; Respectively represent the minimum and maximum water volume in the water release state, δ WA,ch , δ WA,dch They are water storage and water release status;

[0094] The scheduling model of hydrogen storage tanks is:

[0095]

[0096] Where S HY (t) is the remaining hydrogen amount in the hydrogen storage tank at time t, S HY (t-1) is the remaining hydrogen amount in the hydrogen storage tank at time t-1, V HY,ch (t) is the amount of hydrogen stored in the hydrogen storage tank at time t, V HY,dch (t) is the amount of hydrogen released from the hydrogen storage tank at time t, η HY,ch is the hydrogen storage efficiency of the hydrogen storage tank, η HY,dch is the release efficiency of the hydrogen storage tank; is the minimum and maximum hydrogen volume of the hydrogen storage tank; V HY,ch (t), V HY,dch (t) represents the amount of hydrogen charged and discharged from the hydrogen storage tank in the time period t;

[0097] Respectively represent the minimum and maximum hydrogen amounts in the hydrogen storage state; Respectively represent the minimum and maximum amount of hydrogen released, δ BA,ch , δ BA,dch They are water storage and water release states respectively.

[0098] The electricity-fresh water-hydrogen scheduling model in the wind-solar-hydrogen storage system refers to achieving a balance between the supply and demand of electricity, water, and hydrogen networks through optimized scheduling and energy management in the integrated energy system of the offshore energy island. Specifically: the power balance involves the prediction and scheduling of floating wind power and photovoltaic power, as well as the storage and consumption of electricity. The electricity must not only meet the direct electricity demand of the island, but also provide power for water electrolysis and seawater desalination. The fresh water balance involves the operation of the seawater desalination equipment to meet the fresh water load demand. The water output of the seawater desalination device needs to match the fresh water demand on the island and be used for peak shaving and valley filling through water storage tanks. The hydrogen balance involves the operation of the water electrolysis hydrogen production device, as well as the storage and use of hydrogen to meet the hydrogen load demand during peak hours.

[0099] More specifically, the energy island wind-solar-hydrogen storage system scheduling optimization model 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] Among them, F1 is the rolling optimization cycle benefit objective function, F2 is the unit operation penalty cost objective function, s is the typical scenario after scenario reduction, π s is the probability of a typical scenario s, E(t), Q W (t), QG (t) is the amount of electricity, fresh water, and hydrogen sold by the system under the typical scenario s, C Esell 、C Wsell 、C Gsell are the prices of system electricity, water, and hydrogen energy, Δt is the scheduling interval, and λ gas ,λ gt 、E GT (t), E HY (t), E SD (t) are respectively the gas price, gas turbine penalty coefficient, gas turbine output under typical scenario s, power consumption of electrolysis equipment, power consumption of desalination equipment, and N S is the number of typical scenes, N p Optimize the cycle step size for rolling.

[0103] Constraints include power balance constraints, water network balance constraints, and hydrogen network balance constraints;

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

[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] Where, E WT (t) is the output of the wind farm at time t, E PV (t) is 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) is the power consumption of the ocean ranch at time t, E HY (t) is the power consumption of the electrolysis equipment at time t, E SD (t) is the power consumption of the fading equipment at time t, E BA,ch (t), E BA,dch (t) represents the charging and discharging power of the energy storage station in time period t;

[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] Where V SDOUT (t) is the water output of the desalination equipment, V WA,ch (t), V WA,dch (t) represents the amount of water filled and discharged in the water storage tank during the time period t; V WA,load (t) represents the water load.

[0111] The hydrogen network balance constraints are as follows:

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

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

[0114] The wind and light resources in the wind, solar, hydrogen and storage integrated energy system have strong uncertainty, and model predictive control can better cope with the randomness of renewable energy output. At the same time, in the energy island wind, solar, hydrogen and storage integrated energy system, random model predictive control can effectively reduce the impact of uncertain factors on the scheduling results. The random model predictive control scheduling strategy of the energy island wind, solar, hydrogen and storage integrated energy system proposed in this embodiment is as follows: Figure 3 As shown in Figure 1 , this scheduling strategy primarily consists of four components: a prediction model, scenario reduction, rolling optimization, and feedback correction. The prediction model is used to predict the outputs of both the supply and demand sides within the forecast horizon. Based on the resulting forecast results, scenario reduction techniques are used to filter the large number of generated initial scenarios. Then, based on the reduced scenarios, an optimization solution is performed within the forecast horizon, with the objective function of minimizing the system's expected operating costs. The first variable of the resulting control sequence is then used to control the system. Finally, the deviation between the system's actual output and the predicted output is fed back to correct the prediction model for the next step of rolling optimization.

[0115] In order to improve the anti-disturbance and prediction accuracy of the rolling optimization cycle, this embodiment, based on SMPC, innovatively proposes an adaptive control step technology in the rolling optimization link. The purpose of proposing this method is to reduce the prediction deviation between the dispatching instruction value and the actual output value of the wind, solar and hydrogen storage system; by tracking the actual load of the system, the adaptive variable step method is expected to be able to adaptively adjust the step size of the rolling optimization cycle according to the dispatching error, thereby improving the flexibility of the optimization link. By comparing the deviation between the system output dispatching instruction and the actual output of two adjacent rolling cycles, the deviation reference coefficient F is obtained. The specific formula is as follows: The goal of proposing this method is to reduce the prediction deviation between the dispatching instruction value and the actual output value of various energy storage power stations, gas turbine power stations and conversion measurement equipment; the adaptive variable step method is expected to be able to adaptively adjust the step size of the rolling optimization cycle according to the dispatching error, thereby improving the flexibility of the optimization link.

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

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

[0118] S42, based on the forecast data and actual operation 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 length N P ;

[0119] In some embodiments, the S42 specifically includes:

[0120] Based on the forecast data 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] Where, E WT (t) is the output of the wind farm at time t, E PV (t) is 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) is the power consumption of the ocean ranch at time t, E HY (t) is the power consumption of the electrolysis equipment at time t, E SD (t) is the power consumption of the fading equipment at time t, E BA,ch (t), E BA,dch (t) represents the charging and discharging power of the energy storage station in the time period t; VSDOUT (t) is the water output of the desalination equipment, V WA,ch (t), V WA,dch (t) represents the amount of water filled and discharged in the water storage tank during the time period t; V WA,load (t) represents water load; V HYOUT (t) is the outlet hydrogen volume of the electrolysis equipment, V HY,ch (t), V HY,dch (t) represents the amount of hydrogen charged and discharged from the hydrogen storage tank in the time period t, V 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 energy island wind, solar, and hydrogen storage systems at time t and t-1, respectively.

[0124] According to the deviation reference coefficient F, the rolling optimization cycle step N P Perform adaptive changes; when the scheduling prediction error becomes smaller, that is, when F is smaller, the prediction step size N can be increased P , in order to achieve a more global scheduling optimization; when the scheduling prediction error becomes larger, that is, when F is larger, shorten the prediction step length N P , reducing prediction errors and achieving 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 N P , use the variable weight method to distinguish the confidence of each period in the rolling optimization link;

[0127] This embodiment combines the idea of ​​variable weight coefficients on the basis of SMPC. In the traditional SMPC scheduling strategy, the confidence of the prediction results in each optimization cycle is constant. As the step length of the optimization cycle increases, the data at the end of the cycle will affect the prediction results, causing the prediction accuracy to gradually decrease. In the optimization cycle, the farther away from the current moment, the greater the error of the predicted result and the lower the reliability; but shortening the length of the optimization cycle is not conducive to achieving global optimization and cannot comprehensively consider the operating status at future moments. Therefore, it is necessary to reasonably set the length of the optimization cycle, by reducing the confidence at the end of the optimization cycle and increasing the confidence at the front end of the cycle, so as to achieve accurate scheduling control at the current moment. In order to reflect the difference in confidence of the prediction results at different moments in the simulation process, a variable weight method is introduced to distinguish the credibility of the results obtained with different prediction step lengths, that is, as the distance from the current moment increases, the weight coefficient gradually decreases. The specific formula of 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] Wherein, f1 represents the confidence of different step lengths. As the step length increases, the confidence 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 energy island wind, solar, hydrogen and storage system scheduling optimization model is solved to obtain the energy island wind, solar, hydrogen and storage system scheduling results for the period as prediction data, and a scheduling instruction is generated and sent to the energy island wind, solar, hydrogen and storage system for execution.

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

[0132] According to the load forecast value of the previous day, the renewable energy supply side includes wind power, photovoltaic power, and, the power supply mode is to consume all wind power and photovoltaic power, then adjust the gas turbine, and finally compensate by the battery. Therefore, the rolling optimization cycle profit objective function F1 (electricity sales, fresh water, hydrogen revenue) is the highest. Secondly, the system operation process requires the consumption of gas and some electricity that could have been sold (this part of electricity is used for seawater desalination and water electrolysis to produce hydrogen), and the frequent power fluctuations of the gas unit are not conducive to the safe and stable operation of the unit. Therefore, the unit operation penalty cost is introduced to obtain the objective function F2. Δt is the scheduling interval, which is 1 hour in this embodiment. The initial scenario in this embodiment is 100, and Ns and Np are taken as 10 and 8 respectively.

[0133] Feedback correction refers to the process of adjusting the system to bring it closer to the desired output by comparing the difference between actual output and desired output and feeding that difference back to the input. The wind, solar, and hydrogen storage system model takes into account the uncertainties on both the supply and demand sides. It is necessary to add deviations to the day-ahead forecast data to simulate the dispatch results of the wind, solar, and hydrogen storage system. The actual dispatch results are compared and corrected with the forecast model through the feedback link to reduce the model's prediction error on a time scale. This measure reduces the uncertainty of system forecasts and improves the model's prediction accuracy by assuming a normally distributed error of N(0,10) for increasing wind speed and sunlight and N(0,5) for increasing load.

[0134] Simulation verification experiment: Figure 3 As shown in the figure, a stochastic model predictive control scheduling model for a wind-solar-hydrogen storage system with an adaptive variable-step size is established to verify and analyze offshore wind-solar-hydrogen storage scheduling. The region houses a floating photovoltaic power station, a floating wind farm, a seawater desalination and hydrogen production station, a marine ranch, and three energy storage power stations. The region sells electricity to coastal areas at a price calculated based on local time-of-use electricity prices, as shown in Table 1. The wind farm has a rated installed capacity of 264 MW (33 x 8 MW), the photovoltaic power station has a rated capacity of 150 MW, the gas turbine power station has a rated capacity of 300 MW, and the energy storage station has a capacity of 45 MW / 150 MW·h.

[0135] Table 1: Time-of-use electricity price list in a certain place

[0136]

[0137]

[0138] Table 2: Time-of-use water price list in a certain place

[0139]

[0140] Table 3: Time-sharing hydrogen price list in a certain place

[0141]

[0142] 1) Constructing a deviation prediction model that takes into account the uncertainty of wind and solar power:

[0143] Generate a deviation prediction model. The prediction error ΔP of wind power, photovoltaic output and load obeys the normal distribution N~(μ,σ 2 ) where the mean μ is 0, and the variance of wind speed, light, and various loads is σ 2 They are 10, 10, and 5 respectively.

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

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

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

[0147] ④ Find the sample corresponding to the minimum value in the probability distance and delete it, and add the probability of the sample to the one closest to it.

[0148] Sample Office;

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

[0150] 3) Design of 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] ②Use the variable weight method to distinguish the confidence levels of each period in the rolling optimization phase;

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

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

[0155] ④ Bring the actual operation data after executing the scheduling instruction into step ①, and realize

[0156] Feedback correction;

[0157] ⑤Simulation solution:

[0158] Figure 4 This is a schematic diagram of a rolling optimization framework according to an embodiment of the present invention; Figure 5The scheduling results of the wind-solar-hydrogen storage system were 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 the wind-solar-hydrogen storage system in a specific embodiment of the present invention; Figure 7 This is a diagram showing the hydrogen network scheduling results of the wind-solar-hydrogen storage system in a specific embodiment of the present invention; Figure 8 This is a diagram showing the change in the system prediction step size Np in a specific embodiment of the present invention.

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

[0160] The present invention adopts the root mean square deviation I RMSE The prediction performance of the proposed adaptive variable step-size SMPC algorithm was evaluated, and the prediction deviation evaluation indicators proposed for horizontal comparison of the prediction deviations of different model predictive control strategies are as follows:

[0161]

[0162] In addition, in order to prove 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: variable weight stochastic model predictive control method;

[0165] 3. Adaptive variable step size SMPC: that is, the adaptive variable step size stochastic model predictive control method proposed in this application.

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

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

[0168] Strategy Dispatch income (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] It can be quantitatively seen from the table that 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] According to the comparison of indicators, the adaptive variable step-size SMPC strategy proposed in this invention and the adaptive variable step-size SMPC algorithm proposed in this application have increased the dispatch revenue by 0.85% compared with the traditional SMPC algorithm, which has improved the dispatch economy to a certain extent; compared with the variable weight SMPC algorithm, the dispatch cost has decreased slightly, but the decrease is only 0.009%. Its prediction deviation is reduced by 11.2% compared with the traditional SMPC and 24.1% compared with the variable weight SMPC strategy. The adaptive variable step size can adjust the step size according to the prediction deviation, effectively solving the prediction deviation caused by the constant weight and step size problem of the rolling optimization cycle. The results show that the adaptive variable step-size SMPC strategy proposed in this application can effectively reduce the prediction deviation between the dispatch instructions and the actual output of the wind, solar, hydrogen and storage system.

[0173] Example 2

[0174] In a second aspect, this embodiment provides an energy island wind-solar-hydrogen storage system scheduling device 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 perform the method according to embodiment 1.

[0177] Example 3

[0178] In a third aspect, this embodiment provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, the method described in Example 1 is implemented.

[0179] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0180] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0181] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0182] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0183] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for scheduling an energy island wind-solar-hydrogen storage system based on adaptive variable step size, characterized in that: include: S1. Obtain wind power, photovoltaic output and load forecast values, and combine them with the deviation prediction model to obtain the actual values ​​of wind power, photovoltaic output and load; S2. Generate multiple initial scenarios using Monte Carlo simulation based on wind power, photovoltaic output, and actual load values; S3. Use the synchronous back-substitution scenario reduction technique to screen multiple initial scenarios and select typical scenarios that satisfy the overall distribution of random variables and the probability of each typical scenario; S4. Based on typical scenarios and the probabilities of each typical scenario, the energy island wind-solar-hydrogen-storage system scheduling optimization model is solved using the adaptive variable step size SMPC rolling optimization method to obtain the energy island wind-solar-hydrogen-storage system scheduling results for the current period.

2. The method according to claim 1, characterized in that The deviation prediction model includes: The forecast error ΔP of wind power, photovoltaic output and load obeys the normal distribution N~(μ,σ 2 ), where the mean μ is 0, and the variance of wind power, photovoltaic output, and load is σ 2 They are 10, 10, and 5 respectively. The formula for predicting the offset is: Among them, f(ΔP) is the predicted offset of wind power, photovoltaic power and various loads.

3. The method according to claim 1, characterized in that The synchronous back-substitution scenario reduction technology is used to screen multiple initial scenarios and select typical scenarios that meet the overall distribution of random variables and the probability of each typical scenario, including: S31. The generated N initial scenarios are used as sample scenarios. Each sample scenario includes error sequence prediction values ​​of wind power, photovoltaic output, and load in the future prediction time domain. The initial probability of each sample scenario is 1 / N. S32, performing Gaussian normalization processing on the N sample scenes to eliminate the influence of parameter amplitude changes on the results; S33. For each sample scene x i , calculate and other sample scenes x j The probability distance; Where, d ij Represents a sample scene x i and sample scene x j The probability distance between For the sample scene x i The probability of , N is the number of sample scenes; S34. Find the sample scene x corresponding to the minimum value in the required probability distance i And delete it, but add the probability of the sample to the sample scene x closest to it j At this time, the updated sample scene x j Probability for: Where, The sample scenes x before and after the update are j probability; S35. Repeat steps S32-S34 until the number of updated sample scenes is reduced to the required number of typical scenes Ns. The sample scenes at this time are used as typical scenes, and the probability of each typical scene is calculated.

4. The method according to claim 1, wherein The energy island wind-solar-hydrogen storage system scheduling optimization model is constructed based on the energy island wind-solar-hydrogen storage system model. The energy island wind-solar-hydrogen storage system model 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 floating wind farm wake model is a wind farm model that considers the Jensen wake model. The aerodynamic power model of the floating wind turbine considering the wake is: E WT (t) represents the output of the wind farm at time t; n is the number of units, ρ is the air density, R is the radius of the wind turbine blade, C P is the wind energy utilization coefficient of the wind turbine; the pitch angle is α, the bow angle is β, v j Effective incoming wind speed on the wind wheel surface of the unit; (2) The floating photovoltaic power station scheduling model includes: Where: E PV (t) is the output DC power of the photovoltaic module at time t; G t (t) is the radiation power under actual conditions; P pvr is the rated power of the photovoltaic module under standard test conditions; η pv is the power reduction coefficient of the photovoltaic module; β T is the power temperature coefficient of the photovoltaic module; G STC 、T STC is the radiation intensity and battery temperature under standard test conditions, T c is the battery temperature under actual conditions; M is the effective illumination area of ​​the photovoltaic panel; is the oscillation angle; f is the frequency of mechanical oscillation; (3) Gas turbine power plant model includes: Where, E GT (t) is the electric power output of the gas power plant at time t, η GT is the power generation efficiency, q GT,fu is the mass flow rate of fuel, LHV fu is the lower calorific value of the fuel; δ GT is the start and stop state of the gas turbine, are the maximum and minimum outputs of the gas turbine, respectively; (4) The model of the conversion side equipment system includes: the load includes water load and hydrogen load, so there must be equipment to produce fresh water and hydrogen in the system. Here, the osmosis 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: Where E SD (t) is the power consumption of the fading equipment at time t, η SD1 and η SD2 are the power consumption coefficient and desalination coefficient of the desalination equipment, V SDOUT (t) is the water output of the desalination equipment, V SDIN (t) is the water inflow of the desalination equipment, and are the minimum and maximum water inflow of the desalination equipment, is the minimum and maximum power consumption of the desalination equipment, δ SD It is the start and stop status of the dilution equipment; The output and scheduling model for hydrogen production equipment is: Where E HY (t) is the power consumption of the electrolysis equipment at time t, η HY1 and η HY2 They 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) is the water inflow of the electrolysis equipment, and are the minimum and maximum water inflow of the electrolysis equipment, is the minimum and maximum power consumption of the electrolysis equipment, δ HY It is the start and stop status of the electrolysis equipment; (5) Various energy storage station scheduling models include: The energy storage power station dispatch model is: Where S BA (t) is the remaining battery power at time t, S BA (t-1) is the remaining battery power at time t-1, E BA,ch (t) is the electric power stored in the battery at time t, E BA,dch (t) is the electric power released by the battery at time t, Δt is the scheduling interval, η BA,ch is the battery charging efficiency, η BA,dch is the release efficiency of the battery; is the minimum and maximum capacity of the battery; E BA,ch (t), E BA,dch (t) represents the charging and discharging power of the energy storage station in time period t; Respectively represent the minimum and maximum electric power of the charging state; Respectively represent the minimum electric power and maximum electric power in the discharge state, δ BA,ch , δ BA,dch They are charging and discharging states respectively; The fresh water tank scheduling model is: Where S WA (t) is the remaining water in the water tank at time t, S WA (t-1) is the remaining water in the water tank at time t-1, V WA,ch (t) is the amount of water stored in the water tank at time t, V WA,dch (t) is the amount of water released from the water storage tank at time t, η WA,ch is the water storage efficiency of the water tank, η WA,dch is the release efficiency of the water storage tank; is the minimum and maximum water volume of the water storage tank; V WA,ch (t), V WA,dch (t) represents the amount of water filled and discharged in the water storage tank during the time period t; Respectively represent the minimum and maximum water volume of the water storage state; Respectively represent the minimum and maximum water volume in the water release state, δ WA,ch , δ WA,dch They are water storage and water release status; The scheduling model of hydrogen storage tanks is: Where S HY (t) is the remaining hydrogen amount in the hydrogen storage tank at time t, S HY (t-1) is the remaining hydrogen amount in the hydrogen storage tank at time t-1, V HY,ch (t) is the amount of hydrogen stored in the hydrogen storage tank at time t, V HY,dch (t) is the amount of hydrogen released from the hydrogen storage tank at time t, η HY,ch is the hydrogen storage efficiency of the hydrogen storage tank, η HY,dch is the release efficiency of the hydrogen storage tank; is the minimum and maximum hydrogen volume of the hydrogen storage tank; V HY,ch (t), V HY,dch (t) represents the amount of hydrogen charged and discharged from the hydrogen storage tank in the time period t; Respectively represent the minimum and maximum hydrogen amounts in the hydrogen storage state; Respectively represent the minimum and maximum amount of hydrogen released, δ BA,ch , δ BA,dch They are water storage and water release states respectively.

5. The energy island wind-solar-hydrogen storage system scheduling method based on adaptive variable step size according to claim 1 is characterized in that: The energy island wind-solar-hydrogen storage system scheduling optimization model 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: F=max(F1-F2) Among them, F1 is the rolling optimization cycle benefit objective function, F2 is the unit operation penalty cost objective function, s is the typical scenario after scenario reduction, π s is the probability of a typical scenario s, E(t), Q W (t), Q G (t) is the amount of electricity, fresh water, and hydrogen sold by the system under the typical scenario s, C Esell 、C Wsell 、C Gsell are the prices of system electricity, water, and hydrogen energy, Δt is the scheduling interval, and λ gas ,λ gt 、E GT (t), E HY (t), E SD (t) are respectively the gas price, gas turbine penalty coefficient, gas turbine output under typical scenario s, power consumption of electrolysis equipment, power consumption of desalination equipment, and N S is the number of typical scenes, N p Optimize the cycle step size for rolling; Constraints include power balance constraints, water network balance constraints, and hydrogen network balance constraints; The electric power balance constraint is expressed as: E WT (t)+E PV (t)+E GT (t)+E BA,dch (t) =E load (t)+E SF (t)+E HY (t)+E SD (t)+E BA,ch (t) Where, E WT (t) is the output of the wind farm at time t, E PV (t) is 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) is the power consumption of the ocean ranch at time t, E HY (t) is the power consumption of the electrolysis equipment at time t, E SD (t) is the power consumption of the fading equipment at time t, E BA,ch (t), E BA,dch (t) represents the charging and discharging power of the energy storage station in time period t; The water network balance constraints are as follows: V SDOUT (t)-V WA,ch (t)+V WA,dch (t)=V WA,load (t) Where V SDOUT (t) is the water output of the desalination equipment, V WA,ch (t), V WA,dch (t) represents the amount of water filled and discharged in the water storage tank during the time period t; V WA,load (t) represents water load; The hydrogen network balance constraints are as follows: V HYOUT (t)-V HY,ch (t)+V HY,dch (t)=V HY,load (t) Where V HYOUT (t) is the outlet hydrogen volume of the electrolysis equipment, V HY,ch (t), V HY,dch (t) represents the amount of hydrogen charged and discharged from the hydrogen storage tank in the time period t, V HY,load (t) represents the hydrogen load.

6. The energy island wind-solar-hydrogen storage system scheduling method based on adaptive variable step size according to claim 1 is characterized in that: The energy island wind-solar-hydrogen-storage system scheduling optimization model is solved using the adaptive variable-step-size SMPC rolling optimization method to obtain the energy island wind-solar-hydrogen-storage system scheduling results for the current cycle, including: S41. Obtain actual operating data of the energy island wind, solar, and hydrogen storage system after executing the last cycle scheduling instruction; S42, based on the forecast data and actual operation 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 length N P ; S43, based on the rolling optimization cycle step N P , use the variable weight method to distinguish the confidence of each period in the rolling optimization link; S44. Based on typical scenarios and the probabilities of each typical scenario, the energy island wind, solar, hydrogen and storage system scheduling optimization model is solved to obtain the energy island wind, solar, hydrogen and storage system scheduling results for the period as prediction data, and a scheduling instruction is generated and sent to the energy island wind, solar, hydrogen and storage system for execution.

7. The energy island wind-solar-hydrogen storage system scheduling method based on adaptive variable step size according to claim 6 is characterized in that The S42 includes: Based on the forecast data 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; Where, E WT (t) is the output of the wind farm at time t, E PV (t) is 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) is the power consumption of the ocean ranch at time t, E HY (t) is the power consumption of the electrolysis equipment at time t, E SD (t) is the power consumption of the fading equipment at time t, E BA,ch (t), E BA,dch (t) represents the charging and discharging power of the energy storage station in the time period t; V SDOUT (t) is the water output of the desalination equipment, V WA,ch (t), V WA,dch (t) represents the amount of water filled and discharged in the water storage tank during the time period t; V WA,load (t) represents water load; V HYOUT (t) is the outlet hydrogen volume of the electrolysis equipment, V HY,ch (t), V HY,dch (t) represents the amount of hydrogen charged and discharged from the hydrogen storage tank in the time period t, V HY,load (t) represents the hydrogen load; 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; δ(t) and δ(t-1) represent the deviation reference coefficients of the wind, solar, and hydrogen storage systems of the energy island at time t and t-1, respectively; According to the deviation reference coefficient F, the rolling optimization cycle step N P Make adaptive changes; 8. The energy island wind-solar-hydrogen storage system scheduling method based on adaptive variable step size according to claim 6 is characterized in that: The S43 includes: f1=diag(φ i-1 ),0<φ<1,i=1,...,N p , Where f1 represents the confidence level of different step sizes. As the step size increases, the confidence level decreases. φ is a constant between 0 and 1, and i represents different step sizes.

9. An energy island wind-solar-hydrogen storage system scheduling device based on adaptive variable step size, characterized in that: including processors 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 8.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images