Hydrogen production power dual-time scale planning method for green ionization network hydrogen production system

Abstract

The invention relates to a double-time-scale planning method for hydrogen production power of a green ionization network hydrogen production system. The method comprises hour-level hydrogen production power planning based on electric quantity prediction and minute-level hydrogen production power planning based on power prediction. Fitting renewable energy hour-level generating capacity prediction data by adopting a step curve, optimizing a time interval by taking reduction of step times as a target, and ensuring that the time integral quantity of the hour-level generating capacity prediction curve is equal to that of the fitting step curve; and on the basis of the renewable energy minute-level generation power prediction curve, performing in-stage second-round fitting on the step curve after the first-round optimization, and also taking the minimization of the step number as a target optimization time interval. According to the method, under the condition that the electric power and electric quantity balance constraint condition is met, the number of times of adjusting the power of the electrolytic cell is effectively reduced through layered progressive optimization, and the operation stability and economical efficiency of the hydrogen production system under the wind-light fluctuation scene are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of hydrogen production by electrolysis of water using renewable energy, and in particular to a dual-time-scale planning method for hydrogen production power of a green electricity off-grid hydrogen production system. Background Art

[0002] Currently, renewable energy sources such as wind power and photovoltaics have achieved large-scale application. However, with the accelerated development of scenarios such as the Shagohuang New Energy Base and deep-sea renewable energy islands, centralized renewable energy sources face significant challenges in their absorption due to weak or unavailable grids. Hydrogen energy, as a clean and efficient secondary energy source, offers multi-dimensional regulation advantages in terms of energy, time, and space. Using water electrolysis technology to convert random, intermittent, and fluctuating renewable energy into hydrogen energy has become a global technological path to addressing the challenges of large-scale renewable energy absorption, sparking extensive research and practice. Against this backdrop, achieving efficient coupling of fluctuating renewable energy sources with water electrolysis hydrogen production systems through power optimization has become a core technical bottleneck in improving hydrogen energy regulation efficiency.

[0003] Existing methods for electrolyzer hydrogen production power planning can be categorized into two types: grid-supported multi-dimensional collaborative optimization and off-grid threshold dynamic start-stop. In the grid-supported model, a global optimization model for the green electricity hydrogen production system is constructed by integrating multi-dimensional constraints such as wind and solar power forecasts, hydrogen production equipment characteristics, energy storage dynamics, and lifespan models. Intelligent algorithms are used to generate hydrogen production power curves that respond to renewable energy fluctuations and grid dispatch requirements. In off-grid scenarios, electrolyzer hydrogen production power planning faces significant technical challenges due to the operational characteristics of the electrolyzer and the capacity limitations of the energy storage system.

[0004] The researchers proposed a dynamic start-stop strategy for off-grid hydrogen production with power threshold comparison, combining alkaline water electrolysis and solid oxide water electrolysis. However, frequent start-stopping will accelerate the degradation of the electrolyzer membrane and catalyst. Furthermore, the researchers proposed a rotation optimization control strategy for off-grid wind power hydrogen production using alkaline water electrolysis. The start-stop times were optimized by periodically rotating the operating status of the electrolyzer. However, there is a risk of decreased electrolyzer efficiency and increased hydrogen concentration in oxygen during low-power operation. Furthermore, the researchers proposed a large-scale off-grid wind power hydrogen production multi-electrolyzer switching scheduling strategy that takes into account efficiency and start-stop characteristics. By optimizing the operating mode of the electrolyzer, the start-stop times of the electrolyzer were reduced by 93.5%, and the hydrogen production was increased by 44.18%.

[0005] It can be seen that the optimization of the number of starts and stops essentially alleviates the power mutation efficiency loss during the startup phase of the electrolyzer, but its optimization dimension is limited to the threshold control at the startup time, and does not consider the real-time coupling of power fluctuations and efficiency during the operation phase, resulting in the electrolyzer efficiency loss under the dynamic fluctuation scenario of wind and solar power still being high.

[0006] The researchers proposed a graded diagnosis and treatment strategy for off-grid hydrogen production using alkaline water electrolysis and wind power. By balancing the loads of multiple tanks, the low-power operation time of the electrolyzer can be effectively reduced, but it lacks flexibility in severe weather or severe power fluctuations. Furthermore, the researchers proposed an efficiency optimization control method for parallel hydrogen production devices, which uses a mixed integer linear programming algorithm to optimize the optimal power distribution values of different parallel hydrogen production devices. Compared with the operating efficiency under graded switching and power balancing, the proposed efficiency optimization control method can effectively improve the operating efficiency of the parallel system under fluctuating power. Furthermore, the researchers proposed a rule-based coordinated control strategy for multiple electrolyzers for off-grid hydrogen production using PEM. By optimizing steady-state / fluctuating power input, the service life of the electrolyzer can be extended, but the capacity configuration of the energy storage equipment is relatively high.

[0007] However, the existing hydrogen production power planning strategy is not adequately adapted to the coupling characteristics of hourly energy trends and minute-level power mutations of renewable energy, and does not consider the impact of the dynamic power switching process on electrolyzer control and hydrogen production efficiency. This makes it difficult to coordinate operating economy and equipment durability, and the demand for energy storage redundancy remains high.

[0008] Therefore, it is necessary to study a hydrogen production power planning method for green electricity off-grid hydrogen production systems. By considering the impact of electrolyzer power switching and dynamic response constraints, the number of power switching times can be reduced and the response time can be optimized, thereby enhancing the dynamic matching ability of the system with renewable energy fluctuations, achieving efficient energy conversion and optimal economic performance throughout the entire life cycle. Summary of the Invention

[0009] The purpose of the present invention is to provide a dual-time scale planning method for hydrogen production power of a green electricity off-grid hydrogen production system.

[0010] The purpose of the present invention can be achieved by the following technical solutions:

[0011] A dual-time-scale planning method for hydrogen production power of a green off-grid hydrogen production system, including hourly hydrogen production power planning based on electricity forecast and minute-level hydrogen production power planning based on power forecast,

[0012] The hourly hydrogen production power planning based on power forecast includes: using a step curve to fit the hourly power generation forecast data of renewable energy based on an hourly step size to obtain a first step curve, and optimizing the first step curve to obtain a second step curve with the constraint that the integrated areas of the curves before and after optimization are the same and the goal of reducing the number of steps;

[0013] The minute-level hydrogen production power planning based on power prediction includes: for each step of the second step curve, the step curve is used to fit the minute-level power generation prediction data of renewable energy with a step size of minutes to obtain a third step curve, and the third step curve is optimized with the constraint that the integral area of the curve before and after optimization is the same and with the goal of reducing the number of steps to obtain a fourth step curve, and all the fourth step curves are spliced in time sequence to obtain a hydrogen production power curve.

[0014] The step curve is used to fit the hourly power generation forecast data of renewable energy with a step length of one hour to obtain a first step curve, which includes:

[0015] Obtain hourly renewable energy power generation forecast data for the target planning period;

[0016] Drawing a first prediction curve based on hourly power generation forecast data of renewable energy;

[0017] Generate the first fitting interval with a step size of hours;

[0018] Based on the first fitting interval, a step curve is used to fit the first prediction curve to obtain a first step curve, wherein each first fitting interval is fitted into the same step.

[0019] The step of optimizing the first step curve to obtain the second step curve with the constraint that the integral areas of the curves before and after optimization are the same and with the goal of reducing the number of steps further includes:

[0020] According to the parameter constraints and historical data statistics of renewable energy power generation equipment, the power difference threshold for adjacent steps to be merged is set;

[0021] Traversal step: traverse all adjacent steps, calculate the absolute value of the power difference between each adjacent step, screen out the two steps with the smallest absolute value of the power difference, and determine whether the absolute value of the power difference between the two screened steps is less than the power difference threshold. If so, perform the step merging step, otherwise obtain the second step curve;

[0022] In the step of merging steps, if the absolute value of the power difference between the two steps obtained by screening is less than the power difference threshold, the two steps obtained by screening are merged into one step, and the power value of the merged step is determined based on the constraint that the integral area of the curves before and after the merger is the same, and then the traversal step is returned.

[0023] The step curve is used to fit the renewable energy minute-level power generation forecast data with a step length of minutes to obtain a third step curve, including:

[0024] Obtain the minute-level power generation forecast data of renewable energy in the time interval corresponding to any step in the second step curve;

[0025] Drawing a second prediction curve based on the obtained renewable energy minute-level power generation prediction data;

[0026] Generate the second fitting interval with a step size of minutes;

[0027] Based on the second fitting interval, a step curve is used to fit the second prediction curve to obtain a third step curve, wherein each second fitting interval is fitted into the same step.

[0028] The method of optimizing the third step curve to obtain the fourth step curve with the constraint that the integral areas of the curves before and after optimization are the same and with the goal of reducing the number of steps includes:

[0029] According to the parameter constraints and historical data statistics of renewable energy power generation equipment, the power difference threshold for adjacent steps to be merged is set;

[0030] Traversal step: traverse all adjacent steps, calculate the absolute value of the power difference between each adjacent step, screen out the two steps with the smallest absolute value of the power difference, and determine whether the absolute value of the power difference between the two screened steps is less than the power difference threshold. If so, perform the step merging step, otherwise obtain the fourth step curve;

[0031] In the step of merging steps, if the absolute value of the power difference between the two steps obtained by screening is less than the power difference threshold, the two steps obtained by screening are merged into one step, and the power value of the merged step is determined based on the constraint that the integral area of the curves before and after the merger is the same, and then the traversal step is returned.

[0032] The hydrogen production power curve is obtained by splicing all the fourth-step curves in time sequence, including:

[0033] All fourth-step curves are spliced together in time sequence to obtain the transition power curve;

[0034] The hydrogen production power curve is obtained by inserting the linear transition power into the adjacent steps.

[0035] The linear transition power is determined according to the maximum power change rate of the renewable energy power generation equipment and the principle of constant integral area.

[0036] The renewable energy sources include at least wind energy and solar energy.

[0037] A dual-time-scale planning device for hydrogen production power of a green electricity off-grid hydrogen production system comprises a memory, a processor, and a program stored in the memory. When the processor executes the program, the method described above is implemented.

[0038] A storage medium stores a program, which implements the above method when executed.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] 1. Reduce the number of electrolyzer hydrogen production power adjustments and extend equipment life. By fitting hourly and minute-level step curves, the frequently changing power generation power is converted into step-like hydrogen production power instructions. This can reduce the mechanical impact of power step changes on the equipment and reduce losses caused by frequent switching of the control system.

[0041] 2. Enhance the operational stability of the hydrogen production system under fluctuating wind and solar power scenarios. Hourly hydrogen production power planning ensures that the time integral of the predicted power generation curve and the fitted step curve is equal. Minute-by-minute hydrogen production power planning smooths short-term power fluctuations and prevents electrolyzers from triggering protective shutdowns due to instantaneous power exceeding the limit.

[0042] 3. Improve renewable energy absorption capacity and reduce power curtailment. By fitting hourly and minute-level step curves, we ensure a strict time-integrated match between power generation and hydrogen production power, reduce wind and solar curtailment caused by power fluctuations, and maximize the hydrogen conversion efficiency of wind and solar power. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 1 is a flow chart of a dual-time-scale planning method for hydrogen production power in a green electricity off-grid hydrogen production system according to an embodiment of the present invention;

[0044] Figure 2 is a flow chart of a step curve time interval optimization method according to an embodiment of the present invention;

[0045] Figure 3 This is a structural diagram of the wind power off-grid hydrogen production system in Example 1 of the present invention;

[0046] Figure 4 1 is a schematic diagram of dual-time-scale planning of hydrogen production power of an off-grid wind power hydrogen production system in the first embodiment of the present invention;

[0047] Figure 5 This is a schematic structural diagram of a photovoltaic off-grid hydrogen production system in Example 2 of the present invention;

[0048] Figure 6 This is a schematic diagram of dual-time scale planning of hydrogen production power for a photovoltaic off-grid hydrogen production system in the second embodiment of the present invention. DETAILED DESCRIPTION

[0049] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0050] Figure 1This is a flow chart illustrating a dual-timescale planning method for hydrogen production power in a green electricity off-grid hydrogen production system according to an embodiment of the present invention. The method includes hourly hydrogen production power planning based on electricity forecasts and minute-by-minute hydrogen production power planning based on power forecasts. Renewable energy sources include at least wind and solar energy.

[0051] The hourly hydrogen production power planning based on power forecast includes: using a step curve to fit the hourly power generation forecast data of renewable energy with an hourly step size to obtain a first step curve, and then optimizing the first step curve to obtain a second step curve with the constraint that the integrated area of the curves before and after optimization is the same and the goal of reducing the number of steps;

[0052] The minute-level hydrogen production power planning based on power prediction includes: for each step of the second-step curve, the step curve is used to fit the minute-level power generation forecast data of renewable energy with a step size of minutes to obtain the third-step curve, and the third-step curve is optimized to obtain the fourth-step curve with the constraint that the integral area of the curve before and after optimization is the same and the goal of reducing the number of steps, and all the fourth-step curves are spliced in time sequence to obtain the hydrogen production power curve.

[0053] Figure 2 This is a flowchart of the step curve time interval optimization method in an embodiment of the present invention. First, based on device constraint analysis and historical data statistics, a power difference threshold is set for merging adjacent power segments. Second, the adjacent power segments are traversed and the absolute power difference is calculated, screening out the adjacent power segments with the smallest absolute power difference that is below the set threshold. Then, the power values of adjacent time periods are iteratively adjusted to merge the time intervals, and the screening and merging operations are repeated until the difference values of all adjacent power segments exceed the set threshold. Finally, a linear transition power is inserted between adjacent power segments to ensure that it meets the slope conditions of the operating constraints. After calculating the total energy deviation, the power curve is globally shifted to achieve accurate matching of the total energy amount.

[0054] Specifically, in this embodiment, using a step curve to fit the hourly power generation forecast data of renewable energy with a step size of one hour to obtain a first step curve includes:

[0055] Obtain hourly renewable energy power generation forecast data for the target planning period;

[0056] Drawing a first prediction curve based on hourly power generation forecast data of renewable energy;

[0057] Generate the first fitting interval with a step size of hours;

[0058] Based on the first fitting interval, a step curve is used to fit the first prediction curve to obtain a first step curve, wherein each first fitting interval is fitted into the same step.

[0059] In addition, the first step curve is optimized to obtain the second step curve with the constraint that the integral areas of the curves before and after optimization are the same and the goal of reducing the number of steps, and the following is also included:

[0060] According to the parameter constraints and historical data statistics of renewable energy power generation equipment, the power difference threshold for adjacent steps to be merged is set;

[0061] Traversal step: traverse all adjacent steps, calculate the absolute value of the power difference between each adjacent step, screen out the two steps with the smallest absolute value of the power difference, and determine whether the absolute value of the power difference between the two screened steps is less than the power difference threshold. If so, perform the step merging step, otherwise obtain the second step curve;

[0062] In the step of merging steps, if the absolute value of the power difference between the two steps obtained by screening is less than the power difference threshold, the two steps obtained by screening are merged into one step, and the power value of the merged step is determined based on the constraint that the integral area of the curves before and after the merger is the same, and then the traversal step is returned.

[0063] Similarly, the third step curve is obtained by fitting the renewable energy minute-level power generation forecast data with a step size of one minute, including:

[0064] Obtain the minute-level power generation forecast data of renewable energy in the time interval corresponding to any step in the second step curve;

[0065] Drawing a second prediction curve based on the obtained renewable energy minute-level power generation prediction data;

[0066] Generate the second fitting interval with a step size of minutes;

[0067] Based on the second fitting interval, a step curve is used to fit the second prediction curve to obtain a third step curve, wherein each second fitting interval is fitted into the same step.

[0068] Similarly, with the constraint that the integral area of the curves before and after optimization is the same and the goal of reducing the number of steps, the third step curve is optimized to obtain the fourth step curve, including:

[0069] According to the parameter constraints and historical data statistics of renewable energy power generation equipment, the power difference threshold for adjacent steps to be merged is set;

[0070] Traversal step: traverse all adjacent steps, calculate the absolute value of the power difference between each adjacent step, screen out the two steps with the smallest absolute value of the power difference, and determine whether the absolute value of the power difference between the two screened steps is less than the power difference threshold. If so, perform the step merging step, otherwise obtain the fourth step curve;

[0071] In the step of merging steps, if the absolute value of the power difference between the two steps obtained by screening is less than the power difference threshold, the two steps obtained by screening are merged into one step, and the power value of the merged step is determined based on the constraint that the integral area of the curves before and after the merger is the same, and then the traversal step is returned.

[0072] Finally, in this embodiment, all fourth-step curves are spliced in time sequence to obtain a hydrogen production power curve, including:

[0073] All fourth-step curves are spliced together in time sequence to obtain the transition power curve;

[0074] The hydrogen production power curve is obtained by inserting the linear transition power into the adjacent steps.

[0075] The linear transition power is determined based on the maximum power change rate of the renewable energy power generation equipment and the principle of constant integral area.

[0076] Specifically, the following uses wind power off-grid hydrogen production systems and photovoltaic off-grid hydrogen production systems as typical application scenarios to illustrate the implementation process of the hydrogen production power dual-time scale planning method and the step curve time interval optimization method.

[0077] Example 1: Dual-time scale planning of hydrogen production power for an off-grid wind power hydrogen production system.

[0078] like Figure 3 As shown, the wind power off-grid hydrogen production system provided in this embodiment consists of a wind turbine, an energy storage battery, a hydrogen production power supply, and a PEM electrolyzer. The wind turbine uses a 4MW wind turbine from the Huaneng Rudong Baxianjiao offshore wind power project, which is connected to the DC bus through an AC / DC converter, with a real-time power fluctuation range of 0-4.56MW; the energy storage battery is configured with a total capacity of 0.5MW, and the SOC effective working range is 20%-80%; the hydrogen production power supply adopts a three-level IGBT topology, supports a wide voltage input of 150-1000V, a DC ripple ≤1%, and a conversion efficiency ≥97%; the electrolyzer has a rated hydrogen production capacity of 800Nm 3 / h, operating pressure 3.0MPa, and load adjustment range 10%-120%. This diagram clearly shows the energy interaction paths of each component, providing a topological basis for constructing power balance equations and equipment operation constraint models.

[0079] S11. Obtain and analyze hourly power generation forecast data for the current cycle.

[0080] The current cycle refers to the time period during which hydrogen production power planning and scheduling must be implemented. Power generation and power generation forecasts must be completed before the start of the cycle. This embodiment sets October 1, 2022, from 0:00 AM to 12:00 PM as a time period for the off-grid wind power hydrogen production system. The next cycle forecast must be initiated at 11:50 PM at the end of the previous cycle. For example, the forecast task will be executed at 11:50 PM on September 30, 2022. This mechanism reserves sufficient response time by pre-positioning the forecast window, providing reliable input parameters for multi-timescale optimization scheduling while ensuring data timeliness.

[0081] Hourly power generation forecast data can be achieved by analyzing the historical power generation curve of the wind turbine and the historical characteristic data related to the power generation. This embodiment constructs an LSTM-based power generation prediction model based on the historical power generation data of a certain offshore wind farm from January to September 2022 and the historical characteristic data related to the power generation. The power generation forecast value is obtained by inputting the characteristic data of the set current period (i.e., October 1, 2022). On this basis, the confidence interval evaluation method is used to quantify the prediction uncertainty, and finally the refined power generation forecast result is output. The predicted data will serve as the decision-making basis for the electrolyzer hydrogen production power planning and energy dynamic matching, effectively improving the system operation economy and renewable energy absorption efficiency.

[0082] The historical characteristic data may include weather data, temperature data, season data, wind data, etc. This embodiment does not limit the specific type of characteristic data.

[0083] S12. Use a step curve to fit the hourly power generation forecast data of the current cycle.

[0084] Power generation forecast data preprocessing is a necessary prerequisite for step curve fitting. Its core includes two stages: outlier processing and feature analysis. The preprocessing stage needs to filter out outliers such as negative power and over-limit data to ensure the physical rationality of the data and improve model accuracy and control stability. The feature analysis stage divides the power fluctuation level area through volatility analysis and identifies power mutation nodes, thereby establishing the boundaries of non-uniform time segments and providing feature engineering support for subsequent dynamic modeling.

[0085] In this embodiment, the power generation forecast data of the current cycle is divided into 24 time intervals. Each interval uses a constant power value to represent the forecast output of the corresponding time period. The hourly power generation forecast step curve is generated by step curve fitting, such as Figure 4 shown.

[0086] S13. Optimize the step curve time interval with the goal of reducing the number of steps.

[0087] Optimizing the time interval of the hydrogen production power step curve with the goal of reducing the number of steps can reduce the frequency of equipment start-up and shutdown or power fluctuations, thereby improving the reliability and economy of the hydrogen production system.

[0088] The program flow of the step curve time interval optimization method is as follows: Figure 2 As shown, the specific steps include:

[0089] First, the power difference threshold is determined based on the equipment constraint analysis and historical data statistics of the wind power off-grid hydrogen production system. Among them, the power difference threshold is the maximum power difference allowed to be merged between adjacent segments, and its value must be less than the maximum power change rate allowed by the equipment. It is a key parameter for balancing the number of steps and power deviation. This embodiment constructs mathematical models of wind power off-grid hydrogen production systems. These models can accurately characterize the energy transmission characteristics and operating constraint boundaries of each unit, lay a physical foundation for constructing the power balance equation, and provide parameterized constraints. Specifically, the wind turbine model is expressed by the following formula:

[0090]

[0091] Among them, ρ air is the air density; C p is the wind energy conversion efficiency coefficient of the blade; R w is the radius of the wind turbine wheel; v w is the wind speed.

[0092] The energy storage battery electromotive force model is expressed by the following formula:

[0093]

[0094] Among them, N b_s and N b_p are the number of batteries in series and parallel in the battery pack respectively; E0 is the initial internal electromotive force; K is the polarization voltage constant; i(τ) is the charge and discharge current; Q n is the rated capacity of the energy storage battery; A u is the voltage variation coefficient; B c is the capacity variation coefficient; C t is the polarization effect coefficient; T b is the energy storage battery temperature.

[0095] The energy storage battery state of charge model is expressed by the following formula:

[0096]

[0097] The PEM electrolyzer model is represented by the following formula:

[0098]

[0099] Among them, Ncell is the number of electrolytic chambers connected in series; U rev is the standard reversible voltage; R is the ideal gas constant; T is the operating temperature of the electrolytic cell; F is the Faraday constant; are the partial pressures of hydrogen and oxygen, respectively; is the activity of water; α is the charge transfer coefficient; J n is the working current density; J0 is the exchange current density; δ mem is the thickness of the film; σ mem Conductivity of the membrane; J L is the limiting current density.

[0100] The voltage state constraint is expressed by the following formula:

[0101] V min ≤V dc ≤V max

[0102] The current state constraint is expressed by the following formula:

[0103]

[0104] The battery operating power state constraint is expressed by the following formula:

[0105] -P char,max ≤P char ≤0, 0≤P dis ≤P dis,max

[0106] The battery operating capacity state constraint is expressed by the following formula:

[0107] SOC min ≤SOC(t)≤SOC max

[0108] The hydrogen production rate state constraint of the hydrogen production electrolyzer is expressed by the following formula:

[0109]

[0110] When determining the power difference threshold of the off-grid wind power hydrogen production system, this embodiment analyzes key parameters such as the power / capacity constraints of the energy storage battery, the power ramp rate of the electrolyzer, and the start-stop frequency, extracts the maximum allowable power change rate, and reserves a safety margin as the initial threshold benchmark; statistically analyzes the difference distribution characteristics of historical wind power in adjacent time periods, and screens candidate threshold intervals that cover small fluctuations and match the dynamic response capabilities of the equipment; through simulation of system voltage stability, hydrogen production efficiency, and energy deviation indicators under different thresholds, the system is gradually corrected to a balanced threshold that meets the goals of safe equipment operation and step-by-step optimization, and the power difference threshold is determined to be 0.09pu, providing a threshold decision basis for the hydrogen production system that takes into account both operational stability and scheduling economy.

[0111] Secondly, traverse the adjacent power segments and calculate the absolute value of the power difference, and select the adjacent power segments with the smallest absolute value of the power difference and below the set threshold. This embodiment extracts the power values of adjacent time periods in chronological order and calculates their absolute value difference, generates a difference list of all adjacent segment pairs, and then traverses the list to find the smallest power difference and determines whether it is less than or equal to the preset threshold. If the condition is met, the adjacent segment corresponding to the smallest difference is marked as a priority merging object, such as Figure 4 As shown in the figure, the power difference between the 6th hour and the 7th hour is 0.001 pu, which is selected because it is the global minimum and lower than the threshold of 0.09 pu. If there are multiple segments with the same minimum difference, the first segment is selected according to the time order or priority rule. After completing one screening, the list of remaining segments needs to be updated and the power difference between the merged new segment and the adjacent segments needs to be recalculated. The above process is repeated until the difference values of all adjacent segments are higher than the threshold or there are no valid merge candidates.

[0112] Then, the power values of adjacent time periods are iteratively adjusted to merge the time intervals, the number of steps is gradually reduced, the time period parameters are updated and redundant data is deleted, and the screening and merging operations are performed cyclically until the difference values of all adjacent power segments exceed the set threshold. Each round of merging operation must strictly verify the energy integral conservation condition to ensure that the total energy of the step curve is consistent with the original data. Figure 4 As shown, in this embodiment, after the combined power value of the 6th and 7th hours is corrected to 0.218pu, it is merged with the power value of the 5th hour for the second time. The new power value is weighted according to the duration weight of each segment as follows: (0.218×1+0.224×2) / 3=0.222pu. After that, the time period parameters are synchronously updated and redundant data is cleared.

[0113] Finally, adjacent power segments are traversed and linear transition powers are inserted to ensure that the slopes between adjacent segments meet the operational constraints. The total energy deviation between the optimized step curve and the original data is calculated, and the deviation is evenly distributed to each time period through global translation to achieve accurate matching of energy values and eliminate cumulative errors.

[0114] The above steps sequentially implement segmented redundancy elimination, dynamic merging calculation, transition smoothing and energy conservation correction, and generate a hydrogen production power step curve that takes into account both step simplification and dynamic constraints. Figure 4 As shown, this embodiment finally generates a hydrogen production power curve 1 after hourly hydrogen production power planning.

[0115] S21. Obtain and analyze minute-level power generation forecast data for the next period.

[0116] The next time period refers to the adjacent time unit to be scheduled within the current hydrogen production power planning cycle. The power generation forecast must be completed before the start of the time period. In this embodiment, October 1, 2022, 12:00 to 17:00 is set as a time period for the wind power off-grid hydrogen production system. The next cycle forecast must be started 10 minutes before the end of the previous cycle. For example, the forecast task is executed at 11:50 on October 1, 2022, to provide data support for the second round of step curve fitting.

[0117] In this embodiment, the specific method for predicting the minute-level power generation for the next time period is to train a power generation prediction model based on the historical power generation data of the wind turbine and historical feature data related to the power generation, and obtain the power generation prediction value by inputting the feature data of the current period. This embodiment constructs an LSTM-based power generation prediction model based on the historical power generation data and historical feature data related to the power generation of the Huaneng Rudong Baxianjiao Offshore Wind Farm from January to September 2022, and obtains the power generation prediction value by inputting the feature data of the set current period (i.e., October 1, 2022).

[0118] S22: performing a second round of fitting on the first round of step curves within the stage based on the minute-level power generation prediction data.

[0119] In specific implementation, based on Figure 4 The minute-level power generation prediction curve shown is a second round of fitting for the period corresponding to the 12th to 17th hour of the hydrogen production power curve 1.

[0120] S23. Optimize the time interval of the second round of fitting step curves with the goal of reducing the number of steps.

[0121] The method for fitting the time interval of the step curve in the second round is the same as the method for optimizing the time interval of the step curve in step S13. For specific implementation details, please refer to the relevant description of step S13. The difference is that the result of the second round of fitting in statistically analyzing the difference distribution characteristics of adjacent time periods of historical wind power and screening the candidate threshold intervals that cover small fluctuations and match the dynamic response capabilities of the equipment is different from that of the first round of fitting, so the power difference threshold obtained is different; when merging time intervals by iteratively adjusting the power values of adjacent time periods, the time weight calculation of the merging operation adopts minute-level time units. The power difference threshold determined by the second round of fitting in this embodiment is 0.03pu. Example 1 adopts the above-mentioned parameter settings when implementing the second round of fitting, and achieves the step number optimization goal through the step curve time interval optimization method, and obtains the following Figure 4 The hydrogen production power curve 2 is shown.

[0122] The technical solution of the present invention reduces the regulation frequency of the auxiliary equipment of the hydrogen production system, suppresses local hot spots and membrane electrode stress damage caused by power mutation, and thus improves the coordinated scheduling capability of the wind and solar hydrogen production system.

[0123] Example 2: Dual-time scale planning of hydrogen production power for an off-grid photovoltaic hydrogen production system.

[0124] like Figure 5 As shown, the photovoltaic off-grid hydrogen production system provided in this embodiment consists of photovoltaics, energy storage batteries, hydrogen production power supply and PEM electrolyzer. The photovoltaic module has a configuration capacity of 6MW, which is connected to the DC bus through a DC / DC converter, with a real-time power fluctuation range of 0-6.63MW; the energy storage battery has a total capacity of 0.7MW, and the SOC effective working range is 20%-80%; the hydrogen production power supply adopts a three-level IGBT topology, supports a wide voltage input of 150-1000V, DC ripple ≤1%, and conversion efficiency ≥97%; the electrolyzer has a rated hydrogen production capacity of 1000Nm 3 / h, operating pressure 3.0MPa, and load adjustment range 10%-120%. This diagram clearly shows the energy interaction paths of each component, providing a topological basis for constructing power balance equations and equipment operation constraint models.

[0125] S11. Obtain and analyze hourly power generation forecast data for the current cycle.

[0126] In this embodiment, 7:00 to 19:00 on February 15, 2011 is set as a time period of the photovoltaic off-grid hydrogen production system. The next cycle prediction needs to be started 10 minutes before the end of the previous cycle. For example, the prediction task is executed at 6:50 on February 15, 2011.

[0127] The hourly power generation forecast data can be achieved by analyzing the historical power generation curve of the photovoltaic array and the historical characteristic data related to power generation. This embodiment constructs an LSTM-based power generation forecast model based on the historical power generation data and historical characteristic data related to power generation of the Nanjing Mufu Innovation Town photovoltaic power station from February 2010 to January 2011. The power generation forecast value is obtained by inputting the characteristic data of the set current period (i.e., from 7:00 to 19:00 on February 15, 2011). On this basis, the confidence interval evaluation method is used to quantify the prediction uncertainty, and finally outputs a refined power generation forecast result.

[0128] The historical characteristic data may include weather data, temperature data, season data, and radiation intensity data, etc. This embodiment does not limit the specific type of characteristic data.

[0129] S12. Use a step curve to fit the hourly power generation forecast data of the current cycle.

[0130] In this embodiment, the power generation forecast data of the current cycle is divided into 12 time intervals. Each interval uses a constant power value to represent the forecast output of the corresponding time period. The hourly power generation forecast step curve is generated by step curve fitting, such as Figure 6 shown.

[0131] S13. Optimize the step curve time interval with the goal of reducing the number of steps.

[0132] The program flow of the step curve time interval optimization method is as follows: Figure 2 As shown, the specific steps include:

[0133] First, the power difference threshold is determined based on the analysis of equipment constraints and historical data statistics for the photovoltaic off-grid hydrogen production system. This embodiment constructs mathematical models of the photovoltaic off-grid hydrogen production system. These models can accurately characterize the energy transfer characteristics and operational constraint boundaries of each unit, laying a physical foundation for constructing the power balance equation and providing parameterized constraints. Specifically, the photovoltaic array model is represented by the following formula:

[0134]

[0135] Among them, N pv_p With N pv_s are the number of PV modules connected in parallel and in series in the PV array respectively; I sc is the short-circuit current; U oc is the open circuit voltage; C1 and C2 are correction coefficients.

[0136] The mathematical model and constraints of the energy storage battery and PEM electrolyzer are consistent with those of the wind power off-grid hydrogen production system in Example 1. The difference is that the equipment parameters of the photovoltaic off-grid hydrogen production system are different from those of the wind power off-grid hydrogen production system.

[0137] When determining the power difference threshold of the off-grid wind power hydrogen production system, this embodiment analyzes key parameters such as the power / capacity constraints of the energy storage battery, the power ramp rate of the electrolyzer, and the start-stop frequency, extracts the maximum allowable power change rate, and reserves a safety margin as the initial threshold benchmark. The difference distribution characteristics of historical photovoltaic power in adjacent time periods are statistically analyzed to screen candidate threshold intervals that cover small fluctuations and match the dynamic response capabilities of the equipment. Through simulation of system voltage stability, hydrogen production efficiency, and energy deviation indicators under different thresholds, the system is gradually corrected to a balanced threshold that meets the goals of safe equipment operation and step-by-step optimization, and the power difference threshold is determined to be 0.1 pu.

[0138] Secondly, traverse the adjacent power segments and calculate the absolute value of the power difference, and select the adjacent power segments with the smallest absolute value of the power difference and below the set threshold. This embodiment extracts the power values of adjacent time periods in chronological order and calculates their absolute value difference, generates a difference list of all adjacent segment pairs, and then traverses the list to find the smallest power difference and determines whether it is less than or equal to the preset threshold. If the condition is met, the adjacent segment corresponding to the smallest difference is marked as a priority merging object, such as Figure 6 As shown in the figure, the power difference between the 12th hour and the 12th hour is 0.047 pu, which is selected because it is the global minimum and lower than the threshold of 0.1 pu. If there are multiple segments with the same minimum difference, the first segment is selected according to the time order or priority rule. After completing one screening, the list of remaining segments needs to be updated and the power difference between the merged new segment and the adjacent segments needs to be recalculated. The above process is repeated until the difference values of all adjacent segments are higher than the threshold or there are no valid merge candidates.

[0139] Then, the power values of adjacent time periods are iteratively adjusted to merge the time intervals, the number of steps is gradually reduced, the time period parameters are updated and redundant data is deleted, and the screening and merging operations are performed cyclically until the difference values of all adjacent power segments exceed the set threshold. Each round of merging operation must strictly verify the energy integral conservation condition to ensure that the total energy of the step curve is consistent with the original data. Figure 6 As shown, in this embodiment, after the combined power value of the 12th and 13th hours is corrected to 0.712pu, it is merged with the power value of the 11th hour for the second time. The new power value is calculated by weighting the duration of each segment as follows: (0.618×1+0.712×2) / 3=0.681pu. After that, the time period parameters are synchronously updated and redundant data is cleared.

[0140] Finally, adjacent power segments are traversed and linear transition powers are inserted to ensure that the slopes between adjacent segments meet the operational constraints. The total energy deviation between the optimized step curve and the original data is calculated, and the deviation is evenly distributed to each time period through global translation to achieve accurate matching of energy values and eliminate cumulative errors.

[0141] The above steps sequentially implement segmented redundancy elimination, dynamic merging calculation, transition smoothing and energy conservation correction, and generate a hydrogen production power step curve that takes into account both step simplification and dynamic constraints. Figure 6 As shown, this embodiment finally generates a hydrogen production power curve 1 after hourly hydrogen production power planning.

[0142] S21. Obtain and analyze minute-level power generation forecast data for the next period.

[0143] In this embodiment, 11:00 to 14:00 on February 15, 2011 is set as a time period for the photovoltaic off-grid hydrogen production system. The prediction of the next cycle needs to be started 10 minutes before the end of the previous cycle. For example, the prediction task is executed at 10:50 on February 15, 2011 to provide data support for the second round of step curve fitting.

[0144] This embodiment constructs an LSTM-based power generation prediction model based on the historical power generation data of the Nanjing Mufu Innovation Town photovoltaic power station from February 2010 to January 2011 and historical feature data related to power generation. The power generation prediction value is obtained by inputting the feature data of the next set time period (i.e., 11:00 to 14:00 on February 15, 2011).

[0145] S22: performing a second round of fitting on the first round of step curves within the stage based on the minute-level power generation prediction data.

[0146] In specific implementation, based on Figure 6 The minute-level power generation prediction curve shown is a second round of fitting for the period corresponding to the 11th to 14th hour of the hydrogen production power curve 1.

[0147] S23. Optimize the time interval of the second round of fitting step curves with the goal of reducing the number of steps.

[0148] The method for fitting the time interval of the step curve in the second round is the same as the method for optimizing the time interval of the step curve in step S13. For specific implementation details, please refer to the relevant description of step S13. The difference is that the result of the second round of fitting in statistically analyzing the difference distribution characteristics of adjacent time periods of historical photovoltaic power and screening the candidate threshold intervals that cover small fluctuations and match the dynamic response capability of the equipment is different from that of the first round of fitting, so the power difference threshold obtained is different; when merging time intervals by iteratively adjusting the power values of adjacent time periods, the time weight calculation of the merging operation adopts minute-level time units. The power difference threshold determined by the second round of fitting in this embodiment is 0.03pu. Example 1 adopts the above-mentioned parameter settings when implementing the second round of fitting, and achieves the step number optimization goal through the step curve time interval optimization method, and obtains the following Figure 6 The hydrogen production power curve 2 is shown.

[0149] The dual-time-scale planning method for hydrogen production power of the green electricity off-grid hydrogen production system is applicable to alkaline electrolyzers, proton exchange membrane electrolyzers, anion exchange membrane electrolyzers and solid oxide electrolyzers. Its universality stems from the universal modeling of dynamic response characteristics and mechanical stress constraints. For alkaline electrolyzers, the slope constraint suppresses the minute-level power delay caused by electrode polarization and electrolyte circulation inertia, reducing the local temperature rise and oxygen evolution overpotential surge caused by step switching; for anion exchange membrane electrolyzers, the amplitude constraint limits the sudden change of current density, suppresses the fluctuation amplitude within the elastic deformation threshold of the ion membrane, and avoids interface stratification; for solid oxide electrolyzers, the power switching rate constraint and the synchronous control of the thermal gradient can prevent thermal stress cracks in the ceramic electrolyte. Therefore, the dual-time-scale collaborative planning method for hydrogen production power can adapt to the dynamic characteristics of different electrolyzers.

[0150] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0151] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

[0152] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

Claims

1. A dual-time-scale planning method for hydrogen production power of a green off-grid hydrogen production system, including hourly hydrogen production power planning based on electricity forecast and minute-level hydrogen production power planning based on power forecast, characterized in that: The hourly hydrogen production power planning based on power forecast includes: using a step curve to fit the hourly power generation forecast data of renewable energy based on an hourly step size to obtain a first step curve, and optimizing the first step curve to obtain a second step curve with the constraint that the integrated areas of the curves before and after optimization are the same and the goal of reducing the number of steps; The minute-level hydrogen production power planning based on power prediction includes: for each step of the second step curve, the step curve is used to fit the minute-level power generation prediction data of renewable energy with a step size of minutes to obtain a third step curve, and the third step curve is optimized with the constraint that the integral area of the curve before and after optimization is the same and with the goal of reducing the number of steps to obtain a fourth step curve, and all the fourth step curves are spliced in time sequence to obtain a hydrogen production power curve.

2. The dual-time-scale planning method for hydrogen production power of a green off-grid hydrogen production system according to claim 1 is characterized in that: The step curve is used to fit the hourly power generation forecast data of renewable energy with a step length of one hour to obtain a first step curve, which includes: Obtain hourly renewable energy power generation forecast data for the target planning period; Drawing a first prediction curve based on hourly power generation forecast data of renewable energy; Generate the first fitting interval with a step size of hours; Based on the first fitting interval, a step curve is used to fit the first prediction curve to obtain a first step curve, wherein each first fitting interval is fitted into the same step.

3. The dual-time-scale planning method for hydrogen production power of a green off-grid hydrogen production system according to claim 2 is characterized in that: The step of optimizing the first step curve to obtain the second step curve with the constraint that the integral areas of the curves before and after optimization are the same and with the goal of reducing the number of steps further includes: According to the parameter constraints and historical data statistics of renewable energy power generation equipment, the power difference threshold for adjacent steps to be merged is set; Traversal step: traverse all adjacent steps, calculate the absolute value of the power difference between each adjacent step, screen out the two steps with the smallest absolute value of the power difference, and determine whether the absolute value of the power difference between the two screened steps is less than the power difference threshold. If so, perform the step merging step, otherwise obtain the second step curve; In the step of merging steps, if the absolute value of the power difference between the two steps obtained by screening is less than the power difference threshold, the two steps obtained by screening are merged into one step, and the power value of the merged step is determined based on the constraint that the integral area of the curves before and after the merger is the same, and then the traversal step is returned.

4. The method for dual-time-scale planning of hydrogen production power for a green off-grid hydrogen production system according to claim 1, characterized in that: The step curve is used to fit the renewable energy minute-level power generation forecast data with a step length of minutes to obtain a third step curve, including: Obtain the minute-level power generation forecast data of renewable energy in the time interval corresponding to any step in the second step curve; Drawing a second prediction curve based on the obtained renewable energy minute-level power generation prediction data; Generate the second fitting interval with a step size of minutes; Based on the second fitting interval, a step curve is used to fit the second prediction curve to obtain a third step curve, wherein each second fitting interval is fitted into the same step.

5. The dual-time-scale planning method for hydrogen production power of a green off-grid hydrogen production system according to claim 4 is characterized in that: The method of optimizing the third step curve to obtain the fourth step curve with the constraint that the integral areas of the curves before and after optimization are the same and with the goal of reducing the number of steps includes: According to the parameter constraints and historical data statistics of renewable energy power generation equipment, the power difference threshold for adjacent steps to be merged is set; Traversal step: traverse all adjacent steps, calculate the absolute value of the power difference between each adjacent step, screen out the two steps with the smallest absolute value of the power difference, and determine whether the absolute value of the power difference between the two screened steps is less than the power difference threshold. If so, perform the step merging step, otherwise obtain the fourth step curve; In the step of merging steps, if the absolute value of the power difference between the two steps obtained by screening is less than the power difference threshold, the two steps obtained by screening are merged into one step, and the power value of the merged step is determined based on the constraint that the integral area of the curves before and after the merger is the same, and then the traversal step is returned.

6. The dual-time-scale planning method for hydrogen production power of a green off-grid hydrogen production system according to claim 1 is characterized in that: The hydrogen production power curve is obtained by splicing all the fourth-step curves in time sequence, including: All fourth-step curves are spliced together in time sequence to obtain the transition power curve; The hydrogen production power curve is obtained by inserting the linear transition power into the adjacent steps.

7. A dual-time-scale planning method for hydrogen production power of a green off-grid hydrogen production system according to claim 6, characterized in that: The linear transition power is determined according to the maximum power change rate of the renewable energy power generation equipment and the principle of constant integral area.

8. The dual-time-scale planning method for hydrogen production power of a green off-grid hydrogen production system according to claim 1 is characterized in that: The renewable energy sources include at least wind energy and solar energy.

9. A dual-time-scale planning device for hydrogen production power of a green electricity off-grid hydrogen production system, comprising a memory, a processor, and a program stored in the memory, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 8 is implemented.

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

Images