Output prediction method of wind and light power generation system and wind and light hydrogen production scheduling method and system

Abstract

The invention discloses an output prediction method of a wind and light power generation system, and a wind and light hydrogen production scheduling method and system, and relates to the technical field of renewable energy source production control, and the method comprises the steps: obtaining wind and light prediction output data, carrying out the preprocessing, and carrying out the interpolation completion of the missing data; acquiring wind and light predicted output data in the future 0-a hours, and recording the data as first data; acquiring wind and light predicted output data from a to b hours in the future, and recording the data as second data; 0 < a < b; acquiring historical wind and light prediction output data to correct the second data to generate third data; combining the first data and the third data into final predicted output data; outputting final predicted output data in real time; performing real-time rolling correction on the final predicted output data; the real-time rolling correction is to update the final predicted output data by taking a preset time interval as a period. Through real-time feedback and correction of wind and light prediction output data, a continuously updated high-precision basis is provided for a downstream power utilization system.

Description

Technical Field

[0001] This invention relates to the field of renewable energy production control technology, and in particular to a method for predicting the output of wind and solar power generation systems, and a method and system for scheduling wind and solar hydrogen production. Background Technology

[0002] As the global energy structure shifts towards cleaner and lower-carbon energy, renewable energy power generation technologies, represented by wind and solar power, have developed rapidly. Utilizing this "green electricity" to produce "green hydrogen" through water electrolysis is one of the key technological pathways for achieving large-scale, long-term storage and cross-regional consumption of renewable energy. This technology can convert electrical energy into hydrogen energy, which can then be used as fuel or industrial feedstock in multiple fields such as transportation, chemical engineering, and metallurgy, effectively promoting deep decarbonization. Currently, wind and solar power hydrogen production systems mainly include modes such as "partial grid connection of wind and solar power + hydrogen production from surplus electricity" and "completely off-grid independent hydrogen production from wind and solar power."

[0003] Despite its promising prospects, current wind-solar coupled hydrogen production technology still faces a series of pressing technical bottlenecks in its large-scale and efficient application. First, wind and solar energy are inherently volatile, intermittent, and random, with their power generation significantly affected by climate and weather. It is particularly noteworthy that photovoltaic power output exhibits a strong, cyclical, and dramatic fluctuation throughout the day: power begins to climb at sunrise, peaks at midday, and then rapidly declines to zero before sunset. This significant intraday power curve fluctuation, combined with the potentially more continuous wind power output, constitutes an extremely complex and highly volatile mixed energy input. This unstable power output fundamentally contradicts the continuous and stable operation required by downstream chemical production (such as ammonia synthesis and methanol production). Frequent power fluctuations, especially the daily, regular peak-valley changes, force hydrogen electrolyzers into a non-steady-state operating range of repeated start-ups and shutdowns and frequent load increases and decreases, significantly reducing the electro-hydrogen conversion efficiency and equipment lifespan, and posing serious safety challenges.

[0004] Currently, to ensure the stability of downstream production, the system heavily relies on high-precision forecasts of wind and solar power output to enable advance production planning. Current mainstream forecasting methods depend heavily on historical meteorological and power data, generating forecast curves through numerical weather prediction and statistical models. However, this method has inherent limitations: firstly, historical data models struggle to fully capture sudden changes in weather systems, leading to baseline errors in forecasts; secondly, the systems generally lack efficient, automated real-time monitoring and feedback correction mechanisms, failing to dynamically calibrate forecasts and promptly correct deviations based on actual weather changes (such as rapid cloud movement) within ultra-short timescales (e.g., the next 15 minutes to 1 hour). Furthermore, most current systems employ relatively crude production scheduling strategies, failing to adapt and optimize for the dynamic characteristics of wind and solar power output, especially solar power, across different timescales within the day. For example, they fail to coordinate with wind power output based on the typical solar power curve of "morning ramp-midday peak-afternoon decline" to dynamically formulate differentiated hydrogen production operation plans. This results in production scheduling schemes based on forecasts often differing significantly from reality, making it difficult to match energy supply with production load.

[0005] Therefore, from the perspective of system operation economy and energy efficiency, the above problems directly lead to two coexisting adverse phenomena: on the one hand, when wind and solar power output is abundant, the hydrogen production system cannot fully absorb it or the downstream cannot utilize it in time, resulting in "wind and solar curtailment" and wasting green electricity; on the other hand, when wind and solar power output is insufficient or fluctuates drastically, it is unable to provide a stable and sufficient amount of hydrogen to the downstream, resulting in "hydrogen shortage" and affecting continuous production. This contradiction results in low green electricity utilization rate and poor operating economy of the entire system, and chemical equipment faces shortened lifespan and safety hazards due to frequent load fluctuations.

[0006] In summary, there is an urgent need to provide a method for predicting the output of wind and solar power generation systems, and a method and system for scheduling wind and solar hydrogen production, in order to solve the above-mentioned technical problems. Summary of the Invention

[0007] To address the problems existing in the prior art, this invention provides a method for predicting the output of wind and solar power generation systems, a method and system for scheduling hydrogen production from wind and solar power, and aims to improve overall energy efficiency and economy by constructing an ultra-short-term prediction model with real-time feedback correction and formulating a dynamic hydrogen production scheduling strategy based on the prediction results.

[0008] The technical solution of this invention is implemented as follows: The power output prediction method for wind and solar power generation systems includes the following steps: A1. Obtain and preprocess wind and solar power output forecasts to remove outliers; wind and solar power output forecasts are predictions of wind and solar power generation. Specifically, sudden drops in output due to equipment failure can be considered outliers and need to be removed; in practice, outliers can be filtered and removed using threshold ranges.

[0009] A2. Interpolate and complete the missing parts in the wind and solar power prediction output data; A3. Obtain the predicted wind and solar power output data for the next 0 to a hours, and record it as the first data; obtain the predicted wind and solar power output data for the next a to b hours, and record it as the second data; 0 < a < b; obtain the historical predicted wind and solar power output data to correct the second data and generate the third data; merge the first data and the third data into the final predicted power output data; output the final predicted power output data in real time. A4. Perform real-time rolling correction on the final predicted output data; the real-time rolling correction is: repeat step A3 at a preset time interval. Further, the preset time interval is 10 to 20 minutes.

[0010] This invention employs a real-time rolling correction mechanism to dynamically correct and update the predicted power output of wind and solar power, making the predicted curves more closely reflect actual changes. This provides continuously updated, high-precision data for the scheduling plans of downstream power systems, achieving smooth closed-loop control from prediction to execution. This significantly improves the accuracy of renewable energy consumption and the economic efficiency of system operation. By setting a rolling correction cycle on the order of minutes, it ensures prediction accuracy while optimizing computational efficiency and the real-time nature of actual control.

[0011] As a further optimization of the above scheme, the wind and solar power output prediction data is the expected power generation obtained from wind and photovoltaic power generation; the wind and solar power output prediction data includes several data points, each corresponding to the expected power at a specified time. The interpolation completion is represented as follows: ; in, This indicates the time point corresponding to the missing data point. ; This indicates the expected power at a specified time.

[0012] As a further optimization of the above scheme, the correction process is expressed as follows: ; in, This refers to the wind and solar power output data before correction at a specified time t, i.e., the second data; The corrected wind and solar power output data, i.e., the third data; To obtain the prediction error based on the historical wind and solar power output data; For the correction factor, and The value ranges from 0.7 to 1.0.

[0013] As a further optimization of the above scheme, the prediction error is calculated as follows: ; in, This represents the historical wind and solar power output data at a specified time t. This represents the actual wind and solar power output data at a specified time t in history; and Historical data at the same point in time; and The predicted wind and solar power output data for different periods at the same specified time within each period.

[0014] As a further optimization of the above scheme, the final predicted power output data is a plurality of discrete data points fitted according to the preset time interval within an hour from 0 to b, where each data point represents the predicted power output data of wind and solar power at a specified time; the length of the preset time interval is an integer and is a common factor of a and b; b is an integer and is a factor of the period.

[0015] This invention also provides a wind-solar hydrogen production scheduling method, which applies the above-mentioned wind and solar power generation system output prediction method; including the following steps: B1. Based on the final predicted power output data, obtain the predicted wind power and solar power power within 24 hours; B2. Calculate the baseline load for hydrogen production in the electrolyzer based on the wind power forecast; calculate the average wind power forecast based on the baseline load; calculate the daily average hydrogen production of the electrolyzer based on the average wind power forecast; B3. Based on the daily average value, the photovoltaic power prediction is divided into three stages: when the photovoltaic power prediction is higher than the daily average value, it is recorded as the plateau stage; on the same day, the two stages before and after the plateau stage are respectively recorded as the load increase stage and the load decrease stage. Based on the predicted wind power, a first number of electrolytic cells are normally opened; during the plateau phase, a second number of electrolytic cells are opened; during the load increase phase, the number of electrolytic cells opened is gradually increased from the first number based on the increase in the predicted photovoltaic power; during the load decrease phase, the number of electrolytic cells is gradually closed from the second number based on the decrease in the predicted photovoltaic power. Based on the photovoltaic power prediction curve, a step-by-step flexible control is implemented for the electrolyzer. Through predictive smooth transition, frequent start-ups and shutdowns caused by power fluctuations are effectively avoided. This not only mitigates wind power output fluctuations but also improves the electrolyzer's lifespan and operational reliability. B4. Convert the baseline load into a baseline hydrogen production rate; convert the wind power forecast power into a wind power hydrogen production rate; compare the difference between the baseline hydrogen production rate and the wind power hydrogen production rate; obtain the maximum value of the difference as the wind power hydrogen storage demand space; the baseline load is used to measure the fluctuation range when the actual output of wind power is used for hydrogen production, so as to accurately calculate the buffer space of the hydrogen storage tank required to smooth out the fluctuation.

[0016] B5. Obtain the effective capacity of the hydrogen storage tank, subtract the maximum wind power hydrogen storage demand space, and use the result as the photovoltaic hydrogen storage space; convert the photovoltaic hydrogen storage space into a power value; and use the portion of the photovoltaic predicted power that exceeds the power value as the electricity to be fed into the grid. Hydrogen storage tanks are primarily used to balance the discrepancy between hydrogen production and consumption during the day and night. The tanks are primarily used to mitigate the impact of wind power fluctuations on hydrogen production; the remaining space is used to store hydrogen produced from solar power; and only the last portion is used to generate electricity for grid connection, maximizing the local utilization of green energy.

[0017] By dividing photovoltaic power output into three stages—increase, stabilize, and decrease—and coordinating it with wind power, the phased start-up and shutdown of the electrolyzer cluster was optimized. This effectively smoothed the impact of drastic fluctuations in photovoltaic power on the hydrogen production system, maximizing the utilization of green electricity while significantly improving the stability of equipment operation and the overall energy efficiency of the system.

[0018] As a further optimization of the above scheme, the reference load The calculation is expressed as: N represents the number of specified moments within 24 hours. This represents the wind power forecast corresponding to the i-th specified time; specifically, the unit of the baseline load is megawatt-hours (MWh), and the corresponding unit of the wind power forecast is megawatts (MW).

[0019] The average forecast value of wind power The calculation is expressed as: ; The daily average The calculation is expressed as: ; This indicates the energy consumption for hydrogen production in the electrolyzer, that is, the electrical energy consumed to produce a unit volume of hydrogen.

[0020] As a further optimization of the above scheme, it also includes B6, power allocation for each of the electrolyzers: B61. Obtain the first data, the inventory data of the hydrogen storage tank, the real-time power constraint information of the power grid, and the operating conditions of the hydrogen-using equipment, and construct an optimization model; based on the optimization model, output the total power data of the electrolyzer cluster and the hydrogen filling and releasing commands for the hydrogen storage tank; B62. Collect real-time power output data of wind and solar power, calculate the deviation between the real-time power output data of wind and solar power and the predicted power output data of wind and solar power, and perform feedforward-feedback compensation on the total power data to obtain the real-time total power. B63. Based on the load-efficiency characteristic data of the electrolyzer, with the goal of minimizing the total hydrogen production power consumption, the real-time total power is allocated to each electrolyzer, and an optimal power command is generated for each electrolyzer. B64. Repeat steps B61 to B63 with the preset time interval as the cycle.

[0021] By performing second-level load allocation based on the real-time efficiency characteristics of the electrolyzer, the dynamic global optimization of the system's hydrogen production power consumption is achieved. At the same time, the long-term balance and reliability of the equipment are ensured by periodically adjusting the predicted output.

[0022] As a further optimization of the above scheme, in B63, the objective is represented as: The constraints of the objective include: ; Where M represents the number of electrolytic cells, and j represents the j-th electrolytic cell; This represents the power allocated to the j-th electrolytic cell; This represents the real-time total power; and These are the lower and upper limits of the operating power of the j-th electrolytic cell, respectively; This represents the power change of the j-th electrolytic cell; This represents the maximum power change rate of the j-th electrolytic cell, also known as the ramp rate, which is used to limit the drastic power change of each electrolytic cell and prevent excessive power command jumps. This represents the load-efficiency characteristic data of the j-th electrolytic cell, that is, the data of the j-th electrolytic cell when operating at a power of... Energy conversion efficiency at that time.

[0023] The present invention also provides a system that applies the power output prediction method of the wind and solar power generation system described above; including a meteorological data acquisition module, a prediction module, a power generation module, a prediction data processing module, a scheduling module, and a power consumption module; The meteorological data acquisition module collects meteorological data and sends it to the prediction module; the prediction module, based on the meteorological data, combines historical meteorological data and the historical power generation of the power generation module to make predictions and generate the predicted wind and solar power output data; the prediction data processing module executes steps A1 to A4 to generate the final predicted power output data; the scheduling module generates a power scheduling plan for the power consumption module based on the final predicted power output data; the power consumption module executes power consumption according to the power scheduling plan.

[0024] Specifically, the meteorological data acquisition module includes weather stations and irradiance meters. Weather stations measure on-site wind speed, wind direction, temperature, air pressure, and humidity, while irradiance meters measure total solar radiation, direct radiation, and diffuse radiation. The power generation module refers to wind turbines, photovoltaic arrays, and their monitoring systems, used to generate electricity and collect operational data. Specifically, the wind turbine / photovoltaic inverter monitoring system collects real-time active / reactive power, voltage, current, and equipment status information from each power generation unit and the total output of the power station.

[0025] The forecasting module is connected to the meteorological data acquisition module and the power generation module. Based on real-time acquired meteorological data, combined with historical meteorological data and historical power generation data from the power generation module, it calculates and generates initial wind and solar power output data through a forecasting model. The forecast data processing module is connected to the forecasting module.

[0026] The scheduling module is connected to the predictive data processing module. Based on the received final predicted power output data, combined with grid constraints and power demand, it performs optimization calculations to generate the optimal power scheduling plan for the power consumption module.

[0027] The power consumption module is connected to the dispatching module and executes power consumption operations according to the power dispatching plan. The power consumption module includes, but is not limited to, the power grid itself and various adjustable loads, such as the electrolyzer cluster of a hydrogen production plant, which adjusts its total power and internal power distribution by receiving dispatching instructions.

[0028] Compared with the prior art, the present invention achieves the following beneficial effects: (1) This invention, through a real-time rolling correction mechanism, can dynamically correct and update the predicted output of wind and solar power, making the prediction curve closer to actual changes. This provides a continuously updated, high-precision basis for the scheduling plan of downstream power systems, realizing smooth closed-loop control from prediction to execution, and significantly improving the accuracy of renewable energy consumption and the economic efficiency of system operation. By setting a rolling correction cycle at the minute level, the accuracy of prediction is ensured while also taking into account the optimization of computational efficiency and the real-time performance of actual control.

[0029] (2) By dividing photovoltaic power output into three stages—increase, stabilize, and decrease—and coordinating with wind power, the phased start-up and shutdown optimization of the electrolyzer group was achieved. This effectively smoothed the impact of drastic fluctuations in photovoltaic power on the hydrogen production system, and significantly improved the stability of equipment operation and the overall energy efficiency of the system while maximizing the absorption of green electricity.

[0030] (3) Combining the “deviation correction” strategy based on historical data for wind and solar forecast data with the “photovoltaic three-stage allocation” strategy can provide fine and accurate data support for downstream hydrogen production scheduling.

[0031] (4) By allocating load at the second level based on the real-time efficiency characteristics of the electrolyzer, the dynamic global optimization of the hydrogen production power consumption of the system is realized. At the same time, the balance and reliability of the equipment in long-term operation are ensured by periodically adjusting the predicted output.

[0032] (5) This invention constructs a flexible and coordinated control system for the entire process of power generation, hydrogen production, and hydrogen storage. By deeply coupling the flexible operation ranges of each link, a complete operation strategy to cope with random fluctuations in wind and solar power is formed, thereby systematically improving the production resilience and green electricity consumption level of the entire renewable energy hydrogen production chain. Attached Figure Description

[0033] Figure 1 This is a flowchart illustrating the power output prediction method for a wind and solar power generation system provided in an embodiment of the present invention. Figure 2 This is a schematic flowchart of the wind-solar hydrogen production scheduling method provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the module connection of the wind and solar power generation system provided in an embodiment of the present invention. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions in the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0035] like Figure 1 As shown, this embodiment provides a method for predicting the output of a wind and solar power generation system, including the following steps: A1. Obtain and preprocess wind and solar power output forecasts to remove outliers; wind and solar power output forecasts are predictions of wind and solar power generation. Specifically, sudden drops in output due to equipment failure can be considered outliers and need to be removed; in practice, outliers can be filtered and removed using threshold ranges.

[0036] In this embodiment, the wind and solar power forecast output data is the expected power generation obtained from wind and solar power generation; the wind and solar power forecast output data includes several data points, each corresponding to the expected power at a specified time point; specifically, data points are set at 15-minute intervals.

[0037] A2. The missing portions of the wind and solar power output forecast data are interpolated and filled in, specifically as follows: ;in, This indicates the time corresponding to the missing data point. ; This indicates the expected power at a specified time.

[0038] A3. Obtain the predicted wind and solar power output data for the next 0 to a hours, denoted as the first data; obtain the predicted wind and solar power output data for the next a to b hours, denoted as the second data; 0 < a < b. In this embodiment, a 4-hour ultra-short-term and 24-hour short-term wind and solar power forecast structures are used for correction to reduce prediction errors. Therefore, a is set to 4, and b is set to 24. Correspondingly, the wind and solar power output data for one day is divided into 96 data points, corresponding to one time point every 15 minutes.

[0039] Historical wind and solar power output data is used to correct the second data and generate the third data; the first and third data are merged into the final predicted power output data; the final predicted power output data is output in real time.

[0040] In this embodiment, the correction process is represented as follows: , .in, This represents the wind and solar power output data before correction at a specified time t, i.e., the second data. This is the corrected wind and solar power forecast output data, i.e., the third data; To obtain the prediction error based on historical wind power prediction data; For the correction factor, and The value ranges from 0.7 to 1.0, with 0.8 being the specific value used.

[0041] This represents the historical wind and solar power output data at a specified time t. This represents the actual wind and solar power output data at a specified time t in history; and Historical data at the same point in time; and This provides wind and solar power output data for the same specified time within different cycles. In this embodiment, a one-year cycle is used, meaning historical data from the same period of the previous year is used for correction. For example... Given the wind and solar power forecast output data for 10:15 AM on December 20, 2025, then... and These are the predicted and actual wind and solar power output data for 10:15 on December 20, 2024.

[0042] The final predicted power output data consists of multiple discrete data points fitted within a preset time interval over a period of 0-24 hours. Each data point represents the predicted power output data for wind and solar power at a specified time. The preset time interval is 15 minutes in length.

[0043] A4. Perform real-time rolling correction on the final predicted output data; real-time rolling correction is: repeat step A3 at a preset time interval.

[0044] The above describes the entire process of continuously optimizing the 24-hour wind and solar power output curve based on the wind and solar power forecast data for the next 4 hours and the actual wind and solar power output data. For the first 4 hours (t = 1~16), the ultra-short-term forecast values ​​for the next 4 hours are used directly (or slightly smoothed). For the following 20 hours (t = 17~96), historical deviation information is used to correct the original forecast. A hybrid interpolation correction method is used to correct the short-term forecast results, improving the forecast accuracy to over 80%. The corrected 24-hour forecast curve is discretized at 15-minute intervals to form the base power sequence for continuous optimization.

[0045] This invention employs a real-time rolling correction mechanism to dynamically correct and update the predicted power output of wind and solar power, making the predicted curves more closely reflect actual changes. This provides continuously updated, high-precision data for the scheduling plans of downstream power systems, achieving smooth closed-loop control from prediction to execution. This significantly improves the accuracy of renewable energy consumption and the economic efficiency of system operation. By setting a rolling correction cycle on the order of minutes, it ensures prediction accuracy while optimizing computational efficiency and the real-time nature of actual control.

[0046] like Figure 2 As shown, this embodiment also provides a wind-solar hydrogen production scheduling method, which applies the above-mentioned output prediction method; including the following steps: B1. Based on the final predicted power output data, obtain the predicted wind power and solar power output for 24 hours.

[0047] B2. Calculate the baseline load for hydrogen production in the electrolyzer based on the wind power forecast; calculate the average wind power forecast based on the baseline load; calculate the daily average hydrogen production of the electrolyzer based on the average wind power forecast.

[0048] In this embodiment, the reference load The calculation is expressed as: N=96, representing the number of specified moments within 24 hours. This represents the wind power forecast at the i-th specified time; specifically, the unit of the baseline load is megawatt-hours (MWh), and the corresponding unit of the wind power forecast is megawatts (MW).

[0049] Average forecast value of wind power The calculation is expressed as: ; Daily average The calculation is expressed as: ; This indicates the energy consumption of hydrogen production in an electrolyzer, that is, the electrical energy consumed to produce a unit volume of hydrogen.

[0050] B3. Based on the daily average, the photovoltaic power forecast is divided into three stages: when the photovoltaic power forecast is higher than the daily average, it is called the plateau stage; on the same day, the two stages before and after the plateau stage are called the load increase stage and the load decrease stage, respectively. Based on the predicted wind power output, the first number of electrolytic cells is normally opened; during the plateau phase, the second number of electrolytic cells is opened; during the load increase phase, the number of electrolytic cells opened is gradually increased from the first number based on the increase in the predicted photovoltaic power output; during the load decrease phase, the number of electrolytic cells is gradually closed from the second number back to the first number based on the decrease in the predicted photovoltaic power output. If there are a total of 50 electrolytic cells (the second number), then based on the daily stable power generation of the predicted wind power output, 30 of the electrolytic cells (the first number) can be normally opened, and the number of operating electrolytic cells can be gradually increased / decreased based on the diurnal changes in the predicted photovoltaic power output.

[0051] Based on the photovoltaic power prediction curve, the electrolyzer is flexibly controlled in a step-by-step manner. Through predictive smooth transition, frequent start-ups and shutdowns of the equipment caused by power fluctuations are effectively avoided, thereby smoothing out wind power output fluctuations and improving the service life and operational reliability of the electrolyzer.

[0052] B4. Convert the baseline load to the baseline hydrogen production rate; convert the wind power forecast power to the wind power hydrogen production rate; compare the difference between the baseline hydrogen production rate and the wind power hydrogen production rate; obtain the maximum value of the difference as the wind power hydrogen storage demand space; the baseline load is used to measure the fluctuation range when the actual wind power output is used for hydrogen production, so as to accurately calculate the buffer space of the hydrogen storage tank required to smooth out the fluctuation.

[0053] B5. Obtain the effective capacity of the hydrogen storage tank, subtract the maximum wind power hydrogen storage demand space, and use it as the photovoltaic hydrogen storage space; convert the photovoltaic hydrogen storage space into a power value; for the portion of the photovoltaic predicted power that exceeds the power value, use it as the grid-connected electricity.

[0054] In this embodiment, B6 is also included: power allocation for each electrolytic cell. B61. Obtain the first data, the inventory data of the hydrogen storage tank, the real-time power constraint information of the power grid, and the operating conditions of the hydrogen-using equipment, and construct an optimization model; based on the optimization model, output the total power data of the electrolyzer cluster and the hydrogen filling and releasing commands for the hydrogen storage tank; B62. Collect real-time power output data of wind and solar power, calculate the deviation between the data and the predicted power output data of wind and solar power, perform feedforward-feedback compensation on the total power data, and obtain the real-time total power. B63. Based on the load-efficiency characteristic data of the electrolyzers, with the goal of minimizing total hydrogen production power consumption, the real-time total power is allocated to each electrolyzer, and an optimal power command is generated for each electrolyzer; in this embodiment, the goal is expressed as: The constraints on the objective include: ; Where M represents the number of electrolytic cells, and j represents the j-th electrolytic cell; This represents the power allocated to the j-th electrolytic cell; Indicates the real-time total power; and These are the lower and upper limits of the operating power of the j-th electrolytic cell, respectively; This represents the power change of the j-th electrolytic cell; This represents the maximum power change rate of the j-th electrolytic cell, also known as the ramp rate, which is used to limit the drastic power change of each electrolytic cell and prevent excessive power command jumps. This represents the load-efficiency characteristic data of the j-th electrolytic cell, that is, the data of the j-th electrolytic cell when operating at a power of... Energy conversion efficiency at that time.

[0055] B64. Repeat steps B61 to B63 at preset time intervals.

[0056] By performing second-level load allocation based on the real-time efficiency characteristics of the electrolyzer, the dynamic global optimization of the system's hydrogen production power consumption is achieved. At the same time, the long-term balance and reliability of the equipment are ensured by periodically adjusting the predicted output.

[0057] Hydrogen storage tanks are primarily used to balance the discrepancy between hydrogen production and consumption during the day and night. The tanks are primarily used to mitigate the impact of wind power fluctuations on hydrogen production; the remaining space is used to store hydrogen produced from solar power; and only the last portion is used to generate electricity for grid connection, maximizing the local utilization of green energy.

[0058] By dividing photovoltaic power output into three stages—increase, stabilize, and decrease—and coordinating it with wind power, the phased start-up and shutdown of the electrolyzer cluster was optimized. This effectively smoothed the impact of drastic fluctuations in photovoltaic power on the hydrogen production system, maximizing the utilization of green electricity while significantly improving the stability of equipment operation and the overall energy efficiency of the system.

[0059] like Figure 3 As shown, this embodiment also provides a system that applies the above-mentioned power output prediction method; including a meteorological data acquisition module, a prediction module, a power generation module, a prediction data processing module, a scheduling module, and a power consumption module; The meteorological data acquisition module collects meteorological data and sends it to the forecasting module; the forecasting module, based on the meteorological data and combined with historical meteorological data and the historical power generation of the power generation module, makes a forecast to generate wind and solar power output data; the forecast data processing module is used to execute steps A1 to A4 to generate the final forecast power output data; the scheduling module generates a power scheduling plan for the power consumption module based on the final forecast power output data; the power consumption module executes power consumption according to the power scheduling plan.

[0060] Specifically, the meteorological data acquisition module includes weather stations and irradiance meters. Weather stations measure on-site wind speed, wind direction, temperature, air pressure, and humidity, while irradiance meters measure total solar radiation, direct radiation, and diffuse radiation. The power generation module refers to wind turbines, photovoltaic arrays, and their monitoring systems, used to generate electricity and collect operational data. Specifically, the wind turbine / photovoltaic inverter monitoring system collects real-time active / reactive power, voltage, current, and equipment status information from each power generation unit and the total output of the power station.

[0061] The forecasting module is connected to the meteorological data acquisition module and the power generation module. Based on real-time acquired meteorological data, combined with historical meteorological data and historical power generation data from the power generation module, it calculates and generates initial wind and solar power output data through a forecasting model. The forecast data processing module is connected to the forecasting module.

[0062] The scheduling module is connected to the predictive data processing module. Based on the received final predicted power output data, combined with grid constraints and power demand, it performs optimization calculations to generate the optimal power scheduling plan for the power consumption module.

[0063] The power consumption module is connected to the dispatching module and executes power consumption operations according to the power dispatching plan. The power consumption module includes, but is not limited to, the power grid itself and various adjustable loads, such as the electrolyzer cluster of a hydrogen production plant, which adjusts its total power and internal power distribution by receiving dispatching instructions.

[0064] Based on the disclosure and teachings of the foregoing specification, those skilled in the art can make changes and modifications to the above embodiments. Therefore, the present invention is not limited to the specific embodiments disclosed and described above, and some modifications and changes to the present invention should also fall within the protection scope of the claims of the present invention. Furthermore, although some specific terms are used in this specification, these terms are only for convenience of explanation and do not constitute any limitation on the present invention.

Claims

1. A method for predicting the output of a wind and solar power generation system, characterized in that, Includes the following steps: A1. Obtain wind and solar power forecast output data and preprocess it to remove outlier data; A2. Interpolate and complete the missing parts in the wind and solar power prediction output data; A3. Obtain the predicted wind and solar power output data for the next 0 to a hours, and record it as the first data; obtain the predicted wind and solar power output data for the next a to b hours, and record it as the second data; 0 < a < b; Obtain historical wind and solar power output data to correct the second data and generate the third data; Combine the first data and the third data into the final predicted power output data; The final predicted output data is output in real time; A4. Perform real-time rolling correction on the final predicted output data; the real-time rolling correction is: repeat step A3 at a preset time interval.

2. The power output prediction method for a wind and solar power generation system according to claim 1, characterized in that, The wind and solar power output forecast data refers to the expected power generation obtained from wind and solar power generation; the wind and solar power output forecast data includes several data points, each corresponding to the expected power at a specified time. The interpolation completion is represented as follows: ; in, This indicates the time point corresponding to the missing data point. ; This indicates the expected power at a specified time.

3. The power output prediction method for a wind and solar power generation system according to claim 1, characterized in that, The correction process is represented as follows: ; in, This refers to the wind and solar power output data before correction at a specified time t, i.e., the second data; The corrected wind and solar power output data, i.e., the third data; To obtain the prediction error based on the historical wind and solar power output data; For the correction factor, and The value ranges from 0.7 to 1.

0.

4. The power output prediction method for a wind and solar power generation system according to claim 3, characterized in that, The prediction error is calculated as follows: ; in, This represents the historical wind and solar power output data at a specified time t. This represents the actual wind and solar power output data at a specified time t in history; and Historical data at the same point in time; and The predicted wind and solar power output data for different periods at the same specified time within each period.

5. The power output prediction method for a wind and solar power generation system according to claim 4, characterized in that, The final predicted power output data consists of multiple discrete data points fitted within an hour from 0 to b according to the preset time interval. Each data point represents the predicted wind and solar power output data at a specified time. The length of the preset time interval is an integer and is a common factor of a and b. b is an integer and is a factor of the period.

6. A wind-solar hydrogen production scheduling method, which applies the power output prediction method for wind and solar power generation systems as described in any one of claims 1 to 5; characterized in that, Includes the following steps: B1. Based on the final predicted power output data, obtain the predicted wind power and solar power power within 24 hours; B2. Calculate the baseline load for hydrogen production in the electrolyzer based on the predicted wind power output; Calculate the average wind power forecast based on the aforementioned baseline load; Calculate the daily average hydrogen production capacity of the electrolyzer based on the aforementioned wind power average forecast value; B3. Based on the daily average value, the photovoltaic power prediction is divided into three stages: when the photovoltaic power prediction is higher than the daily average value, it is recorded as the plateau stage; on the same day, the two stages before and after the plateau stage are respectively recorded as the load increase stage and the load decrease stage. Based on the predicted wind power, a first number of electrolytic cells are opened normally; during the plateau phase, a second number of electrolytic cells are opened; during the load ramp-up phase, based on the increase in the predicted photovoltaic power, the number of electrolytic cells opened is gradually increased from the first number. During the load reduction phase, the electrolytic cell is gradually shut down from the second quantity based on the decrease in the predicted photoelectric power. B4. Convert the baseline load into a baseline hydrogen production rate; convert the wind power forecast power into wind power hydrogen production rate; compare the difference between the baseline hydrogen production rate and the wind power hydrogen production rate; obtain the maximum value of the difference as the wind power hydrogen storage demand space; B5. Obtain the effective capacity of the hydrogen storage tank, subtract the maximum wind power hydrogen storage space requirement, and use the photovoltaic hydrogen storage space as the photovoltaic hydrogen storage space. The photoelectric hydrogen storage space is converted into a power value; the portion of the photoelectric predicted power that exceeds the power value is used as the electricity to be connected to the grid.

7. The wind-solar hydrogen production scheduling method according to claim 6, characterized in that, The reference load The calculation is expressed as: N represents the number of specified moments within 24 hours. This represents the predicted wind power at the i-th specified time. The average forecast value of wind power The calculation is expressed as: ; The daily average The calculation is expressed as: ; This indicates the energy consumption for hydrogen production in the electrolyzer.

8. The wind-solar hydrogen production scheduling method according to claim 6, characterized in that, It also includes B6, power allocation for each of the aforementioned electrolyzers: B61. Obtain the first data, the inventory data of the hydrogen storage tank, the real-time power constraint information of the power grid, and the operating conditions of the hydrogen-using equipment, and construct an optimization model; based on the optimization model, output the total power data of the electrolyzer cluster and the hydrogen filling and releasing commands for the hydrogen storage tank; B62. Collect real-time power output data of wind and solar power, calculate the deviation between the real-time power output data of wind and solar power and the predicted power output data of wind and solar power, and perform feedforward-feedback compensation on the total power data to obtain the real-time total power. B63. Based on the load-efficiency characteristic data of the electrolyzer, with the goal of minimizing the total hydrogen production power consumption, the real-time total power is allocated to each electrolyzer, and an optimal power command is generated for each electrolyzer. B64. Repeat steps B61 to B63 with the preset time interval as the cycle.

9. The wind-solar hydrogen production scheduling method according to claim 8, characterized in that, In B63, the target is represented as: ; The constraints on the objective include: ; Where M represents the number of electrolytic cells, and j represents the j-th electrolytic cell; This represents the power allocated to the j-th electrolytic cell; This represents the real-time total power; and These are the lower and upper limits of the operating power of the j-th electrolytic cell, respectively; This represents the power change of the j-th electrolytic cell; This represents the maximum power change rate of the j-th electrolytic cell; This represents the load-efficiency characteristic data of the j-th electrolytic cell.

10. A system that applies the power output prediction method for a wind and solar power generation system as described in any one of claims 1 to 5; characterized in that, It includes a meteorological data acquisition module, a forecasting module, a power generation module, a forecast data processing module, a dispatching module, and a power consumption module; The meteorological data acquisition module collects meteorological data and sends it to the prediction module; the prediction module makes predictions based on the meteorological data, combined with historical meteorological data and the historical power generation of the power generation module, and generates the wind and solar predicted power output data; the prediction data processing module is used to execute steps A1 to A4 to generate the final predicted power output data. The scheduling module generates a power scheduling plan for the power consumption module based on the final predicted power output data; The power consumption module executes power consumption according to the power scheduling plan.

Images