Dynamic electric energy scheduling method and system based on hydrogen production data characteristic feedback

Abstract

The invention discloses an electric energy dynamic scheduling method and system based on hydrogen production data characteristic feedback, and relates to the technical field of hydrogen production and electric energy scheduling, and the method comprises the steps: collecting multi-source data of a hydrogen production side (including full-process operation data), a power grid side (dual-path 6kV / 400V grid power and simulated green power fluctuation data), and a power supply side (energy storage emergency guarantee data); constructing a current-hydrogen production linear relation dynamic function covering the whole process; determining a power grid constraint boundary set containing energy storage emergency constraint; evaluating the power supply regulation capability containing the energy storage safe shutdown guarantee; establishing a four-target scheduling model with optimal hydrogen production efficiency, stable power grid, minimum start-stop loss and optimal time coordination, solving an optimal power and time coordination scheme, and generating a multi-system coordination instruction; and executing the instruction, collecting and feeding back in real time, and re-optimizing when the deviation exceeds the limit, thereby realizing closed-loop scheduling. The method adapts to intermittent working conditions, guarantees safe and efficient scheduling, prolongs the service life of equipment, and improves the reliability of a hydrogen production system.

Description

Technical Field

[0001] This invention relates to the field of hydrogen production and power dispatching technology, specifically to a dynamic power dispatching method and system based on hydrogen production data feature feedback. Background Technology

[0002] As green electricity-to-hydrogen technology extends deeper into off-grid intermittent scenarios, such as pilot-scale verification projects that require daily operation under routine conditions of "1 hour preheating + 6 hours full-load operation + 3 hours cooling" and monthly shutdowns for nitrogen replacement, power dispatch has become a core link connecting the demand for unstable green electricity and intermittent hydrogen production with system safety. It not only needs to ensure hydrogen production efficiency to improve green electricity utilization, but also needs to adapt to the characteristics of the entire process to reduce equipment wear. This is of significant practical importance for promoting the industrialization of off-grid green hydrogen and realizing the large-scale application of green electricity-to-hydrogen technology.

[0003] Existing dynamic power dispatch methods based on hydrogen production data feature feedback do not construct a current-hydrogen production correlation model covering the entire process of preheating-full load-cooling-replacement. Data processing ignores key parameters during non-production periods, does not focus on its emergency shutdown protection function, and does not incorporate energy storage emergency parameters into the grid constraint boundary. The dispatch model lacks a coordinated objective for optimal production time throughout the entire process, and the closed-loop dispatch does not cover key feedback parameters such as preheating / cooling rate and nitrogen replacement effect. This results in poor adaptability to operating conditions, frequent start-stop cycles that accelerate equipment degradation, and abnormal deviations that can easily lead to safety risks. It cannot meet the needs of intermittent hydrogen production scenarios. Therefore, it is necessary to provide a dynamic power dispatch method and system based on hydrogen production data feature feedback to solve the above-mentioned problems. Summary of the Invention

[0004] To address the aforementioned technical issues, this paper provides a dynamic power dispatching method and system based on hydrogen production data feature feedback. This solution overcomes the shortcomings of existing dynamic power dispatching methods based on hydrogen production data feature feedback, which fail to construct a current-hydrogen production correlation model covering the entire process from preheating to full load, cooling, and replacement. Furthermore, these methods neglect key parameters during non-production periods in data processing, fail to focus on emergency shutdown protection functions, and do not incorporate emergency storage parameters into grid constraints. The dispatching model lacks a coordinated objective for optimal production time across the entire process, and the closed-loop dispatching does not cover key feedback parameters such as preheating / cooling rates and nitrogen replacement effects. These shortcomings result in poor adaptability to operating conditions, frequent start-stop cycles accelerating equipment degradation, and abnormal deviations easily leading to safety risks, ultimately failing to meet the needs of intermittent hydrogen production scenarios.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A dynamic power dispatching method based on hydrogen production data feature feedback includes: Historical core data on the hydrogen production side and historical supporting data on the power supply side of the alkaline water electrolysis hydrogen production system were collected. At the same time, historical key data on the grid side in off-grid scenarios were collected. Historical data on hydrogen production skid operation, equipment status, characteristic parameters, nitrogen replacement status, alkaline solution circulation status, and standby mode operation were extracted. A dynamic function of the linear relationship between current and hydrogen production was constructed. The historical core data on the hydrogen production side includes daily full-load operation data, preheating period data, cooling period data, and nitrogen replacement data corresponding to long-term shutdown. By combining historical core data from the hydrogen production side, historical supporting data from the power supply side, and historical key data from the grid side, a set of grid constraint boundaries for off-grid-energy storage collaboration is determined. This set of constraint boundaries includes emergency support constraints for energy storage systems. By combining the dynamic function of the current-hydrogen production linear relationship and the set of grid constraint boundaries, the power regulation capability of the grid-type voltage-supported energy storage backup is evaluated, and a regulation capability assessment report is generated. A dynamic scheduling model is established with the goals of optimal hydrogen production efficiency, off-grid grid stability, minimum losses from frequent start-stop operations, and optimal production time coordination throughout the entire process. The dynamic function of the linear relationship between current and hydrogen production, the set of grid constraint boundaries, and the regulation capacity assessment report are used as inputs to solve the optimal power allocation and time coordination scheme that adapts to the entire process of preheating-full load-cooling-displacement. Based on the optimal power allocation and time coordination scheme, the following commands are generated: current regulation command for hydrogen production system, energy storage safety shutdown guarantee command, power output adjustment command, nitrogen replacement control command, and alkali circulation control command. The system executes commands for current regulation of the hydrogen production system, energy storage safety shutdown, power output adjustment, nitrogen replacement control, and alkali circulation control. It also collects feedback data in real time on hydrogen production, grid parameters, energy storage status, alkali concentration / circulation status, nitrogen replacement effect, and preheating / cooling rate. If the feedback data exceeds a preset threshold, the scheduling scheme is re-optimized. A new scheduling scheme is generated by combining the safety control logic of nitrogen replacement, alkali circulation, and soft start-stop procedures to achieve closed-loop scheduling.

[0006] In an optional embodiment, the construction of the current-hydrogen production linear relationship dynamic function specifically includes: Confirm the design life of the hydrogen production system and obtain the core historical data of the hydrogen production side over the past 3 years. The core historical data of the hydrogen production side includes the 1-hour nitrogen replacement operating condition data corresponding to 6 hours of full-load operation, 1 hour of preheating, 3 hours of cooling, and long-term shutdown. Data corresponding to periods of fault shutdown, periods of abnormal fluctuation, and periods of non-experimental shutdown were removed from the historical core data of hydrogen production over the past three years to obtain effective historical core data of hydrogen production, and the duration of effective data was determined simultaneously. Extract historical operation data of hydrogen production skid from effective historical core data of hydrogen production side. The historical operation data of hydrogen production skid includes output current and output voltage records, alkaline solution circulation status records, nitrogen replacement execution records, and temperature change records during the preheating / cooling stage once every 1 minute. According to the operating conditions, the average hourly current and the corresponding actual hydrogen production were obtained. 1000 sets of effective historical core data on the hydrogen production side were selected, which were evenly distributed and covered the preheating stage, 6 hours of full load, cooling stage and nitrogen replacement stage, to determine the initial proportional coefficient. Using historical equipment status data, a comprehensive correction factor is constructed, which includes a temperature correction factor, a pressure correction factor, and a frequent start-stop correction factor. Determine the characteristic correction factor by combining historical data of characteristic parameters; Based on the initial proportional coefficient, comprehensive correction factor and characteristic correction factor, the dynamic proportional coefficient is determined, and the basic form of the current-hydrogen production linear relationship is constructed. From the effective historical core data of hydrogen production, 500 sets of test data that were not involved in the construction of the basic form of the current-hydrogen production linear relationship were selected. The current in the test data was substituted into the basic form of the current-hydrogen production linear relationship to obtain the predicted value of hydrogen production, and the deviation rate of each set was determined. Using the deviation rate of each group, determine the dynamic function of the current-hydrogen production linear relationship that meets the requirements.

[0007] In an optional embodiment, determining the set of power grid constraint boundaries specifically includes: Acquire historical key data from the grid side, historical core data from the hydrogen production side, and historical supporting data from the power supply side over the past three years. The historical key data from the grid side includes voltage / frequency data of dual 6kV / 400V grid power and voltage / frequency data simulating green electricity fluctuations. The historical supporting data from the power supply side includes energy storage system status data and auxiliary power supply data during sudden power outages of energy storage systems. The historical core data from the hydrogen production side includes off-grid auxiliary system load data, historical power data of the hydrogen production system, load response data of the hydrogen production system, nitrogen replacement data, alkali circulation data, and standby mode operation data. The historical key data from the grid side, the historical core data from the hydrogen production side, and the historical supporting data from the power source side over the past three years were unified to a sampling frequency of 1 minute / time and sorted in chronological order to ensure that the timestamps of each data are completely matched. After the timestamps are fully matched, the physical quantities in the historical key data of the grid side, the historical core data of hydrogen production side, and the historical supporting data of the power source side are standardized and the abnormal data are cleaned to obtain the historical key standard data of the grid side, the historical core standard data of hydrogen production side, and the historical supporting standard data of the power source side. Historical data of dual-path 6kV / 400V grid voltage and simulated green electricity fluctuation voltage data were extracted from the historical key standard data of the grid side, and clustered according to the operating condition type to obtain steady-state grid voltage cluster and green electricity fluctuation cluster. Obtain the boundary threshold deviation coefficients of steady-state grid power clusters and green power fluctuation clusters, thereby constructing voltage constraint boundaries; Extract historical frequency data of dual-path 6kV / 400V grid power and simulated green electricity fluctuation frequency data from historical key standard data of the power grid side; A Gaussian mixture model was constructed based on historical frequency data of dual-path 6kV / 400V grid power and simulated green electricity fluctuation frequency data, and decomposed into steady-state components, wave dynamic components and near-limiting state components. Historical frequency data of dual-path 6kV / 400V grid power, simulated green electricity fluctuation frequency data and the Mahalanobis distance of steady-state components are obtained to determine the frequency constraint boundary; Extract energy storage system status data and emergency discharge capacity data from historical supporting standard data on the power supply side, perform energy storage emergency protection efficiency calibration, and simultaneously determine the energy storage emergency protection state of charge range and emergency discharge power limit to obtain energy storage constraint boundaries. Extract the off-grid auxiliary system load data and the historical power data of the hydrogen production system from the historical core standard data of the hydrogen production side, perform auxiliary load statistics, and obtain the average load and peak load of the preheating / cooling / displacement stage and the full-load operation stage; Based on the average and peak loads of each stage, combined with the daily routine operating conditions and the nitrogen replacement operating conditions with monthly shutdowns, the hydrogen production power ratio corresponding to each stage is obtained, thereby determining the power ratio constraint. Voltage / frequency fluctuation data for simulating green electricity are extracted from historical key standard data on the power grid side, and load response data of the hydrogen production system is extracted from historical core standard data on the hydrogen production side. Based on the voltage / frequency fluctuation data of simulated green electricity and the load response data of hydrogen production system, we classify green electricity fluctuations and impose response time constraints, and construct the coupling relationship between green electricity fluctuation amplitude and hydrogen production regulation. By summarizing voltage constraint boundaries, frequency constraint boundaries, energy storage emergency support constraints, and power ratio constraints, and combining them with the coupling relationship between green electricity fluctuation amplitude and hydrogen production regulation, a set of grid constraint boundaries is formed.

[0008] In an optional embodiment, the evaluation of the power regulation capability of grid-type voltage-supported energy storage backup, combining the dynamic function of the current-hydrogen production linear relationship and the set of grid constraint boundaries, and generating a regulation capability evaluation report, specifically includes: Based on the dual-path 6kV / 400V grid power operation log and energy storage BMS system, historical supporting data of the power supply side over the past 3 years were collected, including historical power output data of the two power sources, historical state of charge data of energy storage, emergency discharge power data of energy storage, power output of new energy sources, and power supply reliability records. The average output of each power source was determined based on historical power output data from both sources. Obtain the power output fluctuation range of each power source and determine the power output adjustment range; Based on historical state of charge data of energy storage and emergency support needs, the proportion of time interval of energy storage emergency support state of charge interval is statistically analyzed to determine the effective range of energy storage emergency discharge. Extract the historical maximum emergency discharge power from the energy storage emergency discharge power data, and set the energy storage emergency discharge power limit in combination with emergency support needs; Based on the dynamic function of the linear relationship between current and hydrogen production and the set of grid constraint boundaries, combined with the full-process operating conditions of the hydrogen production system from preheating to full load, cooling to replacement and the frequent start-stop characteristics, the power regulation range of the hydrogen production system at each stage is determined. Determine the basic power compensation capability on the power supply side and construct an evaluation index system for regulation capability; Based on the regulation capacity assessment index system, a regulation capacity assessment report is generated. The regulation capacity assessment report is updated every 24 hours based on the latest power supply side data, clarifying the power supply side regulation capacity and energy storage emergency support status under different intermittent operating conditions and start-stop frequencies.

[0009] In an optional embodiment, the establishment of a dynamic scheduling model aimed at achieving optimal hydrogen production efficiency, off-grid grid stability, minimum losses from frequent start-stop cycles, and optimal production time coordination throughout the entire process involves using the dynamic function of the current-hydrogen production linear relationship, the set of grid constraint boundaries, and the regulation capacity assessment report as inputs to solve for the optimal power allocation and time coordination scheme adapted to the entire process of preheating-full load-cooling-displacement. Specifically, this includes: Based on the dynamic function of the linear relationship between current and hydrogen production, the correlation between hydrogen production and current is transformed into model constraints. Combined with the user's daily target of 6 hours of full-load hydrogen production, it is transformed into hydrogen production power constraints for the entire process of preheating-full-load-cooling. Transform the constraint boundaries in the set of power grid constraint boundaries into model inequality constraints; Based on the regulation capability assessment report, the power output regulation range, energy storage emergency support constraints, and power regulation range of each stage of the hydrogen production system are transformed into model variable constraints. Historical core data from the hydrogen production side, historical key data from the power grid side, and historical supporting data from the power supply side are divided into training set data, validation set data, and test set data. A four-objective optimization function and a genetic algorithm solution model are set. The optimization function incorporates soft start-stop program, standby mode, and nitrogen replacement control logic. A preliminary power allocation and time coordination scheme was obtained; The initial power allocation and time coordination scheme is input into the validation set to determine the hydrogen production efficiency, grid stability index, frequent start-stop loss coefficient and overall time coordination efficiency of the validation set. The schemes that meet the verification requirements are input into the test set, and based on the dynamic function of the current-hydrogen production linear relationship, the hydrogen production power target in the tested schemes is transformed into the hydrogen production current target at each stage, and the optimal power allocation and time coordination scheme is determined.

[0010] In an optional embodiment, the generation of hydrogen production system current regulation commands, energy storage safety shutdown protection commands, power output adjustment commands, nitrogen replacement control commands, and alkali circulation control commands based on the optimal power allocation and time coordination scheme specifically includes: Based on the hydrogen production current targets at each stage in the optimal power allocation and time coordination scheme, and combined with the soft start-stop procedure requirements, the current real-time current on the hydrogen production side is obtained, and the current adjustment difference for preheating load increase, full load stabilization, and cooling load reduction is determined. Based on the maximum allowable adjustment rate of the hydrogen production system, and combined with the soft start-stop requirements of slow preheating and slow cooling, the current adjustment time for each stage is determined. By combining the hydrogen production current target and the maximum allowable adjustment rate of the hydrogen production system, a current adjustment command for the hydrogen production system is generated, which includes step-by-step adjustment parameters for each stage. Based on the energy storage emergency protection requirements in the optimal power allocation and time coordination scheme, the current state of charge of the energy storage is obtained, and an energy storage safety shutdown protection command is generated, which includes emergency triggering conditions, emergency discharge power, state of charge protection threshold, and emergency shutdown timing logic. Based on the output targets of the two power sources in the optimal power allocation and time coordination scheme, and combined with the test requirements for simulating green electricity fluctuations in the later stage, the current real-time output of the power sources is obtained, and the adjustment difference between the output of the first and second power sources is determined. Obtain historical power supply reliability coefficients to determine adjustment priorities; Based on the adjustment priority and output adjustment difference, a power output adjustment command is generated, which includes power supply number, adjustment direction, adjustment power, adjustment step size, adjustment interval, output upper limit protection and green electricity fluctuation simulation adaptation parameters. Based on the nitrogen replacement timing requirements in the optimal power allocation and time coordination scheme, a nitrogen replacement control command is generated. Based on the requirements for heat preservation of alkali solution circulation during the cooling stage, alkali solution circulation control commands are generated. Summarize the current regulation commands of the hydrogen production system, the energy storage safety shutdown protection commands, the power output adjustment commands, the nitrogen replacement control commands, and the alkaline solution circulation control commands, determine the command check codes, and add them to the end of the commands; Generate an execution schedule, clarify the order and connection of instruction execution, including the coordinated execution logic of soft start / stop, nitrogen replacement, and alkali circulation.

[0011] In an optional embodiment, the execution of the hydrogen production system current regulation command, energy storage safety shutdown guarantee command, power output adjustment command, nitrogen replacement control command, and alkali circulation control command involves real-time collection of feedback data on hydrogen production, grid parameters, energy storage status, alkali concentration / circulation status, nitrogen replacement effect, and preheating / cooling rate. If the feedback data exceeds a preset threshold, the scheduling scheme is re-optimized. A new scheduling scheme is generated by combining the safety control logic of nitrogen replacement, alkali circulation, and soft start-stop procedures to achieve closed-loop scheduling. Specifically, this includes: Based on the execution schedule, various instructions are sent to the hydrogen production skid rectifier power controller, energy storage converter controller, power control system, nitrogen replacement control system, and alkali circulation control system via industrial Ethernet, and the instruction sending timestamp is recorded. Set the data acquisition cycle, collect actual current, actual hydrogen production, alkali concentration and circulation status, and preheating / cooling rate through the hydrogen production side monitoring module, collect actual voltage, actual frequency, and total regional load through the grid side monitoring device, collect actual energy storage state of charge, actual emergency discharge power, and actual output of the two power sources through the power supply side monitoring unit, and collect replacement effect data through the nitrogen replacement monitoring unit, and record the acquisition timestamp; Based on the actual current collected and the target current commanded, the absolute value of the current execution deviation is determined; Based on the actual hydrogen production and the dynamic function of the linear relationship between current and hydrogen production, the predicted value of hydrogen production is determined, and the absolute value of the deviation of hydrogen production is obtained simultaneously. Obtain the absolute values ​​of grid voltage deviation, frequency deviation, energy storage state of charge deviation, total power output deviation, nitrogen replacement effect deviation, and preheating / cooling rate deviation. Set a preset deviation threshold and compare each deviation with the preset deviation threshold one by one; If all deviations are within the preset deviation threshold, maintain the current instruction and repeat the deviation acquisition and comparison every 5 minutes; If any deviation exceeds the threshold, an emergency interrupt is immediately triggered, an interrupt signal is sent, and all current running data and deviation status are recorded. Based on all current operational data and deviations, update the real-time datasets for hydrogen production, grid, and power generation, and redetermine the set of grid constraint boundaries. By combining the updated set of power grid constraint boundaries, the regulation capacity is reassessed, and then the optimal solution is re-solved in the dynamic scheduling model. New instructions are generated to replace the original instructions. The new instructions include nitrogen replacement, alkaline solution circulation adjustment parameters and energy storage emergency support parameters for deviations. Repeat the steps of data collection, deviation calculation, comparison and adjustment until all feedback data deviations are within the threshold range, forming a closed-loop scheduling.

[0012] Furthermore, a dynamic power dispatching system based on hydrogen production data feature feedback is proposed to implement the dispatching method described in any of the preceding claims, characterized by comprising: A multi-source data collection module is used to collect historical core data from the hydrogen production side, historical key data from the power grid side, and historical supporting data from the power source side. The data processing and core element construction module is used to construct a dynamic function of the current-hydrogen production linear relationship covering the entire process of preheating-full load-cooling-replacement based on extracted historical operating data, equipment status data, characteristic parameter data, and relevant data on nitrogen replacement, alkaline circulation, and standby mode. This function is then used to analyze and process collected historical key data from the grid side and historical supporting data from the power source side, determining a set of grid constraint boundaries including energy storage emergency support constraints and full-process power constraints. Finally, this module combines the current-hydrogen production linear relationship dynamic function and the grid constraint boundary set to quantitatively analyze historical supporting data from the power source side, assessing the power regulation capability including energy storage safety shutdown protection and generating a regulation capability assessment report, thus providing accurate input elements for the dynamic scheduling model. The dynamic scheduling model solving module is used to establish a dynamic scheduling model with the goals of optimal hydrogen production efficiency, grid stability, minimum frequent start-stop losses, and optimal production time coordination throughout the entire process. The inputs include the dynamic function of the linear relationship between current and hydrogen production, the set of grid constraint boundaries, and relevant data from the regulation capability assessment report. The optimal power allocation and time coordination scheme that adapts to the entire process of preheating-full load-cooling-replacement is obtained through model solving. The instruction generation and execution module is used to generate hydrogen production system current regulation instructions, energy storage safety shutdown guarantee instructions, power output adjustment instructions, nitrogen replacement control instructions and alkali circulation control instructions based on the optimal power allocation and time coordination scheme output by the dynamic scheduling model solution module, and accurately send the instructions to the corresponding execution terminals to drive each system to operate according to the scheduling requirements. The feedback adjustment and closed-loop optimization module is used to collect feedback data in real time during instruction execution, including hydrogen production, grid parameters, energy storage status, nitrogen replacement effect, alkali circulation status, and preheating / cooling rate. The feedback data is compared with preset thresholds. If the feedback data exceeds the preset thresholds, a re-optimization process is triggered, driving the system to return to the data processing or model solving stage to regenerate the scheduling scheme, thereby achieving closed-loop dynamic scheduling.

[0013] In an optional embodiment, the data processing and core element construction module includes: The dynamic function construction unit is used to perform linear relationship fitting and correction based on the historical operation data, equipment status data, and characteristic parameter data of the hydrogen production skid extracted by the multi-source data collection module, combined with the preheating / cooling temperature correction factor and the frequent start-stop correction factor, to construct a dynamic function of current-hydrogen production linear relationship covering the entire process of preheating-full load-cooling-displacement. The constraint boundary determination unit is used to receive the current-hydrogen production linear relationship dynamic function output by the dynamic function construction unit, and combine it with the historical key data of the grid side and the historical supporting data of the power source side collected by the multi-source data collection module. Through data statistics and threshold analysis, the set of grid constraint boundaries including energy storage emergency guarantee constraints and full-process power constraints is determined. The regulation capability assessment unit is used to synchronously receive the current-hydrogen production linear relationship dynamic function from the dynamic function construction unit and the grid constraint boundary set from the constraint boundary determination unit, perform quantitative analysis on the historical supporting data of the power supply side of the multi-source data collection module, assess the power supply regulation capability including energy storage safety shutdown protection, and generate a regulation capability assessment report.

[0014] In an optional embodiment, the dynamic scheduling model solving module includes: The model building unit is used to clarify the four optimization directions of optimal hydrogen production efficiency, grid stability, minimum loss from frequent start-stop, and optimal production time coordination throughout the process. It incorporates control logic of soft start-stop program, standby mode, nitrogen replacement, and alkaline solution circulation, sets objective function and constraints, and establishes a dynamic scheduling model adapted to the coordinated scheduling of hydrogen production, grid and power supply. The model solving unit receives the current-hydrogen production linear relationship dynamic function, grid constraint boundary set, and regulation capability assessment report output by the data processing and core element construction module, and uses them as core input parameters to input into the dynamic scheduling model. It then solves for the optimal power allocation and time coordination scheme by combining the requirements of the entire process of preheating-full load-cooling-displacement.

[0015] Compared with the prior art, the beneficial effects of the present invention are: This proposal presents a dynamic power dispatching method based on hydrogen production data feature feedback. By collecting full-process operation data of hydrogen production side (preheating-full load-cooling-displacement), grid side dual-path 6kV / 400V grid power and simulated green electricity fluctuation data, and power supply side energy storage emergency support data, a dynamic function of current-hydrogen production linear relationship covering the entire process is constructed. The set of grid constraint boundaries including energy storage emergency constraints is determined, realizing the accuracy and scenario-based nature of the dispatching model input elements in off-grid intermittent hydrogen production scenarios. This provides core basis for solving the optimal dispatching scheme in accordance with actual working conditions and avoids the problem of insufficient dispatching adaptability caused by data partiality or constraint deviation. This proposal presents a dynamic power dispatching method based on hydrogen production data feature feedback. By establishing a four-objective dynamic dispatching model that optimizes hydrogen production efficiency, grid stability, minimizes losses from frequent start-stop operations, and achieves optimal production time coordination throughout the entire process, a soft start-stop procedure and multi-system coordination logic are incorporated. Combined with a genetic algorithm, the optimal power allocation and time coordination scheme adapted to the entire process is solved. This achieves a multi-objective balance between hydrogen production demand, grid operation safety, equipment loss control, and time coordination throughout the entire process. It clarifies that energy storage only undertakes the function of ensuring safe shutdown, avoiding the limitations of single-objective optimization and dispatching deviations caused by misjudgment of energy storage function. This solution proposes a dynamic power dispatching method based on hydrogen production data feature feedback. By issuing multi-system collaborative instructions according to the execution schedule, it collects feedback data from the entire hydrogen production process, the power grid, power sources, and auxiliary systems in real time, calculates multi-dimensional deviations and compares them with thresholds. When the deviation exceeds the limit, it triggers an emergency interruption, updates data and constraints, and re-optimizes the scheme through a closed-loop mechanism. This achieves dynamic deviation correction and safe and stable operation of the entire intermittent hydrogen production process, effectively mitigates the damage to electrolyzers caused by sudden temperature changes and frequent start-stops, ensures the anti-interference capability of the dispatching scheme under green power fluctuation scenarios, and extends equipment life. Attached Figure Description

[0016] Figure 1 This is a flowchart of a dynamic power dispatching method based on hydrogen production data feature feedback proposed in this invention; Figure 2 This is a flowchart illustrating the process of obtaining the dynamic function of the current-hydrogen production linear relationship in this invention. Figure 3 This is a flowchart illustrating the process of obtaining the optimal power allocation scheme in this invention. Figure 4 This is a system framework diagram of a dynamic power dispatching system based on hydrogen production data feature feedback proposed in this invention. Detailed Implementation

[0017] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0018] Reference Figure 1 - Figure 4 As shown, a dynamic power dispatching method based on hydrogen production data feature feedback includes: Collect historical core data on the hydrogen production side and historical supporting data on the power supply side of the alkaline water electrolysis hydrogen production system, and at the same time collect historical key data on the grid side in off-grid scenarios; The historical core data of hydrogen production includes daily full-load operation data, preheating period data (about 1 hour), cooling period data (about 3 hours), and nitrogen replacement data corresponding to long-term shutdown (about 1 hour). It also extracts historical data of hydrogen production skid operation, equipment status, characteristic parameters, nitrogen replacement status data, alkali circulation status data, and standby mode operation data to construct a dynamic function of current-hydrogen production linear relationship covering the entire process of preheating-full load-cooling-replacement. By combining historical core data from the hydrogen production side, historical supporting data from the power supply side, and historical key data from the grid side, the set of grid constraint boundaries for off-grid-energy storage synergy is determined. The set of constraint boundaries includes emergency protection constraints for energy storage systems (energy storage only undertakes the function of ensuring safe shutdown and does not participate in daily power regulation). By combining the dynamic function of the current-hydrogen production linear relationship and the set of grid constraint boundaries, the power regulation capability including energy storage safety shutdown protection (excluding daily power regulation assessment of energy storage) is evaluated, and a regulation capability assessment report is generated. A dynamic scheduling model is established with the goals of optimal hydrogen production efficiency, off-grid grid stability, minimum losses from frequent start-stop operations, and optimal production time coordination throughout the entire process. The dynamic function of the linear relationship between current and hydrogen production, the set of grid constraint boundaries, and the regulation capacity assessment report are used as inputs to solve the optimal power allocation and time coordination scheme that adapts to the entire process of preheating-full load-cooling-displacement. Based on the optimal power allocation and time coordination scheme, the following commands are generated: current regulation command for hydrogen production system (including preheating load increase, full load stabilization, and cooling load reduction step parameters), energy storage safety shutdown guarantee command (triggered only in case of sudden power failure), power output adjustment command, nitrogen replacement control command, and alkali circulation control command. The system executes commands for hydrogen production system current regulation, energy storage safety shutdown, power output adjustment, nitrogen replacement control, and alkali circulation control. It also collects real-time feedback data on hydrogen production, grid parameters, energy storage status, alkali concentration / circulation status, nitrogen replacement effect, and preheating / cooling rate. If the feedback data exceeds a preset threshold, the system re-optimizes the scheduling scheme and generates a new scheduling scheme by combining the safety control logic of nitrogen replacement, alkali circulation, and soft start-stop procedures, thus achieving closed-loop scheduling.

[0019] Furthermore, a dynamic function for the linear relationship between current and hydrogen production is constructed, specifically including: Confirm the design life of the hydrogen production system and obtain the core historical data of the hydrogen production side over the past 3 years. The core historical data of the hydrogen production side includes the 1-hour nitrogen replacement operating condition data corresponding to 6 hours of full-load operation, 1 hour of preheating, 3 hours of cooling, and long-term shutdown (every month). Data corresponding to periods of fault shutdown, periods of abnormal fluctuation, and periods of non-experimental shutdown were removed from the historical core data of hydrogen production over the past three years to obtain effective historical core data of hydrogen production, and the duration of effective data was determined simultaneously. Extract historical data of hydrogen production skid operation from effective historical core data of hydrogen production side. Historical data of hydrogen production skid operation includes output current and output voltage records, alkaline solution circulation status records, nitrogen replacement execution records, and temperature change records during the preheating / cooling stage once every 1 minute. According to the operating conditions, the average hourly current and the corresponding actual hydrogen production were obtained. 1000 sets of effective historical core data on the hydrogen production side were selected, which were evenly distributed and covered the preheating stage, 6 hours of full load, cooling stage and nitrogen replacement stage, to determine the initial proportional coefficient. Using historical equipment status data, a comprehensive correction factor is constructed, which includes a temperature correction factor (focusing on the preheating and cooling ranges), a pressure correction factor, and a frequent start-stop correction factor; combined with historical characteristic parameter data, characteristic correction factors are determined. Based on the initial proportional coefficient, comprehensive correction factor and characteristic correction factor, the dynamic proportional coefficient is determined, and the basic form of the current-hydrogen production linear relationship is constructed. 500 sets of test data (including preheating, full load, cooling, and nitrogen purging conditions) that were not involved in the construction of the current-hydrogen production linear relationship were selected from the effective historical core data of hydrogen production. The current in the test data was substituted into the current-hydrogen production linear relationship to obtain the predicted value of hydrogen production, and the deviation rate of each set was determined. Using the deviation rate of each group, determine the dynamic function of the current-hydrogen production linear relationship that meets the requirements (the dynamic coefficient increases with temperature in the preheating stage and decreases with temperature in the cooling stage).

[0020] Specifically, the initial proportional coefficient determination process first requires selecting 1000 sets of data according to the principle of "uniform coverage of operating conditions." For example, under a 6-hour full-load operating condition, 400 sets of data are selected, showing a current stable at 5660A±5% (5377~5943A) and a hydrogen production stable at 200Nm³ / h±5% (190~210Nm³ / h), with 67 sets taken per hour. For a shutdown transition condition, 300 sets of data are selected, showing the process of the current decreasing from full load to 0A (e.g., current 5660A→5000A→4000A→…→0A, with one set taken every 10 minutes). For a standby condition (if there is low-power standby), 300 sets of low-power operating data with a current of 100~200A are selected, with 50 sets taken per hour.

[0021] Understandably, the initial proportional coefficient is calculated using linear regression, with the following formula: in, For the first The hourly average current (A) of the data set. This corresponds to the actual hydrogen production rate per hour (Nm³ / h). The average current of 1000 data sets, The average hydrogen production rate is calculated based on 1000 sets of data, with the result accurate to 0.001 Nm³ / (A·h).

[0022] The comprehensive correction factor is constructed based on historical equipment status data (temperature, pressure, number of start-ups and shutdowns). Each correction factor is calculated separately and then multiplied together to obtain the comprehensive correction factor. For the temperature correction factor, when the electrolytic cell operating temperature T is between 70 and 90℃ (optimal operating condition), k_temp = 1.0; when T < 70℃, k_temp = 1 + (70 − T) × 0.003 (e.g., when T = 65℃, k_temp = 1.015); when T > 90℃, k_temp = 1 − (T − 90) × 0.005 (e.g., when T = 95℃, k_temp = 0.975). For the pressure correction factor, when the system operating pressure P is between 1 and 3 MPa (safe operating condition), k_press = 1.0; when P < 1 MPa, k_press = 1 + (1 − P) × 0.008 (e.g., when P = 0.8 MPa, k_press = 1.016); when P > 3 MPa, k_press = 1 − (P − 3) × 0.01 (e.g., when P = 3.2 MPa, k_press = 0.98). For the frequent start-stop correction factor, the number of start-stop cycles N per month needs to be calculated. When N ≤ 10 times, k_start = 1.0; when 10 < N ≤ 20 times, k_start = 0.98; when N > 20 times, k_start = 0.95 (based on the characteristic that frequent start-stop cycles accelerate electrode attenuation). The final comprehensive correction factor is k_com = k_temp × k_press × k_start. The characteristic correction factor is determined based on historical data of characteristic parameters (operating parameters of the new nickel electrode and the new chamber sealing device), and the characteristic correction factor (k_char) is calculated using the "two-factor product method". The formula for calculating the dynamic proportional coefficient is: k = k_0 × k_com × k_char, where k_0 is the initial proportional coefficient, k_com is the comprehensive correction factor, and k_char is the characteristic correction factor.

[0023] The basic form of the linear relationship between current and hydrogen production is: H = k × I, where H is the hydrogen production rate (Nm³ / h), I is the output current of the electrolyzer (A), and k is the dynamic proportional coefficient (Nm³ / (A·h)).

[0024] Specifically, 500 sets of data that were not used in the initial proportional coefficient calculation were selected from the valid data, covering full load, shutdown transition, and standby (if any) conditions, with no overlap in the timestamps of the data sets. For the prediction of hydrogen production and the acquisition of the deviation rate, the hourly average current I_test from the test data was substituted into the basic form of the current-hydrogen production linear relationship to obtain the predicted hydrogen production value H_pred=k×I_test;

[0025] Through the formula: Calculate the deviation rate for each set of data, where, The actual hydrogen production in the test data was then used to obtain the average deviation rate: .like If ≤3%, then the basic form meets the accuracy requirements; if If the deviation rate is greater than 3%, then increase the number of valid data sets to 1500 and recalculate the initial proportional coefficient and each correction factor until the average deviation rate is ≤3%.

[0026] Understandably, after meeting the deviation rate requirements, the dynamic function is determined as: H(t) = k(t) × I(t), where t is the time dimension (updated daily), k(t) is the daily dynamic ratio coefficient (k_com and k_char are recalculated daily based on the equipment status data and characteristic parameter data of the previous day, and then k(t) is updated), I(t) is the real-time output current (A), and H(t) is the real-time hydrogen production rate (Nm³ / h). Simultaneously, a function update mechanism is set up to recalibrate k_0, k_com, and k_char every 7 days based on historical data from the past 7 days, ensuring that the function adapts to equipment aging and changes in operating conditions (such as electrolyzer temperature fluctuations caused by seasonal temperature changes) and maintains accuracy.

[0027] Furthermore, the set of power grid constraint boundaries is determined, specifically including: Acquire historical key data from the grid side, historical core data from the hydrogen production side, and historical supporting data from the power supply side over the past three years. The historical key data from the grid side includes voltage / frequency data of dual 6kV / 400V grid power and voltage / frequency data simulating green electricity fluctuations. Historical data on the power supply side includes energy storage system status data (state of charge, emergency discharge capacity) and auxiliary power supply data during sudden power outages of energy storage (excluding daily charging and discharging data). The core historical data for hydrogen production includes off-grid auxiliary system load data (including preheating / cooling auxiliary machine load), historical power data of hydrogen production system (by operating condition), hydrogen production system load response data, nitrogen replacement data, alkali circulation data, and standby mode operation data. The historical key data from the grid side, the historical core data from the hydrogen production side, and the historical supporting data from the power source side over the past three years were unified to a sampling frequency of 1 minute / time and sorted in chronological order to ensure that the timestamps of each data are completely matched. After the timestamps are fully matched, the physical quantities in the historical key data of the grid side, the historical core data of hydrogen production side, and the historical supporting data of the power source side are standardized and the abnormal data are cleaned to obtain the historical key standard data of the grid side, the historical core standard data of hydrogen production side, and the historical supporting standard data of the power source side. Historical data of dual-path 6kV / 400V grid voltage and simulated green electricity fluctuation voltage data were extracted from the historical key standard data of the grid side, and clustered according to the operating condition type to obtain steady-state grid voltage cluster and green electricity fluctuation cluster. Obtain the boundary threshold deviation coefficients of steady-state grid power clusters and green electricity fluctuation clusters to construct voltage constraint boundaries; extract historical 6kV / 400V grid frequency data and simulated green electricity fluctuation frequency data from historical key standard data of the grid side; A Gaussian mixture model was constructed based on historical frequency data of dual-path 6kV / 400V grid power and simulated green electricity fluctuation frequency data, and decomposed into steady-state components, wave dynamic components and near-limiting state components. Historical frequency data of dual-path 6kV / 400V grid power, simulated green electricity fluctuation frequency data and the Mahalanobis distance of steady-state components are obtained to determine the frequency constraint boundary; Extract energy storage system status data and emergency discharge capacity data from historical supporting standard data of the power supply side, perform energy storage emergency protection efficiency calibration, and simultaneously determine the energy storage emergency protection charge state range (to ensure that it can support 3 hours of cooling + 1 hour of replacement in the event of a sudden power outage) and emergency discharge power limit (only to meet the operation of auxiliary equipment) to obtain the energy storage constraint boundary. Extract the off-grid auxiliary system load data and the historical power data of the hydrogen production system from the historical core standard data of the hydrogen production side, perform auxiliary load statistics, and obtain the average load and peak load of the preheating / cooling / displacement stage and the full-load operation stage; Based on the average and peak loads of each stage, and combined with the routine operating condition of 1 hour of preheating + 6 hours of full load + 3 hours of cooling per day and the nitrogen replacement operating condition of shutdown every other month, the hydrogen production power ratio of each stage is obtained respectively, so as to determine the power ratio constraint. Voltage / frequency fluctuation data for simulating green electricity are extracted from historical key standard data on the power grid side, and load response data of the hydrogen production system is extracted from historical core standard data on the hydrogen production side. Based on voltage / frequency fluctuation data of simulated green electricity and load response data of hydrogen production system, green electricity fluctuation is classified and response time is constrained, and the coupling relationship between green electricity fluctuation amplitude and hydrogen production regulation is constructed (energy storage does not participate in daily energy replenishment). By summarizing voltage constraint boundaries, frequency constraint boundaries, energy storage emergency support constraints, and power ratio constraints, and combining them with the coupling relationship between green electricity fluctuation amplitude and hydrogen production regulation, a set of grid constraint boundaries is formed.

[0028] Specifically, the process involves comprehensively collecting three types of historical data over the past three years: grid-side data (voltage / frequency data of dual-path 6kV / 400V grid power and simulated green electricity fluctuations), power supply-side data (energy storage system status, emergency discharge capacity, and auxiliary power supply data during sudden power outages), and hydrogen production-side data (including preheating / cooling auxiliary load, power under different operating conditions, nitrogen replacement, and other full-process data). This data is then standardized to a sampling frequency of 1 minute per sampling and timestamps are calibrated. After physical quantity standardization and anomaly data cleaning, voltage constraint boundaries are constructed through operating condition clustering. Frequency constraint boundaries are determined using Gaussian mixture model decomposition and Mahalanobis distance analysis. Energy storage constraint boundaries are clarified through energy storage emergency support efficiency calibration. Power proportion constraints are determined by combining load statistics at each stage. Finally, a coupling relationship between green electricity fluctuation amplitude and hydrogen production regulation is constructed based on green electricity fluctuation data and hydrogen production load response data. All constraints and coupling relationships are then summarized to form a set of grid constraint boundaries. Through standardized processing and precise screening of multi-source data, the reliability of the data source for the constraint boundaries is ensured. Each constraint dimension comprehensively covers the power demand of the entire hydrogen production process, including grid operation, energy storage guarantee, and green electricity fluctuation adaptation scenarios. The constructed coupling relationship between green electricity fluctuation and hydrogen production regulation is adapted to the characteristics of unstable power sources. The resulting set of constraint boundaries provides accurate and practical input for the subsequent dynamic scheduling model, effectively supporting the scheduling scheme to adapt to the power and time coordination requirements of the entire intermittent hydrogen production process while ensuring grid stability.

[0029] Furthermore, by combining the dynamic function of the current-hydrogen production linear relationship and the set of grid constraint boundaries, the power supply regulation capability, including energy storage safety outage protection, is evaluated, and a regulation capability assessment report is generated, specifically including: Based on the dual-path 6kV / 400V grid power operation log and energy storage BMS system, historical supporting data of the power supply side over the past 3 years were collected, including historical power output data of the two power sources, historical state of charge data of energy storage, emergency discharge power data of energy storage, power output of new energy sources, and power supply reliability records. The average output of each power source was determined based on historical power output data from both sources. Obtain the power output fluctuation range of each power source and determine the power output adjustment range; Based on historical state of charge data of energy storage and emergency support needs, the duration of the energy storage emergency support state of charge interval is statistically analyzed to determine the effective range of energy storage emergency discharge (to ensure support for the entire process of cooling and replacement). Extract the historical maximum emergency discharge power from the energy storage emergency discharge power data, and set the energy storage emergency discharge power limit in combination with emergency support needs (only to meet the operation of auxiliary equipment, such as cooling pumps and displacement devices). Based on the dynamic function of the linear relationship between current and hydrogen production and the set of grid constraint boundaries, combined with the full-process operating conditions of the hydrogen production system from preheating to full load, cooling to replacement and the frequent start-stop characteristics, the power adjustment range of the hydrogen production system in each stage is determined (gradually increasing the load in the preheating stage and gradually decreasing the load in the cooling stage). Determine the basic power compensation capacity of the power supply side (excluding daily compensation from energy storage) and construct an evaluation index system for regulation capacity (emphasizing the time-coordinated adaptability of the entire process). Based on the regulation capacity assessment index system, a regulation capacity assessment report is generated. The regulation capacity assessment report is updated every 24 hours based on the latest power supply side data, clarifying the power supply side regulation capacity and energy storage emergency support status under different intermittent operating conditions and start-stop frequencies.

[0030] Furthermore, a dynamic scheduling model is established with the objectives of optimal hydrogen production efficiency, off-grid grid stability, minimum losses from frequent start-stop operations, and optimal production time coordination throughout the entire process. This model is used to solve for the optimal power allocation and time coordination scheme that adapts to the entire process, specifically including: Based on the dynamic function of the linear relationship between current and hydrogen production, the correlation between hydrogen production and current is transformed into a model constraint. Combined with the user's daily target of 6 hours of full-load hydrogen production, it is transformed into a hydrogen production power constraint for the entire process of preheating-full-load-cooling (it is clarified that the hydrogen production during the preheating period is not included in the effective output, and the focus is on optimizing the preheating time and the load increase rate). The constraints in the set of power grid constraint boundaries are transformed into model inequality constraints; based on the regulation capacity assessment report, the power output regulation range, energy storage emergency support constraints, and power regulation range of each stage of the hydrogen production system are transformed into model variable constraints. Historical core data of hydrogen production, historical key data of power grid, and historical supporting data of power supply are divided into training set, validation set and test set. A four-objective optimization function and a genetic algorithm solution model are set. The optimization function incorporates soft start-stop program (preheating and slow load increase, cooling and slow load decrease), standby mode and nitrogen replacement control logic. A preliminary power allocation and time coordination scheme was obtained; The initial power allocation and time coordination scheme is input into the validation set to determine the hydrogen production efficiency, grid stability index, frequent start-stop loss coefficient, and overall time coordination efficiency (such as preheating time and cooling rate). The schemes that meet the verification requirements are input into the test set, and based on the dynamic function of the linear relationship between current and hydrogen production, the hydrogen production power target in the tested schemes is transformed into the hydrogen production current target for each stage, and the optimal power allocation and time coordination scheme is determined (the start and end times of each stage, current / power parameters, and emergency energy storage protection trigger conditions are clearly defined).

[0031] Specifically, using the dual-path 6kV / 400V grid operation logs and the energy storage BMS system as data sources, historical power-side supporting data from the past three years was collected, including historical output of the two power sources, historical state of charge of energy storage, emergency discharge power of energy storage, output of new energy sources, and power reliability records. First, the average output and fluctuation range of each power source were calculated using the historical output data of the two power sources to clarify the power output adjustment range. Then, combined with the historical state of charge data of energy storage and emergency support needs (supporting the entire process of cooling and replacement), the proportion of the energy storage emergency support state of charge interval was statistically analyzed to determine the effective range of emergency discharge. The historical maximum value was extracted from the emergency discharge power data of energy storage and combined with auxiliary equipment. Emergency discharge power limits are set based on operational requirements. Subsequently, based on the established dynamic function of the current-hydrogen production linear relationship and the set of grid constraint boundaries, and combined with the full-process operating conditions of the hydrogen production system ("preheating-full load-cooling-displacement") and frequent start-stop characteristics, the power adjustment range of the hydrogen production system at each stage is divided (gradual load increase during preheating and gradual load decrease during cooling). Finally, the basic power compensation capacity of the power supply side excluding daily compensation from energy storage is determined, and an evaluation index system for adjustment capacity that highlights the time-coordinated adaptability of the entire process is constructed. Based on this, an adjustment capacity evaluation report is generated and updated every 24 hours based on the latest power supply side data, clarifying the power supply side adjustment capacity and energy storage emergency support status under different intermittent operating conditions and start-stop frequencies. Its beneficial effects are as follows: by accurately locating data sources and screening core data, the authenticity and relevance of the power supply regulation capability assessment are ensured; the clear definition of the power output regulation range and energy storage emergency parameters (range and power limit) not only provides input parameters that conform to the actual operating capacity for the subsequent dynamic scheduling model, but also strengthens the functional positioning of energy storage only to ensure safe shutdown, avoiding confusion with daily power regulation; the division of power regulation ranges for each stage of hydrogen production is adapted to the entire process conditions, and the evaluation index system highlights the time-coordinated adaptability, accurately meeting the core needs of intermittent hydrogen production; the daily updated evaluation report can reflect the changes in the power supply and energy storage status in real time, ensuring the timeliness of the evaluation results and the adaptability of the dynamic scheduling scheme, laying the foundation for the safe and efficient solution of the subsequent scheduling scheme.

[0032] Specifically, based on the established dynamic function of the linear relationship between current and hydrogen production, H(t) = k(t) × I(t) (where H is the real-time hydrogen production rate in Nm³ / h; k(t) is the daily updated dynamic proportionality coefficient, with an example value of 0.035 Nm³ / (A·h); and I is the real-time current of the electrolyzer in A), and combined with the rated output voltage of the hydrogen production system U = 176V, the formula relating hydrogen production power and current is derived: P H (t) = U × I(t) = 176 × I(t), which can be further transformed to obtain I(t) = 176P. H (t). Let I(t) = 176P H Substituting (t) into the dynamic function, a direct relationship is formed between hydrogen production and power: H(t) = k(t) × 176P H (t).

[0033] Daily hydrogen production demand target conversion: Assuming the user needs to operate at full load (200 Nm³ / h) for 6 hours and at low load (100 Nm³ / h) for 2 hours daily, the total daily hydrogen production demand is H_r = 200 × 6 + 100 × 2 = 1400 Nm³. Combining this with the above relationship between hydrogen production and power, the result is a daily hydrogen production power integral constraint:

[0034] Meanwhile, the hydrogen production power needs to match the hydrogen production demand during the operating period (8 hours). When k(t)=0.035, the power range is 176×0.035100≈502.8 kW to 176×0.035200≈1005.7 kW.

[0035] The production intervals between months require conversion. If the system operates at standby power during shutdown months (such as June and December each year), and considering the soft start-stop logic, the standby power P_H,d = 26.4kW (corresponding to a standby current of 150A). Therefore, the power constraint during shutdown periods is P_H,d = 26.4kW. H (t)≤26.4 kW. Furthermore, before each monthly shutdown, a process of "reducing load to standby power → nitrogen purging → alkaline solution circulation and pressure maintenance" must be performed, which translates to a constraint: Within one hour before shutdown, P H (t) The operating power needs to be reduced from 1000kW to 26.4kW in steps, and the load reduction rate is ≤20% / min (e.g., it takes ≥5 minutes to reduce from 1000kW to 26.4kW).

[0036] It is understandable to transform the constraint boundaries in the set of power grid constraint boundaries into model inequality constraints. For example, the voltage constraint "6kV power grid: 5.91kV-6.09kV; 400V power grid: 394V-406V" can be transformed into "5.91≤U6kV(t)≤6.09 (unit: kV)" and "394≤U400V(t)≤406 (unit: V)".

[0037] Furthermore, the construction of two of the objectives in the four-objective optimization function includes: Objective 1 (Maximize hydrogen production efficiency): in, The auxiliary system power (such as pure water, cooling system, about 100kW) is T, and the scheduling cycle (24 hours) is T.

[0038] Objective 2 (Maximizing power grid stability): in, (6kV / 400V) (50Hz) is the rated value.

[0039] Constraint embedding is also required. The soft start / stop logic is embedded through "power change rate ≤ 20% / min", and the standby mode is embedded through "stop period". "Embedding ensures that the optimization process conforms to actual operating specifications."

[0040] Furthermore, based on the optimal power allocation and time coordination scheme, current regulation commands for the hydrogen production system, energy storage safety shutdown protection commands, power output adjustment commands, nitrogen replacement control commands, and alkali circulation control commands are generated, specifically including: Based on the hydrogen production current targets at each stage in the optimal power allocation and time coordination scheme, and combined with the soft start-stop procedure requirements, the current real-time current on the hydrogen production side is obtained, and the current adjustment difference for preheating load increase, full load stabilization, and cooling load reduction is determined. Based on the maximum allowable adjustment rate of the hydrogen production system, and combined with the soft start-stop requirements of slow preheating load increase (to avoid sudden temperature rise) and slow cooling load decrease (to avoid sudden temperature drop), the current adjustment time for each stage is determined (1 hour for preheating load increase and 3 hours for cooling load decrease). By combining the hydrogen production current target and the maximum allowable adjustment rate of the hydrogen production system, a current adjustment command for the hydrogen production system is generated, which includes step adjustment parameters for each stage (such as the load increase every 10 minutes in the preheating stage and the load decrease every 20 minutes in the cooling stage). Based on the energy storage emergency protection requirements in the optimal power allocation and time coordination scheme, the current state of charge of the energy storage is obtained, and an energy storage safety shutdown protection command is generated, including emergency triggering conditions (grid power interruption), emergency discharge power, state of charge protection threshold (ensuring support for cooling + replacement), and emergency shutdown timing logic (prioritizing the protection of the cooling system, and then initiating nitrogen replacement). Based on the output targets of the two power sources in the optimal power allocation and time coordination scheme, and combined with the test requirements for simulating green electricity fluctuations in the later stage, the current real-time output of the power sources is obtained, and the adjustment difference between the output of the first and second power sources is determined. Obtain historical power supply reliability coefficients to determine adjustment priorities; Based on the adjustment priority and output adjustment difference, a power output adjustment command is generated, which includes power supply number, adjustment direction, adjustment power, adjustment step size, adjustment interval, output upper limit protection and green electricity fluctuation simulation adaptation parameters. Based on the nitrogen replacement timing requirements in the optimal power allocation and time coordination scheme, nitrogen replacement control commands (including replacement start timing, number of replacements, and replacement duration) are generated. Based on the requirements for alkali circulation and heat preservation during the cooling stage, alkali circulation control instructions (including circulation rate and heat preservation temperature threshold during the cooling stage) are generated. Summarize the current regulation commands of the hydrogen production system, the energy storage safety shutdown protection commands, the power output adjustment commands, the nitrogen replacement control commands, and the alkaline solution circulation control commands, determine the command check codes, and add them to the end of the commands; Generate an execution schedule, clarifying the order and connection of instruction execution (such as automatically switching to full-load operation after preheating, seamlessly starting the cooling program after full-load operation, and triggering nitrogen replacement after cooling during long-term shutdown), including the coordinated execution logic of soft start / stop, nitrogen replacement, and alkali circulation.

[0041] Specifically, based on the optimal power allocation and time coordination scheme, various scheduling instructions are generated in different dimensions. For the hydrogen production system, firstly, based on the hydrogen production current targets at each stage in the scheme and combined with the soft start-stop procedure requirements, the current real-time current is obtained, and the current regulation difference for preheating load increase, full load stabilization, and cooling load reduction is calculated. Then, according to the maximum allowable regulation rate of the hydrogen production system and the requirements of "slow preheating load increase to prevent sudden temperature rise and slow cooling load reduction to prevent sudden temperature drop", the regulation time for 1 hour of preheating load increase and 3 hours of cooling load reduction is determined. Finally, current regulation instructions containing step-by-step regulation parameters for each stage (such as the load increase increment every 10 minutes during preheating and the load reduction increment every 20 minutes during cooling) are generated. For the energy storage system, based on the energy storage emergency protection requirements in the scheme, the current state of charge is obtained, and emergency triggering conditions for "grid power outage" are generated, along with emergency discharge power, state of charge protection thresholds supporting cooling + replacement, and "priority protection for the restart of the cooling system". The system generates safety shutdown instructions based on the "nitrogen replacement" emergency sequence logic. For the power system, it calculates the adjustment difference based on the output targets of the two power sources in the plan and the requirements of the subsequent simulated green electricity fluctuation test. It then determines the adjustment priority by combining the historical power reliability coefficient and generates output adjustment instructions containing power source number, adjustment direction, power, step size, interval, output upper limit protection, and green electricity fluctuation simulation adaptation parameters. For the auxiliary system, it generates control instructions containing start-up timing, number of times, and duration based on the nitrogen replacement sequence requirements in the plan. It also generates control instructions containing circulation rate and insulation temperature threshold based on the alkali circulation insulation requirements during the cooling stage. Finally, it summarizes all instructions and adds check codes to ensure accurate transmission, generating a clear instruction execution sequence and connection relationship (such as automatic switching to full load after preheating, seamless start of cooling after full load, and triggering replacement after cooling during long-term shutdown) and an execution schedule for soft start / stop and auxiliary system coordination logic.

[0042] Understandably, all commands are generated in close alignment with the entire process of "preheating-full load-cooling-replacement" and the emergency support positioning of energy storage. The stepped current regulation command of the hydrogen production system can avoid damage to the electrodes and diaphragms of the electrolyzer due to sudden temperature changes, thus extending equipment life. The energy storage command strictly limits emergency scenarios and support logic, ensuring that energy storage focuses solely on safe shutdown functions and does not interfere with daily scheduling. The power supply command takes into account both reliability priorities and adaptability to green electricity fluctuations, reserving compatibility space for subsequent unstable power supply tests. The auxiliary system command accurately matches the timing requirements of the entire process, avoiding the risks of alkaline contamination and hydrogen-oxygen mixing caused by intermittent shutdowns. The command checksum and execution schedule respectively ensure the accuracy of command transmission and the coordination of execution, avoiding conflicts between multiple system operations. Ultimately, this achieves integrated linkage of hydrogen production, energy storage, power supply, and auxiliary systems, ensuring the safe and efficient implementation of the scheduling plan and adapting to the core scenario requirements of intermittent hydrogen production.

[0043] Furthermore, it executes various instructions and implements closed-loop scheduling, specifically including: Based on the execution schedule, various instructions are sent to the hydrogen production skid rectifier power controller, energy storage converter controller, power control system, nitrogen replacement control system, and alkali circulation control system via industrial Ethernet, and the instruction sending timestamp is recorded. Set the data acquisition cycle, collect actual current, actual hydrogen production, alkali concentration and circulation status, and preheating / cooling rate through the hydrogen production side monitoring module, collect actual voltage, actual frequency, and total regional load through the grid side monitoring device, collect actual energy storage state of charge, actual emergency discharge power (collected only when triggered), and actual output of the two power sources through the power supply side monitoring unit, and collect replacement effect data through the nitrogen replacement monitoring unit, and record the collection timestamp; Based on the actual current collected and the target current commanded, the absolute value of the current execution deviation is determined; Based on the actual hydrogen production and the dynamic function of the linear relationship between current and hydrogen production, the predicted value of hydrogen production is determined, and the absolute value of the deviation of hydrogen production is obtained simultaneously. Obtain the absolute values ​​of grid voltage deviation, frequency deviation, energy storage state of charge deviation, total power output deviation, nitrogen replacement effect deviation, and preheating / cooling rate deviation. Set a preset deviation threshold and compare each deviation with the preset deviation threshold one by one; If all deviations are within the preset deviation threshold, maintain the current instruction and repeat the deviation acquisition and comparison every 5 minutes; If any deviation exceeds the threshold, an emergency interrupt is immediately triggered, an interrupt signal is sent, and all current running data and deviation status are recorded. Based on all current operational data and deviations, update the real-time datasets (including real-time status data of preheating / cooling / displacement) for the hydrogen production side, grid side, and power supply side, and redetermine the set of grid constraint boundaries; By combining the updated set of power grid constraint boundaries, the regulation capacity is reassessed, and then the optimal solution is re-solved in the dynamic scheduling model. New instructions are generated (focusing on adjusting the time coordination parameters and regulation parameters under deviation conditions) to replace the original instructions. The new instructions include nitrogen replacement, alkaline solution circulation adjustment parameters and energy storage emergency support parameters for deviation. Repeat the steps of data collection, deviation calculation, comparison and adjustment until all feedback data deviations are within the threshold range, forming a closed-loop scheduling.

[0044] Specifically, this step first involves sending various commands, such as hydrogen production system current regulation commands and energy storage safety shutdown protection commands, to the corresponding execution terminals, including the hydrogen production skid rectifier power controller and the energy storage converter controller, via industrial Ethernet according to the execution schedule. The timestamps of these commands are recorded synchronously for traceability. Next, a fixed data acquisition cycle is set. The hydrogen production side monitoring module collects actual current, actual hydrogen production, alkali concentration, and preheating / cooling rate. The grid side monitoring device collects actual voltage, frequency, and total regional load. The power supply side monitoring unit collects the actual state of charge of the energy storage (only collecting actual emergency discharge power when the energy storage triggers an emergency) and the actual output of the two power supplies. The nitrogen replacement monitoring unit collects replacement effect data. All collected data is timestamped to ensure timing matching. Then, multi-dimensional deviation calculations are performed: the absolute value of the current execution deviation is determined based on the actual current and the command target current. The predicted hydrogen production value is calculated by combining the actual hydrogen production with the dynamic function of the current-hydrogen production linear relationship, and the absolute value of the hydrogen production deviation is obtained. Simultaneously, the absolute value of the grid voltage / frequency deviation is extracted. The system calculates deviations in the following parameters: deviation of energy storage state of charge, total power output, nitrogen replacement effect, and preheating / cooling rate. Then, it presets deviation thresholds for each parameter and compares the calculated deviations with their corresponding thresholds one by one. If all deviations are within the thresholds, the current command is maintained, and the deviation detection process is repeated every 5 minutes. If any deviation exceeds the threshold, an emergency interruption is immediately triggered, and an interruption signal is sent, simultaneously recording all current operating data and deviation details. Based on the recorded abnormal data, the real-time datasets (including real-time preheating / cooling / replacement status) on the hydrogen production side, grid side, and power supply side are updated. The grid constraint boundary set is redefined, and the power supply regulation capability is reassessed using the updated constraint boundaries. The new assessment results are input into the dynamic scheduling model to resolve the optimal solution, generating key adjustment time coordination parameters, regulation parameters, and new commands including nitrogen replacement / alkali circulation / energy storage emergency support adjustment parameters, which replace the original commands. Finally, the above data acquisition, deviation calculation, and comparison adjustment process is repeated until all feedback data deviations are within the threshold range, completing the closed-loop scheduling.

[0045] Understandably, issuing commands via industrial Ethernet and recording timestamps ensures the stability of command transmission and the traceability of execution trajectories, preventing scheduling disruptions caused by command delays or loss. Multi-faceted, full-parameter data acquisition covers key indicators throughout the hydrogen production process and the core status of the power grid and power sources, especially the "trigger-based acquisition" of emergency discharge power for energy storage, ensuring data integrity while avoiding invalid acquisition. Multi-dimensional deviation calculation accurately locates the differences between command execution and expected targets, providing a clear direction for subsequent adjustments. The "threshold comparison + emergency interruption" mechanism can quickly intercept out-of-tolerance operating conditions, preventing... To prevent deviations from escalating and causing safety risks such as electrolyzer damage and hydrogen-oxygen mixture explosions, ensuring system operational safety; the process of dynamically updating data and constraint boundaries and resolving the model when deviations exceed limits ensures that new instructions can accurately adapt to current abnormal operating conditions, avoiding insufficient adaptability caused by "one-size-fits-all" scheduling; the continuous cyclical deviation detection and adjustment mechanism forms a complete closed loop, ensuring that the system can always dynamically correct deviations and maintain stable operation under complex scenarios such as intermittent hydrogen production and green electricity fluctuations, significantly improving scheduling reliability and anti-interference capabilities, while extending the service life of core components of hydrogen production equipment.

[0046] Furthermore, a dynamic power dispatching system based on hydrogen production data feature feedback is proposed to implement any of the dispatching methods mentioned above, including: The multi-source data collection module is used to collect historical core data from the hydrogen production side (including data from the entire process of preheating / cooling / nitrogen replacement), historical key data from the power grid side (including dual-path 6kV / 400V grid power data and green electricity fluctuation simulation data), and historical supporting data from the power supply side (including energy storage emergency support data). The data processing and core element construction module is used to construct a dynamic function of the current-hydrogen production linear relationship covering the entire process of preheating-full load-cooling-replacement based on extracted historical data of hydrogen production skid operation, equipment status, characteristic parameters, and related data of nitrogen replacement, alkali circulation, and standby mode. This function is used to analyze and process the collected historical key data from the grid side and historical supporting data from the power source side, and to determine the grid constraint boundary set including energy storage emergency support constraints and full-process power constraints. This function is used to quantitatively analyze the historical supporting data from the power source side, evaluate the power regulation capability including energy storage safety shutdown guarantee, and generate a regulation capability assessment report, providing accurate input elements for the dynamic scheduling model. The dynamic scheduling model solving module is used to establish a dynamic scheduling model with the goals of optimal hydrogen production efficiency, grid stability, minimum losses from frequent start-stop operations, and optimal production time coordination throughout the entire process. It takes as input the dynamic function of the linear relationship between current and hydrogen production, the set of grid constraint boundaries, and relevant data from the regulation capability assessment report. Through model solving, it obtains the optimal power allocation and time coordination scheme that adapts to the entire process of preheating-full load-cooling-displacement. The instruction generation and execution module is used to generate the optimal power allocation and time coordination scheme output by the dynamic scheduling model solution module based on the dynamic scheduling model. It generates current regulation instructions for the hydrogen production system (including step parameters for each stage), energy storage safety shutdown guarantee instructions (including emergency triggering conditions), power output adjustment instructions, nitrogen replacement control instructions, and alkali circulation control instructions. The instructions are then accurately sent to the corresponding execution terminals to drive each system to operate according to the scheduling requirements. The feedback adjustment and closed-loop optimization module is used to collect feedback data in real time during command execution, including hydrogen production, grid parameters, energy storage status, nitrogen replacement effect, alkali circulation status, and preheating / cooling rate. The feedback data is compared with preset thresholds. If the feedback data exceeds the preset thresholds, a re-optimization process is triggered, driving the system to return to the data processing or model solving stage to regenerate the scheduling scheme, thereby achieving closed-loop dynamic scheduling.

[0047] Furthermore, the data processing and core element construction module includes: The dynamic function construction unit is used to construct a dynamic function of current-hydrogen production linear relationship covering the entire process of preheating-full load-cooling-displacement based on the hydrogen production skid operation history data (including preheating / cooling / displacement data), equipment status history data, and characteristic parameter history data extracted by the multi-source data collection module, combined with preheating / cooling temperature correction factor and frequent start-stop correction factor. The constraint boundary determination unit receives the current-hydrogen production linear relationship dynamic function output by the dynamic function construction unit, and combines it with the historical key data of the grid side and the historical supporting data of the power source side (including energy storage emergency data) collected by the multi-source data collection module. Through data statistics and threshold analysis, it determines the set of grid constraint boundaries, which includes energy storage emergency guarantee constraints and full-process power constraints. The regulation capability assessment unit is used to synchronously receive the current-hydrogen production linear relationship dynamic function from the dynamic function construction unit and the grid constraint boundary set from the constraint boundary determination unit. It performs quantitative analysis on the historical supporting data of the power supply side of the multi-source data collection module, assesses the power supply regulation capability including energy storage safety shutdown protection, and generates a regulation capability assessment report.

[0048] Furthermore, the dynamic scheduling model solution module includes: The model building unit is used to clarify the four optimization directions of optimal hydrogen production efficiency, grid stability, minimum loss from frequent start-stop, and optimal production time coordination throughout the process. It incorporates the control logic of soft start-stop program (preheating load increase, cooling load decrease), standby mode, nitrogen replacement, and alkaline solution circulation, sets the objective function and constraints, and establishes a dynamic scheduling model adapted to the coordinated scheduling of hydrogen production, grid and power supply. The model solving unit receives the current-hydrogen production linear relationship dynamic function, grid constraint boundary set, and regulation capability assessment report output by the data processing and core element construction module. It then uses these as core input parameters to input into the dynamic scheduling model and solves for the optimal power allocation and time coordination scheme in combination with the requirements of the entire process of preheating-full load-cooling-displacement.

[0049] The advantages of this invention are: it deeply adapts to the entire process of intermittent hydrogen production, from "preheating to full load to cooling to replacement," precisely addressing the core pain points of existing scheduling methods. It focuses on collecting data from non-production periods on the hydrogen production side, fluctuation data from dual-path grid power and green electricity on the grid side, and emergency data from energy storage on the power supply side, constructing a dynamic function of the current-hydrogen production linear relationship covering the entire process. This breaks through the limitation of traditional methods that only focus on full load conditions, providing precise scenario-based input for scheduling. Simultaneously, it clarifies the positioning of energy storage as an "emergency safety shutdown guarantee," correcting the deviation of existing solutions that mistakenly treat it as a daily regulation unit. Both grid constraints and power supply regulation capability assessments are designed around the emergency function of energy storage, aligning with the actual production needs of energy storage to be "safe and not frequently used." Furthermore, a dynamic scheduling model with four objectives—hydrogen production efficiency, grid stability, start-up and shutdown losses, and time coordination—was established. The optimal solution was obtained by verifying the solution with a genetic algorithm and multiple datasets, avoiding the one-sidedness of optimizing a single objective. A closed-loop scheduling mechanism was also constructed for the entire process, collecting feedback parameters in real time during non-production periods. When deviations exceed the limits, rapid interruption and optimization are performed to ensure the system's anti-interference capability, effectively alleviate equipment damage, extend its lifespan, and significantly improve the reliability and feasibility of the off-grid intermittent hydrogen production system.

[0050] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A dynamic power dispatching method based on hydrogen production data feature feedback, characterized in that, include: Historical core data on the hydrogen production side and historical supporting data on the power supply side of the alkaline water electrolysis hydrogen production system were collected. At the same time, historical key data on the grid side in off-grid scenarios were collected. Historical data on hydrogen production skid operation, equipment status, characteristic parameters, nitrogen replacement status, alkaline solution circulation status, and standby mode operation were extracted. A dynamic function of the linear relationship between current and hydrogen production was constructed. The historical core data on the hydrogen production side includes daily full-load operation data, preheating period data, cooling period data, and nitrogen replacement data corresponding to long-term shutdown. By combining historical core data from the hydrogen production side, historical supporting data from the power supply side, and historical key data from the grid side, a set of grid constraint boundaries for off-grid-energy storage collaboration is determined. This set of constraint boundaries includes emergency support constraints for energy storage systems. By combining the dynamic function of the current-hydrogen production linear relationship and the set of grid constraint boundaries, the power regulation capability of the grid-type voltage-supported energy storage backup is evaluated, and a regulation capability assessment report is generated. A dynamic scheduling model is established with the goals of optimal hydrogen production efficiency, off-grid grid stability, minimum losses from frequent start-stop operations, and optimal production time coordination throughout the entire process. The dynamic function of the linear relationship between current and hydrogen production, the set of grid constraint boundaries, and the regulation capacity assessment report are used as inputs to solve the optimal power allocation and time coordination scheme that adapts to the entire process of preheating-full load-cooling-displacement. Based on the optimal power allocation and time coordination scheme, the following commands are generated: current regulation command for hydrogen production system, energy storage safety shutdown guarantee command, power output adjustment command, nitrogen replacement control command, and alkali circulation control command. The system executes commands for current regulation of the hydrogen production system, energy storage safety shutdown, power output adjustment, nitrogen replacement control, and alkali circulation control. It also collects feedback data in real time on hydrogen production, grid parameters, energy storage status, alkali concentration / circulation status, nitrogen replacement effect, and preheating / cooling rate. If the feedback data exceeds a preset threshold, the scheduling scheme is re-optimized. A new scheduling scheme is generated by combining the safety control logic of nitrogen replacement, alkali circulation, and soft start-stop procedures to achieve closed-loop scheduling.

2. The power dynamic dispatching method based on hydrogen production data feature feedback according to claim 1, characterized in that, The construction of the dynamic function for the linear relationship between current and hydrogen production specifically includes: Confirm the design life of the hydrogen production system and obtain the core historical data of the hydrogen production side over the past 3 years. The core historical data of the hydrogen production side includes the 1-hour nitrogen replacement operating condition data corresponding to 6 hours of full-load operation, 1 hour of preheating, 3 hours of cooling, and long-term shutdown. Data corresponding to periods of fault shutdown, periods of abnormal fluctuation, and periods of non-experimental shutdown were removed from the historical core data of hydrogen production over the past three years to obtain effective historical core data of hydrogen production, and the duration of effective data was determined simultaneously. Extract historical operation data of hydrogen production skid from effective historical core data of hydrogen production side. The historical operation data of hydrogen production skid includes output current and output voltage records, alkaline solution circulation status records, nitrogen replacement execution records, and temperature change records during the preheating / cooling stage once every 1 minute. According to the operating conditions, the average hourly current and the corresponding actual hydrogen production were obtained. 1000 sets of effective historical core data on the hydrogen production side were selected, which were evenly distributed and covered the preheating stage, 6 hours of full load, cooling stage and nitrogen replacement stage, to determine the initial proportional coefficient. Using historical equipment status data, a comprehensive correction factor is constructed, which includes a temperature correction factor, a pressure correction factor, and a frequent start-stop correction factor. Determine the characteristic correction factor by combining historical data of characteristic parameters; Based on the initial proportional coefficient, comprehensive correction factor and characteristic correction factor, the dynamic proportional coefficient is determined, and the basic form of the current-hydrogen production linear relationship is constructed. From the effective historical core data of hydrogen production, 500 sets of test data that were not involved in the construction of the basic form of the current-hydrogen production linear relationship were selected. The current in the test data was substituted into the basic form of the current-hydrogen production linear relationship to obtain the predicted value of hydrogen production, and the deviation rate of each set was determined. Using the deviation rate of each group, determine the dynamic function of the current-hydrogen production linear relationship that meets the requirements.

3. The method for dynamic power dispatching based on hydrogen production data feature feedback according to claim 2, characterized in that, The determination of the set of power grid constraint boundaries specifically includes: Acquire historical key data from the grid side, historical core data from the hydrogen production side, and historical supporting data from the power supply side over the past three years. The historical key data from the grid side includes voltage / frequency data of dual 6kV / 400V grid power and voltage / frequency data simulating green electricity fluctuations. The historical supporting data from the power supply side includes energy storage system status data and auxiliary power supply data during sudden power outages of energy storage systems. The historical core data from the hydrogen production side includes off-grid auxiliary system load data, historical power data of the hydrogen production system, load response data of the hydrogen production system, nitrogen replacement data, alkali circulation data, and standby mode operation data. The historical key data from the grid side, the historical core data from the hydrogen production side, and the historical supporting data from the power source side over the past three years were unified to a sampling frequency of 1 minute / time and sorted in chronological order to ensure that the timestamps of each data are completely matched. After the timestamps are fully matched, the physical quantities in the historical key data of the grid side, the historical core data of hydrogen production side, and the historical supporting data of the power source side are standardized and the abnormal data are cleaned to obtain the historical key standard data of the grid side, the historical core standard data of hydrogen production side, and the historical supporting standard data of the power source side. Historical data of dual-path 6kV / 400V grid voltage and simulated green electricity fluctuation voltage data were extracted from the historical key standard data of the grid side, and clustered according to the operating condition type to obtain steady-state grid voltage cluster and green electricity fluctuation cluster. Obtain the boundary threshold deviation coefficients of steady-state grid power clusters and green power fluctuation clusters, thereby constructing voltage constraint boundaries; Extract historical frequency data of dual-path 6kV / 400V grid power and simulated green electricity fluctuation frequency data from historical key standard data of the power grid side; A Gaussian mixture model was constructed based on historical frequency data of dual-path 6kV / 400V grid power and simulated green electricity fluctuation frequency data, and decomposed into steady-state components, wave dynamic components and near-limiting state components. Historical frequency data of dual-path 6kV / 400V grid power, simulated green electricity fluctuation frequency data and the Mahalanobis distance of steady-state components are obtained to determine the frequency constraint boundary; Extract energy storage system status data and emergency discharge capacity data from historical supporting standard data on the power supply side, perform energy storage emergency protection efficiency calibration, and simultaneously determine the energy storage emergency protection state of charge range and emergency discharge power limit to obtain energy storage constraint boundaries. Extract the off-grid auxiliary system load data and the historical power data of the hydrogen production system from the historical core standard data of the hydrogen production side, perform auxiliary load statistics, and obtain the average load and peak load of the preheating / cooling / displacement stage and the full-load operation stage; Based on the average and peak loads of each stage, combined with the daily routine operating conditions and the nitrogen replacement operating conditions with monthly shutdowns, the hydrogen production power ratio corresponding to each stage is obtained, thereby determining the power ratio constraint. Voltage / frequency fluctuation data for simulating green electricity are extracted from historical key standard data on the power grid side, and load response data of the hydrogen production system is extracted from historical core standard data on the hydrogen production side. Based on the voltage / frequency fluctuation data of simulated green electricity and the load response data of hydrogen production system, we classify green electricity fluctuations and impose response time constraints, and construct the coupling relationship between green electricity fluctuation amplitude and hydrogen production regulation. By summarizing voltage constraint boundaries, frequency constraint boundaries, energy storage emergency support constraints, and power ratio constraints, and combining them with the coupling relationship between green electricity fluctuation amplitude and hydrogen production regulation, a set of grid constraint boundaries is formed.

4. The power dynamic dispatching method based on hydrogen production data feature feedback according to claim 3, characterized in that, The dynamic function of the current-hydrogen production linear relationship and the set of grid constraint boundaries are used to evaluate the power regulation capability of grid-type voltage-supported energy storage backup, and a regulation capability assessment report is generated, which specifically includes: Based on the dual-path 6kV / 400V grid power operation log and energy storage BMS system, historical supporting data of the power supply side over the past 3 years were collected, including historical power output data of the two power sources, historical state of charge data of energy storage, emergency discharge power data of energy storage, power output of new energy sources, and power supply reliability records. The average output of each power source was determined based on historical power output data from both sources. Obtain the power output fluctuation range of each power source and determine the power output adjustment range; Based on historical state of charge data of energy storage and emergency support needs, the proportion of time interval of energy storage emergency support state of charge interval is statistically analyzed to determine the effective range of energy storage emergency discharge. Extract the historical maximum emergency discharge power from the energy storage emergency discharge power data, and set the energy storage emergency discharge power limit in combination with emergency support needs; Based on the dynamic function of the linear relationship between current and hydrogen production and the set of grid constraint boundaries, combined with the full-process operating conditions of the hydrogen production system from preheating to full load, cooling to replacement and the frequent start-stop characteristics, the power regulation range of the hydrogen production system at each stage is determined. Determine the basic power compensation capability on the power supply side and construct an evaluation index system for regulation capability; Based on the regulation capacity assessment index system, a regulation capacity assessment report is generated. The regulation capacity assessment report is updated every 24 hours based on the latest power supply side data, clarifying the power supply side regulation capacity and energy storage emergency support status under different intermittent operating conditions and start-stop frequencies.

5. The power dynamic dispatching method based on hydrogen production data feature feedback according to claim 4, characterized in that, The establishment of a dynamic scheduling model aims to achieve optimal hydrogen production efficiency, off-grid grid stability, minimize frequent start-stop losses, and optimal production time coordination throughout the entire process. This model uses the current-hydrogen production linear relationship dynamic function, the grid constraint boundary set, and the regulation capacity assessment report as inputs to solve for the optimal power allocation and time coordination scheme adapted to the entire process of preheating-full load-cooling-displacement. Specifically, this includes: Based on the dynamic function of the linear relationship between current and hydrogen production, the correlation between hydrogen production and current is transformed into model constraints. Combined with the user's daily target of 6 hours of full-load hydrogen production, it is transformed into hydrogen production power constraints for the entire process of preheating-full-load-cooling. Transform the constraint boundaries in the set of power grid constraint boundaries into model inequality constraints; Based on the regulation capability assessment report, the power output regulation range, energy storage emergency support constraints, and power regulation range of each stage of the hydrogen production system are transformed into model variable constraints. Historical core data from the hydrogen production side, historical key data from the power grid side, and historical supporting data from the power supply side are divided into training set data, validation set data, and test set data. A four-objective optimization function and a genetic algorithm solution model are set. The optimization function incorporates soft start-stop program, standby mode, and nitrogen replacement control logic. A preliminary power allocation and time coordination scheme was obtained; The initial power allocation and time coordination scheme is input into the validation set to determine the hydrogen production efficiency, grid stability index, frequent start-stop loss coefficient and overall time coordination efficiency of the validation set. The schemes that meet the verification requirements are input into the test set, and based on the dynamic function of the current-hydrogen production linear relationship, the hydrogen production power target in the tested schemes is transformed into the hydrogen production current target at each stage, and the optimal power allocation and time coordination scheme is determined.

6. The method for dynamic power dispatching based on hydrogen production data feature feedback according to claim 5, characterized in that, The generation of hydrogen production system current regulation commands, energy storage safety shutdown protection commands, power output adjustment commands, nitrogen replacement control commands, and alkali circulation control commands based on the optimal power allocation and time coordination scheme specifically includes: Based on the hydrogen production current targets at each stage in the optimal power allocation and time coordination scheme, and combined with the soft start-stop procedure requirements, the current real-time current on the hydrogen production side is obtained, and the current adjustment difference for preheating load increase, full load stabilization, and cooling load reduction is determined. Based on the maximum allowable adjustment rate of the hydrogen production system, and combined with the soft start-stop requirements of slow preheating and slow cooling, the current adjustment time for each stage is determined. By combining the hydrogen production current target and the maximum allowable adjustment rate of the hydrogen production system, a current adjustment command for the hydrogen production system is generated, which includes step-by-step adjustment parameters for each stage. Based on the energy storage emergency protection requirements in the optimal power allocation and time coordination scheme, the current state of charge of the energy storage is obtained, and an energy storage safety shutdown protection command is generated, which includes emergency triggering conditions, emergency discharge power, state of charge protection threshold, and emergency shutdown timing logic. Based on the output targets of the two power sources in the optimal power allocation and time coordination scheme, and combined with the test requirements for simulating green electricity fluctuations in the later stage, the current real-time output of the power sources is obtained, and the adjustment difference between the output of the first and second power sources is determined. Obtain historical power supply reliability coefficients to determine adjustment priorities; Based on the adjustment priority and output adjustment difference, a power output adjustment command is generated, which includes power supply number, adjustment direction, adjustment power, adjustment step size, adjustment interval, output upper limit protection and green electricity fluctuation simulation adaptation parameters. Based on the nitrogen replacement timing requirements in the optimal power allocation and time coordination scheme, a nitrogen replacement control command is generated. Based on the requirements for heat preservation of alkali solution circulation during the cooling stage, alkali solution circulation control commands are generated. Summarize the current regulation commands of the hydrogen production system, the energy storage safety shutdown protection commands, the power output adjustment commands, the nitrogen replacement control commands, and the alkaline solution circulation control commands, determine the command check codes, and add them to the end of the commands; Generate an execution schedule, clarify the order and connection of instruction execution, including the coordinated execution logic of soft start / stop, nitrogen replacement, and alkali circulation.

7. The power dynamic dispatching method based on hydrogen production data feature feedback according to claim 6, characterized in that, The system executes commands for current regulation of the hydrogen production system, energy storage safety shutdown, power output adjustment, nitrogen replacement control, and alkali circulation control. It also collects real-time feedback data on hydrogen production, grid parameters, energy storage status, alkali concentration / circulation status, nitrogen replacement effect, and preheating / cooling rate. If the feedback data exceeds a preset threshold, the scheduling scheme is re-optimized. A new scheduling scheme is generated by combining the safety control logic of nitrogen replacement, alkali circulation, and soft start / stop procedures to achieve closed-loop scheduling. Specifically, this includes: Based on the execution schedule, various instructions are sent to the hydrogen production skid rectifier power controller, energy storage converter controller, power control system, nitrogen replacement control system, and alkali circulation control system via industrial Ethernet, and the instruction sending timestamp is recorded. Set the data acquisition cycle, collect actual current, actual hydrogen production, alkali concentration and circulation status, and preheating / cooling rate through the hydrogen production side monitoring module, collect actual voltage, actual frequency, and total regional load through the grid side monitoring device, collect actual energy storage state of charge, actual emergency discharge power, and actual output of the two power sources through the power supply side monitoring unit, and collect replacement effect data through the nitrogen replacement monitoring unit, and record the acquisition timestamp; Based on the actual current collected and the target current commanded, the absolute value of the current execution deviation is determined; Based on the actual hydrogen production and the dynamic function of the linear relationship between current and hydrogen production, the predicted value of hydrogen production is determined, and the absolute value of the deviation of hydrogen production is obtained simultaneously. Obtain the absolute values ​​of grid voltage deviation, frequency deviation, energy storage state of charge deviation, total power output deviation, nitrogen replacement effect deviation, and preheating / cooling rate deviation. Set a preset deviation threshold and compare each deviation with the preset deviation threshold one by one; If all deviations are within the preset deviation threshold, maintain the current instruction and repeat the deviation acquisition and comparison every 5 minutes; If any deviation exceeds the threshold, an emergency interrupt is immediately triggered, an interrupt signal is sent, and all current running data and deviation status are recorded. Based on all current operational data and deviations, update the real-time datasets for hydrogen production, grid, and power generation, and redetermine the set of grid constraint boundaries. By combining the updated set of power grid constraint boundaries, the regulation capacity is reassessed, and then the optimal solution is re-solved in the dynamic scheduling model. New instructions are generated to replace the original instructions. The new instructions include nitrogen replacement, alkaline solution circulation adjustment parameters and energy storage emergency support parameters for deviations. Repeat the steps of data collection, deviation calculation, comparison and adjustment until all feedback data deviations are within the threshold range, forming a closed-loop scheduling.

8. A dynamic power dispatching system based on hydrogen production data feature feedback, used to implement the dispatching method as described in any one of claims 1-7, characterized in that, include: A multi-source data collection module is used to collect historical core data from the hydrogen production side, historical key data from the power grid side, and historical supporting data from the power source side. The data processing and core element construction module is used to construct a dynamic function of the current-hydrogen production linear relationship covering the entire process of preheating-full load-cooling-replacement based on extracted historical operating data, equipment status data, characteristic parameter data, and relevant data on nitrogen replacement, alkaline circulation, and standby mode. This function is then used to analyze and process collected historical key data from the grid side and historical supporting data from the power source side, determining a set of grid constraint boundaries including energy storage emergency support constraints and full-process power constraints. Finally, this module combines the current-hydrogen production linear relationship dynamic function and the grid constraint boundary set to quantitatively analyze historical supporting data from the power source side, assessing the power regulation capability including energy storage safety shutdown protection and generating a regulation capability assessment report, thus providing accurate input elements for the dynamic scheduling model. The dynamic scheduling model solving module is used to establish a dynamic scheduling model with the goals of optimal hydrogen production efficiency, grid stability, minimum frequent start-stop losses, and optimal production time coordination throughout the entire process. The inputs include the dynamic function of the linear relationship between current and hydrogen production, the set of grid constraint boundaries, and relevant data from the regulation capability assessment report. The optimal power allocation and time coordination scheme that adapts to the entire process of preheating-full load-cooling-replacement is obtained through model solving. The instruction generation and execution module is used to generate hydrogen production system current regulation instructions, energy storage safety shutdown guarantee instructions, power output adjustment instructions, nitrogen replacement control instructions and alkali circulation control instructions based on the optimal power allocation and time coordination scheme output by the dynamic scheduling model solution module, and accurately send the instructions to the corresponding execution terminals to drive each system to operate according to the scheduling requirements. The feedback adjustment and closed-loop optimization module is used to collect feedback data in real time during instruction execution, including hydrogen production, grid parameters, energy storage status, nitrogen replacement effect, alkali circulation status, and preheating / cooling rate. The feedback data is compared with preset thresholds. If the feedback data exceeds the preset thresholds, a re-optimization process is triggered, driving the system to return to the data processing or model solving stage to regenerate the scheduling scheme, thereby achieving closed-loop dynamic scheduling.

9. A dynamic power dispatching system based on hydrogen production data feature feedback according to claim 8, characterized in that, The data processing and core element construction module includes: The dynamic function construction unit is used to perform linear relationship fitting and correction based on the historical operation data, equipment status data, and characteristic parameter data of the hydrogen production skid extracted by the multi-source data collection module, combined with the preheating / cooling temperature correction factor and the frequent start-stop correction factor, to construct a dynamic function of current-hydrogen production linear relationship covering the entire process of preheating-full load-cooling-displacement. The constraint boundary determination unit is used to receive the current-hydrogen production linear relationship dynamic function output by the dynamic function construction unit, and combine it with the historical key data of the grid side and the historical supporting data of the power source side collected by the multi-source data collection module. Through data statistics and threshold analysis, the set of grid constraint boundaries including energy storage emergency guarantee constraints and full-process power constraints is determined. The regulation capability assessment unit is used to synchronously receive the current-hydrogen production linear relationship dynamic function from the dynamic function construction unit and the grid constraint boundary set from the constraint boundary determination unit, perform quantitative analysis on the historical supporting data of the power supply side of the multi-source data collection module, assess the power supply regulation capability including energy storage safety shutdown protection, and generate a regulation capability assessment report.

10. A dynamic power dispatching system based on hydrogen production data feature feedback according to claim 8, characterized in that, The dynamic scheduling model solving module includes: The model building unit is used to clarify the four optimization directions of optimal hydrogen production efficiency, grid stability, minimum loss from frequent start-stop, and optimal production time coordination throughout the process. It incorporates control logic of soft start-stop program, standby mode, nitrogen replacement, and alkaline solution circulation, sets objective function and constraints, and establishes a dynamic scheduling model adapted to the coordinated scheduling of hydrogen production, grid and power supply. The model solving unit receives the current-hydrogen production linear relationship dynamic function, grid constraint boundary set, and regulation capability assessment report output by the data processing and core element construction module, and uses them as core input parameters to input into the dynamic scheduling model. It then solves for the optimal power allocation and time coordination scheme by combining the requirements of the entire process of preheating-full load-cooling-displacement.

Images