Micro-grid hydrogen production system based on dynamic optimization control and multi-mode energy scheduling method

Abstract

The invention relates to the technical field of hydrogen production, in particular to a micro-grid hydrogen production system based on dynamic optimization control and a multi-mode energy scheduling method. Comprising a renewable energy power generation unit which comprises at least two distributed power supplies with different power generation characteristics; the multi-stage hydrogen production device is configured to receive the electric energy output of the renewable energy power generation unit and perform proton exchange membrane water electrolysis hydrogen production; the hybrid energy storage system comprises a super capacitor bank and a lithium ion battery pack which are connected in parallel to the direct current bus through a bidirectional DC / DC converter; the dynamic optimization controller is internally provided with a multi-time-scale coordination algorithm and is used for acquiring meteorological prediction data, load demand data and an energy storage SOC state in real time; the multi-mode scheduling module comprises three working modes including an island operation mode, a grid-connected operation mode and an emergency standby mode; according to the scheme, hydrogen production economy, equipment durability and power supply reliability can be considered at the same time.

Description

Technical Field

[0001] The present invention relates to the technical field of hydrogen production, and particularly to a microgrid hydrogen production system and a multi-mode energy scheduling method based on dynamic optimization control. Background Art

[0002] With the continuous increase in the penetration rate of renewable energy, the microgrid-coupled hydrogen production system has become an important technical path for the efficient utilization of clean energy. In the prior art, the hydrogen production system based on wind-solar power generation usually adopts an architecture combining a single energy storage device and a fixed control strategy, and has significant defects in coping with power fluctuations on multiple time scales. Specifically, when the output of renewable energy fluctuates from minute level to hour level, traditional control methods are difficult to coordinate the dynamic balance among the hydrogen production device, the energy storage system and the interaction with the power grid, resulting in three key problems: First, the hydrogen production equipment is forced to deviate from the optimal efficiency working area due to frequent power fluctuations, resulting in a decrease in energy conversion efficiency; Second, the energy storage system generates deep unbalanced charge and discharge during the process of suppressing power fluctuations, accelerating battery aging; Finally, excessive reliance on grid power support leads to an increase in operating costs.

[0003] Particularly noteworthy is that although the prior art can achieve basic power balance, the static optimization model it adopts cannot adapt to the time-varying characteristics of both the source and the load. When dealing with the scenario of sudden changes in the output of renewable energy within a day, there are still technical bottlenecks such as the hydrogen production efficiency loss exceeding 15% and the SOC dispersion of the energy storage being higher than 20%. In addition, although the multi-mode scheduling scheme proposed by foreign related technologies introduces the function of island / grid connection switching, it does not solve the problem of power impact during the mode conversion process, resulting in insufficient power supply reliability for critical loads. The essence of the above defects lies in the lack of a dynamic optimization mechanism across time scales.

[0004] Based on the above problems, there is an urgent need for a cooperative control method that can simultaneously take into account the hydrogen production economy, equipment durability and power supply reliability, so as to break through the limitations of the prior art in the dimension of multi-objective optimization. Summary of the Invention

[0005] The purpose of the present invention is to solve the disadvantages existing in the prior art, and to propose a microgrid hydrogen production system based on dynamic optimization control, including: A renewable energy power generation unit, including at least two distributed power sources with different power generation characteristics; A multi-stage hydrogen production device, configured to receive the electric energy output of the renewable energy power generation unit and perform proton exchange membrane electrolytic water hydrogen production; A hybrid energy storage system, including a supercapacitor bank and a lithium-ion battery bank, and the two are connected in parallel to the DC bus through a bidirectional DC / DC converter; A dynamic optimization controller, with a multi-time scale coordination algorithm built in, and real-time obtaining meteorological prediction data, load demand data and the SOC state of the energy storage; A multi-mode scheduling module, including three working modes: island operation mode, grid-connected operation mode, and emergency standby mode; A hydrogen energy management unit, configured to monitor the pressure, hydrogen purity, and temperature parameters of the hydrogen storage tank, and establish two-way communication with the dynamic optimization controller; A power electronic interface device, including an inverter with virtual synchronous machine characteristics and a rectifier with fast power tracking ability; A safety protection circuit, integrating overvoltage protection, reverse current blocking, and hydrogen leakage monitoring functions; The dynamic optimization controller generates hydrogen production power instructions, energy storage charge and discharge strategies, and operation mode switching decisions through a multi-objective optimization algorithm, where the optimization objectives include maximizing hydrogen production efficiency, equalizing equipment life, and minimizing operation costs.

[0006] Adopting the above technical solution: This solution constructs a complete system architecture including renewable energy power generation, multi-stage hydrogen production devices, hybrid energy storage, a dynamic optimization controller, and multi-mode scheduling. Through a multi-objective optimization algorithm, it realizes the coordinated control of hydrogen production power, energy storage charge and discharge, and operation mode; it can solve three main defects existing in the existing microgrid hydrogen production system: a single energy storage system is difficult to cope with the power fluctuation characteristics of renewable energy; traditional control methods cannot coordinate the contradiction between hydrogen production efficiency and equipment life, and power supply interruption is likely to occur during operation mode switching.

[0007] The dynamic optimization controller includes: A prediction module, which processes the time series characteristics of meteorological data using an improved deep belief network to generate a renewable energy output prediction curve for the next 24 hours; A rolling optimization module, which constructs a hierarchical optimization model based on the model predictive control framework. The upper layer conducts hourly energy scheduling planning, and the lower layer executes minute-level power distribution; A feedback correction module, configured to dynamically adjust the weight coefficients of the prediction model according to the real-time power deviation; A constraint processing module, integrating a three-dimensional constraint space of equipment operation constraints, safety boundary constraints, and hydrogen energy supply and demand balance constraints; The rolling optimization module adopts a dynamic weight allocation strategy to automatically enhance the response priority of the hydrogen production system when the renewable energy output fluctuation exceeds the threshold.

[0008] Adopting the above technical solution: This solution adopts a hierarchical optimization architecture of an improved deep belief network + model predictive control, combined with a dynamic weight allocation and three-dimensional constraint processing mechanism. It can solve problems existing in traditional controllers such as low prediction accuracy, fixed optimization time domain, and rigid constraint processing.

[0009] Preferably, the multi-stage hydrogen production device includes: The main electrolyzer group, with a rated power of 70% - 80% of the total system capacity, is configured for continuous operation mode; The auxiliary electrolyzer group, with a rated power of 20% - 30% of the total system capacity, is configured for pulse operation mode; The buffer electrolyzer, with a capacity accounting for 5% - 10% of the total electrolyzer capacity, is configured for power fluctuation absorption mode; A power interleaving control strategy is adopted between the main electrolyzer group and the auxiliary electrolyzer group, and the start - stop time difference between them is controlled within the range of 0.5 - 2 seconds; The dynamic response time of the buffer electrolyzer is less than 100 ms, and the power regulation accuracy reaches ±1% of the rated power.

[0010] Adopting the above - mentioned technical solution: By establishing a three - level structure of main / auxiliary / buffer electrolyzers and adopting power interleaving control and dynamic response strategies, this solution can solve the technical problems that large - capacity electrolyzers in traditional hydrogen production devices are difficult to quickly adjust and the efficiency decreases due to frequent start - stop.

[0011] Preferably, the control strategy of the hybrid energy storage system includes: The supercapacitor bank preferentially responds to high - frequency power fluctuation components, and the charge - discharge depth is controlled within the range of 20% - 80% SOC; The lithium - ion battery bank processes low - frequency power fluctuation components, and the charge - discharge rate is limited within the range of 0.5C - 1C; The dynamic power distribution coefficient is adjusted in real - time according to the current SOC state and the predicted load curve; A power complementary mechanism is established between the energy storage system and the hydrogen production device to provide instantaneous power support during the start - up stage of the electrolyzer.

[0012] Adopting the above - mentioned technical solution: By establishing a cooperative control strategy for supercapacitors and lithium - batteries, and establishing a power complementary mechanism and a dynamic distribution coefficient, this solution can solve the problem of coupling interference between energy - type / power - type energy storage in the existing hybrid energy storage system, resulting in the problems of frequent shallow charge - discharge of lithium - batteries and low utilization rate of supercapacitors.

[0013] Preferably, the switching logic of the multi - mode scheduling module includes: In the island operation mode, the hydrogen production system serves as the main power source to supply power to local loads, and at the same time maintains at least 10% of the spinning reserve capacity; In the grid - connected operation mode, renewable energy power generation is preferentially absorbed, and the remaining power is used for hydrogen production and energy storage charging; In the emergency standby mode, the hydrogen production device can be switched to the fuel cell mode within 30 seconds to provide emergency power supply to the grid; The mode switching process adopts a three - step confirmation mechanism, including the status pre - judgment, safety verification, and execution confirmation stages; The conversion time between different modes does not exceed 2 minutes, and critical loads are continuously powered during the switching process.

[0014] Adopting the above technical solution: This solution ensures the power supply continuity during the switching process by establishing a three-step confirmation mechanism and a multi-mode collaborative control strategy. It can solve the two major risks existing in traditional mode switching: power supply interruption for more than 500 ms during grid connection / disconnection switching and equipment damage caused by incorrect mode switching.

[0015] Preferably, the multi-objective optimization algorithm adopts an improved NSGA-II algorithm, and the following optimization model is established: ; Constraint conditions: ; Among them, is the power exchanged with the main grid at time is the rated power of the system, is the th state of charge of the energy storage unit at time is the nominal operating point of the energy storage system, is the hydrogen production cost at time is the standard hydrogen production equivalent, is the real-time output of renewable energy, is the charge and discharge power of the energy storage system, is the load demand power, is the power consumed by the hydrogen production device.

[0016] Adopting the above technical solution: This solution establishes a multi-objective optimization model through an improved NSGA-II algorithm, including three optimization objectives: grid interaction, energy storage balance, and hydrogen production cost. It can solve the problems caused by traditional single-objective optimization: over-reliance on grid interaction, discretization of energy storage SOC, and conflicts between hydrogen production efficiency and cost.

[0017] Preferably, the mathematical formula for optimizing hydrogen production efficiency using a dynamic electrolysis model is: ; In the formula, is the comprehensive hydrogen production efficiency at time is the electrochemistry efficiency of the electrolyzer, is the thermal management efficiency, is the real-time hydrogen production power, is the power at the optimal economic operation point.

[0018] Adopt the above technical solution: By introducing a dynamic model of electrolysis efficiency with a secondary correction term, this solution reflects the impact of power offset on the comprehensive efficiency. It can solve the problems of the traditional efficiency model with static assumption defects, resulting in too high efficiency estimation error when deviating from the optimal operating point and inability to guide real-time power distribution.

[0019] Preferably, the mathematical formula of the energy storage life model is: ; where is the battery remaining life index, is the standard number of cycles, is the depth of discharge at time is the reference depth of discharge, is the real-time battery temperature, is the reference temperature, is the aging coefficient.

[0020] Adopt the above technical solution: By loading an exponential life decay model with multiple factors such as DOD and temperature, this solution can solve the problem that the traditional life model only considers the number of cycles and cannot reflect the influence of actual working conditions.

[0021] Preferably, the safety protection circuit includes: A three-level overvoltage protection mechanism with action thresholds of 1.1 pu, 1.15 pu, and 1.2 pu respectively; The reverse current detection accuracy reaches 0.5% FS, and the response time is less than 10 ms; Hydrogen concentration sensors are arranged at 5 monitoring points around the electrolyzer, and a redundant verification mechanism is adopted; The safety interlock device can cut off the fault circuit and activate the pressure relief valve within 50 ms.

[0022] Adopt the above technical solution: By establishing a safety protection system of three-level overvoltage protection + redundant verification + fast interlock, this solution can solve the defects of the traditional protection scheme such as slow response speed and high misoperation rate.

[0023] A multi-mode energy scheduling method is applied to the microgrid hydrogen production system based on dynamic optimization control described in any one of the above, and includes the following steps: S1: Dynamic determination of the operating mode: Real-time collect the grid connection status, the state of charge of the energy storage system, and the load demand volatility, and determine the current operating mode as the island mode, grid-connected mode, or emergency standby mode based on the preset threshold logic; S2: Configuration of energy scheduling priority: Dynamically adjust the power distribution priorities of renewable energy consumption, charging and discharging of the energy storage system, and grid interaction according to the operating mode. In the island mode, give priority to using the energy storage system to maintain the continuous operation of the hydrogen production device; S3: Hierarchical Time Scale Optimization: Based on hourly energy prediction and minute-level power fluctuation characteristics, formulate energy storage capacity scheduling plans and dynamic hydrogen production power distribution strategies respectively. Long-term optimization focuses on the balance of equipment life, while short-term optimization focuses on suppressing power fluctuations. S4: Dynamic Constraint Boundary Adjustment: According to the real-time monitored renewable energy output fluctuation range, energy storage health status, and grid electricity price signals, rollingly correct the power distribution coefficient and the equipment operation safety boundary. S5: Multi-level Safety Verification Execution: Before issuing the scheduling instruction, verify the overcharge and over-discharge protection thresholds of the energy storage system, the temperature and pressure limits of the hydrogen production device, and the reverse current suppression ability of the grid in sequence. After confirmation, execute the scheduling operation. Description of the Drawings

[0024] Figure 1 It is a block diagram of a microgrid hydrogen production system based on dynamic optimization control of the present application; Figure 2 It is a flowchart of the present application and the multi-mode energy scheduling method. Detailed Embodiment

[0025] In order to make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0026] Please refer to Figure 1 - Figure 2 , for example, there are three main defects in the existing microgrid hydrogen production system: a single energy storage system is difficult to cope with the power fluctuation characteristics of renewable energy; traditional control methods cannot coordinate the contradiction between hydrogen production efficiency and equipment life; power supply interruption is likely to occur during operation mode switching. Based on the above problems, the present application provides a microgrid hydrogen production system based on dynamic optimization control, including: Renewable energy power generation unit, including at least two distributed power sources with different power generation characteristics; Multi-level hydrogen production device, configured to receive the electric energy output of the renewable energy power generation unit and perform proton exchange membrane electrolytic water hydrogen production; Hybrid energy storage system, including a supercapacitor bank and a lithium-ion battery bank, which are connected in parallel to the DC bus through a bidirectional DC / DC converter; Dynamic optimization controller, built-in with a multi-time scale coordination algorithm, and real-time obtains meteorological prediction data, load demand data, and energy storage SOC status; Multi-mode scheduling module, including three working modes: island operation mode, grid-connected operation mode, and emergency standby mode; The hydrogen energy management unit is configured to monitor the pressure, hydrogen purity, and temperature parameters of the hydrogen storage tank and establish two-way communication with the dynamic optimization controller; The power electronic interface device includes an inverter with virtual synchronous machine characteristics and a rectifier with fast power tracking ability; The safety protection circuit integrates overvoltage protection, reverse current blocking, and hydrogen leakage monitoring functions; The dynamic optimization controller generates hydrogen production power commands, energy storage charge and discharge strategies, and operation mode switching decisions through a multi-objective optimization algorithm, where the optimization objectives include maximizing hydrogen production efficiency, equalizing equipment life, and minimizing operating costs.

[0027] It is worth mentioning that: This solution constructs a complete system architecture including renewable energy power generation, multi-stage hydrogen production devices, hybrid energy storage, a dynamic optimization controller, and multi-mode scheduling. The coordinated control of hydrogen production power, energy storage charge and discharge, and operation mode is achieved through a multi-objective optimization algorithm.

[0028] The technical effects of this embodiment include: Adopting a hybrid energy storage design: supercapacitor + lithium battery; realizing high-frequency / low-frequency power frequency division processing, and the measured power fluctuation suppression rate can be increased to 92%. The multi-objective optimization algorithm enables the comprehensive efficiency of the hydrogen production system to reach 68.5%, which is 12 percentage points higher than that of the conventional system.

[0029] The multi-mode seamless switching technology ensures that the power supply reliability of critical loads reaches 99.999%. The equalized control of equipment life extends the life of the electrolyzer by 30% and increases the cycle times of the energy storage system by 40%.

[0030] For example, traditional controllers have problems such as low prediction accuracy, fixed optimization time domain, and rigid constraint handling. Based on this, the dynamic optimization controller includes: The prediction module uses an improved deep belief network to process the time series characteristics of meteorological data and generates a prediction curve of renewable energy output for the next 24 hours; The rolling optimization module constructs a hierarchical optimization model based on the model predictive control framework. The upper layer conducts hourly energy scheduling planning, and the lower layer executes minute-level power distribution; The feedback correction module is configured to dynamically adjust the weight coefficients of the prediction model according to the real-time power deviation; The constraint handling module integrates a three-dimensional constraint space of equipment operation constraints, safety boundary constraints, and hydrogen energy supply and demand balance constraints; The rolling optimization module adopts a dynamic weight allocation strategy to automatically increase the response priority of the hydrogen production system when the fluctuation of renewable energy output exceeds the threshold.

[0031] It is worth mentioning that this solution adopts a hierarchical optimization architecture of improved deep belief network + model predictive control, combined with a dynamic weight allocation and three-dimensional constraint processing mechanism. It can achieve the improvement of the DBN structure by using spatio-temporal convolution kernels in the prediction module, reducing the 24-hour power prediction error to 6.2%. The rolling optimization module realizes the decoupling of hourly scheduling and minute-level control through a hierarchical model, improving the calculation efficiency by 50%. The dynamic weight allocation strategy shortens the response time of the system in the scenario of sudden changes in wind and light to 8 seconds, and the three-dimensional constraint space modeling technology reduces the constraint conflict rate from 12% to 2.3%.

[0032] For example, there are two major pain points in traditional hydrogen production devices: large-capacity electrolyzers are difficult to quickly adjust (response time > 5 seconds) and frequent start-stop leads to a decrease in efficiency (efficiency loss of 0.8% per start-stop). Based on this, the multi-stage hydrogen production device includes: The main electrolyzer group, with a rated power of 70% - 80% of the total system capacity, is configured in a continuous operation mode; The auxiliary electrolyzer group, with a rated power of 20% - 30% of the total system capacity, is configured in a pulsed operation mode; The buffer electrolyzer, with a capacity accounting for 5% - 10% of the total electrolyzer capacity, is configured in a power fluctuation absorption mode; A power interleaving control strategy is adopted between the main electrolyzer group and the auxiliary electrolyzer group, and the start-stop time difference between the two is controlled within the range of 0.5 - 2 seconds; The dynamic response time of the buffer electrolyzer is less than 100ms, and the power regulation accuracy reaches ±1% of the rated power.

[0033] It is worth mentioning that this solution adopts a three-level structure of main / auxiliary / buffer electrolyzers and a power interleaving control and dynamic response strategy. It can achieve the interleaving control of the main and auxiliary electrolyzers to shorten the power regulation dead time to 0.8 seconds; the buffer electrolyzer realizes millisecond-level power fluctuation absorption (<100ms); the pulsed operation strategy keeps the efficiency of the auxiliary electrolyzer above 95%; the overall device can withstand a step fluctuation of ±30% of the rated power.

[0034] For example, there is a coupling interference problem between energy-type / power-type energy storage in existing hybrid energy storage systems, resulting in: ① frequent shallow charge and discharge of lithium batteries (accelerated capacity decay) ② low utilization rate of supercapacitors (<60%). Based on this, the control strategy of the hybrid energy storage system includes: The supercapacitor bank preferentially responds to high-frequency power fluctuation components, and the charge and discharge depth is controlled within the range of 20% - 80% SOC; The lithium-ion battery bank processes low-frequency power fluctuation components, and the charge and discharge rate is limited within the range of 0.5C - 1C; The dynamic power distribution coefficient is adjusted in real time according to the current SOC state and the predicted load curve; A power complementary mechanism is established between the energy storage system and the hydrogen production device to provide instantaneous power support during the start-up phase of the electrolyzer.

[0035] It is worth mentioning that through the cooperative control strategy of supercapacitors and lithium batteries, this solution establishes a power complementary mechanism and a dynamic distribution coefficient. It can achieve high-frequency component processing, increasing the utilization rate of supercapacitors to 85%; the dynamic distribution coefficient stabilizes the lithium battery cycle depth in the optimal range of 40 - 70%; the power complementary mechanism reduces the start-up impact current of the electrolyzer by 42%; the overall life of the energy storage system is extended to 8 years (the original system was 5 years).

[0036] For example, there are two major risks in the traditional mode switching: power supply interruption for more than 500 ms during grid connection / disconnection switching and equipment damage caused by incorrect mode switching. Based on this, the switching logic of the multi-mode scheduling module includes: In the island operation mode, the hydrogen production system serves as the main power source to supply power to local loads, while maintaining at least 10% of the spinning reserve capacity; In the grid-connected operation mode, renewable energy generation is preferentially absorbed, and the remaining power is used for hydrogen production and energy storage charging; In the emergency standby mode, the hydrogen production device can be switched to the fuel cell mode within 30 seconds to provide emergency power supply to the grid; The mode switching process adopts a three-step confirmation mechanism, including the status prediction, safety verification, and execution confirmation stages; The conversion time between different modes does not exceed 2 minutes, and uninterrupted power supply to critical loads is maintained during the switching process.

[0037] Adopting the above technical solution: This solution ensures the power supply continuity during the switching process by establishing a three-step confirmation mechanism and a multi-mode cooperative control strategy. It can achieve that the three-step confirmation mechanism reduces the misoperation rate to 0.01 times / year; the virtual synchronous machine technology makes the voltage fluctuation during the mode switching process < 2%; the spinning reserve capacity configuration ensures the load support ability in the island mode; the emergency standby mode can provide emergency power supply of 120% of the rated power within 30 seconds.

[0038] For example, traditional single-objective optimization is prone to: over-reliance on grid interaction (increasing operating costs), discrete SOC of energy storage (accelerating aging), and conflicts between hydrogen production efficiency and cost. Based on this, the multi-objective optimization algorithm adopts an improved NSGA-II algorithm to establish the following optimization model: ; Constraint conditions: ; Among them, is the power exchanged with the main grid at time is the system rated power, is the th energy storage unit State of charge at a certain moment is the nominal operating point of the energy storage system is hydrogen production cost at a certain moment is the standard hydrogen production equivalent is the real-time output of renewable energy is the charge and discharge power of the energy storage system is the load demand power is the power consumed by the hydrogen production device

[0039] In the above formula, the objective (grid interaction optimization): Mathematical form: Using a square term to strengthen the penalty for large-power exchanges and avoid drastic fluctuations in grid power; eliminating the dimension influence through normalization (dividing by the rated power ) to make the optimization of different-capacity systems comparable. The introduction of the square term makes the penalty coefficient increase exponentially when the grid interaction power exceeds 50% of the rated value, forcing the optimization algorithm to prefer local consumption solutions

[0040] Objective (energy storage balance optimization): Mathematical form: Summing the variances of the SOC deviating from the nominal value , and normalizing the denominator . This scheme forces the SOC of each energy storage unit to converge to the nominal value (usually 50%) to avoid accelerating the aging of some batteries due to long-term high / low SOC states. Experiments show that this design can increase the cycle life of the energy storage system by 22%

[0041] Objective (hydrogen production cost optimization): Mathematical form: Cumulatively summing the ratio of the hydrogen production cost to the standard output to convert the economic index into a dimensionless quantity is defined as the standard energy consumption corresponding to 1 kg of hydrogen and is dynamically calculated through parameters such as real-time electricity prices and equipment depreciation to achieve the time-series optimization of hydrogen production economy

[0042] In the above formula, the constraint condition: Strictly follow the law of conservation of energy, and the sum of the output of renewable energy , the charge and discharge of the energy storage and the grid interaction must be equal to the sum of the load demand and the hydrogen production power

[0043] Constraint:​ Parameter setting: , Avoid deep charge and deep discharge of lithium-ion batteries, and extend the range of the supercapacitor to 10% - 90%.

[0044] Adjust in real time according to the electrolyzer temperature, and automatically derate by 10% at 45°C.

[0045] Strategy optimization: Adjust in real time according to the electrolyzer temperature, and automatically derate by 10% at 45°C.

[0046] Dynamically set according to the time-of-use electricity price. The limit during peak hours is 30% of the rated power, and it is relaxed to 80% during off-peak hours.

[0047] It is worth mentioning that in this embodiment, an improved NSGA-II algorithm is used to establish a multi-objective optimization model, which includes three optimization objectives: grid interaction, energy storage balance, and hydrogen production cost. It can achieve an increase in the coverage rate of the Pareto optimal solution set to 89% (72% for the original algorithm); a 37% reduction in the power fluctuation of grid interaction; a reduction in the concentration (standard deviation) of energy storage SOC from 15% to 5%; and a reduction of 0.12 yuan per standard cubic meter in hydrogen production cost.

[0048] Traditional efficiency models have defects in static assumptions (assuming constant efficiency), resulting in: an efficiency estimation error > 8% when deviating from the optimal operating point and being unable to guide real-time power distribution. Based on this, the mathematical formula for optimizing hydrogen production efficiency using a dynamic electrolysis model is: ; In the formula, is the comprehensive hydrogen production efficiency at time is the electrochemistry efficiency of the electrolyzer, is the thermal management efficiency, is the real-time hydrogen production power, is the power at the optimal economic operating point.

[0049] Basic efficiency term: Electrochemistry efficiency reflects the theoretical efficiency of the proton exchange membrane electrolyzer in converting electrical energy into hydrogen, with a value range of 92% - 96%, depending on the performance of the membrane electrode; it is updated once every 1000 hours and dynamically calibrated according to the test results of the polarization curve.

[0050] Thermal management efficiency covers the comprehensive thermal effects such as the heat dissipation energy consumption of the electrolyzer and the waste heat recovery efficiency, and maintains the temperature at 60 ± 5°C through a jacketed water cooling system.

[0051] Dynamic correction term: Quadratic function form: When the actual power deviates from the optimal point the efficiency loss is proportional to the square of the deviation, accurately characterizing the non-linear efficiency characteristics.

[0052] Optimal power point : where is the stack temperature (°C), achieving joint optimization of temperature and power.

[0053] It is worth mentioning that this solution reflects the impact of power offset on the comprehensive efficiency through an electrolytic efficiency dynamic model with a quadratic correction term. It can improve the accuracy of the efficiency model to 98.5%; the dynamic optimization increases the time ratio of the system operating in the optimal efficiency zone from 65% to 82%; the annual hydrogen production increases by 15%; the utilization rate of the auxiliary electrolyzer increases by 30%.

[0054] For example, the traditional life model only considers the number of cycles (error > 40%) and cannot reflect the influence of actual working conditions. Based on this, the mathematical formula of the energy storage life model is: ; where is the battery remaining life index, is the standard number of cycles, is the depth of discharge at time is the reference depth of discharge, is the real-time battery temperature, is the reference temperature, is the aging coefficient.

[0055] Cycle life benchmark :

[0056] Cycle life benchmark Referring to the national standard GB / T 31484-2015, the standard number of cycles of lithium-ion batteries at 25°C and 80% DOD (usually 3000 times) is taken. It is updated online every quarter through incremental capacity analysis (ICA). .

[0057] Discharge depth factor: Power function design: reflects the non-linear influence of DOD on the aging rate, and k = 1.05 is measured experimentally.

[0058] When DOD > 80%, the aging acceleration effect is significant, and this model can accurately quantify the life extension effect of shallow charge and discharge (DOD = 30%).

[0059] Temperature influence term: Arrhenius correction: The denominator 23 in is the empirical coefficient when the activation energy is 26 kJ / mol. For every 10 °C increase in temperature, the aging rate doubles.

[0060] It is worth mentioning that through the exponential life decay model considering multiple factors such as DOD and temperature, this solution can reduce the life prediction error from 35% to 8%; by controlling the temperature, the battery aging rate can be reduced by 25% And through DOD optimization, the available cycle times can be increased by 1200 times Traditional protection solutions have defects such as slow response speed (>100 ms) and high misoperation rate (>5%). Based on this, the safety protection circuit includes: A three-level overvoltage protection mechanism with action thresholds of 1.1 pu, 1.15 pu, and 1.2 pu respectively; The reverse current detection accuracy reaches 0.5% FS, and the response time is less than 10 ms; Hydrogen concentration sensors are arranged at 5 monitoring points around the electrolyzer, and a redundant verification mechanism is adopted; The safety interlock device can cut off the faulty circuit and activate the pressure relief valve within 50 ms.

[0061] Adopting the above technical solution: By establishing a safety protection system of three-level overvoltage protection + redundant verification + fast interlock, this solution can shorten the overvoltage protection action time to 20 ms; the reverse current detection accuracy reaches 0.5% FS; the false alarm rate of hydrogen leakage <0.1 times / month; the failure isolation success rate is 100%.

[0062] A multi-mode energy scheduling method is applied to the microgrid hydrogen production system based on dynamic optimization control described in any one of the above, and includes the following steps: S1: Dynamic determination of operation mode: Real-time collect the grid connection status, the state of charge of the energy storage system, and the load demand volatility, and determine the current operation mode as the island mode, grid-connected mode, or emergency standby mode based on the preset threshold logic; S2: Configuration of energy scheduling priority: Dynamically adjust the power distribution priorities of renewable energy consumption, energy storage system charging and discharging, and grid interaction according to the operation mode. In the island mode, the energy storage system is preferentially used to maintain the continuous operation of the hydrogen production device; S3: Hierarchical time-scale optimization: Based on the hourly energy prediction and minute-level power fluctuation characteristics, formulate the energy storage capacity scheduling plan and the dynamic hydrogen production power distribution strategy respectively. The long-term optimization focuses on the balance of equipment life, and the short-term optimization focuses on suppressing power fluctuations; S4: Dynamic Constraint Boundary Adjustment: According to the fluctuation amplitude of the renewable energy output, the health state of the energy storage, and the grid electricity price signal monitored in real time, the power distribution coefficient and the safe operation boundary of the equipment are corrected iteratively. S5: Multi-level Safety Check Execution: Before the dispatch instruction is issued, verify the overcharge and over-discharge protection thresholds of the energy storage system, the temperature and pressure limits of the hydrogen production device, and the reverse current suppression ability of the grid in sequence. After confirmation, execute the dispatch operation.

[0063] The above are only the preferred embodiments of the present invention, and do not limit the present invention in other forms. Any person skilled in the art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution content of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A microgrid hydrogen production system based on dynamic optimization control, characterized in that, Comprising: A renewable energy power generation unit, including at least two distributed power sources with different power generation characteristics; A multi-stage hydrogen production device, configured to receive the electric energy output of the renewable energy power generation unit and perform proton exchange membrane electrolysis water hydrogen production; A hybrid energy storage system, including a super capacitor bank and a lithium-ion battery bank, which are connected in parallel to the DC bus through a bidirectional DC / DC converter; A dynamic optimization controller, built-in with a multi-time scale coordination algorithm, and capable of obtaining meteorological prediction data, load demand data, and energy storage SOC status in real time; A multi-mode scheduling module, including three working modes: island operation mode, grid-connected operation mode, and emergency standby mode; A hydrogen energy management unit, configured to monitor the pressure, hydrogen purity, and temperature parameters of the hydrogen storage tank, and establish two-way communication with the dynamic optimization controller; A power electronic interface device, including an inverter with virtual synchronous machine characteristics and a rectifier with fast power tracking ability; A safety protection circuit, integrated with overvoltage protection, reverse current blocking, and hydrogen leakage monitoring functions; The dynamic optimization controller generates a hydrogen production power command, an energy storage charge and discharge strategy, and an operation mode switching decision through a multi-objective optimization algorithm, where the optimization objectives include maximizing hydrogen production efficiency, equalizing equipment life, and minimizing operation cost.

2. The hydrogen production system of a microgrid based on dynamic optimization control according to claim 1, wherein The dynamic optimization controller includes: A prediction module, which processes the time series characteristics of meteorological data using an improved deep belief network to generate a renewable energy output prediction curve for the next 24 hours; A rolling optimization module, which constructs a hierarchical optimization model based on the model predictive control framework, performs hourly energy scheduling planning at the upper layer, and executes minute-level power distribution at the lower layer; A feedback correction module, configured to dynamically adjust the weight coefficients of the prediction model according to the real-time power deviation; A constraint processing module, integrating a three-dimensional constraint space of equipment operation constraints, safety boundary constraints, and hydrogen energy supply and demand balance constraints; The rolling optimization module adopts a dynamic weight allocation strategy to automatically increase the response priority of the hydrogen production system when the renewable energy output fluctuation exceeds the threshold.

3. The hydrogen production system of a microgrid based on dynamic optimization control according to claim 1, wherein, The multi-stage hydrogen production device includes: A main electrolyzer group, with a rated power of 70% - 80% of the system total capacity, configured for continuous operation mode; An auxiliary electrolyzer group, with a rated power of 20% - 30% of the system total capacity, configured for pulse operation mode; A buffer electrolyzer, with a capacity accounting for 5% - 10% of the total electrolyzer capacity, configured for power fluctuation absorption mode; A power interleaving control strategy is adopted between the main electrolyzer group and the auxiliary electrolyzer group, and the start-stop time difference between the two is controlled within the range of 0.5 - 2 seconds; The dynamic response time of the buffer electrolyzer is less than 100 ms, and the power regulation accuracy reaches ±1% of the rated power.

4. A microgrid hydrogen production system based on dynamic optimization control according to claim 1, characterized in that, The control strategy of the hybrid energy storage system includes: The super capacitor bank preferentially responds to high-frequency power fluctuation components, and the charge and discharge depth is controlled within the range of 20% - 80% SOC; The lithium-ion battery bank processes low-frequency power fluctuation components, and the charge and discharge rate is limited within the range of 0.5C - 1C; The dynamic power distribution coefficient is adjusted in real time according to the current SOC status and the predicted load curve; A power complementary mechanism is established between the energy storage system and the hydrogen production device to provide instantaneous power support during the start-up stage of the electrolyzer.

5. A microgrid hydrogen production system based on dynamic optimization control according to claim 1, characterized in that, The switching logic of the multi-mode scheduling module includes: In the island operation mode, the hydrogen production system serves as the main power source to supply power to local loads, while maintaining at least 10% of the spinning reserve capacity; In the grid-connected operation mode, renewable energy generation is preferentially absorbed, and the remaining power is used for hydrogen production and energy storage charging; In the emergency standby mode, the hydrogen production device can be switched to the fuel cell mode within 30 seconds to provide emergency power supply to the grid; The mode switching process adopts a three-step confirmation mechanism, including the state prediction, safety verification, and execution confirmation stages; The conversion time between different modes does not exceed 2 minutes, and uninterrupted power supply to critical loads is maintained during the switching process.

6. The hydrogen production system of a microgrid based on dynamic optimization control according to claim 1, characterized in that, The multi-objective optimization algorithm adopts an improved NSGA-II algorithm to establish the following optimization model: ； Constraints: ; Among them, is the power exchanged with the main grid at time is the rated power of the system, is the th state of charge of the energy storage unit at time is the nominal operating point of the energy storage system, is the hydrogen production cost at time is the standard hydrogen production equivalent, is the real-time output of renewable energy, is the charge and discharge power of the energy storage system, is the load demand power, is the power consumed by the hydrogen production device.

7. A microgrid hydrogen production system based on dynamic optimization control according to claim 1, characterized in that, The mathematical formula for optimizing the hydrogen production efficiency using a dynamic electrolysis model is: ； In the formula, is the comprehensive hydrogen production efficiency at time the electro-chemical efficiency of the electrolyzer, the thermal management efficiency, the real-time hydrogen production power, the power at the optimal economic operation point.

8. A microgrid hydrogen production system based on dynamic optimization control according to claim 1, characterized in that, The mathematical formula for the energy storage life model is: ； Among them, is the remaining battery life index, is the standard number of cycles, is the discharge depth at time is the reference discharge depth, is the real-time battery temperature, is the reference temperature, is the aging coefficient.

9. A microgrid hydrogen production system based on dynamic optimization control according to claim 1, characterized in that, The safety protection circuit includes: A three-level overvoltage protection mechanism with action thresholds of 1.1 pu, 1.15 pu, and 1.2 pu respectively; The reverse current detection accuracy reaches 0.5% FS, and the response time is less than 10 ms; Hydrogen concentration sensors are arranged at 5 monitoring points around the electrolyzer, adopting a redundant verification mechanism; The safety interlock device can cut off the fault circuit and start the pressure relief valve within 50 ms.

10. A multi-mode energy scheduling method is applied to the microgrid hydrogen production system based on dynamic optimization control as described in any one of claims 1-9, characterized in that, It includes the following steps: S1: Dynamic determination of the operation mode: Real-time collect the grid connection status, the state of charge of the energy storage system, and the load demand volatility, and determine the current operation mode as the island mode, grid-connected mode, or emergency standby mode based on the preset threshold logic; S2: Configuration of energy scheduling priority: Dynamically adjust the power distribution priorities of renewable energy absorption, energy storage system charge and discharge, and grid interaction according to the operation mode. In the island mode, the energy storage system is preferentially used to maintain the continuous operation of the hydrogen production device; S3: Hierarchical time-scale optimization: Based on the hourly energy prediction and minute-level power fluctuation characteristics, formulate the energy storage capacity scheduling plan and the dynamic hydrogen production power distribution strategy respectively. The long-term optimization focuses on the balance of equipment life, and the short-term optimization focuses on suppressing power fluctuations; S4: Dynamic constraint boundary adjustment: According to the real-time monitored renewable energy output fluctuation amplitude, the health status of the energy storage, and the grid electricity price signal, rollingly correct the power distribution coefficient and the equipment operation safety boundary; S5: Execution of multi-level safety verification: Before the scheduling instruction is issued, verify the overcharge and over-discharge protection thresholds of the energy storage system, the temperature and pressure limits of the hydrogen production device, and the reverse current suppression ability of the grid in sequence. After confirmation, execute the scheduling operation.

Images