Renewable energy hydrogen production control system based on dynamic power matching and hydrogen storage cooperation

Abstract

The invention relates to the technical field of new energy and energy storage, and provides a renewable energy hydrogen production control system based on dynamic power matching and hydrogen storage collaboration, which comprises a power conversion system with a constant voltage mode and constant power mode millisecond switching capability; the modularized electrolytic cell array is formed by connecting a plurality of proton exchange membrane electrolytic cells in parallel; the hydrogen storage cooperation unit comprises a high-pressure hydrogen storage tank group, a solid oxide fuel cell and a hydrogen purification device; the intelligent control unit is used for executing a three-stage control strategy; the real-time control layer is used for executing proton exchange membrane electrolytic cell power instruction tracking and power grid frequency deviation compensation; the optimization scheduling layer adopts a reinforcement learning model prediction control algorithm to perform rolling optimization on the power distribution proportion of hydrogen production and power generation; and the planning layer is used for generating a proton exchange membrane electrolytic cell start-stop plan and a hydrogen storage scheduling plan in combination with meteorological prediction and electricity market transaction data. The problems that wind-solar power generation fluctuation is large, and a traditional hydrogen production system is slow in response and insufficient in economical efficiency can be solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of new energy and energy storage technology, and particularly relates to a renewable energy hydrogen production control system based on dynamic power matching and hydrogen storage cooperation. BACKGROUND

[0002] With the continuous expansion of global renewable energy (wind power, photovoltaic) installed capacity, power grid operation is facing double challenges: on the one hand, the intermittency and volatility of wind and solar power generation (such as wind power affected by sudden changes in wind speed, photovoltaic power affected by day and night alternation) lead to intensified power grid power fluctuations, and the problem of "sending out when there is electricity, and not being able to make up for the lack of electricity" is prominent; on the other hand, the response speed of traditional power grid frequency modulation resources (such as thermal power units) is slow, and it is difficult to match the rapid fluctuation demand of renewable energy, and the stability of the power grid is under test.

[0003] Under this background, the use of renewable energy to produce hydrogen (by electrolyzing water to convert electrical energy into hydrogen energy) is considered as an important means to solve wind and light consumption and improve the flexibility of the power grid. However, there are still significant technical bottlenecks in the coordinated control of traditional hydrogen production systems and power grids, which are manifested in the following three aspects: 1. Insufficient dynamic response capability of existing hydrogen production systems The power regulation rate of electrolytic tank, as the core equipment of hydrogen production, directly affects the response capability to wind and light fluctuations. In the traditional scheme, the power regulation of electrolytic tank (especially alkaline electrolytic tank) has obvious lag: for example, the dynamic response time of traditional alkaline electrolytic tank from high load to low load is more than 500ms (some systems even need several minutes), and frequent load change will lead to a decrease in electrolytic efficiency (the regulation rate is only about 2% / s), which is difficult to meet the demand of power grid primary frequency modulation (which requires second-level or even millisecond-level response).

[0004] 2. Lack of economic scheduling of hydrogen storage system and power grid Although hydrogen storage can achieve energy buffering through "hydrogen charging-hydrogen discharging", in the traditional scheme, the interaction between hydrogen storage system and power grid only stays at the level of "passive storage / discharge", and lacks dynamic matching strategies with power grid electricity price and load demand. When the power grid is in the peak electricity price period, the hydrogen storage fails to preferentially generate electricity through fuel cells to participate in peak shaving; in the low valley period, the electrolytic tank also fails to efficiently produce hydrogen to reduce energy cost, resulting in high hydrogen leveling cost (LCOH) (industry average about 22.7 yuan / kg).

[0005] 3. Insufficient safety redundancy design of the system Safety hazards in hydrogen production and storage systems permeate the entire process, including hydrogen leaks, equipment overloads, and abnormal pressure. Traditional solutions suffer from slow fault detection and isolation mechanisms: for example, hydrogen leak detection thresholds are high (typically >500ppm), and the time from detection to shutting down the electrolyzer exceeds one second; when equipment temperature or pressure exceeds limits, the pressure relief valve activation is delayed (>500ms), increasing the risk of fault propagation and failing to meet the basic requirements for safe operation of power grid equipment (such as the IEC61850 standard requiring fault isolation time <1 second).

[0006] In summary, traditional solutions have the following prominent problems in the coordinated control of renewable energy hydrogen production and power grid: The slow dynamic response of the electrolyzer (response time > 500ms, adjustment rate ≤ 2% / s) results in the inability to absorb the fluctuating power of wind and solar power generation in a timely manner, with the curtailment rate of wind and solar power reaching over 15%. The hydrogen storage system and the power grid lack an economical dispatch strategy, and the hydrogen production cost is high while the power grid's support capacity is weak. Insufficient safety redundancy design (fault isolation time > 1 second) results in high system operation risk.

[0007] The aforementioned technological shortcomings severely restrict the large-scale application of hydrogen production from renewable energy sources and the improvement of grid flexibility, and urgently need to be addressed through innovative control strategies and system architectures. Summary of the Invention

[0008] The purpose of this invention is to solve at least one technical problem in the background art and to provide a renewable energy hydrogen production control system based on dynamic power matching and hydrogen storage synergy.

[0009] To achieve the above objectives, the present invention provides a renewable energy hydrogen production control system based on dynamic power matching and hydrogen storage synergy, comprising: Power conversion system: adopts a three-level IGBT converter topology, supports DC / AC bidirectional conversion, DC bus voltage range is 1000-1500VDC, has millisecond-level switching capability between constant voltage mode and constant power mode, and mode switching response time ≤50ms; Modular electrolyzer array: Composed of at least 3 proton exchange membrane electrolyzers connected in parallel, with a rated power of 5-20MW, a load regulation rate of ≥5% / s, and a cold start time of ≤5 minutes; Hydrogen storage and synergy unit: includes a high-pressure hydrogen storage tank group of 35MPa or above, a solid oxide fuel cell and a hydrogen purification device, wherein the power generation efficiency of the solid oxide fuel cell is ≥55% and the hydrogen-to-electricity conversion delay is ≤200ms; Intelligent control unit: integrates edge computing nodes and cloud platform to execute three-level control strategies: Real-time control layer: Based on FPGA hardware circuit, it performs power command tracking of proton exchange membrane electrolyzer and power grid frequency deviation compensation; Optimized scheduling layer: Reinforcement learning model predictive control algorithm is adopted to continuously optimize the power allocation ratio between hydrogen production and power generation; Planning layer: Combining meteorological forecasts and electricity market transaction data, generate start-up and shutdown plans for proton exchange membrane electrolyzers and hydrogen storage dispatch plans.

[0010] According to one aspect of the present invention, the constant voltage mode switching logic of the power conversion system satisfies the following condition: When the mains voltage drops below 0.9 pu, it automatically switches to constant voltage output mode, with output voltage fluctuation ≤ ±1%; In islanded mode, the output frequency accuracy of the power conversion system is 50Hz±0.2Hz, and the total harmonic distortion rate is ≤3%.

[0011] According to one aspect of the invention, the load regulation of the modular electrolytic cell array employs a composite control of feedforward control and feedback control: The feedforward control generates a baseline load rate based on the predicted wind and solar power values, with a prediction time window of 1-10 minutes. The feedback control dynamically corrects the load rate based on the real-time grid frequency deviation Δf. For every 0.01Hz change in Δf, the load rate is adjusted by 1%.

[0012] According to one aspect of the present invention, the SOC hierarchical control strategy of the hydrogen storage co-working unit includes: SOC < 20%: Discharge of solid oxide fuel cells is prohibited; proton exchange membrane electrolyzers should prioritize the utilization of waste electricity; hydrogen production power is limited to 10% of the rated value. 20%≤SOC<80%: Optimize the ratio of hydrogen production to power generation based on real-time electricity price signals, with the objective function being to minimize the levelized cost of hydrogen. SOC≥80%: Solid oxide fuel cells discharge at maximum power, and the proton exchange membrane electrolyzer load rate drops to ≤15%.

[0013] According to one aspect of the present invention, the reinforcement learning model prediction control algorithm includes: State space definition: The system state inputs are grid frequency deviation, state of charge of hydrogen storage co-processing unit, operating efficiency curve of proton exchange membrane electrolyzer, and real-time electricity price; among them, grid frequency deviation represents the supply and demand balance of the power system; state of charge of hydrogen storage co-processing unit reflects the availability of hydrogen storage capacity; operating efficiency curve of proton exchange membrane electrolyzer describes the dynamic mapping relationship between equipment operating efficiency and load rate; real-time electricity price serves as an external variable for economic regulation. Reward Function Design: Constructing a Multi-Objective Optimization Reward Function ; Where Δf is the grid frequency deviation, with a weighting coefficient of 0.6 to penalize frequency fluctuations and enhance system stability control; LCOH is the levelized production cost of hydrogen, with a weighting coefficient of 0.3, achieved by minimizing the difference between LCOH and the theoretical minimum. The deviation drives the optimization of system economy; As a penalty term for hydrogen storage status, a weighting coefficient of 0.1 is used to constrain the operating range of the hydrogen storage co-operation unit and avoid overcharging / over-discharging; Action space definition: Control commands consist of two adjustment dimensions: The load rate adjustment rate of the proton exchange membrane electrolyzer is allowed to be dynamically adjusted within the range of ±5% / s, enabling flexible control of power output. Solid oxide fuel cell output power command: set to a continuously adjustable range of 0-100% of rated power, balancing efficiency and power response.

[0014] According to one aspect of the present invention, the method for generating the hydrogen storage scheduling plan includes: The wind and solar power prediction model based on convolutional neural networks takes meteorological radar data and historical power output curves as inputs and has a prediction error RMSE ≤ 5%. Solve the mixed-integer linear programming model with a 24-hour period. The objective function is: ; in, (t) represents the grid electricity price in the t-th hour; Let t be the electrical power purchased from the grid in hour t. The electricity price for hour t; Let be the electrical power sold to the grid in hour t; The constraints are: The number of times a proton exchange membrane electrolyzer is started and stopped is ≤3 times / day; The pressure of the high-pressure hydrogen storage tank group is ≤45MPa.

[0015] According to one aspect of the present invention, the real-time control layer is provided with a safety fault-tolerance mechanism, the safety fault-tolerance mechanism comprising: Electrical protection: When the DC side voltage exceeds ±10% of the rated value, the power conversion system will be locked within 100ms; when the AC side current exceeds 1.2 times the rated value, the faulty unit will be cleared within 50ms. Hydrogen safety linkage: When the hydrogen leakage concentration is ≥100ppm or the temperature of the high-pressure hydrogen storage tank group is ≥85℃, the power supply of the proton exchange membrane electrolyzer will be shut off and the emergency pressure relief valve will be activated within 200ms.

[0016] According to one aspect of the present invention, the FPGA chip in the real-time control layer performs the following logic: Real-time data acquisition: Synchronously acquire power grid frequency signal, power grid voltage signal and proton exchange membrane electrolyzer temperature signal with a period of 10ms; Frequency anomaly response: When a power grid frequency deviation is detected In such cases, the hydrogen storage coordinating unit should be prioritized for power compensation. Compensation power calculation: Compensation power It consists of two parts, and the formula is: ; Wherein, 10MW / Hz·Δf is the instantaneous compensation term for frequency deviation; This is a cumulative compensation term for frequency deviation, suppressing long-term offset.

[0017] According to one aspect of the present invention, when the SOC of the hydrogen storage co-processing unit is ≥50%, the solid oxide fuel cell supplies power to the grid in a constant voltage and constant frequency mode; The proton exchange membrane electrolyzer is used as a virtual synchronous machine to simulate inertial response, with an inertial time constant H = 3-5 s; When the grid frequency is greater than 50.5 Hz, hydrogen production will be automatically started to absorb excess power.

[0018] According to the solution of the present invention, the present invention can solve the problems of large fluctuations in wind and solar power generation and slow response and insufficient economic efficiency of traditional hydrogen production systems.

[0019] This invention employs a three-level linkage control system of PCS (converter), electrolyzer, and hydrogen storage, which automatically adjusts the working status of the electrolyzer based on the real-time power of wind and solar power generation, ensuring a significant increase in the renewable energy consumption rate. This invention optimizes the hydrogen storage system based on its state of charge (SOC) in three tiers: prioritizing basic hydrogen production needs when the charge is low, balancing cost and efficiency when the charge is medium, and participating in grid peak shaving and valley filling when the charge is high. This invention features a three-level safety mechanism (immediate shutdown upon hydrogen leakage, rapid pressure relief in case of abnormal temperature, and switching to backup mode when battery is low), reducing fault isolation time to within 200ms and significantly improving safety compared to traditional systems.

[0020] This invention enables power grid frequency regulation accuracy to reach ±0.05Hz (regulation response time ≤500ms), effectively supporting power grid stability; the levelized cost of hydrogen (LCOH) is reduced by 18% compared with existing technologies, resulting in better economic efficiency; the electrolyzer load regulation rate is ≥5% / s, with fast dynamic response speed.

[0021] This invention integrates wind and solar power prediction, intelligent control of electrolyzers, SOFC (solid oxide fuel cell) power generation and hydrogen storage management modules, and can be widely used in scenarios such as wind and solar hydrogen production bases and microgrid black start, taking into account energy consumption, grid support and economic needs. Attached Figure Description

[0022] Figure 1 The diagram illustrates a system block diagram of a renewable energy hydrogen production control system based on dynamic power matching and hydrogen storage coordination according to one embodiment of the present invention. Detailed Implementation

[0023] The invention will now be discussed with reference to exemplary embodiments. It should be understood that the described embodiments are merely intended to enable those skilled in the art to better understand and thus implement the invention, and are not intended to imply any limitation on the scope of the invention.

[0024] As used herein, the term "comprising" and its variations are to be interpreted as open-ended terms meaning "including but not limited to". The term "based on" is to be interpreted as "at least partially based on". The terms "one embodiment" and "an embodiment" are to be interpreted as "at least one embodiment".

[0025] Figure 1 This schematic diagram illustrates a system block diagram of a renewable energy hydrogen production control system based on dynamic power matching and hydrogen storage coordination according to one embodiment of the present invention. Figure 1 As shown, in this embodiment, the renewable energy hydrogen production control system based on dynamic power matching and hydrogen storage synergy includes: Power conversion system (PCS): It adopts a three-level IGBT converter topology, supports DC / AC bidirectional conversion, and has a DC bus voltage range of 1000-1500VDC. It has the ability to switch between constant voltage mode and constant power mode in milliseconds, and the mode switching response time is ≤50ms. Modular electrolyzer array: It consists of at least 3 proton exchange membrane (PEM) electrolyzers connected in parallel. The rated power of the proton exchange membrane electrolyzer is 5-20MW, the load regulation rate is ≥5% / s, and the cold start time is ≤5 minutes. Hydrogen storage co-production unit: includes a high-pressure hydrogen storage tank group of 35MPa or above, a solid oxide fuel cell (SOFC) and a hydrogen purification device, wherein the power generation efficiency of the solid oxide fuel cell is ≥55% (LHV) and the hydrogen-to-electricity conversion delay is ≤200ms; Intelligent control unit: integrates edge computing nodes and cloud platform to execute three-level control strategies: Real-time control layer (≤100ms): Based on FPGA hardware circuit, it performs power command tracking of proton exchange membrane electrolyzer and grid frequency deviation compensation; Optimized scheduling layer (1s-1h): The reinforcement learning model predictive control algorithm (MPC) is used to continuously optimize the power allocation ratio between hydrogen production and power generation; Planning layer (≥24h): Combine meteorological forecasts and electricity market transaction data to generate start-up and shutdown plans for proton exchange membrane electrolyzers and hydrogen storage dispatch plans.

[0026] Furthermore, according to one embodiment of the present invention, the constant voltage mode switching logic of the power conversion system satisfies the following conditions: When the mains voltage drops below 0.9 pu, it automatically switches to constant voltage output mode, with output voltage fluctuation ≤ ±1%; In islanded mode, the output frequency accuracy of the power conversion system is 50Hz±0.2Hz, and the total harmonic distortion (THD) is ≤3%.

[0027] Furthermore, according to one embodiment of the present invention, the load regulation of the modular electrolytic cell array adopts a composite control of feedforward control and feedback control: The feedforward control generates a baseline load rate based on the predicted wind and solar power values, with a prediction time window of 1-10 minutes. The feedback control dynamically corrects the load rate based on the real-time grid frequency deviation Δf. For every 0.01Hz change in Δf, the load rate is adjusted by 1%.

[0028] Furthermore, according to one embodiment of the present invention, the SOC hierarchical control strategy of the hydrogen storage co-working unit includes: SOC < 20%: Discharge of solid oxide fuel cells is prohibited; proton exchange membrane electrolyzers should prioritize the utilization of waste electricity; hydrogen production power is limited to 10% of the rated value. 20%≤SOC<80%: Optimize the ratio of hydrogen production to power generation based on real-time electricity price signals, with the objective function being to minimize the levelized cost of hydrogen (LCOH). SOC≥80%: Solid oxide fuel cells discharge at maximum power, and the proton exchange membrane electrolyzer load rate drops to ≤15%.

[0029] Furthermore, according to one embodiment of the present invention, the reinforcement learning model predictive control algorithm (MPC) includes: State space definition: The system state inputs are grid frequency deviation, state of charge (SOC) of the hydrogen storage co-processing unit, operating efficiency curve of the proton exchange membrane electrolyzer, and real-time electricity price. Among them, grid frequency deviation represents the power system supply and demand balance; state of charge of the hydrogen storage co-processing unit reflects the availability of hydrogen storage capacity; operating efficiency curve of the proton exchange membrane electrolyzer describes the dynamic mapping relationship between equipment operating efficiency and load rate; and real-time electricity price serves as an external variable for economic regulation. Reward Function Design: Constructing a Multi-Objective Optimization Reward Function ; Where Δf is the grid frequency deviation, with a weighting coefficient of 0.6 to penalize frequency fluctuations and enhance system stability control; LCOH is the levelized production cost of hydrogen, with a weighting coefficient of 0.3, achieved by minimizing the difference between LCOH and the theoretical minimum. The deviation drives the optimization of system economy; As a penalty term for hydrogen storage status, a weighting coefficient of 0.1 is used to constrain the operating range of the hydrogen storage co-operation unit and avoid overcharging / over-discharging; Action space definition: Control commands consist of two adjustment dimensions: The load rate adjustment rate of the proton exchange membrane electrolyzer is allowed to be dynamically adjusted within the range of ±5% / s, enabling flexible control of power output. Solid oxide fuel cell output power command: set to a continuously adjustable range of 0-100% of rated power, balancing efficiency and power response.

[0030] Furthermore, according to one embodiment of the present invention, the method for generating a hydrogen storage scheduling plan includes: A wind and solar power prediction model based on convolutional neural networks (CNN) takes meteorological radar data and historical power output curves as input and has a prediction error RMSE ≤ 5%. Solve a mixed-integer linear programming (MILP) model with a 24-hour period, using the objective function as follows: ; in, (t) represents the grid electricity price (purchase price, which may include peak-valley price difference) for hour t. Let t be the electrical power purchased from the grid in hour t (decision variable, non-negative). The electricity price for hour t (the selling price, which is usually lower than the buying price or subject to grid constraints). Let t be the electrical power sold to the grid in hour t (decision variable, non-negative, limited by wind and solar power output and energy storage capacity). Core logic: Minimize the difference between the system's total daily electricity purchase cost and electricity sales revenue (i.e., maximize net revenue). The constraints are: The number of start-ups and shutdowns of a proton exchange membrane electrolyzer should be ≤3 times per day. Frequent start-ups and shutdowns of a proton exchange membrane electrolyzer (equipment used for hydrogen production) will exacerbate equipment wear and tear and increase energy consumption. Limiting the number of start-ups and shutdowns can extend its lifespan and reduce maintenance costs. The pressure of the high-pressure hydrogen storage tank group is ≤45Mpa; the upper limit of the safe operating pressure of the high-pressure hydrogen storage tank group (must meet the pressure vessel standard) to avoid the risk of overpressure, while ensuring the reasonable use of hydrogen storage capacity.

[0031] Furthermore, according to one embodiment of the present invention, the real-time control layer is provided with a safety fault-tolerance mechanism, which includes: Electrical protection: When the DC side voltage exceeds ±10% of the rated value, the power conversion system will be locked within 100ms; when the AC side current exceeds 1.2 times the rated value, the faulty unit will be cleared within 50ms. Hydrogen safety linkage: When the hydrogen leakage concentration is ≥100ppm or the temperature of the high-pressure hydrogen storage tank group is ≥85℃, the power supply of the proton exchange membrane electrolyzer will be shut off and the emergency pressure relief valve will be activated within 200ms.

[0032] Furthermore, according to one embodiment of the present invention, the FPGA chip in the real-time control layer executes the following logic: Real-time data acquisition: Synchronously acquire power grid frequency signal, power grid voltage signal and proton exchange membrane electrolyzer temperature signal with a period of 10ms; Frequency anomaly response: When a power grid frequency deviation is detected When the difference between the actual frequency and the rated frequency is ≥0.1Hz, the hydrogen storage co-processing unit will be called first to perform power compensation. Compensation power calculation: Compensation power It consists of two parts, and the formula is: ; Among them, 10MW / Hz⋅Δf is the instantaneous compensation term for frequency deviation, which quickly offsets instantaneous fluctuations; This is a cumulative compensation term for frequency deviation, suppressing long-term offset.

[0033] Furthermore, according to one embodiment of the present invention, the present invention supports the microgrid black-start function, and the execution steps include: When the SOC of the hydrogen storage co-processing unit is ≥50%, the solid oxide fuel cell supplies power to the grid in a constant voltage and constant frequency mode. The proton exchange membrane electrolyzer is used as a virtual synchronous machine to simulate inertial response, with an inertial time constant H = 3-5 s; When the grid frequency is greater than 50.5 Hz, hydrogen production will be automatically started to absorb excess power.

[0034] In this embodiment, for example: Microgrid Black Start: 1. Startup conditions With the main grid losing power, the state of charge (SOC) of hydrogen storage is 60%. Key equipment: SOFC generator set (500kW), electrolytic cell (10MW); 2. Operating Procedures: Main network disconnection → SOFC constant voltage start-up → Electrolyzer VSG mode activation → SOC monitoring; Frequency control: 50Hz±0.2Hz (THD≤3%); Power balance: The electrolytic cell absorbs excess power (maximum rate 3% / s); 3. Performance Indicators: The microgrid frequency is stable at 49.8-50.3Hz; Startup time ≤ 120s; Emergency battery life ≥ 72 hours.

[0035] According to the above-described renewable energy hydrogen production control system based on dynamic power matching and hydrogen storage synergy of the present invention, when applied to wind and solar hydrogen production bases, the system response process includes: When the output of wind and solar power drops by ≥30%, the proton exchange membrane electrolyzer reduces its load to the minimum safe load (10%) within 500ms, and the hydrogen storage fuel cell starts up simultaneously. After the output is restored, the proton exchange membrane electrolyzer ramps up to the target load rate at a slope of 3% / s to avoid step shock.

[0036] In this embodiment, for example: Wind and solar hydrogen production base (50MW wind power + 30MW photovoltaic); 1. System Configuration: Electrolytic cell array: 4×10MW PEM electrolytic cells in parallel; Hydrogen storage system: 5 × 2000 Nm³ hydrogen storage tanks (working pressure 35 MPa); SOFC unit: 3×500kW solid oxide fuel cells; 2. Typical operating conditions: When wind power output suddenly drops by 30%: Real-time control layer: The FPGA reduces the electrolyzer load rate from 80% to 30% within 500ms; Optimize the scheduling layer: adjust hydrogen production capacity based on real-time electricity price (1.2 yuan / kWh during peak periods); Planning level: Wind speed is predicted to decrease in the next 6 hours, and hydrogen storage capacity will be allocated to 60%; 3. Implementation Results: The curtailment rate decreased from 15% to 3%; Frequency modulation revenue increased by 25%; LCOH has been reduced to ¥18.6 / kg.

[0037] According to the above-described solution of the present invention, the present invention can solve the problems of large fluctuations in wind and solar power generation and slow response and insufficient economic efficiency of traditional hydrogen production systems.

[0038] This invention employs a three-level linkage control system of PCS (converter), electrolyzer, and hydrogen storage, which automatically adjusts the working status of the electrolyzer based on the real-time power of wind and solar power generation, ensuring a significant increase in the renewable energy consumption rate. This invention optimizes the hydrogen storage system based on its state of charge (SOC) in three tiers: prioritizing basic hydrogen production needs when the charge is low, balancing cost and efficiency when the charge is medium, and participating in grid peak shaving and valley filling when the charge is high. This invention features a three-level safety mechanism (immediate shutdown upon hydrogen leakage, rapid pressure relief in case of abnormal temperature, and switching to backup mode when battery is low), reducing fault isolation time to within 200ms and significantly improving safety compared to traditional systems.

[0039] This invention enables power grid frequency regulation accuracy to reach ±0.05Hz (regulation response time ≤500ms), effectively supporting power grid stability; the levelized cost of hydrogen (LCOH) is reduced by 18% compared with existing technologies, resulting in better economic efficiency; the electrolyzer load regulation rate is ≥5% / s, with fast dynamic response speed.

[0040] This invention integrates wind and solar power prediction, intelligent control of electrolyzers, SOFC (solid oxide fuel cell) power generation and hydrogen storage management modules, and can be widely used in scenarios such as wind and solar hydrogen production bases and microgrid black start, taking into account energy consumption, grid support and economic needs.

[0041] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in this application.

[0042] It should be understood that the sequence number of each step in the invention and its embodiments does not absolutely imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

Claims

1. A renewable energy hydrogen production control system based on dynamic power matching and hydrogen storage coordination, characterized in that, Comprise: Power conversion system: adopts three-level IGBT converter topology, supports DC / AC bidirectional conversion, DC bus voltage range is 1000-1500VDC, has constant voltage mode and constant power mode millisecond switching capability, mode switching response time ≤50ms; Modular electrolytic cell array: composed of at least 3 proton exchange membrane electrolytic cells in parallel, the rated power of proton exchange membrane electrolytic cell is 5-20MW, load regulation rate ≥5% / s, cold start time ≤5 minutes; Hydrogen storage coordination unit: including 35MPa or above high pressure hydrogen storage tank group, solid oxide fuel cell and hydrogen purification device, the power generation efficiency of the solid oxide fuel cell is ≥55%, and the hydrogen-electric conversion delay is ≤200ms; Intelligent control unit: integrates edge computing node and cloud platform, executes three-level control strategy: Real-time control layer: based on FPGA hardware circuit, executes proton exchange membrane electrolytic cell power instruction tracking and power grid frequency deviation compensation; Optimal scheduling layer: adopts reinforcement learning model predictive control algorithm, rolls to optimize the power distribution ratio of hydrogen production and power generation; Planning layer: combined with weather forecast and power market transaction data, generates proton exchange membrane electrolytic cell start-stop plan and hydrogen storage scheduling plan.

2. The control system for renewable energy based hydrogen production with dynamic power matching and hydrogen storage synergy of claim 1, wherein, The constant voltage mode switching logic of the power conversion system satisfies the following conditions: When the grid voltage drops to 0.9pu or below, automatically switch to constant voltage output mode, output voltage fluctuation ≤±1%; In island mode, the output frequency accuracy of the power conversion system is 50Hz±0.2Hz, and the total harmonic distortion rate is ≤3%.

3. The control system for renewable energy based hydrogen production with dynamic power matching and hydrogen storage synergy of claim 1, wherein, The load regulation of the modular electrolytic cell array adopts composite control of feedforward control and feedback control: The feedforward control generates a reference load rate based on wind and light power prediction value, and the prediction time window is 1-10 minutes; The feedback control dynamically corrects the load rate according to the real-time power grid frequency deviation Δf, and the load rate is adjusted by 1% for every 0.01Hz change of Δf.

4. The control system for renewable energy based hydrogen production with dynamic power matching and hydrogen storage synergy of claim 1, wherein, The SOC grading control strategy of the hydrogen storage coordination unit includes: SOC<20%: prohibit solid oxide fuel cell discharge, proton exchange membrane electrolytic cell preferentially consumes abandoned electricity, hydrogen production power lower limit is 10% of rated value; 20%≤SOC<80%: optimize hydrogen production and power generation ratio according to real-time electricity price signal, target function is minimum hydrogen cost minimization; SOC≥80%: solid oxide fuel cell discharges at maximum power, proton exchange membrane electrolytic cell load rate decreases to ≤15%.

5. The control system for renewable energy based hydrogen production with dynamic power matching and hydrogen storage synergy of claim 1, wherein, The reinforcement learning model predictive control algorithm includes: State space definition: take power grid frequency deviation, hydrogen storage coordination unit state of charge, proton exchange membrane electrolytic cell operation efficiency curve and real-time electricity price as system state input; Among them, the power grid frequency deviation represents the balance state of power supply and demand system; The state of charge of hydrogen storage coordination unit reflects the availability of hydrogen storage capacity; The operation efficiency curve of proton exchange membrane electrolytic cell describes the dynamic mapping relationship between device operation efficiency and load rate; Real-time electricity price as an external variable for economic regulation and control; Reward function design: construct a multi-objective optimization reward function: ； Where Δf is the grid frequency deviation, with a weighting coefficient of 0.6 to penalize frequency fluctuations and enhance system stability control; LCOH is the levelized production cost of hydrogen, with a weighting coefficient of 0.3, achieved by minimizing the difference between LCOH and the theoretical minimum. The deviation drives the optimization of system economy; As a penalty term for hydrogen storage status, a weighting coefficient of 0.1 is used to constrain the operating range of the hydrogen storage co-operation unit and avoid overcharging / over-discharging; Action space definition: control instruction contains two adjustment dimensions: The load rate adjustment rate of the proton exchange membrane electrolyzer is allowed to be dynamically adjusted within ±5% / s, realizing flexible control of power output; The output power instruction of the solid oxide fuel cell is set to a continuous adjustable interval of 0-100% of the rated power, considering both efficiency and power response capability.

6. The control system for renewable energy based hydrogen production with dynamic power matching and hydrogen storage synergy of claim 1, wherein, The method for generating the hydrogen storage scheduling plan comprises: The wind and light power prediction model based on a convolutional neural network has input of meteorological radar data and historical output curves, and a prediction error RMSE≤5%; A mixed integer linear programming model is solved in a 24-hour cycle, and the objective function is: ； wherein, (t) is the electricity grid price at the tth hour; Pb(t) is the electrical power bought from the grid for the tth hour; Pth= electricity selling price for the tth hour; Electric power sold to the grid for the tth hour; The constraint condition is: The start-stop number of the proton exchange membrane electrolyzer is ≤3 times / day; The pressure of the high-pressure hydrogen storage tank group is ≤45 MPa.

7. The control system for renewable energy based hydrogen production with dynamic power matching and hydrogen storage synergy of claim 1, wherein, The real-time control layer is provided with a safety fault-tolerant mechanism, and the safety fault-tolerant mechanism comprises: Electrical protection: when the DC side voltage exceeds ±10% of the rated value, the power conversion system is triggered to be locked within 100 ms; when the AC side current exceeds 1.2 times the rated value, the faulty unit is removed within 50 ms; Hydrogen safety linkage: when the hydrogen leakage concentration is ≥100 ppm or the high-pressure hydrogen storage tank group temperature is ≥85℃, the power supply of the proton exchange membrane electrolyzer is turned off within 200 ms and the emergency pressure relief valve is started.

8. The control system for renewable energy based hydrogen production with dynamic power matching and hydrogen storage synergy of claim 1, wherein, The FPGA chip in the real-time control layer executes the following logic: Real-time data acquisition: the power grid frequency signal, the power grid voltage signal and the proton exchange membrane electrolyzer temperature signal are synchronously collected in a 10 ms cycle; Frequency abnormal response: when the power grid frequency deviation is detected the hydrogen storage coordination unit is preferentially called to perform power compensation; Compensation power calculation: compensation power The formula is composed of two parts: ； wherein 10 MW / Hz·Δf is an immediate compensation term for the frequency deviation; is a cumulative compensation term for the frequency deviation, which suppresses long-term drift.

9. The renewable energy based hydrogen production control system in coordination with dynamic power matching and hydrogen storage as claimed in any of claims 1-8, wherein, When the SOC of the hydrogen storage coordination unit is ≥50%, the solid oxide fuel cell supplies power to the power grid in a constant voltage and constant frequency mode; The proton exchange membrane electrolyzer acts as a virtual synchronous machine, simulates inertia response, and the inertia time constant H=3-5s; When the power grid frequency is >50.5Hz, hydrogen production is automatically started to absorb excess power.

Images