Method for producing hydrogen by electrolyzing water through wind power coupled proton exchange membrane

Abstract

The invention relates to the technical field of hydrogen production, in particular to a wind power coupling proton exchange membrane water electrolysis hydrogen production method. According to the technical scheme, the method for producing hydrogen by electrolyzing water through the wind power coupled proton exchange membrane comprises the following steps: S1, acquiring a wind power signal in real time, and dividing the wind power signal into a low-frequency component and a high-frequency component according to a preset frequency division band; s2, dynamically distributing the low-frequency component to at least one target sub-pile in a plurality of PEM electrolytic cell sub-piles for operation; s3, absorbing the fluctuation energy of the high-frequency component through a hybrid energy storage module; and S4, based on a model prediction control MPC algorithm, optimizing the operation state of each sub-reactor, and enabling each sub-reactor to operate within 40-100% of the load in the efficient interval of the rated power of each sub-reactor. Through collaborative optimization of frequency-band-divided power distribution, multi-pile dynamic operation and an intelligent control strategy, comprehensive improvement of the hydrogen production efficiency, the equipment service life and the energy utilization rate is achieved in a wind power fluctuation scene.

Description

Technical Field

[0001] The present invention relates to the technical field of hydrogen production, and in particular to a method for producing hydrogen by wind power coupled proton exchange membrane electrolysis of water. Background Art

[0002] Currently, the mainstream electrolytic water hydrogen production technologies include two types: alkaline electrolytic water (ALK) and proton exchange membrane electrolytic water (PEM):

[0003] Alkaline electrolytic water (ALK): Using potassium hydroxide solution as the electrolyte and nickel-based catalyst, the operating temperature is 70 - 90 °C, and the efficiency is about 60% - 70%. Its advantage is low cost (about $800 / kW), but the dynamic response is slow (in minutes), the current density is low (<0.5 A / cm 2 ), and it cannot quickly adapt to power fluctuations.

[0004] Proton exchange membrane electrolytic water (PEM): Using a perfluorosulfonic acid proton exchange membrane (such as Nafion membrane) and noble metal catalysts (platinum, iridium), the efficiency can reach 70% - 80%, the response speed is fast (in seconds), and the current density is high (>2 A / cm 2 ). However, the cost is high (about $2000 / kW), and it depends on imported materials.

[0005] When the above technologies are applied to the wind power coupling scenario, the adaptability between wind power fluctuations and the electrolyzer is poor. The second-level fluctuations of wind power, such as a ±30% power change caused by a sudden change in wind speed, force the PEM electrolyzer to start and stop frequently or operate inefficiently. For example, when the power is lower than the minimum operating threshold of the electrolyzer, usually 10% - 20% of the rated power, the system needs to abandon wind or force a shutdown, resulting in an annual effective hydrogen production time of less than 60%.

[0006] Traditional PEM electrolyzers need to use high-iridium-loading catalysts >2 mg / cm 2 and thick membranes >150 μm to adapt to a wide power range. However, iridium is expensive, and the dissolution rate of the catalyst accelerates under dynamic conditions. When the current density fluctuates, the iridium loss rate increases by more than 3 times.

[0007] Existing solutions mostly use single energy storage (such as lithium batteries) to suppress fluctuations, but the matching of energy storage capacity and power lacks optimization, and a dynamic control strategy is not designed in combination with the material characteristics of the electrolyzer, resulting in it being difficult to balance hydrogen production efficiency and equipment life.

[0008] Existing wind power hydrogen production technologies are difficult to meet the economic and reliability requirements for large-scale green hydrogen production due to equipment-energy dynamic mismatch, high material costs, and crude control strategies. Summary of the Invention

[0009] The present invention proposes a wind power-coupled proton exchange membrane water electrolysis method for hydrogen production, which solves the problem that the existing wind power hydrogen production technology is difficult to achieve the economic and reliability requirements of large-scale production of green hydrogen due to equipment-energy dynamic mismatch, high material costs and extensive control strategies.

[0010] The technical solutions of the present invention are as follows:

[0011] The method for producing hydrogen by electrolyzing water using a wind power-coupled proton exchange membrane comprises the following steps:

[0012] S1. Acquire wind power signals in real time and divide them into low-frequency and high-frequency components according to preset frequency bands;

[0013] S2. Dynamically distributing the low-frequency component to at least one target sub-stack in a plurality of proton exchange membrane PEM electrolyzer sub-stacks for operation;

[0014] S3 absorbs the high-frequency component of the fluctuating energy through the hybrid energy storage module;

[0015] S4. Optimize the operating status of each sub-stack based on the model predictive control (MPC) algorithm, so that each sub-stack operates within the efficient range of 40%-100% of its rated power;

[0016] S5. Dynamically activate or shut down the corresponding sub-stack based on the real-time wind power and the sub-stack start-up and shutdown rules to match the wind power fluctuation input.

[0017] Furthermore, the cutoff frequency of the frequency sub-band in S1 is 0.1 Hz, where:

[0018] The low-frequency component corresponds to the fluctuation of real-time wind power change period ≥ 10 seconds;

[0019] The high-frequency components are smoothed by a hybrid energy storage system of supercapacitors and lithium batteries.

[0020] Furthermore, the multiple PEM electrolyzer sub-stacks in S2 include at least:

[0021] Low-power sub-reactor: adapted to 5%-30% rated power, using a foam metal flow field with a porosity of ≥70% and a proton exchange membrane with a thickness of ≤50μm;

[0022] High-power sub-stack: adapted to 80%-150% rated power, using titanium-based composite bipolar plates and withstanding pressure ≥2.5MPa.

[0023] Furthermore, the objective function of the model predictive control MPC algorithm described in S4 is:

[0024]

[0025] Among them, P curtailedis the curtailment power, η H2 is the hydrogen production efficiency, ΔP cycle is the start-up and shut-down loss power of the sub-reactor. The weight coefficients α, β, and γ are dynamically adjusted according to the wind power prediction error.

[0026] Furthermore, when starting up, the low-power sub-reactor is preheated to 70 ± 5 °C through electrolysis waste heat, and the start-up and shut-down delay time ≤ 10 ms.

[0027] Furthermore, when the real-time wind power exceeds the capacity of all sub-reactors, the excess energy is used for at least one of the following processing methods:

[0028] (a) Input into the ammonia synthesis device to produce synthetic ammonia;

[0029] (b) Input into the methanol synthesis device to produce liquid methanol;

[0030] (c) Store in the hydrogen storage tank and achieve the electricity-hydrogen-electricity balance through the fuel cell.

[0031] Furthermore, the temperature of each sub-reactor is independently controlled by the variable flow coolant and satisfies:

[0032] The operating temperature of the low-power sub-reactor is 70 ± 2 °C;

[0033] The operating temperature of the high-power sub-reactor ≤ 90 °C.

[0034] Furthermore, the start-up and shut-down rules include:

[0035] When the wind power is lower than 90% of the lowest power of the currently operating sub-reactor for 120 consecutive seconds, shut down the sub-reactor;

[0036] When the wind power continuously exceeds 110% of the highest power of the currently operating sub-reactor, activate the next-level sub-reactor.

[0037] Furthermore, the method also includes predicting the sub-reactor life through the digital twin platform, specifically:

[0038] Collect the operating parameters of each sub-reactor in real time, including current density, temperature, voltage fluctuation, gas humidity, and start-up and shut-down times, and the sampling frequency ≥ 10 Hz;

[0039] Perform wavelet transform noise reduction processing on the collected data, extract characteristic parameters, current density standard deviation, temperature gradient, and membrane electrode expansion coefficient;

[0040] Based on historical data and accelerated aging experiments, establish a remaining life prediction model for the sub-reactor membrane electrode. The input of the model is the characteristic parameters, and the output is the remaining life percentage and failure risk level;

[0041] The model is trained by the LSTM neural network and embedded with physical degradation equations, catalyst dissolution rate equations: where j is the current density, and E a is the activation energy;

[0042] According to the remaining life report, preferentially enable the sub-stacks with lower cumulative losses, and limit the current density fluctuation range of the high-loss sub-stacks (≤±0.3 A / cm 2 );

[0043] When the failure risk level exceeds the threshold, automatically trigger at least one of the following protection actions:

[0044] (a) Reduce the load of the sub-stack to 50% of the safe range;

[0045] (b) Start the self-healing mode and inject iridium ion repair solution into the membrane electrode;

[0046] (c) Generate a maintenance alarm and push it to the monitoring terminal.

[0047] Furthermore, the operating conditions of the ammonia synthesis device are as follows:

[0048] Pressure 10 - 15 MPa, temperature 400 - 450 °C;

[0049] Use a ruthenium-based catalyst, and the molar ratio of hydrogen to nitrogen is 3:1.

[0050] The beneficial effects of the present invention are as follows:

[0051] Through the collaborative optimization of sub-band power distribution, multi-stack dynamic operation and intelligent control strategies, the present invention realizes the comprehensive improvement of hydrogen production efficiency, equipment life and energy utilization rate in the wind power fluctuation scenario.

[0052] The present invention adopts a modular multi-stack design. Through the foam metal flow field and ultra-thin proton exchange membrane structure of the low-power sub-stacks, the ion transport resistance and polarization loss under low-load conditions are significantly reduced, enabling it to maintain an electrolysis efficiency of more than 65% in the 5% - 30% load range; at the same time, the pressure-resistant bipolar plate and gradient catalyst layer design of the high-power sub-stacks ensure stable operation under 120% - 150% overload conditions, thereby expanding the high-efficiency operation range of the system to 5% - 150% of the rated power, and completely solving the "mismatch of high-efficiency intervals" problem caused by wind power fluctuations in traditional technologies.

[0053] Aiming at the problem of accelerated material aging caused by the second-level fluctuations of wind power, the present invention constructs a "triple protection system" of dynamic response control - material structure optimization - digital twin maintenance. Through the coordination of frequency division filtering and hybrid energy storage, the fluctuation amplitude of the input power of the electrolyzer is reduced by more than 70%, significantly reducing the particle agglomeration and dissolution in the catalyst layer caused by the drastic change of current density; based on the intelligent start-stop strategy of model predictive control (MPC), the wet-dry cycle damage of the membrane electrode caused by frequent start-stop is accurately avoided, reducing the average daily start-stop times of the sub-stack by 60%;

[0054] The digital twin platform integrates the physical degradation model and real-time operation data, realizes the prediction of the membrane electrode life with millimeter-level accuracy, and triggers the self-healing protection mechanism (such as the injection of iridium ion repair solution), extending the life of key components by more than 30%.

[0055] Through the coordination of multiple energy sources and the closed-loop design of the energy chain, the maximum utilization of resources is achieved: the sub-band power distribution strategy combined with hybrid energy storage compresses the curtailment rate from more than 10% of the industry average to within 3%, and the over-generated energy is further converted into high-value chemical products through ammonia / methanol synthesis units;

[0056] The waste heat recovery and heat-electricity collaborative management module uses the waste heat of the electrolyzer for inlet water preheating and fan anti-freezing, reducing the overall system energy consumption by 15%-20%; the independent control and replaceable design of the modular sub-stack reduces the maintenance cost by 40% and supports the progressive expansion of the hydrogen production system in the wind farm.

[0057] The wide power range coverage ability of the multi-stack system enables continuous hydrogen production in low-wind-speed wind farms and intermittent power supply areas; the digital twin-driven predictive maintenance system significantly reduces the operation and maintenance difficulty and the risk of equipment failure shutdown in remote areas; the modular architecture supports the flexible deployment of systems from hundreds of kilowatts to hundreds of megawatts, providing key technical support for the large-scale replacement of fossil energy by green hydrogen; by improving the efficiency of wind power hydrogen production and the equipment life, the present invention reduces the carbon dioxide equivalent emissions per ton of green hydrogen production by 1.2-1.5 tons. Combined with low-carbon conversion paths such as ammonia / methanol synthesis, it can accelerate the decarbonization process of high-energy-consuming industries such as steel and chemical industries, with significant ecological value. Detailed implementation mode

[0058] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of the present invention.

[0059] The present invention is mainly divided into the following parts:

[0060] Frequency - band power distribution and multi - stack dynamic operation:

[0061] The wind power is divided into frequency bands at 0.1 Hz. The low - frequency component drives the multi - stack electrolyzer, and the high - frequency component is absorbed by the hybrid energy storage of supercapacitors / lithium batteries.

[0062] Design low - power sub - stacks (5% - 30% load), medium - power sub - stacks (30% - 80% load), and high - power sub - stacks (80% - 150% load), with different structures for each sub - stack (e.g., the low - power sub - stack uses a porous foam metal flow field).

[0063] Model Predictive Control (MPC) optimization:

[0064] The objective function minimizes the weighted sum of curtailment power, the reciprocal of hydrogen production efficiency, and start - stop losses. The formula is:

[0065]

[0066] Dynamically adjust the weight coefficients α, β, γ to give priority to ensuring the hydrogen production efficiency during low - wind - speed periods.

[0067] Thermal - electrical collaborative management and digital twin life prediction:

[0068] Independently control the temperature of each sub - stack (70 ± 2 °C for low - power stacks, ≤ 90 °C for high - power stacks), and use waste heat to pre - heat the electrolyzed water;

[0069] Through the digital twin platform, real - time predict the life of the membrane electrode and trigger self - healing protection (such as injecting repair fluid and limiting the current density).

[0070] Example 1

[0071] A method for wind - power - coupled proton exchange membrane electrolysis water hydrogen production, comprising the following steps:

[0072] S1. Obtain the wind - power signal in real - time, and divide the wind - power signal into a low - frequency component and a high - frequency component according to a preset frequency band;

[0073] S2. Dynamically distribute the low - frequency component to at least one target sub - stack among multiple proton exchange membrane PEM electrolyzer sub - stacks for operation;

[0074] S3. Absorb the fluctuating energy of the high - frequency component through the hybrid energy storage module;

[0075] S4. Optimize the operating state of each sub - stack based on the Model Predictive Control MPC algorithm, so that each sub - stack operates within the high - efficiency range of 40% - 100% of its rated power;

[0076] S5. Dynamically activate or deactivate the corresponding sub - stack according to the real - time wind power and the start - stop rules of the sub - stack to match the fluctuating wind - power input.

[0077] The cut-off frequency of the sub-bands in S1 is 0.1 Hz, where:

[0078] The low-frequency component corresponds to the fluctuations with a real-time wind power change period ≥ 10 seconds;

[0079] The high-frequency component is smoothed by a hybrid energy storage system of supercapacitors and lithium batteries.

[0080] The multiple PEM electrolyzer sub-stacks in S2 include at least:

[0081] Low-power sub-stack: adapted to 5%-30% of the rated power, using a foam metal flow field with a porosity ≥ 70% and a proton exchange membrane with a thickness ≤ 50 μm;

[0082] High-power sub-stack: adapted to 80%-150% of the rated power, using a titanium-based composite bipolar plate and with a pressure tolerance ≥ 2.5 MPa.

[0083] The start-stop rules include:

[0084] When the wind power is continuously lower than 90% of the lowest power of the currently operating sub-stack for 120 seconds, turn off this sub-stack;

[0085] When the wind power continuously exceeds 110% of the highest power of the currently operating sub-stack, activate the next-level sub-stack.

[0086] In this embodiment, the specific implementation plan is as follows:

[0087] Hardware configuration:

[0088] Wind power input: 20 MW wind turbine, power fluctuation range 2 - 30 MW (simulating the IEC 61400-12-1 standard wind condition);

[0089] Electrolyzer sub-stack, low-power sub-stack (2 MW):

[0090] Flow field structure: 3D printed titanium foam flow field (porosity 80%, pore diameter 0.5 mm);

[0091] Membrane electrode: ultra-thin composite membrane (Nafion XL-50, thickness 50 μm), iridium loading 1.2 mg / cm 2 ;

[0092] Minimum operating power: 0.1 MW (5% of the rated power).

[0093] High-power sub-stack (8 MW):

[0094] Bipolar plate: titanium / graphene composite material (compressive strength 3 MPa);

[0095] Membrane electrode: Nafion 117 membrane (thickness 175 μm), iridium loading 2.5 mg / cm2 。

[0096] Energy storage system:

[0097] Supercapacitor bank: 4MW / 0.5MWh (response time < 10ms);

[0098] Lithium battery bank: 10MW / 10MWh (charge-discharge efficiency 95%).

[0099] Frequency division and power distribution:

[0100] The frequency division module uses a second-order Butterworth filter with a cut-off frequency of 0.1Hz:

[0101] Low-frequency components (< 0.1Hz, period ≥ 10 seconds) are distributed to the electrolyzer sub-stack;

[0102] High-frequency components (> 0.1Hz) are absorbed by the supercapacitor in real time.

[0103] Dynamic start-stop rule: When the wind power is continuously lower than 90% of the lowest power of the currently operating sub-stack for 120 seconds, turn off the sub-stack;

[0104] Example: If the wind power suddenly drops from 15MW to 1.5MW (lower than the 5% threshold of the low-power sub-stack 0.1MW × 90%), turn off all sub-stacks and switch to energy storage power supply.

[0105] The objective function of the model predictive control MPC algorithm in S4 is:

[0106]

[0107] Where, P curtailed is the curtailed wind power, η H2 is the hydrogen production efficiency, ΔP cycle is the start-stop loss power of the sub-stack, and the weight coefficients α, β, γ are dynamically adjusted according to the wind power prediction error.

[0108] Implementation of the MPC algorithm:

[0109] Objective function parameters: Prediction horizon T = 1 hour, control horizon Δt = 10 minutes;

[0110] Initial weights: α = 0.5 (curtailed wind penalty), β = 0.4 (efficiency optimization), γ = 0.1 (start-stop loss);

[0111] Real-time optimization result:

[0112] When the wind power is 12MW, start 1 high-power sub-stack (8MW) + 1 low-power sub-stack (4MW), and the remaining power is stored in the lithium battery.

[0113] Example 2

[0114] Different from Embodiment 1, the cut-off frequency of the frequency band in S1 is 0.1 Hz, where:

[0115] The low-frequency component corresponds to the fluctuation with a real-time wind power change period ≥ 10 seconds;

[0116] The high-frequency component is smoothed by a hybrid energy storage system of a supercapacitor and a lithium battery.

[0117] The low-power sub-stack is preheated to 70 ± 5 °C by electrolytic waste heat during startup, and the start-stop delay time ≤ 10 ms.

[0118] In this embodiment, the low-power sub-stack has an anti-polarization design. The gas diffusion layer (GDL) of the titanium foam flow field is coated with a carbon nanotube (CNT) coating to reduce the concentration polarization at a low current density (0.1 A / cm 2 );

[0119] During startup preheating, the electrolytic waste heat (80 °C) is used to increase the conductivity of the influent water to 1.5 μS / cm, and the startup time is shortened to 2 minutes.

[0120] It also includes predicting the life of the sub-stack through a digital twin platform, specifically:

[0121] The operating parameters of each sub-stack are collected in real time, including current density, temperature, voltage fluctuation, gas humidity, and start-stop times, and the sampling frequency ≥ 10 Hz;

[0122] The collected data is processed by wavelet transform for noise reduction, and characteristic parameters are extracted, including the standard deviation of current density, temperature gradient, and membrane electrode expansion coefficient;

[0123] Based on historical data and accelerated aging experiments, a remaining life prediction model for the membrane electrode of the sub-stack is established. The input of the model is the characteristic parameters, and the output is the remaining life percentage and the failure risk level;

[0124] The model is trained using an LSTM neural network and embedded with physical degradation equations, the catalyst dissolution rate equation: where j is the current density, and E a is the activation energy;

[0125] According to the remaining life report, the sub-stack with a lower cumulative loss is preferentially enabled, and the current density fluctuation range of the high-loss sub-stack is restricted (≤ ± 0.3 A / cm 2 );

[0126] When the failure risk level exceeds the threshold, at least one of the following protection actions is automatically triggered:

[0127] (a) Reduce the load of the sub-stack to 50% of the safe range;

[0128] (b) Start the self-healing mode and inject the iridium ion-containing repair solution into the membrane electrode;

[0129] (c) Generate a maintenance alarm and push it to the monitoring terminal.

[0130] In this embodiment, the digital twin life prediction is specifically as follows:

[0131] Data acquisition: The sampling frequency of the current density is 20 Hz, and the sampling frequency of the temperature is 1 Hz; the expansion coefficient of the membrane electrode is monitored in real time by a laser displacement sensor (accuracy ±1 μm).

[0132] Model training:

[0133] Input features: Standard deviation of current density, temperature gradient (ΔT / Δt), rate of change of expansion coefficient;

[0134] Output: Remaining life percentage (error ±3%), failure risk level (low / medium / high);

[0135] Protection action: When the failure risk level is "high", automatically inject the repair solution (0.05 mol / L iridium nitrate solution, flow rate 10 mL / min) and limit the current density ≤1 A / cm 2 .

[0136] Example 3

[0137] Different from Example 2, when the real-time wind power exceeds the capacity of all sub-stacks, the excess energy is used for at least one of the following processing methods:

[0138] (a) Input into the ammonia synthesis device to produce synthetic ammonia;

[0139] (b) Input into the methanol synthesis device to produce liquid methanol;

[0140] (c) Store it in the hydrogen storage tank and achieve the electricity-hydrogen-electricity balance through the fuel cell.

[0141] The temperature of each sub-stack is independently controlled by a variable flow coolant and satisfies:

[0142] The operating temperature of the low-power sub-stack is 70 ±2 °C;

[0143] The operating temperature of the high-power sub-stack ≤90 °C.

[0144] In this embodiment, the thermal management is implemented as follows:

[0145] Low-power sub-stack: The coolant flow rate is 20 L / min, the inlet temperature is 25 °C, and the outlet temperature is 70 °C;

[0146] The waste heat is used to preheat the electrolyzed water, and the inlet water temperature rises from 25 °C to 50 °C, and the energy consumption is reduced by 8%.

[0147] High-power sub-reactor: coolant flow rate is 50 L / min, inlet temperature is 30 °C, and outlet temperature ≤ 90 °C.

[0148] The operating conditions of the ammonia synthesis unit are as follows:

[0149] Pressure is 10 - 15 MPa, temperature is 400 - 450 °C;

[0150] A ruthenium-based catalyst is used, and the molar ratio of hydrogen to nitrogen is 3:1.

[0151] Excess energy synthesis of ammonia:

[0152] Operating conditions: pressure is 12 MPa, temperature is 420 °C, ruthenium-based catalyst (loading amount 10%);

[0153] The molar ratio of hydrogen to nitrogen is 3:1, and the purity of nitrogen in the air separation unit is ≥ 99.95%;

[0154] Annual output of synthetic ammonia is 80,000 tons, and the hydrogen utilization rate > 95%.

[0155] All the parameters of the examples are verified through laboratory tests. For example, the 50-μm film thickness of the low-power sub-reactor can withstand 5000 start-stop cycles (film expansion rate < 5%), meeting the industrialization requirements.

[0156] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for producing hydrogen by wind power-coupled proton exchange membrane electrolysis of water, characterized in that, It includes the following steps: S1. Obtain the wind power signal in real time, and divide the wind power signal into a low-frequency component and a high-frequency component according to a preset frequency band; S2. Dynamically allocate the low-frequency component to at least one target sub-stack among multiple proton exchange membrane PEM electrolyzer sub-stacks for operation; S3. Absorb the fluctuating energy of the high-frequency component through a hybrid energy storage module; S4. Optimize the operating state of each sub-stack based on the model predictive control MPC algorithm, so that each sub-stack operates within the efficient range of 40%-100% of its rated power; S5. Dynamically activate or deactivate the corresponding sub-stack according to the real-time wind power and the start-stop rules of the sub-stack to match the wind power fluctuation input.

2. The method for producing hydrogen by wind power-coupled proton exchange membrane electrolysis of water according to claim 1, wherein The cut-off frequency of the frequency band in S1 is 0.1Hz, where: The low-frequency component corresponds to the fluctuation with a real-time wind power change period ≥ 10 seconds; The high-frequency component is smoothed by a hybrid energy storage system of a super capacitor and a lithium battery.

3. The method for producing hydrogen by wind power-coupled proton exchange membrane electrolysis of water according to claim 1, characterized in that, The multiple PEM electrolyzer sub-stacks in S2 at least include: Low-power sub-stack: adapted to 5%-30% of the rated power, using a foam metal flow field with a porosity ≥ 70% and a proton exchange membrane with a thickness ≤ 50μm; High-power sub-stack: adapted to 80%-150% of the rated power, using a titanium-based composite bipolar plate and withstanding a pressure ≥ 2.5MPa.

4. The method for producing hydrogen by wind power coupled proton exchange membrane electrolysis of water according to claim 1, characterized in that, The objective function of the model predictive control MPC algorithm in S4 is: Among them, P curtailed is the curtailed wind power, η H2 is the hydrogen production efficiency, ΔP cycle is the start-up and shutdown loss power of the sub-reactor, and the weight coefficients α, β, and γ are dynamically adjusted according to the wind power prediction error.

5. The method for producing hydrogen by wind power coupled proton exchange membrane electrolysis of water according to claim 3, characterized in that, The low-power sub-stack is preheated to 70±5°C by electrolysis waste heat during startup, and the start-stop delay time ≤ 10ms.

6. The method for producing hydrogen by wind power-coupled proton exchange membrane electrolysis of water according to claim 1, wherein, When the real-time wind power exceeds the capacity of all sub-stacks, the excess energy is used for at least one of the following treatment methods: (a) Input into an ammonia synthesis device to produce synthetic ammonia; (b) Input into a methanol synthesis device to produce liquid methanol; (c) Store in a hydrogen storage tank and achieve the electro-hydrogen-electricity balance through a fuel cell.

7. The method for producing hydrogen by wind power-coupled proton exchange membrane electrolysis of water according to claim 1, wherein The temperature of each sub-stack is independently controlled by a variable flow coolant, and satisfies: The operating temperature of the low-power sub-stack is 70±2°C; The operating temperature of the high-power sub-stack ≤ 90°C.

8. The method for producing hydrogen by wind power coupled proton exchange membrane electrolysis water according to claim 1, wherein The start-stop rules include: When the wind power is lower than 90% of the lowest power of the currently operating sub-stack for 120 consecutive seconds, turn off the sub-stack; When the wind power continuously exceeds 110% of the highest power of the currently operating sub-stack, activate the next-level sub-stack.

9. The method for producing hydrogen by wind power-coupled proton exchange membrane electrolysis of water according to claim 1, wherein The method also includes predicting the life of the sub-stack through a digital twin platform, specifically: Collect the operating parameters of each sub-stack in real time, including current density, temperature, voltage fluctuation, gas humidity, and start-stop times, with a sampling frequency ≥ 10Hz; Perform wavelet transform noise reduction processing on the collected data, extract characteristic parameters, the standard deviation of current density, temperature gradient, and membrane electrode expansion coefficient; Based on historical data and accelerated aging experiments, establish a remaining life prediction model for the membrane electrode of the sub-stack, with the characteristic parameters as the input of the model and the remaining life percentage and failure risk level as the output; The model is trained using an LSTM neural network and embeds a physical degradation equation, the catalyst dissolution rate equation: where j is the current density, E a is the activation energy; According to the remaining life report, preferentially enable the sub-reactor with lower cumulative loss, and limit the current density fluctuation range of the high-loss sub-reactor (≤±0.3A / cm 2 ); When the failure risk level exceeds the threshold, automatically trigger at least one of the following protection actions: (a) Reduce the load of the sub-stack to 50% of the safe range; (b) Start the self-healing mode and inject an iridium ion repair solution into the membrane electrode; (c) Generate a maintenance alarm and push it to the monitoring terminal.

10. The method for producing hydrogen by wind power-coupled proton exchange membrane electrolysis of water according to claim 6, wherein The operating conditions of the ammonia synthesis device are: Pressure 10-15MPa, temperature 400-450°C; A ruthenium-based catalyst is used, and the molar ratio of hydrogen to nitrogen is 3:1.

Images