Electrolytic tank cluster cooperative control method adaptive to green electricity fluctuation

Abstract

The invention discloses an adaptive green electricity fluctuation electrolytic bath cluster cooperative control method, which comprises two or more water electrolysis hydrogen production electrolytic baths and a centralized control system for controlling each water electrolysis hydrogen production electrolytic bath, the centralized control system comprises centralized control over starting, stopping and operation load adjustment of all the water electrolysis hydrogen production electrolytic cells. And the centralized control system performs synchronous and collaborative dynamic load adaptive control on the plurality of water electrolysis hydrogen production electrolytic cells, and performs centralized control of synchronous starting, equal load distribution and collaborative adjustment on each water electrolysis hydrogen production electrolytic cell based on real-time change of the power generation amount. And the electrolytic cell is dynamically matched with a green electric power supply in a 30%-100% safe load interval. Synchronous starting, equal load distribution and cooperative adjustment of the multiple electrolytic cells are achieved through the centralized control system, and the load is strictly controlled within the 30%-100% safety interval.

Description

Technical Field

[0001] This invention relates to the field of green electricity hydrogen production technology, specifically to a collaborative control method for an electrolyzer cluster that adapts to green electricity fluctuations. Background Technology

[0002] Unlike traditional power sources, green energy sources such as photovoltaic / wind power are highly volatile, which seriously affects the dynamic adaptability and operational reliability of electrolytic hydrogen production systems. In scenarios with weak sunlight or wind, hydrogen production systems are required to operate under low load conditions to ensure stable operation. This is because in alkaline water electrolysis hydrogen production systems, the purity of the gas increases with increasing current density. However, at lower current densities, the oxygen content produced is lower, which can cause the hydrogen content in the oxygen produced by alkaline water hydrogen production systems operating under low load to exceed the required technical safety limit of 1.5 vol.%.

[0003] In the field of hydrogen production from wind and solar power, the operation and control of multiple electrolyzers in existing technologies still have significant limitations. Traditional systems often adopt a "start-stop-one-cell" mode, such as the centralized control system for hydrogen production from wind and solar power disclosed in prior art application number CN202311533581.0; that is, starting or stopping the electrolyzers one by one according to changes in power generation, such as only operating a single electrolyzer when power generation is low, and gradually putting the rest of the equipment into operation as power generation increases. This control method leads to significant differences in the operating load of each electrolyzer, which in turn causes an imbalance in gas-liquid flow rate, resulting in fluctuations in the pressure difference across the diaphragm inside the electrolyzer. Since the gas production process of alkaline electrolyzers requires extremely high pressure stability, pressure imbalance directly causes hydrogen and oxygen gases to cross each other on both sides of the diaphragm, causing the concentration of hydrogen in oxygen and oxygen in hydrogen to exceed the safety limit of 1.5 vol.%, which not only affects the purity of the gas but also brings serious safety hazards.

[0004] Meanwhile, existing systems are insufficiently adapted to the volatility of green power sources. The load regulation range of conventional alkaline electrolyzers is limited to 30% to 110%. When wind and solar power generation is below 30% of the load, the electrolyzer cannot operate stably due to excessive hydrogen in the oxygen content; while above 110% load, DC energy consumption rises sharply, leading to a significant decrease in energy efficiency. Furthermore, multiple electrolyzers typically share a single gas-liquid separation system. When some electrolyzers start up, stop, or adjust their load, the coupling interference of parameters such as pressure and flow rate within the system intensifies, further reducing the dynamic response capability of the hydrogen production system.

[0005] For the reasons mentioned above, it is necessary to propose an adaptive green electricity fluctuation electrolyzer cluster collaborative control method to solve the above problems. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the existing technology and provide a collaborative control method for electrolyzer clusters that adapts to green electricity fluctuations.

[0007] To achieve the above objectives, the technical solution of the present invention is as follows: An adaptive green electricity fluctuation electrolyzer cluster collaborative control method includes two or more water electrolysis hydrogen production electrolyzers, and a centralized control system for controlling each water electrolysis hydrogen production electrolyzer. The centralized control system includes centralized control of the start-up, shutdown and operating load adjustment of each water electrolysis hydrogen production electrolyzer. Furthermore, the centralized control system performs synchronous and coordinated dynamic load adaptation control on multiple water electrolysis hydrogen production electrolyzers. Based on the real-time changes in power generation, it implements centralized control for synchronous start-up, equal load distribution, and coordinated adjustment of each water electrolysis hydrogen production electrolyzer, enabling the electrolyzers to dynamically match with green power sources within the 30%-100% safe load range.

[0008] Furthermore, the conditions for synchronous startup include that when the power generation reaches a preset startup threshold, all water electrolysis hydrogen production electrolyzers within the control system start simultaneously with equal loads. The formula for calculating the amount of electricity generated when the power generation reaches the activation threshold is: Power generation = Full load power of a single electrolytic cell × Number of electrolytic cells × Preset value of electrolytic cell start-up load (30%).

[0009] Furthermore, the equal load distribution includes simultaneously increasing or decreasing the load of all operating water electrolysis hydrogen production electrolyzers in the same proportion when power generation increases or decreases, so as to always maintain the load consistency of each operating water electrolysis hydrogen production electrolyzer.

[0010] Furthermore, the centralized control system includes a decentralized electrolyzer cluster control method based on a swarm collaboration mechanism, which includes setting an independent edge control module on each of the water electrolysis hydrogen production electrolyzers to form a swarm individual, and several water electrolysis hydrogen production electrolyzers to form an electrolyzer cluster, and also includes a centralized monitoring center and a local communication network.

[0011] Furthermore, the edge control module of each electrolytic cell collects its own operating parameters in real time and broadcasts these parameters to adjacent electrolytic cells via a local communication network. It then compares the parameters based on the broadcasts and acts according to the following interaction rules: The load balancing rule is that when the load of a certain water electrolysis hydrogen production electrolyzer is higher or lower than the average load of the adjacent electrolyzers by a preset range, the load will be automatically reduced or increased to reduce the load deviation of all electrolyzers in the cluster. The disturbance suppression rule states that when the "disturbance coefficient" of a certain electrolyzer exceeds the threshold, the adjacent electrolyzers will synchronously fine-tune the load (±1%) to counteract the impact of local disturbances on the diaphragm pressure difference by coordinating the change in gas-liquid flow rate. Start-up and shutdown coordination rules: When the power generation reaches the start-up threshold, any one electrolyzer sends a synchronous start-up signal, and all electrolyzers respond and start up with a preset start-up load; when the power generation is lower than the shutdown threshold, any one electrolyzer that meets the shutdown conditions sends a synchronous shutdown signal, and the cluster shuts down in coordination.

[0012] Furthermore, it also includes a sensing parameter system based on a multi-dimensional sensing network that collects core parameters in real time from the green electricity side, the environment side, and the system side. Based on the sensing parameters, it predicts and analyzes changes in power generation in advance, and uses the predicted changes in power generation as a basis to optimize and adjust the control logic, so that the entire system operates under the best stable conditions.

[0013] Furthermore, the multi-dimensional sensing network includes a radiance sensor installed on the green electricity side to sense real-time solar radiation intensity, an inverter status monitor to record fluctuations in output power on the green electricity side, and a temperature monitoring system to collect data on the temperature distribution of the photovoltaic array. This includes weather stations set up on the environmental side to collect data on wind speed, cloud movement speed, and atmospheric transparency to predict future trends in sunlight. This includes a power controller installed on the system side to monitor the real-time total power generation and the frequency and amplitude of power fluctuations.

[0014] Furthermore, the optimized control logic includes dynamic determination logic for the start threshold, dynamic determination logic for the stop threshold, and dynamic adjustment strategy for the threshold range.

[0015] Furthermore, the dynamic determination logic for the start-up threshold is as follows: when the power generation first reaches the start-up threshold, the start-up command is not executed immediately, but the following determination process is first entered: S11: Short-term prediction verification based on the prediction model formed by the multi-dimensional sensing network to predict the trend of power generation changes in the future period. If the power generation continues to exceed the activation threshold and the irradiance shows an upward trend, proceed to the next step; otherwise, delay activation. S12: Analyze the fluctuation characteristics of power generation by using the recent power generation data recorded by the multi-dimensional sensing network, and calculate the power fluctuation coefficient ΔP, where ΔP = maximum fluctuation value / average value; if the power fluctuation coefficient ΔP is lower than the preset value, it indicates that the power generation is stable and meets the start-up conditions; if ΔP is higher than the preset value, the buffer judgment mechanism is activated. S13: After the above steps, if it is determined that the real-time power generation is stably higher than the start-up threshold, the short-term prediction verification continues to meet the standard for a period of time, and the power generation fluctuation coefficient is lower than the preset value, and the three conditions are met at the same time, the power controller issues a synchronous start-up command to start all electrolytic cells with the same load.

[0016] Furthermore, the dynamic determination logic for the shutdown threshold is as follows: when the power generation drops to near the shutdown threshold, the following determination process is triggered to avoid accidental shutdown: S21: The weather station set up on the environmental side determines the trend, senses the direction and speed of cloud movement, and determines whether the current decrease in sunlight is a temporary shading or a continuous attenuation. If it is a temporary shading, it enters the power buffering stage; if it is a continuous attenuation, it executes shutdown preparation. S22: The power buffering mechanism includes activating the supercapacitor energy storage module built into the system. When the real-time power generation is lower than the shutdown threshold and the decrease is small, the energy storage module supplements the difference in power to maintain the low-load operation of the electrolyzer and continuously monitor the recovery trend of wind and solar power generation. S23: The synchronous shutdown command is triggered only when the following conditions are met; 1. Real-time power generation is below the shutdown threshold for a continuous period of time; 2. The energy storage module has been deeply discharged; 3. It is predicted that there is no possibility of the price rebounding to the trigger threshold in the near future; Furthermore, the shutdown process employs a gradient load reduction mode to avoid gas-liquid disturbances caused by sudden load changes.

[0017] Furthermore, the dynamic adjustment strategy for the threshold range includes dynamically adjusting the start / stop threshold range based on seasonal and climatic characteristics, which includes the following adjustment methods: S31: During periods of high volatility, the start threshold will be adjusted upwards and the stop threshold will be adjusted downwards to expand the buffer zone; S32: During stable periods, restore the original threshold and improve energy utilization. S33: During the hot phase of the system, if the electrolytic cell is still within the operating temperature range, the prediction time for the start-up determination will be reduced to decrease the energy loss during hot start-up.

[0018] The advantages and beneficial effects of this invention are as follows: 1. This invention achieves synchronous start-up, equal load distribution, and coordinated adjustment of multiple electrolyzers through a centralized control system, strictly controlling the load within the safe range of 30%-100%. It solves the problems of gas-liquid disturbance and hydrogen-oxygen cross-contamination caused by load differences in the traditional "start-stop-one-cell" mode, ensuring that the hydrogen content in the oxygen of the electrolyzer does not exceed the safe limit of 1.5 vol.%; it achieves dynamic matching between wind and solar power generation and hydrogen production load, improves the utilization rate of green electricity, and avoids the safety risks of low-load operation and the problem of surging energy consumption under high load.

[0019] 2. Through decentralized control via swarm collaboration, local information sharing and autonomous adjustment are achieved through edge control modules, while retaining the security constraints of centralized monitoring. This further improves load balancing accuracy (deviation ≤1%), eliminates the risk of single-point failure in centralized control, and increases response speed to within 50ms, making it more adaptable to short-term fluctuations in green power sources. Furthermore, by using disturbance suppression rules to counteract local gas-liquid flow velocity disturbances, the system strengthens hydrogen-oxygen cross-contamination prevention and control, improving system reliability by over 50%.

[0020] 3. Enhance the system with a multi-dimensional sensing network (green energy side, environment side, system side) and power generation prediction to achieve dynamic determination and adaptive adjustment of start-up / shutdown thresholds. Through sensing and prediction, reduce frequent start-ups and shutdowns caused by fluctuations in green power supply (ineffective start-ups and shutdowns reduced by 80%). The supercapacitor buffering mechanism avoids outages caused by short-term power fluctuations. Dynamic thresholds adapt to seasonal and climatic characteristics, ensuring a safe load range of 30%-100% while improving green energy utilization efficiency by 15%-20%, thus strengthening the system's adaptability to the strong fluctuations in green power supply. Attached Figure Description

[0021] Figure 1 This is a logic control diagram of Embodiment 1 of the present invention; Figure 2 This is the logic control diagram of Embodiment 2 of the present invention; Figure 3 This is the logic control diagram of Embodiment 3 of the present invention. Detailed Implementation

[0022] The specific embodiments of the present invention will be further described below with reference to examples. These examples are only used to more clearly illustrate the technical solutions of the present invention and should not be construed as limiting the scope of protection of the present invention.

[0023] Example 1: An adaptive green electricity fluctuation-based electrolyzer cluster collaborative control method is proposed. This scheme addresses the strong fluctuation characteristics of green electricity sources and the problems of hydrogen-oxygen crosstalk in alkaline water electrolysis hydrogen production systems under low load and energy consumption surges under high load. It employs a "synchronous and collaborative dynamic load adaptation control" strategy. Through a centralized control system, it achieves synchronized startup, load equalization, and coordinated regulation of multiple electrolyzers, ensuring that the electrolyzers always operate within a safe load range of 30%-100%, while adapting to dynamic changes in wind and solar power generation. In existing technologies, the load regulation range of conventional alkaline electrolyzers is limited to 30%-110%. Below 30% load, the hydrogen content in oxygen easily exceeds the safe limit of 1.5 vol.%, and above 110% load, DC energy consumption increases significantly. The traditional "one-by-one start-up and shutdown" mode causes gas-liquid flow velocity disturbances due to load differences in the electrolyzers, leading to hydrogen-oxygen crosstalk.

[0024] This implementation involves controlling two or more water electrolysis hydrogen production electrolyzers. The specific number can be matched according to the installed capacity of the green power supply. A centralized control system for controlling each water electrolysis hydrogen production electrolyzer includes a power controller, which is independently connected to each electrolyzer and has start / stop control, load regulation, and status monitoring functions.

[0025] The centralized control system includes centralized control of the start-up, shutdown, and operating load adjustment of each water electrolysis hydrogen production electrolyzer; and the centralized control system performs synchronous and coordinated dynamic load adaptation control of multiple water electrolysis hydrogen production electrolyzers, such as... Figure 1 As shown, the system implements centralized control based on real-time changes in power generation to achieve synchronous startup, equal load distribution, and coordinated adjustment of each water electrolysis hydrogen production electrolyzer, enabling dynamic matching between the electrolyzers and green power sources within a safe load range of 30%-100%. Synchronous startup is achieved only when wind and solar power generation reaches the startup threshold; in this case, the centralized control system triggers a synchronous startup command, and all electrolyzers start simultaneously with equal loads. Specifically, the conditions for synchronous startup include that when the power generation reaches a preset startup threshold, all water electrolysis hydrogen production electrolyzers within the control system start up simultaneously with equal loads. The formula for calculating the amount of electricity generated when the power generation reaches the activation threshold is as follows (Activation Threshold Calculation): Power generation (start-up threshold) = Full load power of a single electrolyzer × Number of electrolyzers × 30% (preset start-up load value).

[0026] Ensure that the load of each electrolytic cell is not less than 30% during startup to meet the requirements for safe operation.

[0027] It is understandable that the operating current of the electrolytic cell varies from manufacturer to manufacturer, and the optimal starting load also differs. Therefore, the starting load can be adjusted as needed. This embodiment takes a starting load of 30% as an example.

[0028] In actual control, during startup, the centralized control system synchronously adjusts the supporting equipment (such as rectifiers and heaters) to ensure that parameters such as electrolyte temperature and pressure quickly reach the reaction conditions, thus avoiding startup delays caused by local parameter imbalances.

[0029] Furthermore, the equal load distribution includes simultaneously increasing or decreasing the load of all operating water electrolysis hydrogen production electrolyzers in the same proportion when power generation increases or decreases, so as to always maintain the load consistency of each operating water electrolysis hydrogen production electrolyzer.

[0030] When power generation increases: the load of each water electrolysis hydrogen production electrolyzer is increased synchronously from 30% load (e.g., 30%→50%→80%→100%) until the total power matches the power generation; when power generation decreases: the load is reduced synchronously from the current load (e.g., 80%→50%→30%), and if it continues to decrease to below the start-up threshold, it is shut down synchronously.

[0031] The regulation logic is that the centralized control system calculates the target load of a single electrolyzer in real time based on the wind and solar power generation, according to the formula: total power generation ÷ number of electrolyzers. Then, it issues the same load regulation command to all electrolyzers to ensure that the load deviation is ≤1%. For example, when the total power generation corresponds to 50% of the target load of a single electrolyzer, the actual load of all electrolyzers is within the range of 49%-51%.

[0032] It is understandable that the above adjustment process inevitably involves the adjustment of PID parameters on the system side. For example, PID control parameters such as pressure and temperature on the system side are adjusted proportionally with the load (such as the pressure setpoint increasing synchronously when the load increases), and the interlock protection threshold (such as the overpressure alarm value) is also dynamically adapted to the current load. Taking a hydrogen production system consisting of four electrolyzers as an example, the full-load power of a single electrolyzer is 100kW (i.e., 100% load corresponds to 100kW). The specific operation process is as follows: 1. Start-up phase, start-up threshold calculation: 100kW (single unit full load) × 4 units × 30% = 120kW; when the wind and solar power generation reaches a stable 120kW, the centralized control system issues a start-up command, and the 4 electrolyzers start up simultaneously at 30% load (30kW / unit), with a total power of 120kW, matching the power generation; at this time, the current density of each electrolyzer is the same, the gas-liquid flow rate is consistent, the pressure difference across the diaphragm is balanced, and the purity of hydrogen and oxygen meets the safety standard (hydrogen in oxygen ≤ 1.5 vol.%).

[0033] 2. During the power generation increase phase, if the wind and solar power generation reaches 200kW, the centralized control system calculates the target load per unit: 200kW ÷ 4 units = 50kW / unit. All electrolyzers simultaneously increase their load from 30% to 50%. If the power generation continues to rise to 400kW, the target load per unit is 100kW / unit. All electrolyzers simultaneously reach 100% load (full load operation), and the total power is perfectly matched with the power generation. During this process, the pressure and PID temperature parameters of the supporting gas-liquid separation system increase proportionally with the load to ensure system stability.

[0034] 3. During the power generation decline phase, if wind and solar power generation drops to 150kW, the target load for a single unit is 37.5kW / unit, and all electrolyzers simultaneously reduce their load from 50% to 37.5%. If power generation drops to 110kW (below the 120kW start-up threshold), the centralized control system issues a shutdown command, and all four electrolyzers stop operating simultaneously to prevent single or partial electrolyzers from operating below 30% load. During the shutdown process, the system simultaneously shuts down the heaters and rectifiers to ensure a smooth transition of internal parameters within the electrolyzers.

[0035] Understandably, a smaller electrolytic cell system can be set up to handle power output below the start-up threshold. For example, if the wind and solar power generation is 100kW and the start-up threshold of 120kW has not been reached, the system can be switched to the smaller system and the electrolytic cells in the smaller system can be used for production. When the system exceeds the start-up threshold, it can be switched to the four parallel electrolytic cell system for production.

[0036] Example 2: This embodiment, based on the "synchronous collaborative dynamic load adaptation" in Embodiment 1, introduces decentralized control logic of swarm collaboration. To address the single-point failure risk and response delay that may exist in traditional centralized control, it achieves autonomous adjustment of the electrolytic cell cluster through individual local interaction and global collaborative optimization. At the same time, it retains the core of safety monitoring, adapts to the strong fluctuation characteristics of green power supply, and ensures stable operation of electrolytic cells within the safe load range (30%-100%).

[0037] This embodiment of the decentralized electrolyzer cluster control method includes setting up independent edge control modules on each of the water electrolysis hydrogen production electrolyzers to form a swarm of individual cells. Several water electrolysis hydrogen production electrolyzers form an electrolyzer cluster. This embodiment is applicable to two or more water electrolysis hydrogen production electrolyzers. Specifically, taking four electrolyzers as an example, each with a full-load power of 100kW, each is equipped with an edge control module. The edge control module has a built-in microprocessor, sensor interface, and communication unit, acting as a "swarm individual" with data acquisition and autonomous decision-making capabilities. It also includes a centralized monitoring center and a local communication network. The centralized monitoring center does not directly participate in load regulation; in this embodiment, it is only responsible for monitoring safety thresholds, such as hydrogen and oxygen purity ≤1.5 vol.%, load ≤100%, and global status recording, which is the system's safety baseline. The local communication network uses industrial wireless technology (such as LoRa) to build neighborhood communication links. Each electrolyzer establishes a real-time connection with 3-4 neighboring electrolyzers, with a communication latency ≤50ms, ensuring status sharing among the "swarm individuals".

[0038] This embodiment takes "distributed collaborative adjustment + centralized security constraints" as its core. The edge control module is responsible for dynamic load balancing, and the centralized monitoring center is responsible for security backup. The two are linked through the communication network, which not only retains the core of "synchronous start-up and equal load" of the first embodiment, but also improves response speed and anti-interference through the swarm mechanism.

[0039] Specifically, such as Figure 2 As shown, the edge control module of each electrolyzer collects its own operating parameters in real time and broadcasts these parameters to adjacent electrolyzers via a local communication network. The parameters are then compared based on the broadcasts. The parameters collected by each electrolyzer's edge control module include: operating load (e.g., 30% or 50% load); gas-liquid flow rate; pressure difference across the diaphragm (reflecting the degree of disturbance); and hydrogen / oxygen purity (accuracy ±0.1 vol.%). These parameters are encoded as "load deviation value" (the difference from the average load of the neighborhood) and "disturbance coefficient" (flow rate fluctuation amplitude) via the local communication network and broadcast to adjacent electrolyzers, simulating a "swarm dance" information transmission method. The load balancing rule stipulates that if the load of a water electrolysis hydrogen production electrolyzer is higher or lower than the average load of adjacent electrolyzers by a preset range, the load will be automatically reduced or increased to minimize the load deviation of all electrolyzers in the cluster. If the load of a certain electrolyzer deviates from the average load of the three adjacent electrolyzers by more than 5% (the preset range is adjustable), the following adjustments will be made automatically: Positive deviation (higher load): Reduce load by 2%; Negative deviation (lower load): Increase load by 2%; Until the load deviation of all electrolytic cells in the cluster is ≤1%, ensure “equal load distribution”.

[0040] The disturbance suppression rule states that when the "disturbance coefficient" of a certain electrolyzer exceeds the threshold, the adjacent electrolyzers will synchronously fine-tune their loads (±1%) to counteract the impact of local disturbances on the diaphragm pressure difference by coordinating the change in gas-liquid flow rate. Specifically, when the "disturbance coefficient" of a certain electrolyzer (such as flow rate fluctuation > 10%) exceeds the threshold, it is determined to be a local gas-liquid disturbance. The adjacent electrolyzers will synchronously fine-tune their loads (±1%) to balance the pressure difference across the diaphragm by coordinating the change in gas-liquid flow rate, thus preventing hydrogen-oxygen crosstalk.

[0041] Start-up and shutdown coordination rules: When the power generation reaches the start-up threshold, any one electrolyzer sends a synchronous start-up signal, and all electrolyzers respond and start up with a preset start-up load; when the power generation is lower than the shutdown threshold, any one electrolyzer that meets the shutdown conditions sends a synchronous shutdown signal, and the cluster shuts down in coordination.

[0042] The centralized monitoring center receives the status of each electrolytic cell in real time and intervenes forcibly when the following situations occur: When the hydrogen in oxygen of any electrolytic cell > 1.4 vol.% (close to the 1.5 vol.% limit): Broadcast a "warning signal" to the cluster, and all electrolytic cells reduce the load by 5% synchronously according to the rules. When the load exceeds 100% or is lower than 30%: Send a direct instruction to correct it to ensure that the safe load range is not exceeded. Emergency failure (such as oxygen in hydrogen exceeding the standard): Forcefully cut off the power supply of all electrolytic cells and terminate the distributed regulation. This embodiment can realize the linkage of the supporting system: The edge control module sends instructions to the PID controller, heater, etc. synchronously, and makes them adjust according to the load change. For example, when the load increases, the PID pressure set value increases proportionally. And it is also adapted to the green power supply: The edge control module receives the total power signal of the green power supply in real time and converts it into a "target load reference". For example, when the total power is 200 kW, the target load of 4 units is 50%, which is used as a reference for neighborhood regulation to ensure that the total power of the cluster matches the power generation.

[0043] Specifically, taking a cluster composed of 4 electrolytic cells (single unit full load 100 kW) as an example, the specific operation process is as follows: Startup stage: Startup threshold: 100 kW × 4 units × 30% = 120 kW; When the wind and solar power generation stably reaches 120 kW, an electrolytic cell (becoming the "queen bee" through neighborhood voting) broadcasts a startup signal, and the edge control modules of the four electrolytic cells respond and start at a 30% load (30 kW / unit) simultaneously; After startup, each edge control module immediately shares the load information through the local network. After the first adjustment, the load deviation is controlled within ±0.5%, the gas-liquid flow rate is consistent, and the diaphragm pressure difference is balanced.

[0044] During the power generation increase stage, such as from 120 kW to 200 kW: The total power generation of 200 kW corresponds to a single unit target load of 50%, and the "queen bee" broadcasts the reference load signal; Assume that after the initial adjustment, the load of electrolytic cell A is 51%, and the average loads of adjacent B, C, and D are 49%. The deviation of A is +2% (not exceeding 5%) and remains stable; If the load of A rises to 56% and the deviation is +7% (exceeding the preset range), then A automatically reduces by 2% to 54%, and B, C, and D synchronously increase by 1% to 50%. Finally, the cluster load is stable at 52% - 53%, and the deviation ≤ 1%; During this period, due to local flow rate fluctuations in an electrolytic cell, the "perturbation coefficient" reaches 12% (exceeding the 10% threshold), and the adjacent 3 units synchronously fine-tune the load (+1%) to offset the perturbation by jointly changing the flow rate, and the diaphragm pressure difference resumes balance.

[0045] During the power generation decline phase, such as from 200kW to 150kW: a total power generation of 150kW corresponds to a single unit target load of 37.5%; the load of electrolyzer B drops to 36%, the adjacent average load is 38%, the deviation is -2%, electrolyzer B automatically increases by 1% to 37%, the neighboring area synchronously decreases by 0.5%, and the final load is stable at 37%-38%; if the power generation drops to 100kW (below the shutdown threshold of 108kW), the "queen bee" broadcasts a shutdown signal, and all four electrolyzers shut down simultaneously to avoid single unit operating at low load.

[0046] Example of safety monitoring intervention: During operation, the hydrogen concentration in oxygen in electrolyzer C rises to 1.45 vol.% (close to the 1.5 vol.% limit), and the central monitoring center broadcasts an early warning signal; all electrolyzers reduce their load by 5% (e.g., from 50% to 45%) in coordination through the edge control module, and the hydrogen concentration in oxygen subsequently drops to 1.2 vol.%, restoring a safe state.

[0047] Example 3: Based on the "synchronous and collaborative dynamic load adaptation" in Example 1, this embodiment introduces a multi-dimensional sensing network and predictive control. To address the problems of frequent system start-ups and shutdowns and load adjustment lags caused by the strong fluctuations of green power sources, the "sensing-prediction-dynamic judgment" logic is used to optimize the start / stop threshold determination, ensuring that the electrolyzer operates stably within the safe load range (30%-100%), while improving the utilization rate of green power sources.

[0048] Specifically, this includes two or more water electrolysis hydrogen production electrolyzers; this embodiment uses four as an example. The centralized control system includes a power controller responsible for starting, stopping, and regulating the load of the electrolyzers. This embodiment adds a "multi-dimensional sensing network" to receive and process sensing data. The multi-dimensional sensing network includes a cluster of sensing devices covering the green electricity side, the environmental side, and the system side, such as... Figure 3 As shown, it specifically includes: Green electricity side: Irradiance sensor, sampling frequency 1Hz, used to monitor solar radiation intensity; inverter status monitor, used to record output power fluctuations, accuracy ±0.5%; temperature monitoring system, to collect photovoltaic array temperature distribution, resolution 0.5℃.

[0049] On the environmental side: the weather station can collect data on wind speed (0-30 m / s), cloud movement speed (±0.2 m / s), and atmospheric transparency to predict changes in sunlight.

[0050] On the system side: The power controller has a built-in monitoring module that monitors the total power generation in real time with a sampling period of 100ms; it also calculates the frequency and amplitude of power fluctuations.

[0051] This embodiment takes "sensing parameter acquisition → power generation prediction → dynamic threshold determination → coordinated load adjustment" as the core process. Based on "synchronous start-up and equal load" in Embodiment 1, it realizes "intelligent" threshold determination through a multi-dimensional sensing network, avoiding the limitations of traditional fixed thresholds.

[0052] I. The multi-dimensional perception and prediction mechanism implemented in this paper is as follows: 1. Sensing parameter system After preprocessing (Kalman filtering to remove noise), the core parameters collected by each sensing device are converted into three types of feature data: Real-time status data: current irradiance, photovoltaic output power, electrolytic cell load, ambient temperature and humidity, etc., for immediate judgment; Fluctuation characteristic data: standard deviation σ of power fluctuation over the past 30 seconds and slope of irradiance change, used to analyze stability; Predictive input data includes cloud movement trajectories, atmospheric transparency, and historical power generation for the same period, which are used to train the predictive model.

[0053] 2. Power generation prediction model This second embodiment adopts the "predictive analysis" approach, constructs an LSTM (Long Short-Term Memory) prediction model, inputs multi-dimensional sensing data, and outputs short-term (3 minutes) and medium-term (15 minutes) power generation change trends, with a prediction accuracy of ±8%, which is better than the ±15% of the traditional model.

[0054] II. Dynamic Threshold Determination Logic: 1. Dynamic determination of the start-up threshold: The calculation formula for the start-up threshold remains "single unit full load power × number of units × preset start-up load value", for example: four units × 100kW × 30% = 120kW. However, the determination process adds three levels of verification to avoid false compliance. Specifically: S11: Short-term forecast verification. When the real-time power generation first reaches 120kW, it is not started immediately. Instead, the power generation trend for the next 3 minutes is output through the forecast model. If the forecast result meets the conditions of "continuously above 120kW" and "irradiance slope > 0" (showing an upward trend), then proceed to the next step; if the forecast is that it will be below 120kW within 3 minutes (such as when clouds are about to block the light), then the start-up is delayed.

[0055] S12: Fluctuation characteristic analysis, calculate the power fluctuation coefficient ΔP = (maximum fluctuation value / average value) × 100% in the past 10 seconds. If ΔP < 10% (preset value), it indicates that the green power supply output is stable; if ΔP ≥ 10% (such as power jump caused by instantaneous gusts), the buffer mechanism is activated - wait for 2 minutes and then re-detect, during which the sensing data is continuously updated.

[0056] S13: Start-up conditions: The centralized control system will issue a synchronous start-up command and start all electrolytic cells at 30% load only when the three conditions are met simultaneously: "real-time power generation is stable at ≥120kW for 30 consecutive seconds", "predicted to meet the standard for 3 minutes", and "ΔP < 10%".

[0057] III. Dynamic determination of shutdown threshold: The shutdown threshold is set to 90% of the startup threshold (e.g., 120kW × 90% = 108kW). When the power generation drops to around 108kW, a level-three judgment is triggered to avoid accidental shutdown. S21: Trend determination, using environmental meteorological station data to determine the type of decreased sunlight: Brief obstruction, such as an isolated cloud passing by, with an expected obstruction time of less than 1 minute: enter the power buffering phase; Continuous attenuation, such as at sunset or when there is extensive cloud cover, with an estimated attenuation time > 5 minutes: Execute shutdown preparation.

[0058] S22: Power buffering mechanism. The supercapacitor energy storage module is activated, with a capacity configured at 5% of rated power × 1 minute, i.e., four units × 100kW × 5% × 1 minute = 20kWh. When the real-time power generation is below 108kW but the decrease is less than 15% (e.g., dropping to 100kW), the energy storage module supplements the difference by 8kW, maintaining the electrolyzer at 30% load operation, while simultaneously monitoring the recovery trend of green power supply.

[0059] S23: Shutdown execution conditions: A synchronous shutdown command will only be issued when all three conditions are met: real-time power generation is less than 108kW for 60 consecutive seconds, the energy storage module discharge depth is greater than 80%, and it is predicted that there is no possibility of the power generation rising to 120kW in the next 5 minutes. The shutdown process adopts a "gradual load reduction" method, reducing the load from 30% to 0 within 5 seconds to avoid gas-liquid disturbances.

[0060] IV. Dynamic adjustment strategy for threshold range: Dynamically optimize startup / shutdown thresholds based on season and system status to improve adaptability: S31: During periods of high volatility, such as cloudy summer days: the start-up threshold is increased by 5% (120kW→126kW), and the shutdown threshold is decreased by 5% (108kW→102kW), expanding the buffer zone and reducing invalid start-ups and shutdowns; S32: During stable periods, such as sunny days in winter: restore the original threshold (120kW / 108kW) to improve the utilization rate of green electricity; S33: System hot phase, such as within 30 minutes after shutdown: If the electrolytic cell temperature is still >60℃ (operating temperature range), the startup prediction time is reduced from 3 minutes to 1 minute, reducing the energy consumption of hot start-up. The energy consumption of hot start-up is 20% lower than that of cold start-up.

[0061] Specifically, taking a system consisting of four electrolytic cells (each with a full load of 100kW) as an example, the specific operation process is as follows: (1) Start-up phase Start-up threshold: 100kW × 4 units × 30% = 120kW; Scenario: Cloudy weather in summer (high fluctuation period), threshold dynamically adjusted to 126kW; process: The real-time power generation rose to 126kW, but the prediction model showed that it would drop to 120kW (<126kW) after 3 minutes due to cloud cover, triggering a delayed start-up. Two minutes later, the cloud cover moved away, and the power generation rebounded to 130kW. It is predicted to remain at ≥126kW for three minutes, with an irradiance slope of 0.5kW / min (upward trend). The calculation shows that ΔP = 8% (< 10%) over the last 10 seconds, which satisfies the stability condition. The centralized control system issues a command, and the four electrolytic cells start up simultaneously at 30% load (30kW / cell), and the supporting systems enter the working state in sync.

[0062] (2) Stable operation phase With wind and solar power generation increased to 200kW, the centralized control system adjusts the load of each electrolyzer to 50% (50kW / unit) in accordance with the principle of "equal distribution". Real-time monitoring by multi-dimensional sensing network: stable irradiance, power fluctuation ΔP=3%, hydrogen and oxygen purity in electrolyzer ≤1.2 vol.%, system maintains current load.

[0063] (3) Phase of declining power generation Shutdown threshold (dynamically adjusted): 126kW × 90% × 95% = 109kW (5% lower during periods of high volatility); Scenario: Sudden cloud cover reduces power generation to 105kW (<109kW). process: The environmental weather station determined it to be a "brief obstruction" (cloud movement speed 10m / s, obstruction time expected 40 seconds). Start the supercapacitor energy storage module to supplement the 4kW difference (105kW+4kW=109kW) and maintain the electrolyzer at 30% load; Forty seconds later, the cloud cover moved away, the power generation rebounded to 130kW, the energy storage module stopped discharging, and the system load simultaneously rose to 32.5% (130kW ÷ 4 units = 32.5kW / unit).

[0064] (4) Shutdown phase Scenario: During sunset, power generation continues to drop to 100kW (<109kW). process: The weather station determined it to be "continuous decay" (solar altitude angle <10°, predicted to drop to 80kW within 5 minutes); The energy storage module has a discharge depth of 90% (exceeding the 80% threshold). The centralized control system issued a shutdown command, and the four electrolytic cells gradually decreased from 30% to 0 load within 5 seconds, with the supporting system shutting down simultaneously.

[0065] This embodiment reduces the number of invalid start-stop cycles by more than 80% through multi-dimensional perception and prediction, solving the problem of system instability caused by the fluctuation of green power supply, and demonstrating strong anti-fluctuation capability. The combination of dynamic thresholds and buffering mechanisms ensures that the electrolyzer always operates within the 30%-100% range, maintaining a 100% hydrogen and oxygen purity qualification rate and providing appropriate safety redundancy. Hot start optimization and stable-period threshold recovery effectively improve energy utilization, increasing the utilization rate of green power supply by 15%-20%. Dynamic adjustment of the threshold range adapts to different seasons and climates, enhancing versatility and improving the environmental adaptability of the equipment.

[0066] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for adaptive green electricity fluctuation-based collaborative control of an electrolyzer cluster, characterized in that, It includes two or more water electrolysis hydrogen production electrolyzers, and a centralized control system for controlling each water electrolysis hydrogen production electrolyzer. The centralized control system includes centralized control of the start-up, shutdown and operating load adjustment of each water electrolysis hydrogen production electrolyzer. Furthermore, the centralized control system performs synchronous and coordinated dynamic load adaptation control on multiple water electrolysis hydrogen production electrolyzers. Based on the real-time changes in power generation, it implements centralized control for synchronous start-up, equal load distribution, and coordinated adjustment of each water electrolysis hydrogen production electrolyzer, enabling the electrolyzers to dynamically match with green power sources within a safe load range.

2. The adaptive green electricity fluctuation electrolyzer cluster collaborative control method according to claim 1, characterized in that, The conditions for synchronous startup include: when the power generation reaches a preset startup threshold, all water electrolysis hydrogen production electrolyzers in the control system start up simultaneously with equal loads. The formula for calculating the green electricity generation corresponding to the preset start-up threshold is: Power generation = Full load power of a single electrolytic cell × Number of electrolytic cells × Preset value of electrolytic cell start-up load.

3. The adaptive green electricity fluctuation electrolyzer cluster collaborative control method according to claim 1, characterized in that, The equal load distribution includes simultaneously increasing or decreasing the load of all operating water electrolysis hydrogen production electrolyzers in the same proportion when power generation increases or decreases, so as to always maintain the same load of each operating water electrolysis hydrogen production electrolyzer.

4. The adaptive green electricity fluctuation electrolyzer cluster collaborative control method according to claim 1, characterized in that, The centralized control system includes a decentralized electrolyzer cluster control method based on a swarm collaboration mechanism. This method involves setting an independent edge control module on each of the water electrolysis hydrogen production electrolyzers to form a swarm of individual cells. Several water electrolysis hydrogen production electrolyzers form an electrolyzer cluster. The system also includes a centralized monitoring center and a local communication network.

5. The adaptive green electricity fluctuation electrolyzer cluster collaborative control method according to claim 4, characterized in that, Each electrolytic cell's edge control module collects its own operating parameters in real time and broadcasts these parameters to adjacent electrolytic cells via a local communication network. The modules then compare the parameters based on the broadcasts and act according to the following interaction rules: The load balancing rule is that when the load of a certain water electrolysis hydrogen production electrolyzer is higher or lower than the average load of the adjacent electrolyzers by a preset range, the load will be automatically reduced or increased to reduce the load deviation of all electrolyzers in the cluster. Start-up and shutdown coordination rules: When the power generation reaches the start-up threshold, any one electrolyzer sends a synchronous start-up signal, and all electrolyzers respond and start up with the preset start-up load; When the power level falls below the shutdown threshold, any electrolytic cell that meets the shutdown conditions will send a synchronous shutdown signal, and the entire cluster will shut down in a coordinated manner.

6. The adaptive green electricity fluctuation electrolyzer cluster collaborative control method according to claim 1, characterized in that, By using sensing technology to predict and analyze changes in power generation in advance, the entire system is adjusted according to the predicted and analyzed changes in power generation to ensure that it operates under the best safe and energy-saving conditions.

7. The adaptive green electricity fluctuation electrolyzer cluster collaborative control method according to claim 6, characterized in that, The sensing technology includes a multi-dimensional sensing network formed by real-time collection of core parameters from the green electricity side, the environment side, and the system side.

8. The adaptive green electricity fluctuation electrolyzer cluster collaborative control method according to claim 7, characterized in that, The system is adjusted based on the predicted changes in power generation, including dynamic determination logic for start-up thresholds, dynamic determination logic for shutdown thresholds, and dynamic adjustment strategies for threshold ranges.

9. The adaptive green electricity fluctuation electrolyzer cluster collaborative control method according to claim 8, characterized in that, The steps for determining the threshold in a multi-dimensional sensing network include: sensing parameter acquisition, power generation prediction, dynamic threshold determination, and coordinated load adjustment.

Images