ALK-PEM series-parallel system control method and device based on wind and light fluctuation

Abstract

The invention provides an ALK-PEM series-parallel system control method and device based on wind and light fluctuation. Determining a to-be-distributed total power target value of the target electrolytic cell series-parallel system according to the external fluctuating power supply input power; constructing a corresponding particle coding sequence according to the number of the electrolytic cell units; using a preset particle swarm optimization model to perform iterative optimization processing on the particle coding sequence according to the operation physical boundary parameters, and determining target particle position information; determining a start-stop state combination and a power distribution proportion of each electrolytic cell unit in the target electrolytic cell series-parallel system by using a preset power distribution rule according to the target particle position information; and according to the start-stop state combination, the power distribution proportion and the to-be-distributed total power target value, the operation power value of each electrolytic cell unit in the target electrolytic cell series-parallel system is determined. Therefore, the target value of the total power to be distributed is accurately determined according to the input power of the external fluctuating power supply, and the technical defect of power response lag is overcome.

Description

Technical Field

[0001] This manual pertains to the field of renewable energy hydrogen production technology, and particularly relates to the control method and device for ALK-PEM hybrid systems based on wind and solar power fluctuations. Background Technology

[0002] As the hydrogen energy industry evolves towards large-scale production, hybrid systems composed of heterogeneous electrolyzers have become the core path for absorbing fluctuating new energy sources and simultaneously achieving large-scale hydrogen production. However, existing control strategies lack a coordinated scheduling mechanism to address the differences in dynamic characteristics of heterogeneous electrolyzers during power allocation, resulting in technical defects such as limited adjustment range and delayed power response when facing real-time power fluctuations.

[0003] There is currently no effective solution to the above problems. Summary of the Invention

[0004] This manual provides a control method and device for an ALK-PEM hybrid system based on wind and solar power fluctuations. It solves the technical problems of existing technologies in the power distribution of heterogeneous electrolyzer hybrid systems, which suffer from limited adjustment range and lag in power response when facing real-time power fluctuations due to the lack of dynamic characteristics for coordinated scheduling.

[0005] This manual provides a control method for the ALK-PEM hybrid system based on wind and solar fluctuations, including: The system acquires the input power of an external fluctuating power source and the operating parameters of the target electrolyzer hybrid system; wherein the operating parameters include the number of each electrolyzer unit and the corresponding operating physical boundary parameters; the target electrolyzer hybrid system includes at least an alkaline electrolyzer and a proton exchange membrane electrolyzer; Based on the external fluctuating power input, determine the target total power to be allocated for the target electrolytic cell hybrid system; Construct a corresponding particle coding sequence based on the number of electrolytic cell units; Using a pre-defined particle swarm optimization model, the particle encoding sequence is iteratively optimized based on the physical boundary parameters to determine the target particle position information; wherein, the iterative optimization process includes: dynamically adjusting the search step size using a linearly decreasing inertial weight adjustment strategy, and performing optimization processing using a multi-objective function as the evaluation criterion; Using a preset power allocation rule, the start-stop state combination and power allocation ratio of each electrolytic cell unit in the target electrolytic cell hybrid system are determined based on the target particle position information; Based on the start / stop state combination, the power allocation ratio, and the target total power to be allocated, the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system is determined.

[0006] In one embodiment, the step of using a preset particle swarm optimization model to iteratively optimize the particle encoding sequence based on the operational physical boundary parameters to determine the target particle position information includes: The initial position of each particle in the particle swarm is determined based on the particle encoding sequence and the physical boundary parameters. Based on the current position of each particle and the multi-objective function, determine the current generation value of each particle; Based on the current generation value, determine the individual target position of each particle and the group target position of the particle swarm; The group center position of the particle swarm is determined based on the individual target positions of all particles. The current inertia weight parameters are determined based on the current iteration number, the preset iteration stop threshold, and the linearly decreasing inertia weight adjustment strategy. Based on the individual target position, the group target position, the group center position, and the current inertial weight parameter, the updated position of each particle is determined using the quantum behavior optimization criterion. Repeat the steps from determining the current generation value to determining the update position until the number of iterations reaches the preset iteration stop threshold, and obtain the target particle position information based on the group target position.

[0007] In one embodiment, determining the current generation value of each particle based on its current position and the multi-objective function includes: Based on the current position of each particle and the physical boundary parameters of each electrolytic cell unit, determine the operating status and power allocation value of each electrolytic cell unit. Based on the operating status, the allocated power value, and the target value of the total power to be allocated, the evaluation values ​​of each evaluation factor in the multi-objective function are determined; wherein, the evaluation factors include power tracking factor, start-stop factor, hydrogen production efficiency factor, and power quality factor; Based on the evaluation values ​​of each evaluation factor and the preset weights corresponding to each evaluation factor, the current generation value of each particle is determined by fusion calculation using the multi-objective function.

[0008] In one embodiment, determining the start-stop state combination and power allocation ratio of each electrolytic cell unit in the target electrolytic cell hybrid system based on the target particle position information using a preset power allocation rule includes: Based on the target particle position information and the preset state switching threshold, the operating state of each electrolytic cell unit is determined, and the start-stop state combination is obtained. Based on the start-stop state combination and the corresponding operating physical boundary parameters of each electrolytic cell unit, the basic operating power of each electrolytic cell unit in the start state is determined. The remaining power to be allocated is determined based on the total target power to be allocated and the sum of the basic operating powers. Based on the proportional allocation criteria in the preset power allocation rules, the numerical proportional relationship between each element in the target particle position information, and the remaining power to be allocated, the optimized incremental power corresponding to each electrolytic cell unit is determined. Based on the constraint correction criteria in the preset power allocation rules, the ramp rate limit and power operation range corresponding to each electrolytic cell unit, the sum of the basic operating power and the optimized incremental power is corrected to determine the power allocation ratio corresponding to each electrolytic cell unit.

[0009] In one embodiment, determining the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system based on the start-stop state combination, the power allocation ratio, and the target total power to be allocated includes: Based on the start / stop state combination, the target electrolytic cell unit in the start state is determined; Based on the physical boundary parameters, determine the lower limit of the operating power for each of the target electrolytic cells. The remaining total power to be allocated is determined based on the difference between the sum of the aforementioned operating power lower limits and the target value of the total power to be allocated; Using the power allocation ratio corresponding to each target electrolytic cell unit, the remaining total power to be allocated is incrementally allocated to obtain the additional power allocation value corresponding to each target electrolytic cell unit; The operating power value of each electrolytic cell unit is determined based on the start / stop state combination, the lower limit of operating power corresponding to each target electrolytic cell unit, and the additional power allocation value.

[0010] In one embodiment, determining the target total power to be allocated for the target electrolytic cell hybrid system based on the external fluctuating power input includes: Obtain the converter efficiency corresponding to the input power of the external fluctuating power supply, and the auxiliary operating power consumption of the target electrolytic cell hybrid system; Based on the converter efficiency, determine the initial power value for the input power of the external fluctuating power supply; The target value of the total power to be allocated is determined based on the difference between the preliminary power value and the power consumption of the auxiliary machine.

[0011] In one embodiment, after determining the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system based on the start-stop state combination, the power allocation ratio, and the target total power to be allocated, the method further includes: Based on the operating power value corresponding to each electrolytic cell unit, determine the modulation control parameters of the power conversion equipment corresponding to each electrolytic cell unit; The DC output current of the power conversion equipment is determined based on the modulation control parameters. The electrolytic cell unit is driven to perform electrolytic hydrogen production based on the DC output current.

[0012] In one embodiment, after obtaining the external fluctuating power input and the operating parameters of the target electrolytic cell hybrid system, the method further includes: The dynamic hydrogen price in the current hydrogen market associated with the target electrolyzer hybrid system is obtained through the communication interface.

[0013] This manual provides a control device for an ALK-PEM hybrid system based on wind and solar fluctuations, including: The data acquisition module is used to acquire the input power of the external fluctuating power source and the operating parameters of the target electrolyzer hybrid system; wherein, the operating parameters include the number of each electrolyzer unit and the corresponding operating physical boundary parameters; the target electrolyzer hybrid system includes at least an alkaline electrolyzer and a proton exchange membrane electrolyzer; The total power determination module is used to determine the target total power to be allocated for the target electrolytic cell hybrid system based on the input power of the external fluctuating power supply. The sequence determination module is used to construct a corresponding particle coding sequence based on the number of electrolytic cell units; The position information determination module is used to determine the target particle position information by using a preset particle swarm optimization model and iteratively optimizing the particle encoding sequence according to the running physical boundary parameters; wherein, the iterative optimization process includes: dynamically adjusting the search step size using a linearly decreasing inertial weight adjustment strategy, and performing optimization processing with a multi-objective function as the evaluation criterion. The start / stop allocation determination module is used to determine the start / stop state combination and power allocation ratio of each electrolytic cell unit in the target electrolytic cell hybrid system based on the target particle position information using a preset power allocation rule. The power allocation module is used to determine the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system based on the start / stop state combination, the power allocation ratio, and the target total power to be allocated.

[0014] This specification also provides an electronic device, including a processor and a memory for storing processor-executable instructions, wherein the processor, when executing the instructions, implements an ALK-PEM hybrid system control method based on wind and solar fluctuations.

[0015] This specification also provides a computer-readable storage medium storing computer instructions that, when executed, implement a control method for an ALK-PEM hybrid system based on wind and solar fluctuations.

[0016] Based on the ALK-PEM hybrid system control method based on wind and solar fluctuations provided in this specification, the input power of the external fluctuating power source and the operating parameters of the target electrolyzer hybrid system are obtained. The operating parameters include the number of each electrolyzer unit and the corresponding operating physical boundary parameters. The target electrolyzer hybrid system includes at least an alkaline electrolyzer and a proton exchange membrane electrolyzer. Based on the input power of the external fluctuating power source, the target total power to be allocated to the target electrolyzer hybrid system is determined. Based on the number of electrolyzer units, a corresponding particle coding sequence is constructed. Using a preset particle swarm optimization model, based on the operating physical boundaries... The parameters are used to iteratively optimize the particle encoding sequence to determine the target particle position information. The iterative optimization process includes: dynamically adjusting the search step size using a linearly decreasing inertial weight adjustment strategy, and performing optimization using a multi-objective function as the evaluation criterion; using a preset power allocation rule, determining the start-stop state combination and power allocation ratio of each electrolytic cell unit in the target electrolytic cell hybrid system based on the target particle position information; and determining the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system based on the start-stop state combination, the power allocation ratio, and the target total power to be allocated. In this way, by acquiring the operating physical boundary parameters of the target electrolytic cell hybrid system and using a preset particle swarm optimization model to iteratively optimize the particle encoding sequence, the search step size is dynamically adjusted through a linearly decreasing inertial weight adjustment strategy, and global space optimization is performed according to the multi-objective function evaluation criterion and the quantum behavior optimization criterion. This solves the problem of limited adjustment range in the face of real-time power fluctuations in existing technologies. By determining the target particle position information and using preset power allocation rules to determine the start-stop state combination, power allocation ratio and operating power value of each electrolytic cell unit, the accurate determination of the total power target value to be allocated based on the external fluctuating power input power is achieved, overcoming the technical defect of power response lag. Attached Figure Description

[0017] To more clearly illustrate the embodiments of this specification, the accompanying drawings used in the embodiments will be briefly introduced below. The drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating an embodiment of the ALK-PEM hybrid system control method based on wind-solar fluctuations provided in this specification. Figure 2 This is a schematic diagram of the electronic device structure provided in one embodiment of this specification; Figure 3 This is a schematic diagram of the structural composition of the ALK-PEM hybrid system control device based on wind and solar fluctuations, provided in one embodiment of this specification. Figure 4 This is a performance comparison diagram provided as an embodiment of this specification. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0020] See Figure 1 As shown in the embodiments of this specification, a control method for an ALK-PEM hybrid system based on wind and solar fluctuations is provided, wherein the method is specifically applied to the server side. In specific implementation, the method may include the following: S101: Obtain the input power of the external fluctuating power source and the operating parameters of the target electrolyzer hybrid system; wherein, the operating parameters include the number of each electrolyzer unit and the corresponding operating physical boundary parameters; the target electrolyzer hybrid system includes at least an alkaline electrolyzer and a proton exchange membrane electrolyzer; S102: Determine the target total power to be allocated for the target electrolytic cell hybrid system based on the external fluctuating power input; S103: Construct a corresponding particle coding sequence based on the number of electrolytic cell units; S104: Using a preset particle swarm optimization model, the particle encoding sequence is iteratively optimized according to the physical boundary parameters to determine the target particle position information; wherein, the iterative optimization process includes: dynamically adjusting the search step size using a linearly decreasing inertial weight adjustment strategy, and performing optimization processing with a multi-objective function as the evaluation criterion. S105: Using a preset power allocation rule, determine the start-stop state combination and power allocation ratio of each electrolytic cell unit in the target electrolytic cell hybrid system according to the target particle position information; S106: Determine the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system based on the start / stop state combination, the power allocation ratio, and the target total power to be allocated.

[0021] The aforementioned external fluctuating power input can be the real-time output power of a wind power system or a photovoltaic power system, the value of which fluctuates non-stationarily within a time scale of seconds to minutes depending on weather conditions.

[0022] The aforementioned target electrolyzer hybrid system can be a coupled hydrogen production array comprising at least one alkaline electrolyzer (ALK) unit and at least one proton exchange membrane electrolyzer (PEM) unit, which can operate in a complementary manner by utilizing the response characteristics of different electrolyzers.

[0023] The above operating parameters can be the total number of devices describing the scale of the electrolytic cell hybrid system, as well as a set of physical quantities that determine the operating safety boundary and dynamic performance indicators of each device.

[0024] The aforementioned physical boundary parameters can be the allowable power adjustment range for each electrolytic cell unit, the power ramp-up rate limit per unit time, and the depreciation cost incurred during a single equipment start-up.

[0025] The aforementioned target value for the total power to be allocated can be determined based on the real-time measured external fluctuating power input, representing the total power demand that needs to be absorbed by all electrolytic cell units in the hybrid system at the current moment.

[0026] The aforementioned particle encoding sequence can be a set of continuous numerical vectors representing the start-stop combination states of all electrolytic cell units in the hybrid system. The mapping of discrete states is achieved by thresholding the vector elements.

[0027] The aforementioned pre-defined particle swarm optimization model can be an adaptive quantum behavior particle swarm optimization (AQPSO) algorithm model that integrates quantum behavior optimization criteria and adaptive parameter adjustment strategies.

[0028] The aforementioned target particle position information can be the optimal coordinate vector determined by the algorithm after completing the optimization iteration in the multi-dimensional solution space, which enables the multi-objective function to reach a convergent state.

[0029] The aforementioned linearly decreasing inertia weight adjustment strategy can be an adjustment rule that allows the algorithm's inertia weight to increase with the number of iterations, linearly decreasing from a preset initial weight value to a preset final weight value, in order to balance the global search capability and local exploitation capability in the optimization process.

[0030] The aforementioned multi-objective function can be a comprehensive cost assessment model that takes into account multiple dimensions such as system power tracking deviation, system-level start-stop conversion frequency, total hydrogen production efficiency, and power quality level.

[0031] The aforementioned preset power allocation rule can be a power scheduling principle that follows the principle of "prioritizing minimum power to ensure equipment operation, and optimizing the allocation of remaining power according to the proportion of particle positions" while also taking into account ramping constraints.

[0032] The above start / stop state combination can be a logical sequence obtained after state decoding based on the target particle position, reflecting whether each electrolytic cell unit in the hybrid system is in the open or closed state at the current moment.

[0033] The aforementioned power allocation ratio can be a weighted allocation coefficient for secondary optimization of the remaining power to be allocated among the various operating electrolytic cell units after removing the basic operating power of each unit.

[0034] The aforementioned operating power values ​​can be the final determined power command values ​​actually executed by each electrolytic cell unit, and their sum matches the total power target value to be allocated.

[0035] In some embodiments, obtaining the external fluctuating power input and the operating parameters of the target electrolyzer hybrid system may specifically include: According to the external power monitoring command, the real-time power sequence of the external fluctuating power source within a preset sampling period is obtained; Based on the configuration information of the target electrolyzer hybrid system, the number of electrolyzer units and the equipment type distribution corresponding to each electrolyzer unit are determined; wherein, the equipment type distribution includes at least alkaline electrolyzer units and proton exchange membrane electrolyzer units; Based on the equipment type distribution, determine the operating physical boundary parameters corresponding to each electrolytic cell unit; The physical boundary parameters of operation include the rated power range, power ramp rate limit, and single system-level start-up and shutdown cost for each electrolytic cell unit.

[0036] Specifically, the first step is to perform data acquisition and processing of the external fluctuating power source. Specifically, the controller, based on preset external power monitoring commands, retrieves real-time output data from wind or solar power generation via a communication interface and organizes it into a real-time power sequence within a preset sampling period. For example, in a 24-hour simulation scenario, the sampling resolution can be set to 1 minute to obtain an input power curve that reflects the high-frequency fluctuation characteristics of renewable energy, serving as the raw input for subsequent optimization calculations.

[0037] Subsequently, based on the pre-stored target electrolyzer hybrid configuration information, the total number of electrolyzer units currently participating in scheduling and their specific equipment type distribution are identified. In this embodiment, the hybrid configuration consists of a heterogeneous electrolyzer array, explicitly including proton exchange membrane (PEM) electrolyzer units with fast response capabilities and alkaline (ALK) electrolyzer units with economies of scale. The specific configuration can be further subdivided into one PEM electrolyzer, one small alkaline electrolyzer (ALKs), and one medium-sized alkaline electrolyzer (ALKm) to construct a heterogeneous cluster with multi-level regulation capabilities.

[0038] Finally, based on the identified equipment type, the operating physical boundary parameters corresponding to each electrolytic cell unit are matched and determined from the parameter database. These parameters include, but are not limited to: the rated power range of each unit (e.g., PEM 0.05-1.5MW, ALKm 1.05-5.25MW), power ramp-up rate limits (e.g., dynamic constraints within the range of 1.0MW / min to 1.5MW / min), and single-secondary start-up and shutdown costs. By parameterizing these physical constraints, an accurate optimization solution space boundary is provided for the subsequent AQPSO algorithm, ensuring that the generated control commands conform to the hardware physical characteristics of the equipment.

[0039] By performing high-frequency sampling of external fluctuating power sources and refining the equipment type distribution of heterogeneous electrolytic cell units, a multi-dimensional physical boundary model was established, including rated power, ramp rate, and system-level start-up and shutdown costs. This provides a precise data foundation and physical constraints for subsequent intelligent power allocation. By parametrically decoupling the fast response characteristics of PEM with the operating boundary of ALK, the difficulties in coordinated scheduling caused by the lack of heterogeneous equipment characteristic descriptions in existing technologies are effectively solved, ensuring the operational safety and regulation accuracy of the hybrid system when absorbing real-time power fluctuations from the source.

[0040] In some embodiments, constructing a corresponding particle coding sequence based on the number of electrolytic cell units may specifically include: The sequence dimension of a single particle in the particle swarm optimization model is determined based on the number of electrolytic cell units. Based on the sequence dimension, an initial encoding vector containing multiple consecutive value elements is generated; Based on the preset mapping relationship between each electrolytic cell unit and each continuous value element in the initial encoding vector, each continuous value element is associated with the corresponding electrolytic cell unit to obtain the particle encoding sequence.

[0041] Specifically, the mathematical dimension of the search space is first determined based on the total number of identified electrolytic cell units. For example, if the target electrolytic cell hybrid includes one proton exchange membrane (PEM) electrolytic cell, one small alkaline electrolytic cell (ALKs), and one medium alkaline electrolytic cell (ALKm), then the sequence dimension of a single particle is determined to be 3. This one-to-one correspondence of dimensions ensures that the subsequent algorithm search process can cover every independent physical device in the process.

[0042] Next, based on the determined sequence dimension, an initial encoding vector is generated within a preset numerical range. In practice, this encoding vector consists of a set of continuous floating-point values, for example, randomly generated within the interval [0,1]. The purpose of using continuous value encoding instead of directly using 0 / 1 discrete encoding is to enable the optimization process to perform a smooth search in the continuous solution space using the AQPSO algorithm, thereby improving the convergence accuracy of the algorithm when facing complex fluctuating conditions.

[0043] Finally, according to the preset device mapping rules, each element in the initial encoding vector is bound to a physical device. In this example, the first element of the vector is designated as the state and power weighting factor of the PEM electrolyzer, the second element is designated as the factor of ALKs, and the third element is designated as the factor of ALKm. Through this mapping relationship, the algorithm can directly transform the abstract numerical evolution into a precise description of the operating state of a specific physical device, ultimately completing the construction of the particle encoding sequence.

[0044] By mapping discrete electrolytic cell start-up and shutdown states and continuous power allocation weights to a continuous numerical space, the AQPSO algorithm can utilize the optimization characteristics of continuous functions to perform global optimization in a multi-dimensional solution space. This effectively solves the problems of low search efficiency and easy trapping in extreme values ​​in traditional discrete coding when dealing with heterogeneous coupled control. This construction method, which dynamically determines the dimension based on the number of devices, not only improves the compatibility of the control scheme with hybrid connections of different scales, but also ensures that each electrolytic cell unit can obtain accurate scheduling instructions during the global optimization process through high-dimensional parameter expression, ultimately forming a particle coding sequence with high robustness and high descriptive accuracy.

[0045] In some embodiments, a mapping process from the algorithm encoding space to the physical device state is first performed. Specifically, each particle in the particle swarm uses a continuous value encoding method, where the values ​​of each dimension of its position vector represent the candidate control weights of the corresponding electrolytic cell unit. A state decoding process is performed based on a preset state switching threshold (e.g., set to 0.5): when the position value of a particle is greater than the state switching threshold, the corresponding electrolytic cell unit is determined to be in the on state (i.e., state value 1); otherwise, the unit is determined to be in the off state (i.e., state value 0).

[0046] Furthermore, the state decoding process follows the following logical principles:

[0047] in, This is the running status value. This represents the particle's position value.

[0048] This mechanism of mapping continuous values ​​to discrete states enables the AQPSO algorithm to perform global optimization within a smooth solution space, effectively avoiding the problem of low search efficiency caused by directly using discrete encoding.

[0049] After determining the start / stop state combination, the intelligent power distributor executes specific instruction allocation according to the principle of "basic support and incremental optimization". First, based on the identified start-up state units, it retrieves their corresponding operating physical boundary parameters to determine the basic operating power of each electrolytic cell unit in the start-up state. This step reflects the allocation principle of "minimum power priority", that is, prioritizing the allocation of the minimum power allowance to each operating electrolytic cell to maintain its electrochemical reaction stability and thermal balance, ensuring that the equipment's lifespan is not damaged due to underpower operation.

[0050] Subsequently, a refined allocation of the remaining power is performed. The remaining power to be allocated is determined by subtracting the target total power from the sum of the base operating power of each activated unit. Next, based on the proportional relationships between the values ​​in each dimension of the particle position information, a power allocation weight coefficient is determined for each target electrolytic cell unit, and the remaining power to be allocated is incrementally distributed according to this coefficient. For example, if a particle has a larger value in a certain dimension, it means that the unit is given a higher power absorption weight under the current operating conditions, thus obtaining more optimized incremental power.

[0051] Finally, physical safety verification and correction are performed. The base operating power and optimized incremental power of each unit are summed, and constraint corrections are strictly performed according to the corresponding power operating range of the unit (such as the maximum rated power limit) and the power ramp-up rate limit. If the calculated power value exceeds the physical adjustment boundary at the sampling time, the power value is forcibly locked at the allowed extreme point. Through this allocation logic that combines particle swarm optimization weights with real-time hardware physical constraints, the power allocation ratio of each electrolytic cell unit is finally determined, ensuring that the control commands meet the global optimal objective while possessing strict physical feasibility.

[0052] By introducing a decoding mechanism of "continuous encoding-threshold discrimination" and a two-level power allocation criterion of "basic support + proportional increment", a deep decoupling between the algorithm search flexibility and the operational constraints of heterogeneous equipment is achieved. This solves the technical defects of existing technologies, such as poor optimization accuracy and delayed command response, when dealing with discrete start-stop states and continuous power allocation and coordinated scheduling. By prioritizing the minimum operating power and allocating the remaining power in combination with particle weights, the load ratio can be dynamically adjusted according to the real-time status of each electrolytic cell unit. This not only significantly improves the ability to absorb external fluctuating power, but also effectively suppresses the impact of frequent and large power fluctuations on the lifespan of the electrolytic cell through strict ramp rate and operating range constraints. Thus, energy conversion efficiency is optimized while ensuring operational stability.

[0053] In some embodiments, an improved particle swarm optimization algorithm integrating quantum behavior optimization and adaptive parameter adjustment is introduced to construct a collaborative scheduling mechanism among heterogeneous electrolyzer units, significantly improving the overall conversion efficiency of the system for fluctuating wind and solar power. Under the precise guidance strategy of multi-objective functions, this specification achieves a significant increase in hydrogen production while effectively suppressing frequent invalid actions of equipment using system-level start-stop judgment criteria, greatly reducing hardware wear and long-term operation and maintenance costs. In addition, the algorithm has excellent dynamic response characteristics and convergence accuracy, and can match the drastic fluctuations of external input power in real time and accurately, ensuring a deep integration of steady-state performance and economic benefits of the electrolyzer hybrid system under complex operating conditions.

[0054] In some embodiments, before performing iterative optimization, the operating physical boundary parameters of each electrolytic cell unit are first corrected through a dynamic health status sensing module. Specifically, the controller acquires the cumulative runtime, start-stop count, and current reactor temperature of each electrolytic cell unit in real time, and calculates the real-time health status index (SOH) of each unit. Based on the SOH, a preset rated power range is dynamically scaled, with the following correction criteria:

[0055] in, This is the current corrected power limit. This is the upper limit of the equipment's rated power. The decay coefficient is based on health status and temperature. Through this dynamic boundary correction, the load weight of units with faster decay can be automatically reduced in the mid-to-late stage of operation, avoiding accelerated equipment aging caused by forced absorption of fluctuating power, and providing a more physically realistic solution space constraint for subsequent optimization.

[0056] Furthermore, to further improve the economics of start-up and shutdown, this embodiment introduces a multi-step time-series smoothing prediction term into the multi-objective cost function. The external fluctuating power supply prediction sequence for the next k sampling times (e.g., the next 5 minutes) is obtained through a communication gateway. When determining the start-up / shutdown state combination at the current moment, the intelligent power divider not only evaluates the power matching degree at the current moment but also uses a sliding window mechanism to calculate the average fluctuation intensity within the future sequence.

[0057] If the predicted sequence shows that the duration of the current power spike is lower than the preset "inertial start threshold", the algorithm will suppress the start command of large inertial devices (such as ALK electrolyzers) even if it calculates that the start-stop factor is dominant at the current moment, by adding a timing penalty weight. This rolling optimization mechanism, which takes into account the continuity of timing, effectively solves the problem of false start-stop caused by "nearsighted" optimization in the prior art, and ensures the smoothness of state switching and low loss during long-term operation.

[0058] Finally, the AQPSO algorithm incorporates the modified dynamic power boundary and multi-step temporal penalty term into the iterative evolution process. In each optimization round, the particle not only needs to find the position with the lowest cost at the current moment, but also needs to satisfy the dynamically changing physical envelope constraint. Through this dual-coupling optimization of space (heterogeneous devices) and time (prediction window), the final determined target particle position information can guide the hybrid grid to achieve high-quality real-time absorption of wind and solar power while ensuring the health of the equipment throughout its entire life cycle.

[0059] By introducing dynamic health status perception and multi-step timing smoothing prediction mechanisms, this approach achieves a leap from "static constraints" to "dynamic optimization throughout the entire lifecycle" in control strategies. It overcomes the technical bottlenecks of existing technologies, such as the ambiguity of safety boundaries during the later stages of heterogeneous equipment aging and the frequent false start-stops in the face of high-frequency fluctuations. Through dynamic correction of power boundaries, this solution effectively extends the electrolytic reactor's lifespan by approximately 15%, and, combined with timing rolling optimization, further reduces the number of invalid start-stops by 12%. This not only ensures the system's robustness under extreme fluctuation conditions but also achieves a dual breakthrough in energy consumption accuracy and long-term system operation and maintenance economy through in-depth analysis of the equipment's physical characteristics.

[0060] In some embodiments, the step of using a preset particle swarm optimization model to iteratively optimize the particle encoding sequence based on the operational physical boundary parameters to determine the target particle position information may further include the following: S1: Determine the initial position of each particle in the particle swarm based on the particle encoding sequence and the physical boundary parameters. S2: Determine the current generation value of each particle based on its current position and the multi-objective function; S3: Based on the current generation value, determine the individual target position of each particle and the group target position of the particle swarm; S4: Determine the group center position of the particle swarm based on the individual target positions of all particles; S5: Determine the current inertia weight parameters based on the current iteration number, the preset iteration stop threshold, and the linearly decreasing inertia weight adjustment strategy; S6: Based on the individual target position, the group target position, the group center position, and the current inertial weight parameter, the updated position of each particle is determined using the quantum behavior optimization criterion; S7: Repeat the steps from determining the current generation value to determining the update position until the number of iterations reaches the preset iteration stop threshold, and obtain the target particle position information based on the group target position.

[0061] The aforementioned preset particle swarm optimization model can be a model constructed using the Adaptive Quantum-behaved Particle Swarm Optimization (AQPSO) algorithm, which is used to perform global optimization of the solution space based on quantum probability distribution properties.

[0062] Specifically, the initial positions of each particle in the particle swarm are first determined based on the dimension of the particle encoding sequence (i.e., the number of electrolytic cell units) and the physical boundary parameters of each unit. Specifically, a set of position vectors is randomly generated within a preset search space boundary (e.g., the [0,1] interval) to ensure that the initial search points cover possible device combination schemes, laying a diverse foundation for subsequent global optimization.

[0063] In each iteration, based on the current position of each particle and combined with evaluation criteria including power tracking, start-stop loss, hydrogen production efficiency, and power quality, the current generation value of each particle is determined. Then, the current generation value of each particle is compared with the historical records: if the current value is better than the historical records, the individual target position of that particle is updated; simultaneously, the position with the best generation value performance is selected from the entire swarm and updated as the swarm's target position. Through this two-layer selection, the algorithm achieves real-time locking of "superior genes" during the search process.

[0064] To further incorporate swarm synergy, a mean-based calculation is performed based on the individual target positions of all particles to determine the swarm's center position (i.e., the average optimal position). This center position reflects the swarm's overall perception of the optimal solution distribution. Subsequently, based on the current iteration count and a preset iteration stopping threshold, a linearly decreasing inertia weight adjustment strategy is used to calculate the current inertia weight parameter. This parameter decreases as the iteration progresses, enabling the algorithm to possess strong global exploration capabilities in the early stages and precise local exploration capabilities in the later stages.

[0065] In the position evolution stage, based on the individual target position, the group target position, the group center position, and the current inertial weight parameters, the updated position of each particle is determined according to the quantum behavior optimization criterion. Under the influence of quantum behavior, particles no longer have a definite velocity, but instead perform a jump-like search in the solution space through a probability distribution function. This allows the algorithm to effectively escape local optima traps and perform optimization processing in a broader dimension.

[0066] Finally, it is determined whether the current iteration count has reached the preset iteration stopping threshold: if not, the updated position is used as the current position for the next cycle, and the steps of determining the cost value and updating the position are repeated; if the threshold has been reached, the loop is stopped, and the final target particle position information is determined based on the current group target position. Through this structured closed-loop evolution, it is possible to quickly converge to the optimal operating point of the hybrid system within a minute-level sampling period.

[0067] By introducing the swarm center position and quantum behavior optimization criteria, particles can perform nondeterministic jump search in the high-dimensional solution space, effectively solving the technical bottleneck of traditional particle swarm optimization algorithms that are prone to getting trapped in local optima when dealing with complex coupling constraints of heterogeneous electrolyzers, and significantly improving the success rate of global optimization. Combined with the inertia weight adjustment strategy that decreases linearly with the iteration process, a balance is achieved between the search step size and the dynamic response speed and steady-state convergence accuracy, ensuring that the hybrid system can quickly lock the target particle position information that takes into account energy utilization, equipment life and hydrogen production efficiency when facing drastic fluctuations in wind and solar power, thereby improving the overall operating benefits while ensuring safe operation.

[0068] In some embodiments, the intelligent control method provided in this specification operates within a hierarchical optimization architecture composed of multiple functional modules. The core of this architecture lies in achieving closed-loop, refined management of the hybrid electrolytic cells through deep coupling between software logic modules and physical constraint modules. Specifically, an adaptive parameter regulator is first activated to dynamically evolve the algorithm parameters according to the current iteration stage. This regulator uses a linearly decreasing inertia weight adjustment strategy to correct the search step size and contraction / expansion coefficients in real time during the optimization process, enabling the algorithm to automatically switch search intensity based on different wind and solar fluctuation intensities, ensuring the algorithm's robustness under complex operating conditions.

[0069] Subsequently, the particle swarm optimization engine, guided by the parameter regulator, performs a swarm intelligence optimization search. This engine constructs a multidimensional solution space corresponding to the number of electrolyzer units and uses quantum behavior optimization criteria to drive the particle swarm to perform probabilistic jump searches within this space. During the search process, the engine not only records the position information of each particle but also guides the swarm to rapidly converge towards the optimal solution interval by maintaining the swarm center position, thereby providing a sequence of candidate control parameters for heterogeneous power allocation.

[0070] In the evaluation phase, a multi-objective cost evaluator performs comprehensive performance quantification on the candidate solutions generated by the engine. This evaluator integrates evaluation criteria across four dimensions: power point tracking, production efficiency, equipment start-up and shutdown, and power quality. By calculating the comprehensive cost-effectiveness of each candidate solution within the current sampling period, the evaluator can identify a balanced solution that accurately follows external power fluctuations while also ensuring long-term operation of the electrolyzer. This multi-dimensional evaluation mechanism effectively avoids problems such as frequent equipment start-ups and shutdowns or low operating efficiency caused by single-objective optimization.

[0071] Finally, the intelligent power allocator transforms the optimized abstract particle positions into specific physical execution commands. Upon receiving the optimization results, the allocator automatically combines the operational physical boundary parameters of each electrolyzer unit (such as ramp rate and power range) to execute a dynamic power allocation strategy. This strategy includes state discrimination processing and incremental power allocation processing, ensuring that the final issued power operating value satisfies multi-objective optimization while strictly conforming to the heterogeneous physical constraints of the alkaline electrolyzer and the proton exchange membrane electrolyzer, thus completing a full control closed loop from algorithm optimization to hardware-driven operation.

[0072] By constructing a hierarchical optimization architecture consisting of an optimization engine, a power distributor, a cost evaluator, and a parameter regulator, the system achieves deep integration of adaptive algorithm evolution and physical constraints of heterogeneous equipment. This addresses the technical shortcomings of existing control strategies, such as low modularity and poor dynamic adaptability, when dealing with fluctuating power sources. Through collaborative scheduling among modules, the system can dynamically adjust the search step size based on real-time operating conditions and comprehensively weigh multiple operating indicators. This not only significantly improves the global optimization efficiency and tracking accuracy of power distribution but also effectively ensures the operational stability and equipment lifespan of the hybrid system when absorbing severe power fluctuations through the strict control of physical boundaries by the intelligent power distributor. Ultimately, this achieves the optimal balance between the overall system operating cost and hydrogen production revenue.

[0073] In some embodiments, the Adaptive Quantum Behavior Particle Swarm Optimization (AQPSO) algorithm provided in this specification introduces a linearly decreasing inertia weight mechanism to dynamically adjust the search characteristics of particles during algorithm iteration. Specifically, the control device pre-sets the initial inertia weight value, the final inertia weight value, and the maximum number of iterations for the algorithm. In this embodiment, the initial inertia weight value is set to 0.9 to give the particles strong global search momentum in the early stages of the algorithm; the final inertia weight value is set to 0.4 to enable the particles to have more refined local adjustment capabilities in the later stages of the algorithm; and the maximum number of iterations is set to 50 to ensure convergence is completed within the time window that meets the real-time control requirements.

[0074] Specifically, during the iterative evolution process, the inertia weight parameter is adjusted in real time according to the current iteration progress. The linearly decreasing inertia weight mechanism follows the following criteria:

[0075] in, For real-time inertial weight parameters, The initial inertia weight threshold, To terminate the inertia weight threshold, This represents the current iteration number. This is the threshold for the maximum number of iterations.

[0076] During the iterative optimization process, the weights are linearly decayed based on the ratio of the current iteration count to the maximum iteration count. As the number of iterations gradually increases, the current inertia weight decreases uniformly from the initial 0.9 to 0.4. This dynamic evolution process allows the algorithm to perform large-scale "prospecting" across the entire multidimensional solution space with a large step size in the early stages of optimization, effectively increasing the probability of discovering the global optimum and preventing particles from getting trapped in local solution regions too early.

[0077] As the algorithm enters the later stages of iteration, the search range of particles is gradually compressed as the inertia weight decreases linearly, thus entering a high-precision "development" phase. Lower inertia weights suppress disordered particle jumping, prompting the particle swarm to perform fine-tuning searches within the locked preferred region until the preset convergence condition is met. Through this strategy of smoothly transitioning from "wide-area search" to "fine-grained convergence" with the iterative process, the optimal target particle position information can be quickly and stably locked under the complex physical boundary constraints of heterogeneous electrolyzer hybrid systems.

[0078] By introducing an adaptive inertia weight adjustment strategy that linearly decreases from 0.9 to 0.4, real-time optimization of the dynamic characteristics of the particle optimization process is achieved. This solves the technical defects of traditional optimization algorithms in handling multivariate coupled control of electrolytic cells, such as slow convergence speed, easy trapping in extreme values, and insufficient steady-state accuracy. By balancing global exploration and local development capabilities during the iteration process, the algorithm's adaptability to fluctuating wind and solar power conditions is significantly improved. This ensures that the algorithm can quickly converge to the optimal scheduling scheme within a limited 50 iterations, thereby further improving the accuracy and operational economy of hybrid power allocation while ensuring the effectiveness of the absorption response.

[0079] In some embodiments, the method for determining the current generation value of each particle based on its current position and the multi-objective function may further include the following: S1: Determine the operating status and power allocation value of each electrolytic cell unit based on the current position of each particle and the operating physical boundary parameters of each electrolytic cell unit. S2: Based on the operating status, the allocated power value, and the target value of the total power to be allocated, determine the evaluation value of each evaluation factor in the multi-objective function; wherein, the evaluation factors include power tracking factor, start-stop factor, hydrogen production efficiency factor, and power quality factor; S3: Based on the evaluation values ​​of each evaluation factor and the preset weights corresponding to each evaluation factor, the current generation value of each particle is determined by performing a fusion calculation using the multi-objective function.

[0080] Specifically, in each iteration of the particle swarm optimization evaluation, a mapping process from the encoding space to the physical operating space is first performed. Specifically, based on the current position coordinates of each particle and a preset state switching criterion (e.g., a threshold of 0.5), the continuous particle position vectors are decoded into the operating state (on or off) of each electrolytic cell unit. Simultaneously, based on the operating physical boundary parameters of each electrolytic cell unit (e.g., the PEM power range of 0.05-1.5MW and ramp-up limits), the allocated power value for each active unit under the current candidate scheme is determined. This mapping method enables the algorithm to search for the optimal discrete equipment combination scheme within the continuous solution space.

[0081] Subsequently, based on the obtained operating status, allocated power value, and current target total power to be allocated, four evaluation factors in the multi-objective function are calculated. When calculating the power tracking factor, the absolute deviation between the sum of the allocated power of each unit and the target total power to be allocated is calculated. When calculating the start-up / shutdown factor, a level-based judgment criterion is adopted, meaning that a start-up cost is only included when the state changes from all electrolyzers shut down to at least one electrolyzer operating (e.g., from three cells shut down to one cell running), thus avoiding excessive penalties for frequent fluctuations in a single device. When calculating the hydrogen production efficiency factor, the real-time total hydrogen production is determined based on the power of each unit, and its negative value is taken to meet the cost minimization objective. When calculating the power quality factor, a quantitative evaluation is performed based on the power factor of the current equipment combination.

[0082] Finally, based on the preset weights of each component, a multi-objective function is used to perform a fusion calculation. For example, in this embodiment, a higher weight (e.g., 500.0) can be assigned to the power tracking component to ensure the accuracy of grid absorption, and a significant weight (e.g., 300.0) can be assigned to the hydrogen production efficiency component to balance economic benefits, while auxiliary weights (e.g., 5.0) are assigned to the start-up and shutdown components and the power quality component. By weighted summing of the evaluation values ​​of each evaluation factor, the current generation value of each particle is determined. This generation value serves as the sole benchmark for evaluating the "quality" of a particle, directly guiding the particle swarm towards a balance point that balances response speed, equipment lifespan, and operating efficiency, ultimately determining the optimal target particle position information.

[0083] By establishing a four-dimensional evaluation criterion covering power point tracking, stage start-stop, hydrogen production efficiency, and power quality, the performance of ALK-PEM hybrid operation has been fully quantified, solving the problem that existing technologies struggle to balance the lifespan of heterogeneous equipment with overall efficiency during power allocation. In particular, by introducing stage start-stop judgment criteria and target conversion processing for hydrogen production, the algorithm can effectively filter out minor disturbances in wind and solar power, significantly reducing frequent start-stop losses while ensuring real-time power tracking accuracy, and achieving deep synergistic optimization of energy utilization and equipment operation and maintenance costs.

[0084] In some embodiments, the step of determining the current generation value of each particle by performing a fusion calculation using the multi-objective function based on the evaluation values ​​of each evaluation factor and the preset weights corresponding to each evaluation factor may specifically include: The power tracking cost component is determined based on the evaluation value of the power tracking factor and the preset power deviation penalty weight. The start-stop loss cost component is determined based on the evaluation value of the start-stop factor and the preset start-stop loss weight. Based on the evaluation value of the hydrogen production efficiency factor and the preset production benefit weight, the efficiency gain cost component is determined. The power quality cost component is determined based on the evaluation value of the power quality factor and the preset quality constraint weights. The current generation value of each particle is determined by weighted fusion processing based on the power tracking cost component, the start-stop loss cost component, the efficiency gain cost component, and the power quality cost component.

[0085] By introducing multi-dimensional cost components and their differentiated weighted fusion processing, a unified evaluation system capable of simultaneously characterizing physical constraints, operational losses, and economic benefits is constructed. This solves the scheduling imbalance problem caused by the single evaluation dimension in existing control strategies when dealing with ALK-PEM hybrid systems. By assigning high weights to power point tracking and hydrogen production efficiency, and in conjunction with system-level start-up and shutdown cost constraints, the AQPSO algorithm can effectively suppress the ineffective actions of heterogeneous electrolyzers while ensuring the accuracy of wind and solar power integration. This significantly improves the system's energy utilization rate and equipment lifespan, thereby guiding the system to achieve global performance optimization through the final determined current generation value of each particle.

[0086] In some embodiments, the multi-objective function is used to comprehensively evaluate the following four dimensions: Power Tracking Dimension (Preset Weight: 500.0): The power tracking cost is determined based on the deviation between the external fluctuating power input and the total allocated power, and a high-weight penalty term is used to ensure the real-time accuracy of absorption.

[0087] Specifically, firstly, the real-time input power of the external fluctuating power source (such as a photovoltaic array) is acquired, and simultaneously, the total power allocation of each electrolyzer unit in the target electrolyzer hybrid system under the current candidate scheme is calculated. Then, the absolute deviation value between the two is determined to characterize the degree of power imbalance in the system at the current sampling time. Finally, this absolute deviation value is multiplied by a preset power deviation penalty weight to obtain the final power point tracking cost. Its calculation logic can be expressed as follows:

[0088] in, For power tracking cost, For external fluctuating power input, The total power absorbed by the system. The power deviation penalty weight.

[0089] Level Start-Stop Dimension (Preset Weight: 5.0): Introduces level start-stop judgment rules, only when the hybrid connection switches from a completely stopped state to an on state where at least one device is running, the single level start-up loss cost is included, in order to suppress frequent invalid actions.

[0090] Hydrogen production efficiency dimension (preset weight: 300.0): Based on the efficiency gain cost determined by the total real-time hydrogen production of each unit, the goal of maximizing hydrogen production is transformed into the goal of minimizing cost through negative correlation processing.

[0091] Furthermore, an efficiency gain cost component is introduced to quantify the production benefits of the hybrid system under the current power allocation scheme. Specifically, the system first calculates the real-time hydrogen production of each unit based on the real-time power allocation of each electrolyzer unit using a preset electrolysis efficiency model, and then sums them to obtain the total real-time hydrogen production of the system, as follows:

[0092] in, For efficiency gains, cost This represents the total hydrogen production in real time.

[0093] Power quality dimension (preset weight: 5.0): The quality constraint cost is determined based on the real-time power factor and the electrical characteristics of the heterogeneous equipment combination, in order to ensure the operational stability of the grid connection.

[0094] Specifically, in the evaluation phase of particle optimization, a comprehensive cost calculation is performed on each candidate power allocation scheme using a preset multi-objective function. First, the power tracking cost is quantified. Specifically, the absolute deviation between the current external fluctuating power input and the sum of the operating power values ​​of all operating units in the target electrolyzer hybrid system is calculated. Then, this deviation is multiplied by a preset power deviation penalty weight (e.g., 500.0) to determine the power tracking cost component. This high-weight configuration ensures that the algorithm always prioritizes "accurate absorption of fluctuating power" during the optimization process, effectively reducing wind and solar power curtailment.

[0095] Secondly, the determination of start-up and shutdown costs at the execution level. Unlike traditional start-up and shutdown counting based on individual devices, this embodiment adopts a state transition criterion at the execution level: it retrieves the global operating state at the previous sampling time and compares it with the current candidate start-up and shutdown state combinations. Only when the state changes from "all electrolyzer units are in the off state" to "at least one electrolyzer unit is in the on state" is a start-up determined, and the corresponding start-up loss cost is included. This determination method fully considers the energy consumption and preheating costs of auxiliary equipment during the overall start-up of a large-scale hydrogen production station, effectively filtering out frequent start-up and shutdown interference of individual devices caused by small power fluctuations, and protecting the hardware lifespan.

[0096] Subsequently, the efficiency gain cost and power quality cost are evaluated. Based on the electrolysis efficiency model of each electrolyzer unit at the current power, the expected total hydrogen production is calculated. To meet the requirement of the particle swarm optimization algorithm to find the minimum cost point, the negative value of the total hydrogen production is taken as the efficiency gain cost component; the higher the hydrogen production, the smaller the corresponding cost component. Simultaneously, the overall power factor is evaluated based on the current equipment combination scheme (e.g., one PEM with two ALK units), and the power quality cost component is determined accordingly to ensure that harmonics and reactive power are within a controlled range under high-power operating conditions.

[0097] Finally, based on the preset weights of each component (e.g., 500.0, 5.0, 300.0, 5.0), a weighted summation is performed on the cost components of the four dimensions to ultimately determine the current generation value of the particle. Through this multi-dimensional fusion evaluation, the algorithm can automatically find the globally optimal balance point among "tracking accuracy," "equipment wear and tear," "production revenue," and "operational quality," ensuring that the final determined target particle position information can guide the hybrid system to maximize overall benefits.

[0098] By constructing a multi-objective evaluation system consisting of power point tracking, stage start-stop, hydrogen production efficiency, and power quality, a comprehensive quantification and in-depth synergistic optimization of the performance of heterogeneous electrolyzers operating in parallel was achieved. This solved the technical problem of equipment lifespan damage or poor economic benefits caused by the single index in power allocation in existing technologies. In particular, by introducing stage start-stop judgment criteria and differentiated weight allocation, it can maintain extremely high power point tracking accuracy to absorb green electricity when facing drastic fluctuations in wind and solar power, and effectively identify and suppress meaningless frequent start-stop switching. Under the premise of ensuring grid operation quality, it significantly improves the overall hydrogen production efficiency and equipment operation safety.

[0099] In some embodiments, the method of determining the start-stop state combination and power allocation ratio of each electrolytic cell unit in the target electrolytic cell hybrid system based on the target particle position information using a preset power allocation rule may further include the following: S1: Based on the target particle position information and the preset state switching threshold, determine the operating state of each electrolytic cell unit and obtain the start-stop state combination; S2: Determine the basic operating power of each electrolytic cell unit in the on state based on the combination of start and stop states and the corresponding physical boundary parameters of each electrolytic cell unit. S3: Determine the remaining power to be allocated based on the total power target value to be allocated and the sum of the basic operating powers; S4: Based on the proportional allocation criteria in the preset power allocation rules, the numerical proportional relationship between each element in the target particle position information, and the remaining power to be allocated, determine the optimized incremental power corresponding to each electrolytic cell unit. S5: Based on the constraint correction criteria in the preset power allocation rules, the ramp rate limit and power operation range of each electrolytic cell unit, the sum of the basic operating power and the optimized incremental power is corrected to determine the power allocation ratio of each electrolytic cell unit.

[0100] Specifically, after determining the target particle position information, the first step is to perform a mapping process from the encoding space to the physical device state. Specifically, the control unit performs a judgment based on the values ​​corresponding to each dimension in the target particle position information, combined with a preset state switching threshold (e.g., 0.5): when the value of a certain dimension exceeds the state switching threshold, the corresponding electrolytic cell unit is determined to be in the on state; otherwise, it is determined to be in the off state. Through this threshold judgment rule, continuous optimization results are transformed into discrete combinations of start and stop states, achieving preliminary scheduling of the alkaline electrolytic cell (ALK) and proton exchange membrane (PEM) electrolytic cell arrays.

[0101] Subsequently, the basic power support and remaining power calculations are performed. Based on the determined start-stop state combination, the operating physical boundary parameters corresponding to each electrolytic cell unit in the start state are retrieved, and the lower limit of power operation for each unit is determined as the basic operating power. On this basis, the deviation between the target total power to be allocated and the sum of the basic operating powers of all start-up units is calculated to obtain the remaining power to be allocated. This processing step follows the principle of "basic operation priority" to ensure that all electrolytic cell equipment put into operation can operate within the preset stable electrochemical performance range.

[0102] During the incremental allocation phase, dynamic weight calculation is performed based on a preset proportional allocation criterion. Specifically, based on the element values ​​representing each electrolytic cell unit in the target particle position information, the numerical proportional relationship between each activated device is calculated and determined as the power allocation weight coefficient for each unit. Subsequently, the remaining power to be allocated is distributed to each unit according to the weight coefficient, determining the optimized incremental power for each unit. This dynamic proportional allocation method based on particle position allows the power allocation scheme to evolve in real time following the global equilibrium point obtained by the AQPSO algorithm.

[0103] Finally, the allocation results are physically verified according to preset constraint correction criteria. The initial allocation value is obtained by summing the base operating power and the optimized incremental power of each unit. This initial allocation value is then corrected based on the ramp rate limit and power operating range (e.g., the maximum rated power limit) corresponding to that unit. If the power change rate caused by the initial allocation value exceeds the ramp rate limit, it is corrected to the allowable variation boundary; if the total exceeds the operating limit, a limit truncation is performed. Through this correction process, the power allocation ratio for each electrolytic cell unit is ultimately determined, ensuring that the issued control commands are both economical and meet the physical safety constraints of the hardware.

[0104] By constructing a multi-level power allocation rule consisting of state discrimination, basic support, proportional allocation, and constraint correction, deep collaborative scheduling of the start-up and shutdown states and power output of the ALK-PEM hybrid system was achieved. This solved the problems of low adjustment accuracy and response lag caused by the lack of correlation of heterogeneous equipment characteristics when facing real-time power fluctuations in existing technologies. In particular, by introducing corrections for basic operating power guarantees and ramp-up rate limits, it ensured that control commands strictly adhered to the hardware physical boundaries of the electrolyzer unit while meeting the optimization objectives. This significantly improved the absorption capacity of wind and solar power curtailment and hydrogen production efficiency, while effectively avoiding the negative impact of drastic power fluctuations on the electrolyzer's lifespan. Thus, a balance between operational safety and operational benefits was achieved under complex fluctuation conditions.

[0105] In some embodiments, the method for determining the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system based on the start-stop state combination, the power allocation ratio, and the target total power to be allocated may further include the following: S1: Based on the start / stop state combination, determine the target electrolytic cell unit that is in the start state; S2: Determine the lower limit of the operating power for each of the target electrolytic cells based on the physical boundary parameters of operation; S3: Determine the remaining total power to be allocated based on the difference between the sum of the lower limits of the operating power and the target value of the total power to be allocated; S4: Using the power allocation ratio corresponding to each target electrolytic cell unit, perform incremental allocation calculation on the remaining total power to be allocated to obtain the additional power allocation value corresponding to each target electrolytic cell unit; S5: Determine the operating power value of each electrolytic cell unit based on the start / stop state combination, the lower limit of the operating power corresponding to each target electrolytic cell unit, and the additional power allocation value.

[0106] Specifically, after obtaining the optimal control parameters, the instruction generation process for the actual execution power of each electrolyzer unit is performed. First, the controller identifies the target electrolyzer units that are in the on state during the current sampling period according to the start-stop state combination determined in the foregoing steps. For example, if the combined state indicates that the proton exchange membrane (PEM) electrolyzer and the small alkaline electrolyzer (ALKs) are on, while the medium alkaline electrolyzer (ALKm) is off, then the PEM and ALKs are locked as the target objects for this power consumption execution.

[0107] Subsequently, the extraction of the basic operation boundary and the difference calculation are performed. According to the operation physical boundary parameters corresponding to each target electrolyzer unit, the respective operation power lower limit values are determined. Due to the different electrochemical characteristics of the alkaline electrolyzer and the PEM electrolyzer, there are significant differences in their lower limit values. For example, the ALK unit usually needs to maintain above 20% of the rated power to ensure the stability of the operation temperature and pressure. Calculate the sum of the operation power lower limit values of all the units in the on state, and perform a difference operation with the total power target value to be allocated, so as to determine the remaining total power to be allocated.

[0108] In the incremental allocation link, the incremental allocation calculation is performed on the remaining total power to be allocated by using the power allocation ratios corresponding to each target electrolyzer unit determined by particle optimization before. Specifically, the remaining total power to be allocated is used as the multiplicand, and multiplied by the allocation weight coefficient corresponding to each unit to obtain the additional power allocation value corresponding to each target electrolyzer unit. This two-level allocation criterion of "lower limit guarantee + proportional increment" ensures that the allocation result not only meets the minimum operation physical constraints of the equipment, but also realizes the optimal ratio of the remaining energy of the wind-solar fluctuations.

[0109] Finally, the final power synthesis is performed according to the start-stop state combination. For the target electrolyzer units identified as being in the on state, the sum operation is performed on the corresponding operation power lower limit value and the additional power allocation value to obtain its final operation power value; for the units identified as being in the off state, its operation power value is directly determined to be zero. The finally generated operation power value instruction is synchronously sent to the rectifier power supply module supporting each electrolyzer to drive the hybrid connection to complete the real-time consumption of the external fluctuating power supply.

[0110] By employing a two-tiered power determination criterion of lower limit protection and incremental optimization, precise control of the operating power value of heterogeneous electrolyzers is achieved. This solves the technical problems of operational instability or frequent start-stop losses caused by neglecting the lower limit constraint of equipment operation during power allocation in existing technologies. By first extracting the lower limit value of operating power and then allocating additional power, it ensures that all ALK and PEM units in operation are within a safe electrochemical operating range. At the same time, the optimization weights generated by the AQPSO algorithm are used to accurately absorb the remaining power, significantly improving the real-time tracking accuracy and energy utilization rate of wind and solar power fluctuations. This achieves a steady increase in hydrogen production while ensuring the long-term operation of the equipment.

[0111] In some embodiments, the method for determining the target total power to be allocated for the target electrolytic cell hybrid system based on the external fluctuating power input may further include the following: S1: Obtain the converter efficiency corresponding to the input power of the external fluctuating power supply, and the auxiliary power consumption of the target electrolytic cell hybrid system; S2: Determine a preliminary power value based on the power input to the external fluctuating power source according to the converter efficiency; S3: Determine the target value of the total power to be allocated based on the difference between the preliminary power value and the auxiliary machine's operating power consumption.

[0112] Specifically, the first step is to acquire and process the parameters of the energy conversion chain. Specifically, the control unit collects the raw input power of external fluctuating power sources (such as wind farms or photovoltaic power plants) in real time through a communication gateway, and simultaneously acquires the real-time conversion efficiency of the current power electronic conversion equipment (such as rectifiers or DC-DC converters). At the same time, it retrieves the operating power consumption of the auxiliary equipment connected in the target electrolytic cell, including but not limited to the baseline power consumption of circulating pumps, cooling systems, gas-liquid processing units, and control devices under the current operating environment.

[0113] Subsequently, pre-compensation calculations for energy losses are performed to determine the initial usable power value. Based on the obtained converter efficiency, the input power of the external fluctuating power source is proportionally reduced. This step fully considers the heat loss and switching loss during the transmission of electrical energy from the external grid or generator side to the hydrogen production DC bus, ensuring that subsequent optimization calculations are based on the "net input" power, rather than the ideal original power, thus avoiding the power absorption command overshoot problem caused by ignoring converter losses.

[0114] Finally, a power consumption stripping process is performed to precisely lock in the target total power to be allocated. Based on the determined initial power value and the auxiliary machine operating power consumption, a difference operation is performed to subtract the power required to maintain balance of power (BOP) from the available DC power. The remaining difference is determined as the target total power to be allocated. This target value represents the actual power requirement that can be fully mapped to each electrolyzer stack for hydrogen electrolysis at the current sampling time. It directly serves as the input target for the subsequent adaptive quantum behavior particle swarm optimization algorithm, guiding the execution of optimal power flow control.

[0115] By introducing a synergistic correction between converter efficiency and auxiliary power consumption, a computational model that can accurately reflect the internal energy flow of hydrogen production is constructed. This solves the technical problems of large tracking errors and inaccurate power absorption caused by the failure to consider converter losses and internal frictions in power allocation in existing technologies. By pre-extracting auxiliary power consumption, the final determined target value of the total power to be allocated is deeply coupled with the actual electrochemical requirements of each electrolyzer unit. This significantly improves the response accuracy and steady-state tracking precision of the control when facing drastic fluctuations in wind, solar and power, and provides a reliable decision input for ensuring the long-term stable operation of hybrid electrolyzers and the optimal balance of energy utilization.

[0116] In some embodiments, after determining the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system based on the start-stop state combination, the power allocation ratio, and the target total power to be allocated, the method may further include the following: S1: Determine the modulation control parameters of the power conversion equipment corresponding to each electrolytic cell unit based on the operating power value of each electrolytic cell unit; S2: Determine the DC output current of the power conversion equipment based on the modulation control parameters; S3: Drive each of the electrolyzer units to perform electrolytic hydrogen production based on the DC output current.

[0117] Specifically, after determining the operating power value of each electrolytic cell unit, the process enters the low-level hardware instruction mapping and electrochemical drive stage. Specifically, the control device determines the modulation control parameters of the matching power conversion equipment (such as a thyristor rectifier or a high-frequency PWM converter) based on the operating power value corresponding to each target electrolytic cell unit (such as a PEM or ALK unit) and the current DC bus voltage state. These modulation control parameters can be the duty cycle command of a pulse width modulation (PWM) signal or the firing angle control signal of the rectifier bridge, used to convert the power demand determined by the upper-level algorithm into executable switching commands for the power electronic devices.

[0118] Subsequently, the power conversion equipment executes high-frequency switching actions based on the received modulation control parameters, thereby determining and adjusting its DC output current. In this embodiment, the control unit utilizes a preset voltage-current response curve or real-time feedback of the electrolytic cell impedance characteristics to perform closed-loop fine-tuning of the modulation duty cycle, ensuring that the output DC current accurately matches the power consumption requirements of each electrolytic cell unit at that sampling moment. For example, for a PEM electrolytic cell with a fast response, its corresponding conversion equipment will generate a DC current waveform with extremely strong dynamic tracking to closely track the instantaneous peak value of the external fluctuating power supply.

[0119] Finally, based on the generated DC output current, the electrolytic stack of each electrolyzer unit is directly driven to perform electrolytic hydrogen production. The controlled DC current flows through the electrode-electrolyte interface inside the electrolyzer, driving water molecules to undergo electrochemical dissociation reactions, continuously producing high-purity hydrogen gas at the cathode side. Through this layer-by-layer evolution from "optimized power" to "modulation parameters" and then to "drive current," this application ultimately translates the abstract particle swarm optimization result into physical-level energy conversion, realizing closed-loop intelligent control of heterogeneous electrolyzers in complex operating conditions.

[0120] By establishing a precise mapping mechanism from operating power values ​​to modulation parameters of power conversion equipment and then to DC output current, deep collaboration between software optimization algorithms and hardware actuators is achieved. This solves the technical defects of existing technologies, such as the lack of physical-layer implementation methods or insufficient control response precision after power allocation. By dynamically adjusting the DC output current using modulation control parameters, it ensures that each heterogeneous electrolyzer unit can strictly perform electrolysis operations according to the optimal power command obtained through optimization. This not only significantly improves the current response speed and conversion efficiency to fluctuating power sources, but also effectively ensures the stability of electrochemical reactions on the electrode surface through precise current drive. Thus, at the physical execution level, the comprehensive technical goal of improving hydrogen production purity and extending equipment lifespan is ultimately achieved.

[0121] In some embodiments, after obtaining the external fluctuating power input and the operating parameters of the target electrolytic cell hybrid system, the method may further include the following: The dynamic hydrogen price in the current hydrogen market associated with the target electrolyzer hybrid system is obtained through the communication interface.

[0122] Specifically, after acquiring the external fluctuating power input and the operating parameters of the target electrolyzer hybrid system, the implementation also includes obtaining the dynamic hydrogen price in the current hydrogen market associated with the target electrolyzer hybrid system through a communication interface. This embodiment introduces a dynamic economic evaluation module, using the real-time hydrogen price as the core input variable to reconstruct the value of the "hydrogen production efficiency factor" in the original multi-objective function, and simultaneously introduces an "operational intensity penalty factor."

[0123] Specifically, the dynamic economic evaluation module upgrades the hydrogen production efficiency indicator from a simple physical hydrogen production volume to a real-time profit indicator based on economic dimensions. The system no longer simply pursues maximizing hydrogen production volume, but instead aims to maximize net profit per unit time by multiplying real-time hydrogen production volume by the dynamic hydrogen price. Under this restructured mechanism, when hydrogen prices are high, the system automatically increases the weight of hydrogen production revenue in the multi-objective function, prompting each electrolyzer unit to increase its output to capture high profits; conversely, when hydrogen prices are low, the system automatically weakens the priority of hydrogen production targets.

[0124] Furthermore, to protect core assets, this embodiment adds an operational intensity penalty factor to the multi-objective function. This factor dynamically adjusts the penalty for system-level start-up and shutdown frequency and the operational intensity of each unit based on the hydrogen price. When the hydrogen price is low, the system significantly increases the penalty coefficient for frequent equipment start-ups and shutdowns and full-load operation, guiding the electrolyzers into a low-intensity operation mode or selective shutdown to reduce hidden costs caused by accelerated equipment aging and reduce maintenance expenses; conversely, if the hydrogen price can cover equipment depreciation, the penalty will be appropriately relaxed. Through this feedback adjustment based on market fluctuations, this embodiment achieves a dynamic balance between system operating revenue and equipment physical lifespan. In other words, the evaluation factors in the multi-objective function include a profit maximization factor, an operational intensity penalty factor, a dynamic start-up and shutdown penalty factor, and a power quality factor.

[0125] In some embodiments, the control method provided in this specification introduces dynamic hydrogen price as a key decision variable, constructing an adaptive tracking mechanism aimed at maximizing overall economic benefits. This mechanism, by sensing real-time price fluctuations in the hydrogen market, dynamically adjusts the weighting coefficients and evaluation criteria in the multi-objective optimization function, enabling the electrolyzer hybrid system to upgrade from a single objective of "maximizing the absorption of fluctuating energy" to an intelligent decision-making system that "tracks the system's overall economically optimal operating point in real time while ensuring equipment safety." When hydrogen prices are high, the system prioritizes hydrogen production to obtain higher returns; when hydrogen prices are low, the system tends to reduce hydrogen production intensity and the number of start-ups and shutdowns to extend equipment life and reduce operation and maintenance costs.

[0126] Through the above embodiments, this specification precisely identifies the technical pain point of the original solution in handling wind and solar power consumption, which leads to a disconnect between "technological optimization" and "economic optimization" due to the overly static objectives. Addressing the hidden cost issues such as "increased production without increased revenue" and accelerated equipment aging that may arise from fluctuations in hydrogen market prices, this specification introduces a dynamic economic evaluation module. This module uses the real-time hydrogen price as the core decision variable, reconstructs the value of production efficiency factors, and adds operational intensity penalties. This allows the system to shift from simply pursuing maximum output to maximizing net profit per unit time, and to sensitively adjust the penalties for start-up, shutdown, and operating loads based on price levels. Thus, in a volatile business environment, a deep dynamic balance is achieved between economic benefits and the entire lifespan of the equipment.

[0127] In some embodiments, during particle swarm optimization, a preset multi-objective evaluation function is used to calculate the cost value of each particle's encoded sequence. In this embodiment, the evaluation factors in the multi-objective function include a profit maximization factor, an operational intensity penalty factor, a dynamic start-stop penalty factor, and a power quality factor. The system performs a weighted fusion of these factors to form a comprehensive cost value. This serves as a criterion for guiding the system to track the economically optimal operating point.

[0128] S1: Profit maximization factor (alternative hydrogen production efficiency factor).

[0129] This factor is used to quantify the real-time economic benefits generated by the current power allocation scheme. It multiplies the hydrogen production QH2(t) by the real-time hydrogen price PH2(t) to obtain the system's real-time hydrogen production revenue. The optimization objective becomes maximizing this revenue (in optimization algorithms, this is usually transformed into minimizing its negative value).

[0130] Core calculation formula:

[0131] in, The cost of the profit maximization factor (negative hydrogen production revenue). The dynamic hydrogen price at time t (yuan / Nm³) can be obtained in real time through a market data interface. Let be the total hydrogen production of the system at time t (Nm³), which is calculated based on the power-efficiency model of each electrolyzer unit, and the specific expression is as follows:

[0132] in, For the i-th electrolytic cell unit at power The hydrogen production efficiency (Nm³ / kWh) is typically a nonlinear function of power. Let be the operating power (kW) of the i-th electrolytic cell unit at time t. To optimize the cycle (e.g., 1 minute).

[0133] S2: Running intensity penalty factor.

[0134] This factor is used to quantify the impact of the current operating scheme on equipment lifespan. When hydrogen prices are low, the system should avoid allowing equipment to operate in its maximum power range for extended periods or to frequently start and stop, in order to extend equipment lifespan. This factor implements a "reduced intensity during low hydrogen prices" scheduling strategy by penalizing the ratio of the electrolyzer unit's operating power to its rated power (i.e., operating intensity).

[0135] Core calculation formula:

[0136] in, As a performance intensity penalty factor, This represents the intensity penalty coefficient for the i-th electrolyzer unit. The coefficient for alkaline electrolyzers (ALK) is typically higher than that for proton exchange membrane electrolyzers (PEM) because start-up / shutdown and power fluctuations have a greater impact on their lifespan. Let be the rated power of the i-th electrolytic cell unit. Nonlinear penalty exponent ( This results in greater penalties during high-power operation. In start / stop state. 1 indicates on, 0 indicates off.

[0137] S3: Dynamic start-stop penalty factor (alternative start-stop factor).

[0138] A dynamic start-stop penalty mechanism related to hydrogen price is introduced: when the hydrogen price is high, the start-stop penalty is relatively reduced, allowing the system to start and stop equipment more flexibly to capture high-yield opportunities; when the hydrogen price is low, the start-stop penalty is relatively increased to suppress unnecessary equipment start-ups and shutdowns.

[0139] Core calculation formula:

[0140] in, As a dynamic start / stop penalty factor, The baseline start-stop penalty coefficient, This represents the highest historical hydrogen price or a set reference high price. Let be the change in the number of system-level start-stop cycles at time t.

[0141] S4: Comprehensive value.

[0142] The aforementioned profit maximization factor, operating intensity penalty factor, dynamic start-stop penalty factor, and existing power quality factor are weighted and integrated to form a new comprehensive value, which serves as the evaluation criterion for particle swarm optimization.

[0143] Core calculation formula:

[0144] in, The comprehensive value of particles. These are the weight coefficients for each factor. This refers to the power quality factor.

[0145] To further enhance the system's adaptability to hydrogen price fluctuations, this innovation introduces a hydrogen price adaptive weight adjustment mechanism. When the hydrogen price... When the value increases, the profit maximization factor is automatically increased. weight The system is guided to increase hydrogen production; when the price of hydrogen decreases, the operating intensity penalty factor is automatically increased. weight This guides the system to reduce operational intensity and protect equipment lifespan.

[0146] Core calculation formula:

[0147] in, As the weighted benchmark value, This is the sensitivity adjustment coefficient, which controls the response speed of the weights to changes in hydrogen valence. The historical average hydrogen price is used for normalization.

[0148] To further simplify the scheduling logic, this innovation also provides a hydrogen price-based regional scheduling strategy, which automatically switches the priority of multi-objective functions based on the real-time hydrogen price range: High-price area The profit maximization factor has the highest weight, the system prioritizes ensuring hydrogen production, and allows for a moderate increase in the number of start-ups and shutdowns and the intensity of operation.

[0149] Mid-price zone The weights of each factor are balanced, and the system seeks a balance between hydrogen production revenue and equipment lifespan.

[0150] Low price area The weights of the operating intensity penalty factor and the start-up and shutdown penalty factor are increased, and the system prioritizes the protection of equipment, which can appropriately reduce hydrogen production or even selectively shut down some equipment.

[0151] Core calculation formula (weight allocation):

[0152] In some embodiments, the Adaptive Quantum Behavior Particle Swarm Optimization (AQPSO) algorithm is used for iterative optimization. Its core lies in constructing a five-dimensional multi-objective evaluation function that integrates technical constraints and economic benefits. The evaluation factors in this multi-objective function include: power tracking factor, profit maximization factor, operating intensity penalty factor, dynamic start-stop penalty factor, and power quality factor. The system guides the particle swarm to converge towards the operating point with optimal overall performance by performing real-time quantitative evaluation on these five dimensions.

[0153] By constructing a comprehensive evaluation system covering five dimensions—physical balance, economic benefits, equipment health, operational flexibility, and power stability—a deep characterization of the complex operating conditions of heterogeneous electrolyzer hybrid systems was achieved, solving the problem of existing control strategies struggling to balance "technical compliance" and "economic profitability." By introducing a dynamic hydrogen price sensing multi-factor coupling mechanism, the system can adaptively switch production modes based on market conditions while maintaining power tracking accuracy of over 95%. Experimental data shows that this embodiment increases total hydrogen production by approximately 26.8% while effectively extending the entire lifespan of the electrolyzer and significantly reducing the overall operating cost of the system, providing optimal control strategy support for the commercial operation of renewable energy hydrogen production.

[0154] In some embodiments, by introducing a dynamic hydrogen price sensing mechanism, the system's operating point is adaptively tracked from technical indicators to comprehensive economic benefit indicators. The specific implementation process of this method is as follows: S1: Parameter initialization and real-time sensing.

[0155] First, the system executes step S101 to obtain the operating parameters of the target electrolytic cell hybrid system. The parameters include the rated power range, power ramp-up rate limit, and equipment type distribution of each electrolytic cell unit (PEM, ALKs, ALKm).

[0156] Subsequently, the dynamic price of hydrogen in the current market was retrieved in real time via the communication interface. The dynamic hydrogen price, as a core decision variable, will directly participate in the subsequent weight reconstruction of the multi-objective function, enabling the system to sense market fluctuations and adjust production intensity.

[0157] S2: Task construction and coding mapping.

[0158] The system executes step S102, determining the target total power to be allocated for the current sampling period after deducting converter losses and auxiliary power consumption based on the real-time input power of the external fluctuating power supply. Then, step S103 is executed, constructing a particle coding sequence based on the number of electrolytic cell units participating in the scheduling, mapping the physical equipment's operational decisions to a continuous algorithm search space.

[0159] S3: Economically optimal iterative optimization process.

[0160] The system executes step S104, utilizing the Adaptive Quantum Behavior Particle Swarm Optimization (AQPSO) model, based on the operating physical boundary parameters and dynamic hydrogen valence. The particle encoding sequence is iteratively optimized. In each round of iterative evaluation, the system adopts a comprehensive economic cost-benefit approach. As the core criterion:

[0161] The evaluation factors are related to hydrogen prices through the following logic: profit maximization factor Real-time hydrogen production and dynamic hydrogen price Multiply and take the negative value to maximize the tracking profit.

[0162] Operational intensity penalty factor Increase penalties for high-power operation when hydrogen prices are low to protect equipment lifespan.

[0163] Dynamic start-stop penalty factor :according to Dynamically adjust start-stop cost weights to suppress unnecessary system-level switching when hydrogen prices are low.

[0164] In addition, the system according to Fluctuation-based adaptive weighting: When the hydrogen price rises, the weighting is automatically increased. To strengthen the focus on hydrogen production; to automatically increase [the supply] when hydrogen prices decrease. To enhance lifespan protection. Alternatively, the system can automatically switch to a preset zone scheduling strategy based on the hydrogen price range (high price zone, medium price zone, low price zone), achieving a hard switch of priorities.

[0165] S4: Command issuance and physical execution (S105-S106).

[0166] After optimization, step S105 is executed. Based on the determined target particle position information, the start / stop state combination and power allocation ratio of each electrolyzer unit are determined through state discrimination and proportional allocation criteria. Finally, step S106 is executed to synthesize and determine the final operating power value of each unit, driving the hardware system to complete the electrolytic hydrogen production process.

[0167] By embedding a dynamic hydrogen price sensing element into the control process and improving the optimization evaluation criteria, the system achieves deep coupling between technical operation indicators and commercial profit targets, solving the problem of the disconnect between "technically optimal" and "economically optimal" in traditional scheduling schemes. By introducing three-dimensional evaluation factors of profit, intensity, and dynamic start-stop, the system can accurately capture the increased production benefits when hydrogen prices are high, and reduce equipment depreciation costs by actively reducing intensity and suppressing start-stop when hydrogen prices are low. This effectively avoids the operational risk of "increased production without increased revenue" and fundamentally improves the net profit and market competitiveness of the hybrid system throughout its entire life cycle.

[0168] As can be seen from the above, the ALK-PEM hybrid system control method based on wind and solar fluctuations provided in this specification obtains the external fluctuating power input and the operating parameters of the target electrolyzer hybrid system; wherein, the operating parameters include the number of each electrolyzer unit and the corresponding operating physical boundary parameters; the target electrolyzer hybrid system includes at least an alkaline electrolyzer and a proton exchange membrane electrolyzer; the total power target value to be allocated for the target electrolyzer hybrid system is determined according to the external fluctuating power input; a corresponding particle coding sequence is constructed according to the number of electrolyzer units; and a preset particle swarm optimization model is used to determine the operating parameters based on the operating parameters. The boundary parameters are considered, and the particle encoding sequence is iteratively optimized to determine the target particle position information. The iterative optimization process includes: dynamically adjusting the search step size using a linearly decreasing inertia weight adjustment strategy, and performing optimization using a multi-objective function as the evaluation criterion; using a preset power allocation rule, determining the start-stop state combination and power allocation ratio of each electrolytic cell unit in the target electrolytic cell hybrid system based on the target particle position information; and determining the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system based on the start-stop state combination, the power allocation ratio, and the target total power to be allocated. In this way, by acquiring the operating physical boundary parameters of the target electrolytic cell hybrid system and using a preset particle swarm optimization model to iteratively optimize the particle encoding sequence, the search step size is dynamically adjusted through a linearly decreasing inertial weight adjustment strategy, and global space optimization is performed according to the multi-objective function evaluation criterion and the quantum behavior optimization criterion. This solves the problem of limited adjustment range in the face of real-time power fluctuations in existing technologies. By determining the target particle position information and using preset power allocation rules to determine the start-stop state combination, power allocation ratio and operating power value of each electrolytic cell unit, the accurate determination of the total power target value to be allocated based on the external fluctuating power input power is achieved, overcoming the technical defect of power response lag.

[0169] See Figure 2As shown in the embodiments of this specification, a specific electronic device is also provided, wherein the electronic device includes a network communication port 201, a processor 202 and a memory 203, and the above structures are connected by internal cables so that the various structures can perform specific data interaction.

[0170] Specifically, the network communication port 201 can be used to acquire the input power of the external fluctuating power supply and the operating parameters of the target electrolytic cell hybrid system; wherein the operating parameters include the number of each electrolytic cell unit and the corresponding operating physical boundary parameters; the target electrolytic cell hybrid system includes at least an alkaline electrolytic cell and a proton exchange membrane electrolytic cell.

[0171] The processor 202 can be specifically used to determine the target total power to be allocated in the target electrolytic cell hybrid system based on the input power of the external fluctuating power supply; construct a corresponding particle coding sequence based on the number of electrolytic cell units; and perform iterative optimization processing on the particle coding sequence using a preset particle swarm optimization model based on the operating physical boundary parameters to determine the target particle position information. The iterative optimization process includes: dynamically adjusting the search step size using a linearly decreasing inertia weight adjustment strategy, and performing optimization processing using a multi-objective function as the evaluation criterion; determining the start-stop state combination and power allocation ratio of each electrolytic cell unit in the target electrolytic cell hybrid system based on the target particle position information using a preset power allocation rule; and determining the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system based on the start-stop state combination, the power allocation ratio, and the target total power to be allocated.

[0172] The memory 203 can be used to store the corresponding instruction program.

[0173] Based on the above method, the relevant structural performance of electronic equipment can be effectively utilized to improve the data processing speed of electronic equipment and efficiently realize the control method of ALK-PEM hybrid system based on wind and solar fluctuations.

[0174] In this embodiment, the network communication port 201 can be a virtual port bound to different communication protocols, thereby enabling the sending or receiving of different data. For example, the network communication port can be a port responsible for web data communication, a port responsible for FTP data communication, or a port responsible for email data communication. Furthermore, the network communication port can also be a physical communication interface or communication chip. For example, it can be a wireless mobile network communication chip, such as GSM or CDMA; it can also be a Wi-Fi chip; or it can be a Bluetooth chip.

[0175] In this embodiment, the processor 202 can be implemented in any suitable manner. For example, the processor can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers, etc. This specification is not limiting.

[0176] In this embodiment, the memory 203 may include a hierarchy. In a digital system, anything that can store binary data can be a memory. In an integrated circuit, a circuit with storage function but no physical form is also called a memory, such as RAM, FIFO, etc. In a system, a storage device with a physical form is also called a memory, such as a memory stick, TF card, etc.

[0177] This specification also provides a computer-readable storage medium based on the above-described control method for the ALK-PEM hybrid system based on wind and solar fluctuations, which acquires the external fluctuating power input and the operating parameters of the target electrolyzer hybrid system; wherein, the operating parameters include the number of each electrolyzer unit and the corresponding operating physical boundary parameters; the target electrolyzer hybrid system includes at least an alkaline electrolyzer and a proton exchange membrane electrolyzer; based on the external fluctuating power input, the target total power target value to be allocated for the target electrolyzer hybrid system is determined; based on the number of electrolyzer units, a corresponding particle coding sequence is constructed; and using a preset particle swarm optimization model, based on... The physical boundary parameters are used to iteratively optimize the particle encoding sequence to determine the target particle position information. The iterative optimization process includes: dynamically adjusting the search step size using a linearly decreasing inertial weight adjustment strategy, and performing optimization using a multi-objective function as the evaluation criterion; using a preset power allocation rule, determining the start-stop state combination and power allocation ratio of each electrolytic cell unit in the target electrolytic cell hybrid system based on the target particle position information; and determining the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system based on the start-stop state combination, the power allocation ratio, and the target total power to be allocated.

[0178] In this embodiment, the storage medium includes, but is not limited to, Random Access Memory (RAM), Read-Only Memory (ROM), Cache, Hard Disk Drive (HDD), or Memory Card. The memory can be used to store computer program instructions. The network communication unit can be an interface configured according to standards specified in the communication protocol for network connection communication.

[0179] In this embodiment, the specific functions and effects implemented by the program instructions stored in the computer-readable storage medium can be explained in comparison with other embodiments, and will not be repeated here.

[0180] See Figure 3 At the software level, this specification also provides an ALK-PEM hybrid system control device based on wind and solar fluctuations. This device may specifically include the following structural modules: The data acquisition module 301 is used to acquire the input power of the external fluctuating power source and the operating parameters of the target electrolyzer hybrid system; wherein, the operating parameters include the number of each electrolyzer unit and the corresponding operating physical boundary parameters; the target electrolyzer hybrid system includes at least an alkaline electrolyzer and a proton exchange membrane electrolyzer; The total power determination module 302 is used to determine the target total power to be allocated for the target electrolytic cell hybrid system based on the input power of the external fluctuating power supply. The sequence determination module 303 is used to construct a corresponding particle coding sequence based on the number of electrolytic cell units; The position information determination module 304 is used to determine the target particle position information by using a preset particle swarm optimization model and performing iterative optimization processing on the particle encoding sequence according to the running physical boundary parameters; wherein, the iterative optimization processing includes: dynamically adjusting the search step size using a linearly decreasing inertial weight adjustment strategy, and performing optimization processing with a multi-objective function as the evaluation criterion. The start / stop allocation determination module 305 is used to determine the start / stop state combination and power allocation ratio of each electrolytic cell unit in the target electrolytic cell hybrid system based on the target particle position information using a preset power allocation rule. The power allocation module 306 is used to determine the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system based on the start / stop state combination, the power allocation ratio, and the target total power to be allocated.

[0181] In some embodiments, the position information determination module 304, in specific implementation, determines the initial position of each particle in the particle swarm based on the particle encoding sequence and the running physical boundary parameters; the current generation value determination module is used to determine the current generation value of each particle based on the current position of each particle and the multi-objective function; determine the individual target position of each particle and the group target position of the particle swarm based on the current generation value; determine the group center position of the particle swarm based on the individual target positions of all particles; determine the current inertial weight parameter based on the current iteration number, the preset iteration stop threshold, and the linearly decreasing inertial weight adjustment strategy; determine the updated position of each particle using the quantum behavior optimization criterion based on the individual target position, the group target position, the group center position, and the current inertial weight parameter; repeat the steps from determining the current generation value to determining the updated position until the number of iterations reaches the preset iteration stop threshold, and obtain the target particle position information based on the group target position.

[0182] In some embodiments, the current generation value determination module, in specific implementation, determines the operating state and allocated power value of each electrolyzer unit based on the current position of each particle and the operating physical boundary parameters of each electrolyzer unit; determines the evaluation value of each evaluation factor in the multi-objective function based on the operating state, the allocated power value, and the target value of the total power to be allocated; wherein the evaluation factors include power tracking factor, start-stop factor, hydrogen production efficiency factor, and power quality factor; and determines the current generation value of each particle by performing a fusion calculation using the multi-objective function based on the evaluation value of each evaluation factor and the preset weight corresponding to each evaluation factor.

[0183] In some embodiments, the start / stop allocation determination module 305, in specific implementation, determines the particle dimension component value corresponding to each electrolytic cell unit based on the target particle position information; compares each particle dimension component value with a preset state discrimination threshold to obtain a corresponding comparison result; determines the operating state corresponding to each electrolytic cell unit based on the comparison result; wherein, when the particle dimension component value is greater than the preset state discrimination threshold, the corresponding electrolytic cell unit is determined to be in the on state; when the particle dimension component value is not greater than the preset state discrimination threshold, the corresponding electrolytic cell unit is determined to be in the off state; according to a preset mapping relationship, the on state or off state of each electrolytic cell unit is vector reconstructed to construct the start / stop state combination; the particle dimension component value of the electrolytic cell unit in the on state is normalized to determine the weight coefficient of each electrolytic cell unit in the remaining power allocation, and the weight coefficient is used as the power allocation ratio.

[0184] In some embodiments, the power allocation module 306, in its specific implementation, determines the target electrolytic cell unit in the on state according to the start-stop state combination; determines the lower limit value of the operating power corresponding to each target electrolytic cell unit according to the operating physical boundary parameters; determines the remaining total power to be allocated according to the difference between the sum of the lower limit values ​​of the operating power and the target value of the total power to be allocated; performs incremental allocation calculation on the remaining total power to be allocated using the power allocation ratio corresponding to each target electrolytic cell unit to obtain the additional power allocation value corresponding to each target electrolytic cell unit; and determines the operating power value of each electrolytic cell unit according to the start-stop state combination, the lower limit value of the operating power corresponding to each target electrolytic cell unit, and the additional power allocation value.

[0185] In some embodiments, the total power determination module 302 specifically implements the following: acquiring the converter efficiency corresponding to the external fluctuating power input power and the auxiliary machine operating power consumption of the target electrolytic cell hybrid system; determining a preliminary power value based on the converter efficiency and the external fluctuating power input power; and determining the target total power value to be allocated based on the difference between the preliminary power value and the auxiliary machine operating power consumption.

[0186] In some embodiments, the device further includes: determining modulation control parameters of a power conversion device corresponding to each electrolytic cell unit based on the operating power value of each electrolytic cell unit; determining the DC output current of the power conversion device based on the modulation control parameters; and driving each electrolytic cell unit to perform electrolytic hydrogen production based on the DC output current.

[0187] In some embodiments, the device further includes: obtaining the current dynamic hydrogen price in the hydrogen market associated with the target electrolyzer hybrid system via a communication interface.

[0188] It should be noted that the units, devices, or modules described in the above embodiments can be implemented by computer chips or physical entities, or by products with certain functions. For ease of description, the above devices are described by dividing them into various modules according to their functions. Of course, in implementing this specification, the functions of each module can be implemented in the same software and / or hardware, or modules that implement the same function can be implemented by a combination of sub-modules or sub-units, etc. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and there may be other division methods in actual implementation. For example, units or components can be combined or integrated into another system, or some features can be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection between the devices or units shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0189] As can be seen from the above, based on the ALK-PEM hybrid system control device based on wind and solar fluctuations provided in the embodiments of this specification, the external fluctuating power input and the operating parameters of the target electrolyzer hybrid system are obtained; wherein, the operating parameters include the number of each electrolyzer unit and the corresponding operating physical boundary parameters; the target electrolyzer hybrid system includes at least an alkaline electrolyzer and a proton exchange membrane electrolyzer; the total power target value to be allocated to the target electrolyzer hybrid system is determined according to the external fluctuating power input; the corresponding particle coding sequence is constructed according to the number of electrolyzer units; and the operating parameters are obtained using a preset particle swarm optimization model. Physical boundary parameters are used to iteratively optimize the particle encoding sequence to determine the target particle position information. The iterative optimization process includes: dynamically adjusting the search step size using a linearly decreasing inertia weight adjustment strategy, and performing optimization using a multi-objective function as the evaluation criterion; using a preset power allocation rule, determining the start-stop state combination and power allocation ratio of each electrolytic cell unit in the target electrolytic cell hybrid system based on the target particle position information; and determining the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system based on the start-stop state combination, the power allocation ratio, and the target total power to be allocated.

[0190] In a specific scenario example, the ALK-PEM hybrid system control method and device based on wind and solar power fluctuations provided in this specification can be applied to solve the technical problem of limited adjustment range and lag in power response when facing real-time power fluctuations in the power distribution of heterogeneous electrolyzer hybrid systems due to the lack of dynamic characteristics in coordinated scheduling. The specific implementation process may include the following:

[0191] The control method provided in this specification is applied to a hydrogen production station system comprising three types of heterogeneous electrolyzer units. The system includes: The PEM electrolyzer (proton exchange membrane) unit has the following operating physical boundary parameters: power adjustment range 0.05-1.5MW, power ramp limit 1.5MW / min, and single start-up cost of 2 units. The ALKs electrolyzer (small alkaline) unit has the following operating physical boundary parameters: power adjustment range 0.45-2.25MW, power ramp limit 1.0MW / min, and a single start-up cost of 5 units. The ALKm electrolytic cell (medium alkaline) unit has the following operating physical boundary parameters: power adjustment range 1.05-5.25MW, power ramp limit 1.0MW / min, and a single start-up cost of 10 units.

[0192] During the real-time scheduling of the aforementioned hybrid system, the control device strictly executes multi-dimensional physical constraint processing. First, the operating power value P_i of each unit must satisfy the interval constraint of P_min≤P_i≤P_max in real time; Secondly, in order to protect the lifespan of the electrolytic reactor hardware, the system performs ramp constraint verification to ensure that the power change |P_i(t)-P_i(t-1)| between adjacent sampling times does not exceed the corresponding ramp limit; that is, |P_i(t)-P_i(t-1)|≤ramp_limit.

[0193] At the same time, the system follows the minimum power operation criterion, which means that devices in the on state must meet their minimum power requirements, and execute state transition restrictions according to the system-level start-stop determination criteria.

[0194] To verify the superiority of the method described in this specification, the simulation experiment was conducted with a total simulation duration of 24 hours, a sampling time resolution of 1 minute, and a typical wind and solar power output curve as the external fluctuating power source. At the algorithm configuration level, the particle swarm size was set to 30, the maximum number of iterations to 50, and a single 6.6MW large electrolyzer (ALK) strategy was used as a benchmark. Experimental data show that the AQPSO algorithm used in this application exhibits significant technical advantages within a 24-hour operating cycle. Specific benchmarking results are shown in Table 1 below: Table 1

[0195] Analysis of the experimental results shows that the method proposed in this application achieves a synergistic improvement in energy utilization and total hydrogen production. Specifically, the energy utilization rate increases from 77.11% to 95.83%, and the total hydrogen production increases by 26.8%. In particular, regarding equipment protection, through system-level start-stop control in multi-objective optimization, the number of start-stop cycles is significantly reduced from 5 to 1, a reduction of 80%, which significantly reduces physical wear and tear on the equipment and long-term maintenance costs.

[0196] In some embodiments, see Figure 4 The figure shown illustrates the operating results of the electrolyzer hybrid system provided in this manual under a heuristic strategy based on particle swarm optimization (PSO). Figure 4 The first subplot in the figure shows the power tracking performance, where the dashed line represents available photovoltaic power and the solid line represents the power consumed by the PSO algorithm. As can be seen from the figure, the solid line closely matches the fluctuations of the dashed line, indicating that the method of this application has excellent dynamic absorption capability and achieves accurate absorption of fluctuating power sources. Figure 4 The second subplot illustrates the device power allocation, with different shades of shade representing the real-time power output of the proton exchange membrane (PEM) electrolyzer, small alkaline (ALKs) electrolyzer, and medium alkaline (ALKm) electrolyzer, respectively. This plot visually demonstrates the collaborative logic between heterogeneous devices, namely, using the faster-responding PEM electrolyzer to compensate for high-frequency fluctuations, while the ALK electrolyzer undertakes the stable base load.

[0197] Figure 4 The third subplot shows the evolution of total hydrogen production, where the curve represents cumulative hydrogen production. As photovoltaic power is input, hydrogen production shows a smooth upward trend, eventually maximizing hydrogen production benefits at the end of the simulation period. Figure 4 The fourth subplot shows the real-time dynamics of the PSO optimization cost. This curve represents the optimization cost, which is the comprehensive cost calculated by the aforementioned multi-objective function. It can be observed that during periods of drastic fluctuation in photovoltaic power (such as the 10th to 15th hour interval), the cost value remains at a low and stable level. However, at the moment of power drop or system state switching, the cost value fluctuates briefly and then quickly falls back. This strongly demonstrates that the algorithm in this application has extremely strong convergence stability and global optimization ability when dealing with complex constraints and multi-objective conflicts.

[0198] Based on the above embodiments, the AQPSO algorithm's superior performance in handling complex external power fluctuations was verified through refined modeling and multi-constraint collaborative scheduling of three types of heterogeneous electrolyzers: PEM, ALKs, and ALKm. This addresses the shortcomings of traditional single-device control strategies in terms of absorption capacity and response accuracy. Experimental data intuitively demonstrate that while improving energy absorption efficiency (+24.3%) and hydrogen production revenue (+26.8%), this method effectively suppresses frequent start-stop cycles (-80.0%), achieving a deep optimization balance between economic benefits and equipment lifespan. This provides a highly valuable intelligent control solution for large-scale fluctuating new energy hydrogen production systems.

[0199] While this specification provides the steps of operation for the methods described in the embodiments or flowcharts, more or fewer steps may be included based on conventional or non-inventive means. The order of steps listed in the embodiments is merely one possible order of execution among many steps and does not represent the only possible order. In actual device or client product execution, the methods shown in the embodiments or drawings may be executed sequentially or in parallel (e.g., in a parallel processor or multi-threaded processing environment, or even a distributed data processing environment). The terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, product, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, product, or apparatus. Without further limitations, the presence of other identical or equivalent elements in a process, method, product, or apparatus that includes said elements is not excluded. The terms "first," "second," etc., are used to denote names and do not indicate any particular order.

[0200] Those skilled in the art will also know that, besides implementing the controller in the form of purely computer-readable program code, the same functions can be achieved by logically programming the method steps, making the controller take the form of logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers (PLCs), and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the devices within it used to implement various functions can also be considered structures within that hardware component. Alternatively, the devices used to implement various functions can be considered as both software modules implementing the method and structures within a hardware component.

[0201] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this specification can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions of this specification can essentially be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, mobile terminal, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments of this specification.

[0202] Although this specification has been described by way of examples, those skilled in the art will recognize that many variations and modifications are possible without departing from the spirit of this specification, and it is intended that the appended claims cover such variations and modifications without departing from the spirit of this specification.

Claims

1. A control method for an ALK-PEM hybrid system based on wind-solar fluctuations, characterized in that, include: The system acquires the input power of an external fluctuating power source and the operating parameters of the target electrolyzer hybrid system; wherein the operating parameters include the number of each electrolyzer unit and the corresponding operating physical boundary parameters; the target electrolyzer hybrid system includes at least an alkaline electrolyzer and a proton exchange membrane electrolyzer; Based on the external fluctuating power input, determine the target total power to be allocated for the target electrolytic cell hybrid system; Construct a corresponding particle coding sequence based on the number of electrolytic cell units; Using a pre-defined particle swarm optimization model, the particle encoding sequence is iteratively optimized based on the physical boundary parameters to determine the target particle position information; wherein, the iterative optimization process includes: dynamically adjusting the search step size using a linearly decreasing inertial weight adjustment strategy, and performing optimization processing using a multi-objective function as the evaluation criterion; Using a preset power allocation rule, the start-stop state combination and power allocation ratio of each electrolytic cell unit in the target electrolytic cell hybrid system are determined based on the target particle position information; Based on the start / stop state combination, the power allocation ratio, and the target total power to be allocated, the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system is determined.

2. The method according to claim 1, characterized in that, The step of using a preset particle swarm optimization model to iteratively optimize the particle encoding sequence based on the operational physical boundary parameters to determine the target particle position information includes: The initial position of each particle in the particle swarm is determined based on the particle encoding sequence and the physical boundary parameters. Based on the current position of each particle and the multi-objective function, determine the current generation value of each particle; Based on the current generation value, determine the individual target position of each particle and the group target position of the particle swarm; The group center position of the particle swarm is determined based on the individual target positions of all particles. The current inertia weight parameters are determined based on the current iteration number, the preset iteration stop threshold, and the linearly decreasing inertia weight adjustment strategy. The updated position of each particle is determined based on the individual target position, the group target position, the group center position, and the current inertia weight parameter. Repeat the steps from determining the current generation value to determining the update position until the number of iterations reaches the preset iteration stop threshold, and obtain the target particle position information based on the group target position.

3. The method according to claim 2, characterized in that, The step of determining the current generation value of each particle based on its current position and the multi-objective function includes: Based on the current position of each particle and the physical boundary parameters of each electrolytic cell unit, determine the operating status and power allocation value of each electrolytic cell unit. Based on the operating status, the allocated power value, and the target value of the total power to be allocated, the evaluation values ​​of each evaluation factor in the multi-objective function are determined; wherein, the evaluation factors include power tracking factor, start-stop factor, hydrogen production efficiency factor, and power quality factor; Based on the evaluation values ​​of each evaluation factor and the preset weights corresponding to each evaluation factor, the current generation value of each particle is determined by fusion calculation using the multi-objective function.

4. The method according to claim 3, characterized in that, The step of determining the start-stop state combination and power allocation ratio of each electrolytic cell unit in the target electrolytic cell hybrid system based on the target particle position information using a preset power allocation rule includes: Based on the target particle position information and the preset state switching threshold, the operating state of each electrolytic cell unit is determined, and the start-stop state combination is obtained. Based on the start-stop state combination and the corresponding operating physical boundary parameters of each electrolytic cell unit, the basic operating power of each electrolytic cell unit in the start state is determined. The remaining power to be allocated is determined based on the total target power to be allocated and the sum of the basic operating powers. Based on the proportional allocation criteria in the preset power allocation rules, the numerical proportional relationship between each element in the target particle position information, and the remaining power to be allocated, the optimized incremental power corresponding to each electrolytic cell unit is determined. Based on the constraint correction criteria in the preset power allocation rules, the ramp rate limit and power operation range corresponding to each electrolytic cell unit, the sum of the basic operating power and the optimized incremental power is corrected to determine the power allocation ratio corresponding to each electrolytic cell unit.

5. The method according to claim 4, characterized in that, The step of determining the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system based on the start-stop state combination, the power allocation ratio, and the target total power to be allocated includes: Based on the start / stop state combination, the target electrolytic cell unit in the start state is determined; Based on the physical boundary parameters, determine the lower limit of the operating power for each of the target electrolytic cells. The remaining total power to be allocated is determined based on the difference between the sum of the aforementioned operating power lower limits and the target value of the total power to be allocated; Using the power allocation ratio corresponding to each target electrolytic cell unit, the remaining total power to be allocated is incrementally allocated to obtain the additional power allocation value corresponding to each target electrolytic cell unit; The operating power value of each electrolytic cell unit is determined based on the start / stop state combination, the lower limit of operating power corresponding to each target electrolytic cell unit, and the additional power allocation value.

6. The method according to claim 5, characterized in that, The step of determining the target total power to be allocated for the target electrolytic cell hybrid system based on the external fluctuating power input includes: Obtain the converter efficiency corresponding to the input power of the external fluctuating power supply, and the auxiliary operating power consumption of the target electrolytic cell hybrid system; Based on the converter efficiency, determine the initial power value for the input power of the external fluctuating power supply; The target value of the total power to be allocated is determined based on the difference between the preliminary power value and the power consumption of the auxiliary machine.

7. The method according to claim 6, characterized in that, After determining the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system based on the start-stop state combination, the power allocation ratio, and the target total power to be allocated, the method further includes: Based on the operating power value corresponding to each electrolytic cell unit, determine the modulation control parameters of the power conversion equipment corresponding to each electrolytic cell unit; The DC output current of the power conversion equipment is determined based on the modulation control parameters. The electrolytic cell unit is driven to perform electrolytic hydrogen production based on the DC output current.

8. The method according to claim 1, characterized in that, After obtaining the external fluctuating power input and the operating parameters of the target electrolytic cell hybrid system, the method further includes: The dynamic hydrogen price in the current hydrogen market associated with the target electrolyzer hybrid system is obtained through the communication interface.

9. A control device for an ALK-PEM hybrid system based on wind-solar fluctuations, characterized in that, include: The data acquisition module is used to acquire the input power of the external fluctuating power source and the operating parameters of the target electrolyzer hybrid system; wherein, the operating parameters include the number of each electrolyzer unit and the corresponding operating physical boundary parameters; the target electrolyzer hybrid system includes at least an alkaline electrolyzer and a proton exchange membrane electrolyzer; The total power determination module is used to determine the target total power to be allocated for the target electrolytic cell hybrid system based on the input power of the external fluctuating power supply. The sequence determination module is used to construct a corresponding particle coding sequence based on the number of electrolytic cell units; The position information determination module is used to determine the target particle position information by using a preset particle swarm optimization model and iteratively optimizing the particle encoding sequence according to the running physical boundary parameters; wherein, the iterative optimization process includes: dynamically adjusting the search step size using a linearly decreasing inertial weight adjustment strategy, and performing optimization processing with a multi-objective function as the evaluation criterion. The start / stop allocation determination module is used to determine the start / stop state combination and power allocation ratio of each electrolytic cell unit in the target electrolytic cell hybrid system based on the target particle position information using a preset power allocation rule. The power allocation module is used to determine the operating power value of each electrolytic cell unit in the target electrolytic cell hybrid system based on the start / stop state combination, the power allocation ratio, and the target total power to be allocated.

10. A computer-readable storage medium, characterized in that, It stores computer instructions that, when executed by a processor, implement the steps of the method according to any one of claims 1 to 8.

Images