A method, system, device and medium for optimizing operation of a cluster of electrolytic cells

Abstract

The application provides an electrolytic tank cluster operation optimization method, system, device and medium, comprising the following steps: based on the obtained wind and light output and the running state of the electrolytic tank cluster, adopting a multi-type power distribution strategy from the power angle, determining a day-ahead start-stop optimization strategy; based on the day-ahead start-stop optimization strategy, solving a day-ahead operation optimization model from the current angle, determining an electrolytic tank day-ahead plan; based on the electrolytic tank day-ahead plan and the obtained daily wind and light output prediction value, adopting an intra-day optimization model to perform intra-day rolling tracking optimization, obtaining an electrolytic tank intra-day plan; adopting a collaborative control method to perform collaborative control on the electrolytic tank day-ahead plan and the electrolytic tank intra-day plan, completing the operation optimization of the electrolytic tank cluster; the current density, temperature data and power are taken as core control variables and are integrated into the optimization logic, precise collaborative matching of the current-temperature-power is realized, the operation optimization dimension is widened, and the power matching precision and the whole-process optimization capability are improved.

Description

Technical Field

[0001] This invention belongs to the field of energy and control technology, and specifically relates to a method, system, equipment and medium for optimizing the operation of an electrolytic cell cluster. Background Technology

[0002] In the field of new energy hydrogen production, the efficient operation of electrolyzer clusters is the core link to improve energy conversion efficiency and reduce hydrogen production costs. Among them, PEM (proton exchange membrane) electrolyzers and alkaline electrolyzers, due to their respective technical characteristics (PEM electrolyzers have fast response speed and alkaline electrolyzers have low operating costs), often adopt a cooperative operation mode to adapt to the fluctuation characteristics of wind and solar energy, which is a key technical path for large-scale hydrogen production from new energy sources.

[0003] Currently, some studies are exploring the coordinated operation of the two types of electrolyzers, focusing on optimizing the cluster power allocation strategy, and attempting to achieve dynamic adjustment of the overall cluster load by leveraging the characteristics and advantages of the two types of electrolyzers. Meanwhile, some studies have also preliminarily constructed a multi-state operation model for electrolyzers, dividing it into basic states such as production, hot standby, and shutdown, providing a preliminary basis for cluster operation control.

[0004] However, existing methods for optimizing the operation of electrolyzer clusters still have technical shortcomings. They are unable to fully leverage the synergistic advantages of the two types of electrolyzers, nor can they accurately adapt to the actual engineering needs of wind and solar energy with strong volatility and dynamic changes in hydrogen production load. This results in insufficient cluster operation efficiency, flexibility, and stability. The specific pain points are as follows: Existing collaborative operation optimization schemes are mostly based on the breakdown of the total power of the cluster. They only determine the number of two types of electrolyzers to be put into operation, the start-up and shutdown scheme, and the load allocation ratio through power parameters, resulting in a single dimension of operation optimization. Summary of the Invention

[0005] To overcome the shortcomings of the prior art, this invention proposes a method for optimizing the operation of an electrolytic cell cluster, comprising: Based on the obtained wind and solar power output and the operating status of the electrolyzer cluster, a multi-type power allocation strategy is adopted from a power perspective to determine the day-ahead start-stop optimization strategy; the day-ahead start-stop optimization strategy includes the day-ahead set of electrolyzer numbers and the day-ahead set of electrolyzer start-stop states; Based on the aforementioned day-ahead start-up and shutdown optimization strategy, the pre-built day-ahead operation optimization model is solved from the perspective of current to determine the day-ahead schedule of the electrolyzer. Based on the daily plan of the electrolyzer and the obtained intraday wind and solar power output forecast, an intraday optimization model is used to perform intraday rolling tracking optimization to obtain the intraday plan of the electrolyzer. A collaborative control method is adopted to coordinate the daily plan and intraday plan of the electrolyzers, thereby optimizing the operation of the electrolyzer cluster. The current-day operation optimization model includes a first-level objective function constructed with the objectives of minimizing system energy consumption, minimizing lifespan decay, and maximizing economic benefits, respectively. Under the premise that the first-level objective function is optimal, a second-level objective function is constructed with the objective of maximizing the total system benefit. The first constraint conditions are set for the first-level objective function and the second-level objective function with current density and temperature data as key parameters.

[0006] Preferably, the step of determining the day-ahead start-up and shutdown optimization strategy based on the obtained wind and solar power output and the operating status of the electrolyzer cluster, using multiple power allocation strategies from a power perspective, includes: Based on the wind and solar power output and load demand of wind farms and photovoltaic power stations in the operating scenario, the set of day-ahead electrolyzer numbers is determined by using volatility assessment methods and intermittent assessment methods. Based on the set of electrolyzers for the day, a multi-type power allocation strategy is adopted from a power perspective to determine the power of the electrolyzers for the day. Based on the power of the electrolyzers for the day and a pre-constructed multi-state operation characterization model of the electrolyzers, the set of start-up and shutdown states of the electrolyzers for the day is determined. The set of the number of electrolyzers and the set of the start-up and shutdown states of the electrolyzers are used as the start-up and shutdown optimization strategy. The set of electrolyzer quantities for the current day includes: multiple combinations of electrolyzer quantities; the combination of electrolyzer quantities includes the current day alkaline electrolyzer quantity and the current day proton exchange membrane electrolyzer quantity; the multiple power allocation strategies include chain allocation strategy and average allocation strategy.

[0007] Preferably, the multi-type power allocation strategy includes a chain allocation strategy and an average allocation strategy; the step of determining the day-ahead electrolytic cell power by adopting the multi-type power allocation strategy from a power perspective based on the day-ahead set of electrolytic cell numbers, and determining the day-ahead set of electrolytic cell start-up and shutdown states based on the day-ahead electrolytic cell power combined with a pre-constructed multi-state operation characterization model of the electrolytic cells, includes: Based on the day-ahead set of electrolyzers, the wind and solar power output, and the load demand, the chain allocation strategy is adopted. The alkaline electrolyzers in each electrolyzer combination are used as the basic load units, and the proton exchange membrane electrolyzers are used as the fluctuation response units. The alkaline electrolyzers are started at full capacity in a stepwise manner, and the proton exchange membrane electrolyzers take over the remaining power in units of maximum power per cell. The day-ahead electrolyzer power of each electrolyzer combination is obtained. Based on the day-ahead electrolyzer power and the multi-state operation characterization model of the electrolyzers, the start-up and shutdown status of each electrolyzer in each electrolyzer combination in the chain allocation strategy at each time is obtained. Based on the day-ahead set of electrolyzers, the wind and solar power output and load demand, the average allocation strategy is adopted. The alkaline electrolyzers in each electrolyzer combination are used as the basic load units, and the proton exchange membrane electrolyzers are used as the fluctuation response units. The alkaline electrolyzers are started at full capacity in a stepped manner, and the proton exchange membrane electrolyzers are started at the minimum safe power of a single cell. The remaining power of the started proton exchange membrane electrolyzers is evenly allocated to each electrolyzer combination to obtain the day-ahead electrolyzer power of each electrolyzer combination. Based on the day-ahead electrolyzer power and the multi-state operation characterization model of the electrolyzers, the start-up and shutdown status of each electrolyzer in each electrolyzer combination in the average allocation strategy at each time is obtained. Based on the start-up and shutdown status of each electrolytic cell in all combinations of the number of electrolytic cells at each time, a set of start-up and shutdown statuses of electrolytic cells for each day is generated.

[0008] Preferably, the construction of the electrolyzer's operational characterization model includes: An electrolytic cell temperature model is constructed based on the heat generated by the fuel cell stack, the heat flowing into the fuel cell stack, the heat flowing out of the fuel cell stack, and the heat dissipation from the environment. Based on the training data of the electrolytic cell parameter identification model, an electrolytic cell parameter identification model is constructed by training a deep learning artificial intelligence model architecture. Based on the electrolytic cell temperature model, the electrolytic cell parameter identification model, and the internal and external equipment of the electrolytic cell, the electrolytic cell identification parameters and operating equipment are determined. Combined with the electrolytic cell operating state mapping function, the operating state of the electrolytic cell is determined, and the construction of the multi-state operating characterization model of the electrolytic cell is completed. The deep learning artificial intelligence model architecture includes an attention mechanism, an encoder layer, and a decoder layer; the training data includes one or more of the following: historical electrolyzer inlet temperature, historical electrolyzer outlet temperature, historical hydrogen-side pressure, historical oxygen-side pressure, historical water-side pressure, historical hydrogen in oxygen, historical oxygen in hydrogen, historical conductivity, and historical reactor feed water flow rate; the internal and external equipment of the electrolyzer includes one or more of the following: circulating pump, makeup water pump, heater, pneumatic valve, gas separator, pure water machine, cooling system, instrument gas supply system, and current rectifier; the operating state includes one or more of the following: production state, thermal state, idle state, and fault state.

[0009] Preferably, the step of determining the electrolytic cell identification parameters and operating equipment based on the electrolytic cell temperature model, the electrolytic cell parameter identification model, and the internal and external equipment of the electrolytic cell, and determining the operating state of the electrolytic cell by combining the electrolytic cell operating state mapping function, thereby completing the construction of the multi-state operating characterization model of the electrolytic cell, includes: The electrolyzer temperature model and electrolyzer parameter identification model that have been constructed are used to determine the electrolyzer identification parameters; The operating equipment is determined based on the operating status of the internal and external equipment of the electrolytic cell; Based on the electrolytic cell identification parameters and the operating equipment, the operating state of the electrolytic cell is determined using the electrolytic cell operating state characterization mapping function, thus completing the construction of a multi-state operating characterization model for the electrolytic cell.

[0010] Preferably, in the heat engine state, a dynamic and accurate characterization model of the heat engine state is constructed based on current density, electrolytic cell temperature, and electrolytic cell state switching parameters. The expression for the dynamic and precise representation model is as follows:

[0011] In the formula, To be at current density and electrolytic cell temperature Under the hot-engine state, For conceptual functions, To be at current density and electrolytic cell temperature The set of parameters for identifying the lower electrolytic cell. To be at current density and electrolytic cell temperature The lower electrolytic cell contains a collection of internal and external equipment.

[0012] Preferably, the step of solving the pre-built daily operation optimization model from the perspective of current based on the day-to-day start-up and shutdown optimization strategy to determine the daily schedule of the electrolyzer includes: Based on the day-ahead start-stop optimization strategy, from the perspective of current, under the condition of satisfying the first constraint, one of the first-level objective functions of the day-ahead operation optimization model is selected for solution to obtain the optimal actual power value of the electrolyzer for the selected first-level objective function. Under the premise of obtaining the optimal actual power value of the electrolyzer by solving the selected first-level objective function, based on the operation and maintenance cost of the current-day set of electrolyzers, the second-level objective function is solved from the perspective of current, under the condition of satisfying the first constraint, to obtain the optimal number of electrolyzers and the optimal start-up and shutdown state of the electrolyzers. Based on the actual power value of the optimal electrolytic cell, the optimal number of electrolytic cells, and the optimal start-up and shutdown status of the electrolytic cells, a daily plan for the electrolytic cells is generated. The first-level objective function includes a first-level objective function aiming to minimize system energy consumption, a first-level objective function aiming to minimize lifetime degradation, and a first-level objective function aiming to maximize economic benefits; the second-level objective function is constructed based on the operation and maintenance costs of the day-ahead set of electrolyzers, with the goal of maximizing total revenue; the first constraint conditions include one or more of the following: wind and solar power output constraints, grid connection and disconnection constraints, electrolyzer constraints, hydrogen storage tank constraints, and energy storage constraints; the electrolyzer constraints include: power constraints, temperature constraints, overload time constraints, ramp rate constraints, and operating state constraints; the operating state constraints include: operating state constraints and transfer state constraints.

[0013] Preferably, the step of using an intraday optimization model to perform intraday rolling tracking optimization based on the day-ahead plan of the electrolyzer and the obtained intraday wind and solar power output forecast values ​​to obtain the intraday plan of the electrolyzer includes: Based on the daytime wind and solar power output data, the predicted value of the daytime wind and solar power output is obtained through a pre-constructed intraday wind and solar power output prediction model; With a preset time interval as the scheduling cycle, when the first scheduling cycle arrives, starting from the daily plan of the electrolyzer and the predicted value of the daily wind and solar power output, and with the next predicted time domain as the optimization window, the intraday optimization model is solved under the condition of satisfying the second constraint to obtain the control signal of the first scheduling cycle. A feedback correction mechanism is adopted to feed back the predicted value of daily wind and solar power output of the previous scheduling cycle to the predicted value of daily wind and solar power output of the next scheduling cycle. Based on the daily plan of the electrolyzer and the predicted daily wind and solar power output for each scheduling cycle, the daily tracking optimization is performed by rolling the solution of the daily optimization model to obtain the control signal for each scheduling cycle. When the number of scheduling rounds reaches the preset number, the optimal control signal sequence is obtained according to the control signal of each scheduling cycle, which serves as the daily plan of the electrolyzer. The intraday wind and solar power output prediction model is constructed based on historical operating data, including a distributed photovoltaic power output model, a wind power output model, an electrolysis hydrogen production system model, and an energy storage model, through discretization and difference equations.

[0014] Preferably, the construction of the intraday optimization model includes: A daily optimization objective function is constructed with the goal of minimizing the deviation between the daily plan and the intraday plan for the electrolyzer. A second constraint is set for the intraday optimization objective function; The second constraint includes one or more of the following: electrolyzer input power constraint, energy storage power constraint, hydrogen storage power constraint, and incremental constraint; The deviations include one or more of the following: energy consumption deviation, lifespan deviation, economic deviation, alkaline electrolyzer power deviation, proton exchange membrane electrolyzer power deviation, and energy storage power deviation.

[0015] Preferably, the method of employing a collaborative control approach to coordinate the daily plan and intraday plan of the electrolyzers, thereby optimizing the operation of the electrolyzer cluster, includes: A fixed-type electrolytic reactor balanced operation method is adopted to coordinate the daily plan and intraday plan of the electrolytic cells, thereby optimizing the operation of the electrolytic cell cluster. Alternatively, a lifetime-based electrolytic reactor balanced operation method can be adopted to coordinate the daily and intraday plans of the electrolytic cells, thereby optimizing the operation of the electrolytic cell cluster. The fixed-type electrolytic reactor balancing operation method involves numbering the electrolytic cells according to their performance, dividing them into multiple operating scenarios based on the relationship between the total power of multiple electrolytic cells and the power threshold of each cell, allocating power accordingly, and periodically renumbering the cells based on the number of operating conditions in each scenario for rotation. The operating scenarios include: startup operation, optimal power operation, rated power operation, and overload operation. The lifetime-based electrolytic reactor balancing operation method uses the lifetime decay of each electrolytic cell as an optimization target, dynamically allocating power through day-ahead double-layer optimization and intraday rolling optimization to make the cumulative lifetime decay of each electrolytic cell tend to be consistent.

[0016] Based on the same inventive concept, the present invention also provides an electrolytic cell cluster operation optimization system, including: a day-ahead start-stop optimization module, an electrolytic cell day-ahead plan generation module, an electrolytic cell intraday plan generation module, and a collaborative control module; Based on the obtained wind and solar power output and the operating status of the electrolyzer cluster, a multi-type power allocation strategy is adopted from a power perspective to determine the day-ahead start-stop optimization strategy; the day-ahead start-stop optimization strategy includes the day-ahead set of electrolyzer numbers and the day-ahead set of electrolyzer start-stop states; Based on the aforementioned day-ahead start-up and shutdown optimization strategy, the pre-built day-ahead operation optimization model is solved from the perspective of current to determine the day-ahead schedule of the electrolyzer. Based on the daily plan of the electrolyzer and the obtained intraday wind and solar power output forecast, an intraday optimization model is used to perform intraday rolling tracking optimization to obtain the intraday plan of the electrolyzer. A collaborative control method is adopted to coordinate the daily plan and intraday plan of the electrolyzers, thereby optimizing the operation of the electrolyzer cluster. The current-day operation optimization model includes a first-level objective function constructed with the objectives of minimizing system energy consumption, minimizing lifespan decay, and maximizing economic benefits, respectively. Under the premise that the first-level objective function is optimal, a second-level objective function is constructed with the objective of maximizing the total system benefit. The first constraint conditions are set for the first-level objective function and the second-level objective function with current density and temperature data as key parameters.

[0017] Preferably, the daytime start / stop optimization module includes: The day-ahead electrolyzer quantity determination submodule is used to determine the set of day-ahead electrolyzer quantities based on the wind and solar power output and load demand of wind farms and photovoltaic power plants in the operating scenario, using volatility assessment methods and intermittent assessment methods. The daytime electrolyzer start-up / shutdown state determination submodule is used to determine the daytime electrolyzer power based on the set of daytime electrolyzer quantities and adopt a multi-type power allocation strategy from a power perspective. Based on the daytime electrolyzer power and a pre-built multi-state operation characterization model of the electrolyzer, the daytime electrolyzer start-up / shutdown state set is determined. The day-ahead start-stop optimization strategy generation submodule is used to take the day-ahead electrolyzer quantity set and the day-ahead electrolyzer start-stop state set as the day-ahead start-stop optimization strategy; The set of electrolyzer quantities for the current day includes: multiple combinations of electrolyzer quantities; the combination of electrolyzer quantities includes the current day alkaline electrolyzer quantity and the current day proton exchange membrane electrolyzer quantity; the multiple power allocation strategies include chain allocation strategy and average allocation strategy.

[0018] Preferably, the multi-type power allocation strategy includes a chain allocation strategy and an average allocation strategy; the day-ahead electrolyzer start / stop status determination submodule is specifically used for: Based on the day-ahead set of electrolyzers, the wind and solar power output, and the load demand, the chain allocation strategy is adopted. The alkaline electrolyzers in each electrolyzer combination are used as the basic load units, and the proton exchange membrane electrolyzers are used as the fluctuation response units. The alkaline electrolyzers are started at full capacity in a stepwise manner, and the proton exchange membrane electrolyzers take over the remaining power in units of maximum power per cell. The day-ahead electrolyzer power of each electrolyzer combination is obtained. Based on the day-ahead electrolyzer power and the multi-state operation characterization model of the electrolyzers, the start-up and shutdown status of each electrolyzer in each electrolyzer combination in the chain allocation strategy at each time is obtained. Based on the day-ahead set of electrolyzers, the wind and solar power output and load demand, the average allocation strategy is adopted. The alkaline electrolyzers in each electrolyzer combination are used as the basic load units, and the proton exchange membrane electrolyzers are used as the fluctuation response units. The alkaline electrolyzers are started at full capacity in a stepped manner, and the proton exchange membrane electrolyzers are started at the minimum safe power of a single cell. The remaining power of the started proton exchange membrane electrolyzers is evenly allocated to each electrolyzer combination to obtain the day-ahead electrolyzer power of each electrolyzer combination. Based on the day-ahead electrolyzer power and the multi-state operation characterization model of the electrolyzers, the start-up and shutdown status of each electrolyzer in each electrolyzer combination in the average allocation strategy at each time is obtained. Based on the start-up and shutdown status of each electrolytic cell in all combinations of the number of electrolytic cells at each time, a set of start-up and shutdown statuses of electrolytic cells for each day is generated.

[0019] Preferably, the system further includes an electrolytic cell state operation characterization model construction module; the electrolytic cell state operation characterization model construction module includes: The electrolytic cell temperature model construction submodule is used to construct the electrolytic cell temperature model based on the heat generated by the fuel cell stack, the heat flowing into the fuel cell stack, the heat flowing out of the fuel cell stack, and the heat dissipation from the environment. The electrolytic cell parameter identification model construction submodule is used to construct the electrolytic cell parameter identification model based on the training data of the electrolytic cell parameter identification model by training a deep learning artificial intelligence model architecture. The electrolytic cell operation status characterization submodule is used to determine the electrolytic cell identification parameters and operating equipment based on the electrolytic cell temperature model, the electrolytic cell parameter identification model, and the internal and external equipment of the electrolytic cell. Combined with the electrolytic cell operation status mapping function, it determines the operation status of the electrolytic cell and completes the construction of the multi-state operation characterization model of the electrolytic cell. The deep learning artificial intelligence model architecture includes an attention mechanism, an encoder layer, and a decoder layer; the training data includes one or more of the following: historical electrolyzer inlet temperature, historical electrolyzer outlet temperature, historical hydrogen-side pressure, historical oxygen-side pressure, historical water-side pressure, historical hydrogen in oxygen, historical oxygen in hydrogen, historical conductivity, and historical reactor feed water flow rate; the internal and external equipment of the electrolyzer includes one or more of the following: circulating pump, makeup water pump, heater, pneumatic valve, gas separator, pure water machine, cooling system, instrument gas supply system, and current rectifier; the operating state includes one or more of the following: production state, thermal state, idle state, and fault state.

[0020] Preferably, the electrolytic cell operating status characterization submodule is specifically used for: The electrolyzer temperature model and electrolyzer parameter identification model that have been constructed are used to determine the electrolyzer identification parameters; The operating equipment is determined based on the operating status of the internal and external equipment of the electrolytic cell; Based on the electrolytic cell identification parameters and the operating equipment, the operating state of the electrolytic cell is determined using the electrolytic cell operating state characterization mapping function, thus completing the construction of a multi-state operating characterization model for the electrolytic cell.

[0021] Preferably, in the heat engine state, a dynamic and accurate characterization model of the heat engine state is constructed based on current density, electrolytic cell temperature, and electrolytic cell state switching parameters. The expression for the dynamic and precise representation model is as follows:

[0022] In the formula, To be at current density and electrolytic cell temperature Under the hot-engine state, For conceptual functions, To be at current density and electrolytic cell temperature The set of parameters for identifying the lower electrolytic cell. To be at current density and electrolytic cell temperature The lower electrolytic cell contains a collection of internal and external equipment.

[0023] Preferably, the electrolytic cell day-ahead planning generation module is specifically used for: Based on the day-ahead start-stop optimization strategy, from the perspective of current, under the condition of satisfying the first constraint, one of the first-level objective functions of the day-ahead operation optimization model is selected for solution to obtain the optimal actual power value of the electrolyzer for the selected first-level objective function. Under the premise of obtaining the optimal actual power value of the electrolyzer by solving the selected first-level objective function, based on the operation and maintenance cost of the current-day set of electrolyzers, the second-level objective function is solved from the perspective of current, under the condition of satisfying the first constraint, to obtain the optimal number of electrolyzers and the optimal start-up and shutdown state of the electrolyzers. Based on the actual power value of the optimal electrolytic cell, the optimal number of electrolytic cells, and the optimal start-up and shutdown status of the electrolytic cells, a daily plan for the electrolytic cells is generated. The first-level objective function includes a first-level objective function aiming to minimize system energy consumption, a first-level objective function aiming to minimize lifetime degradation, and a first-level objective function aiming to maximize economic benefits; the second-level objective function is constructed based on the operation and maintenance costs of the day-ahead set of electrolyzers, with the goal of maximizing total revenue; the first constraint conditions include one or more of the following: wind and solar power output constraints, grid connection and disconnection constraints, electrolyzer constraints, hydrogen storage tank constraints, and energy storage constraints; the electrolyzer constraints include: power constraints, temperature constraints, overload time constraints, ramp rate constraints, and operating state constraints; the operating state constraints include: operating state constraints and transfer state constraints.

[0024] Preferably, the electrolytic cell daily plan generation module includes: Based on the daytime wind and solar power output data, the predicted value of the daytime wind and solar power output is obtained through a pre-constructed intraday wind and solar power output prediction model; With a preset time interval as the scheduling cycle, when the first scheduling cycle arrives, starting from the daily plan of the electrolyzer and the predicted value of the daily wind and solar power output, and with the next predicted time domain as the optimization window, the intraday optimization model is solved under the condition of satisfying the second constraint to obtain the control signal of the first scheduling cycle. A feedback correction mechanism is adopted to feed back the predicted value of daily wind and solar power output of the previous scheduling cycle to the predicted value of daily wind and solar power output of the next scheduling cycle. Based on the daily plan of the electrolyzer and the predicted daily wind and solar power output for each scheduling cycle, the daily tracking optimization is performed by rolling the solution of the daily optimization model to obtain the control signal for each scheduling cycle. When the number of scheduling rounds reaches the preset number, the optimal control signal sequence is obtained according to the control signal of each scheduling cycle, which serves as the daily plan of the electrolyzer. The intraday wind and solar power output prediction model is constructed based on historical operating data, including a distributed photovoltaic power output model, a wind power output model, an electrolysis hydrogen production system model, and an energy storage model, through discretization and difference equations.

[0025] Preferably, the system further includes: an intraday optimization model construction module; the intraday optimization model construction module is used for: A daily optimization objective function is constructed with the goal of minimizing the deviation between the daily plan and the intraday plan for the electrolyzer. A second constraint is set for the intraday optimization objective function; The second constraint includes one or more of the following: electrolyzer input power constraint, energy storage power constraint, hydrogen storage power constraint, and incremental constraint; The deviations include one or more of the following: energy consumption deviation, lifespan deviation, economic deviation, alkaline electrolyzer power deviation, proton exchange membrane electrolyzer power deviation, and energy storage power deviation.

[0026] Preferably, the collaborative control module is specifically used for: A fixed-type electrolytic reactor balanced operation method is adopted to coordinate the daily plan and intraday plan of the electrolytic cells, thereby optimizing the operation of the electrolytic cell cluster. Alternatively, a lifetime-based electrolytic reactor balanced operation method can be adopted to coordinate the daily and intraday plans of the electrolytic cells, thereby optimizing the operation of the electrolytic cell cluster. The fixed-type electrolytic reactor balancing operation method involves numbering the electrolytic cells according to their performance, dividing them into multiple operating scenarios based on the relationship between the total power of multiple electrolytic cells and the power threshold of each cell, allocating power accordingly, and periodically renumbering the cells based on the number of operating conditions in each scenario for rotation. The operating scenarios include: startup operation, optimal power operation, rated power operation, and overload operation. The lifetime-based electrolytic reactor balancing operation method uses the lifetime decay of each electrolytic cell as an optimization target, dynamically allocating power through day-ahead double-layer optimization and intraday rolling optimization to make the cumulative lifetime decay of each electrolytic cell tend to be consistent.

[0027] Based on the same inventive concept, the present invention also provides an electronic device, comprising: at least one processor and a memory; wherein the memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, an electrolytic cell cluster operation optimization method as described above is implemented.

[0028] Based on the same inventive concept, the present invention also provides a readable storage medium having a computer program stored thereon and an executable program stored thereon, wherein when the executable program is executed, it implements the electrolytic cell cluster operation optimization method as described above.

[0029] Compared with the closest existing technology, the present invention has the following beneficial effects: This invention provides a method, system, equipment, and medium for optimizing the operation of an electrolytic cell cluster, comprising: determining a day-ahead start-stop optimization strategy based on acquired wind and solar power output and the operating status of the electrolytic cell cluster, employing multiple power allocation strategies from a power perspective; the day-ahead start-stop optimization strategy includes a day-ahead set of electrolytic cell numbers and a day-ahead set of electrolytic cell start-stop states; based on the day-ahead start-stop optimization strategy, solving a pre-constructed day-ahead operation optimization model from a current perspective to determine the day-ahead schedule for the electrolytic cells; based on the day-ahead schedule for the electrolytic cells and the acquired intraday wind and solar power output forecasts, performing intraday rolling tracking optimization using an intraday optimization model to obtain the intraday schedule for the electrolytic cells; and employing a collaborative control method to collaboratively control the day-ahead schedule and the intraday schedule for the electrolytic cells to complete the operation optimization of the electrolytic cell cluster; wherein, the day-ahead operation optimization model includes a first-level objective function constructed with the objectives of minimizing system energy consumption, minimizing lifetime degradation, and maximizing economic benefits, respectively, and under the premise of optimizing the first-level objective function, maximizing the total system benefit. The invention constructs a secondary objective function with high power as the objective and sets first constraints for the primary and secondary objective functions using current density and temperature data as key parameters. Based on power, the invention determines the daily set of electrolytic cell quantity and electrolytic cell start / stop status. From the perspective of current, it determines the daily plan for electrolytic cells and sets first constraints for the daily operation optimization model using current density and temperature data as key parameters. Therefore, in the full-process optimization architecture of daily optimization, intraday tracking optimization, and collaborative control, the invention considers not only power but also the impact of current density and temperature data in the electrolytic cell cluster operation optimization. It integrates current density, temperature data, and power as core control variables into the optimization logic, constructing a systematic constraint system and objective function to support refined control, achieving precise collaborative matching of current, temperature, and power, broadening the dimensions of operation optimization, improving power matching accuracy and full-process optimization capabilities, effectively enhancing the electrolytic cell cluster's ability to track wind and solar power, and significantly improving overall operational efficiency and system adaptability. Attached Figure Description

[0030] Figure 1 This invention provides a schematic flowchart of an electrolytic cell cluster operation optimization method. Figure 2 A schematic diagram of the chained allocation strategy provided by the present invention; Figure 3 A schematic diagram of the average allocation strategy provided by the present invention; Figure 4 This is a schematic diagram of electrolytic cell parameter identification provided by the present invention; Figure 5 The electrolytic cell operation status characterization diagram provided by the present invention; Figure 6 This is a schematic diagram of the fixed electrolytic reactor equalization operation method provided by the present invention; Figure 7 Flowchart of the operation optimization method provided by the present invention; Figure 8 A schematic diagram of an electrolytic cell cluster operation optimization system provided by the present invention; Figure 9 This is a schematic diagram of a device structure provided by the present invention. Detailed Implementation

[0031] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0032] Example 1: This invention provides a method for optimizing the operation of an electrolytic cell cluster, such as... Figure 1 As shown, it includes: Step 1: Based on the obtained wind and solar power output and the operating status of the electrolyzer cluster, adopt multiple power allocation strategies from a power perspective to determine the day-ahead start-stop optimization strategy; the day-ahead start-stop optimization strategy includes the day-ahead set of electrolyzer numbers and the day-ahead set of electrolyzer start-stop states; Step 2: Based on the day-ahead start-stop optimization strategy, solve the pre-built day-ahead operation optimization model from the perspective of current to determine the day-ahead plan of the electrolyzer; Step 3: Based on the daily plan of the electrolyzer and the obtained intraday wind and solar power output forecast, an intraday optimization model is used to perform intraday rolling tracking optimization to obtain the intraday plan of the electrolyzer; Step 4: Employ a collaborative control method to coordinate the daily and intraday plans of the electrolyzers, thereby optimizing the operation of the electrolyzer cluster. The current-day operation optimization model includes a first-level objective function constructed with the objectives of minimizing system energy consumption, minimizing lifespan decay, and maximizing economic benefits, respectively. Under the premise that the first-level objective function is optimal, a second-level objective function is constructed with the objective of maximizing the total system benefit. The first constraint conditions are set for the first-level objective function and the second-level objective function with current density and temperature data as key parameters.

[0033] In one implementation, step 1 above, based on the obtained wind and solar power output and the operating status of the electrolyzer cluster, employs multiple power allocation strategies from a power perspective to determine the day-ahead start-up and shutdown optimization strategy, including: Based on the wind and solar power output and load demand of wind farms and photovoltaic power stations in the operating scenario, the set of day-ahead electrolyzer numbers is determined by using volatility assessment methods and intermittent assessment methods. For example, the electrolyzer cluster includes a pure proton exchange membrane (PEM) electrolyzer cluster, a pure alkaline electrolyzer cluster, and a mixed cluster of PEM and pure alkaline electrolyzers; the technical solution of this invention is applicable to the above three types of electrolyzer clusters; based on the wind and solar power output and load demand of wind farms and photovoltaic power plants in the operating scenario, and comprehensively considering the hydrogen storage tank capacity parameters, the fluctuation amplitude and fluctuation rate of wind and solar power output calculated using the fluctuation assessment method, and the duty cycle of abrupt events calculated using the intermittent assessment method, the number of multiple alkaline electrolyzers (ALK) and proton exchange membrane (PEM) electrolyzers is determined to generate a day-ahead electrolyzer quantity set. Simultaneously, to protect the electrolyzers from power fluctuations, batteries are installed to protect them, thus also generating a battery quantity set; the battery quantity set includes multiple combinations of photovoltaic and wind power battery quantities. The wave-based qualitative assessment method is used to calculate the fluctuation amplitude and fluctuation rate of wind and solar power output. Fluctuation range The calculation formula is as follows:

[0034] In the formula, For fluctuation range, The standard deviation of the wind and solar power output; This represents the average power output of wind and solar power. The larger the value, the greater the fluctuation in output power.

[0035] The formula for calculating the fluctuation rate is as follows:

[0036] In the formula, For fluctuation rate, Let be the output value of the new energy source at time t; for The output value of new energy sources at any given time. Logarithmic data on contributions to new energy sources; Indicates a time interval. The larger the value, the faster the output fluctuation rate. An intermittent evaluation method is used to calculate the duty cycle of abrupt changes in wind and solar power output per unit time. The formula for calculating the duty cycle of abrupt changes is as follows:

[0037] in, Duty cycle for sudden change events; The total duration of the abrupt change event; This represents the total observation duration. When... If the output of new energy exceeds a given threshold within a given time interval, a sudden change in new energy output is considered to have occurred, i.e.:

[0038] in, for The output value of new energy sources at any given time. Let be the output value of the new energy source at time t. Threshold for new energy output; At the same time, when the output of new energy sources changes by more than 10% of the rated installed capacity within 15 minutes, it is also considered that a sudden change event has occurred.

[0039] Based on the set of electrolyzers for the day, a multi-type power allocation strategy is adopted from a power perspective to determine the power of the electrolyzers for the day. Based on the power of the electrolyzers for the day and a pre-constructed multi-state operation characterization model of the electrolyzers, the set of start-up and shutdown states of the electrolyzers for the day is determined. The set of the number of electrolyzers and the set of the start-up and shutdown states of the electrolyzers are used as the start-up and shutdown optimization strategy. The day-ahead electrolyzer set includes multiple combinations of electrolyzer quantities; these combinations consist of day-ahead alkaline electrolyzers and day-ahead proton exchange membrane electrolyzers; the multiple power allocation strategies include chain allocation and average allocation strategies. Based on the assessment of wind and solar power fluctuations and intermittency, this invention rationally determines the combination of alkaline and proton exchange membrane electrolyzer quantities, and combines chain and average power allocation strategies to accurately match wind and solar power output fluctuations, reduce wind and solar curtailment, lower start-up and shutdown losses, and improve the economy and adaptability of day-ahead start-up and shutdown strategies.

[0040] In one implementation, the multi-type power allocation strategy includes a chained allocation strategy and an average allocation strategy; the step of determining the day-ahead electrolytic cell power by adopting the multi-type power allocation strategy from a power perspective based on the day-ahead set of electrolytic cell numbers, and determining the day-ahead set of electrolytic cell start-up and shutdown states based on the day-ahead electrolytic cell power combined with a pre-constructed multi-state operation characterization model of the electrolytic cells, includes: Based on the day-ahead set of electrolyzers, the wind and solar power output, and the load demand, the chain allocation strategy is adopted. The alkaline electrolyzers in each electrolyzer combination are used as the basic load units, and the proton exchange membrane electrolyzers are used as the fluctuation response units. The alkaline electrolyzers are started at full capacity in a stepwise manner, and the proton exchange membrane electrolyzers take over the remaining power in units of maximum power per cell. The day-ahead electrolyzer power of each electrolyzer combination is obtained. Based on the day-ahead electrolyzer power and the multi-state operation characterization model of the electrolyzers, the start-up and shutdown status of each electrolyzer in each electrolyzer combination in the chain allocation strategy at each time is obtained. For example, the base load unit maintains stable long-term operation to ensure the stability of hydrogen production; the fluctuation response unit has fast response characteristics and can flexibly adapt to random fluctuations in green electricity output. The power distribution method, which involves the alkaline electrolyzer starting at full capacity in a stepped manner and the proton exchange membrane electrolyzer sequentially receiving the remaining power in units of maximum power per cell, helps extend the lifespan of the proton exchange membrane electrolyzer. Figure 2 The diagram shows a chain-based allocation strategy; the strategy prioritizes ALK electrolyzers at full capacity, with PEM electrolyzers filling in with higher-power cells. Details are as follows: The ALK electrolytic cell is started in a stepped manner: "single cell at full capacity → multiple cells at full capacity → full upper limit". Only when the ALK electrolytic cell cannot absorb the total power will the PEM electrolytic cell start at the maximum power of a single cell. The remaining power is taken over by multiple PEM electrolytic cells, giving priority to ensuring the stable full-capacity operation of the ALK electrolytic cell. The PEM electrolytic cell is only used to make up for the high power.

[0041] In the low power range ( P el ≤40MW, ALK not at full rated capacity, of which, P el (Total power of the electrolytic cell) The total power of the ALK electrolyzer is calculated as follows:

[0042] in, This represents the total power of the ALK electrolytic cell; P el This represents the total power of the electrolytic cell; This refers to the rated power of the ALK electrolytic cell; This refers to the number of ALK electrolytic cells; This represents the maximum power of the ALK electrolytic cell. The power breakdown calculation for the PEM electrolytic cell is as follows:

[0043] in, The power of the j-th PEM electrolyzer in the low-power range, This represents the maximum power of the PEM electrolyzer. The total power of the PEM electrolyzer. This is the minimum power for a PEM electrolyzer.

[0044] In the high power range ( P el >40MW, ALK electrolyzer reaches its upper limit) The total power of the ALK electrolyzer is calculated as follows:

[0045] in, The total power of the ALK electrolytic cell, This represents the total power of the electrolytic cell; This refers to the rated power of the ALK electrolytic cell; This refers to the number of ALK electrolytic cells; This represents the maximum power of the ALK electrolytic cell.

[0046] The power breakdown calculation for the PEM electrolytic cell is as follows:

[0047] in, The power of the j-th PEM electrolyzer in the high-power range during the chain-distribution strategy; Based on the day-ahead set of electrolyzers, the wind and solar power output and load demand, the average allocation strategy is adopted. The alkaline electrolyzers in each electrolyzer combination are used as the basic load units, and the proton exchange membrane electrolyzers are used as the fluctuation response units. The alkaline electrolyzers are started at full capacity in a stepped manner, and the proton exchange membrane electrolyzers are started at the minimum safe power of a single cell. The remaining power of the started proton exchange membrane electrolyzers is evenly allocated to each electrolyzer combination to obtain the day-ahead electrolyzer power of each electrolyzer combination. Based on the day-ahead electrolyzer power and the multi-state operation characterization model of the electrolyzers, the start-up and shutdown status of each electrolyzer in each electrolyzer combination in the average allocation strategy at each time is obtained. For example, such as Figure 3 The diagram shows a schematic of the average power allocation strategy. This strategy prioritizes full capacity for ALK electrolyzers and distributes power evenly among PEM electrolyzers with lower capacity. This power allocation method helps reduce the unit hydrogen production energy consumption of proton exchange membrane electrolyzers, as detailed below: The ALK electrolyzer's stepped full-capacity start-up logic is consistent with the chain-based allocation strategy. The PEM electrolyzer starts up with the minimum power of a single cell, and the remaining power is evenly distributed among multiple PEM electrolyzers. Priority is given to ensuring that multiple PEM electrolyzers operate in parallel at low power, rather than a single cell at high power.

[0048] In the low power range ( P el ≤40MW, ALK electrolyzer not at full rated capacity) Total power of ALK electrolytic cell:

[0049] in, This represents the total power of the ALK electrolytic cell; This represents the total power of the electrolytic cell; This refers to the rated power of the ALK electrolytic cell; This refers to the number of ALK electrolytic cells; This represents the maximum power of the ALK electrolytic cell.

[0050] The power breakdown calculation for the PEM electrolytic cell is as follows:

[0051] in, The power of the j-th PEM electrolyzer in the low-power range; This is the minimum power for a PEM electrolyzer; This represents the total power of the PEM electrolyzer.

[0052] In the high power range ( P el >40MW, ALK reaches its limit) The total power of the ALK electrolytic cell is calculated as follows:

[0053] in, This represents the total power of the ALK electrolytic cell; This represents the total power of the electrolytic cell; This refers to the rated power of the ALK electrolytic cell; This refers to the number of ALK electrolytic cells; This represents the maximum power of the ALK electrolytic cell.

[0054] The power breakdown of the PEM electrolyzer is calculated as follows:

[0055] in, The power of the j-th PEM electrolyzer in the high-power range is allocated in the average distribution strategy; This represents the total power of the PEM electrolyzer. This refers to the number of PEM electrolytic cells.

[0056] Based on the start-up and shutdown status of each electrolyzer at each time point in all combinations of electrolyzer quantities, a set of electrolyzer start-up and shutdown statuses for each day is generated. This invention establishes a multi-state operation characterization model for electrolyzers, considers the differences in specifications of electrolyzers from different manufacturers and performs targeted adaptation processing, and combines chain-type and average multi-type power allocation strategies to determine the start-up and shutdown status of electrolyzers through regularized power allocation.

[0057] In one implementation, the construction of the electrolytic cell state operation characterization model includes: An electrolytic cell temperature model is constructed based on the heat generated by the fuel cell stack, the heat flowing into the fuel cell stack, the heat flowing out of the fuel cell stack, and the heat dissipation from the environment. For example, according to the principle of energy conservation, the total power input to the electrolytic cell... Convert to hydrogen energy and thermal energy, The calculation formula is as follows:

[0058] in, This represents the total power of the electrolytic cell. Where is the voltage of the electrolytic cell, I is the current density, and N is the number of electrolytic cells; The calculation formula is as follows:

[0059] in, Hydrogen energy, U ocv The thermal neutral voltage of the electrolytic cell is set to 1.48V; I For current density, N This refers to the number of electrolytic cells; The formula for calculating the heat generated by the fuel cell stack in the electrolytic cell is as follows:

[0060] in, This refers to the heat generated by the fuel cell stack in the electrolytic cell; The voltage of the electrolytic cell is a function of the current and the stack temperature, and is calculated using an empirical formula:

[0061] In the formula, This refers to the voltage of the electrolytic cell. U rev It is a reversible voltage; A st This represents the electrochemically active area; k 1. k 2. k All three are voltage parameters; T st This refers to the temperature of the electrolytic cell.

[0062] The temperature of the electrolytic cell is determined by the heat generated by the fuel cell stack within the electrolytic cell. Heat flowing into the fuel cell stack Heat flowing out of the fuel cell Environmental heat dissipation The expression is determined by four parts as follows:

[0063]

[0064]

[0065]

[0066] In the formula, This is the equivalent heat capacity of the fuel cell stack; T st This refers to the temperature of the electrolytic cell; The heat generated by the fuel cell stack; This refers to the heat flowing into the fuel cell stack; To dissipate the heat from the fuel cell stack; Heat dissipation for the environment; R st Equivalent thermal resistance; T ex The heat exchanger outlet temperature is set to a constant value. c This is the specific heat capacity of water; The density of water; This refers to the circulating water flow rate; The ambient temperature.

[0067] Based on the training data of the electrolytic cell parameter identification model, an electrolytic cell parameter identification model is constructed by training a deep learning artificial intelligence model architecture. For example, the deep learning artificial intelligence model architecture adopts the Transformer network. The Transformer network consists of an attention mechanism and encoder / decoder layers, abandoning the traditional CNN (Convolutional Neural Network) and recurrent neural network structures. The encoder consists of a multi-head attention layer and a feedforward neutral network (FNN) layer. When one input is encoded, the other inputs are also considered by the multi-head attention layer. The data undergoes nonlinear transformation in the FNN layer, further improving the network's training capability. The decoder network consists of two multi-head attention layers and an FNN layer. The first multi-head attention layer adds a masking operation to prevent training data leakage. The second multi-head attention layer receives the output of the encoder network and calculates the mapping features between input and output variables. Residual connections and layer normalization are performed on the output of each layer to improve the network's convergence performance. Since some parameters of the electrolytic cell cannot be represented by a mechanistic model, this invention uses the Transformer architecture to identify electrolytic cell pressure, safety, and other parameters, completing the construction of an electrolytic cell parameter identification model. like Figure 4 The diagram shows the electrolytic cell parameter identification. By constructing the electrolytic cell parameter identification model, the electrolytic cell identification parameters can be identified based on the input data. The input data includes the current density, voltage, and temperature of the electrolytic cell. The output data of the electrolytic cell parameter identification model includes the electrolytic cell identification parameters, including pressure, safety, and other parameters.

[0068] Based on the electrolytic cell temperature model, the electrolytic cell parameter identification model, and the internal and external equipment of the electrolytic cell, the electrolytic cell identification parameters and operating equipment are determined. Combined with the electrolytic cell operating state mapping function, the operating state of the electrolytic cell is determined, and the construction of the multi-state operating characterization model of the electrolytic cell is completed. The deep learning artificial intelligence model architecture includes an attention mechanism, an encoder layer, and a decoder layer. The training data includes one or more of the following: historical electrolyzer inlet temperature, historical electrolyzer outlet temperature, historical hydrogen-side pressure, historical oxygen-side pressure, historical water-side pressure, historical hydrogen in oxygen, historical oxygen in hydrogen, historical conductivity, and historical feed water flow rate. The internal and external equipment of the electrolyzer includes one or more of the following: circulating pump, makeup water pump, heater, pneumatic valve, gas separator, pure water machine, cooling system, instrument gas supply system, and current rectifier. The operating state includes one or more of the following: production state, thermal state, idle state, and fault state. Existing multi-state models lack systematic representation methods and have insufficient state identification accuracy. Existing research defines the state of the electrolyzer only by rough thresholds of a few key parameters such as temperature and power. It does not integrate the collaborative judgment logic of multiple parameters, nor does it associate the operating state of auxiliary equipment in the cluster (such as circulating pumps, separators, etc.). It only achieves state division through a single parameter threshold, resulting in blurred state boundaries and large identification errors. Meanwhile, existing technologies lack efficient parameter identification methods, making it impossible to accurately obtain key operating parameters such as electrolytic cell pressure and safety. Furthermore, they fail to establish a dual-dimensional "parameter-equipment" status judgment logic, thus failing to accurately reflect the actual operating conditions of the electrolytic cell and auxiliary systems. This impacts the relevance and effectiveness of optimization strategies such as power allocation and status switching, and ultimately fails to provide reliable operating condition support for cluster operation optimization. Figure 5 This invention provides a characterization diagram of the electrolyzer's operating status. The electrolyzer parameter identification model significantly improves the accuracy and reliability of electrolyzer operating status identification, providing reliable operating condition support for cluster optimization. Addressing the shortcomings of existing technologies that rely solely on single parameter thresholds for state classification, suffer from large identification errors, and lack effective parameter identification methods, this invention proposes a systematic multi-state operation characterization method for electrolyzers. On one hand, it establishes an electrolyzer temperature model based on current perspective and employs a Transformer-based intelligent identification method for multi-type electrolysis hydrogen production coordinated control processes to accurately identify electrolyzer pressure, safety, and other key parameters, solving the problem of inaccurate parameter acquisition in existing technologies. On the other hand, a two-dimensional state characterization model of "parameter discrimination + equipment discrimination" is established. Parameter discrimination integrates multiple operating parameters such as temperature, pressure, and safety, while equipment discrimination relates the operating status of equipment within the system and external auxiliary equipment. Only when both discrimination conditions are met simultaneously is the electrolytic cell determined to be in a certain state of production, heat engine, idle, or fault. This effectively eliminates the limitations of single parameter threshold judgment, significantly reduces state identification errors, and provides accurate operating condition basis for optimization strategies such as cluster power allocation, state switching, and equipment deployment planning, avoiding the failure of optimization strategies due to state misjudgment.

[0069] In one implementation, the step of determining the electrolytic cell identification parameters and operating equipment based on the electrolytic cell temperature model, the electrolytic cell parameter identification model, and the internal and external equipment of the electrolytic cell, and determining the operating state of the electrolytic cell by combining the electrolytic cell operating state mapping function, thereby completing the construction of a multi-state operating characterization model of the electrolytic cell, includes: The electrolyzer temperature model and electrolyzer parameter identification model that have been constructed are used to determine the electrolyzer identification parameters; For example, parameter discrimination is used to identify the electrolytic cell identification parameters through the electrolytic cell temperature model and the electrolytic cell parameter identification model.

[0070] The operating equipment is determined based on the operating status of the internal and external equipment of the electrolytic cell; For example, the device identification method is used to determine which devices inside and outside the tank are running. The running device is set to 1, and the non-running device is set to 0.

[0071] Based on the electrolytic cell identification parameters and the operating equipment, the operating state of the electrolytic cell is determined using the electrolytic cell operating state characterization mapping function, thus completing the construction of a multi-state operating characterization model for the electrolytic cell.

[0072] For example, parameter discrimination is used. The electrolytic cell temperature is obtained through an electrolytic cell temperature model. The electrolytic cell temperature, voltage, and current density are then used to obtain electrolytic cell identification parameters through an electrolytic cell parameter identification model. Equipment discrimination is then employed, combining internal system equipment and external auxiliary equipment to establish a multi-state operation characterization model for the electrolytic cell. When the electrolytic cell meets both parameter requirements (e.g., the required temperature is room temperature) and equipment requirements (e.g., the circulating pump is on), it can be determined that the electrolytic cell is in one of the following states: production state, hot-engine state, idle state, or fault state. The operating state expression of the electrolytic cell is as follows:

[0073] In the formula, This refers to the operating status of the electrolytic cell. A mapping function characterizing the operating state of the electrolyzer. Identify the parameter set for the electrolytic cell. It is a collection of internal and external equipment for an electrolytic cell; in,

[0074] in, This refers to the operating status of the electrolytic cell. In production status. In hot state, In idle state When the electrolytic cell is in a fault state, the operating state of the cell is set to display 1, and the other operating states are set to display 0. Parameter discrimination set ; in, Identify the parameter set for the electrolytic cell. This refers to the inlet temperature of the electrolytic cell. This refers to the outlet temperature of the electrolytic cell. The feed water flow rate is set; and the identification parameters for which type of electrolyzer are identified are set, i.e., the identification parameter for that electrolyzer is set to 1, and the identification parameters for other electrolyzers are set to 0. Device discrimination set ; in, It is a collection of internal and external equipment for an electrolytic cell. For circulation pump, For water replenishment pump, It is a current rectifier; and the running device is set to 1, and the non-running device is set to 0.

[0075] Table 1 shows a discrimination diagram of the multi-state operation characterization model of an electrolyzer; Table 1

[0076] This invention obtains key identification parameters based on the electrolytic cell temperature model and parameter identification model, and combines the operating status of internal and external equipment. By comprehensively determining the operating status through a state mapping function, it can achieve accurate and comprehensive characterization of the electrolytic cell operating status, effectively avoid misjudgment based on a single parameter threshold, and improve the reliability of state identification and the credibility of cluster optimization decisions.

[0077] In one implementation, in the heat engine state, a dynamic and accurate characterization model of the heat engine state is constructed based on current density, electrolytic cell temperature, and electrolytic cell state switching parameters. The expression for the dynamic and precise representation model is as follows:

[0078] In the formula, To be at current density and electrolytic cell temperature Under the hot-engine state, For conceptual functions, To be at current density and electrolytic cell temperature The set of parameters for identifying the lower electrolytic cell. To be at current density and electrolytic cell temperature The lower electrolytic cell contains a collection of internal and external equipment.

[0079] For example, the electrolytic cell state switching parameters include a set of electrolytic cell identification parameters and a set of internal and external equipment of the electrolytic cell; the present invention also includes: a dynamic and accurate characterization model in Tst The hot start time at a given temperature is calculated as follows:

[0080] In the formula, To be at current density and electrolytic cell temperature T st Hot start time at temperature; t min Minimum warm-up time; k st This is the temperature-time correlation coefficient; The ambient temperature; Existing state models define hot standby status too rigidly, lacking adaptability and flexibility, and failing to meet the coordinated response requirements of two types of electrolyzers. In existing models, hot standby status is often determined by the electrolyzer's rated temperature, simplifying it into a single discrete state, ignoring the continuous nature of temperature, and failing to differentiate between PEM and alkaline electrolyzers based on their different thermal characteristics. Furthermore, the correlation between hot standby temperature and state switching parameters, hot start-up time, and energy consumption is insufficient, lacking an effective mapping relationship. This results in a lack of refined basis for hot standby status control, making it impossible to dynamically adjust hot standby strategies based on wind and solar power fluctuations, hindering the balance between rapid start-up and shutdown response and energy consumption optimization, and failing to fully adapt to the instantaneous fluctuations in wind and solar energy, further reducing the cluster's wind and solar energy absorption capacity and operational flexibility. This invention achieves refined identification and dynamic control of the thermal engine state, significantly enhancing the adaptability and response flexibility of the electrolyzer cluster to wind and solar energy fluctuations, and meeting the coordinated response requirements of the two types of electrolyzers. To address the shortcomings of existing technologies that limit the thermal engine state to a discrete state at a fixed temperature, fail to consider the differences in thermal characteristics between the two types of electrolyzers, and do not correlate key parameters for state switching, this invention proposes a temperature-corrected refined thermal engine state identification method. This method combines real-time electrolyzer temperature and state switching parameters to establish a dynamic and accurate thermal engine state model, extending the thermal engine state to a continuous temperature range. Simultaneously, it correlates thermal engine temperature with key parameters such as start-up time and energy consumption, establishing a mapping relationship between the two. This method can dynamically adjust the thermal engine temperature according to the fluctuation rhythm of wind and solar power. A higher thermal engine temperature is used for rapid start-up when wind and solar power briefly rebounds, while a lower thermal engine temperature is used to reduce energy consumption when wind and solar power remain low. This effectively shortens the start-up response time and reduces energy waste during the thermal engine process. Furthermore, combined with two rotation strategies for collaborative control, it further enhances the wind and solar power absorption capacity and operational flexibility of the cluster, achieving a balance between rapid start-up and shutdown response and energy consumption optimization, fully adapting to the collaborative response requirements of PEM and alkaline electrolyzers.

[0081] In one implementation, the step of solving a pre-built daily operation optimization model from a current perspective based on the day-to-day start-up and shutdown optimization strategy to determine the daily schedule of the electrolyzer includes: Based on the day-ahead start-stop optimization strategy, from the perspective of current, under the condition of satisfying the first constraint, one of the first-level objective functions of the day-ahead operation optimization model is selected for solution to obtain the optimal actual power value of the electrolyzer for the selected first-level objective function. Under the premise of obtaining the optimal actual power value of the electrolyzer by solving the selected first-level objective function, based on the operation and maintenance cost of the current-day set of electrolyzers, the second-level objective function is solved from the perspective of current, under the condition of satisfying the first constraint, to obtain the optimal number of electrolyzers and the optimal start-up and shutdown state of the electrolyzers. Based on the actual power value of the optimal electrolytic cell, the optimal number of electrolytic cells, and the optimal start-up and shutdown status of the electrolytic cells, a daily plan for the electrolytic cells is generated. The first-level objective function includes a first-level objective function aiming to minimize system energy consumption, a first-level objective function aiming to minimize lifetime degradation, and a first-level objective function aiming to maximize economic benefits; the second-level objective function is constructed based on the operation and maintenance costs of the day-ahead set of electrolyzers, with the goal of maximizing total revenue; the first constraint conditions include one or more of the following: wind and solar power output constraints, grid connection and disconnection constraints, electrolyzer constraints, hydrogen storage tank constraints, and energy storage constraints; the electrolyzer constraints include: power constraints, temperature constraints, overload time constraints, ramp rate constraints, and operating state constraints; the operating state constraints include: operating state constraints and transfer state constraints.

[0082] Example 1) Wind and solar power output constraints

[0083]

[0084] In the formula, for t The theoretical minimum output of wind power at any given time. They are respectively t The theoretical maximum output of wind power at any given time. for t Wind power value at any given time for t The theoretical minimum output of photovoltaic power at any given time. for t Real-time photovoltaic output value They are respectively t The theoretical maximum output of photovoltaic power at any given time.

[0085] 2) Network top and bottom constraints

[0086]

[0087] In the formula, For internet access power consumption, For offline power consumption, For internet access coefficient, This is the offline coefficient. Let be the wind and solar power input at time t. T For the operating cycle.

[0088] 3) Electrolytic cell constraints Electrolytic cell constraints include power, temperature, overload time, and operating status constraints.

[0089] The power constraints are as follows:

[0090]

[0091] In the formula, This is the minimum power of the electrolytic cell; for t Power at time Hydrogen production efficiency; This is the lower limit of the electrolytic cell ramp rate. These represent the upper limits of the ramp rate of the electrolytic cell. This is the maximum power of the electrolytic cell. The actual power of the electrolytic cell at time t; This represents the actual power of the electrolytic cell at time t-1.

[0092] Temperature constraint

[0093] In the formula, This is the lower limit temperature of the electrolytic cell; This is the upper limit temperature of the electrolytic cell. Let t be the temperature of the electrolytic cell at time t.

[0094] Overload Time Constraint The overload limit controller is dynamic. For ALK electrolyzers, if the overload exceeds 100%, it will limit the overload to the nominal level after 15 minutes. For PEM electrolyzers, if the overload exceeds 150%, it will limit the overload to the nominal level after 15 minutes.

[0095] Running state constraints Running state constraints include running state constraints and transition state constraints.

[0096] For runtime constraints, only one of the four runtime states can exist at any given time:

[0097] In the formula, S 1( t ) represents the 0-1 variable of the production state at time t; S 2(t () represents the 0-1 variable of the hot standby state at time t; S 3( t () represents the 0-1 variable of the idle state at time t; S 4( t () represents a 0-1 variable indicating the fault state at time t; for example, when the electrolytic cell is in production mode, i.e. S 1( t Display 1 if the condition is not met, otherwise display 0. There is a certain time interval between the two idle states:

[0098] In the formula, For the first k The end time of the next idle state; For the first k +1 times the start time of the idle state; This is the minimum interval time for the idle state.

[0099] The number of times an idle state exists within a cycle is limited:

[0100] In the formula, For statistical periods; This represents the number of times the idle state occurred. This represents the maximum number of idle times allowed within a period. This is the initial time. For transition state constraints, only one of the six transition states can exist at any given time:

[0101] In the formula, Y c ( t )for t The cold start state is constantly monitored, and the state is represented by 0-1 variables. Y w ( t )for t The state is constantly in a hot start state, and the state is represented by 0-1 variables; Z n ( t )for t The system is always in a stopped state, and the state is represented by 0-1 variables. Z e ( t )for t The system is in a constant state of emergency shutdown, and the state is represented by 0-1 variables. M ( t )for t The status is constantly monitored and represented by 0-1 variables. P( t )for t The operational status is constantly being monitored, and the status is represented by 0-1 variables.

[0102] The state transition has a time constraint:

[0103]

[0104] In the formula, Transition time under different states; For the set of transition times; Cold start time; This refers to the warm start time; This refers to the downtime. This is for emergency shutdown time; For maintenance time; Let X be the transition time in state X; Indicates any; It is a set of transition states.

[0105] Furthermore, only one of the running state and the transition state can exist at any given time:

[0106] In the formula, R ( t )for t The 0-1 variables in the runtime state at any given time; T ( t )for t The 0-1 variables of the transition state at time step.

[0107] 4) Hydrogen storage tank constraints The main constraint on hydrogen storage tanks is capacity constraint. Additionally, the amount of hydrogen stored in the tank must remain constant throughout a given period.

[0108]

[0109] In the formula, for t Hydrogen storage capacity at all times; This is the maximum capacity of the hydrogen storage tank; This is the minimum capacity for the hydrogen storage tank. T For the running cycle; for Hydrogen storage capacity at all times; 5) Energy storage constraints

[0110] In the formula, This is the upper limit of energy storage capacity. Let be the capacity of the hydrogen storage tank at time t; First-level objective function The primary objective function includes minimizing system energy consumption, minimizing lifetime degradation, and maximizing economic benefits.

[0111] 1) Minimize system energy consumption

[0112]

[0113] In the formula, The objective function is a first-order function that minimizes system energy consumption. This refers to system energy consumption; For pump energy consumption; Energy consumption for other auxiliary equipment; DC power consumption; This refers to system losses.

[0114] ① Pump energy consumption Pump energy consumption refers to the electrical energy consumed by various pumps in the system (such as circulating water pumps, makeup water pumps, hydrogen / oxygen transfer pumps, etc.) to drive fluid flow. The calculation formula is as follows:

[0115] In the formula: For the power of the circulating water pump, For the power of the water pump, For the power of the hydrogen transfer pump, For the work of the oxygen delivery pump, This refers to the running time of the pumps in the system.

[0116] The circulating water pump is mainly used to ensure the uniform distribution of electrolyte within the electrolytic cell and to promptly remove the heat generated by the electrolysis reaction, thus ensuring consistent reaction conditions within the cell and maintaining stable system operation. The calculation formula is as follows:

[0117] In the formula, The density of water, This refers to the circulating water flow rate; For the head of the circulating system; ρ is the efficiency of the circulating pump; g is the acceleration due to gravity.

[0118] The makeup water pump is used to continuously replenish pure water to maintain the stable operation of the electrolysis reaction. The makeup water pump must be adapted to ultrapure water conditions to ensure a stable supply of pure water to the system and guarantee the continuous and efficient electrolysis reaction.

[0119] The calculation formula is as follows:

[0120] In the formula, For water replenishment flow; To replenish water head; Water supply pump efficiency; The density of pure water, g is the acceleration due to gravity.

[0121] Hydrogen transfer pumps are used to transfer hydrogen in hydrogen production systems, as expressed by the following formula:

[0122]

[0123] In the formula, The density of the hydrogen delivery pump, This refers to the hydrogen flow rate; To increase the lifting height for hydrogen delivery; For the efficiency of the hydrogen transfer pump; This refers to the current in the electrolytic cell; This represents the molar volume of a gas under standard conditions. For current efficiency; It is Faraday's constant; This represents the number of individual cells connected in series.

[0124] An oxygen delivery pump is used to transport oxygen produced by water electrolysis, ensuring the flow of oxygen from the electrolyzer to subsequent utilization / discharge stages. The formula is as follows:

[0125]

[0126] In the formula, Oxygen density, Oxygen flow rate; For oxygen delivery head; ρ is the efficiency of the oxygen delivery pump; g is the acceleration due to gravity.

[0127] ② Energy consumption of other auxiliary machines Other auxiliary energy consumption refers to the electrical energy consumed by auxiliary equipment other than pumps that ensure the stable operation of the system. This typically includes pure water, cooling, purification, and instrumentation equipment. The calculation formula is as follows:

[0128] In the formula, For the energy consumption of other auxiliary machines, P pure Power of the pure water preparation equipment; P cool Power for cooling equipment; P puri Power of the hydrogen purification equipment;P ins The power of the main monitoring system for instruments and meters, This refers to the auxiliary machine's operating time.

[0129] Pure water preparation equipment provides pure water that meets purity requirements to the electrolyzer. Its energy consumption is directly related to the product water flow rate and water recovery rate, as expressed by the following formula:

[0130]

[0131] In the formula, P pure The power of the pure water preparation equipment, E pure Energy consumption per unit of water production; Q pure Pure water equipment production flow rate; Pure water equipment efficiency; This is the operating voltage for the pure water system; F It is Faraday's constant; The conductivity of the influent; The conductivity of the effluent; The average molar conductivity at infinite dilution of ions; Cooling equipment can remove excess heat generated during system operation. Energy consumption is directly related to the cooling load, as expressed by the following formula:

[0132] In the formula, P cool For cooling equipment power, Q cool This represents the total cooling load of the system. COP The coefficient of performance (COP) of the cooling equipment; This is for the efficiency of cooling equipment. In general, in engineering projects, cooling equipment operates at a fixed power.

[0133] Hydrogen purification equipment purifies crude hydrogen produced by an electrolyzer to a target purity. Energy consumption is directly related to the purification volume, as expressed by the following formula:

[0134]

[0135] In the formula, For the power of hydrogen purification equipment, Energy consumption per unit of purification; This refers to the amount of hydrogen purified. To improve the efficiency of purification equipment; This refers to the current in the electrolytic cell; This represents the molar volume of a gas under standard conditions. For current efficiency; It is Faraday's constant; This represents the number of individual cells connected in series.

[0136] The instruments and meters mainly monitor system parameters and realize automatic control. Energy consumption is based on the rated power of the equipment, and the formula is expressed as follows:

[0137] In the formula: P ins The power of the main monitoring system for instruments and meters, For the first i The rated power of the instrument is h, where h is the number of instruments.

[0138] ③ DC power consumption The formula for calculating the rated input power of an electrolytic cell is as follows:

[0139] In the formula, P 0 represents the rated input power of the electrolytic cell; This refers to the voltage of the electrolytic cell; I 0 represents the current in the electrolytic cell.

[0140] The formula for calculating the DC power consumption of an electrolytic cell is as follows:

[0141] In the formula, DC power consumption per unit of hydrogen production; This refers to the voltage of the electrolytic cell; I 0 represents the current in the electrolytic cell. For testing time; Q This represents the hydrogen production during the test.

[0142] ④ System losses System losses refer to the rectifier conversion losses and DC line losses that exist during the process of converting the power supply of the electrolytic cell from AC to DC.

[0143] Alternating current (AC) from the power grid needs to be converted to direct current (DC) by a rectifier before being input into the electrolytic cell. Efficiency losses occur in the rectifier due to conduction losses and switching losses in the semiconductor devices. The calculation formula is as follows:

[0144] In the formula, This refers to the rectifier conversion loss; To output DC power to the rectifier; This refers to the rectifier efficiency.

[0145] The power loss of DC cables due to resistance follows Joule's law, and the calculation formula is as follows:

[0146]

[0147] In the formula: For line loss; This refers to the current in the electrolytic cell; The resistance of the wire; The resistivity of the conductor; The length of the conductor; This represents the cross-sectional area of ​​the conductor.

[0148] The formula for calculating the total system loss is as follows:

[0149] In the formula, For system losses; This refers to the system uptime.

[0150] 2) With the least lifespan decay

[0151]

[0152] In the formula, The objective function is a first-order function that minimizes lifetime decay. For lifespan decay, Due to the degradation of service life; For variable load life decay; This reduces the lifespan of the start-stop system.

[0153] The degradation of service life under stable operation at the same power can be represented by a power function model:

[0154] In the formula, For service life degradation, This is the upper limit of decay; t L This refers to the actual running time. for t The maximum lifespan of the electrolytic cell at any given time; k L The operating attenuation coefficient, For the running cycle; When operating at different power levels, the curve can be represented by a parabolic model:

[0155] In the formula, For the maximum lifespan of the electrolytic cell, b LDifferent operating power attenuation coefficients; This represents the theoretical maximum lifespan of the electrolytic cell. This is the actual power per unit value; This is the per-unit value of the rated power.

[0156] The attenuation of variable load life is positively correlated with the attenuation coefficient of variable load and the power variation range, as expressed by the following formula:

[0157] In the formula, For variable load life decay, This is the variable load penalty factor; Electrolytic cell power The base attenuation; For the electrolytic cell power from arrive The variable load attenuation coefficient; For the number of load variations; Let be the electrolytic cell power at time i; Let be the electrolytic cell power at time i+1.

[0158] The variable load attenuation coefficient is expressed using an exponential model of the variable load amplitude:

[0159] In the formula, For decay growth parameters; For the maximum load range, For natural numbers, For decay growth parameters; Start-up and shutdown life decay refers to the life decay of an electrolyzer as it changes from an idle state, expressed by the following formula:

[0160] In the formula, To reduce start-stop lifespan, This is the start-stop attenuation coefficient; Power after startup The base attenuation; Number of starts and stops; 3) Based on the highest economic benefits

[0161]

[0162] In the formula, The objective function is the first-order objective function that yields the highest economic benefits. For economic benefits, For the benefit of hydrogen production, Cost over life; The profitability of hydrogen production depends primarily on the price of hydrogen and the amount of hydrogen produced, as shown in the following formula:

[0163]

[0164] In the formula, For the benefit of hydrogen production, For hydrogen valence; V m The standard molar volume of the gas is 22.414 L / mol. F The Faraday constant is 96485 C / mol. For the running cycle; Hydrogen production; This refers to the power of the electrolytic cell; This represents the molar volume of a gas under standard conditions. For electrolytic cell efficiency; U ocv The thermal neutral voltage of the electrolytic cell is set to 1.48V; This refers to the temperature of the electrolytic cell.

[0165] The efficiency of an electrolytic cell is determined by both current efficiency and voltage efficiency, as expressed by the following formula:

[0166]

[0167] In the formula, For the efficiency of the electrolytic cell, The current efficiency is determined through actual measurements. For voltage efficiency, Where is the voltage of the electrolytic cell, and I is the current density; Lifetime cost mainly depends on the lifespan of the electrolyzer and equipment depreciation, as expressed by the following formula:

[0168] In the formula, For life-cycle cost, Let be the depreciation factor at time t. For the operating cycle, For lifespan decay, The lifetime decays at time t; The depreciation factor has an exponential relationship with the percentage of decay; depreciation costs are lower at low decay rates and higher at high decay rates. The formula is as follows:

[0169]

[0170] In the formula, for t Depreciation factor at any time This is the depreciation magnitude factor; This is the depreciation growth rate coefficient; For the cost of electrolytic cell equipment, This is the depreciation factor; Second-order objective function Under the premise of optimizing the first-level objective function, and considering the operation and maintenance costs of different numbers of electrolytic cells, the total system revenue is maximized, as expressed by the following formula:

[0171]

[0172] In the formula, The objective function is a second-order objective function. For the operation and maintenance costs of electrolytic cells; This represents the operation and maintenance cost coefficient of the electrolytic cell; To contribute practically; As a benchmark, The optimal value of the first-order objective function; For the running cycle; ALK electrolytic cell; PEM electrolytic cell. It is an electrolytic cell.

[0173] This invention solves for the first-level objective function based on the current perspective, accurately matching the current-temperature-power coupling characteristics of electrolyzers, effectively reducing energy consumption, delaying lifespan degradation, or improving economic efficiency. Building upon this, a second-level objective function incorporating operation and maintenance costs further optimizes the number of electrolyzers and their start-up and shutdown states, maximizing the overall system benefit. Multi-dimensional constraints (wind and solar power, grid connection and disconnection, electrolyzer power / temperature / ramp-up / overload / state transition, hydrogen storage, and energy storage) comprehensively ensure operational safety and power balance. This solution achieves a two-layer collaboration from single-unit power allocation to cluster combination optimization, significantly improving the economy, adaptability, and reliability of day-ahead planning.

[0174] In one implementation, the step of using an intraday optimization model to perform intraday rolling tracking optimization based on the day-ahead plan of the electrolyzer and the obtained intraday wind and solar power output forecast values ​​to obtain the intraday plan of the electrolyzer includes: Based on the daytime wind and solar power output data, the predicted value of the daytime wind and solar power output is obtained through a pre-constructed intraday wind and solar power output prediction model; For example, in the modeling stage of the intraday wind and solar power output prediction model, a distributed photovoltaic power output model, a wind power output model, an electrolysis hydrogen production system model, and an energy storage model are established based on historical operating data. The variation characteristics of wind and solar power output are described through discretization and difference equations to provide data support for subsequent optimization. After the intraday wind and solar power output prediction model is completed, the intraday wind and solar power output prediction value is obtained based on the day-ahead wind and solar power output data.

[0175] With a preset time interval as the scheduling cycle, when the first scheduling cycle arrives, starting from the daily plan of the electrolyzer and the predicted value of the daily wind and solar power output, and with the next predicted time domain as the optimization window, the intraday optimization model is solved under the condition of satisfying the second constraint to obtain the control signal of the first scheduling cycle. For example, intraday tracking optimization is scheduled at 15-minute intervals, with each 4-hour period constituting a prediction cycle. Optimization is performed for the next 4 hours within each prediction cycle. The scheduling instruction for the first time interval of each scheduling cycle is immediately transmitted to the hybrid electrolysis hydrogen production system, enabling continuous optimization. A total of 6 optimizations are performed within a 24-hour intraday period. During the rolling optimization phase, optimization objectives are set, including rapidly and accurately tracking wind and solar power output, considering factors such as grid integration, lifespan, and economics to formulate intraday plans for the electrolyzers, while also considering constraints such as energy storage limitations, hydrogen production limitations, and wind and solar power limitations to ensure the safety and economy of system operation. When the first scheduling cycle arrives, the intraday wind and solar power output prediction value for the first scheduling cycle is obtained based on the daytime wind and solar power output data from the previous time period. After optimization, control signals are generated and applied to the controlled objects, including the ALK electrolyzer, PEM electrolyzer, and energy storage, to achieve reasonable power allocation.

[0176] A feedback correction mechanism is adopted to feed back the predicted value of daily wind and solar power output of the previous scheduling cycle to the predicted value of daily wind and solar power output of the next scheduling cycle. For example, after control execution, the system feeds back the intraday wind and solar power output forecast value from the previous scheduling cycle to the intraday wind and solar power output forecast model, and continuously corrects and optimizes the value, enabling the system to dynamically adapt to changes in photovoltaic power and improve scheduling accuracy. This optimized control process, combining wind and solar power output forecasting, rolling optimized scheduling, and feedback correction mechanisms, realizes intelligent power allocation between the electrolysis hydrogen production system and the energy storage system, improves the wind and solar power absorption rate, and ensures the safe and stable operation of the electrolyzer and energy storage equipment.

[0177] Based on the daily plan of the electrolyzer and the predicted daily wind and solar power output for each scheduling cycle, the daily tracking optimization is performed by rolling the solution of the daily optimization model to obtain the control signal for each scheduling cycle. When the number of scheduling rounds reaches the preset number, the optimal control signal sequence is obtained according to the control signal of each scheduling cycle, which serves as the daily plan of the electrolyzer. The intraday wind and solar power output prediction model is based on historical operating data and includes distributed photovoltaic power output, wind power output, electrolysis hydrogen production system, and energy storage models, constructed through discretization and difference equations. This invention solves the first-level objective function from the current perspective, accurately matching the current-temperature-power coupling characteristics of the electrolyzer, effectively reducing energy consumption, delaying lifespan degradation, or improving economic efficiency. Furthermore, a second-level objective function based on operation and maintenance costs is introduced to further optimize the number of electrolyzers and their start-up and shutdown states, maximizing the total system benefit. Multi-dimensional constraints (wind and solar, grid connection and disconnection, electrolyzer power / temperature / ramp-up / overload / state transition, hydrogen storage, and energy storage) comprehensively ensure operational safety and power balance. This solution achieves a two-layer collaboration from single-unit power allocation to cluster combination optimization, significantly improving the economy, adaptability, and reliability of day-ahead planning.

[0178] In one implementation, the construction of the intraday optimization model includes: A daily optimization objective function is constructed with the goal of minimizing the deviation between the daily plan and the intraday plan for the electrolyzer. For example, during the intraday tracking optimization phase, the optimization objective is to minimize the deviation from the previous day's optimized scheduling results. The intraday tracking optimization timescale is every 15 minutes, and the prediction time domain is set to 4 hours. The objective function is shown in the following equation:

[0179] In the formula, To optimize the objective function within the day, This represents the change in performance indicators between the daily plan and the intraday plan for the electrolytic cell. This represents the change in the control vector between the day-ahead plan and the day-intraday plan for the electrolytic cell. These are the weighting coefficients; For performance index deviation, To control vector deviation.

[0180] In the objective function The difference between the day-ahead plan and the intraday plan for electrolyzers, which includes energy consumption, lifespan, and economics, corresponds to the day-ahead optimized scheduling strategy. The formula is expressed as follows:

[0181] In the formula, This represents the change in performance indicators between the daily plan and the intraday plan for the electrolytic cell. This represents the deviation from the energy consumption objective function. This represents the deviation from the lifetime objective function. This represents the deviation from the economic objective function.

[0182]

[0183] In the formula, This represents the deviation from the energy consumption objective function. Energy consumption deviation; For a specific moment; For time intervals; To optimize the number of time periods within a forecast period.

[0184]

[0185] In the formula, This is due to lifespan deviation;

[0186] In the formula, For economic deviation; In the objective function This represents the change in the control vector, specifically the difference between the intraday and day-ahead scheduling plans for ALK power, PEM power, and energy storage power. The formula is as follows:

[0187] In the formula, This represents the change in the control vector between the day-ahead plan and the day-intraday plan for the electrolytic cell. For ALK electrolytic cell deviation; For PEM electrolytic cell deviation; This is due to energy storage deviation; To optimize the number of time periods within a forecast period; Introducing weight coefficients into the objective function W This allows for a more rational allocation of surplus power to different control vectors when there is a deviation between the planned power output and the actual power output during the day. The formula is expressed as follows:

[0188] In the formula, W is Weighting coefficients The weight of the ALK electrolyzer. The weight of the PEM electrolyzer. The weighting for energy storage.

[0189] When there is a significant deviation between the actual photovoltaic power output and the planned power output on a given day, the PEM (Power Regulator) should be prioritized for power adjustment to minimize fluctuations in ALK (Alternating Current Generation) and energy storage power, thereby improving hydrogen production efficiency and ensuring that energy storage follows the day-ahead scheduling plan. Therefore, the weighting coefficients of the control vector should be set as follows: .

[0190] A second constraint is set for the intraday optimization objective function; The second constraint includes one or more of the following: electrolyzer input power constraint, energy storage power constraint, hydrogen storage power constraint, and incremental constraint; Example, running constraints During the intraday optimization and scheduling phase, constraints need to be placed on the state vector in the prediction model to ensure it remains within a reasonable operating range. Specifically, this requires limiting the input power range, energy storage power range, and hydrogen storage power range for both types of electrolytic hydrogen production cells. The formula is as follows:

[0191] In the formula, This represents the minimum power of the ALK electrolytic cell; This refers to the power of the ALK electrolytic cell; This represents the maximum power of the ALK electrolytic cell; This represents the minimum power output of the PEM electrolyzer. Power of the PEM electrolyzer. To optimize the number of time periods within a forecast period; For a specific moment; For time intervals; This represents the maximum power of the PEM electrolyzer. This represents the maximum energy storage capacity. Energy storage value, This is the minimum energy storage value; Hydrogen storage capacity; This represents the maximum hydrogen storage capacity.

[0192] At the same time, the increment must be limited to not exceed the maximum or minimum value. The formula is expressed as follows:

[0193] In the formula, This represents the minimum change in the ALK electrolytic cell. This represents the change in the ALK electrolytic cell. To optimize the number of time periods within a forecast period; For a specific moment; For time intervals; This represents the maximum change in the ALK electrolytic cell. This represents the minimum change in the PEM electrolytic cell. This represents the change in the PEM electrolytic cell. This represents the maximum change in the PEM electrolytic cell. This represents the maximum value of the energy storage change. This represents the change in energy storage. This represents the minimum change in hydrogen storage. This represents the change in hydrogen storage. This represents the maximum change in hydrogen storage.

[0194] The deviations include one or more of the following: energy consumption deviation, lifetime deviation, economic deviation, alkaline electrolyzer power deviation, proton exchange membrane electrolyzer power deviation, and energy storage power deviation. This invention constructs an intraday optimization objective function based on minimizing the deviation between the day-ahead and intraday plans for the electrolyzer, comprehensively considering multi-dimensional deviations of energy consumption, lifetime, economy, and control vectors (ALK power, PEM power, and energy storage power), and introduces weighting coefficients and sets... This system can prioritize the rapid adjustment of PEM electrolyzers during wind and solar power fluctuations, maintaining stable ALK and energy storage power. This achieves optimal tracking between day-ahead planning and actual intraday operation while ensuring hydrogen production efficiency and energy storage tracking accuracy. Simultaneously, by setting secondary constraints such as electrolyzer power, energy storage power, hydrogen storage capacity, and incremental change rate, the safety and feasibility of intraday adjustment processes are ensured. This scheme achieves a smooth transition from day-ahead planning to intraday execution, significantly improving the system's adaptability to wind and solar power fluctuations, operational stability, and economic efficiency.

[0195] In one implementation, the method of employing a collaborative control approach to coordinate the daily plan and intraday plan of the electrolyzers, thereby optimizing the operation of the electrolyzer cluster, includes: A fixed-type electrolytic reactor balanced operation method is adopted to coordinate the daily plan and intraday plan of the electrolytic cells, thereby optimizing the operation of the electrolytic cell cluster. Example, (a) fixed type like Figure 6 The diagram illustrates a fixed-reactor equalization operation method. Since each reactor operates under different conditions—startup, optimal, rated, overload, or fluctuating—reactors with higher serial numbers start up first and are most prone to overload and performance degradation. To balance the operating states of the reactors, a fixed-reactor equalization operation method is proposed. The reactor with the best performance is designated as number 1. By comparing the number of times each reactor operates under different conditions within a given period, the reactors are renumbered, achieving reactor rotation and preventing excessive differences between reactors that could lead to premature failure of some reactors.

[0196] Assuming the total operating power of multi-stacking hydrogen production There are multiple electrolytic cells, each with a rated power of [missing information]. Based on the total operating power of multiple electrolysis hydrogen production units The size is divided into 4 scenarios: 1. Start running In this scenario One electrolysis hydrogen production unit is insufficient to start all operations, that is To reduce the number of start-ups and shutdowns, the operating strategy aims to maximize the number of electrolyzers started. The number of hydrogen produced by electrolysis during startup is defined as... a Electrolytic reactor 2 to electrolytic reactor a All reactors operate at startup power, with electrolytic reactor 1 handling the remaining power to prevent any reactor from operating outside of safety constraints. The operating power of each reactor is calculated using the following formula:

[0197] In the formula, The power of the first electrolytic cell; For the power of the second electrolytic cell, For the first a Power of each electrolytic cell For the power of the third electrolytic cell, This refers to the rated power of the electrolytic cell. For the first a+2 The power of each electrolytic cell For the first Power of each electrolytic cell; This represents the total power of the electrolytic cell; The number of electrolytic cells operating at startup power.

[0198] 2. Optimal power operation In this scenario One electrolysis hydrogen production unit is sufficient to start all operations, but not enough for all operations to run at their optimal power point. The operating strategy aims to maximize the number of electrolyzers operating at their optimal power point. The hydrogen production rate at the optimal power point is defined as... b Electrolytic reactor 1 to electrolytic reactor b All are operating at optimal power, electrolytic reactor b +2 to electrolytic reactor Still operating at startup power, electrolytic reactor b +1 provides the remaining power, and the operating power of each electrolytic reactor is calculated according to the following formula:

[0199] In the formula, For the power of the b-th electrolytic cell, For the (b+1)th electrolytic cell power, This represents the power of the (b+2)th electrolytic cell; For the (b+3)th electrolytic cell, For the first The power of each electrolytic cell This refers to the rated power of the electrolytic cell; This represents the total power of the electrolytic cell; The quantity of hydrogen produced by electrolysis is sufficient to start all operations under optimal power operating conditions; This represents the number of electrolytic cells operating at their optimal power point.

[0200] 3. Rated power operation In this scenario Each electrolytic hydrogen production unit can operate at its optimal power, but there is still surplus power that needs to be allocated. This surplus power can be supplied without overload. The operating strategy allocates surplus power to preceding reactors to operate at rated power, while the remaining reactors operate at optimal power, minimizing the number of electrolyzers operating at suboptimal power. The amount of hydrogen produced by electrolysis at rated power is defined as... c Electrolytic reactor 1 to electrolytic reactor c All are operating at rated power, electrolytic reactor c +2 to electrolytic reactor Still operating at optimal power, the electrolytic reactor c +1 provides the remaining power, and the operating power of each electrolytic reactor is calculated according to the following formula:

[0201] In the formula, The power of the c-th electrolytic cell; For the (c+2)th electrolytic cell power, For the (c+3)th electrolytic cell power, For the first The power of each electrolytic cell The number of electrolytic cells operating at rated power; The quantity of hydrogen produced by electrolysis under rated power operation scenarios can meet the optimal power operation requirements; 4. Overload operation In this scenario Each electrolysis hydrogen production unit is operating at its rated power, yet this is still insufficient to meet the total input power requirement. The operating strategy allocates input power exceeding the rated power to as few electrolyzers as possible, reducing the number of electrolyzers operating at overload power. The amount of hydrogen produced by electrolysis at rated power is defined as... d Electrolytic reactor 1 to electrolytic reactor d All are operating at overload power, electrolytic reactor d +2 to electrolytic reactor Still operating at rated power, the electrolytic reactor d +1 provides the remaining power, and the operating power of each electrolytic reactor is calculated according to the following formula:

[0202] In the formula, For the (d+1)th electrolytic cell power, The power of the (d+2)th electrolytic cell The power of the (d+3)th electrolytic cell For the first The power of each electrolytic cell This refers to the number of electrolytic cells operating at overload power.

[0203] Alternatively, a lifetime-based electrolytic reactor balanced operation method can be adopted to coordinate the daily and intraday plans of the electrolytic cells, thereby optimizing the operation of the electrolytic cell cluster. For example, in a wind-solar coupled hydrogen production system, if multiple electrolyzers operate in tandem, an uneven distribution of power within the electrolyzers can easily occur during the rotation cycle if an effective power allocation strategy is lacking. Specifically, some electrolyzers bear high loads for extended periods, with frequent start-ups, shutdowns, and load changes, resulting in a significantly higher rate of power decay than other equipment; while other electrolyzers remain at low loads or idle for extended periods, leading to insufficient resource utilization. This imbalance not only shortens the overall average lifespan of the electrolyzers and increases maintenance and replacement costs, but also reduces system reliability and availability, hindering the synergistic improvement of wind and solar power absorption capacity and hydrogen production economics. To address these issues, a two-layer collaborative control framework of "day-ahead optimized operation + intraday optimized operation" is used to achieve a dynamic balance between electrolyzer power allocation and lifespan decay. First, in day-ahead optimized operation, an optimization model is constructed based on multi-source data such as day-ahead wind and solar power output forecasts, time-of-use electricity prices, and hydrogen load demand, with the goal of "lowest hydrogen production cost + balanced lifespan decay." By rationally allocating the baseline output plan for each electrolyzer, the system avoids single electrolyzers operating under high load or extreme conditions for extended periods, laying the foundation for balanced lifespan over a long timescale. Secondly, in intraday tracking and optimization, the power commands for each electrolyzer are rapidly corrected to address uncertainties such as fluctuations in wind and solar power output and sudden changes in hydrogen load during real-time operation. The rapid load-changing capability of the PEM electrolyzers smooths out short-term power fluctuations, while strictly controlling the number of start-ups and shutdowns and the rate of load change to prevent excessive accumulation of start-up / shutdown attenuation and load-change attenuation. Finally, the lifespan consistency of the overall electrolyzers is ensured through lifespan deviation indices between electrolyzers.

[0204] In the formula, D i For the first i The lifespan of each electrolytic cell decreases; L This is an indicator of the lifespan deviation between electrolytic cells; N This represents the number of electrolytic cells.

[0205] By employing the above-mentioned balanced operation method, while ensuring the economic efficiency of hydrogen production and the capacity for wind and solar energy absorption, the differences in the lifespan degradation of each electrolyzer can be effectively mitigated, the overall lifespan of the electrolyzer stack can be extended, and the long-term operational efficiency of the system can be improved.

[0206] The fixed electrolytic reactor balancing operation method involves numbering the electrolytic cells according to their performance, dividing them into multiple operating scenarios based on the relationship between the total power of multiple electrolytic cells and the power threshold of each cell, and allocating power accordingly. The cells are periodically renumbered based on the number of operating conditions in each scenario for rotation. These operating scenarios include: start-up operation, optimal power operation, rated power operation, and overload operation. The lifetime-based electrolytic reactor balancing operation method uses the lifetime decay of each cell as an optimization objective, dynamically allocating power through day-ahead double-layer optimization and intraday rolling optimization to ensure that the cumulative lifetime decay of each cell tends to be consistent. Existing technologies have a single optimization dimension, lacking refined control of current and temperature coupling, and insufficient accuracy in matching power allocation strategies with wind and solar power. Existing collaborative operation optimization schemes mostly revolve around splitting the total power of the cluster, determining the number of cells, start-up and shutdown schemes, and load allocation ratios for the two types of electrolytic cells only through power parameters. They fail to integrate the coupling relationship between current density, temperature, and power as core control variables into the optimization logic, and do not construct a systematic constraint system and objective function to support refined control. The electrolysis efficiency and equipment losses of both PEM and alkaline electrolyzers exhibit non-linear relationships with current density, and temperature directly affects the current adaptation range and power output characteristics of the electrolyzer. Optimization from a single power perspective cannot achieve synergistic matching of current, temperature, and power, making it difficult for the electrolyzer's output power to accurately track fluctuations in wind and solar power, easily leading to wind and solar curtailment or energy waste. Furthermore, the lack of tiered target guidance makes it impossible to balance economic benefits, equipment lifespan, and energy consumption control, thus restricting the overall operational efficiency of the cluster. These technical problems, combined, prevent existing electrolyzer cluster operation optimization methods from fully leveraging the synergistic advantages of PEM and alkaline electrolyzers, resulting in low cluster operating efficiency, insufficient wind and solar power absorption capacity, high equipment maintenance costs, and poor system safety and economy, thus hindering the efficient application of electrolyzer clusters in the field of new energy hydrogen production.

[0207] like Figure 7The diagram shows the operation optimization method of this invention. This method addresses the technical problems of existing electrolyzer cluster operation optimization methods, such as a single optimization dimension, insufficient accuracy of multi-state model representation, and rigid hot standby state definition. It proposes a continuous temperature range dynamic control mode for the thermal engine state. This is achieved by constructing a full-process optimization architecture coupled with current, temperature, and power, establishing a multi-state representation method with a two-dimensional "parameter-equipment" structure, and combining temperature correction to achieve refined identification and dynamic control of the thermal engine state. Ultimately, this improves the electrolyzer cluster's ability to track and adapt to wind and solar power, balancing operational economy, equipment lifespan, and energy consumption optimization. It fully leverages the synergistic advantages of PEM and alkaline electrolyzers to meet the needs of new energy hydrogen production.

[0208] By broadening the dimensions of operational optimization, improving power matching accuracy and full-process optimization capabilities, this invention effectively enhances the electrolyzer cluster's ability to track wind and solar power, significantly improving overall operational efficiency and system adaptability. Addressing the shortcomings of existing technologies that only optimize clusters based on power parameters and do not incorporate the coupling control dimensions of current density and temperature, this invention constructs a full-process optimization architecture of "day-ahead dual-layer optimization + intraday tracking optimization + collaborative control." The day-ahead dual-layer optimization includes start-up and shutdown optimization and operational optimization. Start-up and shutdown optimization determines the number of operating devices and their start-up / shutdown states based on power, while operational optimization determines the operational strategy based on current. Simultaneously, it constructs operational constraints and two-level objective functions that consider the influence of current and temperature. The first-level objective function includes three categories: optimal economic benefit, minimum lifespan degradation, and minimum system energy consumption, all considering the influence of current and temperature, used to determine the actual power of the electrolyzer. The second-level objective function integrates the results of day-ahead start-up and shutdown and operational optimization to ensure maximum overall system benefit. At the same time, by combining the hybrid power allocation strategy, precise coordination and matching of current, temperature and power can be achieved, which not only avoids the waste of wind and solar power, but also reduces equipment operating losses. Furthermore, it can effectively improve the overall energy conversion efficiency of the cluster, the service life of equipment, and the economy and safety of the system for a variety of typical application scenarios.

[0209] Example 2: Based on the same inventive concept, this invention also provides an electrolytic cell cluster operation optimization system, such as... Figure 8 As shown, it includes: a day-ahead start / stop optimization module, a day-ahead plan generation module for electrolyzers, a day-intraday plan generation module for electrolyzers, and a collaborative control module; Based on the obtained wind and solar power output and the operating status of the electrolyzer cluster, a multi-type power allocation strategy is adopted from a power perspective to determine the day-ahead start-stop optimization strategy; the day-ahead start-stop optimization strategy includes the day-ahead set of electrolyzer numbers and the day-ahead set of electrolyzer start-stop states; Based on the aforementioned day-ahead start-up and shutdown optimization strategy, the pre-built day-ahead operation optimization model is solved from the perspective of current to determine the day-ahead schedule of the electrolyzer. Based on the daily plan of the electrolyzer and the obtained intraday wind and solar power output forecast, an intraday optimization model is used to perform intraday rolling tracking optimization to obtain the intraday plan of the electrolyzer. A collaborative control method is adopted to coordinate the daily plan and intraday plan of the electrolyzers, thereby optimizing the operation of the electrolyzer cluster. The current-day operation optimization model includes a first-level objective function constructed with the objectives of minimizing system energy consumption, minimizing lifespan decay, and maximizing economic benefits, respectively. Under the premise that the first-level objective function is optimal, a second-level objective function is constructed with the objective of maximizing the total system benefit. The first constraint conditions are set for the first-level objective function and the second-level objective function with current density and temperature data as key parameters.

[0210] Preferably, the daytime start / stop optimization module includes: The day-ahead electrolyzer quantity determination submodule is used to determine the set of day-ahead electrolyzer quantities based on the wind and solar power output and load demand of wind farms and photovoltaic power plants in the operating scenario, using volatility assessment methods and intermittent assessment methods. The daytime electrolyzer start-up / shutdown state determination submodule is used to determine the daytime electrolyzer power based on the set of daytime electrolyzer quantities and adopt a multi-type power allocation strategy from a power perspective. Based on the daytime electrolyzer power and a pre-built multi-state operation characterization model of the electrolyzer, the daytime electrolyzer start-up / shutdown state set is determined. The day-ahead start-stop optimization strategy generation submodule is used to take the day-ahead electrolyzer quantity set and the day-ahead electrolyzer start-stop state set as the day-ahead start-stop optimization strategy; The set of electrolyzer quantities for the current day includes: multiple combinations of electrolyzer quantities; the combination of electrolyzer quantities includes the current day alkaline electrolyzer quantity and the current day proton exchange membrane electrolyzer quantity; the multiple power allocation strategies include chain allocation strategy and average allocation strategy.

[0211] Preferably, the multi-type power allocation strategy includes a chain allocation strategy and an average allocation strategy; the day-ahead electrolytic cell start / stop status determination submodule is specifically used for: Based on the day-ahead set of electrolyzers, the wind and solar power output, and the load demand, the chain allocation strategy is adopted. The alkaline electrolyzers in each electrolyzer combination are used as the basic load units, and the proton exchange membrane electrolyzers are used as the fluctuation response units. The alkaline electrolyzers are started at full capacity in a stepwise manner, and the proton exchange membrane electrolyzers take over the remaining power in units of maximum power per cell. The day-ahead electrolyzer power of each electrolyzer combination is obtained. Based on the day-ahead electrolyzer power and the multi-state operation characterization model of the electrolyzers, the start-up and shutdown status of each electrolyzer in each electrolyzer combination in the chain allocation strategy at each time is obtained. Based on the day-ahead set of electrolyzers, the wind and solar power output and load demand, the average allocation strategy is adopted. The alkaline electrolyzers in each electrolyzer combination are used as the basic load units, and the proton exchange membrane electrolyzers are used as the fluctuation response units. The alkaline electrolyzers are started at full capacity in a stepped manner, and the proton exchange membrane electrolyzers are started at the minimum safe power of a single cell. The remaining power of the started proton exchange membrane electrolyzers is evenly allocated to each electrolyzer combination to obtain the day-ahead electrolyzer power of each electrolyzer combination. Based on the day-ahead electrolyzer power and the multi-state operation characterization model of the electrolyzers, the start-up and shutdown status of each electrolyzer in each electrolyzer combination in the average allocation strategy at each time is obtained. Based on the start-up and shutdown status of each electrolytic cell in all combinations of the number of electrolytic cells at each time, a set of start-up and shutdown statuses of electrolytic cells for each day is generated.

[0212] Preferably, the system further includes an electrolytic cell state operation characterization model construction module; the electrolytic cell state operation characterization model construction module includes: The electrolytic cell temperature model construction submodule is used to construct the electrolytic cell temperature model based on the heat generated by the fuel cell stack, the heat flowing into the fuel cell stack, the heat flowing out of the fuel cell stack, and the heat dissipation from the environment. The electrolytic cell parameter identification model construction submodule is used to construct the electrolytic cell parameter identification model based on the training data of the electrolytic cell parameter identification model by training a deep learning artificial intelligence model architecture. The electrolytic cell operation status characterization submodule is used to determine the electrolytic cell identification parameters and operating equipment based on the electrolytic cell temperature model, the electrolytic cell parameter identification model, and the internal and external equipment of the electrolytic cell. Combined with the electrolytic cell operation status mapping function, it determines the operation status of the electrolytic cell and completes the construction of the multi-state operation characterization model of the electrolytic cell. The deep learning artificial intelligence model architecture includes an attention mechanism, an encoder layer, and a decoder layer; the training data includes one or more of the following: historical electrolyzer inlet temperature, historical electrolyzer outlet temperature, historical hydrogen-side pressure, historical oxygen-side pressure, historical water-side pressure, historical hydrogen in oxygen, historical oxygen in hydrogen, historical conductivity, and historical reactor feed water flow rate; the internal and external equipment of the electrolyzer includes one or more of the following: circulating pump, makeup water pump, heater, pneumatic valve, gas separator, pure water machine, cooling system, instrument gas supply system, and current rectifier; the operating state includes one or more of the following: production state, thermal state, idle state, and fault state.

[0213] Preferably, the electrolytic cell operating status characterization submodule is specifically used for: The electrolyzer temperature model and electrolyzer parameter identification model that have been constructed are used to determine the electrolyzer identification parameters; The operating equipment is determined based on the operating status of the internal and external equipment of the electrolytic cell; Based on the electrolytic cell identification parameters and the operating equipment, the operating state of the electrolytic cell is determined using the electrolytic cell operating state characterization mapping function, thus completing the construction of a multi-state operating characterization model for the electrolytic cell.

[0214] Preferably, in the heat engine state, a dynamic and accurate characterization model of the heat engine state is constructed based on current density, electrolytic cell temperature, and electrolytic cell state switching parameters. The expression for the dynamic and precise representation model is as follows:

[0215] In the formula, To be at current density and electrolytic cell temperature Under the condition of a hot engine, For conceptual functions, To be at current density and electrolytic cell temperature The set of parameters for identifying the lower electrolytic cell. To be at current density and electrolytic cell temperature The lower electrolytic cell contains a collection of internal and external equipment.

[0216] Preferably, the electrolytic cell day-ahead planning generation module is specifically used for: Based on the day-ahead start-stop optimization strategy, from the perspective of current, under the condition of satisfying the first constraint, one of the first-level objective functions of the day-ahead operation optimization model is selected for solution to obtain the optimal actual power value of the electrolyzer for the selected first-level objective function. Under the premise of obtaining the optimal actual power value of the electrolyzer by solving the selected first-level objective function, based on the operation and maintenance cost of the current-day set of electrolyzers, the second-level objective function is solved from the perspective of current, under the condition of satisfying the first constraint, to obtain the optimal number of electrolyzers and the optimal start-up and shutdown state of the electrolyzers. Based on the actual power value of the optimal electrolytic cell, the optimal number of electrolytic cells, and the optimal start-up and shutdown status of the electrolytic cells, a daily plan for the electrolytic cells is generated. The first-level objective function includes a first-level objective function aiming to minimize system energy consumption, a first-level objective function aiming to minimize lifetime degradation, and a first-level objective function aiming to maximize economic benefits; the second-level objective function is constructed based on the operation and maintenance costs of the day-ahead set of electrolyzers, with the goal of maximizing total revenue; the first constraint conditions include one or more of the following: wind and solar power output constraints, grid connection and disconnection constraints, electrolyzer constraints, hydrogen storage tank constraints, and energy storage constraints; the electrolyzer constraints include: power constraints, temperature constraints, overload time constraints, ramp rate constraints, and operating state constraints; the operating state constraints include: operating state constraints and transfer state constraints.

[0217] Preferably, the electrolytic cell daily plan generation module includes: Based on the daytime wind and solar power output data, the predicted value of the daytime wind and solar power output is obtained through a pre-constructed intraday wind and solar power output prediction model; With a preset time interval as the scheduling cycle, when the first scheduling cycle arrives, starting from the daily plan of the electrolyzer and the predicted value of the daily wind and solar power output, and with the next predicted time domain as the optimization window, the intraday optimization model is solved under the condition of satisfying the second constraint to obtain the control signal of the first scheduling cycle. A feedback correction mechanism is adopted to feed back the daytime wind and solar power output value of the previous scheduling cycle to the intraday wind and solar power output prediction model, and correct it to obtain the intraday wind and solar power output prediction value of the next scheduling cycle. Based on the daily plan of the electrolyzer and the predicted daily wind and solar power output for each scheduling cycle, the daily tracking optimization is performed by rolling the solution of the daily optimization model to obtain the control signal for each scheduling cycle. When the number of scheduling rounds reaches the preset number, the optimal control signal sequence is obtained according to the control signal of each scheduling cycle, which serves as the daily plan of the electrolyzer. The intraday wind and solar power output prediction model is constructed based on historical operating data, including a distributed photovoltaic power output model, a wind power output model, an electrolysis hydrogen production system model, and an energy storage model, through discretization and difference equations.

[0218] Preferably, the system further includes: an intraday optimization model construction module; the intraday optimization model construction module is used for: A daily optimization objective function is constructed with the goal of minimizing the deviation between the daily plan and the intraday plan for the electrolyzer. A second constraint is set for the intraday optimization objective function; The second constraint includes one or more of the following: electrolyzer input power constraint, energy storage power constraint, hydrogen storage power constraint, and incremental constraint; The deviations include one or more of the following: energy consumption deviation, lifespan deviation, economic deviation, alkaline electrolyzer power deviation, proton exchange membrane electrolyzer power deviation, and energy storage power deviation.

[0219] Preferably, the collaborative control module is specifically used for: A fixed-type electrolytic reactor balanced operation method is adopted to coordinate the daily plan and intraday plan of the electrolytic cells, thereby optimizing the operation of the electrolytic cell cluster. Alternatively, a lifetime-based electrolytic reactor balanced operation method can be adopted to coordinate the daily and intraday plans of the electrolytic cells, thereby optimizing the operation of the electrolytic cell cluster. The fixed-type electrolytic reactor balancing operation method involves numbering the electrolytic cells according to their performance, dividing them into multiple operating scenarios based on the relationship between the total power of multiple electrolytic cells and the power threshold of each cell, allocating power accordingly, and periodically renumbering the cells according to the number of times each operating scenario occurs. The operating scenarios include: startup operation, optimal power operation, rated power operation, and overload operation. The lifetime-based electrolytic reactor balancing operation method uses the lifetime decay of each electrolytic cell as an optimization target, dynamically allocating power through day-ahead double-layer optimization and intraday rolling optimization to make the cumulative lifetime decay of each electrolytic cell tend to be consistent.

[0220] Example 3 like Figure 9 As shown, the present invention also provides an electronic device, which may be a computer device, a microcontroller device, a smart mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, processor, and transceiver component are connected via a bus; the memory can be used to store executable programs, and an exemplary executable program may include instructions; the processor is used to execute the instructions stored in the memory. The memory can also be used to store data, which can be accessed and / or modified when instructions are executed.

[0221] The processor may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and it is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions in the storage medium to implement the corresponding method flow or corresponding function, so as to realize the steps of the electrolytic cell cluster operation optimization method in the above embodiments.

[0222] Example 4 Based on the same inventive concept, this invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory). This readable storage medium is a memory device within an electronic device used to store programs and data. It is understood that the storage medium here can include both built-in storage media within the electronic device and extended storage media supported by the electronic device. The storage medium provides storage space, which stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more executable programs (including program code). It should be noted that the storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. Loading and executing one or more instructions stored in the storage medium by the processor can implement the steps of the electrolytic cell cluster operation optimization method described in the above embodiments.

[0223] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0224] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0225] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0226] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0227] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit its scope of protection. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that after reading the present invention, they can still make various changes, modifications or equivalent substitutions to the specific implementation methods of the application, but these changes, modifications or equivalent substitutions are all within the scope of protection of the claims pending approval.

Claims

1. A method for optimizing the operation of an electrolytic cell cluster, characterized in that, include: Based on the obtained wind and solar power output and the operating status of the electrolyzer cluster, multiple power allocation strategies are adopted from a power perspective to determine the day-ahead start-up and shutdown optimization strategy; The day-ahead start-stop optimization strategy includes a set of day-ahead electrolyzer counts and a set of day-ahead electrolyzer start-stop states. Based on the aforementioned day-ahead start-up and shutdown optimization strategy, the pre-built day-ahead operation optimization model is solved from the perspective of current to determine the day-ahead schedule of the electrolyzer. Based on the daily plan of the electrolyzer and the obtained intraday wind and solar power output forecast, an intraday optimization model is used to perform intraday rolling tracking optimization to obtain the intraday plan of the electrolyzer. A collaborative control method is adopted to coordinate the daily plan and intraday plan of the electrolyzers, thereby optimizing the operation of the electrolyzer cluster. The current-day operation optimization model includes a first-level objective function constructed with the objectives of minimizing system energy consumption, minimizing lifespan decay, and maximizing economic benefits, respectively. Under the premise that the first-level objective function is optimal, a second-level objective function is constructed with the objective of maximizing the total system benefit. The first constraint conditions are set for the first-level objective function and the second-level objective function with current density and temperature data as key parameters.

2. The method as described in claim 1, characterized in that, Based on the obtained wind and solar power output and the operating status of the electrolyzer cluster, the day-ahead start-up and shutdown optimization strategy is determined by employing multiple power allocation strategies from a power perspective, including: Based on the wind and solar power output and load demand of wind farms and photovoltaic power stations in the operating scenario, the set of day-ahead electrolyzer numbers is determined by using volatility assessment methods and intermittent assessment methods. Based on the set of electrolyzers for the day, a multi-type power allocation strategy is adopted from a power perspective to determine the power of the electrolyzers for the day. Based on the power of the electrolyzers for the day and a pre-constructed multi-state operation characterization model of the electrolyzers, the set of start-up and shutdown states of the electrolyzers for the day is determined. The set of the number of electrolyzers and the set of the start-up and shutdown states of the electrolyzers are used as the start-up and shutdown optimization strategy. The set of electrolyzer quantities for the current day includes: multiple combinations of electrolyzer quantities; the combination of electrolyzer quantities includes the current day alkaline electrolyzer quantity and the current day proton exchange membrane electrolyzer quantity; the multiple power allocation strategies include chain allocation strategy and average allocation strategy.

3. The method as described in claim 2, characterized in that, The multi-type power allocation strategy includes a chain allocation strategy and an average allocation strategy; the multi-type power allocation strategy is adopted based on the set of electrolyzer numbers for each day, and the power of the electrolyzers for each day is determined from a power perspective. Based on the power of the electrolyzers for each day and a pre-constructed multi-state operation characterization model of the electrolyzers, the set of start-up and shutdown states of the electrolyzers for each day is determined, including: Based on the day-ahead set of electrolyzers, the wind and solar power output, and the load demand, the chain allocation strategy is adopted. The alkaline electrolyzers in each electrolyzer combination are used as the basic load units, and the proton exchange membrane electrolyzers are used as the fluctuation response units. The alkaline electrolyzers are started at full capacity in a stepwise manner, and the proton exchange membrane electrolyzers take over the remaining power in units of maximum power per cell. The day-ahead electrolyzer power of each electrolyzer combination is obtained. Based on the day-ahead electrolyzer power and the multi-state operation characterization model of the electrolyzers, the start-up and shutdown status of each electrolyzer in each electrolyzer combination in the chain allocation strategy at each time is obtained. Based on the day-ahead set of electrolyzers, the wind and solar power output and load demand, the average allocation strategy is adopted. The alkaline electrolyzers in each electrolyzer combination are used as the basic load units, and the proton exchange membrane electrolyzers are used as the fluctuation response units. The alkaline electrolyzers are started at full capacity in a stepped manner, and the proton exchange membrane electrolyzers are started at the minimum safe power of a single cell. The remaining power of the started proton exchange membrane electrolyzers is evenly allocated to each electrolyzer combination to obtain the day-ahead electrolyzer power of each electrolyzer combination. Based on the day-ahead electrolyzer power and the multi-state operation characterization model of the electrolyzers, the start-up and shutdown status of each electrolyzer in each electrolyzer combination in the average allocation strategy at each time is obtained. Based on the start-up and shutdown status of each electrolytic cell in all combinations of the number of electrolytic cells at each time, a set of start-up and shutdown statuses of electrolytic cells for each day is generated.

4. The method as described in claim 2 or 3, characterized in that, The construction of the electrolyzer state operation characterization model includes: An electrolytic cell temperature model is constructed based on the heat generated by the fuel cell stack, the heat flowing into the fuel cell stack, the heat flowing out of the fuel cell stack, and the heat dissipation from the environment. Based on the training data of the electrolytic cell parameter identification model, an electrolytic cell parameter identification model is constructed by training a deep learning artificial intelligence model architecture. Based on the electrolytic cell temperature model, the electrolytic cell parameter identification model, and the internal and external equipment of the electrolytic cell, the electrolytic cell identification parameters and operating equipment are determined. Combined with the electrolytic cell operating state mapping function, the operating state of the electrolytic cell is determined, and the construction of the multi-state operating characterization model of the electrolytic cell is completed. The deep learning artificial intelligence model architecture includes an attention mechanism, an encoder layer, and a decoder layer; the training data includes one or more of the following: historical electrolyzer inlet temperature, historical electrolyzer outlet temperature, historical hydrogen-side pressure, historical oxygen-side pressure, historical water-side pressure, historical hydrogen in oxygen, historical oxygen in hydrogen, historical conductivity, and historical reactor feed water flow rate; the internal and external equipment of the electrolyzer includes one or more of the following: circulating pump, makeup water pump, heater, pneumatic valve, gas separator, pure water machine, cooling system, instrument gas supply system, and current rectifier; the operating state includes one or more of the following: production state, thermal state, idle state, and fault state.

5. The method as described in claim 4, characterized in that, The process involves determining the electrolytic cell identification parameters and operating equipment based on the electrolytic cell temperature model, the electrolytic cell parameter identification model, and the internal and external equipment of the electrolytic cell. Then, by combining this with the electrolytic cell operating state mapping function, the operating state of the electrolytic cell is determined, thus completing the construction of a multi-state operating characterization model for the electrolytic cell. This includes: The electrolyzer temperature model and electrolyzer parameter identification model that have been constructed are used to determine the electrolyzer identification parameters; The operating equipment is determined based on the operating status of the internal and external equipment of the electrolytic cell; Based on the electrolytic cell identification parameters and the operating equipment, the operating state of the electrolytic cell is determined using the electrolytic cell operating state characterization mapping function, thus completing the construction of a multi-state operating characterization model for the electrolytic cell.

6. The method as described in claim 4, characterized in that, In the aforementioned thermodynamic state, a dynamic and accurate characterization model of the thermodynamic state is constructed based on current density, electrolytic cell temperature, and electrolytic cell state switching parameters. The expression for the dynamic and precise representation model is as follows: In the formula, To be at current density and electrolytic cell temperature Under the condition of a hot engine, For conceptual functions, To be at current density and electrolytic cell temperature The set of parameters for identifying the lower electrolytic cell. To be at current density and electrolytic cell temperature The lower electrolytic cell contains a collection of internal and external equipment.

7. The method as described in claim 1, characterized in that, The step of solving the pre-built daily operation optimization model from the perspective of current based on the day-to-day start-up and shutdown optimization strategy to determine the daily schedule of the electrolyzer includes: Based on the day-ahead start-stop optimization strategy, from the perspective of current, under the condition of satisfying the first constraint, one of the first-level objective functions of the day-ahead operation optimization model is selected for solution to obtain the optimal actual power value of the electrolyzer for the selected first-level objective function. Under the premise of obtaining the optimal actual power value of the electrolyzer by solving the selected first-level objective function, based on the operation and maintenance cost of the current-day set of electrolyzers, the second-level objective function is solved from the perspective of current, under the condition of satisfying the first constraint, to obtain the optimal number of electrolyzers and the optimal start-up and shutdown state of the electrolyzers. Based on the actual power value of the optimal electrolytic cell, the optimal number of electrolytic cells, and the optimal start-up and shutdown status of the electrolytic cells, a daily plan for the electrolytic cells is generated. The first-level objective function includes a first-level objective function aiming to minimize system energy consumption, a first-level objective function aiming to minimize lifetime degradation, and a first-level objective function aiming to maximize economic benefits; the second-level objective function is constructed based on the operation and maintenance costs of the day-ahead set of electrolyzers, with the goal of maximizing total revenue; the first constraint conditions include one or more of the following: wind and solar power output constraints, grid connection and disconnection constraints, electrolyzer constraints, hydrogen storage tank constraints, and energy storage constraints; the electrolyzer constraints include: power constraints, temperature constraints, overload time constraints, ramp rate constraints, and operating state constraints; the operating state constraints include: operating state constraints and transfer state constraints.

8. The method as described in claim 1, characterized in that, The process involves using an intraday optimization model to perform rolling tracking optimization based on the day-ahead plan of the electrolyzer and the obtained intraday wind and solar power output forecasts, resulting in the intraday plan for the electrolyzer, including: Based on the daytime wind and solar power output data, the predicted value of the daytime wind and solar power output is obtained through a pre-constructed intraday wind and solar power output prediction model; With a preset time interval as the scheduling cycle, when the first scheduling cycle arrives, starting from the daily plan of the electrolyzer and the predicted value of the daily wind and solar power output, and with the next predicted time domain as the optimization window, the intraday optimization model is solved under the condition of satisfying the second constraint to obtain the control signal of the first scheduling cycle. A feedback correction mechanism is adopted to feed back the predicted value of daily wind and solar power output of the previous scheduling cycle to the predicted value of daily wind and solar power output of the next scheduling cycle. Based on the daily plan of the electrolyzer and the predicted daily wind and solar power output for each scheduling cycle, the daily tracking optimization is performed by rolling the solution of the daily optimization model to obtain the control signal for each scheduling cycle. When the number of scheduling rounds reaches the preset number, the optimal control signal sequence is obtained according to the control signal of each scheduling cycle, which serves as the daily plan of the electrolyzer. The intraday wind and solar power output prediction model is constructed based on historical operating data, including a distributed photovoltaic power output model, a wind power output model, an electrolysis hydrogen production system model, and an energy storage model, through discretization and difference equations.

9. The method as described in claim 1 or 8, characterized in that, The construction of the intraday optimization model includes: A daily optimization objective function is constructed with the goal of minimizing the deviation between the daily plan and the intraday plan for the electrolyzer. A second constraint is set for the intraday optimization objective function; The second constraint includes one or more of the following: electrolyzer input power constraint, energy storage power constraint, hydrogen storage power constraint, and incremental constraint; The deviations include one or more of the following: energy consumption deviation, lifespan deviation, economic deviation, alkaline electrolyzer power deviation, proton exchange membrane electrolyzer power deviation, and energy storage power deviation.

10. The method as described in claim 1, characterized in that, The method of employing a collaborative control approach to coordinate the daily and intraday plans of the electrolytic cells, thereby optimizing the operation of the electrolytic cell cluster, includes: A fixed-type electrolytic reactor balanced operation method is adopted to coordinate the daily plan and intraday plan of the electrolytic cells, thereby optimizing the operation of the electrolytic cell cluster. Alternatively, a lifetime-based electrolytic reactor balanced operation method can be adopted to coordinate the daily and intraday plans of the electrolytic cells, thereby optimizing the operation of the electrolytic cell cluster. The fixed-type electrolytic reactor balancing operation method involves numbering the electrolytic cells according to their performance, dividing them into multiple operating scenarios based on the relationship between the total power of multiple electrolytic cells and the power threshold of each cell, allocating power accordingly, and periodically renumbering the cells based on the number of operating conditions in each scenario for rotation. The operating scenarios include: startup operation, optimal power operation, rated power operation, and overload operation. The lifetime-based electrolytic reactor balancing operation method uses the lifetime decay of each electrolytic cell as an optimization target, dynamically allocating power through day-ahead double-layer optimization and intraday rolling optimization to make the cumulative lifetime decay of each electrolytic cell tend to be consistent.

11. An electrolytic cell cluster operation optimization system, characterized in that, include: The module includes a day-ahead start / stop optimization module, a day-ahead plan generation module for electrolytic cells, an intraday plan generation module for electrolytic cells, and a collaborative control module. Based on the obtained wind and solar power output and the operating status of the electrolyzer cluster, multiple power allocation strategies are adopted from a power perspective to determine the day-ahead start-up and shutdown optimization strategy; The day-ahead start-stop optimization strategy includes a set of day-ahead electrolyzer counts and a set of day-ahead electrolyzer start-stop states. Based on the aforementioned day-ahead start-up and shutdown optimization strategy, the pre-built day-ahead operation optimization model is solved from the perspective of current to determine the day-ahead schedule of the electrolyzer. Based on the daily plan of the electrolyzer and the obtained intraday wind and solar power output forecast, an intraday optimization model is used to perform intraday rolling tracking optimization to obtain the intraday plan of the electrolyzer. A collaborative control method is adopted to coordinate the daily plan and intraday plan of the electrolyzers, thereby optimizing the operation of the electrolyzer cluster. The current-day operation optimization model includes a first-level objective function constructed with the objectives of minimizing system energy consumption, minimizing lifespan decay, and maximizing economic benefits, respectively. Under the premise that the first-level objective function is optimal, a second-level objective function is constructed with the objective of maximizing the total system benefit. The first constraint conditions are set for the first-level objective function and the second-level objective function with current density and temperature data as key parameters.

12. The system as claimed in claim 11, characterized in that, The aforementioned start / stop optimization module includes: The day-ahead electrolyzer quantity determination submodule is used to determine the set of day-ahead electrolyzer quantities based on the wind and solar power output and load demand of wind farms and photovoltaic power plants in the operating scenario, using volatility assessment methods and intermittent assessment methods. The daytime electrolyzer start-up / shutdown state determination submodule is used to determine the daytime electrolyzer power based on the set of daytime electrolyzer quantities and adopt a multi-type power allocation strategy from a power perspective. Based on the daytime electrolyzer power and a pre-built multi-state operation characterization model of the electrolyzer, the daytime electrolyzer start-up / shutdown state set is determined. The day-ahead start-stop optimization strategy generation submodule is used to take the day-ahead electrolyzer quantity set and the day-ahead electrolyzer start-stop state set as the day-ahead start-stop optimization strategy; The set of electrolyzer quantities for the current day includes: multiple combinations of electrolyzer quantities; the combination of electrolyzer quantities includes the current day alkaline electrolyzer quantity and the current day proton exchange membrane electrolyzer quantity; the multiple power allocation strategies include chain allocation strategy and average allocation strategy.

13. The system as described in claim 12, characterized in that, The multi-type power allocation strategy includes a chain allocation strategy and an average allocation strategy; the day-ahead electrolytic cell start / stop status determination submodule is specifically used for: Based on the day-ahead set of electrolyzers, the wind and solar power output, and the load demand, the chain allocation strategy is adopted. The alkaline electrolyzers in each electrolyzer combination are used as the basic load units, and the proton exchange membrane electrolyzers are used as the fluctuation response units. The alkaline electrolyzers are started at full capacity in a stepwise manner, and the proton exchange membrane electrolyzers take over the remaining power in units of maximum power per cell. The day-ahead electrolyzer power of each electrolyzer combination is obtained. Based on the day-ahead electrolyzer power and the multi-state operation characterization model of the electrolyzers, the start-up and shutdown status of each electrolyzer in each electrolyzer combination in the chain allocation strategy at each time is obtained. Based on the day-ahead set of electrolyzers, the wind and solar power output and load demand, the average allocation strategy is adopted. The alkaline electrolyzers in each electrolyzer combination are used as the basic load units, and the proton exchange membrane electrolyzers are used as the fluctuation response units. The alkaline electrolyzers are started at full capacity in a stepped manner, and the proton exchange membrane electrolyzers are started at the minimum safe power of a single cell. The remaining power of the started proton exchange membrane electrolyzers is evenly allocated to each electrolyzer combination to obtain the day-ahead electrolyzer power of each electrolyzer combination. Based on the day-ahead electrolyzer power and the multi-state operation characterization model of the electrolyzers, the start-up and shutdown status of each electrolyzer in each electrolyzer combination in the average allocation strategy at each time is obtained. Based on the start-up and shutdown status of each electrolytic cell in all combinations of the number of electrolytic cells at each time, a set of start-up and shutdown statuses of electrolytic cells for each day is generated.

14. The system as described in claim 12 or 13, characterized in that, The system further includes an electrolytic cell state operation characterization model construction module; the electrolytic cell state operation characterization model construction module includes: The electrolytic cell temperature model construction submodule is used to construct the electrolytic cell temperature model based on the heat generated by the fuel cell stack, the heat flowing into the fuel cell stack, the heat flowing out of the fuel cell stack, and the heat dissipation from the environment. The electrolytic cell parameter identification model construction submodule is used to construct the electrolytic cell parameter identification model based on the training data of the electrolytic cell parameter identification model by training a deep learning artificial intelligence model architecture. The electrolytic cell operation status characterization submodule is used to determine the electrolytic cell identification parameters and operating equipment based on the electrolytic cell temperature model, the electrolytic cell parameter identification model, and the internal and external equipment of the electrolytic cell. Combined with the electrolytic cell operation status mapping function, the operation status of the electrolytic cell is determined, and the construction of the multi-state operation characterization model of the electrolytic cell is completed. The deep learning artificial intelligence model architecture includes an attention mechanism, an encoder layer, and a decoder layer; the training data includes one or more of the following: historical electrolyzer inlet temperature, historical electrolyzer outlet temperature, historical hydrogen-side pressure, historical oxygen-side pressure, historical water-side pressure, historical hydrogen in oxygen, historical oxygen in hydrogen, historical conductivity, and historical reactor feed water flow rate; the internal and external equipment of the electrolyzer includes one or more of the following: circulating pump, makeup water pump, heater, pneumatic valve, gas separator, pure water machine, cooling system, instrument gas supply system, and current rectifier; the operating state includes one or more of the following: production state, thermal state, idle state, and fault state.

15. The system as described in claim 14, characterized in that, The electrolytic cell operating status characterization submodule is specifically used for: The electrolyzer temperature model and electrolyzer parameter identification model that have been constructed are used to determine the electrolyzer identification parameters; The operating equipment is determined based on the operating status of the internal and external equipment of the electrolytic cell; Based on the electrolytic cell identification parameters and the operating equipment, the operating state of the electrolytic cell is determined using the electrolytic cell operating state characterization mapping function, thus completing the construction of a multi-state operating characterization model for the electrolytic cell.

16. The system as described in claim 14, characterized in that, In the aforementioned thermodynamic state, a dynamic and accurate characterization model of the thermodynamic state is constructed based on current density, electrolytic cell temperature, and electrolytic cell state switching parameters. The expression for the dynamic and precise representation model is as follows: In the formula, To be at current density and electrolytic cell temperature Under the condition of a hot engine, For conceptual functions, To be at current density and electrolytic cell temperature The set of parameters for identifying the lower electrolytic cell. To be at current density and electrolytic cell temperature The lower electrolytic cell contains a collection of internal and external equipment.

17. The system as claimed in claim 11, characterized in that, The electrolytic cell day-ahead planning generation module is specifically used for: Based on the day-ahead start-stop optimization strategy, from the perspective of current, under the condition of satisfying the first constraint, one of the first-level objective functions of the day-ahead operation optimization model is selected for solution to obtain the optimal actual power value of the electrolyzer for the selected first-level objective function. Under the premise of obtaining the optimal actual power value of the electrolyzer by solving the selected first-level objective function, based on the operation and maintenance cost of the current-day set of electrolyzers, the second-level objective function is solved from the perspective of current, under the condition of satisfying the first constraint, to obtain the optimal number of electrolyzers and the optimal start-up and shutdown state of the electrolyzers. Based on the actual power value of the optimal electrolytic cell, the optimal number of electrolytic cells, and the optimal start-up and shutdown status of the electrolytic cells, a daily plan for the electrolytic cells is generated. The first-level objective function includes a first-level objective function aiming to minimize system energy consumption, a first-level objective function aiming to minimize lifetime degradation, and a first-level objective function aiming to maximize economic benefits; the second-level objective function is constructed based on the operation and maintenance costs of the day-ahead set of electrolyzers, with the goal of maximizing total revenue; the first constraint conditions include one or more of the following: wind and solar power output constraints, grid connection and disconnection constraints, electrolyzer constraints, hydrogen storage tank constraints, and energy storage constraints; the electrolyzer constraints include: power constraints, temperature constraints, overload time constraints, ramp rate constraints, and operating state constraints; the operating state constraints include: operating state constraints and transfer state constraints.

18. The system as claimed in claim 11, characterized in that, The electrolytic cell daily plan generation module includes: Based on the daytime wind and solar power output data, the predicted value of the daytime wind and solar power output is obtained through a pre-constructed intraday wind and solar power output prediction model; With a preset time interval as the scheduling cycle, when the first scheduling cycle arrives, starting from the daily plan of the electrolyzer and the predicted value of the daily wind and solar power output, and with the next predicted time domain as the optimization window, the intraday optimization model is solved under the condition of satisfying the second constraint to obtain the control signal of the first scheduling cycle. A feedback correction mechanism is adopted to feed back the predicted value of daily wind and solar power output of the previous scheduling cycle to the predicted value of daily wind and solar power output of the next scheduling cycle. Based on the daily plan of the electrolyzer and the predicted daily wind and solar power output for each scheduling cycle, the daily tracking optimization is performed by rolling the solution of the daily optimization model to obtain the control signal for each scheduling cycle. When the number of scheduling rounds reaches the preset number, the optimal control signal sequence is obtained according to the control signal of each scheduling cycle, which serves as the daily plan of the electrolyzer. The intraday wind and solar power output prediction model is constructed based on historical operating data, including a distributed photovoltaic power output model, a wind power output model, an electrolysis hydrogen production system model, and an energy storage model, through discretization and difference equations.

19. The system as described in claim 11 or 18, characterized in that, The system further includes: an intraday optimization model construction module; the intraday optimization model construction module is used for: A daily optimization objective function is constructed with the goal of minimizing the deviation between the daily plan and the intraday plan for the electrolyzer. A second constraint is set for the intraday optimization objective function; The second constraint includes one or more of the following: electrolyzer input power constraint, energy storage power constraint, hydrogen storage power constraint, and incremental constraint; The deviations include one or more of the following: energy consumption deviation, lifespan deviation, economic deviation, alkaline electrolyzer power deviation, proton exchange membrane electrolyzer power deviation, and energy storage power deviation.

20. The system as claimed in claim 11, characterized in that, The collaborative control module is specifically used for: A fixed-type electrolytic reactor balanced operation method is adopted to coordinate the daily plan and intraday plan of the electrolytic cells, thereby optimizing the operation of the electrolytic cell cluster. Alternatively, a lifetime-based electrolytic reactor balanced operation method can be adopted to coordinate the daily and intraday plans of the electrolytic cells, thereby optimizing the operation of the electrolytic cell cluster. The fixed-type electrolytic reactor balancing operation method involves numbering the electrolytic cells according to their performance, dividing them into multiple operating scenarios based on the relationship between the total power of multiple electrolytic cells and the power threshold of each cell, allocating power accordingly, and periodically renumbering the cells based on the number of operating conditions in each scenario for rotation. The operating scenarios include: startup operation, optimal power operation, rated power operation, and overload operation. The lifetime-based electrolytic reactor balancing operation method uses the lifetime decay of each electrolytic cell as an optimization target, dynamically allocating power through day-ahead double-layer optimization and intraday rolling optimization to make the cumulative lifetime decay of each electrolytic cell tend to be consistent.

21. An electronic device, characterized in that, include: At least one processor and memory; The memory and processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, an electrolytic cell cluster operation optimization method as described in any one of claims 1 to 10 is implemented.

22. A readable storage medium, characterized in that, It contains an executable program, which, when executed, implements an electrolytic cell cluster operation optimization method as described in any one of claims 1 to 10.

Images