A source-grid hydrogen-ammonia coordinated hydrogen energy load side grid interaction control method

Abstract

This invention discloses a method for coordinated control of hydrogen energy load-side grid interaction in a source-grid manner. First, a hydrogen-ammonia load aggregation model is constructed, comprising a dynamic model of electro-thermal-mass coupling of the water electrolysis hydrogen production unit, a mass balance model of the hydrogen buffer unit, and a thermal inertia response model of the ammonia synthesis unit. Based on this, a multi-timescale hierarchical control architecture consisting of a scheduling layer, a coordination layer, and a local layer is established to achieve decoupled coordination between dynamic tracking and command execution. A cross-timescale coordinated control strategy based on a dynamic safety domain is designed, calculating the maximum adjustable power range in real time, and developing differentiated coordination logic for scenarios such as rapid grid frequency regulation, new energy fluctuation smoothing, and long-term peak shaving. Under the premise of strictly adhering to process constraints such as electrolyzer temperature, ramp rate, hydrogen storage tank pressure, and ammonia feed flow rate and temperature change rate, this invention significantly improves the regulation accuracy and response speed of hydrogen energy load-side grid interaction, ensuring the safety of chemical processes and equipment lifespan.

Description

Technical Field

[0001] This invention belongs to the field of power system operation and control, and particularly relates to a method for interactive control of hydrogen energy load-side power grid in a source-grid hydrogen-ammonia coordinated manner. Background Technology

[0002] With the continuous expansion of new energy power generation capacity, the demand for flexible adjustment resources in the power system is becoming increasingly urgent. The water electrolysis-to-hydrogen coupled ammonia synthesis system, as a large-scale adjustable power load, can convert fluctuating renewable energy into hydrogen and ammonia for storage or utilization. It features large response capacity and high adjustment potential, and is considered an important resource for participating in grid peak shaving, frequency regulation, and providing ancillary services. In recent years, research on the integration of the water electrolysis-to-hydrogen-to-ammonia synthesis system into the power system as a power load has gradually increased. Related results mainly focus on resource allocation and capacity planning, and some research has also been conducted on the system's participation in power-side energy storage from an energy management perspective.

[0003] However, most current analytical methods have low time resolution, relying primarily on steady-state or quasi-steady-state models and typically using fixed efficiency coefficients to describe the conversion relationships between electrical energy, hydrogen energy, and ammonia energy. While this simplified approach facilitates long-term planning or energy balance analysis, it fails to characterize the impact of dynamic process constraints on system control characteristics, such as temperature changes during electrolyzer startup, thermal inertia of the ammonia synthesis reactor, and pressure fluctuations in the gas buffer zone. In reality, the processes of water electrolysis for hydrogen production and ammonia synthesis exhibit significant multi-timescale characteristics. Fluctuations on the power side occur on the order of seconds to minutes, while the thermodynamic and material balances of chemical processes demonstrate inertial responses on the order of minutes to hours. When the system participates in grid interaction as a flexible load, using only steady-state models for control decisions will fail to accurately predict the actual regulatory capacity boundaries of the system during dynamic processes and will also make it difficult to coordinate the coupling relationships between devices with different timescales (fast and slow). This limitation makes it difficult for existing technologies to effectively solve the technical problems of cross-timescale balance matching and overall system load balancing control. It fails to fully unleash the system's regulatory potential and may also cause operational safety issues such as excessive thermal stress in the electrolyzer and damage to the ammonia synthesis catalyst when responding quickly to grid commands.

[0004] Therefore, how to establish a refined model that can reflect the dynamic process characteristics of the water electrolysis hydrogen production and ammonia synthesis system, and on this basis design a control method that takes into account the coordination of multiple time scales, so that it can participate in grid interaction safely and flexibly as an electrical load, has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method for interactive control of the hydrogen energy load-side power grid in a source-grid hydrogen-ammonia coordinated manner, the details of which are as follows: A method for source-grid-hydrogen-ammonia coordinated control of hydrogen energy load-side grid interaction includes the following steps: S1. Construct a hydrogen-ammonia loading polymerization model that includes dynamic process constraints, wherein the hydrogen-ammonia loading polymerization model includes: An electro-thermal-mass coupled dynamic model of a water electrolysis hydrogen production unit is used to characterize the nonlinear behavior of the electrolyzer under different power levels and temperature conditions, and to extract the operating power range constraints, ramp rate constraints, start-up and shutdown time constraints, and temperature limits of the electrolyzer. A dynamic model of the hydrogen buffer unit is used to describe the dynamic relationship between hydrogen storage and pressure in the hydrogen storage tank, and to extract the pressure limit, flow rate change constraint and minimum / maximum flow constraint of the hydrogen storage tank. The lumped parameter dynamic model of the ammonia synthesis unit is used to describe the dynamic response relationship between the feed hydrogen flow rate and the output ammonia production, and to extract the minimum feed flow rate constraint, temperature change rate constraint, temperature operating range constraint and pressure fluctuation constraint of the ammonia synthesis unit. S2. Establish a multi-timescale hierarchical control architecture for source-grid hydrogen-ammonia coordination, including: The scheduling layer, with a control cycle of minutes to hours, is used to receive grid scheduling plans and formulate the total power plan of the electrolyzer cluster and the hydrogen charge state reference trajectory of the hydrogen storage tank with the goal of minimizing system operating costs. The coordination layer, with a control cycle of seconds to minutes, receives the planned trajectory issued by the scheduling layer, uses model predictive control algorithms to track power grid regulation commands, and coordinates the power distribution between the electrolyzer cluster and the ammonia synthesis unit under the dynamic process constraints. The local layer, with an execution cycle of milliseconds to seconds, is used to execute instructions issued by the coordination layer, collect status data and feed it back to the coordination layer, and execute safety interlock protection logic. S3. Design a cross-timescale coordinated control strategy based on dynamic security domains, including: The coordination layer calculates the dynamic security domain in real time during each control cycle. The dynamic security domain includes the maximum adjustable power range in which the system can safely respond in the short future time domain and its duration. Based on the comparison results between the grid regulation command and the dynamic safety domain, the corresponding coordinated control logic is executed. This includes adjusting the electrolyzer power to track the command and dynamically adjusting the hydrogen charge state reference trajectory of the hydrogen storage tank to absorb hydrogen production fluctuations when the command value is within the dynamic safety domain. When the command value exceeds the dynamic safety domain, the response is based on the safety boundary and information on insufficient regulation capacity is generated and uploaded to the scheduling layer.

[0006] Furthermore, the electro-thermal-mass coupling dynamic model of the water electrolysis hydrogen production unit includes: The electrolytic cell model is used to describe the relationship between the total voltage and current density of the electrolyzer. The thermal model uses the lumped parameter method to establish thermal dynamic equations to describe the temperature changes in the electrolytic cell; An efficiency characteristic model is established by fitting offline experimental data to create a two-dimensional interpolation table between power, temperature and efficiency. The operating power P of the electrolytic cell EL The interval constraint is P EL,min (T EL ) ≤ P EL ≤ P EL,max , where P EL,max For rated power, P EL,min (T EL The minimum stable operating power is ); the ramp rate constraint is . ,in, This represents the change in the active power of the electrolytic cell. This indicates the limit for the ramp rate of the electrolytic cell. This indicates the time interval corresponding to the power change; the temperature limit is the value that the electrolytic cell operating temperature is maintained within a safe range.

[0007] Furthermore, the dynamic model of the hydrogen buffer unit includes: The state variables of the hydrogen storage tank are defined using hydrogen quantity and pressure as state variables, defining the hydrogen charge state. The ideal gas law is used to describe the relationship between the hydrogen pressure and the amount of hydrogen in the hydrogen storage tank. The mass balance dynamic equation is used to describe the relationship between the change in hydrogen inventory in the hydrogen storage tank and the inflow and outflow rates. The pressure limit of the hydrogen storage tank is... ,in, This indicates the minimum permissible operating pressure of the hydrogen storage tank. This indicates the maximum permissible operating pressure of the hydrogen storage tank. This indicates the real-time operating pressure of the hydrogen storage tank; the flow rate change constraint includes the flow rate change constraint of the inlet compressor and the flow rate change constraint of the outlet pressure regulating valve.

[0008] Furthermore, the lumped parameter dynamic model of the ammonia synthesis unit includes: The dynamic model for ammonia production uses a first-order inertial element plus a pure time delay to describe the dynamic relationship between the feed hydrogen flow rate and the liquid ammonia production. A reactor temperature dynamic model is used to describe the balance between reactor temperature, ammonia production, and heat of reaction. The minimum feed flow rate constraint for the ammonia synthesis unit is: The temperature change rate constraint is The temperature operating range is constrained as follows: The pressure fluctuation constraint is ,in, This indicates the feed hydrogen flow rate into the ammonia synthesis unit. This is the minimum allowable flow rate of the feed hydrogen. This indicates the average temperature of the ammonia synthesis reactor. This represents the rate of change of reactor temperature over time. This indicates the upper limit of the allowable rate of change of reactor temperature. and These represent the minimum and maximum allowable temperatures for the ammonia synthesis reactor, respectively. This indicates the real-time temperature of the reactor. This represents the rate of change of pressure over time. This represents the upper limit of the allowable rate of pressure change.

[0009] Furthermore, the scheduling layer adopts a model predictive control framework to continuously optimize the operation plan for the next 4 hours with a control cycle of 15 minutes. The optimization objectives include electricity cost, electrolyzer aging cost, and hydrogen storage tank hydrogen charge state deviation penalty term. The constraints include electrolyzer power constraints, electrolyzer ramp-up constraints, hydrogen storage tank hydrogen charge state constraints, and ammonia synthesis unit association constraints.

[0010] Furthermore, the coordination layer adopts a model predictive control algorithm with a control cycle of 1 minute. Its internal predictive model adopts the joint dynamic model of the electrolyzer, hydrogen storage tank and ammonia synthesis unit established in step S1, which is represented in discrete-time state-space form. The optimization objectives include grid command tracking error penalty, control quantity change penalty and constraint over-limit soft penalty. The constraints include electrolyzer power limit, electrolyzer ramp-up constraint, hydrogen storage tank pressure constraint, ammonia synthesis feed flow constraint, ammonia synthesis feed flow rate change constraint, ammonia synthesis temperature constraint and power balance constraint.

[0011] Furthermore, the calculation method for the dynamic safety domain in step S3 is as follows: the coordination layer calculates the maximum adjustable power range in the future short time domain based on the dynamic model established in step S1 and the real-time status fed back by the local layer in the prediction time domain of 5 minutes. The upward adjustment capability is taken as the minimum value of the adjustment capability limited by the electrolyzer cluster ramping capability constraint, the upper limit constraint of the hydrogen storage tank pressure constraint, and the electrolyzer temperature constraint. The downward adjustment capability is taken as the minimum value of the adjustment capability limited by the electrolyzer cluster downward ramping capability constraint, the lower limit constraint of the hydrogen storage tank pressure constraint, and the minimum operating power constraint of the electrolyzer.

[0012] Furthermore, the coordination control logic described in step S3 includes: When the power grid issues a fast frequency regulation command, if the command value is within the dynamic safety domain, the coordination layer model predicts and controls the power tracking command of the electrolyzer cluster, and dynamically adjusts the hydrogen charge state reference trajectory of the hydrogen storage tank to absorb hydrogen production fluctuations; if the command value exceeds the dynamic safety domain, it responds according to the safety boundary and generates information on insufficient regulation capacity, which is then uploaded to the scheduling layer. When local renewable energy power fluctuates, the dispatch layer sends a fluctuation absorption instruction to the coordination layer. The coordination layer's model predictive control tracks the dynamically adjusted target value by adjusting the electrolyzer power and proactively adjusts the feed flow of the ammonia synthesis unit to reserve buffer space. When the power grid issues a long-term peak-shaving command, the dispatching layer converts the peak-shaving plan into a reference trajectory of the total power of the electrolyzer cluster and a target trajectory of the hydrogen charge state of the hydrogen storage tank, and sends it to the coordination layer. The coordination layer uses model prediction and control to prioritize the adjustment of the electrolyzer power and simultaneously and gradually adjust the feed flow rate of the ammonia synthesis unit, constrain the rate of change of the ammonia synthesis feed flow rate, and control the hydrogen charge state of the hydrogen storage tank to change according to the target trajectory.

[0013] Furthermore, the local layer is deployed in the field controller, including power modules, compressor frequency converters, regulating valve positioners and sensors, which collect data with a sampling period of 100ms to 1s, communicate with the coordination layer through industrial Ethernet or fieldbus, and independently execute safety interlock protection logic including electrolyzer temperature over-limit, hydrogen storage tank pressure over-limit, and ammonia synthesis unit temperature and pressure over-limit.

[0014] Compared with the prior art, the present invention has at least the following beneficial effects: This invention addresses the common problems in existing technologies regarding the interaction of water electrolysis hydrogen production coupled with ammonia synthesis systems with the power grid, such as coarse-grained models, inaccurate dynamic responses, and difficulties in coordination across time scales. It proposes a systematic solution. Unlike traditional methods that use fixed efficiency coefficients and simplified approaches based primarily on steady-state or quasi-steady-state models, this invention first constructs a refined hydrogen-ammonia load polymerization model that includes a dynamic model of the electro-thermal-mass coupling of the water electrolysis hydrogen production unit, a mass balance model of the hydrogen buffer unit, and a thermal inertia response model of the ammonia synthesis unit. This model accurately characterizes key dynamic process constraints such as temperature changes during electrolyzer startup, thermal inertia of the ammonia synthesis reactor, and pressure fluctuations in the gas buffer zone, thus realistically reflecting the actual regulatory capacity boundaries of the system under different operating conditions.

[0015] Building upon this foundation, this invention further establishes a multi-timescale hierarchical control architecture comprising a scheduling layer, a coordination layer, and a local layer. This architecture decouples and coordinates minute-to-hour-level economic optimization, second-to-minute-level dynamic tracking and constraint management, and millisecond-to-second-level instruction execution, effectively resolving the contradiction that a single control layer cannot simultaneously address rapid grid response and the stable operation of slow chemical processes. More importantly, this invention designs a cross-timescale coordinated control strategy based on a dynamic safety domain. By calculating in real-time the system's maximum adjustable power range and its duration in the short-term future time domain, and combining this with typical scenarios such as rapid grid frequency regulation, new energy fluctuation mitigation, and long-term peak shaving, differentiated coordinated control logic is formulated. This enables the system to flexibly respond to grid commands and fully leverage its regulatory potential while strictly adhering to process constraints such as electrolyzer temperature limits, ramp-up rates, start-up and shutdown time constraints, hydrogen storage tank pressure safety ranges, and ammonia feed flow rate and temperature change rate.

[0016] In summary, this invention, through the organic combination of refined dynamic modeling, hierarchical collaborative control, and dynamic safety domain constraint management, not only significantly improves the regulation accuracy and response speed of the water electrolysis hydrogen production and ammonia synthesis system as a flexible power load participating in grid interaction, but also effectively ensures the operational safety and service life of chemical process equipment, achieving coordinated optimization and win-win results between the power system and the chemical system at multiple time scales. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0018] Figure 1 This is a schematic diagram of the technical framework provided by an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, other embodiments obtained by those skilled in the art without creative effort are all within the scope of protection of the present invention.

[0020] This embodiment provides a source-grid hydrogen-ammonia coordinated control method for the hydrogen energy load-side grid interaction. This method is applied to a system comprising a 20MW wind power generation, 10MW photovoltaic power generation, a 5MW alkaline water electrolysis hydrogen production system (including four 1.25MW electrolyzers), and a 2000Nm³ / h grid. 3The system comprises hydrogen storage tanks (pressure range 1.6-3.2 MPa) and a Haber-Bosch ammonia synthesis plant with an annual capacity of 1,000 tons. Connected to the regional power grid via a grid connection point, this system can both absorb local renewable energy and serve as a flexible load in response to grid dispatch commands.

[0021] like Figure 1 As shown, the control method in this embodiment mainly includes the following steps: Step 1: Construct a hydrogen ammonia loading polymerization model with dynamic process constraints. Since the processes of hydrogen production from water electrolysis and ammonia synthesis are inherently highly nonlinear and inertial chemical dynamic processes, traditional steady-state models using fixed efficiency coefficients cannot describe their true adjustable boundaries. Therefore, this step first establishes a refined aggregate model capable of characterizing the dynamic response of the system, providing an accurate predictive basis for subsequent control.

[0022] 1. Dynamic model of water electrolysis hydrogen production unit In this embodiment, the water electrolysis hydrogen production unit adopts alkaline water electrolysis hydrogen production technology, comprising four alkaline electrolyzers with a rated power of 1.25MW. To accurately describe its dynamic response characteristics, this embodiment establishes a dynamic mechanism model of electro-thermal-mass coupling. This model consists of three parts: an electroelectronic model, a thermal model, and an efficiency characteristic model, which can characterize the nonlinear behavior of the electrolyzer under different power levels and temperature conditions.

[0023] (1) Electrical sub-model The relationship between the total voltage and current density of an alkaline electrolyzer can be simplified to the following form:

[0024] in, U stack The total voltage of the electrolytic cell (V); N cell This represents the number of electrolytic cells connected in series. R The ideal gas constant is taken as 8.314 J / (mol·K); T The operating temperature of the electrolytic cell (K); α This is the charge transfer coefficient, typically ranging from 0.5 to 0.7. F The value is Faraday's constant, taken as 96485 C / mol; j Current density (A / m) 2 ); j 0 represents the exchange current density (A / m). 2 ), which increases with increasing temperature; r ohm Ω·m² is the ohmic resistivity; m and n are empirical coefficients used to fit the diffusion overpotential.

[0025] Active power of the electrolytic cell P EL With DC current I stack The relationship is:

[0026] in, , A cell The effective area of ​​a single electrolytic cell (m²) 2 ).

[0027] (2) Thermal model During operation, the electrolytic cell generates Joule heat, while the electrochemical reaction absorbs heat; both factors jointly determine the temperature change of the electrolytic cell. This embodiment uses the lumped parameter method to establish a thermal dynamic model, assuming a uniform temperature distribution inside the electrolytic cell.

[0028] In the formula, T EL The temperature of the electrolytic cell (K); The heat capacity (J / K) of the electrolytic cell can be obtained by weighting the mass of each component with its specific heat capacity. For electrolytic cell efficiency; Heat dissipation (W) represents the heat loss from the electrolytic cell to the surrounding environment. Heat dissipation is calculated using Newton's law of cooling.

[0029] in, h conv The equivalent convective heat transfer coefficient (W / (m) 2 K), can be obtained through experimental calibration or by consulting the equipment manual; A surf The heat dissipation area of ​​the outer surface of the electrolytic cell (m²) 2 ); T amb The ambient temperature (K) is taken as 298.15K in this embodiment.

[0030] (3) Efficiency characteristic model The efficiency of an electrolyzer is defined as the ratio of the chemical energy of the produced hydrogen to the input electrical energy:

[0031] In the formula, Hydrogen production flow rate of the electrolyzer (Nm³) 3 / s), and current density j Following Faraday's law:

[0032] in, V m Let m be the molar volume of the gas, taken as 0.022414 m³ under standard conditions. 3 / mol; ε Faraday efficiency is the efficiency that characterizes the leakage loss of current. It is related to current density and operating temperature, and is usually close to 1 at high current densities and decreases at low current densities. The lower heating value of hydrogen is taken as 120 MJ / kg or 3.00 kWh / Nm³. 3 .

[0033] In practical applications, efficiency Because it is difficult to calculate accurately directly using analytical expressions, this embodiment uses an offline experimental data fitting method to establish the power... P EL and temperature T EL A two-dimensional interpolation table relating efficiency to thermal efficiency is used. Specifically, under different temperature conditions (e.g., 60°C, 70°C, 80°C, 90°C), the hydrogen production rate of the electrolyzer at different power points (e.g., 10%PN, 20%PN, ..., 100%PN) is tested, and the corresponding efficiency values ​​are calculated to form an efficiency characteristic surface. In the control system, the efficiency value under the current operating condition is obtained in real time through table lookup and linear interpolation for use in thermal model calculations and energy balance analysis.

[0034] (4) Key dynamic process constraints Based on the above model, this embodiment extracts the following key dynamic process constraints as the operating boundaries that subsequent control strategies must adhere to: ① Operating power range: The operating power of the electrolytic cell is limited by temperature, i.e., P EL,min (T EL ) ≤ P EL ≤ P EL,max , where P EL,max Rated power, determined by the equipment nameplate; P EL,min (T EL The minimum stable operating power is 0.5, which varies with temperature. When the electrolytic cell is in a hot state (T... EL At ≥ 75℃, P EL,min It can be reduced to 10% of rated power; when in a cold state (T EL At ≤ 60℃, P EL,min It needs to be maintained at 20% or more of the rated power to ensure thermal balance and gas purity requirements.

[0035] ② Rate of increase constraint: The rate of power change in the electrolytic cell is limited by thermal stress and pressure balancing capacity, satisfying... In this embodiment, R is taken as... EL= 0.2 pu / min, meaning the power change per minute does not exceed 20% of the rated power. This value can be calibrated according to the specific model of the electrolytic cell and the operation manual provided by the manufacturer.

[0036] ③ Start-up and shutdown time constraints: It takes about 15 minutes for the electrolyzer to go from a cold start to stable operation, during which it needs to go through stages such as preheating, airtightness check, and pressure build-up; after shutdown, it needs to cool down for about 5 minutes before it can be restarted to prevent safety risks caused by thermal stress damage and residual hydrogen accumulation.

[0037] ④ Temperature limits: The operating temperature of the electrolytic cell must be maintained within a safe range, typically 60℃ ≤ T EL ≤ 90°C. Efficiency decreases significantly below 60°C, and gas purity is difficult to guarantee; above 90°C, diaphragm aging and electrode corrosion may be accelerated, affecting equipment lifespan.

[0038] The above models together constitute a complete dynamic description of the water electrolysis hydrogen production unit. Through this model, the control system can predict the temperature change trend under different power commands and assess in advance whether process constraints are violated, thereby achieving safe and efficient control decisions.

[0039] 2. Dynamic Model of Hydrogen Buffer Unit In this embodiment, the hydrogen buffer unit serves as a crucial intermediate link connecting the water electrolysis hydrogen production unit and the ammonia synthesis unit, playing a vital role in energy buffering, flow regulation, and pressure stabilization. Because the hydrogen production rate of the electrolyzer can change rapidly (on the order of seconds to minutes), while the ammonia synthesis unit's demand for feed hydrogen changes relatively slowly (on the order of minutes to hours), directly and rigidly coupling the two would lead to frequent fluctuations in the ammonia synthesis unit's operation, affecting its operational stability and catalyst lifespan. Therefore, this embodiment incorporates a hydrogen buffer unit, using the "peak shaving and valley filling" effect of the hydrogen storage tank to decouple the upstream and downstream processes in terms of time and flow.

[0040] The hydrogen buffer unit in this embodiment consists of a 2000Nm 3 The system consists of a stationary high-pressure hydrogen storage tank, an inlet compressor, and an outlet pressure regulating valve. The hydrogen storage tank is a steel pressure vessel with a design pressure of 3.2 MPa, a minimum operating pressure of 1.6 MPa, and an operating temperature of ambient temperature (approximately 25°C). The following details the construction process of its dynamic model.

[0041] (1) Definition of state variables of hydrogen storage tank To describe the dynamic characteristics of the hydrogen storage tank, the following state variables are defined in this embodiment: Hydrogen inventory: The total amount of hydrogen stored in the hydrogen storage tank is expressed as the volume of hydrogen under standard conditions (0°C, 101.325 kPa). This variable is a key indicator for system energy management, reflecting the "state of charge" of the buffer system.

[0042] Hydrogen storage tank pressure: Ptank (unit: Pa or MPa), is the absolute pressure inside the hydrogen storage tank, which is determined by the amount of hydrogen stored and the temperature.

[0043] Normalized index for hydrogen inventory: The state of charge (SoHSoH) of hydrogen in a hydrogen storage tank is defined as follows:

[0044] in, The volume of hydrogen gas under standard conditions (0°C, 101.325 kPa) (unit: Nm³) 3 ), and Corresponding to the minimum working pressure and the highest work pressure Hydrogen reserves (Nm³) below. SoH The value range is [0,1]. SoH = 0 indicates that the hydrogen storage tank is at the minimum allowable pressure. SoH = 1 indicates that the pressure is at the maximum allowable pressure.

[0045] (2) Equation of state for gases This embodiment uses the ideal gas law to describe the relationship between the pressure and quantity of hydrogen in the hydrogen storage tank. Since the maximum working pressure is 3.2 MPa, hydrogen can be approximated as an ideal gas at this pressure, with a compressibility factor Z≈1, and the error is within an acceptable range for engineering applications.

[0046] For a hydrogen storage tank with a fixed volume, the gas law is:

[0047] in, V tank The geometric volume (m³) of the hydrogen storage tank 3 In this embodiment, ), is a fixed value; n is the amount of hydrogen (mol); R is the ideal gas constant, taken as 8.314 J / (mol·K); T tank The temperature (K) of the gas inside the hydrogen storage tank is assumed in this embodiment to be in thermal equilibrium with the surrounding environment, and the temperature is constant at the ambient temperature. T amb = 298.15K.

[0048] When converting the amount of substance to volume under standard conditions, using the relationship that 1 mol of gas has a volume of 0.022414 m³ under standard conditions (273.15 K, 101.325 kPa), we can obtain:

[0049] Combining the two equations, we obtain the conversion relationship between pressure and standard volume:

[0050] To simplify calculations, this embodiment uses an engineering approximation formula:

[0051] in, k p The conversion factor can be obtained from the equipment nameplate parameters or through on-site calibration; P 0 is a constant. Under the hydrogen storage tank parameters of this embodiment, it was calculated that: when hour, MPa; when hour, MPa. Therefore. k p = 0.0008MPa / Nm 3 , P 0 = 1.6 MPa.

[0052] (3) Dynamic equation of mass balance The change in the hydrogen level in the hydrogen storage tank is determined by both the inflow and outflow rates, and follows the law of conservation of mass.

[0053] In the formula, Hydrogen flow rate (Nm³) entering the hydrogen storage tank 3 / s), the hydrogen output from the water electrolysis hydrogen production unit; Hydrogen flow rate (Nm³) from the hydrogen storage tank 3 / s), supplied as raw material to the ammonia synthesis unit.

[0054] By combining the relationship between pressure and inventory, the dynamic equation for pressure can be obtained:

[0055] (4) Key dynamic process constraints To ensure the safe operation of the hydrogen storage tank and the stable operation of subsequent equipment, the following dynamic process constraints are set in this embodiment: ① Pressure limit: The working pressure of the hydrogen storage tank must be maintained within a safe range, that is:

[0056] In this embodiment, =1.6MPa, =3.2MPa. When the pressure is below the lower limit, it may cause unstable gas flow or fail to meet the minimum inlet pressure requirements of the ammonia synthesis unit; when the pressure is above the upper limit, it will trigger the safety valve to release gas, resulting in hydrogen loss and safety risks.

[0057] ② Flow rate change constraint: Both the intake compressor and the outlet pressure regulating valve have mechanical inertia, limiting their flow regulation rate. This embodiment sets the following constraints:

[0058] in, The upper limit of the intake compressor flow rate change rate is set at 0.05 Nm. 3 / s 2 (i.e., a change of 3 Nm per minute) 3 / s); The upper limit of the flow rate change rate of the outlet pressure regulating valve is set at 0.1 Nm. 3 / s 3 (i.e., a change of 6 Nm per minute) 3 These parameters can be calibrated based on the actual response characteristics of the compressor and valves.

[0059] ③Minimum / maximum flow constraints:

[0060]

[0061] In this embodiment, =0 (Shutdown allowed) =0.5 Nm3 / s (corresponding to the maximum hydrogen production capacity of the electrolyzer); =0.02Nm 3 / s (corresponding to the minimum stable operating flow rate of the ammonia synthesis unit). =0.4Nm 3 / s (corresponding to the maximum processing capacity of the ammonia synthesis unit).

[0062] 3. Dynamic model of ammonia synthesis unit The ammonia synthesis unit is the final stage of the energy chain in this embodiment, responsible for reacting hydrogen and nitrogen supplied by the hydrogen buffer unit to produce liquid ammonia. Due to the significant thermal inertia and chemical equilibrium characteristics of the ammonia synthesis reaction, its response speed is much slower than that of the water electrolysis hydrogen production unit. If its dynamic process cannot be accurately described, control decisions will become disconnected from actual operating conditions, leading to catalyst thermal stress damage or reactor instability. Therefore, this embodiment establishes a lumped parameter dynamic model of the ammonia synthesis unit to characterize the dynamic response relationship between its input hydrogen flow rate and output ammonia yield, as well as the variation patterns of key process parameters.

[0063] The ammonia synthesis unit in this embodiment employs the Haber-Bosch process and includes a fixed-bed catalytic reactor, a gas recirculation compressor, a heat exchanger, and an ammonia separator. The reactor is filled with an iron-based or ruthenium-based catalyst, and the reaction pressure is 15-20 MPa while the reaction temperature is 400-500°C. To meet the real-time requirements of the control system, this embodiment simplifies the ammonia synthesis unit, using a lumped parameter model to describe its overall dynamic characteristics, rather than performing a three-dimensional model of the detailed flow field and concentration distribution inside the reactor.

[0064] (1) Dynamic model of ammonia production The input variable for the ammonia synthesis unit is the feed hydrogen flow rate. (Unit: Nm³ / s), the output variable is the liquid ammonia production. (Unit: kg / s). The dynamic relationship between the two is approximately that of a first-order inertial element plus a pure time-delay element:

[0065] in: The time constant (in seconds) for the ammonia synthesis unit characterizes the system's response speed to changes in input. This parameter is determined by the reactor's heat capacity, the total mass of the gas, and the heat exchanger's heat transfer capacity. For a 1000-ton-per-year plant in this embodiment, The typical value range is 600~1800 seconds (10~30 minutes), and the specific value can be obtained through step response experiments or equipment parameter identification. In this embodiment, we take... =1200 seconds. The static gain coefficient (unit: kg / Nm³) of the ammonia synthesis unit represents the stable ammonia production that can be generated per unit hydrogen flow input.

[0066] The pure time lag (in seconds) characterizes the time delay from a change in hydrogen flow rate to the onset of ammonia production response. This lag is primarily caused by the gas flow time within the pipes and reactor, the thermal inertia of the heat exchanger, and the delay in the ammonia separation process. For the apparatus of this embodiment, The typical value range is 60~300 seconds (1~5 minutes), and this embodiment takes... =180 seconds.

[0067] (2) Dynamic model of reactor temperature The temperature of the ammonia synthesis reactor is a key parameter affecting catalyst activity and reaction efficiency, and it is also a process variable that the control system must strictly monitor. This embodiment establishes a lumped parameter dynamic model of the reactor temperature to describe its balance relationship with ammonia yield and heat of reaction.

[0068] in,T syn This indicates the average temperature of the reactor. In this embodiment, the normal operating temperature range is 400℃ ≤ T syn ≤500℃. If the temperature is too low, the reaction rate is insufficient and the conversion rate decreases; if the temperature is too high, the catalyst is deactivated by accelerated sintering, and the equilibrium conversion rate decreases instead. C th,syn The total heat capacity of the reactor (unit: J / K) includes the total heat capacity of the catalyst bed, reactor shell, heat exchanger, and internal gases.

[0069] Q react The exothermic power of the reaction is determined by the ammonia production and the enthalpy change of the reaction.

[0070] in, ΔH For the enthalpy change of the reaction, 92.4 kJ / mol was used; η therm The thermal efficiency coefficient characterizes the proportion of reaction heat absorbed by the reactor, and is usually taken as 0.8~0.95. In this example, it is taken as 0.9.

[0071] Q cool To recover the heat removed by the cooling system, circulating cooling water or boiler feedwater is typically used. This embodiment assumes the cooling system can maintain the reactor temperature within a set range, and its dynamic response is controlled by a cooling water flow regulating valve, which can be approximated as:

[0072] in, UA It is the product of the overall heat transfer coefficient of the heat exchanger and its area (W / K). T cool,in The inlet temperature of the cooling medium is (K). u cool The opening degree of the cooling valve (0~1) can be adjusted by the upper-level control system as a control variable.

[0073] (3) Key dynamic process constraints To ensure the safe and stable operation of the ammonia synthesis unit, the following dynamic process constraints are set in this embodiment. These constraints will serve as the operating boundaries that must be followed in subsequent control strategies: ①Minimum feed flow rate constraint: To prevent gas circulation interruption and catalyst bed fluidization failure, the feed hydrogen flow rate must be maintained above the minimum value. In this embodiment, F min = 0.02Nm 3 / s corresponds to a minimum stable operating load of approximately 20% of the rated load for the ammonia synthesis unit. Operating below this value may result in uneven gas distribution, hotspot shift, or disruption of the catalyst reduction atmosphere.

[0074] ② Temperature change rate constraint: To protect the catalyst from thermal shock, the rate of change of reactor temperature must be limited to a certain range. In this embodiment, R T,syn = 0.083℃ / s (i.e., a change of no more than 5°C per minute). This value can be determined according to the technical manual provided by the catalyst manufacturer. Exceeding this limit may cause thermal stress cracking or a decrease in catalyst activity.

[0075] ③ Temperature operating range constraints: In this embodiment, T syn,min = 400℃, T syn,max =500 C. Below the lower limit, the reaction rate is insufficient; above the upper limit, the catalyst deactivates rapidly and the equilibrium conversion rate decreases.

[0076] ④ Pressure fluctuation constraint: The rate of pressure change in the synthesis tower is limited to avoid seal failure or equipment fatigue.

[0077] In this embodiment, R P,syn = 0.5 MPa / min. Excessive pressure may trigger the safety interlock, while a sudden drop in pressure may cause the catalyst bed to loosen.

[0078] Step 2: Establish a multi-timescale hierarchical control architecture for source-grid-hydrogen-ammonia synergy. The electrical response of an electrolyzer can reach the millisecond to second level, but its thermal dynamics affect the available power range on a minute-level scale. Pressure changes in a hydrogen storage tank depend on the minute-level cumulative effect of inlet and outlet gas flow rates. Temperature and yield changes in ammonia synthesis reactors exhibit hour-level inertial responses, with a time constant of approximately 20 minutes. If a single control layer operates on a second-level cycle, while it can quickly respond to grid commands, it struggles to handle the long-term economic optimization required by the scheduling layer. If it operates on an hour-level cycle, it cannot participate in the grid's rapid regulation services at all. Therefore, this embodiment adopts a layered decoupling approach, assigning control tasks of different time scales to different control layers. These layers collaborate through limited information exchange, ensuring both rapid response to grid commands and stable operation of the chemical process.

[0079] 1. Scheduling layer (minute to hour level, control cycle 15 minutes) The dispatch layer is deployed within the Energy Management System (EMS) and is responsible for handling optimization decision-making problems on a timescale from minutes to hours. This layer receives the day-ahead power plan (formulated 24 hours in advance, with a time resolution of 15 minutes) and the intraday rolling correction plan (updated 4 hours in advance, with a time resolution of 5-15 minutes) issued by the power grid dispatch center. At the same time, it acquires local renewable energy generation forecast data and the current operating status of the system. With the goal of minimizing system operating costs, it formulates the total power plan of the electrolyzer cluster and the SoH reference trajectory of the hydrogen storage tanks for a future period of time.

[0080] (1) Optimization Model The scheduling layer employs a model predictive control framework, using a 15-minute control cycle to continuously optimize the runtime plan for the next 4 hours (i.e., 16 control cycles). The optimization problem is described as follows: ① Decision variables: P EL,set ( k ): No. k The total power setpoint of the electrolytic cell cluster within each control cycle is a continuous decision variable, and its range is limited by the total capacity of the electrolytic cell cluster.

[0081] SoH ref ( k ): No. k The reference target value of the hydrogen state of charge in the hydrogen storage tank at the end of each control cycle is a continuous decision variable used to guide the coordination layer to maintain the healthy operation of the hydrogen storage tank while meeting the needs of grid interaction.

[0082] ②Objective function:

[0083] in, Np = 16 represents the prediction time domain (corresponding to the next 4 hours).

[0084] C elec ( k ): No. k The electricity cost (in yuan) for each cycle is calculated using the following formula:

[0085] In the formula, Δt = 0.25 hours is the duration of the control cycle; ρ elec ( k ) is the first k The electricity price (RMB / kWh) for each cycle can be obtained from the time-of-use electricity price published by the power grid or the spot market price. This cost-incentive system encourages higher electricity consumption during off-peak hours and lower consumption or participation in demand response during peak hours.

[0086] C degr ( k ): No. k The aging cost (in yuan) of the electrolyzer over one cycle is used to characterize the impact of power fluctuations on equipment lifespan. A linear model related to the rate of power change is employed.

[0087] In the formula, α degr The aging cost coefficient (RMB / kW) is obtained by fitting life loss data provided by the electrolyzer manufacturer; in this embodiment, it is set to 0.005 RMB / kW. This penalty term suppresses drastic power fluctuations and extends the equipment's service life.

[0088] C soh ( k ): No. k The SoH deviation penalty term (amount) for each cycle is used to guide the hydrogen storage tank to maintain a reasonable state of charge, reserving adjustment capacity for subsequent grid interaction:

[0089] In the formula, β soh The deviation penalty coefficient (in yuan) is set to 100 yuan in this embodiment; SoH target The target state of charge is typically set to 0.5 (i.e., the hydrogen storage tank pressure is at the middle value) to retain bidirectional regulation capability.

[0090] ③Constraints: Electrolytic cell power constraints: P EL,min,agg ( k ) ≤ P EL,set ( k ) ≤ P EL,max,agg in, P EL,max,agg = 5000kW is the total rated power of the electrolytic cell cluster; P EL,min,agg ( k The minimum stable operating power of the cluster is denoted as , which varies with temperature and is predicted by the thermal model in step one. The calculation formula is as follows:

[0091] In the formula, T EL,i ( k ) is the first iTaiwan Electrolytic Cells k The predicted temperature at any given time is obtained recursively from the dynamic model in step one.

[0092] Electrolytic cell ramping constraints:

[0093] in, R EL,agg = 1000kW / h is the limit of the total ramp rate of the cluster, which corresponds to the ramp rate of 0.2 pu / min of a single electrolytic cell multiplied by 4 cells.

[0094] SoH constraint of hydrogen storage tank: SoH min ≤ SoH ref ( k ) ≤ SoH max in, SoH min = 0, SoH max = 1, corresponding to a lower pressure limit of 1.6 MPa and an upper pressure limit of 3.2 MPa, respectively. Furthermore, the change in SoH between adjacent cycles must satisfy a rate constraint determined by the maximum inlet and outlet gas flow rates:

[0095] in, =0.5Nm 3 / s, =0.4Nm 3 / s, =2000Nm 3 , SoH range = 1. The calculated maximum change in SoH per cycle is approximately 0.4.

[0096] Ammonia synthesis unit related constraints: The scheduling layer needs to ensure that the SoH reference trajectory is energy-balanced with the feed demand of the ammonia synthesis unit. Assuming the ammonia synthesis unit operates at rated conditions or planned load, the feed hydrogen flow rate... Given a known quantity. The change in SoH in the hydrogen storage tank should satisfy:

[0097] in, Hydrogen production flow rate of the electrolyzer, and electrolyzer power. P EL,set ( k and efficiency η ELThe relevant parameters are calculated using the efficiency characteristic model from step one. This constraint couples the electrolyzer power with the hydrogen storage tank state to ensure energy balance.

[0098] (2) Rolling optimization execution process of the scheduling layer The scheduling layer performs the following steps at the beginning of each control cycle (15 minutes): Status Acquisition: Obtain the current actual status of the system from the coordination layer, including the temperature of each electrolytic cell. T EL,i Hydrogen storage tank pressure P tank (or SoH), temperature of the ammonia synthesis reactor T syn wait.

[0099] Predictive Model Update: Based on the current state and the dynamic model established in Step 1, predict the evolution trend of each state variable over the next 4 hours. The prediction assumes that the coordination layer can perfectly track the scheduling layer's plan, but considers the possible range of actual tracking errors.

[0100] Optimization Solution: Use a commercial solver (such as Gurobi, CPLEX) or an open-source solver (such as Ipopt) to solve the above optimization problem and obtain the results for the next 4 hours. P EL,set ( k )and SoH ref ( k ) sequence (k=1,...,16).

[0101] Instruction issuance: The optimized plan for the first control cycle (P) will be issued. EL,set (1) and SoH ref (1) The reference trajectory is sent to the coordination layer. In the next control cycle, the optimization window is updated on a rolling basis, and the above steps are repeated to achieve closed-loop correction.

[0102] 2. Coordination layer (second to minute level, control cycle of 1 minute) The coordination layer, deployed within the regional controller (PLC), operates on a 1-minute control cycle and serves as the hub connecting slow scheduling and rapid execution. This layer uses the planned trajectory issued by the scheduling layer as a long-term reference and the dynamic model constructed in step one as its internal predictive model. Employing a model predictive control (MPC) algorithm, it tracks real-time grid fast adjustment commands (such as AGC signals and frequency modulation commands) and coordinates power allocation between the electrolyzer cluster and the ammonia synthesis unit, ensuring precise load response while meeting dynamic process constraints.

[0103] (1) Basic framework of the coordination layer MPC The coordination layer MPC employs a rolling time-domain optimization strategy, performing optimization calculations once per control cycle (1 minute). The core elements of MPC include three stages: predictive model, rolling optimization, and feedback correction.

[0104] ①Prediction Model The internal prediction model for MPC employs a simultaneous dynamic model of the electrolyzer, hydrogen storage tank, and ammonia synthesis unit established in step one, represented in discrete-time state-space form. The input to the prediction model is the control variable u(k), the output is the controlled variable y(k), and the state variable is x(k). The model is discretized using the forward Euler method, with a discrete time step size of [missing information]. Δt MPC = 1 minute.

[0105] Control variable u(k):

[0106] For the first i Active power (kW) of the Taiwan Electrolytic Cell. This refers to the feed hydrogen flow rate for the ammonia synthesis unit.

[0107] State variable x(k):

[0108] in, For the temperature of each electrolytic cell, For the hydrogen storage tank pressure, The temperature of the ammonia synthesis reactor. The yield of liquid ammonia is expressed in kg / s.

[0109] Controlled variable y(k):

[0110] in, The total power of the electrolyzer cluster (kW) is a direct regulating quantity that participates in grid interaction; The hydrogen charge state (dimensionless) in the hydrogen storage tank reflects the energy state of the buffer system.

[0111] The state transition equation of the prediction model is obtained by discretizing the dynamic model in step one, and its specific form is as follows: Electrolyzer temperature update (based on thermal model):

[0112] in, Based on the efficiency characteristic model of step one and Interpolation is obtained; .

[0113] Hydrogen storage tank pressure update (based on mass balance):

[0114] in, This represents the total hydrogen production flow rate of the electrolyzer. This represents the feed flow rate for synthetic ammonia. The pressure-to-stock conversion factor is 0.0008 MPa / Nm³ in this embodiment.

[0115] Ammonia synthesis temperature and yield updates (based on a first-order inertial plus pure lag model, lag processing needs to be considered): To handle pure time delay =180 seconds (i.e., 3 control cycles). This embodiment uses a delayed state vector to store historical inputs. Define the delay variable. ,but:

[0116]

[0117] in, ; To account for the heat removed by the cooling system, this embodiment assumes that the cooling system can maintain a set temperature, the dynamics of which are guaranteed by the lower-level controller, approximated in MPC as follows: ,in The cooling valve opening can be optimized in MPC as an additional control variable.

[0118] ② Scrolling optimization In each control cycle, MPC solves a quadratic programming problem in a finite time domain. The optimization time domain is Np = 15 minutes (15 control cycles), and the control time domain is Nc = 5 minutes (5 control cycles). It is assumed that the control quantity remains unchanged after Nc.

[0119] Objective function:

[0120] The first item: tracking error penalty. For reference trajectory, where This is a value set by the power grid command or the dispatching layer. This is the SoH reference trajectory issued by the scheduling layer. For the weight matrix, this embodiment takes... q P = 1 (kW) - ²), q SoH = 100, reflecting a higher priority for SoH tracking, because deviation from SoH will affect the system's future adaptability.

[0121] The second item: Penalty for changes in the control quantity. The change in control quantity between adjacent cycles This is a penalty weight matrix used to suppress drastic changes in control variables and extend equipment life. In this embodiment, the penalty is based on changes in electrolytic cell power. r i = 0.01, Penalty for changes in ammonia synthesis feed flow rate r 5 = 0.1.

[0122] The third item: Soft penalties for exceeding limits. To constrain the sum of quantities exceeding the limit, λ =1000 is the penalty coefficient. When process constraints cannot be strictly met (such as when the predicted temperature exceeds the limit), a soft penalty mechanism is used to allow slight deviations from the limit but impose a high penalty to ensure the feasibility of the solution.

[0123] Constraints: Electrolytic cell power limits:

[0124] in, It is determined by the temperature-dependent minimum power function in step one.

[0125] Electrolytic cell ramping constraints:

[0126] in, R EL =0.2pu / min=250kW / min.

[0127] Hydrogen storage tank pressure constraints: P tank,min ≤ P tank (k) ≤ P tank,max In this embodiment, P tank,min =1.6MPa, P tank,max =3.2MPa.

[0128] Ammonia synthesis feed flow rate constraints:

[0129] In this embodiment, =0.02Nm 3 / s, =0.4Nm 3 / s.

[0130] Constraint on the rate of change of ammonia synthesis feed flow:

[0131] in, Rfeed = 0.1Nm 3 / s 2 (i.e., a change of 6 Nm per minute) 3 / s).

[0132] Temperature constraints for ammonia synthesis:

[0133]

[0134] in, =400℃, =500℃, R T,syn = 0.083℃ / s (i.e., a change of 5°C per minute).

[0135] Power balance constraints:

[0136] This constraint ensures consistency in the total power calculation.

[0137] ③ Solving and Execution The MPC optimization problem is a nonlinear programming problem (NLP) because of its efficiency characteristics. η EL (P,T) and heat loss Q loss (T) is a nonlinear function. This embodiment uses the Sequential Quadratic Programming (SQP) algorithm or the interior-point method for solving the problem, with the solver being the open-source library CasaADi combined with Ipopt. The optimization solution is completed within each control cycle (1 minute), and the first value u(1) of the optimal control sequence is sent to the local layer for execution. In the next control cycle, the system feeds back the actual state value. x (k+1), MPC continuously updates the optimization window and solves the problem again.

[0138] 3. In-situ layer (milliseconds to seconds) The local layer is deployed within the field controller and includes low-level execution devices such as electrolytic cell power modules (thyristor rectifiers or IGBT converters), compressor frequency converters, and regulating valve positioners, as well as measurement devices such as temperature sensors, pressure transmitters, and flow meters. This layer is responsible for rapidly executing commands issued by the coordination layer, collecting high-frequency status data and feeding it back to the coordination layer, while also executing independent safety interlock protection logic.

[0139] (1) Instruction execution The local stratum receives setpoint instructions from the coordination layer, including: Active power setpoints for each electrolytic cell P EL,i,set (kW); Hydrogen flow rate setpoint for ammonia synthesis feed FH2,feed,set (Nm) 3 / s).

[0140] These commands are sent to the field controller via industrial Ethernet (such as Profinet, EtherNet / IP) or fieldbus (such as Modbus TCP) with a communication cycle of 100ms. The field controller uses PID control algorithms or more advanced internal model control algorithms to adjust the power electronic devices or valve openings, enabling the actual value to quickly track the set value. The response time of the electrolytic cell power module is typically in the millisecond range, and the response time of the flow control valve is in the second range, with an overall execution accuracy of ±1%.

[0141] (2) Data collection and feedback The following data were collected from the formation at a sampling period of 100ms to 1s: Electrolytic cell: DC voltage U stack,i DC current I stack,i ,temperature T EL,i (Multi-point thermocouples), electrolyte level, and pressure; Hydrogen storage tank: Pressure P tank Temperature T tank Inlet and outlet air flow rate F H2,in F H2,out ; Ammonia synthesis unit: reactor temperature Tsyn (multi-point), pressure Psyn, ammonia production Q NH3 (Calculated by mass flow meter or liquid level change).

[0142] After being filtered, the collected data is uploaded to the coordination layer via the communication network at a rate of 1 second to 1 minute, serving as status feedback for the MPC.

[0143] (3) Safety interlock protection The ground-level control system executes safety interlock logic independently of the upper-level control system to ensure safe equipment shutdown under abnormal operating conditions. Interlock conditions include, but are not limited to: Electrolytic cell: temperature T EL,i Emergency shutdown may be triggered by conditions such as ≥ 95℃, pressure ≥ 1.0MPa (the working pressure of an electrolyzer is usually low), low electrolyte level, or excessive hydrogen concentration.

[0144] Hydrogen storage tank: Pressure P tank ≥ 3.3MPa triggers safety valve release, P tank ≥ 3.4MPa triggers emergency shutdown.

[0145] Ammonia synthesis unit: Temperature Tsyn≥520℃ or ≤380℃, pressure Psyn≥22MPa triggers interlock shutdown.

[0146] The safety interlock logic is implemented in a PLC or dedicated safety controller and has the highest execution priority, ensuring that personal and equipment safety is prioritized under any circumstances.

[0147] The three-layer architecture described above achieves collaboration through the following mechanisms: In terms of time scale, the scheduling layer processes economic optimization on a 15-minute cycle, the coordination layer processes dynamic tracking and constraint management on a 1-minute cycle, and the local layer processes instruction execution on a millisecond to second level. The time scales of each layer differ by one to two orders of magnitude, effectively avoiding performance degradation caused by large time scales for a single controller. Regarding information transmission, the scheduling layer outputs planned trajectories (slow-changing references), the coordination layer outputs control instructions (fast-changing instructions), and the local layer outputs execution signals (instantaneous instructions). Uplink feedback gradually decreases in frequency from the local layer to the coordination layer and then to the scheduling layer, achieving effective information compression and transmission. In terms of constraint management, the scheduling layer handles macroscopic power and SoH constraints, the coordination layer adds dynamic process constraints (such as temperature change rate and ramp rate) on top of the scheduling layer constraints, and the local layer adds equipment-level safety interlock protection on top of the coordination layer constraints. This progressively refined constraint management method ensures that the final execution instructions are feasible under all levels of constraints.

[0148] Step 3: Design a cross-timescale coordinated control strategy based on dynamic security domains Building upon the layered architecture established in Step Two, this step further designs the core coordination and control strategy. The key challenge this strategy addresses is ensuring that the slow-speed ammonia synthesis unit does not deviate from its safe operating range due to frequent adjustments, while simultaneously responding rapidly to grid commands. To this end, this step introduces the concept of a dynamic safety domain, serving as a crucial link between the coordination layer and the scheduling layer, and coordinating the fast and slow dynamics.

[0149] 1. Real-time calculation of dynamic security domains At the start of each control cycle (1 minute), the MPC of the coordination layer quickly calculates the maximum and minimum adjustable power within a short time domain, based on the dynamic model established in step one and the real-time status feedback from the local layer. In this embodiment, the prediction time domain is set to 5 minutes. This duration is chosen based on the following considerations: firstly, 5 minutes is sufficient to cover the electrolyzer's ramp-up process and significant pressure changes in the hydrogen storage tank; secondly, this duration is shorter than the time constant of the ammonia synthesis unit (20 minutes), therefore, the ammonia feed flow rate can be considered an adjustable variable within a short time domain, while the ammonia temperature change can be approximately ignored or treated as a boundary constraint. The core idea of ​​the calculation is to solve for the maximum power increment and maximum power decrement that the system can safely respond to within a specified time window, provided all dynamic process constraints are met. These two quantities together constitute the dynamic safety domain, which mathematically is a time-varying interval. ,in This indicates the maximum power that the system can currently increase (in kW). This indicates the maximum power reduction that the system can currently achieve (in kW; a positive value indicates the extent of downward adjustment capability).

[0150] (1) Upward adjustment ability Calculation Upward adjustment capability is defined as the maximum power that the system can safely increase within the next 5 minutes, provided all process constraints are met. Its calculation can be formalized as a constrained optimization problem, with the decision variables being the power trajectory of each electrolyzer and the ammonia feed flow trajectory within the next 5 minutes. To meet the real-time requirements of online calculation, this embodiment employs a rule-based approximate calculation method, decomposing the upward adjustment capability into the adjustment capability limited by three independent constraints, and taking the minimum value among them.

[0151] The first constraint is the ramp-up capability constraint of the electrolytic cell cluster. Within a 5-minute time window, the increase in electrolytic cell power is limited by its maximum ramp-up rate. Let the total power of the electrolytic cell cluster at the current moment be... The upper limit of the ramp rate for each electrolytic cell is R EL = 250kW / min, then the total ramping capacity of the 4 electrolytic cells is R EL,agg = 1000kW / min. Within 5 minutes, the maximum power increase limited only by the climb rate is: P up,ramp = R EL,agg T window = 1000 × 5 = 5000kW in, T window= 5min is the prediction time domain. Since the total rated power of the electrolyzer is 5000kW, this constraint is usually not a limiting factor in practice, but the remaining capacity needs to be considered when the power is close to full capacity.

[0152] The second constraint is the upper limit constraint on the hydrogen storage tank pressure. An increase in electrolyzer power leads to an increase in hydrogen production. If the hydrogen storage tank pressure is already close to its upper limit, the ability to adjust upwards will be limited by the hydrogen buffer capacity. Let the current hydrogen storage tank pressure be... P tank ( t 0), the upper limit of pressure is P tank,max = 3.2MPa, the equivalent capacity factor of the hydrogen storage tank is kp = 0.0008MPa / Nm 3 Therefore, the additional amount of hydrogen that the hydrogen storage tank can hold is:

[0153] Assume the current ammonia synthesis feed flow rate is Its maximum value is =0.4Nm 3 / s, then within 5 minutes (300 seconds), the additional amount of hydrogen that the ammonia synthesis unit can consume is:

[0154] Without changing the ammonia synthesis feed flow rate, the upward adjustment capability is limited by the remaining capacity of the hydrogen storage tank; that is, the increase in hydrogen production by the electrolyzer within 5 minutes must not exceed [a certain limit]. Let the average hydrogen production efficiency of the electrolyzer be... Hydrogen has a low calorific value (LHV). H2 = 3.00kWh / Nm 3 The upward adjustment power, limited by the capacity of the hydrogen storage tank, is:

[0155] If the ammonia synthesis feed flow rate is adjusted synchronously, the effective hydrogen storage capacity can be expanded to [value missing]. This enhances the ability to adjust upwards.

[0156] The third constraint is the maximum feed flow rate constraint for ammonia synthesis. When the ammonia feed flow rate has reached its limit, the additional hydrogen produced by the electrolyzer cannot be consumed in time and must be buffered by the hydrogen storage tank. Therefore, this constraint is implicitly included in the hydrogen storage tank constraint. In addition, the electrolyzer temperature constraint also needs to be considered: if the current electrolyzer temperature is close to its limit (e.g., 85°C), further increasing the power may cause the temperature to exceed the limit (90°C). In this case, the temperature rise rate needs to be predicted using a thermal model to ensure that the temperature does not exceed the limit within 5 minutes. The calculation of the temperature constraint involves the thermal model in step one, and a linear approximation is used here: Let the temperature rise rate of the electrolyzer under the current operating conditions be dT / dP (unit: °C / kW), then the allowable power increment is (T... max T current ) / (dT / dP).

[0157] Taking into account the above constraints, the final value of the upward adjustment capability is:

[0158] in, P up,temp This refers to the power increment constrained by temperature. In this embodiment, a lookup table of temperature rise rates under different temperatures and power levels is established through offline simulation, and the power increase rate is quickly obtained online through table lookup and interpolation. P up,temp .

[0159] (2) Downward adjustment capability Calculation Down-regulation capability is defined as the maximum power that the system can safely reduce within the next 5 minutes, provided all process constraints are met. Its calculation is symmetrical to that of up-regulation capability, and is also decomposed into three independent constraints.

[0160] The first constraint is the downward ramp rate constraint of the electrolytic cell cluster. Assume that the upper limit of the downward ramp rate for each electrolytic cell is the same as the upward ramp rate, which is... R EL = 250kW / min, then the maximum power reduction within 5 minutes is also 5000kW.

[0161] The second constraint is the lower limit constraint on the hydrogen storage tank pressure. Reducing the electrolyzer power will lead to a decrease in hydrogen production. If the hydrogen storage tank pressure is already close to the lower limit, further reducing the power may cause the tank pressure to bottom out, affecting the gas supply stability of the ammonia synthesis unit. Let the current hydrogen storage tank pressure be P. tank (t0), the lower limit of pressure is P tank,min = 1.6MPa, then the amount of hydrogen that the hydrogen storage tank can release is:

[0162] Assume the current ammonia synthesis feed flow rate is Its minimum value is =0.02Nm 3 If the hydrogen consumption is / s, then the reduction in hydrogen consumption by the ammonia synthesis unit within 5 minutes is:

[0163] The downward adjustment power, limited by the hydrogen storage tank's gas supply capacity, is:

[0164] If we consider simultaneously reducing the ammonia synthesis feed flow rate, the effective gas supply capacity can be expanded to: This enhances the ability to adjust downwards.

[0165] The third constraint is the minimum operating power constraint for the electrolytic cell. The power of the electrolytic cell cannot be lower than the minimum stable operating power corresponding to its current temperature. P EL,min ( T EL,i This value increases as the temperature decreases. Let the total power of the current electrolytic cell cluster be... P EL,total (t0), the minimum power of the cluster corresponding to the current temperature is P. EL,min,agg If (t0), then the allowable power reduction is:

[0166] This constraint is particularly critical when the electrolyzer is in a cold state, because the electrolyzer needs to maintain a high power to keep the thermal balance.

[0167] Taking into account the above constraints, the final value of the downward adjustment capability is:

[0168] (3) Calculation of the duration In addition to providing amplitude information of the regulation capability, the dynamic safety domain also needs to provide an estimate of the duration. The duration is defined as the longest time the system can maintain its maximum regulation capability without violating any constraints under the current state and optimal control strategy. Its calculation method is based on the aforementioned optimization problem, extending the prediction time domain until a constraint is triggered; the difference between the trigger time and the current time is the duration. For example, in an upward regulation scenario, if the limiting factor is the upper limit of the hydrogen storage tank pressure, the duration is the time required to rise from the current pressure to the upper limit, which can be calculated inversely using the dynamic equations of the hydrogen storage tank. In this embodiment, the coordination layer will use the calculated regulation capability range... Together with the corresponding upward sustainability time and downward sustainability time, it is uploaded to the scheduling layer through the uplink communication channel.

[0169] 2. Coordination and control logic Based on the dynamic security domain, this embodiment designs coordination control logic for three typical scenarios, corresponding to three grid interaction needs: rapid frequency regulation, new energy fluctuation smoothing, and long-term peak shaving. These three scenarios cover different time scales from seconds to hours, demonstrating the core advantage of this method in cross-time scale coordination.

[0170] Scenario 1: The power grid issues a rapid frequency adjustment command (on a timescale of seconds to minutes). When the power grid dispatch center issues Automatic Generation Control (AGC) or frequency regulation commands, these commands are updated in real time on a second- or minute-by-minute basis, requiring the system to adjust the power at grid connection points within a very short period. At the beginning of each control cycle (1 minute), the coordination layer MPC first receives the latest frequency regulation command. P AGC ( t ), and calculate the required power change ΔP. req =P AGC (t) P actual (t), where P actual (t) represents the actual power at the current grid connection point. Subsequently, MPC determines whether this change is within the current dynamic safety domain: if ΔP req If ≥0 (requiring increased power), then determine... If ΔP req If the value is less than 0 (requiring a reduction in power), then determine... .

[0171] If the command value is within the dynamic security domain, it indicates that the system currently possesses security response capabilities, and MPC will execute the following coordination control logic. First, MPC will assign a higher weight (q in this embodiment) to the tracking error term of the power grid command in the objective function. P =1), driving rapid power tracking of electrolytic cell clusters P AGC ( t The reason for prioritizing the adjustment of the electrolyzer cluster power is that the electrolyzer has the fastest electrical response speed (milliseconds to seconds), which can meet the speed requirements of frequency modulation commands. Secondly, the MPC dynamically adjusts the reference trajectory SoH of the hydrogen storage tank. ref The buffering capacity of the hydrogen storage tank is utilized to absorb fluctuations in hydrogen production caused by rapid adjustments. Specifically, when the electrolyzer rapidly increases its power, the hydrogen production rises instantaneously, and the MPC will temporarily increase the SoH. ref The setpoint allows for a moderate increase in the hydrogen storage tank pressure within a safe range (e.g., from 0.5 to 0.55), thereby temporarily storing excess hydrogen instead of immediately requiring the ammonia synthesis unit to increase its feed. This adjustment is achieved by modifying the SoH value in the MPC objective function. ref(k) is implemented, and its specific value is determined by MPC through rolling optimization based on the prediction model. Conversely, when the electrolyzer rapidly reduces its power, MPC will temporarily reduce the SoH. ref This allows the hydrogen storage tank to release hydrogen to meet the continuous demand of the ammonia synthesis unit. This proactive approach of utilizing the buffering capacity of the hydrogen storage tank by adjusting the SoH reference trajectory enables the ammonia synthesis unit to operate under relatively stable conditions, avoiding frequent adjustments caused by rapid grid commands.

[0172] If the command value exceeds the dynamic safety domain, it indicates that the system cannot fully respond to the command without violating process constraints under the current operating conditions. In this case, MPC will respond according to the safety boundary, meaning the actual power change is limited to the boundary value of the dynamic safety domain: if Then according to Response; if Then according to In response, MPC generates a "insufficient adjustment capacity" message and uploads it to the scheduling layer. This message includes the value of the current security domain. and The specific reasons for the limitation (such as "hydrogen storage tank pressure is close to the upper limit" or "electrolyzer temperature is too low, resulting in minimum power limitation") can be identified. After receiving this information, the dispatch layer can adjust the reference trajectory appropriately in subsequent optimization plans or report the limitation of system regulation capacity to the power grid dispatch center.

[0173] Scenario 2: Significant fluctuations in renewable energy power (on a minute-by-minute scale) When the output of local renewable energy sources (wind or solar) fluctuates significantly, if it is not absorbed locally, the fluctuating power will be injected into the grid, affecting the power quality and stability of the grid. In this embodiment, the dispatch layer monitors the difference between renewable energy output and the total system load in real time, defining the net load as P. net (t)=P load (t) P renewable (t), where P load (t) represents the internal load of the system (including plant power consumption outside the electrolytic cell, etc.), P renewable (t) represents the power generation of new energy sources. When the net load fluctuation is detected to exceed the set threshold (such as 10% of the rated power), the dispatch layer actively sends an internal instruction to the coordination layer to "absorb the fluctuation", requiring the system to smooth the net load fluctuation by adjusting the power of the electrolyzer, so as to achieve local balance between source and load.

[0174] After the coordination layer MPC receives the fluctuation absorption command, the tracking reference P in its objective function... ref (t) is no longer a fixed planned value for the power grid, but a target value that is dynamically adjusted based on the output of new energy sources. Specifically, let the target power exchange between the grid connection point and the power grid be... Pgrid,target (Usually a fixed value or determined by the power grid dispatch), then the target power that the electrolyzer should track is: P ref (t) = P grid,target (P load (t) P renewable (t))+P base Among them, P base This represents the base power of the electrolyzer. In practice, this formula is equivalent to requiring the electrolyzer power to decrease inversely with the output of new energy: when the output of new energy increases, the target power of the electrolyzer increases to absorb excess electricity; when the output of new energy decreases, the target power of the electrolyzer decreases to compensate for the power deficit. Within the dynamic safety domain, MPC actively tracks this dynamic target by adjusting the electrolyzer power, achieving on-site absorption of new energy fluctuations.

[0175] Meanwhile, the MPC proactively adjusts the feed flow rate of the ammonia synthesis unit to reserve buffer space for potential fluctuations. This proactive adjustment is achieved through the MPC's predictive function: when predicting future conditions, the MPC assesses the possible trends of new energy output fluctuations in the future period. This trend information is input into the MPC from ultra-short-term forecast data (new energy output forecasts for the next 15 minutes to 4 hours) provided by the scheduling layer. The MPC incorporates a penalty term for the hydrogen storage tank's SoH in its optimization objective to maintain it within a reasonable range capable of handling future fluctuations. For example, if a new energy output fluctuation is predicted to occur upward within the next 15 minutes, the MPC will moderately reduce the ammonia synthesis feed flow rate in advance to maintain the hydrogen storage tank pressure at a lower level (e.g., SoH=0.4), reserving sufficient hydrogen storage space for the upcoming upward fluctuation; conversely, if a downward fluctuation is predicted, the ammonia synthesis feed flow rate will be increased in advance to maintain the hydrogen storage tank pressure at a higher level (e.g., SoH=0.6), reserving space for hydrogen release. This proactive adjustment is automatically achieved through the MPC's rolling optimization, without the need for additional rule-based judgments.

[0176] Scenario 3: The power grid issues long-term peak-shaving instructions (hourly level) When the power grid requires system participation in peak shaving services, the dispatch center issues a peak shaving plan in advance, requiring the system to adjust the electricity load to a specified level within the next few hours. These instructions have a timescale on the hourly level and a slow rate of change, but a long duration, placing high demands on the system's energy balance capabilities. After receiving the peak shaving plan, the dispatch layer converts it into a reference trajectory P of the total power of the electrolyzer cluster. ref (t) and the target trajectory of the hydrogen storage tank SoH SoH ref(t), and then distributed to the coordination layer. The design of the SoH target trajectory follows the energy balance principle: during peak shaving periods, the hydrogen storage tank, acting as an energy buffer, should be able to support adjustments to the electrolyzer power due to changes in its SoH. Specifically, the peak shaving period is set from t0 to... t f The planned power of the electrolytic cell is P. ref (t), the planned feed flow rate for synthetic ammonia is F. H2,feed (t) (determined by the scheduling layer based on the production plan), then the SoH trajectory should satisfy:

[0177] The scheduling layer uses this relationship to reverse-calculate SoH. ref (t) ensures that SoH remains within a safe range during peak shaving periods.

[0178] The coordination layer (MPC) performs slow-time-scale optimization based on the reference trajectory issued by the scheduling layer. Unlike the rapid response in Scenario 1, in this scenario, MPC prioritizes the smoothness of the adjustment process. When the peak-shaving command requires a reduction in system power load, the MPC's optimization strategy is as follows: First, the MPC prioritizes reducing the power of the electrolyzer, which is the most direct and effective adjustment method. However, reducing the electrolyzer power will lead to a decrease in hydrogen production. If the ammonia synthesis unit maintains its original feed flow rate, the hydrogen storage tank pressure will drop rapidly, potentially hitting the bottom limit. Therefore, while reducing the electrolyzer power, the MPC simultaneously and gradually reduces the feed flow rate of the ammonia synthesis unit. This process utilizes the thermal inertia of the ammonia synthesis unit (time constant τ). syn =1200 seconds) of slow adjustment, so that hydrogen supply and demand are rematched at a new equilibrium point. MPC is achieved by constraining the rate of change of the ammonia synthesis feed flow rate ( , where R feed =0.1Nm 3 / s 2 This ensures a smooth adjustment process. Simultaneously, the MPC precisely controls the SoH of the hydrogen storage tank according to the SoH issued by the scheduling layer. ref (t) Trajectory changes, with a high weight (q) applied to the SoH tracking error in the objective function. SoH =100), ensuring that the hydrogen storage tank pressure neither hits the bottom nor exceeds the limit at the end of the peak shaving period, leaving sufficient adjustment space for subsequent operation.

[0179] Conversely, when peak-shaving orders require an increase in system power load, the MPC prioritizes increasing the electrolyzer power and simultaneously and gradually increases the feed flow rate of the ammonia synthesis unit, similarly utilizing the buffering capacity of the hydrogen storage tank for a smooth transition. Throughout the peak-shaving process, the MPC consistently treats dynamic process constraints as hard constraints, especially the ammonia synthesis temperature change rate constraint. (℃ / s), by calculating the temperature change trend in the MPC prediction model and applying constraints in the optimization, it is ensured that the reactor temperature change rate does not exceed 5°C per minute, effectively protecting the catalyst lifetime.

[0180] The coordinated control logic of the three scenarios described above fully demonstrates the core role of the dynamic safety domain as a coordination link across time scales. In the fast frequency regulation scenario, the dynamic safety domain provides the MPC with a real-time regulation capability boundary, ensuring that rapid response does not violate process constraints, and decoupling of fast and slow processes is achieved by adjusting the SoH reference trajectory. In the renewable energy fluctuation scenario, the MPC actively smooths fluctuations using the regulation capability within the dynamic safety domain, and achieves forward-looking coordination through ultra-short-term forecasting. In the long-term peak shaving scenario, dynamic safety domain information helps the scheduling layer formulate feasible SoH trajectories, while the coordination layer achieves energy balance through smooth regulation and constraint management. Together, these three scenarios constitute a complete cross-time scale coordinated control strategy, enabling the water electrolysis hydrogen production and ammonia synthesis system to participate in grid interaction safely and flexibly.

[0181] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0182] 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 them; under the concept of the present invention, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the present invention as described above, which are not provided in detail for the sake of brevity; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for source-grid hydrogen-ammonia coordinated hydrogen energy load-side grid interaction control, characterized in that, Including the following steps: S1. Construct a hydrogen-ammonia loading polymerization model that includes dynamic process constraints, wherein the hydrogen-ammonia loading polymerization model includes: An electro-thermal-mass coupled dynamic model of a water electrolysis hydrogen production unit is used to characterize the nonlinear behavior of the electrolyzer under different power levels and temperature conditions, and to extract the operating power range constraints, ramp rate constraints, start-up and shutdown time constraints, and temperature limits of the electrolyzer. A dynamic model of the hydrogen buffer unit is used to describe the dynamic relationship between hydrogen storage and pressure in the hydrogen storage tank, and to extract the pressure limit, flow rate change constraint and minimum / maximum flow constraint of the hydrogen storage tank. The lumped parameter dynamic model of the ammonia synthesis unit is used to describe the dynamic response relationship between the feed hydrogen flow rate and the output ammonia production, and to extract the minimum feed flow rate constraint, temperature change rate constraint, temperature operating range constraint and pressure fluctuation constraint of the ammonia synthesis unit. S2. Establish a multi-timescale hierarchical control architecture for source-grid hydrogen-ammonia coordination, including: The scheduling layer, with a control cycle of minutes to hours, is used to receive grid scheduling plans and formulate the total power plan of the electrolyzer cluster and the hydrogen charge state reference trajectory of the hydrogen storage tank with the goal of minimizing system operating costs. The coordination layer, with a control cycle of seconds to minutes, receives the planned trajectory issued by the scheduling layer, uses model predictive control algorithms to track power grid regulation commands, and coordinates the power distribution between the electrolyzer cluster and the ammonia synthesis unit under the dynamic process constraints. The local layer, with an execution cycle of milliseconds to seconds, is used to execute instructions issued by the coordination layer, collect status data and feed it back to the coordination layer, and execute safety interlock protection logic. S3. Design a cross-timescale coordinated control strategy based on dynamic security domains, including: The coordination layer calculates the dynamic security domain in real time during each control cycle. The dynamic security domain includes the maximum adjustable power range in which the system can safely respond in the short future time domain and its duration. Based on the comparison results between the grid regulation command and the dynamic safety domain, the corresponding coordinated control logic is executed. This includes adjusting the electrolyzer power to track the command and dynamically adjusting the hydrogen charge state reference trajectory of the hydrogen storage tank to absorb hydrogen production fluctuations when the command value is within the dynamic safety domain. When the command value exceeds the dynamic safety domain, the response is based on the safety boundary and information on insufficient regulation capacity is generated and uploaded to the scheduling layer.

2. The hydrogen energy load-side power grid interaction control method according to claim 1, wherein The electro-thermal-mass coupling dynamic model of the water electrolysis hydrogen production unit includes: The electrolytic cell model is used to describe the relationship between the total voltage and current density of the electrolyzer. The thermal model uses the lumped parameter method to establish thermal dynamic equations to describe the temperature changes in the electrolytic cell; An efficiency characteristic model is established by fitting offline experimental data to create a two-dimensional interpolation table between power, temperature and efficiency. The operating power P of the electrolyzer EL The interval constraint is P EL,min (T EL ) ≤ P EL ≤ P EL,max , wherein P EL,max is the rated power, P EL,min (T EL ) is the minimum stable operating power; the ramp rate constraint is , wherein represents the change amount of the active power of the electrolyzer, represents the limit value of the ramp rate of the electrolyzer, represents the time interval corresponding to the power change; and the temperature limit value is that the working temperature of the electrolyzer is maintained within a safe range.

3. The method of claim 1, wherein, The dynamic model of the hydrogen buffer unit includes: The state variables of the hydrogen storage tank are defined using hydrogen quantity and pressure as state variables, and the state of charge of hydrogen is defined. The ideal gas law is used to describe the relationship between the hydrogen pressure and the amount of hydrogen in the hydrogen storage tank. The mass balance dynamic equation is used to describe the relationship between the change in hydrogen inventory in the hydrogen storage tank and the inflow and outflow rates. The pressure limit of the hydrogen storage tank is wherein, represents the minimum allowable working pressure of the hydrogen storage tank, represents the maximum allowable working pressure of the hydrogen storage tank, represents the real-time working pressure of the hydrogen storage tank; the flow rate change rate constraint includes an inlet gas compressor flow rate change rate constraint and an outlet gas pressure regulating valve flow rate change rate constraint.

4. The hydrogen energy load-side grid interaction control method as described in claim 1, characterized in that, The lumped parameter dynamic model of the ammonia synthesis unit includes: The dynamic model for ammonia production uses a first-order inertial element plus a pure time delay to describe the dynamic relationship between the feed hydrogen flow rate and the liquid ammonia production. A reactor temperature dynamic model is used to describe the balance between reactor temperature, ammonia production, and heat of reaction. The minimum feed flow rate constraint for the ammonia synthesis unit is: The temperature change rate constraint is The temperature operating range is constrained as follows: The pressure fluctuation constraint is ,in, This indicates the feed hydrogen flow rate into the ammonia synthesis unit. This is the minimum allowable flow rate of the feed hydrogen. This indicates the average temperature of the ammonia synthesis reactor. This represents the rate of change of reactor temperature over time. This indicates the upper limit of the allowable rate of change of reactor temperature. and These represent the minimum and maximum allowable temperatures for the ammonia synthesis reactor, respectively. This indicates the real-time temperature of the reactor. This represents the rate of change of pressure over time. This represents the upper limit of the allowable rate of pressure change.

5. The method as described in claim 1, characterized in that, The scheduling layer adopts a model predictive control framework to continuously optimize the operation plan for the next 4 hours with a control cycle of 15 minutes. The optimization objectives include electricity cost, electrolyzer aging cost, and hydrogen storage tank hydrogen charge state deviation penalty term. The constraints include electrolyzer power constraints, electrolyzer ramp-up constraints, hydrogen storage tank hydrogen charge state constraints, and ammonia synthesis unit association constraints.

6. The method as described in claim 1, characterized in that, The coordination layer adopts a model predictive control algorithm with a control cycle of 1 minute. Its internal predictive model adopts the joint dynamic model of the electrolyzer, hydrogen storage tank and ammonia synthesis unit established in step S1, which is represented in discrete time state space form. The optimization objectives include grid command tracking error penalty, control quantity change penalty and constraint over-limit soft penalty. The constraints include electrolyzer power limit, electrolyzer ramp-up constraint, hydrogen storage tank pressure constraint, ammonia synthesis feed flow constraint, ammonia synthesis feed flow rate change constraint, ammonia synthesis temperature constraint and power balance constraint.

7. The method as described in claim 1, characterized in that, The calculation method of the dynamic safety domain in step S3 is as follows: Based on the dynamic model established in step S1 and the real-time status fed back by the local layer, the coordination layer calculates the maximum adjustable power range in the future short time domain in the prediction time domain of 5 minutes. The upward adjustment capability is taken as the minimum value of the adjustment capability limited by the electrolyzer cluster ramping capability constraint, the upper limit constraint of the hydrogen storage tank pressure constraint and the electrolyzer temperature constraint. The downward adjustment capability is taken as the minimum value of the adjustment capability limited by the electrolyzer cluster downward ramping capability constraint, the lower limit constraint of the hydrogen storage tank pressure constraint and the minimum operating power constraint of the electrolyzer.

8. The method as described in claim 1, characterized in that, The coordination control logic in step S3 includes: When the power grid issues a fast frequency regulation command, if the command value is within the dynamic safety domain, the coordination layer model predicts and controls the power tracking command of the electrolyzer cluster, and dynamically adjusts the hydrogen charge state reference trajectory of the hydrogen storage tank to absorb hydrogen production fluctuations; if the command value exceeds the dynamic safety domain, it responds according to the safety boundary and generates information on insufficient regulation capacity, which is then uploaded to the scheduling layer. When local renewable energy power fluctuates, the dispatch layer sends a fluctuation absorption instruction to the coordination layer. The coordination layer's model predictive control tracks the dynamically adjusted target value by adjusting the electrolyzer power and proactively adjusts the feed flow of the ammonia synthesis unit to reserve buffer space. When the power grid issues a long-term peak-shaving command, the dispatching layer converts the peak-shaving plan into a reference trajectory of the total power of the electrolyzer cluster and a target trajectory of the hydrogen charge state of the hydrogen storage tank, and sends it to the coordination layer. The coordination layer uses model prediction and control to prioritize the adjustment of the electrolyzer power and simultaneously and gradually adjust the feed flow rate of the ammonia synthesis unit, constrain the rate of change of the ammonia synthesis feed flow rate, and control the hydrogen charge state of the hydrogen storage tank to change according to the target trajectory.

9. The method as described in claim 1, characterized in that, The local layer is deployed in the field controller and includes power modules, compressor frequency converters, regulating valve positioners and sensors. It collects data with a sampling period of 100ms to 1s, communicates with the coordination layer through industrial Ethernet or fieldbus, and independently executes safety interlock protection logic including electrolyzer temperature over-limit, hydrogen storage tank pressure over-limit, and ammonia synthesis unit temperature and pressure over-limit.

10. A source-grid hydrogen-ammonia coordinated hydrogen load-side grid interaction control system, characterized in that, Control is performed using the method described in any one of claims 1 to 9.

Images