Online Optimization Operation Method and System of Hydrogen Energy System under Uncertainty Environment

Abstract

The present invention discloses an online optimization operation method and system for a hydrogen energy system under an uncertain environment. The method steps are as follows: (1) Under the uncertain environment of renewable energy output, establish a mathematical optimization model for minimizing the long-term operation cost of the hydrogen energy system; (2) Use the Lyapunov optimization framework to decompose the established problem of minimizing the long-term operation cost of the hydrogen energy system into multiple single-time-slot minimization sub-problems; (3) Based on the observed system state parameters, solve the current time-slot minimization sub-problem, and use the optimal solution of this sub-problem for the operation control of the hydrogen energy system; (4) Update the virtual queue related to the hydrogen energy storage system; (5) Repeat steps (3)-(4) until the current time-slot number is greater than the specified number of optimization time-slots and then terminate. The method of the present invention does not require any prior information about the uncertain parameters and can effectively reduce the operation cost and carbon emissions of the hydrogen energy system.

Description

Technical Field

[0001] The present invention relates to an online optimal operation method and system for a hydrogen - containing energy system under an uncertain environment, belonging to the cross - field of the economic operation of the hydrogen - containing energy system and artificial intelligence. Background Technique

[0002] As is well known, the large - scale exploration and utilization of fossil fuels have caused serious energy and environmental problems. To alleviate these problems, the use of clean energy is a promising way. As a clean energy, hydrogen energy has attracted much attention due to its wide sources, convenient transportation and storage, etc. In a hydrogen - containing energy system, by coordinating the hydrogen energy storage system and renewable energy generation, such as photovoltaic power generation, the local utilization rate of renewable energy can be improved. Therefore, it is very important to study the optimal operation of the hydrogen - containing energy system.

[0003] Existing studies have proposed many methods to optimize the operation of hydrogen - containing energy systems, such as stochastic programming, robust optimization, model predictive control, alternating direction multiplier method, and mixed - integer non - linear programming. Although the above - mentioned methods have made positive progress, they require prior information about uncertain parameters or prediction of uncertain parameters. To avoid predicting uncertain parameters or obtaining their prior information, some data - driven methods have been proposed, such as deep reinforcement learning. Some work optimizes the energy management and economic operation in a hydrogen - containing energy system by using and improving deep reinforcement learning methods and combines them with industrial co - generation systems. In addition, some work also adopts a multi - agent deep reinforcement learning - based method to coordinate the operation of a hydrogen - containing energy building system on the premise of considering building thermodynamics. Although the deep reinforcement learning - based methods do not require any prior information about uncertain parameters and can operate online. However, these methods require a large amount of training data and lack interpretability, that is, the reasons for the performance of the methods are unknown. Summary of the Invention

[0004] Aiming at the deficiencies of the existing technology, the technical problem to be solved by the present invention is: reducing the long - term operation cost of the hydrogen - containing energy system and improving the economy of system operation.

[0005] An online optimal operation method for a hydrogen - containing energy system under an uncertain environment includes the following steps:

[0006] (1) Under the uncertain environment of the output of renewable energy generators, establish a mathematical optimization model for minimizing the long - term operation cost of the hydrogen - containing energy system;

[0007] (2) Use the Lyapunov optimization framework to decompose the established optimization problem of the long - term operation cost of the hydrogen - containing energy system into multiple single - time - slot optimization sub - problems;

[0008] (3) Solve the optimization sub - problem for the current time slot based on the observed system state parameters, and use the optimal solution of the sub - problem for the operation control of the hydrogen - containing energy system;

[0009] (4) Update the virtual queue related to the hydrogen energy storage system to obtain the virtual queue length for the next time slot;

[0010] (5) Repeat steps (3) - (4) until the current time - slot number is greater than the specified number of optimization time slots and then terminate.

[0011] Furthermore, the objective function of the mathematical optimization model for minimizing the long - term operating cost of the hydrogen - containing energy system is as follows:

[0012]

[0013] Where: \(C\) 1,t represents the cost of purchasing electricity at time slot \(t\), \(1\leq t\leq T\), and \(T\) represents the specified number of optimization time slots; \(C\) 2,t represents the cost of photovoltaic curtailment at time slot \(t\); \(C\) 3,t represents the cost of hydrogen purchase at time slot \(t\); \(C\) 4,t represents the cost of the operation / shutdown status and start - stop of each device at time slot \(t\); \(C\) 5,t represents the carbon emission cost at time slot \(t\), is the expectation operator.

[0014] The specific expressions of each operating cost in the objective function are as follows:

[0015] \(C\) 1,t = \(S\) g,t \(P\) grid,t \(\Delta t(2)\)

[0016]

[0017] \(C\) 3,t = \(S\) h,t \(m\) buy,t (4)

[0018]

[0019] \(C\) 5,t = \(\alpha\) ce \((\mu\) g \(P\) grid,t \(\Delta t+\mu\) pv \(P\) pv,t \(\Delta t+\mu\) h \(m\) buy,t )(6)

[0020] Where: \(\Delta t\) is the time interval for the regulation of the hydrogen energy system; \(S\) g,t is the electricity price at time slot \(t\), \(P\) grid,t is the electricity quantity purchased from the power grid at time slot \(t\); \(\alpha\)pv is the PV curtailment penalty factor, is the maximum PV power generation in time slot t, P pv,t is the actual PV power generation in time slot t; S h，t is the hydrogen price in time slot t, m buy,t is the amount of hydrogen purchased from the hydrogen market in time slot t; X ∈ {el, fc} represent the electrolyzer and fuel cell devices respectively, are the costs of device X for operation, startup, and shutdown respectively, are the states of device X for operation, shutdown, and startup / shutdown respectively; α ce is the carbon emission penalty factor, μ g represents the carbon emission rate related to the electricity purchased from the main grid, μ pv represents the carbon emission rate of the PV system, μ h is the carbon emission coefficient of the hydrogen purchased from the hydrogen market.

[0021] Furthermore, the constraint conditions of the mathematical optimization model for minimizing the long-term operating cost of the hydrogen-containing energy system include PV power generation system constraints, hydrogen energy storage system constraints, and system power supply-demand balance constraints.

[0022] The PV power generation system constraint expression is as follows:

[0023]

[0024]

[0025] Where: η pv is the PV power generation efficiency, h pv is the total radiation area of the solar panels, l t is the solar radiation intensity in time slot t; is the maximum power demand in the hydrogen-containing energy system, represents all.

[0026] The hydrogen energy storage system constraint expression is as follows:

[0027]

[0028]

[0029]

[0030]

[0031]

[0032]

[0033] Where: Ht+1 , H t are the hydrogen storage levels at time slots t+1 and t respectively; ω el , ω fc are the conversion coefficients of the electrolyzer and fuel cell respectively; P el,t , P fc,t are the input power of the electrolyzer and the output power of the fuel cell at time slot t respectively; m buy,t is the amount of hydrogen purchased from the hydrogen energy market at time slot t; H min , H max are the minimum and maximum storage levels of the hydrogen storage tank respectively; and M B are the maximum input power of the electrolyzer, the maximum output power of the fuel cell, and the maximum purchase amount of hydrogen that can be bought from the primary energy to the hydrogen market respectively.

[0034] The power supply-demand balance constraint expression of the system is as follows:

[0035]

[0036]

[0037] where: P load,t is the power demand at time slot t; P grid,t is the amount of electricity purchased from the main power grid at time slot t, is the maximum amount of electricity that can be purchased from the main power grid at time slot t.

[0038] Furthermore, the decision variables of the mathematical optimization model for minimizing the long-term operating cost of the hydrogen-containing energy system are the set {P gird,t , P pv,t , P el,t , P fc,t , m buy,t}}.

[0039] Furthermore, the single-time-slot optimization sub-problem P2 is as follows:

[0040] (P2) min Q t χ t + Vy t (17)

[0041] s.t. (7)-(8), (11)-(16)(18)

[0042] The process of transforming the long-term operating cost optimization problem of the hydrogen-containing energy system into a single-time-slot optimization sub-problem includes the following 4 steps:

[0043] 21) Construct a virtual queue Q related to the time-coupling constraint (9) in the hydrogen storage system t = Ht +W H , where: W H is an important control parameter;

[0044] 22) Introduce the Lyapunov function and define the Lyapunov drift-penalty function;

[0045] 23) Transform the solution of the Lyapunov drift-penalty function into a problem of minimizing the upper bound of the Lyapunov drift-penalty function, that is , where: V is a weight parameter;

[0046] 24) Further relax the problem of minimizing the upper bound of the Lyapunov drift-penalty function into the form of sub-problem P2, where:

[0047]

[0048] The important control parameter W in steps 21) and 23) H and the weight parameter V need to satisfy and

[0049] Let the variable The variable and

[0050] The variable

[0051] where is the highest electricity price.

[0052] and are defined as follows:

[0053] If then and

[0054] If then and

[0055] Furthermore, the optimal solution of the single-time slot optimization sub-problem includes which are the optimal solution of the electricity purchase quantity, the optimal solution of the actual photovoltaic power generation, the optimal solution of the electrolyzer input power, the optimal solution of the fuel cell output power, and the optimal solution of the hydrogen purchase quantity, respectively.

[0056] Furthermore, for the virtual queue related to the hydrogen energy storage system, its update process is as follows:

[0057]

[0058] An online optimal operation system for a hydrogen energy system under an uncertain environment, comprising the following modules:

[0059] Model establishment module: Establish a mathematical optimization model for minimizing the long-term operation cost of the hydrogen energy system under the uncertain environment of the output of renewable energy generators;

[0060] Single-time slot optimization module: Use the Lyapunov optimization framework to decompose the established optimization problem of the long-term operation cost of the hydrogen energy system into multiple single-time slot optimization sub-problems;

[0061] Control module: Based on the observed system state parameters, solve the current time slot optimization sub-problem, and use the optimal solution of the sub-problem for the operation control of the hydrogen energy system;

[0062] Virtual queue module: Update the virtual queue related to the hydrogen energy storage system to obtain the virtual queue length of the next time slot;

[0063] Circulation module: Repeat the working processes of the control module and the virtual queue module until the current time slot number is greater than the specified number of optimization time slots and then terminate.

[0064] A computer-readable storage medium for storing the above-mentioned online optimal operation system and method for a hydrogen energy system under an uncertain environment.

[0065] The beneficial effects achieved by the present invention: Compared with the long-term operation optimization methods for hydrogen energy systems based on stochastic programming, dynamic programming, model predictive control, etc., the method of the present invention only relies on the currently observed system parameters for decision-making and does not need to know any prior information about the uncertain parameters, so it has stronger generality. Compared with the rule-based method, the method of the present invention can effectively reduce the operation cost and carbon emissions, and can flexibly adjust the carbon emissions according to the carbon emission penalty coefficient. Compared with the method based on deep reinforcement learning, the method of the present invention does not require any training data, can provide approximate optimal performance, and the performance quality is interpretable. Description of the Drawings

[0066] Figure 1 is the flowchart of the online optimal operation method for the hydrogen energy system provided by the present invention;

[0067] Figure 2 is the comparison chart of the operation cost of the hydrogen energy system between the embodiment of the method of the present invention and other schemes under different carbon emission penalty coefficients;

[0068] Figure 3 is the comparison chart of the carbon emissions during the operation of the hydrogen energy system between the embodiment of the method of the present invention and other schemes under different carbon emission penalty coefficients;

[0069] Figure 4It is a comparison chart of the operating costs of the hydrogen-containing energy system in the embodiments of the method of the present invention and other solutions under different hydrogen prices;

[0070] Figure 5 It is a comparison chart of the carbon emissions during the operation of the hydrogen-containing energy system in the embodiments of the method of the present invention and other solutions under different hydrogen prices;

[0071] Figure 6 It is a comparison chart of the operating costs of the hydrogen-containing energy system in the embodiments of the method of the present invention and other solutions under different ratios of total photovoltaic power generation to total load;

[0072] Figure 7 It is a comparison chart of the carbon emissions during the operation of the hydrogen-containing energy system in the embodiments of the method of the present invention and other solutions under different ratios of total photovoltaic power generation to total load. Detailed implementation manners

[0073] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only used to illustrate the technical solutions of the present invention more clearly, and cannot be used to limit the protection scope of the present invention.

[0074] As Figure 1 shown, the design flow chart of the online optimal operation method of the hydrogen-containing energy system provided by the present invention includes the following steps:

[0075] Step 1: Establish a mathematical optimization model for minimizing the long-term operating cost of the hydrogen-containing energy system in the environment of uncertain output of renewable energy generators;

[0076] Step 2: Use the Lyapunov optimization framework to decompose the established long-term operating cost optimization problem of the hydrogen-containing energy system into multiple single-time-slot optimization sub-problems;

[0077] Step 3: Solve the current time-slot optimization sub-problem based on the observed system state parameters, and use the optimal solution of this sub-problem for the operation control of the hydrogen-containing energy system;

[0078] Step 4: Update the virtual queue related to the hydrogen energy storage system to obtain the virtual queue length of the next time slot;

[0079] Step 5: Repeat Steps 3-4 until the current time-slot serial number is greater than the specified number of optimization time slots and then terminate.

[0080] In the above Step 1, the established mathematical optimization model for minimizing the long-term operating cost of the hydrogen-containing energy system mainly includes three parts: an objective function, constraint conditions, and decision variables.

[0081] The objective function of the mathematical optimization model can be expressed as follows:

[0082]

[0083] Where: C 1,t represents the cost of purchasing electricity for time slot t, where 1 ≤ t ≤ T and T represents the specified number of optimization time slots; C 2,t represents the cost of PV curtailment for time slot t; C 3,t represents the cost of hydrogen purchase for time slot t; C 4,t represents the cost of the operating / turning-off states, starting and stopping of each device for time slot t; C 5,t represents the carbon emission cost for time slot t, is the expectation operator.

[0084] The specific expressions of each operating cost in the objective function are as follows:

[0085] C 1,t = S g,t P grid,t Δt(2)

[0086]

[0087] C 3,t = S h,t m buy,t (4)

[0088]

[0089] C 5,t = α ce (μ g P grid,t Δt + μ pv P pv,t Δt + μ h m buy,t )(6)

[0090] Where: Δt is the time interval for the hydrogen energy system to adjust; S g,t is the electricity price for time slot t, P grid,t is the electricity quantity purchased from the power grid for time slot t; α pv is the PV curtailment penalty coefficient, is the maximum PV power generation for time slot t, P pv,t is the actual PV power generation for time slot t; S h.t is the hydrogen price for time slot t, m buy,t is the hydrogen quantity purchased from the hydrogen market for time slot t; are the costs of device X operating, starting, and shutting down respectively, are the operating, turning-off, starting and stopping states of device X respectively; α ce is the carbon emission penalty coefficient, μ gDenotes the carbon emission rate related to the electricity purchased from the main power grid, μ pv Denotes the carbon emission rate of the photovoltaic system, μ h Is the carbon emission coefficient of the hydrogen purchased from the hydrogen market.

[0091] The constraint conditions of the mathematical optimization model include photovoltaic power generation system constraints, hydrogen energy storage system constraints, and system power supply and demand balance constraints.

[0092] The constraints of the photovoltaic power generation system can be expressed as follows:

[0093]

[0094]

[0095] Where: η pv Is the photovoltaic power generation efficiency, h pv Is the total radiation area of the solar panels, l t Is the solar radiation intensity at time slot t; Is the maximum power demand in the hydrogen-containing energy system.

[0096] The constraints of the hydrogen energy storage system can be expressed as follows:

[0097]

[0098]

[0099]

[0100]

[0101]

[0102]

[0103] Where: H t+1 、H t Are the hydrogen energy storage levels at time slots t + 1 and t respectively; ω el 、ω fc Are the conversion coefficients of the electrolyzer and fuel cell respectively; P el,t 、P fc,t Are the input power of the electrolyzer and the output power of the fuel cell at time slot t respectively; m buy,t Is the amount of hydrogen purchased from the hydrogen market at time slot t; H min 、H max Are the minimum and maximum storage levels of the hydrogen storage tank respectively; And M BThey are the maximum input power of the electrolyzer, the maximum output power of the fuel cell, and the maximum hydrogen purchase volume that can be bought from the hydrogen market with primary energy. The above hydrogen energy storage system constraint model is a time-coupled constraint.

[0104] The power supply-demand balance constraint of the system can be expressed as follows:

[0105]

[0106]

[0107] Where: P load,t is the power demand at time slot t; P grid,t is the electricity quantity purchased from the main power grid at time slot t, is the maximum electricity quantity that can be purchased from the main power grid at time slot t.

[0108] The decision variables of the mathematical optimization model are {P gird,t , P pv,t , P el,t , P fc,t , m buy,t}.

[0109] In step 2 above, the long-term operation cost optimization problem of the established hydrogen energy system is decomposed into multiple single-time-slot optimization sub-problems by using the Lyapunov optimization framework. The specific steps are as follows:

[0110] (1) To make the system controllable, assume that the following conditions hold:

[0111] μ g ≥ μ pv (17)

[0112]

[0113]

[0114]

[0115] Where: Assuming that (17) is satisfied because the carbon emission rate of photovoltaic is usually less than that of the main power grid, and other assumptions can ensure that the control parameter W H and the weight parameter V are positive numbers.

[0116] By converting the time-coupled constraint (9) in the hydrogen energy storage system into a time-average constraint, the established operation cost minimization problem of the hydrogen energy system can be relaxed into the following optimization problem:

[0117]

[0118] s.t. (7)-(8), (11)-(16)(22)

[0119]

[0120] wherein: is defined as follows:

[0121]

[0122]

[0123]

[0124] Construct a virtual queue Q(t) related to the time-average constraint (23) in the hydrogen energy storage system and its dynamic update process as follows:

[0125] Q t = H t + W H (27)

[0126]

[0127] wherein: W H is a control parameter. Based on the Lyapunov optimization framework, the constraint condition (28) can be equivalently transformed into the following virtual queue stability problem:

[0128]

[0129] s.t. (7)-(8), (11)-(16)(30)

[0130] Q t Average rate is stable (31)

[0131] wherein: In the constraint condition (31), the definition of Q t Average rate stability is defined as follows:

[0132] (2) Introduce the Lyapunov function and define the Lyapunov drift function

[0133]

[0134] From

[0135] Furthermore, it can be obtained that wherein:

[0136] Therefore, the Lyapunov drift-penalty function is expressed as follows:

[0137]

[0138] Wherein: V is a weight parameter.

[0139] (3) According to Equation (32), by minimizing the upper bound of the Lyapunov drift-penalty function, the following problem of minimizing the upper bound of the Lyapunov drift-penalty function can be obtained:

[0140]

[0141] s.t. (7)-(8), (11)-(16)(34)

[0142] (4) To find the appropriate control parameter W H and the weight parameter V to ensure that the hydrogen energy storage system constraint condition H min ≤H t ≤H max is satisfied, the optimization problem is further transformed into the following form:

[0143] min Q t χ t +Vy t (35)

[0144] s.t. (7)-(8), (11)-(16)(36)

[0145] Wherein:

[0146] The said control parameter W H and the weight parameter V need to satisfy and

[0147] Let the variable Variable and

[0148] Variable

[0149] and can be defined as follows:

[0150] If then and

[0151] If then and

[0152] In the above step 3, the control module of the hydrogen-containing energy system reads the current time slot status data including the virtual queue Q t and the maximum photovoltaic power generation System power demand P load,t 、 electricity price S g,t 、 hydrogen price S h,t 、 carbon emission rate μ related to purchasing electricity from the main power grid g 、 carbon emission rate μ related to the photovoltaic power generation system pv and carbon emission rate μ related to purchasing hydrogen from the hydrogen market h ; At the same time, considering the control parameter W H and the value of the weight parameter V, solve to obtain the grid power purchase quantity P gird,t 、 actual photovoltaic power generation P pv,t 、 electrolyzer input power P el,t 、 fuel cell output power P fc,t 、 and hydrogen purchase quantity m buy,t , and use these decision data for the operation control of the hydrogen - containing energy system.

[0153] In step 4 above, the virtual queue module updates the virtual queue to obtain the virtual queue length of the next time slot.

[0154] In step 5 above, the cycle module of the hydrogen - containing energy system determines whether the current time slot number is greater than the specified optimization time slot number. If the current time slot number is less than the specified optimization time slot number, repeat steps 3 - 4 above; if the current time slot number is greater than the specified optimization time slot number, the optimization terminates.

[0155] An online optimization operation system for a hydrogen - containing energy system under an uncertain environment, characterized by including the following modules:

[0156] Model establishment module: Establish a mathematical optimization model for minimizing the long - term operation cost of the hydrogen - containing energy system under the uncertainty of the output of renewable energy generators;

[0157] Single - time - slot optimization module: Use the Lyapunov optimization framework to decompose the established long - term operation cost optimization problem of the hydrogen - containing energy system into multiple single - time - slot optimization sub - problems;

[0158] Control module: Based on the observed system state parameters, solve the current time - slot optimization sub - problem and use the optimal solution of the sub - problem for the operation control of the hydrogen - containing energy system;

[0159] Virtual queue module: Update the virtual queue related to the hydrogen energy storage system to obtain the virtual queue length of the next time slot;

[0160] Cycle module: Repeat the working processes of the control module and the virtual queue module until the current time slot number is greater than the specified optimization time slot number and then terminate.

[0161] A computing-readable storage medium for storing the above-mentioned online optimization operation system and method of a hydrogen energy system under an uncertain environment.

[0162] Compared with the prior art, the method proposed in the embodiments of the present invention is universal. Since the control strategy of the hydrogen energy system for each time slot is obtained only based on the observed state of the current time slot, this method does not need to know any prior information about the parameters of the uncertain system; the method proposed in the present invention is efficient. Compared with the existing optimization operation methods, such as Figure 2 As shown, it is a comparison chart of the operation costs of the hydrogen energy system under different carbon emission penalty coefficients between the embodiment of the method of the present invention and other schemes. Scheme 1 is an optimization method based on rules. Compared with Scheme 1, the method of the present invention can save 0.298%-7.582% of the total operation cost.

[0163] As Figure 3 shown, it is a comparison chart of the carbon emissions during the operation of the hydrogen energy system under different carbon emission penalty coefficients between the embodiment of the method of the present invention and other schemes. Compared with Scheme 1, the method of the present invention can flexibly adjust the carbon emissions by adjusting the carbon emission penalty coefficient. Specifically, compared with Scheme 1, the method of the present invention can reduce the carbon emissions by 13.95%-23.83%.

[0164] As Figure 4 shown, it is a comparison chart of the operation costs of the hydrogen energy system under different hydrogen prices between the embodiment of the method of the present invention and other schemes. Compared with Scheme 1, the proposed method can save 4.387%-13.436% of the total operation cost.

[0165] As Figure 5 shown, it is a comparison chart of the carbon emissions during the operation of the hydrogen energy system under different hydrogen prices between the embodiment of the method of the present invention and other schemes. Compared with Scheme 1, the proposed method can reduce the carbon emissions by 22.17%-43.44%.

[0166] As Figure 6 shown, it is a comparison chart of the operation costs of the hydrogen energy system under different ratios of total photovoltaic power generation to total load between the embodiment of the method of the present invention and other schemes. Compared with Scheme 1, the proposed method can save 1.655%-8.917% of the total operation cost.

[0167] As Figure 7 shown, it is a comparison chart of the carbon emissions during the operation of the hydrogen energy system under different ratios of total photovoltaic power generation to total load between the embodiment of the method of the present invention and other schemes. Compared with Scheme 1, the proposed method can reduce the carbon emissions by 14.011%-25.315%.

[0168] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: it is still possible to modify the specific implementation manners of the present invention or make equivalent substitutions, and any modification or equivalent substitution that does not depart from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. An online optimal operation method for a hydrogen - containing energy system under uncertain environment, characterized in that, it includes the following steps: Step 1: Under the uncertain environment of the output of renewable energy generators, establish a mathematical optimization model for minimizing the long - term operation cost of the hydrogen - containing energy system; In Step 1, the objective function expression of the mathematical optimization model is as follows: Among them: C 1,t represents the cost of purchasing electricity in time slot t, where 1 ≤ t ≤ T and T represents the specified number of optimization time slots; C 2,t represents the cost of PV curtailment in time slot t; C 3,t represents the cost of hydrogen purchase in time slot t; C 4,t represents the cost of the operation / turn-off state and start / stop of each device in time slot t; C 5,t represents the carbon emission cost in time slot t, and E is the expectation operator; The specific expressions of each operation cost in the objective function are as follows: C 1,t = S g,t P grid,t Δt (2) C 3,t = S h,t m buy , t (4) C 5,t = α ce (μ g P grid,t Δt + μ pv P pv,t Δt + μ h m buy,t ) (6) Where: Δt is the time interval for the regulation of the hydrogen - containing energy system; S g,t is the electricity price in time slot t, P grid,t is the electricity quantity purchased from the power grid in time slot t; α pv is the PV curtailment penalty coefficient, is the maximum PV power generation in time slot t, P pv,t is the actual PV power generation in time slot t; S h,t is the hydrogen price in time slot t, m buy,t is the hydrogen quantity purchased from the hydrogen market in time slot t; are the costs of device X for operation, startup, and shutdown respectively, are the states of device X for operation, shutdown, startup - shutdown respectively; α ce is the carbon emission penalty coefficient, μ g represents the carbon emission rate related to the electricity purchased from the main power grid, μ pv represents the carbon emission rate of the PV system, μ h is the carbon emission coefficient of the hydrogen purchased from the hydrogen market; The constraint conditions of the mathematical optimization model include: Photovoltaic power generation system constraint, the expression is as follows: Where: η pv is the photovoltaic power generation efficiency, h pv is the total radiation area of the solar panel, l t is the solar radiation intensity at time slot t; is the maximum power demand in the hydrogen-containing energy system; Hydrogen energy storage system constraint, the expression is as follows: Where: H t+1 and H t are the hydrogen storage levels at time slots t + 1 and t respectively; ω el and ω fc are the conversion coefficients of the electrolyzer and fuel cell respectively; P el,t and P fc,t are the input power of the electrolyzer and the output power of the fuel cell at time slot t respectively; m buy,t is the amount of hydrogen purchased from the hydrogen market at time slot t, H min and H max are the minimum and maximum storage levels of the hydrogen storage tank respectively; and M B are the maximum input power of the electrolyzer, the maximum output power of the fuel cell, and the maximum purchase amount of hydrogen that can be bought from the primary energy to the hydrogen market respectively; System power supply - demand balance constraint, the expression is as follows: Where: P load,t is the power demand in time slot t; P grid,t is the amount of electricity purchased from the main power grid in time slot t, is the maximum amount of electricity that can be purchased from the main power grid in time slot t; The decision variables of the mathematical optimization model are: {P gird,t , P pv,t , P el,t , P fc,t , m buy,t} set; Step 2: Use the Lyapunov optimization framework to decompose the established long - term operation cost optimization problem of the hydrogen - containing energy system into multiple single - time - slot optimization sub - problems; In Step 2, the single - time - slot optimization sub - problem P2 is as follows: (P2)min Q t χ t +Vy t (17) s.t.(7)-(8),(11)-(16) (18) The process of decomposing the established long - term operation cost optimization problem of the hydrogen - containing energy system into multiple single - time - slot optimization sub - problems includes the following steps: (21) Construct a virtual queue Q related to the time-coupling constraint (9) in the hydrogen energy storage system t = H t + W H where W H is a control parameter;​ 22) Define the Lyapunov drift - penalty function; 23) Transform the solution of the Lyapunov drift-penalty function into the problem of minimizing the upper bound of the Lyapunov drift-penalty function, that is where V is the weight parameter; 24) Further relax the problem of minimizing the upper bound of the Lyapunov drift - penalty function into the form of sub - problem P2, where: The control parameter W in step 21) and step 23) H and the weight parameter V satisfy and Let the variable variable and Variable and are defined as follows: If then and If then and Step 3: Based on the observed system state parameters, solve the current - time - slot optimization sub - problem, and use the optimal solution of this sub - problem for the operation control of the hydrogen - containing energy system; Step 4: Update the virtual queue related to the hydrogen energy storage system to obtain the virtual queue length of the next time - slot; Step 5: Repeat Step 3 - Step 4 until the current time - slot number is greater than the specified optimization time - slot number and then terminate.

2. According to the online optimal operation method for a hydrogen - containing energy system under uncertain environment described in claim 1, characterized in that, In step 3, the optimal solutions of the single-slot optimization sub-problem include the optimal solution of the electricity purchase quantity, the optimal solution of the actual photovoltaic power generation quantity, the optimal solution of the electrolyzer input power, the optimal solution of the fuel cell output power, and the optimal solution of the hydrogen purchase quantity, respectively.

3. According to the online optimal operation method for a hydrogen - containing energy system under uncertain environment described in claim 1, characterized in that, In Step 4, update the virtual queue related to the hydrogen energy storage system, and the dynamic update process is as follows:

4. An online optimal operation system for a hydrogen - containing energy system under uncertain environment, characterized in that, it includes the following modules: Model establishment module: Under the uncertain environment of the output of renewable energy generators, establish a mathematical optimization model for minimizing the long - term operation cost of the hydrogen - containing energy system; The objective function expression of the mathematical optimization model is as follows: Where: C 1,t represents the cost of purchasing electricity in time slot t, where 1 ≤ t ≤ T and T represents the specified number of optimization time slots; C 2,t represents the cost of PV curtailment in time slot t; C 3,t represents the cost of hydrogen purchase in time slot t; C 4,t represents the cost of the operation / shutdown status, start / stop of each device in time slot t; C 5,t represents the carbon emission cost in time slot t, and E is the expectation operator; The specific expressions of each operation cost in the objective function are as follows: C 1,t = S g,t P grid,t Δt (2) C 3,t = S h,t m buy,t (4) C 5,t = α ce (μ g P grid,t Δt + μ pv P pv,t Δt + μ h m buy,t ) (6) where: Δt is the time interval for the regulation of the hydrogen - containing energy system; S g,t is the electricity price at time slot t, P grid,t is the electricity quantity purchased from the power grid at time slot t; α pv is the PV curtailment penalty coefficient, is the maximum PV power generation at time slot t, P pv,t is the actual PV power generation at time slot t; S h,t is the hydrogen price at time slot t, m buy,t is the hydrogen quantity purchased from the hydrogen market at time slot t; are the costs of equipment X for operation, startup, and shutdown respectively, are the states of equipment X for operation, shutdown, startup - shutdown respectively; α ce is the carbon emission penalty coefficient, μ g represents the carbon emission rate related to the electricity purchased from the main power grid, μ pv represents the carbon emission rate of the PV system, μ h is the carbon emission coefficient of the hydrogen purchased from the hydrogen market; The constraint conditions of the mathematical optimization model include: Photovoltaic power generation system constraint, the expression is as follows: Where: η pv is the photovoltaic power generation efficiency, h pv is the total radiation area of the solar panel, l t is the solar radiation intensity at time slot t; is the maximum power demand in the hydrogen - containing energy system; Hydrogen energy storage system constraint, the expression is as follows: Where: H t+1 and H t are the hydrogen storage levels at time slots t + 1 and t respectively; ω el and ω fc are the conversion coefficients of the electrolyzer and fuel cell respectively; P el,t and P fc,t are the input power of the electrolyzer and the output power of the fuel cell at time slot t respectively; m buy,t is the amount of hydrogen purchased from the hydrogen market at time slot t, and H min and H max are the minimum and maximum storage levels of the hydrogen storage tank respectively; and M B are the maximum input power of the electrolyzer, the maximum output power of the fuel cell, and the maximum purchase amount of hydrogen that can be bought from the primary energy to the hydrogen market respectively; System power supply - demand balance constraint, the expression is as follows: Where: P load,t is the power demand in time slot t; P grid,t is the electricity quantity purchased from the main power grid in time slot t, is the maximum electricity quantity that can be purchased from the main power grid in time slot t; The decision variables of the mathematical optimization model are: {P gird,t ,P pv,t ,P el,t ,P fc,t ,m buy,t} set; Single - time - slot optimization module: Use the Lyapunov optimization framework to decompose the established long - term operation cost optimization problem of the hydrogen - containing energy system into multiple single - time - slot optimization sub - problems; The single - time - slot optimization sub - problem P2 is as follows: (P2)min Q t χ t +Vy t (17) s.t.(7)-(8),(11)-(16) (18) The process of decomposing the established long - term operation cost optimization problem of the hydrogen - containing energy system into multiple single - time - slot optimization sub - problems includes the following steps: (21) Construct the virtual queue Q related to the time-coupling constraint (9) in the hydrogen energy storage system related to the virtual queue Q t = H t + W H , W H is a control parameter; 22) Define the Lyapunov drift-penalty function; 23) Transforming the solution of the Lyapunov drift-penalty function into the problem of minimizing the upper bound of the Lyapunov drift-penalty function, that is where V is the weight parameter; 24) Further relax the problem of minimizing the upper bound of the Lyapunov drift-penalty function into the form of sub-problem P2, where: The control parameter W in the steps 21) and 23) H and the weight parameter V satisfy and Let the variable variable and Variable and are defined as follows: If then and If then and Control module: Based on the observed system state parameters, solve the optimization sub-problem for the current time slot, and use the optimal solution of the sub-problem for the operation control of the hydrogen energy system; Virtual queue module: Update the virtual queue related to the hydrogen energy storage system to obtain the virtual queue length in the next time slot; Loop module: Repeat the working processes of the control module and the virtual queue module until the current time slot number is greater than the specified number of optimized time slots and then terminate.

5. A computer-readable storage medium for storing the online optimization operation method of the hydrogen energy system under uncertain environment according to any one of claims 1-3.

Images