VMD-MPC-based off-grid hydrogen production system energy management scheduling method

Abstract

The invention relates to an off-grid hydrogen production system energy management scheduling method based on VMD-MPC, and relates to the field of new energy power generation energy management. According to the technical scheme, a low-frequency trend component and a high-frequency disturbance component in a power signal are stripped through variational mode decomposition VMD, a layered optimization framework of day-ahead global planning-intraday dynamic correction is constructed, and the pain point that an existing single time scale model lags in response to dynamic change of a system is effectively solved. In addition, according to the scheme, the input signal quality is optimized through VMD preprocessing, and the input stability of the model predictive control MPC scheduling model is remarkably improved. Besides, a deviation dynamic self-adaptive penalty mechanism is introduced into an intra-day rolling optimization link, so that a scheduling strategy can automatically adjust the tracking intensity according to the actual fluctuation amplitude and flexibly adapt to the change of working conditions, and the anti-interference ability of the system to deal with the uncertainty of wind and light output and load is greatly enhanced.

Description

Technical Field

[0001] This invention relates to the field of energy management technology for new energy power generation, and in particular to an energy management and scheduling method for off-grid hydrogen production systems based on VMD-MPC. Background Technology

[0002] In recent years, in response to the national strategy for sustainable energy development, the installed capacity of photovoltaic and wind power generation has continued to climb, strongly driving the rapid development of off-grid renewable energy hydrogen production systems. These systems are characterized by multi-energy flow coupling, significant multi-timescale features, and complex operating conditions. They must simultaneously address multiple challenges, including fluctuations in wind and solar power output, dynamic load changes, and differences in equipment response characteristics, to achieve economical and efficient operation while ensuring stable system operation.

[0003] In the field of off-grid system scheduling model research, existing technologies are mostly limited to day-ahead single-timescale optimization. However, the volatility of renewable energy in off-grid wind-solar-hydrogen-storage systems exhibits significant heterogeneity across different time scales, making single-time-scale scheduling strategies difficult to adapt to this characteristic. Therefore, in-depth research on multi-time-scale collaborative optimization scheduling schemes is urgently needed. On the other hand, existing optimization scheduling techniques mostly employ model predictive control (MPC) algorithms alone. However, wind, solar, and load signals are highly non-stationary, and the optimization effect of MPC is highly dependent on the accuracy of the input signals, significantly limiting its scheduling accuracy and applicability.

[0004] Traditional energy management methods generally adopt a single time-scale dispatching model. Under actual operating conditions where wind and solar forecasting errors are large and equipment dynamic response capabilities vary significantly, a series of problems are likely to occur: First, dispatching costs increase, as forecasting errors lead to frequent start-ups and shutdowns and power adjustments of electrolyzers and energy storage equipment, significantly increasing equipment maintenance and operating energy costs. Second, wind and solar curtailment is severe, as the system struggles to respond quickly to highly fluctuating power changes and cannot fully absorb renewable energy, resulting in energy waste. Third, system stability decreases, as intraday dispatching uses fixed weight settings, making it difficult to adapt to dynamic changes in operating conditions, leading to increased system operational fluctuations. Fourth, dispatching robustness is insufficient, as directly using raw, non-stationary forecasting signals as dispatching input without effective preprocessing significantly reduces the accuracy and reliability of dispatching decisions. Summary of the Invention

[0005] The purpose of this invention is to provide an energy management and scheduling method for off-grid hydrogen production systems based on VMD-MPC, so as to solve the problems existing in the prior art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides an energy management and scheduling method for an off-grid hydrogen production system based on VMD-MPC, the method comprising: Modeling is performed on the renewable energy module, electricity storage module, hydrogen storage module, and energy consumption module of the off-grid new energy hydrogen production system. Variational mode decomposition (VMD) is performed on the power forecast values ​​of wind power, photovoltaic power, electric load and hydrogen load to obtain finite bandwidth subsequences with different frequency characteristics. The finite bandwidth subsequences include low-frequency components that characterize the overall trend of change and high-frequency components that characterize random fluctuations. A day-ahead optimization scheduling model is established based on the low-frequency components to generate power scheduling plans for each device. A penalty factor based on dynamic adjustment of deviation is introduced, and an intraday rolling optimization scheduling model is established based on the high-frequency component and the power scheduling plan. The model predicts and controls the MPC to perform rolling optimization and outputs closed-loop termination instructions. The off-grid new energy hydrogen production system is controlled by the closed-loop step command to achieve energy balance and stable control.

[0007] In some implementations, the modeling of the renewable energy module, electricity storage module, hydrogen storage module, and energy consumption module of the off-grid new energy hydrogen production system includes: Establish a lithium battery energy storage model, wherein the lithium battery energy storage model satisfies: The state-of-charge expression for a lithium battery at each time t is: ; In the formula, The state of charge of the lithium battery at time t; This refers to the rated capacity of the lithium battery. For time intervals; The charging efficiency of lithium batteries; The charging power of the lithium battery at time t; Let be the discharge power of the lithium battery at time t; The discharge efficiency of lithium batteries; At each time t, the charging and discharging power constraints of the lithium battery are: ; In the formula, This represents the maximum discharge power of the lithium battery. The value is a 0-1 variable representing the charging and discharging status of the lithium battery at time t. A value of 1 indicates that the lithium battery is in the discharging state, while a value of 0 indicates that the lithium battery is in the charging state. This refers to the maximum charging power of the lithium battery. At each time t, the state of charge constraint of the lithium battery is: ; In the formula, This represents the state of charge limit of the lithium battery at time t; This represents the upper limit of the state of charge of the lithium battery at time t; The state of charge of the lithium battery at the last moment of the day; This represents the state of charge of the lithium battery at the beginning of the day.

[0008] In some embodiments, the step of modeling the renewable energy module, electricity storage module, hydrogen storage module, and energy consumption module of the off-grid new energy hydrogen production system further includes: Establish an electrolytic cell model, wherein the electrolytic cell model satisfies: The power consumption expression for an electrolytic cell is: ; In the formula, Let be the electrical power consumption of the i-th electrolytic cell at time t; Let be the hydrogen production of the i-th electrolyzer at time t; The hydrogen production efficiency of the electrolyzer; The power constraint of the electrolytic cell is: ; In the formula, This represents the maximum value of the power ramp-up variation in the electrolytic cell; This represents the maximum output power of the electrolytic cell; Let t be the 0-1 variable representing the operating condition of the i-th electrolytic cell at time t. If 1, it indicates that the cell is started. This represents the minimum output power of the electrolytic cell; The start-up and shutdown constraints of the electrolytic cell are: ; In the formula, Let be a 0-1 variable representing the startup status of the i-th electrolytic cell at time t. If 1, it indicates startup. This is a 0-1 variable representing the shutdown status of the i-th electrolytic cell at time t. A value of 1 indicates shutdown.

[0009] In some embodiments, the step of modeling the renewable energy module, electricity storage module, hydrogen storage module, and energy consumption module of the off-grid new energy hydrogen production system further includes: Establish a hydrogen storage tank model, wherein the hydrogen storage tank model satisfies: The hydrogen storage state expression of the hydrogen storage tank at each time t is: ; In the formula, The hydrogen storage state of the hydrogen storage tank at time t; The value represents the mass of hydrogen absorbed or released by the hydrogen storage tank at time t. A positive value indicates the absorption of hydrogen, and a negative value indicates the release of hydrogen. This refers to the rated capacity of the hydrogen storage tank. At each time t, the hydrogen storage state constraint of the hydrogen storage tank is: ; In the formula, This represents the minimum value of the hydrogen storage state in the hydrogen storage tank; This represents the maximum value of the hydrogen storage state in the hydrogen storage tank; This represents the final hydrogen storage state of the hydrogen storage tank within one day. This represents the initial hydrogen storage state of the hydrogen storage tank within one day.

[0010] In some embodiments, performing variational mode decomposition (VMD) on the power predictions of wind power, photovoltaic power, electrical load, and hydrogen load includes: Obtain power forecasts for wind power, photovoltaic power, electrical load, and hydrogen load at 24 specific times. The modal components K and the quadratic penalty factor α of VMD were determined using a grid search method. The power prediction value is decomposed into four finite bandwidth subsequences by VMD, where the first two subsequences are low-frequency components used for day-ahead optimization models, and the latter two subsequences are high-frequency components used for intraday rolling optimization models.

[0011] In some implementations, the step of establishing a day-ahead optimized scheduling model based on the low-frequency components and generating power scheduling plans for each device includes: The optimization period of the day-ahead optimization scheduling model is set to 24 hours, the optimization step size is 1 hour, and the input is the low-frequency mode component obtained by VMD decomposition. The optimization objective is to minimize the total scheduling cost of the off-grid renewable energy hydrogen production system. This total scheduling cost includes system operation and maintenance costs, electrolyzer start-up and shutdown costs, energy loss costs from the electricity-to-hydrogen conversion, and costs associated with wind and solar power curtailment. The expression for this cost is: ; In the formula, To optimize scheduling costs in the near future; For system operation and maintenance costs; Costs associated with starting and stopping the electrolytic cell; Cost of energy loss during the electro-hydrogen conversion; Costs associated with wind and solar power curtailment; including system operation and maintenance costs. This is the sum of the operation and maintenance costs of wind and solar turbines, lithium batteries, and electrolytic cells; Set constraints, including equipment operating state constraints, electrolyzer operating power constraints, battery SOC boundary constraints, and hydrogen supply and demand balance constraints. Solve the day-ahead optimization scheduling model and output the hourly power scheduling plan for electrolyzers and energy storage devices as a reference benchmark for intraday scheduling.

[0012] In some implementations, the penalty factor is a dynamic function that varies with power deviation, expressed as: ; In the formula, As a penalty factor; Basic weights; This is the adjustment coefficient; The high-frequency predicted power after VMD decomposition; This represents the day-ahead scheduling plan value for the corresponding time period.

[0013] In some implementations, when the high-frequency predicted power after VMD decomposition... The daily scheduling plan value for the corresponding time period When the deviation is large, the off-grid new energy hydrogen production system automatically increases the penalty factor. Increase the tracking intensity to ensure stable system operation; when the high-frequency prediction power after VMD decomposition... The daily scheduling plan value for the corresponding time period When the deviation is small, the off-grid new energy hydrogen production system weakens the tracking, reduces unnecessary frequent adjustments, and optimizes equipment lifespan and economy.

[0014] In some implementations, the step of establishing an intraday rolling optimization scheduling model based on the high-frequency components and the power scheduling plan, and executing rolling optimization and outputting closing-loop instructions through model prediction control MPC includes: The optimization interval of the intraday rolling optimization model is set to 15 minutes, the scheduling cycle is 4 hours, and the input is the high-frequency component obtained from VMD decomposition and the day-ahead power scheduling plan. Construct the optimization objective function: ; In the formula, The total cost of system scheduling; This refers to the system power or equipment adjustment amount obtained from intraday model optimization. This represents the day-ahead scheduling plan value for the corresponding time period; Solve the optimization objective function within each rolling window, and execute only the scheduling result of the first time interval as the closing node instruction; Enter the next round of rolling optimization, continuously update until full coverage of all-day scheduling is achieved, forming MPC closed-loop feedback control.

[0015] In some implementations, the off-grid new energy hydrogen production system adopts a common DC bus configuration. Wind and photovoltaic power generation supply electrical loads and electrolyzers through AC / DC converters and DC / DC converters. Lithium batteries mitigate short-term fluctuations in wind and photovoltaic power generation. Electrolyzers convert electricity into hydrogen energy. Hydrogen storage tanks provide physical storage of hydrogen and buffer against hydrogen energy fluctuations.

[0016] The beneficial effects of the technical solution provided by this invention include at least the following: This technical solution uses Variational Mode Decomposition (VMD) to remove low-frequency trend components and high-frequency disturbance components from the power signal, constructing a hierarchical optimization framework of day-ahead global planning and intraday dynamic correction. This effectively addresses the pain point of existing single-timescale models' lagging response to system dynamic changes. Furthermore, to address the problem that traditional methods are difficult to directly and effectively utilize due to the strong non-stationarity of the original wind, solar, and load signals, this solution optimizes the input signal quality through VMD preprocessing, significantly improving the input stability of the Model Predictive Control (MPC) scheduling model. In addition, a dynamic adaptive penalty mechanism for deviation is introduced in the intraday rolling optimization stage, allowing the scheduling strategy to autonomously adjust the tracking intensity according to the actual fluctuation amplitude, flexibly adapting to changes in operating conditions and greatly enhancing the system's anti-interference capability against uncertainties in wind and solar power output and load. 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 The diagram illustrates a flowchart of an off-grid hydrogen production system energy management and scheduling method based on VMD-MPC, provided by an exemplary embodiment of the present invention.

[0019] Figure 2 The diagram illustrates an optimized scheduling method for an off-grid hydrogen production system based on VMD-MPC, provided by an exemplary embodiment of the present invention.

[0020] Figure 3 The diagram illustrates a solution for an off-grid hydrogen production system energy management and scheduling method based on VMD-MPC, provided by an exemplary embodiment of the present invention.

[0021] Figure 4 The diagram illustrates a structural block diagram of an off-grid new energy hydrogen production energy management system based on VMD-MPC, provided by an exemplary embodiment of the present invention. Detailed Implementation

[0022] 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, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0024] Figure 1 The diagram illustrates a flowchart of an off-grid hydrogen production system energy management and scheduling method based on VMD-MPC, according to an exemplary embodiment of the present invention. Figure 2 This illustration shows an optimized scheduling diagram of an off-grid hydrogen production system energy management and scheduling method based on VMD-MPC, provided by an exemplary embodiment of the present invention. Figure 3 The diagram illustrates a solution schematic for an off-grid hydrogen production system energy management and scheduling method based on VMD-MPC, provided by an exemplary embodiment of the present invention. This VMD-MPC-based off-grid hydrogen production system energy management and scheduling method includes: Step 101: Model the renewable energy module, electricity storage module, hydrogen storage module, and energy consumption module of the off-grid new energy hydrogen production system respectively.

[0025] In some embodiments, the renewable energy module, electricity storage module, hydrogen storage module, and energy consumption module of the off-grid new energy hydrogen production system are modeled respectively, including: A lithium battery energy storage model is established. Lithium batteries possess rapid response characteristics, effectively mitigating power fluctuations caused by wind and solar power output. To ensure stable operation of lithium batteries and delay their performance degradation, key performance characteristics such as charge / discharge power, state of charge (SOC), and SOC continuity must be comprehensively considered when establishing the equipment model. The lithium battery energy storage model satisfies the following: The state-of-charge expression for a lithium battery at each time t is: ; In the formula, The state of charge of the lithium battery at time t; This refers to the rated capacity of the lithium battery. For time intervals; The charging efficiency of lithium batteries; The charging power of the lithium battery at time t; Let be the discharge power of the lithium battery at time t; The discharge efficiency of lithium batteries; At each time t, the charging and discharging power constraints of the lithium battery are: ; In the formula, This represents the maximum discharge power of the lithium battery. The value is a 0-1 variable representing the charging and discharging status of the lithium battery at time t. A value of 1 indicates that the lithium battery is in the discharging state, while a value of 0 indicates that the lithium battery is in the charging state. This refers to the maximum charging power of the lithium battery. At each time t, the state of charge constraint of the lithium battery is: ; In the formula, This represents the state of charge limit of the lithium battery at time t; This represents the upper limit of the state of charge of the lithium battery at time t; The state of charge of the lithium battery at the last moment of the day; This represents the state of charge of the lithium battery at the beginning of the day.

[0026] In this embodiment, lithium battery energy storage serves as a mechanism to mitigate fluctuations in wind and solar power output. By quantifying parameters such as charge / discharge power, state of charge (SOC), and continuous constraints, it can match the rapid response characteristics of lithium batteries with the system's fluctuation regulation requirements. This ensures that lithium batteries can promptly replenish or release energy when wind and solar power output surges or drops. Furthermore, by using upper and lower limits for SOC and charge / discharge power threshold constraints, it avoids battery performance degradation or safety risks caused by overcharging and over-discharging, thus extending the equipment's service life. The setting of continuous SOC constraints in the model, consistent with the initial and final states throughout the day, ensures that lithium battery operation meets the energy balance requirements of the entire day, preventing energy storage imbalances after the end of a single day's scheduling from affecting the operation the following day. The introduction of 0-1 variables for charge / discharge conditions provides a decision boundary for subsequent optimization algorithms to distinguish charge / discharge modes and allocate power.

[0027] In some embodiments, the renewable energy module, electricity storage module, hydrogen storage module, and energy consumption module of the off-grid new energy hydrogen production system are modeled separately, and the modeling also includes: An electrolyzer model is established. The electrolyzer is the foundation of green hydrogen production, and its performance affects the economics and stability of green hydrogen. The electrolyzer model satisfies the following: The power consumption expression for an electrolytic cell is: ; In the formula, Let be the electrical power consumption of the i-th electrolytic cell at time t; Let be the hydrogen production of the i-th electrolyzer at time t; The hydrogen production efficiency of the electrolyzer; The power constraint of the electrolytic cell is: ; In the formula, This represents the maximum value of the power ramp-up variation in the electrolytic cell; This represents the maximum output power of the electrolytic cell; Let t be the 0-1 variable representing the operating condition of the i-th electrolytic cell at time t. If 1, it indicates that the cell is started. This represents the minimum output power of the electrolytic cell; The start-up and shutdown constraints of the electrolytic cell are: ; In the formula, Let be a 0-1 variable representing the startup status of the i-th electrolytic cell at time t. If 1, it indicates startup. This is a 0-1 variable representing the shutdown status of the i-th electrolytic cell at time t. A value of 1 indicates shutdown.

[0028] In this embodiment, by constructing a quantitative correlation between power consumption and hydrogen production, the energy conversion law of the electrolyzer is clarified. This helps the system calculate the hydrogen production efficiency under different power inputs, avoiding energy waste or insufficient hydrogen production caused by blind adjustment. The setting of power constraints and ramp limits fully adapts to the physical operating characteristics of the electrolyzer. Since the power adjustment of the electrolyzer has inertia, excessive or rapid power fluctuations can easily lead to equipment damage and decreased hydrogen production efficiency. By quantifying the upper and lower limits of power and the maximum ramp value, the model can effectively avoid such risks and ensure that the electrolyzer operates within a safe operating range. At the same time, the introduction of the 0-1 variable of the operating condition defines the start-up and shutdown state boundaries of the electrolyzer, providing a clear decision basis for the coordinated scheduling of multiple electrolyzers, and facilitating the system to flexibly allocate the number of operating electrolyzers according to the fluctuation of wind and solar power output. The start-up and shutdown state constraints further standardize the operating logic of the electrolyzer. By limiting the mutual exclusivity of start-up and shutdown states at the same time, the mechanical losses and energy consumption caused by frequent start-up and shutdown of equipment are avoided.

[0029] In some embodiments, the renewable energy module, electricity storage module, hydrogen storage module, and energy consumption module of the off-grid new energy hydrogen production system are modeled separately, and the modeling also includes: A hydrogen storage tank model was established. The hydrogen storage tank can mitigate fluctuations in hydrogen energy through the filling and releasing of hydrogen. Since the technology of gaseous hydrogen storage tanks is mature and has high dynamic response characteristics, gaseous hydrogen storage tanks were chosen as the hydrogen storage device. The hydrogen storage tank model satisfies the following: The hydrogen storage state expression of the hydrogen storage tank at each time t is: ; In the formula, The hydrogen storage state of the hydrogen storage tank at time t; The value represents the mass of hydrogen absorbed or released by the hydrogen storage tank at time t. A positive value indicates the absorption of hydrogen, and a negative value indicates the release of hydrogen. This refers to the rated capacity of the hydrogen storage tank. At each time t, the hydrogen storage state constraint of the hydrogen storage tank is: ; In the formula, This represents the minimum value of the hydrogen storage state in the hydrogen storage tank; This represents the maximum value of the hydrogen storage state in the hydrogen storage tank; This represents the final hydrogen storage state of the hydrogen storage tank within one day. This represents the initial hydrogen storage state of the hydrogen storage tank within one day.

[0030] In this embodiment, by quantifying the correlation between hydrogen storage status and the mass of hydrogen absorbed and released, the instantaneous fluctuations in hydrogen production can be smoothed out through the regulation of hydrogen absorption and release from the hydrogen storage tank. This avoids energy waste caused by excessive hydrogen production or load shortages caused by insufficient hydrogen production, thus achieving a dynamic balance between hydrogen supply and demand. The setting of hydrogen storage status constraints, by clearly defining the upper and lower limits of hydrogen storage, can effectively avoid safety risks caused by overpressure or underpressure operation of the hydrogen storage tank. Furthermore, the constraint of consistent hydrogen storage status throughout the day ensures the periodic steady-state operation of the system, preventing imbalances in hydrogen storage status after the end of a single day's scheduling from affecting the hydrogen supply and demand allocation the following day.

[0031] In some embodiments, see Figure 4 The off-grid new energy hydrogen production system adopts a common DC bus configuration. Wind and photovoltaic power generation supply electrical loads and electrolyzers through AC / DC converters and DC / DC converters. Lithium batteries mitigate short-term fluctuations in wind and photovoltaic power generation. Electrolyzers convert electricity into hydrogen energy. Hydrogen storage tanks enable physical storage of hydrogen and buffer against hydrogen energy fluctuations.

[0032] In this embodiment, compared to the traditional AC bus architecture, the common DC bus significantly simplifies the energy conversion link. Through the matching of AC / DC and DC / DC converters, wind and solar power can be directly supplied to the electrical load and electrolyzer in DC form, reducing energy loss caused by multiple AC-DC conversions and significantly improving energy transmission efficiency. Lithium batteries, with their fast response characteristics, smooth out short-term high-frequency fluctuations in wind and solar power output. Hydrogen storage tanks, through the storage and release of hydrogen energy, address medium- to long-term fluctuations in hydrogen supply and demand, while also absorbing excess electrical energy that cannot be absorbed by lithium batteries (converted into hydrogen energy for storage through the electrolyzer), achieving cross-period energy allocation and improving energy utilization. Furthermore, when wind and solar power output is sufficient, electrical energy can be prioritized for supply to the electrical load, with the surplus converted into hydrogen energy for storage through the electrolyzer; when wind and solar power output is insufficient, hydrogen storage resources can provide reverse energy replenishment (or the lithium battery can directly release energy), ensuring a stable supply to the load.

[0033] Step 102: Perform variational mode decomposition (VMD) on the power prediction values ​​of wind power, photovoltaic, electric load and hydrogen load to obtain finite bandwidth subsequences with different frequency characteristics. The finite bandwidth subsequences include low-frequency components that characterize the overall trend of change and high-frequency components that characterize random fluctuations.

[0034] In some embodiments, variational mode decomposition (VMD) is performed on the power forecasts of wind power, photovoltaic power, electrical load, and hydrogen load, including: Obtain power forecasts for wind power, photovoltaic power, electrical load, and hydrogen load at 24 specific times. The modal components K and the quadratic penalty factor α of VMD were determined using a grid search method. The power prediction values ​​are decomposed into four finite bandwidth subsequences by VMD. The first two subsequences are low-frequency components, which are used for the day-ahead optimization model, and the last two subsequences are high-frequency components, which are used for the intraday rolling optimization model.

[0035] In this embodiment, VMD is a completely non-recursive, variable-mode, adaptive quasi-orthogonal signal decomposition method with strong harmonic separation capabilities. It can solve the mode aliasing and endpoint effects problems existing in empirical mode decomposition and ensemble empirical mode decomposition, significantly reducing the non-stationarity of time series with strong nonlinearity and high complexity. Using VMD to decompose wind and solar power output and load sequences, K modal components with different frequency characteristics can be obtained. Using the VMD method, the power prediction values ​​of wind power, photovoltaic, electrical load, and hydrogen load at 24 full-hour intervals are decomposed, and a grid search method is used to determine the modal components K and the quadratic penalty factor α. The power forecasts for wind power, photovoltaic power, and electricity and hydrogen loads at 24 full hours are decomposed using VMD to obtain four finite bandwidth subsequences representing different frequency scales. Subsequences 1 and 2 represent the overall trend of the curve and will be used for slow power scheduling within off-grid renewable energy hydrogen production systems, i.e., day-ahead optimization model. Subsequences 3 and 4 represent the random fluctuations of the curve at different frequency scales and will be used for rapid power scheduling within off-grid renewable energy hydrogen production systems, i.e., intraday rolling optimization model.

[0036] Step 103: Establish a day-ahead optimization scheduling model based on low-frequency components to generate power scheduling plans for each device.

[0037] In some embodiments, a day-ahead optimization scheduling model is established based on low-frequency components to generate power scheduling plans for each device, including: The optimization period of the day-ahead optimization scheduling model is set to 24 hours, the optimization step size is 1 hour, and the input is the low-frequency mode component obtained by VMD decomposition. The optimization objective is to minimize the total scheduling cost of off-grid renewable energy hydrogen production systems. The total scheduling cost includes system operation and maintenance costs, electrolyzer start-up and shutdown costs, energy loss costs from electricity-to-hydrogen conversion, and costs associated with wind and solar curtailment. The expression is: ; In the formula, To optimize scheduling costs in the near future; For system operation and maintenance costs; Costs associated with starting and stopping the electrolytic cell; Cost of energy loss during the electro-hydrogen conversion; Costs associated with wind and solar power curtailment; including system operation and maintenance costs. This is the sum of the operation and maintenance costs of wind and solar turbines, lithium batteries, and electrolytic cells; Set constraints, including equipment operating status constraints, electrolyzer operating power constraints, battery SOC boundary constraints, and hydrogen supply and demand balance constraints. Solve the day-ahead optimization scheduling model and output the hourly power scheduling plan for electrolyzers and energy storage devices as a reference benchmark for intraday scheduling.

[0038] In this embodiment, compared to planning directly based on the original signal, introducing the low-frequency component after VMD decomposition as input can filter out the interference of high-frequency random disturbances on scheduling decisions, focus on the overall trend of energy supply and demand changes, and make day-ahead scheduling planning more in line with the long-term operating pattern of the system. The setting of multi-dimensional constraints defines the safe operating boundary for the scheduling plan, ensuring that the power allocation scheme is compatible with the physical characteristics of the equipment and the supply and demand balance requirements, and avoiding risks such as equipment over-limit operation and hydrogen supply and demand imbalance.

[0039] Step 104: Introduce a penalty factor based on dynamic adjustment of deviation, establish an intraday rolling optimization scheduling model based on high-frequency components and power scheduling plan, and execute rolling optimization and output closed-loop point instructions through model predictive control MPC.

[0040] In some embodiments, the penalty factor is a dynamic function that varies with power deviation, expressed as: ; In the formula, As a penalty factor; Basic weights; This is the adjustment coefficient; The high-frequency predicted power after VMD decomposition; This represents the day-ahead scheduling plan value for the corresponding time period.

[0041] High-frequency predicted power after VMD decomposition The daily scheduling plan value for the corresponding time period When the deviation is large, the off-grid new energy hydrogen production system automatically increases the penalty factor. Increase the tracking intensity to ensure stable system operation; when the high-frequency prediction power after VMD decomposition... The daily scheduling plan value for the corresponding time period When the deviation is small, the off-grid new energy hydrogen production system weakens the tracking, reduces unnecessary frequent adjustments, and optimizes equipment lifespan and economy.

[0042] In this embodiment, a high-frequency component is introduced as input to specifically capture short-term random fluctuations in wind and solar loads. Combined with the rolling optimization characteristics of MPC, real-time correction of scheduling decisions is achieved, ensuring that instructions closely match dynamic changes in operating conditions. The dynamic penalty factor enables adaptive adjustment of the control intensity. By dynamically adjusting the weight according to the power deviation, tracking is strengthened when the deviation is large to ensure system stability; and the adjustment is weakened when the deviation is small to reduce ineffective start-ups and shutdowns and losses of equipment.

[0043] In some embodiments, an intraday rolling optimization scheduling model is established based on high-frequency components and power scheduling plans. Model predictive control (MPC) executes rolling optimization and outputs closing-loop termination instructions, including: The optimization interval of the intraday rolling optimization model is set to 15 minutes, the scheduling cycle is 4 hours, and the input is the high-frequency component obtained from VMD decomposition and the day-ahead power scheduling plan. Construct the optimization objective function: ; In the formula, The total cost of system scheduling; This refers to the system power or equipment adjustment amount obtained from intraday model optimization. This represents the day-ahead scheduling plan value for the corresponding time period; Solve the optimization objective function within each rolling window, and execute only the scheduling result of the first time interval as the closing node instruction; Enter the next round of rolling optimization, continuously update until full coverage of all-day scheduling is achieved, forming MPC closed-loop feedback control.

[0044] In this embodiment, the combination of a 15-minute optimization interval and a 4-hour scheduling cycle ensures the real-time nature of scheduling decisions, enabling rapid response to sudden changes in operating conditions, and achieving forward-looking control through cycle coverage. The objective function focuses on controlling the deviation between the actual adjustment amount and the daily plan. Combined with the logic of solving using a rolling window and executing only the first interval instruction, it effectively filters high-frequency noise interference and improves scheduling accuracy. The MPC closed-loop feedback mechanism achieves dynamic correction of scheduling instructions through continuous iterative updates, significantly enhancing the system's ability to resist fluctuations.

[0045] Step 105: Control the operation of each module of the off-grid new energy hydrogen production system according to the closed-loop step instructions to achieve energy balance and stable control.

[0046] In this embodiment, the closed-loop control instructions have been verified through MPC rolling optimization and are fully adapted to high-frequency fluctuations and equipment characteristic constraints. By being distributed to each module, coordinated linkage between wind and solar power generation, energy storage, hydrogen production, and energy consumption can be achieved. In this case, power imbalances caused by fluctuations in wind and solar power output or sudden load changes are avoided, and ineffective equipment actions are reduced through unified control.

[0047] In summary, this technical solution uses Variational Mode Decomposition (VMD) to remove low-frequency trend components and high-frequency disturbance components from the power signal, constructing a hierarchical optimization framework of day-ahead global planning and intraday dynamic correction. This effectively addresses the pain point of existing single-timescale models' lagging response to system dynamic changes. Furthermore, addressing the problem that traditional methods are difficult to directly and effectively utilize due to the strong non-stationarity of the original wind, solar, and load signals, this solution optimizes the input signal quality through VMD preprocessing, significantly improving the input stability of the Model Predictive Control (MPC) scheduling model. In addition, a dynamic adaptive penalty mechanism for deviation is introduced in the intraday rolling optimization stage, allowing the scheduling strategy to autonomously adjust the tracking intensity according to the actual fluctuation amplitude, flexibly adapting to changes in operating conditions and greatly enhancing the system's anti-interference capability against uncertainties in wind and solar power output and load.

[0048] It is understood that the specific examples in this document are only intended to help those skilled in the art better understand this disclosure, and are not intended to limit the scope of the invention.

[0049] It is understood that in the various embodiments of this specification, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of this disclosure.

[0050] It is understood that the various implementation methods described in this specification can be implemented individually or in combination, and this disclosure does not limit them.

[0051] Unless otherwise stated, all technical and scientific terms used in this disclosure have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of this specification. The term "and / or" as used in this specification includes any and all combinations of one or more of the associated listed items. The singular forms "a," "the," and "the" as used in this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0052] The above description is merely a specific embodiment of this specification, but the scope of protection of this invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this specification should be included within the scope of protection of this specification. Therefore, the scope of protection of this invention should be determined by the scope of the claims.

Claims

1. An energy management and scheduling method for an off-grid hydrogen production system based on VMD-MPC, characterized in that, The method includes: Modeling is performed on the renewable energy module, electricity storage module, hydrogen storage module, and energy consumption module of the off-grid new energy hydrogen production system. Variational mode decomposition (VMD) is performed on the power forecast values ​​of wind power, photovoltaic power, electric load and hydrogen load to obtain finite bandwidth subsequences with different frequency characteristics. The finite bandwidth subsequences include low-frequency components that characterize the overall trend of change and high-frequency components that characterize random fluctuations. A day-ahead optimization scheduling model is established based on the low-frequency components to generate power scheduling plans for each device. A penalty factor based on dynamic adjustment of deviation is introduced, and an intraday rolling optimization scheduling model is established based on the high-frequency component and the power scheduling plan. The model predicts and controls the MPC to perform rolling optimization and outputs closed-loop termination instructions. The off-grid new energy hydrogen production system is controlled by the closed-loop step command to achieve energy balance and stable control.

2. The energy management and scheduling method for off-grid hydrogen production systems based on VMD-MPC according to claim 1, characterized in that, The modeling of the renewable energy module, electricity storage module, hydrogen storage module, and energy consumption module of the off-grid new energy hydrogen production system includes: Establish a lithium battery energy storage model, wherein the lithium battery energy storage model satisfies: The state-of-charge expression for a lithium battery at each time t is: ； In the formula, The state of charge of the lithium battery at time t; This refers to the rated capacity of the lithium battery. For time intervals; The charging efficiency of lithium batteries; The charging power of the lithium battery at time t; Let be the discharge power of the lithium battery at time t; The discharge efficiency of lithium batteries; At each time t, the charging and discharging power constraints of the lithium battery are: ； In the formula, This represents the maximum discharge power of the lithium battery. The value is a 0-1 variable representing the charging and discharging status of the lithium battery at time t. A value of 1 indicates that the lithium battery is in the discharging state, while a value of 0 indicates that the lithium battery is in the charging state. This refers to the maximum charging power of the lithium battery. At each time t, the state of charge constraint of the lithium battery is: ； In the formula, This represents the state of charge limit of the lithium battery at time t; This represents the upper limit of the state of charge of the lithium battery at time t; The state of charge of the lithium battery at the last moment of the day; This represents the state of charge of the lithium battery at the beginning of the day.

3. The energy management and scheduling method for off-grid hydrogen production systems based on VMD-MPC according to claim 1, characterized in that, The modeling of the renewable energy module, electricity storage module, hydrogen storage module, and energy consumption module of the off-grid new energy hydrogen production system also includes: Establish an electrolytic cell model, wherein the electrolytic cell model satisfies: The power consumption expression for an electrolytic cell is: ； In the formula, Let be the electrical power consumption of the i-th electrolytic cell at time t; Let be the hydrogen production of the i-th electrolyzer at time t; The hydrogen production efficiency of the electrolyzer; The power constraint of the electrolytic cell is: ； In the formula, This represents the maximum value of the power ramp-up variation in the electrolytic cell; This represents the maximum output power of the electrolytic cell; Let t be the 0-1 variable representing the operating condition of the i-th electrolytic cell at time t. If 1, it indicates that the cell is started. This represents the minimum output power of the electrolytic cell; The start-up and shutdown constraints of the electrolytic cell are: ； In the formula, Let be a 0-1 variable representing the startup status of the i-th electrolytic cell at time t. If 1, it indicates startup. This is a 0-1 variable representing the shutdown status of the i-th electrolytic cell at time t. A value of 1 indicates shutdown.

4. The energy management and scheduling method for off-grid hydrogen production systems based on VMD-MPC according to claim 1, characterized in that, The modeling of the renewable energy module, electricity storage module, hydrogen storage module, and energy consumption module of the off-grid new energy hydrogen production system also includes: Establish a hydrogen storage tank model, wherein the hydrogen storage tank model satisfies: The hydrogen storage state expression of the hydrogen storage tank at each time t is: ； In the formula, The hydrogen storage state of the hydrogen storage tank at time t; The value represents the mass of hydrogen absorbed or released by the hydrogen storage tank at time t. A positive value indicates the absorption of hydrogen, and a negative value indicates the release of hydrogen. This refers to the rated capacity of the hydrogen storage tank. At each time t, the hydrogen storage state constraint of the hydrogen storage tank is: ； In the formula, This represents the minimum value of the hydrogen storage state in the hydrogen storage tank; This represents the maximum value of the hydrogen storage state in the hydrogen storage tank; This represents the final hydrogen storage state of the hydrogen storage tank within one day. This represents the initial hydrogen storage state of the hydrogen storage tank within one day.

5. The energy management and scheduling method for off-grid hydrogen production systems based on VMD-MPC according to claim 1, characterized in that, The variational mode decomposition (VMD) is performed on the power prediction values ​​of wind power, photovoltaic power, electric load, and hydrogen load, including: Obtain power forecasts for wind power, photovoltaic power, electrical load, and hydrogen load at 24 specific times. The modal components K and the quadratic penalty factor α of VMD were determined using a grid search method. The power prediction value is decomposed into four finite bandwidth subsequences by VMD, where the first two subsequences are low-frequency components used for day-ahead optimization models, and the latter two subsequences are high-frequency components used for intraday rolling optimization models.

6. The energy management and scheduling method for off-grid hydrogen production systems based on VMD-MPC according to claim 1, characterized in that, The step of establishing a day-ahead optimized scheduling model based on the low-frequency components and generating power scheduling plans for each device includes: The optimization period of the day-ahead optimization scheduling model is set to 24 hours, the optimization step size is 1 hour, and the input is the low-frequency mode component obtained by VMD decomposition. The optimization objective is to minimize the total scheduling cost of the off-grid renewable energy hydrogen production system. This total scheduling cost includes system operation and maintenance costs, electrolyzer start-up and shutdown costs, energy loss costs from the electricity-to-hydrogen conversion, and costs associated with wind and solar power curtailment. The expression for this cost is: ； In the formula, To optimize scheduling costs in the near future; For system operation and maintenance costs; Costs associated with starting and stopping the electrolytic cell; Cost of energy loss during the electro-hydrogen conversion; Costs associated with wind and solar power curtailment; including system operation and maintenance costs. This is the sum of the operation and maintenance costs of wind and solar turbines, lithium batteries, and electrolytic cells; Set constraints, including equipment operating state constraints, electrolyzer operating power constraints, battery SOC boundary constraints, and hydrogen supply and demand balance constraints. Solve the day-ahead optimization scheduling model and output the hourly power scheduling plan for electrolyzers and energy storage devices as a reference benchmark for intraday scheduling.

7. The energy management and scheduling method for off-grid hydrogen production systems based on VMD-MPC according to claim 1, characterized in that, The penalty factor is a dynamic function that varies with power deviation, and its expression is: ； In the formula, As a penalty factor; Basic weights; This is the adjustment coefficient; The high-frequency predicted power after VMD decomposition; This represents the day-ahead scheduling plan value for the corresponding time period.

8. The energy management and scheduling method for off-grid hydrogen production systems based on VMD-MPC according to claim 7, characterized in that, High-frequency predicted power after VMD decomposition The daily scheduling plan value for the corresponding time period When the deviation is large, the off-grid new energy hydrogen production system automatically increases the penalty factor. Increase the tracking intensity to ensure stable system operation; when the high-frequency prediction power after VMD decomposition... The daily scheduling plan value for the corresponding time period When the deviation is small, the off-grid new energy hydrogen production system weakens the tracking, reduces unnecessary frequent adjustments, and optimizes equipment lifespan and economy.

9. The energy management and scheduling method for off-grid hydrogen production systems based on VMD-MPC according to claim 7, characterized in that, The step of establishing an intraday rolling optimization scheduling model based on the high-frequency components and the power scheduling plan, and using model prediction to control MPC to perform rolling optimization and output closing loop instructions includes: The optimization interval of the intraday rolling optimization model is set to 15 minutes, the scheduling cycle is 4 hours, and the input is the high-frequency component obtained from VMD decomposition and the day-ahead power scheduling plan. Construct the optimization objective function: ； In the formula, The total cost of system scheduling; This refers to the system power or equipment adjustment amount obtained from intraday model optimization. This represents the day-ahead scheduling plan value for the corresponding time period; Solve the optimization objective function within each rolling window, and execute only the scheduling result of the first time interval as the closing node instruction; Enter the next round of rolling optimization, continuously update until full coverage of all-day scheduling is achieved, forming MPC closed-loop feedback control.

10. The energy management and scheduling method for off-grid hydrogen production systems based on VMD-MPC according to claim 1, characterized in that, The off-grid new energy hydrogen production system adopts a common DC bus configuration. Wind and photovoltaic power generation supply electrical loads and electrolyzers through AC / DC converters and DC / DC converters. Lithium batteries mitigate short-term fluctuations in wind and photovoltaic power generation. Electrolyzers convert electricity into hydrogen energy. Hydrogen storage tanks provide physical storage of hydrogen and buffer against hydrogen energy fluctuations.

Images