Digital twin driven photovoltaic hydrogen production economic dispatch method and device

Abstract

The present application relates to hydrogen energy and electric power system coupling control technical field, especially in kind of digital twin driven photovoltaic hydrogen production economic scheduling method and device, wherein the method comprises the following steps: according to the physical structure of hydrogen production system and electrolytic tank performance attenuation data, the digital twin virtual mapping is established and the twin data interface parameter is output; combining equipment investment, operation and maintenance consumption, attenuation cost and hydrogen yield data, the twin data interface parameter is substituted into economic evaluation calculation, the whole cycle cost function and income function are obtained; and according to the cost function and income function, determine the first combination, the first combination is the ratio sequence of marginal income and attenuation cost in unit scheduling period; the hydrogen production safe operation constraint and low load adaptive control rule are converted into the scheduling hard limit condition set of power upper and lower limit and purity boundary; input photovoltaic time series output data to the digital twin virtual mapping.

Description

Technical Field

[0001] This invention relates to the field of hydrogen energy and power system coupling control technology, and in particular to a digital twin-driven photovoltaic hydrogen production economic dispatch method and device. Background Technology

[0002] Photovoltaic power generation hydrogen production systems effectively address the issue of renewable energy consumption and produce green hydrogen by using photovoltaic power for water electrolysis. Economic dispatching of a hydrogen production system refers to the rational allocation of electrolyzer operating power at different times, while meeting safety constraints, to maximize hydrogen production or minimize operating costs.

[0003] Currently, there is considerable research in the field of economic dispatch of photovoltaic hydrogen production. Common technical approaches include: establishing an electro-hydrogen conversion efficiency model for the electrolyzer, combining it with photovoltaic power output prediction, and using model predictive control or heuristic optimization algorithms (such as particle swarm optimization and genetic algorithms) to solve for the optimal power allocation sequence. In terms of modeling, most methods approximate the efficiency of the electrolyzer as a constant, or simply obtain the static efficiency curve from a table based on the load rate. The objective function typically includes equipment investment, operation and maintenance costs, and hydrogen sales revenue; some studies also consider start-up and shutdown penalties. Constraints generally include the rated power range of the electrolyzer, ramp-up rate, etc.

[0004] However, the aforementioned existing technologies still have the following problems in practical engineering applications: Electrolyzers experience performance degradation over long-term operation due to electrode corrosion, membrane degradation, and other factors, leading to a gradual decrease in hydrogen production per unit of energy consumption. Most existing scheduling methods ignore this time-varying characteristic, employing fixed or static efficiency models. This causes optimized scheduling schemes to deviate from actual operating conditions throughout the entire lifecycle, potentially resulting in overuse of the electrolyzer in the early stages and decreased returns later, thus distorting economic assessments.

[0005] Second, existing scheduling strategies are mostly open-loop offline optimizations, meaning that a scheduling plan is calculated and executed based on forecast data up to the day, lacking a closed-loop correction mechanism based on measured data. When the actual photovoltaic output deviates significantly from the forecast or the electrolyzer performance changes beyond expectations, the offline plan cannot be adjusted in real time, resulting in poor robustness and easily leading to power distribution imbalances or resource waste.

[0006] Third, electrolyzers face gas purity safety risks when operating at low loads. When the operating power is too low, the concentration of hydrogen impurities in oxygen may rise rapidly, posing an explosion hazard once it exceeds the safety limit. Simultaneously, frequent and large power fluctuations accelerate electrolyzer degradation. Existing optimization models often only set fixed lower power limits, failing to transform the purity boundary and adaptive protection rules under low loads into quantifiable hard constraints. This can lead to optimization results potentially hitting safety red lines during actual operation.

[0007] Fourth, the application of digital twin technology in energy systems is still in its early stages. Existing research mostly uses it for real-time monitoring of single devices, lacking a unified framework that deeply couples physical attenuation models, economic evaluation models, and security constraint models. The parameters of each model are isolated from each other, making it difficult to achieve data sharing and collaborative optimization in economic dispatch. Summary of the Invention

[0008] Based on this, it is necessary to provide a digital twin-driven photovoltaic hydrogen production economic scheduling method and device to address the above-mentioned technical problems. By incorporating the performance degradation cost of the electrolyzer into the optimization objective and guiding power allocation with the ratio of marginal revenue to degradation cost, this method avoids excessive equipment wear and tear in pursuit of short-term gains, effectively delays electrolyzer aging, and reduces the total life cycle operating cost.

[0009] This invention provides a digital twin-driven method for the economic dispatch of photovoltaic hydrogen production, the method comprising: Based on the physical structure of the hydrogen production system and the performance degradation data of the electrolyzer, a digital twin virtual mapping is established and the twin data interface parameters are output. By combining equipment investment, operation and maintenance costs, decay costs and hydrogen revenue data, the twin data interface parameters are substituted into the economic evaluation calculation to obtain the full-cycle cost function and revenue function; and a first combination is determined based on the cost function and revenue function, which is a sequence of the ratios of marginal revenue and decay cost within a unit scheduling period. The constraints on safe operation of hydrogen production and the adaptive control rules under low load are converted into a set of hard constraints for scheduling based on power upper and lower limits and purity boundaries. The photovoltaic time-series output data is input into the digital twin virtual mapping, and simulation is performed with the set of scheduling hard constraints as the boundary to obtain the state prediction sequence; and a second combination is determined based on the state prediction sequence and the set of scheduling hard constraints, the second combination being the set of correlation parameters between power regulation and purity margin in each time period; With the goal of maximizing economic benefits, a scheduling scheme table for long-term global operation is calculated by combining the state prediction sequence, the first combination, and the second combination. The scheduling scheme table is sent to the physical system. After collecting the measured data, the parameters of the digital twin virtual mapping and the coefficients in the economic evaluation calculation are iteratively updated based on the measured data.

[0010] In one embodiment, establishing a digital twin virtual mapping and outputting twin data interface parameters based on the physical structure of the hydrogen production system and the performance degradation data of the electrolyzer includes: Collect data on the geometric dimensions, voltage-current curves, temperature and pressure ranges, and decay data of the hydrogen production system. Group the decay data by runtime segment and take the median of each group as the decay value for that time segment. After aligning the median decay value with other data, store it as a mapping table to output the predicted values ​​of electrolyzer efficiency, decay factor, and hydrogen production rate.

[0011] In one embodiment, by combining equipment investment, operation and maintenance costs, degradation costs, and hydrogen revenue data, the parameters of the twin data interface are substituted into the economic evaluation calculation to obtain a full-cycle cost function and a revenue function; and a first combination is determined based on the cost function and the revenue function, including: Calculate the ratio of marginal revenue to decay cost for each scheduling period to obtain the original ratio sequence; The original ratio sequence is compared with the difference between the previous and next time periods. If the difference exceeds the set percentage of the ratio of the previous time period, the current value is replaced with the average ratio of the two time periods. Arrange the corrected ratios in chronological order to obtain the first combination.

[0012] In one embodiment, the conversion of hydrogen production safety operation constraints and low-load adaptive control rules into a set of hard scheduling constraints based on power upper and lower limits and purity boundaries includes: Extract the minimum power, maximum power, lower limit of hydrogen purity, and lower limit of oxygen purity of the electrolytic cell; The minimum power limit for the current period is dynamically adjusted based on the cumulative attenuation of the electrolyzer in the previous period. The dynamically adjusted power upper and lower limits and purity boundaries are merged into a set of constraints.

[0013] In one embodiment, the input photovoltaic time-series output data is mapped to the digital twin virtual map, and simulation is performed using the set of scheduling hard constraints as boundaries to obtain a state prediction sequence; and a second combination is determined based on the state prediction sequence and the set of scheduling hard constraints, including: By substituting the photovoltaic output into the virtual mapping time by time period, we can obtain the operating power, hydrogen production, hydrogen purity, and cumulative degradation for each time period. Calculate the power adjustment amount for each time period and determine whether the purity margin is lower than the set first threshold; if it is lower, mark the time period as a margin alarm. The power regulation amount, purity margin, and alarm flags are stored together as a second combination.

[0014] In one embodiment, the step of calculating a scheduling scheme table for long-term global operation, with the goal of maximizing economic benefits and combining the state prediction sequence, the first combination, and the second combination, includes: With the goal of maximizing total revenue, power is allocated to periods with high ratios in the first combination, sorted from high to low, and power is allocated to periods with high ratios first. If a high ratio period is marked as a margin alarm in the second combination, its priority is reduced to the lowest and it is scheduled to be postponed to the next combination. Power allocation is determined for each time period until all time periods are allocated, and a scheduling scheme table is output.

[0015] In one embodiment, the step of distributing the scheduling scheme table to the physical system, collecting measured data, and iteratively updating the parameters of the digital twin virtual mapping and the coefficients in the economic evaluation calculation based on the measured data includes: Calculate the difference between the measured power and the predicted power, and the difference between the measured hydrogen production and the predicted hydrogen production; If the power difference exceeds the set second threshold and the hydrogen production difference exceeds the set third threshold, the measured values ​​will directly replace the efficiency and attenuation values ​​for the corresponding time period in the virtual mapping. If any difference does not exceed the threshold, it is corrected by the weighted average of the previous value and the measured value; at the same time, the cost coefficient in the economic calculation is corrected according to the same rule.

[0016] The present invention also provides a digital twin-driven photovoltaic hydrogen production economic dispatch device, applied to the digital twin-driven photovoltaic hydrogen production economic dispatch method described in any of the above embodiments, the device comprising: The module is used to establish a digital twin virtual mapping and output twin data interface parameters based on the physical structure of the hydrogen production system and the performance degradation data of the electrolyzer. The first generation module is used to combine equipment investment, operation and maintenance consumption, decay cost and hydrogen revenue data, substitute the twin data interface parameters into the economic evaluation calculation, and obtain the full-cycle cost function and revenue function; and determine the first combination based on the cost function and revenue function, wherein the first combination is a sequence of the ratio of marginal revenue to decay cost within a unit scheduling period. The conversion module is used to convert hydrogen production safety operation constraints and low-load adaptive control rules into a set of scheduling hard constraints for power upper and lower limits and purity boundaries; The second generation module is used to input photovoltaic time-series output data into the digital twin virtual mapping, perform simulation and deduction with the set of scheduling hard constraints as the boundary, and obtain a state prediction sequence; and determine a second combination based on the state prediction sequence and the set of scheduling hard constraints, wherein the second combination is a set of correlation parameters between power regulation and purity margin in each time period; The optimization module is used to calculate the scheduling scheme table for long-term global operation with the goal of maximizing economic benefits, by combining the state prediction sequence, the first combination, and the second combination. The update module is used to send the scheduling scheme table to the physical system, collect measured data, and iteratively update the parameters of the digital twin virtual mapping and the coefficients in the economic evaluation calculation based on the measured data.

[0017] The present invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the digital twin-driven photovoltaic hydrogen production economic dispatch method as described above.

[0018] The present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, implements the digital twin-driven photovoltaic hydrogen production economic scheduling method as described above.

[0019] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the digital twin-driven photovoltaic hydrogen production economic scheduling method as described above.

[0020] The aforementioned digital twin-driven photovoltaic hydrogen production economic dispatch method and device incorporates the performance degradation cost of the electrolyzer into the optimization objective and guides power allocation with the ratio of marginal revenue to degradation cost. This avoids excessive equipment wear and tear in pursuit of short-term gains, effectively delays electrolyzer aging, and reduces the overall operating cost.

[0021] The gas purity boundary and dynamic minimum power lower limit under low load operation are transformed into hard constraints, and alarm periods are actively avoided in the optimization process to eliminate the risk of explosion caused by excessive hydrogen in oxygen and improve the engineering feasibility of the scheduling scheme.

[0022] By using measured data to make weighted corrections or direct replacements to the parameters and economic coefficients of the twin model, the prediction accuracy and scheduling quality can be continuously improved, and the waste of resources and economic losses caused by model mismatch can be reduced.

[0023] Without the need for high-precision photovoltaic output prediction, the system can still provide a suboptimal feasible solution when there are prediction errors through the coordinated scheduling mechanism of the first and second combinations, thereby enhancing the system's adaptability to light fluctuations. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0025] Figure 1 A flowchart of the digital twin-driven photovoltaic hydrogen production economic dispatch method provided by the present invention; Figure 2 A module diagram of a digital twin-driven photovoltaic hydrogen production economic dispatch device provided by the present invention; Figure 3An internal structural diagram of the computer device provided by the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, 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.

[0027] The following is combined with Figures 1 to 3 This invention describes a digital twin-driven photovoltaic hydrogen production economic dispatch method and apparatus.

[0028] In one embodiment, a digital twin-driven method for the economic dispatch of photovoltaic hydrogen production includes the following steps: Step S100: Based on the physical structure of the hydrogen production system and the performance degradation data of the electrolyzer, establish a digital twin virtual mapping and output the twin data interface parameters.

[0029] Step S200: Combining equipment investment, operation and maintenance costs, decay costs and hydrogen revenue data, the twin data interface parameters are substituted into the economic evaluation calculation to obtain the full-cycle cost function and revenue function; and a first combination is determined based on the cost function and revenue function, the first combination being a sequence of the ratios of marginal revenue to decay costs within a unit scheduling period.

[0030] Step S300 converts the hydrogen production safety operation constraints and low-load adaptive control rules into a set of scheduling hard constraints for power upper and lower limits and purity boundaries.

[0031] Step S400: Input photovoltaic time-series output data into the digital twin virtual mapping, perform simulation and deduction with the set of scheduling hard constraints as the boundary, and obtain the state prediction sequence; and determine the second combination based on the state prediction sequence and the set of scheduling hard constraints, the second combination being the set of correlation parameters between power regulation and purity margin in each time period.

[0032] Step S500: With the goal of maximizing economic benefits, a scheduling scheme table for long-term global operation is calculated by combining the state prediction sequence, the first combination, and the second combination.

[0033] Step S600: The scheduling scheme table is sent to the physical system. After collecting the measured data, the parameters of the digital twin virtual mapping and the coefficients in the economic evaluation calculation are iteratively updated based on the measured data.

[0034] The aforementioned digital twin-driven photovoltaic hydrogen production economic dispatch method uses a digital twin as a unified central hub, mapping the physical structure and attenuation data of the hydrogen production system into an iteratively updatable twin virtual model and outputting standardized interface parameters. Utilizing equipment investment, operation and maintenance, attenuation costs, and hydrogen revenue data, it calculates the ratio of marginal revenue to attenuation cost for each dispatch period, forming a first combination to quantify the economic benefit tendency of each period. Simultaneously, the low-load safe operation rules are transformed into a set of hard constraints on power upper and lower limits and purity boundaries. Combined with photovoltaic output data, simulations are performed in the twin model to obtain a set of correlation parameters (second combination) between power adjustment and purity margin for each period, used to identify periods of safety risk. The optimization module prioritizes the first combination and uses the second combination as a safety penalty to perform long-term global power allocation, generating a dispatch scheme table that balances economy and safety. It then uses measured data to close-loop correct the twin model parameters and economic coefficients, forming a complete closed loop from state perception to benefit quantification to safety constraints to global optimization and finally to feedback correction.

[0035] By incorporating the cost of electrolytic cell performance degradation into the optimization objective and guiding power allocation with the ratio of marginal revenue to degradation cost, excessive equipment wear and tear in pursuit of short-term gains is avoided, effectively delaying electrolytic cell aging and reducing the overall operating cost.

[0036] The gas purity boundary and dynamic minimum power lower limit under low load operation are transformed into hard constraints, and alarm periods are actively avoided in the optimization process to eliminate the risk of explosion caused by excessive hydrogen in oxygen and improve the engineering feasibility of the scheduling scheme.

[0037] By using measured data to make weighted corrections or direct replacements to the parameters and economic coefficients of the twin model, the prediction accuracy and scheduling quality can be continuously improved, and the waste of resources and economic losses caused by model mismatch can be reduced.

[0038] Without the need for high-precision photovoltaic output prediction, the system can still provide a suboptimal feasible solution when there are prediction errors through the coordinated scheduling mechanism of the first and second combinations, thereby enhancing the system's adaptability to light fluctuations.

[0039] In one embodiment, establishing a digital twin virtual mapping and outputting twin data interface parameters based on the physical structure of the hydrogen production system and the performance degradation data of the electrolyzer includes the following steps: Step S110: Collect the geometric dimensions, voltage and current curves, temperature and pressure ranges, and decay data of the hydrogen production system.

[0040] Specifically, for a 1MW alkaline electrolyzer hydrogen production system, the following data were collected using sensors and a historical operating database: Geometric dimensions: number of electrolytic cell chambers, effective electrode area, and diaphragm thickness.

[0041] Voltage and current curves: the actual voltage and current correspondence under different load rates (e.g., 20%, 40%, 60%, 80%, 100%).

[0042] Temperature and pressure range: The operating temperature range is 60℃-90℃, and the pressure range is 1.6MPa-3.0MPa.

[0043] Decay data: The efficiency (kWh / Nm³ H2) of the electrolyzer was recorded in the operation logs of the past 12 months after each shutdown maintenance. The rate of decrease in efficiency is the decay data.

[0044] Step S120: Group the attenuation data by runtime period and take the median of each group as the attenuation value for that period.

[0045] Specifically, the 12 months of decay data were grouped by month to obtain 12 groups of data.

[0046] For example, the decay value sequence in the 3rd month group is [0.12%, 0.15%, 0.14%, 0.18%, 0.13%], and the median of 0.14% is taken as the representative decay value for the 3rd month.

[0047] If the data volume for a given month is even, take the average of the two middle values. This approach can effectively avoid model fluctuations caused by occasional outliers from sensors.

[0048] Step S130: Align the median decay value with other data and store it as a mapping table, then output the predicted values ​​of electrolyzer efficiency, decay factor, and hydrogen production rate.

[0049] Specifically, the median decay value of each time period obtained in step S120 is aligned with the average load rate, average temperature and other parameters of the corresponding time period to form a multidimensional mapping table.

[0050] Each row in the mapping table corresponds to a runtime segment and contains the following fields: [time period index, average load rate, average temperature, median decay value].

[0051] Electrolyzer efficiency: defined as the ratio of actual hydrogen production to theoretical hydrogen production in the current period, which can be directly queried from the mapping table.

[0052] Attenuation factor: defined as the ratio of efficiency attenuation in the current time period to that in the initial time period, calculated as current time period efficiency / initial time period efficiency.

[0053] Hydrogen production rate prediction: Input the power value for the current period, and output the predicted hydrogen production rate in Nm³ / h by linearly interpolating the voltage and current curves and combining efficiency correction.

[0054] For example, suppose the input power P = 500kW during a certain period. Looking up the mapping table, the electrolyzer voltage U = 200V at this power. Then the current I = P / U = 2500A. According to Faraday's law, the theoretical hydrogen production rate = (I × 3600) / (2 × 96485) × 22.4 / 1000 ≈ 0.418 Nm³ / h·kA, that is, the theoretical hydrogen production ≈ 1.045 Nm³ / h. Multiplying this by the current efficiency during the current period (e.g., 0.85), we get the final predicted hydrogen production ≈ 0.89 Nm³ / h.

[0055] In one embodiment, the process of combining equipment investment, operation and maintenance costs, degradation costs, and hydrogen revenue data, substituting the twin data interface parameters into the economic evaluation calculation, yields a full-cycle cost function and a revenue function; and determining a first combination based on the cost function and revenue function includes the following steps: Step S210: Calculate the ratio of marginal revenue to decay cost for each scheduling period to obtain the original ratio sequence.

[0056] Specifically, the scheduling period is set to 1 hour. For the t-th time period: Marginal revenue = (hydrogen price) × (the amount of additional hydrogen produced for every 1kW increase in power during this period) Degradation cost = (Electrolyzer replacement cost) × (Percentage of lifespan loss due to degradation per hour of operation during this period) Assuming the price of hydrogen is 30 yuan / kg, and that an additional 1kW of power can produce 0.2kg more hydrogen during a certain period, the marginal revenue is 6 yuan. If the operation during this period leads to a 0.001% reduction in the electrolyzer's lifespan, and the replacement cost is 1 million yuan, then the attenuation cost is 1 yuan. The ratio for this period is 6 / 1 = 6. The original ratio sequence is obtained by calculating each period, for example: [2.5, 6, 1.8, 4.2,...].

[0057] Step S220: Compare the difference between the previous and next time periods in the original ratio sequence. If the difference exceeds the set percentage of the ratio of the previous time period, replace the current value with the average ratio of the two time periods.

[0058] Specifically, the percentage is set at 50%. If the three adjacent time periods in the original ratio sequence are [3, 10, 4].

[0059] Compare the difference between 10 and the previous time 3: (10-3) / 3 ≈ 233%, which exceeds 50%. The ratio 10 in the second time period is determined to be an outlier.

[0060] The original value 10 is replaced with the average ratio of the two time periods (time period 1 and time period 3), which is (3+4) / 2 = 3.5. The corrected sequence becomes [3, 3.5, 4]. This correction process can suppress extreme scheduling instructions caused by data noise or prediction mutations.

[0061] Step S230: Arrange the corrected ratios in chronological order to obtain the first combination.

[0062] Specifically, the adjusted ratios for all scheduling periods (e.g., the next 24 hours) are arranged in chronological order (t=1,2,...,24) into a sequence, denoted as the first combination. = [r1, r2, ..., r 24 This sequence directly reflects the relative attractiveness of each time period in terms of the trade-off between unit marginal benefit and decay cost; the higher the value, the more suitable the time period is for scheduling high-power operation.

[0063] In one embodiment, converting hydrogen production safety operation constraints and low-load adaptive control rules into a set of hard scheduling constraints based on power upper and lower limits and purity boundaries includes the following steps: Step S310: Extract the minimum power, maximum power, lower limit of hydrogen purity, and lower limit of oxygen purity of the electrolytic cell.

[0064] Specifically, extract the information from the electrolytic cell equipment manual: Minimum power: 0.2 MW (20% of rated power).

[0065] Maximum power: 1.0 MW (100% of rated power).

[0066] Lower limit of hydrogen purity: 99.5% (volume fraction).

[0067] Lower limit of oxygen purity: 98.5% (volume fraction).

[0068] Step S320: Dynamically adjust the minimum power limit for the current period based on the cumulative attenuation of the electrolytic cell in the previous period.

[0069] Specifically, after 1000 hours of cumulative operation, the electrolyzer has experienced a cumulative efficiency degradation of 5% of its initial efficiency. At this point, the originally planned minimum power of 0.2MW will cause the hydrogen concentration in the oxygen to rise too rapidly. According to the preset adaptive rules: Dynamic minimum power = Original minimum power × (1 + Cumulative attenuation × Coefficient).

[0070] Assuming a coefficient of 2, the dynamic minimum power = 0.2 × (1 + 0.05 × 2) = 0.22 MW. As the attenuation intensifies, the lower limit of the minimum power dynamically increases to mitigate the safety risks under low load.

[0071] Step S330: The dynamically adjusted upper and lower power limits and purity boundaries are merged into a set of constraint conditions.

[0072] Specifically, the maximum power and purity lower limit from step S310 are combined with the dynamic minimum power calculated in step S320 to form a set of hard constraints for the scheduling period, for example: .

[0073] In one embodiment, the input photovoltaic time-series output data is mapped to the digital twin virtual map, and simulation is performed using the set of scheduling hard constraints as boundaries to obtain a state prediction sequence; and a second combination is determined based on the state prediction sequence and the set of scheduling hard constraints, including the following steps: In step S410, the photovoltaic output is substituted into the virtual mapping for each time period to obtain the operating power, hydrogen production, hydrogen purity, and cumulative degradation for each time period.

[0074] Specifically, the predicted photovoltaic output for the next 24 hours (e.g., [0.3, 0.5, 0.8, …, 0.1] MW) is sequentially substituted into the digital twin mapping table established in step S130. Through table lookup and interpolation calculations, the output for each time period is determined as follows: Operating power (limited by the constraint set in step S330, if photovoltaic output is lower than...) If so, it will be forced to 0 or standby.

[0075] Hydrogen production (Nm³).

[0076] Hydrogen purity (%).

[0077] Updated cumulative decay (%).

[0078] Step S420: Calculate the power adjustment amount for each time period and determine whether the purity margin is lower than the set first threshold; if it is lower, mark the time period as a margin alarm.

[0079] Specifically, the power adjustment amount is defined as the absolute value of the difference between the power in the current time period and the power in the previous time period. The first threshold is set to 1%.

[0080] Calculate the purity margin = current hydrogen purity - 99.5% (lower purity limit).

[0081] For example, if the calculated hydrogen purity is 99.8% for a certain period, the purity margin is 0.3%, which is lower than the 1% threshold. The system will mark this period as a "margin alarm". This alarm means that if the system operates at the original photovoltaic power, the gas purity will approach the safety limit, and this needs to be avoided or adjusted in subsequent optimizations.

[0082] In step S430, the power regulation amount, purity margin, and alarm flag are stored together as a second combination.

[0083] Specifically, for each time period, the power regulation (MW), purity margin (%), and alarm flag (True / False) are stored as a tuple. All time period tuples form a second set, denoted as... = [(ΔP1, margin1,alert1), (ΔP2, margin2, alert2), ..., (ΔP 24 margin 24 alert 24 )).

[0084] In one embodiment, the step of calculating the scheduling scheme table for long-term global operation, with the goal of maximizing economic benefits, by combining the state prediction sequence, the first combination, and the second combination, includes the following steps: Step S510: With the goal of maximizing total revenue, sort the first combination by ratio from high to low, and prioritize allocating power to periods with high ratios.

[0085] Specifically, the first group After sorting the ratios, the highest ratio is 8.5 for period 8, and the second highest is 7.2 for period 3. Given the limited total photovoltaic power generation, priority is given to allocating as much power as possible to period 8 (not exceeding its power limit and photovoltaic output), and the remaining power is then allocated to period 3, and so on.

[0086] In step S520, if the high ratio period is marked as a margin alarm in the second combination, its priority is reduced to the lowest and the arrangement is postponed to the next period.

[0087] Specifically, although the ratio is high in time slot 8, its alert8 value is True (margin alarm) in the second combination. To ensure safety, the priority of this time slot in the power allocation queue is reduced to the lowest level. The power originally scheduled for allocation is postponed to the next safe time slot with a higher ratio (e.g., time slot 12). If alarms occur in all subsequent time slots, a forced power reduction operation mode is entered, reducing the power to near the dynamic minimum power and triggering a safety review.

[0088] Step S530: Determine power allocation for each time period until all time periods are allocated, and output the scheduling scheme table.

[0089] Specifically, after the above priority sorting and alarm processing, the final planned power is allocated to each time period. The [time, planned power, expected hydrogen production, purity] of all time periods are compiled into a table, which is the scheduling plan table, used for issuance and execution.

[0090] In one embodiment, the step of distributing the scheduling scheme table to the physical system, collecting measured data, and iteratively updating the parameters of the digital twin virtual mapping and the coefficients in the economic evaluation calculation based on the measured data includes the following steps: Step S610: Calculate the difference between the measured power and the predicted power, and the difference between the measured hydrogen production and the predicted hydrogen production.

[0091] Specifically, one hour after the scheduling scheme was implemented, measured data was collected: measured power 0.52 MW, measured hydrogen production 105 Nm³. The predicted power for the corresponding period was 0.50 MW, and the predicted hydrogen production was 110 Nm³. Calculation: Power difference = |0.52 - 0.50| = 0.02 MW.

[0092] Difference in hydrogen production = |105 - 110| = 5 Nm³.

[0093] In step S620, if the power difference exceeds the set second threshold and the hydrogen production difference exceeds the set third threshold, the efficiency value and attenuation value of the corresponding time period in the virtual mapping are directly replaced with the measured values.

[0094] Specifically, the second threshold is set to 0.05 MW, and the third threshold is 3 Nm³. Here, the power difference of 0.02 MW does not exceed 0.05 MW, and therefore does not meet the direct replacement condition. If the power difference reaches 0.1 MW and the hydrogen production difference reaches 10 Nm³ in a certain period, the model is considered severely mismatched, and the measured efficiency (hydrogen production / power consumption) and the calculated attenuation value are directly written into the mapping table to replace the original parameters.

[0095] In step S630, if any difference does not exceed the threshold, the weighted average of the previous value and the measured value is used for correction; at the same time, the cost coefficient in the economic calculation is corrected according to the same rule.

[0096] Specifically, since the current power difference and hydrogen production difference do not both exceed the threshold, a weighted average is used for slow correction.

[0097] Let the weighting factors be α = 0.7 (previous value weight) and β = 0.3 (measured value weight).

[0098] Updated efficiency = 0.7 × original efficiency + 0.3 × measured efficiency.

[0099] Meanwhile, the degradation cost coefficient in the economic calculation is also corrected using the same rule: new coefficient = 0.7 × original coefficient + 0.3 × (new coefficient estimated based on measured degradation). Through this closed-loop update, the twin model can gradually adapt to the long-term performance evolution of the electrolyzer.

[0100] The digital twin-driven photovoltaic hydrogen production economic dispatch device provided by the present invention will be described below. The digital twin-driven photovoltaic hydrogen production economic dispatch device described below and the digital twin-driven photovoltaic hydrogen production economic dispatch method described above can be referred to each other.

[0101] In one embodiment, a digital twin-driven photovoltaic hydrogen production economic dispatch device is applied to the digital twin-driven photovoltaic hydrogen production economic dispatch method described in any of the above embodiments, including a construction module, a first generation module, a conversion module, a second generation module, an optimization module, and an update module.

[0102] The building module is used to establish a digital twin virtual mapping based on the physical structure of the hydrogen production system and the performance degradation data of the electrolyzer, and output the twin data interface parameters.

[0103] The first generation module is used to combine equipment investment, operation and maintenance costs, decay costs and hydrogen revenue data, substitute the twin data interface parameters into the economic evaluation calculation, and obtain the full-cycle cost function and revenue function; and determine the first combination based on the cost function and revenue function, the first combination being a sequence of the ratio of marginal revenue to decay cost within a unit scheduling period.

[0104] The conversion module is used to convert hydrogen production safety operation constraints and low-load adaptive control rules into a set of scheduling hard constraints for power upper and lower limits and purity boundaries.

[0105] The second generation module is used to input photovoltaic time-series output data into the digital twin virtual mapping, perform simulation and deduction with the set of scheduling hard constraints as the boundary, and obtain a state prediction sequence; and determine a second combination based on the state prediction sequence and the set of scheduling hard constraints, the second combination being a set of correlation parameters between power regulation and purity margin in each time period.

[0106] The optimization module is used to calculate a scheduling scheme table for long-term global operation with the goal of maximizing economic benefits, by combining the state prediction sequence, the first combination, and the second combination.

[0107] The update module is used to send the scheduling scheme table to the physical system, collect measured data, and iteratively update the parameters of the digital twin virtual mapping and the coefficients in the economic evaluation calculation based on the measured data.

[0108] Figure 3 This example illustrates a schematic diagram of the physical structure of an electronic device, which can be a smart terminal. Its internal structure diagram can be as follows: Figure 3As shown, this electronic device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a digital twin-driven photovoltaic hydrogen production economic dispatch method.

[0109] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the electronic device to which the present invention is applied. A specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0110] On the other hand, the present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, implements a digital twin-driven photovoltaic hydrogen production economic scheduling method.

[0111] On another front, a computer program product or computer program is provided, comprising computer instructions stored in a computer storage medium. A processor of an electronic device reads the computer instructions from the computer storage medium, and when the processor executes the computer instructions, it implements a digital twin-driven photovoltaic hydrogen production economic dispatch method.

[0112] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory.

[0113] By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0114] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0115] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the appended claims.

Claims

1. A digital twin-driven economic dispatch method for photovoltaic hydrogen production, characterized in that, The method includes: Based on the physical structure of the hydrogen production system and the performance degradation data of the electrolyzer, a digital twin virtual mapping is established and the twin data interface parameters are output. By combining equipment investment, operation and maintenance costs, decay costs and hydrogen revenue data, the twin data interface parameters are substituted into the economic evaluation calculation to obtain the full-cycle cost function and revenue function; and a first combination is determined based on the cost function and revenue function, which is a sequence of the ratios of marginal revenue and decay cost within a unit scheduling period. The constraints on safe operation of hydrogen production and the adaptive control rules under low load are converted into a set of hard constraints for scheduling based on power upper and lower limits and purity boundaries. The photovoltaic time-series output data is input into the digital twin virtual mapping, and simulation is performed with the set of scheduling hard constraints as the boundary to obtain the state prediction sequence; and a second combination is determined based on the state prediction sequence and the set of scheduling hard constraints, the second combination being the set of correlation parameters between power regulation and purity margin in each time period; With the goal of maximizing economic benefits, a scheduling scheme table for long-term global operation is calculated by combining the state prediction sequence, the first combination, and the second combination. The scheduling scheme table is sent to the physical system. After collecting the measured data, the parameters of the digital twin virtual mapping and the coefficients in the economic evaluation calculation are iteratively updated based on the measured data.

2. The digital twin-driven photovoltaic hydrogen production economic dispatch method according to claim 1, characterized in that, The process of establishing a digital twin virtual mapping and outputting twin data interface parameters based on the physical structure of the hydrogen production system and the performance degradation data of the electrolyzer includes: Collect data on the geometric dimensions, voltage-current curves, temperature and pressure ranges, and decay data of the hydrogen production system. Group the decay data by runtime segment and take the median of each group as the decay value for that time segment. After aligning the median decay value with other data, store it as a mapping table to output the predicted values ​​of electrolyzer efficiency, decay factor, and hydrogen production rate.

3. The digital twin-driven photovoltaic hydrogen production economic dispatch method according to claim 1, characterized in that, By combining equipment investment, operation and maintenance costs, attenuation costs and hydrogen revenue data, and substituting the twin data interface parameters into the economic evaluation calculation, the full-cycle cost function and revenue function are obtained. And determine the first combination based on the cost function and the revenue function, including: Calculate the ratio of marginal revenue to decay cost for each scheduling period to obtain the original ratio sequence; The original ratio sequence is compared with the difference between the previous and next time periods. If the difference exceeds the set percentage of the ratio of the previous time period, the current value is replaced with the average ratio of the two time periods. Arrange the corrected ratios in chronological order to obtain the first combination.

4. The digital twin-driven photovoltaic hydrogen production economic dispatch method according to claim 1, characterized in that, The process of converting hydrogen production safety operation constraints and low-load adaptive control rules into a set of hard scheduling constraints based on power upper and lower limits and purity boundaries includes: Extract the minimum power, maximum power, lower limit of hydrogen purity, and lower limit of oxygen purity of the electrolytic cell; The minimum power limit for the current period is dynamically adjusted based on the cumulative attenuation of the electrolyzer in the previous period. The dynamically adjusted power upper and lower limits and purity boundaries are merged into a set of constraints.

5. The digital twin-driven photovoltaic hydrogen production economic dispatch method according to claim 1, characterized in that, The input photovoltaic time-series output data is mapped to the digital twin virtual mapping, and simulation is performed with the set of scheduling hard constraints as the boundary to obtain the state prediction sequence; And determine a second combination based on the state prediction sequence and the set of scheduling hard constraints, including: By substituting the photovoltaic output into the virtual mapping time by time period, we can obtain the operating power, hydrogen production, hydrogen purity, and cumulative degradation for each time period. Calculate the power adjustment amount for each time period and determine whether the purity margin is lower than the set first threshold; if it is lower, mark the time period as a margin alarm. The power regulation amount, purity margin, and alarm flags are stored together as a second combination.

6. The digital twin-driven photovoltaic hydrogen production economic dispatch method according to claim 1, characterized in that, The method, aiming at maximizing economic benefits, calculates a scheduling scheme table for long-term global operation by combining the state prediction sequence, the first combination, and the second combination, including: With the goal of maximizing total revenue, power is allocated to periods with high ratios in the first combination, sorted from high to low, and power is allocated to periods with high ratios first. If a high ratio period is marked as a margin alarm in the second combination, its priority is reduced to the lowest and it is scheduled to be postponed to the next combination. Power allocation is determined for each time period until all time periods are allocated, and a scheduling scheme table is output.

7. The digital twin-driven photovoltaic hydrogen production economic dispatch method according to claim 1, characterized in that, The step of distributing the scheduling scheme table to the physical system, collecting measured data, and iteratively updating the parameters of the digital twin virtual mapping and the coefficients in the economic evaluation calculation based on the measured data includes: Calculate the difference between the measured power and the predicted power, and the difference between the measured hydrogen production and the predicted hydrogen production; If the power difference exceeds the set second threshold and the hydrogen production difference exceeds the set third threshold, the measured values ​​will directly replace the efficiency and attenuation values ​​for the corresponding time period in the virtual mapping. If any difference does not exceed the threshold, it is corrected by the weighted average of the previous value and the measured value; at the same time, the cost coefficient in the economic calculation is corrected according to the same rule.

8. A digital twin-driven photovoltaic hydrogen production economic dispatch device, applied to the digital twin-driven photovoltaic hydrogen production economic dispatch method according to any one of claims 1 to 7, characterized in that, The device includes: The module is used to establish a digital twin virtual mapping and output twin data interface parameters based on the physical structure of the hydrogen production system and the performance degradation data of the electrolyzer. The first generation module is used to combine equipment investment, operation and maintenance consumption, decay cost and hydrogen revenue data, substitute the twin data interface parameters into the economic evaluation calculation, and obtain the full-cycle cost function and revenue function; and determine the first combination based on the cost function and revenue function, wherein the first combination is a sequence of the ratio of marginal revenue to decay cost within a unit scheduling period. The conversion module is used to convert hydrogen production safety operation constraints and low-load adaptive control rules into a set of scheduling hard constraints for power upper and lower limits and purity boundaries; The second generation module is used to input photovoltaic time-series output data into the digital twin virtual mapping, perform simulation and deduction with the set of scheduling hard constraints as the boundary, and obtain a state prediction sequence; and determine a second combination based on the state prediction sequence and the set of scheduling hard constraints, wherein the second combination is a set of correlation parameters between power regulation and purity margin in each time period; The optimization module is used to calculate the scheduling scheme table for long-term global operation with the goal of maximizing economic benefits, by combining the state prediction sequence, the first combination, and the second combination. The update module is used to send the scheduling scheme table to the physical system, collect measured data, and iteratively update the parameters of the digital twin virtual mapping and the coefficients in the economic evaluation calculation based on the measured data.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images