Multi-energy complementary wind, solar, thermal and hydrogen integrated control methods, platforms and media

Abstract

This invention discloses a multi-energy complementary integrated control method, platform, and medium for wind, solar, thermal, and hydrogen power generation, relating to the field of new energy power generation control technology. The method includes: constructing a power generation expectation curve; establishing a wind-solar power generation fusion curve; obtaining the characteristics of excess wind and solar output, equal wind and solar output, and insufficient wind and solar output modes; performing wind and solar curtailment suppression regulation to obtain a first energy complementarity control strategy; performing wind and solar fluctuation smoothing compensation regulation on the multi-energy modules to obtain a second energy complementarity control strategy; and performing energy supplementation optimization regulation on the multi-energy modules under thermal power generation ratio limitations to obtain a third energy complementarity control strategy. This invention solves the technical problems of intermittency and volatility in existing wind and solar power generation, leading to poor power supply stability and low energy utilization. It achieves precise supply and demand matching and differentiated control of multi-energy modules, improving the stability and energy utilization of integrated wind, solar, thermal, and hydrogen power generation.

Description

Technical Field

[0001] This invention relates to the field of new energy power generation control technology, specifically to a multi-energy complementary integrated control method, platform, and medium for wind, solar, thermal, and hydrogen power generation. Background Technology

[0002] Against the backdrop of rapid development in new energy power generation, wind and photovoltaic power generation have been widely used due to their clean and renewable advantages. However, affected by natural conditions, their output exhibits significant intermittency, fluctuation, and uncertainty, leading to frequent wind and solar curtailment and low energy utilization. Meanwhile, while traditional thermal power generation can provide stable output, its high carbon emission intensity does not meet the requirements of low-carbon development. Hydrogen energy, as a clean and efficient energy storage and supply carrier, is gradually becoming an important component of multi-energy complementary systems. Currently, integrated wind-solar-thermal-hydrogen power generation systems lack precise supply and demand forecasting and time-series alignment analysis mechanisms for predetermined time windows. Insufficient coordinated control among multiple energy modules makes it difficult to implement differentiated adjustments based on different modal characteristics such as excess, equal, and insufficient wind and solar output. This results in poor power supply stability, difficulty in rationally controlling the proportion of thermal power generation, and an inability to fully leverage the comprehensive advantages of multi-energy complementarity.

[0003] Existing wind and solar power generation technologies suffer from intermittency and volatility, resulting in poor power supply stability and low energy utilization. Summary of the Invention

[0004] This application provides a multi-energy complementary wind, solar, thermal and hydrogen integrated control method, platform and medium to address the technical problems of intermittent and fluctuating wind and solar power generation in the prior art, which leads to poor power supply stability and low energy utilization.

[0005] In view of the above problems, this application provides a multi-energy complementary wind, solar, thermal and hydrogen integrated control method, platform and medium.

[0006] The first aspect of this application provides a multi-energy complementary integrated control method for wind, solar, thermal, and hydrogen energy, the method comprising:

[0007] Based on a predetermined time zone window, the power generation demand of the power plant is predicted under a tolerance compensation, and a power generation expectation curve is constructed. The power plant includes multiple energy modules. Based on the predetermined time zone window, the wind and solar power generation forecasts of the multiple energy modules are fused to establish a wind and solar power fusion curve. Based on the power generation expectation curve, the wind and solar power fusion curve is analyzed for multi-dimensional supply and demand characteristics under time-series alignment to obtain wind and solar power surplus mode characteristics, wind and solar power equal output mode characteristics, and wind and solar power shortage mode characteristics. Based on the wind and solar power surplus mode characteristics, the multiple energy modules are regulated to suppress wind and solar curtailment, obtaining a first energy complementarity control strategy. Based on the wind and solar power equal output mode characteristics, the multiple energy modules are regulated to smooth wind and solar power fluctuations, obtaining a second energy complementarity control strategy. Based on the wind and solar power shortage mode characteristics, the multiple energy modules are regulated to optimize energy supplementation under thermal power generation ratio restrictions, obtaining a third energy complementarity control strategy.

[0008] A second aspect of this application provides a multi-energy complementary integrated wind, solar, thermal, and hydrogen control platform, the platform comprising:

[0009] The system includes a power generation expectation curve construction unit, used to predict the power generation demand of a power station under tolerance compensation based on a predetermined time zone window, and construct the power generation expectation curve. The power station includes multiple energy modules. A wind-solar power generation fusion curve establishment unit is used to perform wind and solar power generation prediction fusion on the multiple energy modules based on the predetermined time zone window, and establish a wind-solar power generation fusion curve. A modal characteristic acquisition unit is used to perform multi-dimensional supply and demand characteristic analysis of the wind-solar power generation fusion curve under time alignment based on the power generation expectation curve, and obtain the wind and solar power output surplus modal characteristic and the wind and solar power output equalization modal characteristic. The system includes: a first control strategy acquisition unit, used to adjust the multi-energy module to suppress wind and solar curtailment based on the wind and solar excess output mode characteristics, and acquire a first energy complementarity control strategy; a second control strategy acquisition unit, used to adjust the multi-energy module to smooth wind and solar fluctuations based on the wind and solar equal output mode characteristics, and acquire a second energy complementarity control strategy; and a third control strategy acquisition unit, used to adjust the multi-energy module to optimize energy supplementation under the thermal power generation ratio limit based on the wind and solar insufficient output mode characteristics, and acquire a third energy complementarity control strategy.

[0010] In a third aspect, this application provides a computer-readable storage medium storing a computer program for executing the multi-energy complementary wind, solar, thermal, and hydrogen integrated control method provided in this application.

[0011] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0012] Based on a predetermined time zone window, the power generation demand of the power plant is predicted under a tolerance compensation, and a power generation expectation curve is constructed. Based on the predetermined time zone window, the wind and solar power generation predictions of the multi-energy modules are fused to establish a wind-solar power generation fusion curve. The wind-solar power generation fusion curve is analyzed for multi-dimensional supply and demand characteristics under time-series alignment to obtain the wind and solar power output surplus mode characteristics, wind and solar power output equal mode characteristics, and wind and solar power output insufficient mode characteristics. Wind and solar curtailment suppression and regulation are applied to the multi-energy modules to obtain the first energy complementarity control strategy. Wind and solar fluctuation smoothing compensation and regulation are applied to the multi-energy modules to obtain the second energy complementarity control strategy. Energy supplementation optimization and regulation are applied to the multi-energy modules under the thermal power generation ratio limit to obtain the third energy complementarity control strategy. This achieves the technical effect of precise supply and demand matching and differentiated regulation of the multi-energy modules, improving the stability and energy utilization rate of integrated wind, solar, thermal, and hydrogen power generation. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 A flowchart illustrating the multi-energy complementary integrated wind, solar, thermal, and hydrogen control method provided in this application embodiment.

[0015] Figure 2 A schematic diagram of the structure of the multi-energy complementary wind, solar, thermal and hydrogen integrated control platform provided in the embodiments of this application.

[0016] Figure labeling: 10 is the expected power generation curve construction unit, 20 is the wind and solar power generation integration curve establishment unit, 30 is the modal characteristic acquisition unit, 40 is the first control strategy acquisition unit, 50 is the second control strategy acquisition unit, and 60 is the third control strategy acquisition unit. Detailed Implementation

[0017] This application provides a multi-energy complementary integrated control method, platform, and medium for wind, solar, thermal, and hydrogen power generation, which addresses the technical problems of intermittent and fluctuating wind and solar power generation, resulting in poor power supply stability and low energy utilization.

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0019] Example 1, as Figure 1 As shown, this application provides a multi-energy complementary wind, solar, thermal, and hydrogen integrated control method, the method comprising:

[0020] Step S100: Based on the predetermined time zone window, predict the power generation demand of the power station under tolerance compensation, and construct the expected power generation curve. The power station includes multiple energy modules.

[0021] Specifically, for power plants that include wind power, photovoltaic power, thermal power, and hydrogen power modules, power demand forecasting is conducted using a predetermined time zone window as the time benchmark. First, the overall power demand of the power plant is predicted based on this predetermined time zone window, forming a power demand benchmark curve. Then, knowledge graph learning is performed on the historical power loss records of the power plant to uncover loss-related patterns and construct a power loss analytical graph. Subsequently, based on this analytical graph, the power demand benchmark curve is analyzed for multi-dimensional loss characteristics, accurately extracting loss characteristic parameters for multiple nodes. Finally, combined with these node loss characteristics, the power demand benchmark curve is adjusted with tolerance compensation to ultimately generate a power demand expectation curve that fits the actual power generation scenario and takes into account loss factors.

[0022] Step S200: Perform wind and solar power generation prediction and fusion on the multi-energy modules according to the predetermined time zone window, and establish a wind and solar power generation fusion curve.

[0023] Specifically, using a predetermined time zone window consistent with the expected power generation curve as a unified time benchmark, for multi-energy modules including wind power, photovoltaic power, thermal power, and hydrogen power, the power generation conditions of wind power and photovoltaic power modules are first focused separately: historical power generation data and real-time operating parameters of the two types of modules within the predetermined time zone window are collected, such as wind speed, solar intensity, equipment load rate, and environmental impact factors. Power prediction models adapted to the characteristics of wind and solar power generation are used to independently predict the power generation of wind power and photovoltaic power, obtaining two sets of accurate single-module power generation prediction results. Subsequently, based on the temporal synergy and complementary laws of wind and solar power generation, the two sets of prediction results are redundancy elimination, deviation correction, and synergistic integration through data fusion algorithms. The temporal correlation and fluctuation complementarity of the two types of energy output are fully considered, and finally, a wind and solar power fusion curve that can comprehensively and accurately reflect the integrated wind and solar power generation capacity within the predetermined time zone window is established.

[0024] Step S300: Based on the expected power generation curve, perform multi-dimensional supply and demand characteristic analysis on the wind and solar power generation fusion curve under time alignment to obtain the wind and solar power output surplus mode characteristics, wind and solar power output equal mode characteristics, and wind and solar power output insufficient mode characteristics.

[0025] Specifically, the constructed expected power generation curve and the wind-solar power generation fusion curve are first aligned according to a predetermined time zone window to ensure consistency between the two curves in the time dimension, thereby generating a power comparison chart that intuitively reflects the supply and demand relationship. Then, based on this power comparison chart, multi-node supply and demand deviation identification is performed to accurately capture supply and demand differences at different time points and under different energy output scenarios, forming a supply and demand deviation vector set containing information such as deviation value, deviation duration, and deviation impact range. Next, using the supply and demand deviation vector set as the core analysis basis, feature extraction is performed on the power comparison chart: for nodes where wind and solar power generation output exceeds demand, features such as the intensity and duration of excess output are captured to generate wind and solar power excess output modal characteristics; for nodes where wind and solar power generation output is basically matched with demand, features such as the stability and fluctuation amplitude of equal output are extracted to generate wind and solar power equal output modal characteristics; for nodes where wind and solar power generation output is lower than demand, features such as the size and distribution of the insufficient output gap are explored to generate wind and solar power insufficient output modal characteristics, providing accurate modal data support for subsequent targeted energy complementarity control strategies.

[0026] Step S400: Based on the excess wind and solar power output mode characteristics, the multi-energy module is adjusted to suppress wind and solar curtailment, thereby obtaining the first energy complementarity control strategy.

[0027] Specifically, firstly, based on the surplus power characteristics of wind and solar power modules, the total surplus wind and solar energy within a predetermined time window is accurately calculated. Then, real-time operating status data of the hydrogen energy modules, such as equipment operating conditions and energy storage capacity, is collected. Combined with constraints on hydrogen production capacity, such as maximum hydrogen production power and energy consumption limits, the energy absorption characteristics are predicted to determine the maximum energy capacity that the hydrogen energy modules can absorb. Subsequently, it is determined whether this absorbable energy is greater than or equal to the surplus wind and solar power. If so, the hydrogen energy module's hydrogen production and storage decision-making is optimized based on the surplus wind and solar power, generating the first control for energy complementarity. The strategy involves the following steps: If the absorbable electrical energy is less than the surplus wind and solar power, first calculate the new surplus electrical energy after deducting the absorbable electrical energy. Simultaneously, formulate a first hydrogen production and storage decision group based on the absorbable electrical energy. Then, filter the hydrogen production capacity constraints to obtain a second hydrogen production and storage decision group and optimize the hydrogen production efficiency to determine the optimal hydrogen production and storage scheme. At the same time, implement flexible power generation restriction and adjustment for wind and solar energy modules to obtain a flexible power generation restriction and adjustment scheme. Finally, integrate the optimal hydrogen production and storage scheme with the flexible power generation restriction and adjustment scheme to form an energy complementary first control strategy that can effectively suppress wind and solar power curtailment.

[0028] Step S500: Based on the equal modal characteristics of wind and solar power output, perform wind and solar fluctuation smoothing compensation adjustment on the multi-energy module to obtain the second energy complementary control strategy.

[0029] Specifically, based on the equal modal characteristics of wind and solar power output, the multi-dimensional fluctuation characteristics of multiple energy modules under the scenario where wind and solar power output and demand are basically matched within a predetermined time window are analyzed in depth, including fluctuation frequency, amplitude, duration and impact range, and the equal modal fluctuation characteristics are extracted. Then, based on these fluctuation characteristics, a coordinated following regulation scheme is formulated for the thermal energy module and the hydrogen energy module in the multi-energy module. The dynamic response of the two modules is coordinated to offset the random fluctuations of wind and solar power generation, forming a fluctuation compensation regulation decision set. At the same time, the historical power supply fluctuation risk record set accumulated by the power plant is used for deep learning training to build a power supply fluctuation risk prediction network that can accurately predict the potential risks of the regulation scheme. Finally, the fluctuation compensation regulation decision set is input into the prediction network. Through multiple rounds of power supply fluctuation risk iteration optimization, the optimal regulation scheme that takes into account the fluctuation compensation effect and power supply stability is selected, generating an energy complementary second control strategy that can achieve smooth compensation of wind and solar fluctuations and ensure stable power supply quality.

[0030] Step S600: Based on the insufficient wind and solar power output mode characteristics, perform energy supplementation optimization adjustment of the multi-energy module under the thermal power generation ratio limit to obtain the third energy complementarity control strategy.

[0031] Specifically, based on the characteristics of insufficient wind and solar power output, the magnitude, distribution, and timing of the power output gap within a predetermined time window are accurately determined. Based on this, a combined thermal and hydrogen power output supplementation decision is made for multi-energy modules including wind, solar, thermal, and hydrogen power. The power generation response speed of the thermal power module, the energy storage and release capacity of the hydrogen power module, and their synergistic complementary effects are comprehensively considered to generate a group of supplementation schemes covering different thermal-hydrogen output ratios. Subsequently, the thermal power generation corresponding to each scheme within the supplementation scheme group is calculated one by one. The proportions are used to form a complete sequence of thermal power generation proportions. Then, based on the preset thermal power generation limit proportions, such as the upper limit of the power generation proportion corresponding to environmental constraints and energy consumption standards, the energy supplementation scheme group is screened, and schemes that exceed the limit proportions are eliminated, and the energy supplementation optimization space that meets the requirements is defined. Finally, with the core power supply quality indicators such as power supply voltage stability, frequency compliance rate, and power supply continuity as the targets, multiple rounds of iterative optimization are carried out within the energy supplementation optimization space to select the optimal energy supplementation scheme that takes into account the energy supplementation efficiency, thermal power generation proportion compliance and power supply quality stability, and generate the third control strategy for energy complementarity.

[0032] In one possible implementation, step S100 further includes:

[0033] Step S110: Based on the predetermined time zone window, predict the power generation demand of the power plant and obtain the power demand baseline curve.

[0034] Step S120: Perform knowledge graph learning on the power generation loss record set of the power plant to construct a power generation loss analytical graph.

[0035] Step S130: Analyze the loss characteristics of the power demand baseline curve based on the power generation loss analysis map to obtain the loss characteristics of multiple nodes.

[0036] Step S140: Perform tolerance compensation on the power demand baseline curve based on the loss characteristics of the multiple nodes to generate the expected power generation curve.

[0037] Specifically, the process begins by collecting historical load data from power plants within the same time window, output correlation data from multiple energy modules, regional customer types, and industry electricity consumption patterns. Simultaneously, it integrates external dynamic factors such as meteorological forecasts (wind speed, sunlight, temperature), macroeconomic electricity consumption trends, and grid dispatch instructions. Data standardization is performed using Z-score normalization, outliers are removed according to the 3σ principle, and missing values ​​are filled using interpolation to form a well-structured dataset. Next, feature extraction is performed. The historical load data is decomposed into trend, periodic, and random components using the STL time series decomposition algorithm to extract time-series features. Finally, the gradient boosting tree algorithm is used to mine the correlation between external factors and power generation. The nonlinear correlation characteristics of power demand are analyzed, and two types of features are fused to form a high-dimensional feature matrix. Subsequently, a long short-term memory network based on an attention mechanism is constructed as the core prediction model. The high-dimensional feature matrix is ​​input into the model, and the long-term dependence of load demand is captured through the input gate, forget gate, and output gate of the long short-term memory network. The attention mechanism is used to strengthen the weight ratio of key factors in critical periods, thereby improving the model's ability to sensitively capture demand fluctuations. Finally, the Kalman filter algorithm is used to dynamically correct the preliminary prediction results of the model output, eliminating the bias caused by random interference. The final output is a power demand baseline curve that can accurately reflect the time-series changes in power generation demand within a predetermined time window.

[0038] First, a data set of power generation losses accumulated during the operation of the power plant is collected. This data set covers loss data of wind power modules, photovoltaic power modules, thermal power modules, and hydrogen power modules under different operating conditions, time periods, and environmental conditions. This includes multi-dimensional information such as the time of loss occurrence, loss type, loss value, related energy module operating parameters, and external influencing factors. Then, the data set undergoes preprocessing, including data cleaning to remove invalid data and correcting abnormal data. Entity recognition technology is used to extract core entities related to losses, such as each energy module, loss type, and influencing factors. Finally, a relation extraction algorithm is used to mine these entities. The system identifies relationships between data points, such as the correspondence between specific modules and certain loss types, and the influence of environmental factors on the degree of loss. Based on the preprocessed structured data, knowledge graph learning is conducted to construct a three-layer knowledge graph consisting of an entity layer, a relationship layer, and an attribute layer. The entity layer clarifies the core elements related to various types of loss, the relationship layer defines the logical connections between elements, and the attribute layer supplements the specific characteristic parameters of each entity. Finally, knowledge fusion eliminates data redundancy and conflicts, and knowledge reasoning refines the implicit relationships in the graph, forming a power generation loss analysis graph that clearly presents the patterns of power generation loss, module relationships, and influencing factors.

[0039] First, the power demand baseline curve is divided into time nodes according to a predetermined time zone window. The power generation demand value corresponding to each time node in the curve and the associated energy module operating scenario are identified. Then, the power generation loss analysis map is used as the core analysis basis. This map has structured knowledge such as loss type, loss cause, loss degree and entity correlation under different operating conditions and time periods of each energy module. Subsequently, for each node of the power demand baseline curve, the corresponding loss correlation rules and historical data are matched from the power generation loss analysis map to analyze the possible loss mode, loss influencing factor and loss value range of each energy module under that node. At the same time, combined with the power generation demand scale and energy output ratio of the node, the actual impact of loss on the power demand of the node is quantitatively analyzed. Finally, through multi-dimensional feature extraction, key parameters such as loss rate, loss duration, loss sensitivity factor and loss compensable range corresponding to each node are obtained to form the loss characteristics of multiple nodes.

[0040] Hierarchical clustering algorithm is used to classify the loss characteristics of multiple nodes, with loss rate, loss sensitivity factor, and compensable range as core clustering indicators. All time-series nodes are divided into three categories: high loss level, medium loss level, and low loss level, clarifying the distribution pattern and loss characteristic differences of nodes at different levels. Subsequently, a dynamic compensation coefficient model based on multiple linear regression is constructed, using node loss level, corresponding node power generation demand scale, and historical compensation effect data as model input variables. The model parameters are solved using the least squares method to establish the mapping relationship between input variables and compensation coefficients. For each node, a personalized tolerance compensation coefficient adapted to its loss characteristics is calculated to ensure that the compensation intensity accurately matches the actual impact of loss. Then, according to the time sequence of a predetermined time zone window, the personalized compensation coefficient of each node is applied one by one to the node power value corresponding to the power demand baseline curve to complete the point-by-point compensation adjustment. Finally, a moving average smoothing algorithm is used to iteratively optimize the compensated curve to eliminate curve fluctuations caused by compensation differences between nodes. At the same time, the curve is verified in combination with the actual operating constraints of the power plant to eliminate abnormal compensation points that exceed the reasonable range. Finally, a power generation expectation curve that takes into account both accurate loss compensation and reasonable demand time sequence is generated.

[0041] In one possible implementation, step S300 further includes:

[0042] Step S310: Perform time-series alignment based on the expected power generation curve and the wind-solar power generation fusion curve to obtain a power comparison chart.

[0043] Step S320: Based on the power comparison diagram, identify the supply and demand deviation of multiple nodes and obtain the supply and demand deviation vector set.

[0044] Step S330: Capture the wind and solar power surplus node features of the power comparison map according to the supply and demand deviation vector set, and generate the wind and solar power surplus mode characteristics.

[0045] Step S340: Capture the wind and solar power output equal node features of the power comparison map according to the supply and demand deviation vector set, and generate the wind and solar power output equal modal characteristics.

[0046] Step S350: Capture the features of insufficient wind and solar power output nodes in the power comparison map according to the supply and demand deviation vector set, and generate the modal characteristics of insufficient wind and solar power output.

[0047] Specifically, using the predetermined time zone window upon which the expected power generation curve and the wind-solar power generation fusion curve are constructed as a unified time reference, the time axis scales and data acquisition frequencies of the two curves are first clearly defined to ensure consistency in the time dimension. Then, a time axis calibration algorithm is used to match each time point of the expected power generation curve (reflecting the expected power demand of power plants within the predetermined time zone window) with the wind-solar power generation fusion curve (reflecting the integrated wind and solar power generation capacity within the predetermined time zone window). This ensures that the two curves correspond to a unique power value under the same time marker, namely the demand power value and the actual wind and solar power generation value. Finally, through data visualization technology, the two matched curves are integrated and plotted on the same coordinate system, with the horizontal axis representing the time series of the predetermined time zone window and the vertical axis representing the power value, forming a power comparison chart that allows for a direct comparison of the actual output of wind and solar power generation versus the expected demand at different times.

[0048] First, all time-series nodes within the predetermined time zone window in the power comparison chart are identified. Each node corresponds to the demand power value of the expected power generation curve and the actual power generation value of the wind-solar power generation fusion curve. Then, a multi-dimensional deviation calculation method is used to traverse each time-series node and calculate the absolute deviation between the actual power generation value and the demand power value, i.e., the difference between the actual value and the demand value, and the relative deviation rate, i.e., the ratio of the absolute deviation to the demand power value. At the same time, the duration of the deviation, the specific time period in which the deviation occurs, and the operating conditions of the energy module of the corresponding node are recorded. Next, the calculation results of each node are feature-integrated to form vector data containing multi-dimensional information such as absolute deviation, relative deviation rate, deviation duration, deviation time period, and operating conditions. Finally, the vector data of all time-series nodes are arranged in chronological order to construct a complete supply-demand deviation vector set. This vector set can comprehensively and accurately quantify the supply-demand differences of each node.

[0049] First, a criterion for determining wind and solar power overcapacity is established. When the actual wind and solar power output at a node in the supply-demand deviation vector set exceeds the demand power value of the expected power output curve, and the relative deviation rate between the two exceeds a preset overcapacity threshold, that node is identified as a wind and solar power overcapacity node. Then, based on a power comparison map and combined with multi-dimensional information such as the absolute deviation, duration of deviation, and time period of deviation corresponding to the overcapacity node in the supply-demand deviation vector set, a feature extraction algorithm is used to mine the core features of the overcapacity node. These features include the peak size of the overcapacity output, the specific time point of the peak, the duration of the overcapacity state, the growth and decay rates of the overcapacity output, and the distribution density of the overcapacity output in different time periods. Finally, the extracted multi-dimensional features are integrated and structured to clarify the key parameters and variation patterns of the overcapacity mode, generating wind and solar power overcapacity mode characteristics that can accurately characterize the overcapacity situation.

[0050] A quantitative standard for determining equivalence was established, defining the absolute deviation between actual wind and solar power generation and demand power as ±5% of rated power and the relative deviation rate threshold as 3%. By traversing the supply-demand deviation vector set, time-series nodes that meet this standard were automatically selected, locking the time distribution range of equivalence nodes in the power comparison chart. Subsequently, a feature extraction algorithm based on a sliding window was used to analyze the power comparison chart data corresponding to equivalence nodes in a 15-minute time window, quantitatively extracting multi-dimensional core features, including the continuous and stable duration of output matching, the peak range of power fluctuation within the window, the fluctuation frequency, the distribution ratio of equivalence output in different time periods, and related operating condition parameters such as the minimum stable output of thermal energy modules and the energy storage maintenance power of hydrogen energy modules under equivalence conditions. Finally, through a feature structuring algorithm, the above quantitative features were logically classified and integrated according to basic characteristics, fluctuation characteristics, and operating condition adaptation characteristics, forming structured data containing key parameters, change patterns, and operating condition adaptation requirements, generating wind and solar power output equivalence modal characteristics that accurately depict the state of wind and solar power output and demand equilibrium matching.

[0051] A clear quantitative standard for determining insufficiency is established, using the following thresholds: actual wind and solar power generation value is lower than the demand value of the expected power generation curve, absolute deviation exceeds -8% of rated power, and relative deviation rate is lower than -5%. By traversing the supply-demand deviation vector set, time-series nodes that meet this standard are automatically selected, locking the time distribution of insufficiency nodes and the corresponding basic data of power gap in the power comparison chart. Subsequently, a sliding window algorithm with a 5-minute step size is used to perform segmented analysis on the time-series data of the power comparison chart corresponding to the insufficiency nodes, quantifying and extracting multi-dimensional core features, including the peak size of the power gap, the continuous duration of the gap, the growth and decay rate of the gap, the distribution density of the gap at different time periods, and the actual power decay ratio of wind power modules and photovoltaic energy modules under the gap state. Finally, through a feature-structured modeling algorithm, the above-mentioned quantitative features are logically classified and integrated according to the gap basic parameters, time-series change characteristics, and module correlation characteristics, forming structured data containing key parameters, change patterns, and operating condition adaptation requirements, generating wind and solar power output insufficiency modal characteristics that can accurately characterize the wind and solar power output gap situation.

[0052] In one possible implementation, step S400 further includes:

[0053] Step S410: Calculate the surplus wind and solar power energy based on the excess wind and solar power output mode characteristics.

[0054] Step S420: Based on the real-time status data of the hydrogen energy module and the hydrogen production capacity constraints, predict the energy absorption characteristics to obtain the absorbable energy.

[0055] Step S430: Determine whether the absorbable electrical energy is greater than or equal to the surplus wind and solar power energy.

[0056] Step S440: If the absorbable electrical energy is greater than or equal to the surplus wind and solar energy, perform hydrogen production and energy storage decision optimization for the hydrogen energy module based on the surplus wind and solar energy to generate the first energy complementarity control strategy.

[0057] Specifically, the core parameters quantified in the surplus power mode characteristics of wind and solar power output are first extracted, including the actual wind and solar power generation of each surplus node, the expected power demand of the corresponding node, the duration and distribution time of the surplus node, and the calculation granularity of 1 minute in the predetermined time zone window. All surplus time-series nodes are traversed. Then, the difference between the actual wind and solar power generation and the demand power of each node is calculated sequentially through the node-by-node power difference calculation model, which is the instantaneous surplus power of a single node. The instantaneous surplus power of a single node is then multiplied by the corresponding time interval, and 1 minute is converted into 1 / 60 hour to obtain the surplus power of a single node during the time period. Finally, the surplus power of all surplus nodes during the time period is accumulated and summed. At the same time, a line transmission loss correction coefficient is introduced, which is determined based on the loss rate of the corresponding time period in the power generation loss analysis map. The accumulated result is dynamically corrected, and finally the total amount and time-series distribution data of the surplus wind and solar power accumulated within the predetermined time zone window are accurately calculated.

[0058] Real-time status data of the hydrogen energy module is collected, including the number of operating electrolyzers, current hydrogen production efficiency, remaining capacity of the storage tank, equipment operating temperature and pressure, cumulative number of electrolyzer start-ups and shutdowns, and explicit hydrogen production capacity constraints, namely, the maximum / minimum electrolysis power in the hydrogen production power constraint parameters, the upper limit of continuous operation time and the upper limit of the number of simultaneous starts for a single unit in the electrolyzer operation constraint parameters, and the upper limit of hydrogen storage pressure. All data is standardized, and parameters such as power, capacity, and pressure are mapped to the [0, 1] interval. Outliers are removed, and missing data is filled in using interpolation to form a regular input dataset. Subsequently, feature engineering is performed, using the real-time status data and constraints as basic features, and derivative features such as the ratio of remaining capacity to upper limit of hydrogen storage pressure and the product of current hydrogen production efficiency and maximum electrolysis power are generated through feature cross-pollination. The top 20 features strongly correlated with energy absorption capacity were selected using the mutual information method, and a high-dimensional feature matrix was constructed. Next, a gradient boosting regression model was built, using a decision tree as the base learner. The decision tree depth was set to 6, the learning rate to 0.1, and the number of base learners to 100. Mean squared error was used as the loss function. Forward step-by-step iterative training was employed, adding a new base learner in each training round to fit the prediction residuals of the previous round model. L2 regularization was introduced to suppress overfitting. Finally, the preprocessed feature matrix was input into the trained GBR model. The model integrates the prediction results of all base learners and outputs the maximum energy absorption value of the hydrogen energy module within a predetermined time window, i.e., the absorbable energy. This value strictly adheres to the hydrogen production capacity constraints to ensure it matches the actual operating capacity of the hydrogen energy module.

[0059] First, the total amount and time-series distribution data of surplus wind and solar energy calculated are retrieved and accurately extracted with the absorbable energy value obtained by the gradient boosting regression model to establish a data comparison module. Then, a numerical magnitude logic judgment algorithm is used to directly compare the total amount of absorbable energy with the total amount of surplus wind and solar energy. At the same time, the time-series distribution characteristics of the two are combined to help verify the matching of absorbable energy and corresponding surplus wind and solar energy in different periods. Finally, a binary judgment result is output, that is, to determine whether the absorbable energy is greater than or equal to the surplus wind and solar energy.

[0060] With the objective functions of maximizing hydrogen production efficiency, minimizing energy storage costs, and minimizing equipment losses, and combining the total amount and time-series distribution data of surplus wind and solar power, a hydrogen production capacity constraint model is constructed using the upper and lower limits of hydrogen production power, the limit on the number of electrolyzers in operation, and the upper limit of hydrogen storage pressure as constraint variables. Subsequently, the particle swarm optimization algorithm parameters are initialized, setting the particle population size to 50, the maximum number of iterations to 100, and the inertia weight to decrease linearly with an initial value of 0.9 and a final value of 0.4. The learning factors are 1.5 and 1.7, respectively. The number of electrolyzers started, the hydrogen production power allocation per electrolyzer, and the energy storage... The temporal planning, which uses the allocation of hydrogen storage capacity in each time period as a particle optimization dimension, iterates the search for the optimal solution in the solution space using a particle swarm optimization algorithm. During the process, the optimal positions of individual particles and the optimal positions of the group are continuously updated until the iteration termination condition is met. Finally, the key control parameters corresponding to the optimal solution are extracted, including the start-up and shutdown sequence of the electrolyzer, the set value of hydrogen production power in each time period, and the charging rhythm planning of the hydrogen storage tank. These parameters are integrated in a logical structure according to the equipment control command - temporal execution scheme - safety constraint threshold to generate an energy complementary first control strategy that can directly guide the operation of the hydrogen energy module, ensuring that surplus wind and solar energy is fully and efficiently converted into hydrogen energy storage.

[0061] In one possible implementation, step S430 further includes:

[0062] Step S431: If the absorbable electrical energy is less than the surplus wind and solar power, calculate and update the surplus electrical energy, and simultaneously make hydrogen production and energy storage decisions for the hydrogen energy module based on the absorbable electrical energy to obtain the first hydrogen production and energy storage decision group.

[0063] Step S432: Filter the first hydrogen production and energy storage decision group according to the hydrogen production capacity constraints to obtain the second hydrogen production and energy storage decision group.

[0064] Step S433: Based on the second hydrogen production and energy storage decision group, perform the optimization of hydrogen production energy efficiency to determine the optimal hydrogen production and energy storage scheme.

[0065] Step S434: Based on the updated surplus power, perform flexible power limiting adjustment on the wind power module and photovoltaic power module to obtain a flexible power limiting adjustment scheme.

[0066] Step S435: Based on the hydrogen production and storage optimization scheme and the flexible emission restriction and adjustment scheme, the first energy complementarity control strategy is generated through coordinated integration.

[0067] Specifically, firstly, the system performs precise calculations using a difference calculation model to calculate the surplus electrical energy = surplus wind and solar power - absorbable electrical energy. It then divides the predetermined time zone window into 1-minute time granularities, simultaneously outputting the total amount and time-series distribution curve of the surplus electrical energy. Simultaneously, using absorbable electrical energy as a rigid constraint upper limit, and combining it with real-time status data of the hydrogen energy module (i.e., the number of operating electrolyzers, current hydrogen production efficiency, remaining capacity of the hydrogen storage tank, and equipment operating temperature / pressure), a decision enumeration algorithm is used to generate multi-scenario solutions around core decision dimensions. Electrolyzer startup combinations cover different start-stop pairings of 1 to 4 devices; hydrogen production power allocation is divided into 10 dynamic adjustment levels with a 5% gradient; hydrogen storage timing planning distinguishes between three modes: peak-priority charging, balanced charging, and off-peak replenishment; and equipment start-stop rhythm is set with three response intervals of 5 minutes, 10 minutes, and 15 minutes. These dimensions are cross-combined to form 80 differentiated hydrogen production and storage solutions. After integration, the first hydrogen production and storage decision group is obtained, ensuring that the decisions cover various scenario requirements such as safe equipment operation and energy efficiency optimization.

[0068] First, the constraints on hydrogen production capacity were quantified into calculable numerical thresholds, clarifying the hydrogen production power constraints: the power range of a single electrolyzer is 50~200kW, the total hydrogen production power does not exceed 800kW, the electrolyzer operation constraints are: the number of electrolyzers started simultaneously is ≤4, the continuous operation time of a single electrolyzer is ≤12 hours, and the hydrogen storage pressure constraints are: the upper limit of the hydrogen storage tank pressure is ≤35MPa. A multi-dimensional constraint verification model was constructed. Then, the constraints of each of the 80 schemes in the first hydrogen production and energy storage decision group were verified one by one. The model automatically compared the matching of the number of electrolyzers started, hydrogen production power allocation, operation time planning, hydrogen storage pressure prediction value and quantified constraint thresholds of each scheme. Invalid schemes that exceeded the upper limit of the number of start-ups, exceeded the range of hydrogen production power, violated the operation time regulations or may exceed the hydrogen storage pressure limit were eliminated. Compliant schemes that met all constraints were retained and finally integrated to form the second hydrogen production and energy storage decision group to ensure that the basic schemes for subsequent optimization have technical feasibility and equipment safety.

[0069] Construct a quantitative model for hydrogen production energy efficiency, based on the amount of hydrogen produced per unit of electricity (Nm³). 3Using ( / kWh) as the core evaluation index, this model integrates key influencing factors such as electrolyzer operating efficiency, hydrogen storage tank charging loss, and equipment start-up and shutdown energy consumption. It substitutes parameters from each scheme in the second hydrogen production and storage decision-making group, including the number of electrolyzers started, hydrogen production power allocation at different times, and hydrogen storage time sequence planning, into the model to accurately calculate the hydrogen production energy efficiency value of each scheme. Subsequently, an improved particle swarm optimization algorithm framework is built, using electrolyzer start-up combination, dynamic hydrogen production power allocation ratio, and hydrogen storage priority as optimization dimensions. The particle population size is set to 40, the maximum number of iterations to 60, and the inertia weight adopts a linear decreasing strategy, initially 0.9, decreasing to 0.4 at the end of the iteration. Learning factors are set to 1.6 and 1.8 respectively. With maximizing hydrogen production energy efficiency as the objective function, the optimal solution is searched through iterative updates of particle position and velocity. A congestion factor is introduced to avoid local optimum traps. The iteration terminates when the maximum number of iterations is reached or the energy efficiency value fluctuation for three consecutive generations is less than 0.008 Nm. 3 When the value is / kWh, the parameter combination corresponding to the current optimal particle is extracted to form a hydrogen production and storage optimization scheme that includes the optimal number of electrolyzers in operation, the hydrogen production power setting value for each time period, and the hydrogen storage tank charging sequence planning, so as to ensure that the hydrogen production energy efficiency is optimal under the constraint of absorbable electrical energy.

[0070] A flexible power limiting regulation scheme is formulated by adopting an implementation approach of time-series decomposition of surplus power, module characteristic adaptation, and quantification of power limiting parameters. First, with a time granularity of 1 minute, the time-series distribution curve of surplus power is decomposed into the power gap to be reduced in each time period. Combining the real-time output data and regulation response characteristics of wind power modules and photovoltaic power modules, the upper limit of the power regulation rate of wind power modules is 0.5MW / min, and the upper limit of the power regulation rate of photovoltaic modules is 0.3MW / min. A proportional allocation algorithm is used to determine the power limiting responsibility ratio of the two modules, which is dynamically allocated according to the current actual output ratio. Then, the power limiting ratio is divided into 5% gradients, and the optimal power limiting ratio of the two modules in each time period is determined by iterative calculation. At the same time, a power limiting smoothing constraint is set, and the fluctuation of the power limiting ratio between adjacent time periods does not exceed 3% to avoid power abrupt changes affecting grid stability. Finally, the power limiting start time, duration, dynamic adjustment power curve, and safe operation threshold of each module are defined to form a structured flexible power limiting regulation scheme covering power limiting time sequence, power allocation, and regulation rate, ensuring that surplus power is reduced smoothly and efficiently.

[0071] A time-constrained collaborative modeling approach is adopted to generate the first control strategy for energy complementarity. First, using a 1-minute time granularity within a predetermined time window as a benchmark, the start-up and shutdown sequence of the electrolyzer in the hydrogen production and storage optimization scheme, and the hydrogen production power curves for each time period, are time-aligned with the wind and solar power curtailment start-up times and dynamic adjustment power in the flexible power curtailment regulation scheme. This ensures that the power adjustment directions of both are consistent and conflict-free at the same time node. Then, a parameter-constrained collaborative model is constructed to verify whether the superposition value of hydrogen production power and wind / solar curtailment power meets the grid acceptance threshold, while also incorporating real-time status data from the hydrogen energy module. By optimizing the adjustment response characteristics of wind and solar modules, key parameters such as the hydrogen production power adjustment rate are synchronized with the wind and solar power curtailment rate to avoid sudden power fluctuations. Finally, through a structured integration algorithm, the coordinated hydrogen production and storage parameters, namely the number of electrolyzers in operation, the hydrogen production power setpoint, the hydrogen storage sequence, and the flexible curtailment parameters, namely the curtailment ratio, adjustment duration, and safety threshold of each module, are logically classified and sorted according to the equipment control command, the sequence execution plan, and the safety constraint standard. This forms a structured document covering the rules for coordinated operation of multiple energy modules, parameter configuration schemes, and emergency adjustment mechanisms, generating a first control strategy for energy complementarity that can be directly implemented.

[0072] In one possible implementation, step S500 further includes:

[0073] Step S510: Perform multidimensional fluctuation characteristic analysis based on the equal modal characteristics of wind and solar power output to obtain the equal modal fluctuation characteristics.

[0074] Step S520: Based on the equal modal fluctuation characteristics, the thermal power module and the hydrogen energy module are coordinated and adjusted to obtain the fluctuation compensation adjustment decision set.

[0075] Step S530: Perform deep learning based on the historical power supply fluctuation risk record set of the power station to obtain a power supply fluctuation risk prediction network.

[0076] Step S540: Based on the power supply fluctuation risk prediction network, perform iterative optimization of the fluctuation compensation adjustment decision set to generate the energy complementarity second control strategy.

[0077] Specifically, using a predetermined time zone window of 1 minute as the minimum analysis granularity, time-series power data from the equivalent modal characteristics of wind and solar power output are extracted. A sliding window algorithm, with a window size of 5 minutes and a step size of 1 minute, is used to segment the data. Subsequently, multi-dimensional fluctuation characteristics are quantified and analyzed for each segment. Power fluctuation amplitude is calculated by subtracting the minimum power from the maximum power within the segment, while simultaneously calculating the fluctuation percentages within the ±1%, ±3%, and ±5% rated power ranges. Fluctuation frequency is calculated by counting the number of times power exceeds the ±2% rated power threshold per unit time, converted to the number of fluctuations per hour. Fluctuation duration characteristics are calculated through the accumulation of continuous fluctuation duration, the statistics of the longest single fluctuation duration, and the average duration of fluctuations with different amplitudes. Additional features include the fluctuation trend slope (i.e., the slope of linear fitting of power within a segment) and the temporal distribution of fluctuation peaks (i.e., the time points when fluctuation peaks occur in each time period). All quantified features are categorized and integrated according to amplitude characteristics, frequency characteristics, duration characteristics, and trend characteristics to form structured equivalent modal fluctuation characteristics.

[0078] Based on the power fluctuation amplitude, frequency, duration, and trend characteristics of equal-modal fluctuations, a quantitative model for fluctuation compensation demand is established. Three compensation scenarios are categorized: small-amplitude high-frequency fluctuations, medium-amplitude sustained fluctuations, and large-amplitude sudden fluctuations. The priority and accuracy requirements of compensation responses under different scenarios are clearly defined. Subsequently, a collaborative following adjustment model for thermal power modules and hydrogen energy modules is constructed. The thermal power module is set to primarily perform basic compensation and slow adjustment, with a power adjustment range of 50-300MW, a response delay ≤10 seconds, and an adjustment rate ≤15MW / minute. The hydrogen energy module is set to secondarily perform rapid replenishment and precise smoothing, with an energy storage release power range of 10-80MW, a response delay ≤2 seconds, and an adjustment rate ≤20MW / minute. Collaboration is achieved through dynamic allocation of compensation weights. Finally, a decision enumeration algorithm is used to generate 40 differentiated compensation schemes covering different fluctuation scenarios, including basic compensation schemes, rapid response schemes, and redundant backup schemes, based on key parameters such as the power output adjustment range of the thermal power module, the energy storage release rhythm of the hydrogen energy module, and the time difference of their collaborative response. These schemes are integrated to obtain a fluctuation compensation adjustment decision set, ensuring coverage of various fluctuation smoothing needs.

[0079] Historical power supply fluctuation risk records of power plants were collected, covering historical fluctuation compensation and adjustment decision parameters, corresponding power fluctuation amplitude / frequency / duration, grid voltage / frequency deviation data, equipment operating status, and final fluctuation risk level, quantified into risk values ​​from 0 to 10. The data was cleaned, outliers were removed, missing values ​​were filled, and the data was standardized. All parameters were mapped to the [0, 1] interval and divided into training and validation sets in a 7:3 ratio. Subsequently, a prediction network architecture based on a deep neural network (DNN) was constructed. The input layer had 20 neurons corresponding to multidimensional feature parameters, and the hidden layers had 3 layers containing 128, 64, and 32 neurons respectively. Activation functions were... The model employs ReLU, with one neuron in the output layer to output a quantified value of power supply fluctuation risk. The activation function is Sigmoid. The mean squared error (MSE) is used as the loss function, and the Adam optimizer is used with a learning rate of 0.002. The network is iteratively trained using the training set, and the model accuracy is evaluated using the validation set after each training round. An early stopping mechanism is introduced: training stops if the loss value on the validation set does not decrease after 5 consecutive rounds. L2 regularization is used to suppress overfitting. When the model training is complete and the prediction error on the validation set is below 3%, the final power supply fluctuation risk prediction network is obtained. This network can accurately predict the corresponding power supply fluctuation risk by adjusting the decision parameters based on the input fluctuation compensation.

[0080] Each decision set in the fluctuation compensation and adjustment decision set, including parameters such as the output adjustment range and response rate of the thermal power module, the energy release power of the hydrogen energy module, the energy replenishment rhythm, and the coordination time difference between the two, is input one by one into the power supply fluctuation risk prediction network. The network outputs a power supply fluctuation risk quantification value corresponding to each decision set, ranging from 0 to 100 points, with lower scores indicating lower risk. Subsequently, an iterative optimization process is initiated. In the first round, the top 15 low-risk candidate schemes are selected by sorting them in ascending order of risk quantification value. In the second round, the feasibility of the candidate schemes is verified by combining the operating constraints of multiple energy modules, such as the adjustment limit of the thermal power module, the upper limit of the energy storage capacity of the hydrogen energy module, and the grid acceptance threshold, and schemes that do not meet the constraints are eliminated. In the third round, the fluctuation compensation accuracy of the remaining schemes is reviewed to ensure that the actual compensation error is ≤ ±1% of the rated power. Finally, the optimal decision with the minimum power supply fluctuation risk and meeting all constraints is locked, and the coordination response details of the thermal power and hydrogen energy modules are further optimized. The logic structure of module control command - timing execution plan - safety constraint standard is integrated to generate a second energy complementary control strategy that can accurately smooth wind and solar fluctuations and minimize power supply risks.

[0081] In one possible implementation, step S600 further includes:

[0082] Step S610: Based on the insufficient wind and solar power output mode characteristics, make a fire-hydrogen coordinated power output decision for the multi-energy module to obtain a group of power output supplementation schemes.

[0083] Step S620: Calculate the thermal power generation ratio for each thermal power generation scheme within the energy replenishment scheme group to obtain the thermal power generation ratio sequence.

[0084] Step S630: Based on the thermal power generation ratio sequence, filter the energy replenishment scheme group according to the thermal power generation limit ratio to obtain the energy replenishment optimization space.

[0085] Step S640: Perform power supply quality iterative optimization based on the energy replenishment optimization space to generate the third energy complementarity control strategy.

[0086] Specifically, core parameters such as the total power deficit, time-series distribution curve, duration of a single deficit, and deficit growth rate are extracted from the characteristics of insufficient wind and solar power output. A deficit quantification model is used to decompose the energy replenishment demand into precise energy replenishment power for each time period. Then, using the operating characteristics of thermal power modules (adjustment range 50-300MW, response delay ≤10 seconds, minimum output adjustment range 5MW) and hydrogen energy modules (energy storage and release power range 10-80MW, response delay ≤2 seconds, adjustment rate ≤20MW / minute) as constraints, a thermal-hydrogen coordinated energy replenishment model is constructed to clarify... The priority of energy replenishment is determined by prioritizing the use of hydrogen energy modules for rapid replenishment during sudden bursts of power shortages, and thermal energy modules for basic replenishment during sustained and stable shortages. Finally, a decision enumeration algorithm is used to generate 100 differentiated energy replenishment schemes covering different energy replenishment scenarios. These schemes are combined with key parameters such as the output level of thermal power generation (divided into 25 levels in 10MW increments), the energy replenishment power allocation ratio of hydrogen energy modules (divided into 20 levels in 5% increments), and the time difference of coordinated response between the two (divided into 1-second intervals within the 0-5 second range).

[0087] The process iterates through each energy replenishment scheme in the group, extracts the predicted thermal power generation corresponding to each scheme, and calculates it based on the output level and operating time of the thermal energy module. This is then compared with the predicted total power generation, which includes the sum of the actual predicted output of wind power, photovoltaic power, and thermal-hydrogen replenishment output. Subsequently, the thermal power generation ratio is calculated for each scheme using the formula: thermal power generation ratio = predicted thermal power generation / predicted total power generation, with a precision unit of 0.01 to ensure the accuracy of the calculation results. Finally, the thermal power generation ratios of all energy replenishment schemes are arranged sequentially according to the original order of the schemes in the group, generating a structured thermal power generation ratio sequence containing the ratio value, the corresponding scheme number, and the calculation timestamp.

[0088] Define the preset thermal power generation limit ratio, including upper and lower limits, such as an upper limit of 40% and a lower limit of 15%, and quantify it into directly comparable numerical thresholds. Then, establish a correlation mapping between the ratio sequence and the energy replenishment scheme group. Iterate through each ratio value in the thermal power generation ratio sequence in sequence, and simultaneously retrieve the corresponding related schemes in the energy replenishment scheme group. Use a threshold verification algorithm to determine whether the ratio value is within the limit ratio range, i.e., greater than or equal to the lower limit and less than or equal to the upper limit. Eliminate energy replenishment schemes whose ratio values ​​exceed the limit range one by one, and retain all compliant schemes that meet the thermal power generation ratio constraints. Classify and organize them according to the original scheme logic to form a structured energy replenishment optimization space, providing a basic scheme pool that meets the ratio limit for subsequent power supply quality optimization.

[0089] A multi-dimensional quantitative evaluation system for power supply quality was constructed, covering core indicators such as voltage deviation (allowable range ±2%), frequency fluctuation (allowable range 50Hz ± 0.2Hz), harmonic content (total harmonic distortion ≤ 5%), and power supply stability (uninterrupted continuous operation duration). The weights of each indicator were determined using the analytic hierarchy process (AHP), and a comprehensive power supply quality score ranging from 0 to 100 was calculated through weighted calculation. Subsequently, with optimal power supply quality as the objective function, an improved particle swarm optimization algorithm framework was built. The optimization dimensions included the allocation of thermal power supply power, the release rate of hydrogen energy modules, and the time difference between their coordinated responses in the energy replenishment optimization space. The population size was set to 50, the maximum number of iterations to 70, and the inertia weight adopted a non-linear decreasing strategy. The learning factor was initially set to 0.9 and decreased to 0.3 at the end of the iteration, with the learning factor set to 1.7 and 1.9 respectively. A mutation mechanism was introduced to avoid getting trapped in local optima. The parameters of each scheme were optimized through algorithm iteration to continuously improve the comprehensive power supply quality score. The optimization was terminated when the maximum number of iterations was reached or the score fluctuation was less than 0.3 points for four consecutive generations. Finally, the key control parameters corresponding to the optimal scheme were extracted, including the dynamic allocation curve of the thermal power module's supplementary power, the response rate setpoint, and the safe operation threshold. These parameters were then logically integrated according to the module supplementary power instruction, the timing execution plan, and the quality constraint standard to generate a third energy complementarity control strategy. This strategy ensured that the output gap was accurately filled and the optimal power supply quality was guaranteed while meeting the thermal power generation ratio limit.

[0090] In one possible implementation, step S600 further includes:

[0091] The multi-energy module includes a wind power module, a photovoltaic power module, a thermal power module, and a hydrogen power module.

[0092] Specifically, the multi-energy modules are the core components supporting the operation of the integrated wind, solar, thermal, and hydrogen power plant. They encompass four types of energy modules with clearly defined functions and complementary synergies: the wind power module, relying on wind turbine generators, converts wind energy into electricity, possessing clean and pollution-free characteristics and output influenced by wind speed, making it a core force in renewable energy generation; the photovoltaic energy module captures solar energy through solar panels and converts it into electricity, with output closely related to sunlight intensity and weather conditions, forming an important combination with the wind power module for renewable energy generation; the thermal power module uses fossil fuels, such as coal and natural gas, as its energy source, providing stable and reliable basic output through combustion power generation, possessing advantages such as a wide adjustment range and stable response, serving as a key support for energy supplementation and peak shaving; the hydrogen energy module combines energy storage and power generation functions, capable of storing surplus electricity through water electrolysis to achieve the utilization of wind and solar power curtailment, and releasing electricity through hydrogen fuel cells when needed, with fast response and flexible adjustment, serving as an important link for achieving multi-energy synergy and complementarity. The coordinated operation of these four modules provides diversified energy support for integrated control.

[0093] Example 2, based on the same inventive concept as the multi-energy complementary wind-solar-fire-hydrogen integrated control method in the foregoing examples, such as... Figure 2 As shown, this application provides a multi-energy complementary integrated control platform for wind, solar, thermal, and hydrogen energy. The platform and method embodiments in this application are based on the same inventive concept. The platform includes:

[0094] The power generation expectation curve construction unit 10 is used to predict the power generation demand of the power station under tolerance compensation according to a predetermined time zone window, and construct the power generation expectation curve. The power station includes a multi-energy module.

[0095] The wind and solar power generation fusion curve establishment unit 20 is used to perform wind and solar power generation power prediction and fusion of the multi-energy modules according to the predetermined time zone window, and establish a wind and solar power generation fusion curve.

[0096] The modal characteristic acquisition unit 30 is used to perform multi-dimensional supply and demand characteristic analysis on the wind and solar power generation fusion curve under time alignment based on the expected power generation curve, and to acquire the modal characteristics of wind and solar power output surplus, wind and solar power output equal, and wind and solar power output deficiency.

[0097] The first control strategy acquisition unit 40 is used to perform wind and solar curtailment suppression and adjustment on the multi-energy module according to the wind and solar power excess output mode characteristics, and acquire the first control strategy for energy complementarity.

[0098] The second control strategy acquisition unit 50 is used to perform wind and solar fluctuation smoothing compensation adjustment on the multi-energy module according to the equal modal characteristics of wind and solar output, and acquire the energy complementary second control strategy.

[0099] The third control strategy acquisition unit 60 is used to perform energy supplementation optimization adjustment of the multi-energy module under the thermal power generation ratio limit according to the insufficient wind and solar power output mode characteristics, and to acquire the energy complementarity third control strategy.

[0100] Furthermore, the platform is also used to implement the following functions:

[0101] Based on the predetermined time zone window, the power generation demand of the power plant is predicted to obtain a power demand baseline curve; knowledge graph learning is performed on the power generation loss record set of the power plant to construct a power generation loss analysis graph; loss characteristics are analyzed on the power demand baseline curve based on the power generation loss analysis graph to obtain multiple node loss characteristics; tolerance compensation is applied to the power demand baseline curve based on the multiple node loss characteristics to generate the expected power generation curve.

[0102] Furthermore, the platform is also used to implement the following functions:

[0103] The power generation expectation curve and the wind-solar power generation fusion curve are time-series aligned to obtain a power comparison map; multi-node supply-demand deviations are identified based on the power comparison map to obtain a supply-demand deviation vector set; the power comparison map is then subjected to wind and solar power excess output node feature capture based on the supply-demand deviation vector set to generate the wind and solar power excess output modal characteristics; the power comparison map is then subjected to wind and solar power equal output node feature capture based on the supply-demand deviation vector set to generate the wind and solar power equal output modal characteristics; and the power comparison map is then subjected to wind and solar power insufficient output node feature capture based on the supply-demand deviation vector set to generate the wind and solar power insufficient output modal characteristics.

[0104] Furthermore, the platform is also used to implement the following functions:

[0105] The surplus wind and solar power is calculated based on the excess wind and solar power output mode characteristics; the energy absorption characteristics are predicted based on the real-time status data of the hydrogen energy module and the hydrogen production capacity constraints to obtain the absorbable energy; it is determined whether the absorbable energy is greater than or equal to the surplus wind and solar power; if the absorbable energy is greater than or equal to the surplus wind and solar power, the hydrogen energy module is optimized for hydrogen production and energy storage based on the surplus wind and solar power to generate the first energy complementarity control strategy.

[0106] Furthermore, the platform is also used to implement the following functions:

[0107] If the absorbable electrical energy is less than the surplus wind and solar power, the surplus electrical energy is calculated and updated. Simultaneously, hydrogen production and storage decisions are made for the hydrogen energy module based on the absorbable electrical energy, resulting in a first hydrogen production and storage decision group. The first hydrogen production and storage decision group is filtered according to the hydrogen production capacity constraints to obtain a second hydrogen production and storage decision group. The hydrogen production energy efficiency is maximized based on the second hydrogen production and storage decision group to determine the optimal hydrogen production and storage scheme. Flexible power generation limiting is applied to the wind power module and photovoltaic energy module based on the updated surplus electrical energy to obtain a flexible power generation limiting scheme. The optimal hydrogen production and storage scheme and the flexible power generation limiting scheme are then integrated to generate the first energy complementarity control strategy.

[0108] Furthermore, the platform is also used to implement the following functions:

[0109] Based on the equivalent modal characteristics of wind and solar power output, multidimensional fluctuation characteristics are analyzed to obtain equivalent modal fluctuation characteristics; based on the equivalent modal fluctuation characteristics, the thermal power module and hydrogen energy module are coordinated and adjusted to obtain a fluctuation compensation adjustment decision set; deep learning is performed on the historical power supply fluctuation risk record set of the power plant to obtain a power supply fluctuation risk prediction network; based on the power supply fluctuation risk prediction network, the fluctuation compensation adjustment decision set is iteratively optimized to generate the energy complementarity second control strategy.

[0110] Furthermore, the platform is also used to implement the following functions:

[0111] Based on the insufficient wind and solar power output mode characteristics, the multi-energy module performs a thermal-hydrogen coordinated supplementary power output decision to obtain a supplementary power scheme group; based on each supplementary power scheme group, the thermal power generation ratio is calculated to obtain a thermal power generation ratio sequence; based on the thermal power generation ratio sequence, the supplementary power scheme group is screened according to the thermal power generation limit ratio to obtain a supplementary power optimization space; based on the supplementary power optimization space, power supply quality is iteratively optimized to generate the third energy complementarity control strategy.

[0112] Furthermore, the platform is also used to implement the following functions:

[0113] The multi-energy module includes a wind power module, a photovoltaic power module, a thermal power module, and a hydrogen power module.

[0114] Example 3: Based on the same inventive concept as the multi-energy complementary wind-solar-fire-hydrogen integrated control method in the foregoing examples, this example provides a computer-readable storage medium for storing software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the multi-energy complementary wind-solar-fire-hydrogen integrated control method in this application. The processor executes the software programs, instructions, and modules stored in the memory to perform various functional applications and data processing of the computer device, thereby realizing the aforementioned multi-energy complementary wind-solar-fire-hydrogen integrated control method.

[0115] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0116] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0117] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A multi-energy complementary wind, solar, thermal, and hydrogen integrated control method, characterized in that, The method includes: Based on the predetermined time zone window, the power generation demand of the power station is predicted under a tolerance compensation, and a power generation expectation curve is constructed. The power station includes multiple energy modules. Based on the predetermined time zone window, the wind and solar power generation power prediction and fusion of the multi-energy modules are performed to establish a wind and solar power generation fusion curve. Based on the expected power generation curve, the wind and solar power generation fusion curve is analyzed for multi-dimensional supply and demand characteristics under time alignment to obtain the wind and solar power output surplus mode characteristics, wind and solar power output equal mode characteristics, and wind and solar power output insufficient mode characteristics. Based on the wind and solar power surplus mode characteristics, the multi-energy module is adjusted to suppress wind and solar curtailment, thereby obtaining the first energy complementarity control strategy. Based on the equal modal characteristics of wind and solar power output, the multi-energy module is adjusted to smooth wind and solar fluctuations to obtain a second energy complementary control strategy. Based on the insufficient wind and solar power output mode characteristics, the multi-energy module is optimized for supplementary energy under the thermal power generation ratio limit to obtain the third energy complementarity control strategy. Based on the equivalent modal characteristics of wind and solar power output, the multi-energy modules are subjected to wind and solar power fluctuation smoothing compensation adjustment to obtain a second energy complementary control strategy, including: Based on the equivalent modal characteristics of wind and solar power output, multidimensional fluctuation characteristics are analyzed to obtain the equivalent modal fluctuation characteristics; Based on the aforementioned equal modal fluctuation characteristics, the thermal power module and the hydrogen energy module are coordinated and adjusted to obtain a fluctuation compensation adjustment decision set. Deep learning is performed on the historical power supply fluctuation risk record set of the power station to obtain a power supply fluctuation risk prediction network; Based on the power supply fluctuation risk prediction network, the fluctuation compensation and adjustment decision set is iteratively optimized to generate the energy complementary second control strategy. Based on the insufficient wind and solar power output mode characteristics, the multi-energy module is subjected to energy supplementation optimization adjustment under the thermal power generation ratio limit to obtain a third energy complementarity control strategy, including: Based on the insufficient wind and solar power output mode characteristics, the multi-energy module is used to make a fire-hydrogen coordinated supplementary power output decision to obtain a group of supplementary power schemes. The thermal power generation ratio is calculated based on each thermal power generation scheme in the energy replenishment scheme group to obtain the thermal power generation ratio sequence. Based on the thermal power generation ratio sequence, the energy replenishment scheme group is screened according to the thermal power generation limit ratio to obtain the energy replenishment optimization space; Based on the energy replenishment optimization space, power supply quality is iteratively optimized to generate the third energy complementarity control strategy.

2. The multi-energy complementary wind, solar, thermal, and hydrogen integrated control method as described in claim 1, characterized in that, Based on the power demand forecast of the power plant under a tolerance compensation according to the predetermined time zone window, a power generation expectation curve is constructed, including: Based on the predetermined time zone window, the power generation demand of the power plant is predicted, and a power demand baseline curve is obtained. Knowledge graph learning is performed on the power generation loss record set of the power plant to construct an analytical graph of power generation loss; Based on the power generation loss analysis map, the power demand baseline curve is analyzed for loss characteristics to obtain the loss characteristics of multiple nodes. The power demand baseline curve is compensated for based on the loss characteristics of the multiple nodes to generate the expected power generation curve.

3. The multi-energy complementary wind, solar, thermal, and hydrogen integrated control method as described in claim 1, characterized in that, Based on the expected power generation curve, a multi-dimensional supply and demand characteristic analysis of the wind and solar power integration curve under time-series alignment is performed, including: Based on the expected power generation curve and the wind-solar power generation fusion curve, a power comparison chart is obtained by time-series alignment. Based on the power comparison diagram, multi-node supply and demand deviations are identified to obtain a supply and demand deviation vector set. Based on the supply-demand deviation vector set, the power comparison map is used to capture the features of wind and solar power surplus nodes, and the wind and solar power surplus mode characteristics are generated. Based on the supply-demand deviation vector set, the power comparison map is subjected to wind and solar power output equal node feature capture to generate the wind and solar power output equal modal characteristics; Based on the supply-demand deviation vector set, the insufficient wind and solar power output node features are captured in the power comparison map to generate the insufficient wind and solar power output modal characteristics.

4. The multi-energy complementary wind, solar, thermal, and hydrogen integrated control method as described in claim 1, characterized in that, Based on the aforementioned wind and solar power surplus mode characteristics, the multi-energy modules are adjusted to suppress wind and solar curtailment, thereby obtaining a first energy complementarity control strategy, including: Calculate the surplus wind and solar power energy based on the aforementioned excess wind and solar power output mode characteristics; Based on the real-time status data of the hydrogen energy module and the constraints of hydrogen production capacity, the characteristics of electricity absorption are predicted to obtain the absorbable electricity. Determine whether the absorbable electrical energy is greater than or equal to the surplus wind and solar power energy; If the absorbable electrical energy is greater than or equal to the surplus wind and solar energy, the hydrogen energy module is optimized for hydrogen production and energy storage based on the surplus wind and solar energy to generate the first energy complementarity control strategy.

5. The multi-energy complementary wind, solar, thermal, and hydrogen integrated control method as described in claim 4, characterized in that, Determining whether the absorbable electrical energy is greater than or equal to the surplus wind and solar power includes: If the absorbable electrical energy is less than the surplus wind and solar power, the surplus electrical energy is calculated and updated, and hydrogen production and energy storage decisions are made for the hydrogen energy module based on the absorbable electrical energy to obtain the first hydrogen production and energy storage decision group. The first hydrogen production and energy storage decision group is filtered according to the hydrogen production capacity constraints to obtain the second hydrogen production and energy storage decision group. Based on the second hydrogen production and energy storage decision group, the hydrogen production energy efficiency is maximized and the optimal hydrogen production and energy storage scheme is determined. Based on the updated surplus power, a flexible power limiting adjustment scheme is obtained by adjusting the power generation of wind power modules and photovoltaic power modules. The first energy complementarity control strategy is generated by synergistically integrating the hydrogen production and energy storage optimization scheme and the flexible emission restriction and regulation scheme.

6. The multi-energy complementary wind, solar, thermal, and hydrogen integrated control method as described in claim 1, characterized in that, The multi-energy module includes a wind power module, a photovoltaic power module, a thermal power module, and a hydrogen power module.

7. A multi-energy complementary integrated control platform for wind, solar, thermal, and hydrogen energy, characterized in that: The platform is used to implement the multi-energy complementary wind, solar, thermal, and hydrogen integrated control method according to any one of claims 1-6, and the platform comprises: The power generation expectation curve construction unit is used to predict the power generation demand of the power station under tolerance compensation based on a predetermined time zone window, and to construct the power generation expectation curve. The power station includes a multi-energy module. The wind and solar power generation fusion curve establishment unit is used to perform wind and solar power generation power prediction and fusion of the multi-energy modules according to the predetermined time zone window, and establish a wind and solar power generation fusion curve. The modal characteristic acquisition unit is used to perform multi-dimensional supply and demand characteristic analysis on the wind and solar power generation fusion curve under time alignment based on the expected power generation curve, and to acquire the modal characteristics of wind and solar power output surplus, wind and solar power output equal, and wind and solar power output insufficient. The first control strategy acquisition unit is used to adjust the wind and solar curtailment suppression of the multi-energy module according to the wind and solar power excess mode characteristics, and acquire the first energy complementary control strategy. The second control strategy acquisition unit is used to perform wind and solar fluctuation smoothing compensation adjustment on the multi-energy module according to the wind and solar output equal modal characteristics, and acquire the energy complementary second control strategy. The third control strategy acquisition unit is used to perform energy supplementation optimization adjustment of the multi-energy module under the thermal power generation ratio limit based on the insufficient wind and solar power output mode characteristics, and to acquire the third control strategy for energy complementarity.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the multi-energy complementary wind, solar, thermal and hydrogen integrated control method as described in any one of claims 1-6.

Images