Drainage basin water, wind and light medium and long term scheduling decision-making method based on double-layer model and terminal

Abstract

The invention discloses a basin water-wind-light medium-and-long-term scheduling decision-making method based on a double-layer model and a terminal. According to the method, medium-and-long-term resource allocation and short-term operation simulation are separated, and a double-layer decision framework of dynamic feedback is constructed: an upper-layer model focuses on medium-and-long-term operation cost minimization, and a lower-layer model optimizes short-term operation cost for diversified water, wind and light resource scenes; and the short-term energy abandoning penalty cost of the lower-layer model is fed back to the upper-layer model so as to correct the medium-and-long-term strategy. According to the method, cooperation of medium-and-long-term economy and short-term operation robustness is ensured, the uncertainty of wind and light resources is effectively dealt with through scene processing, the economy, reliability and renewable energy consumption level of overall operation of the system are remarkably improved, and self-adaptive optimization configuration under multiple time scales is achieved.

Description

Technical Field

[0001] This invention relates to the field of renewable energy system planning and operation technology, and in particular to a watershed water, wind and solar medium- and long-term scheduling decision-making method and terminal based on a two-layer model. Background Technology

[0002] As the proportion of renewable energy sources such as wind and solar power in the power system increases, the intermittency and volatility of their output pose challenges to the stable operation of the system. Hydropower stations, with their flexible regulation and rapid response, have natural conditions for coordinated operation with wind and solar power stations on a watershed scale, forming a multi-energy complementary system of hydropower, wind power, and solar power, thereby improving the overall absorption capacity of renewable energy.

[0003] Currently, optimization studies for this system mostly focus on short-term timescales, such as day-ahead or intraday scheduling. However, the effectiveness of short-term optimization depends on the boundary conditions provided by medium- and long-term optimization. Medium- and long-term optimization must address the long-term forecast uncertainties of water inflow, wind and solar power output, and load demand, and its optimization results directly affect the economic efficiency and reliability of system operation.

[0004] In existing technologies, some studies have proposed medium- to long-term multi-objective stochastic scheduling methods that consider the complementarity of water, wind, and solar power. For example, typical scenario sets are generated using Markov chains, and multi-objective optimization models are constructed to solve the resource allocation strategy for reservoir groups. However, these methods have the following shortcomings: First, in terms of model architecture, existing methods mostly adopt a single-layer optimization structure, directly embedding stochastic programming into medium- to long-term models, resulting in unclear model responsibility boundaries and difficulty in adapting to the needs of hierarchical decision-making in actual scheduling. Second, in terms of uncertainty handling, existing methods mostly rely on one-time scenario generation and static optimization, lacking dynamic feedback mechanisms, and are insufficiently adaptable and robust to diverse real-world scenarios. Third, in terms of time-scale coordination, existing models have a weak characterization of the coupling relationship between medium- to long-term and short-term scenarios, lacking a two-way interaction mechanism, which easily leads to a disconnect between planning and execution.

[0005] Therefore, an optimization configuration method is needed that can effectively cope with medium- and long-term uncertainties, has a dynamic feedback mechanism, and achieves tight coupling across multiple time scales in order to improve the overall system performance. Summary of the Invention

[0006] The technical problem to be solved by this invention is to provide a medium- and long-term scheduling decision-making method and terminal for watershed hydro-wind-solar systems based on a two-layer model, which can realize the dynamic optimization configuration of watershed hydro-wind-solar multi-energy systems on a medium- and long-term time scale, effectively improving the economy, reliability and renewable energy consumption level of system operation.

[0007] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: A watershed water, wind, and solar power scheduling decision-making method based on a two-layer model includes the following steps: Obtain historical watershed operation data for the target watershed; Based on the historical watershed operation data, an upper-level model is constructed with the goal of minimizing the medium- and long-term operating costs of the multi-energy power generation system. Based on the historical watershed operation data, a lower-level model is constructed with the objective of minimizing the short-term operating costs of the multi-energy power generation system under different water, wind, and solar resource scenarios. The medium- and long-term resource allocation strategy output by the upper-level model is coupled to the lower-level model, and the energy curtailment penalty cost output by the lower-level model is coupled to the upper-level model to obtain a two-layer optimization model. The optimal resource allocation strategy for the target watershed in the medium to long term is obtained by iteratively solving the two-layer optimization model. The multi-energy power generation system in the target watershed is scheduled based on the optimal resource allocation strategy.

[0008] To solve the above-mentioned technical problems, another technical solution adopted by the present invention is as follows: The basin-wide water, wind, and solar medium- and long-term scheduling decision-making terminal based on a two-layer model includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements each step of the above-mentioned basin-wide water, wind, and solar medium- and long-term scheduling decision-making method based on a two-layer model.

[0009] The beneficial effects of this invention are as follows: First, by acquiring historical watershed operation data of the target watershed, it provides a basic input for the construction of subsequent models, ensuring that the model parameters can reflect actual operating patterns, thereby enhancing the adaptability to uncertainties in water, wind, and solar resources. Second, based on historical watershed operation data, upper-level and lower-level models are constructed respectively. The upper-level model aims to minimize the medium- and long-term operating costs of the multi-energy power generation system, focusing on medium- and long-term economic optimization, avoiding the problem of ambiguous decision boundaries caused by mixed objectives in single-level models. The lower-level model optimizes short-term operating costs for different water, wind, and solar resource scenarios, improving robustness to diverse actual conditions through scenario-based processing. This application realizes medium- and long-term scheduling decisions through a two-level optimization architecture, separating the decision-making levels and establishing a two-way dynamic feedback mechanism to cope with long-term forecast uncertainties and the need for multi-timescale coordination, thereby realizing dynamic optimization configuration of the watershed's water, wind, and solar multi-energy system on medium- and long-term timescales, effectively improving the system's economic efficiency, reliability, and renewable energy consumption level. Attached Figure Description

[0010] Figure 1 A flowchart of a watershed water, wind, and solar medium- and long-term scheduling decision-making method based on a two-layer model provided in an embodiment of the present invention; Figure 2 The interaction flowchart of the two-layer optimization model provided in the embodiments of the present invention; Figure 3 A schematic diagram of the structure of a watershed water, wind, and solar medium- and long-term scheduling decision-making terminal based on a two-layer model provided in an embodiment of the present invention; Label Explanation: 100. Long-term scheduling decision-making terminal for watershed water, wind and solar power based on a two-layer model; 101. Memory; 102. Processor. Detailed Implementation

[0011] To explain in detail the technical content, objectives, and effects of the present invention, the following description is provided in conjunction with the embodiments and accompanying drawings.

[0012] Embodiments of the present invention provide a medium- to long-term scheduling decision-making method for watershed water, wind, and solar power based on a two-layer model, comprising the following steps: Obtain historical watershed operation data for the target watershed; Based on the historical watershed operation data, an upper-level model is constructed with the goal of minimizing the medium- and long-term operating costs of the multi-energy power generation system. Based on the historical watershed operation data, a lower-level model is constructed with the objective of minimizing the short-term operating costs of the multi-energy power generation system under different water, wind, and solar resource scenarios. The medium- and long-term resource allocation strategy output by the upper-level model is coupled to the lower-level model, and the energy curtailment penalty cost output by the lower-level model is coupled to the upper-level model to obtain a two-layer optimization model. The optimal resource allocation strategy for the target watershed in the medium to long term is obtained by iteratively solving the two-layer optimization model. The multi-energy power generation system in the target watershed is scheduled based on the optimal resource allocation strategy.

[0013] As described above, the beneficial effects of this invention are as follows: First, by acquiring historical watershed operation data of the target watershed, it provides a basic input for the construction of subsequent models, ensuring that the model parameters can reflect actual operating patterns, thereby enhancing the adaptability to uncertainties in water, wind, and solar resources. Second, based on historical watershed operation data, upper-level and lower-level models are constructed respectively. The upper-level model aims to minimize the medium- and long-term operating costs of the multi-energy power generation system, focusing on medium- and long-term economic optimization, avoiding the problem of ambiguous decision boundaries caused by mixed objectives in single-level models. The lower-level model optimizes short-term operating costs for different water, wind, and solar resource scenarios, improving robustness to diverse actual conditions through scenario-based processing. This application realizes medium- and long-term scheduling decisions through a two-level optimization architecture, separating the decision-making levels and establishing a two-way dynamic feedback mechanism to cope with long-term forecast uncertainties and multi-timescale coordination needs, thereby realizing dynamic optimization configuration of the watershed's water, wind, and solar multi-energy system on medium- and long-term timescales, effectively improving the system's economic efficiency, reliability, and renewable energy consumption level.

[0014] Furthermore, the upper-level model constructed based on the historical watershed operation data, with the objective of minimizing the medium- and long-term operating costs of the multi-energy power generation system, includes: The startup cost of non-renewable energy generation in the target basin and the predicted basin operation data are determined based on the historical basin operation data. The medium- and long-term operating costs of the multi-energy power generation system are determined based on the startup costs and the expected commissioning status of the non-renewable energy generator units. Based on the predicted watershed operation data, upper-level constraints are constructed, and based on the upper-level constraints, an upper-level model is constructed with the goal of minimizing the medium- and long-term operation costs.

[0015] As described above, firstly, quantifying start-up costs and predicting basin operation data based on historical watershed operation data avoids biases caused by theoretical assumptions, ensuring the authenticity and reliability of cost parameters and prediction data. Secondly, combining start-up costs with the expected operational status of non-renewable energy generating units to dynamically calculate medium- and long-term operating costs, rather than using static fixed values, allows the cost model to adapt to changes in unit start-up and shutdown during different scheduling cycles, enhancing its adaptability to actual operating scenarios. Finally, defining upper-level constraints using predicted output data ensures that the constraint boundaries closely reflect the actual uncertainties of water, wind, and solar resources, thereby minimizing costs while avoiding the risk of infeasible solutions and significantly improving the model's robustness in complex watershed environments.

[0016] Furthermore, based on the historical watershed operation data, a lower-level model is constructed with the objective of minimizing short-term operating costs under different water, wind, and solar resource scenarios, including: Based on the historical watershed operation data, the fuel cost of non-renewable energy power generation in the target watershed, the energy curtailment penalty cost of the multi-energy power generation system under different water, wind and solar resource scenarios, and the predicted output data are determined respectively. The short-term operating costs of the multi-energy power generation system under different water, wind and solar resource scenarios are determined based on the fuel cost and the energy curtailment penalty cost, respectively. Based on the predicted power output data, lower-level constraints are constructed, and based on the lower-level constraints, a lower-level model is constructed with the objective of minimizing the short-term operating cost.

[0017] As described above, firstly, determining fuel costs, curtailment penalty costs, and predicted output data based on historical watershed operation data captures the long-term fluctuation patterns of water, wind, and solar resources, providing reliable data support for subsequent optimization. Secondly, combining fuel costs and curtailment penalty costs forms a quantitative expression of short-term operating costs. This not only highlights the importance of curtailment penalties in the cost structure but also guides scheduling decisions to prioritize reducing curtailment. Finally, constructing lower-level constraints and forming an optimization model based on predicted output data ensures that the optimization process occurs within the physically feasible boundary, while strengthening the feedback relationship between medium- and long-term strategies and short-term execution. Through these technical solutions, the robustness and adaptability of overall scheduling are significantly improved, thus better addressing short-term uncertainties and achieving close coupling across multiple time scales.

[0018] Furthermore, the two-layer optimization model includes an upper-layer optimization model and a lower-layer optimization model; The medium- to long-term resource allocation strategy output by the upper-level model is coupled to the lower-level model, and the energy curtailment penalty cost output by the lower-level model is coupled to the upper-level model, resulting in a two-layer optimization model including: Based on the medium- and long-term resource allocation strategy output by the upper-level model, the coupling constraints of the lower-level model are constructed to obtain the lower-level optimization model; The lower-level optimization model is used to conduct short-term operation simulations of multiple water, wind and solar resource scenarios to obtain the energy curtailment penalty cost of the multiple water, wind and solar resource scenarios under the coupling constraint conditions. The aggregated penalty cost is obtained by aggregating the energy curtailment penalty costs of the multiple water, wind and solar resource scenarios. The upper-level optimization model is obtained by updating the medium- and long-term operating costs of the multi-energy power generation system based on the aggregated penalty cost. The multi-energy power generation system is simulated in the medium and long term using the upper-level optimization model to obtain the medium and long-term resource allocation strategy of the multi-energy power generation system under the aggregation penalty cost.

[0019] As described above, firstly, the coupling constraints of the lower-level model are constructed based on the medium- and long-term resource allocation strategy output by the upper-level model. This process ensures that short-term operation strictly adheres to key parameters such as the reservoir's target water level, avoiding deviation from the overall planning objectives. Secondly, the lower-level optimization model conducts short-term operational simulations of multiple water, wind, and solar resource scenarios. A diverse set of scenarios generated from historical data covers the full range of fluctuations in wind and solar power output, significantly improving the model's adaptability to uncertainty. Subsequently, the energy curtailment penalty costs of multiple scenarios are aggregated, and the probability weights of different scenarios are combined to form a unified feedback signal. Based on this aggregated penalty cost, the medium- and long-term operating costs are updated, internalizing potential losses in short-term operation into medium- and long-term decisions, prompting medium- and long-term strategies to proactively avoid high-risk states. Finally, the upper-level optimization model conducts medium- and long-term operational simulations under the updated costs, forming a closed-loop iterative process. This strengthens the synergy across multiple time scales, ensuring that the optimization results meet both economic requirements and practical feasibility. This process effectively solves the problem of the disconnect between medium- and long-term optimization and short-term optimization, improving the system's adaptability to uncertainties in water, wind, and solar resources and its overall optimization robustness.

[0020] Furthermore, the medium- and long-term resource allocation strategy output by the upper-level model includes the target water storage level of the reservoir; The coupling constraints for constructing the lower-level model based on the medium- and long-term resource allocation strategy output by the upper-level model include: ; in, represents the final water level of the reservoir on the last day of the medium-to-long-term operating cycle t under the water, wind and solar resources scenario s; K represents the set of all short-term operating cycles under the medium-to-long-term operating cycle t. This indicates the target water level of the reservoir during its medium- to long-term operating cycle t.

[0021] As described above, the target water level output by the upper-level model provides clear boundary conditions for the lower-level model, enabling short-term simulations to strictly meet medium- and long-term water level requirements while considering the uncertainties of wind and solar resources. This approach effectively prevents energy waste or supply shortages caused by uncontrolled water levels. Furthermore, the definition of key variables allows for quantifiable execution of constraints, supporting precise optimization under different scenarios and enhancing the system's adaptability to fluctuations in water, wind, and solar resources. In this way, a two-way interaction between medium- and long-term optimization and short-term operation is achieved, comprehensively improving the system's operational performance and robustness.

[0022] Furthermore, the aggregated penalty cost, obtained by aggregating the energy curtailment penalties from the multiple water, wind, and solar resource scenarios, includes: ; in, This represents the aggregation penalty cost during the medium-to-long-term operating cycle t; Let represent the probability of occurrence of water, scenery, and light resource scene s, and let S represent the set of all water, scenery, and light resource scenes; This represents the penalty cost coefficient for wind and solar power curtailment; This represents the wind curtailment power of a multi-energy power generation system on day k under the scenario of water, wind, and solar resources; Let represent the curtailed solar power of a multi-energy power generation system on day k under the scenario of water, wind, and solar resources s; K represents the set of all short-term operating cycles under the medium-to-long-term operating cycle t.

[0023] As described above, the introduction of scenario occurrence probabilities gives greater weight to the energy curtailment penalty cost of high-probability scenarios during the aggregation process. This avoids the dilution of the impact of high-probability scenarios caused by simple averaging, thus ensuring that the aggregated penalty cost more accurately reflects the characteristics of uncertainty distribution in actual operation. Simultaneously, the combination of the wind and solar curtailment penalty cost coefficients with curtailed wind and solar power accurately captures the energy waste of multi-energy power generation systems on specific operating days, avoiding biases caused by subjective assumptions. Furthermore, the aggregation process covers all short-term operating cycles under the medium-to-long-term operating cycle t. By incorporating daily energy curtailment data throughout the entire cycle, it ensures that the aggregated penalty cost reflects the full-cycle operating characteristics, preventing local short-term fluctuations from misleading medium-to-long-term decisions. Ultimately, this aggregation mechanism based on scenario probabilities, energy curtailment data, and full-cycle coverage enables the upper-level model to dynamically respond to scenario uncertainties fed back by the lower-level model when updating medium-to-long-term operating costs, significantly improving the adaptability of the two-layer optimization model to fluctuations in water, wind, and solar resources.

[0024] Furthermore, updating the medium- to long-term operating costs of the multi-energy power generation system based on the aggregated penalty cost includes: ; in, This represents the medium- to long-term operating cost of the updated multi-energy power generation system; U represents the decision variables of the upper-level optimization model. Indicates the cost of aggregation penalty; This indicates the start-up cost of non-renewable energy generation; This represents the expected operational status of a non-renewable energy generator unit under a medium- to long-term operating cycle t; T represents the set of all medium- to long-term operating cycles.

[0025] As described above, firstly, updating the medium- and long-term operating costs based on aggregated penalty costs allows the upper-level model to absorb the curtailment risk information reflected in the short-term operation of the lower-level model, avoiding cost underestimation due to ignoring resource scenario uncertainties. Secondly, cost updates are performed based on the decision variable U of the upper-level optimization model. U, as the carrier of key strategies such as reservoir water level, ensures a close correlation between cost calculation and medium- and long-term resource allocation strategies, enabling feedback information to directly guide the direction of strategy optimization. Thirdly, updates are combined with aggregated penalty costs, which aggregate wind and solar curtailment losses from different hydropower, wind, and solar resource scenarios. Incorporating these costs into the medium- and long-term cost framework allows the upper-level model to proactively weigh short-term curtailment risks against long-term economic efficiency during the optimization process, improving the adaptability of decision-making. Furthermore, the start-up costs of non-renewable energy generation are integrated and linked to the expected commissioning status. As fixed expenditures for unit start-up and shutdown, the start-up costs are dynamically matched to the actual unit operation plan by multiplying them with the expected commissioning status, avoiding incorrect cost inclusion during non-operational periods and ensuring that the cost structure conforms to physical operating laws. Finally, the expected commissioning status under the medium-to-long-term operating cycle t is considered. This status indicates the planned start-up and shutdown status of the unit in each cycle. By refining the process by cycle, cost updates can respond to dynamic changes in the time scale, ensuring the continuity and accuracy of economic assessment.

[0026] Furthermore, by iteratively solving the two-layer optimization model, the optimal resource allocation strategy for the target watershed in the medium to long term is obtained, including: The two-layer optimization model is iteratively solved, and after each iteration, the convergence condition is checked based on the medium- and long-term resource allocation strategy output by the upper-layer optimization model in two consecutive iterations. If satisfied, the medium-to-long-term resource allocation strategy output by the last iteration of the upper-level optimization model is marked as the optimal resource allocation strategy for the target watershed in the medium-to-long-term time scale. If the conditions are not met, then return to iteratively solving the two-layer optimization model.

[0027] As described above, the convergence criterion is based on the medium- to long-term resource allocation strategy output by the upper-level optimization model in two consecutive iterations, rather than relying on a fixed number of iterations or a single indicator. Since the randomness of water, wind, and solar resource scenarios may cause fluctuations in the results of a single iteration, the similarity between two consecutive strategies can reliably distinguish between true convergence and accidental fluctuations, thus ensuring the adaptability of the detection mechanism to diverse scenarios. Marking the strategy output by the upper-level model in the last iteration as the optimal resource allocation strategy directly links the convergence state to the final decision output, ensuring that the selected strategy is within a stable optimization range and eliminating interference from intermediate transitional strategies. If the convergence condition is not met, the iteration loop is returned, forming a dynamic feedback loop. By continuously adjusting the coupling relationship between the upper-level strategy and the lower-level penalty cost, the model gradually approaches the global optimum, significantly improving its robustness against medium- to long-term uncertainties.

[0028] Furthermore, the short-term operating costs of the multi-energy power generation system under different hydro, wind, and solar resource scenarios are determined based on the fuel cost and the energy curtailment penalty cost, respectively, including: ; in, This represents the short-term operating cost of a multi-energy power generation system under different water, wind, and solar resource scenarios (s). This represents all decision variables of the lower-level model under different water, wind, and solar resource scenarios s; This represents the fuel cost coefficient for non-renewable energy generation. This represents the power output data of non-renewable energy generation on day k under the scenario of water, wind, and solar resources; This represents the penalty cost coefficient for wind and solar power curtailment; This represents the wind curtailment power of a multi-energy power generation system on day k under the scenario of water, wind, and solar resources; Let represent the curtailed solar power of a multi-energy power generation system on day k under the scenario of water, wind, and solar resources s; K represents the set of all short-term operating cycles under the medium-to-long-term operating cycle t.

[0029] As described above, firstly, the fuel cost calculation based on the output data of non-renewable energy generation reflects the direct correlation between power generation and fuel consumption, ensuring that the cost calculation closely reflects the actual operating conditions and avoiding biases caused by empirical estimations. Secondly, the calculation of the curtailment penalty cost using wind and solar power curtailment captures the real-time economic losses from insufficient renewable energy absorption, enabling the model to dynamically respond to fluctuations in wind and solar resources. Thirdly, by summing the accumulated costs over the short-term operating cycle set K, the model integrates the operating data throughout the entire cycle, eliminating the interference of daily random fluctuations and providing a stable and comprehensive cost assessment perspective. Finally, the formula organically integrates fuel costs and curtailment penalty costs into an optimizable objective function, enabling the lower-level model to solve based on clear mathematical logic, significantly improving the feasibility and decision-making quality of multi-timescale collaborative optimization.

[0030] Another embodiment of the present invention provides a watershed water, wind and solar medium- and long-term scheduling decision-making terminal based on a two-layer model, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements each step of the above-described watershed water, wind and solar medium- and long-term scheduling decision-making method based on a two-layer model.

[0031] As described above, the beneficial effects of this invention are as follows: First, by acquiring historical watershed operation data of the target watershed, it provides a basic input for the construction of subsequent models, ensuring that the model parameters can reflect actual operating patterns, thereby enhancing the adaptability to uncertainties in water, wind, and solar resources. Second, based on historical watershed operation data, upper-level and lower-level models are constructed respectively. The upper-level model aims to minimize the medium- and long-term operating costs of the multi-energy power generation system, focusing on medium- and long-term economic optimization, avoiding the problem of ambiguous decision boundaries caused by mixed objectives in single-level models. The lower-level model optimizes short-term operating costs for different water, wind, and solar resource scenarios, improving robustness to diverse actual conditions through scenario-based processing. This application realizes medium- and long-term scheduling decisions through a two-level optimization architecture, separating the decision-making levels and establishing a two-way dynamic feedback mechanism to cope with long-term forecast uncertainties and multi-timescale coordination needs, thereby realizing dynamic optimization configuration of the watershed's water, wind, and solar multi-energy system on medium- and long-term timescales, effectively improving the system's economic efficiency, reliability, and renewable energy consumption level.

[0032] The watershed hydro-wind-solar multi-energy complementary system provided by the embodiments of the present invention, based on a two-layer model, can be applied to the watershed hydro-wind-solar multi-energy complementary system, improving the economy and reliability of watershed hydro-wind-solar multi-energy power generation system dispatch. The following is a detailed description of the specific implementation: Please refer to Figures 1 to 2 Embodiment 1 of the present invention is as follows: The watershed water, wind, and solar power scheduling decision-making method based on a two-layer model specifically includes steps S1-S5: S1. Obtain historical watershed operation data for the target watershed.

[0033] Acquiring historical watershed operational data for the target watershed can be understood as extracting a data set related to water, wind, and solar resources from the actual operational records of the target watershed. This can be achieved through manual compilation of historical records, automatic collection of sensor data, or access to stored information in a database. The primary purpose is to provide the basic inputs required for model building, ensuring that the optimization process reflects actual operational patterns. Specifically, historical watershed operational data refers to various types of data related to the operation of the power generation system recorded within a certain timeframe in the target watershed. This data can include hydrological data, meteorological data, wind and solar power station data, and power system data. Hydrological data includes reservoir characteristic data (such as reservoir capacity curves (water level-capacity relationship curves), reservoir area curves, tailwater level-discharge relationship curves, design dead water level, normal storage water level, flood control limit water level, and flood control high water level) and watershed hydrological station data (such as daily runoff or flow data from major tributary and main stream control hydrological stations, and watershed areal precipitation data). Meteorological data includes conventional meteorological information (such as wind speed, wind direction, temperature, relative humidity, atmospheric pressure, cloud cover, and precipitation) and solar radiation data (such as total horizontal radiation, direct solar radiation, and direct solar radiation). Radiation and diffuse radiation); Wind and solar power plant data includes power plant metadata (such as power plant geographical location, total installed capacity, wind turbine / photovoltaic module model, quantity, layout diagram, inverter efficiency curve, wind turbine power curve, photovoltaic module conversion efficiency) and wind and solar power output data (such as total active power, theoretical maximum output, planned shutdown / maintenance records, and power curtailment order records); Power system data includes load data (such as historical load data of the entire complementary system power supply area or the grid connected to it), grid operation data (such as grid purchase price, wind and solar curtailment penalty cost, and load deficit penalty cost) and other power generation data (such as the installed capacity of non-renewable energy power generation, minimum technical output, ramp rate, start-up and shutdown costs, fuel costs, and historical output curves).

[0034] S2. Construct an upper-level model based on the historical watershed operation data, with the goal of minimizing the medium- and long-term operating costs of the multi-energy power generation system.

[0035] Specifically, step S2 includes the following steps S21-S23: S21. Determine the start-up cost of non-renewable energy power generation in the target basin and predict the basin operation data based on the historical basin operation data.

[0036] Start-up cost refers to the expense incurred when a non-renewable energy generator unit switches from a shutdown state to an operating state. It can be calculated by analyzing historical start-up and shutdown records combined with actual unit parameters, ensuring that the start-up cost reflects the true dynamic characteristics of the system. In this application, non-renewable energy power generation refers to fossil fuel combustion power generation. Predicted watershed operation data refers to estimates of the watershed's operational status over a future period based on historical data. This can be achieved through time series analysis, machine learning prediction, or statistical regression, aiming to capture long-term resource fluctuation patterns and provide a foundation for modeling. Predicted watershed operation data may include long-term inflow forecast sequences for reservoirs, long-term wind and solar power forecast sequences for wind farms and photovoltaic power plants, long-term system load demand forecast sequences, and operating parameters and economic cost parameters for various types of generator units such as hydropower and thermal power.

[0037] S22. Determine the medium- and long-term operating costs of the multi-energy power generation system based on the startup costs and the expected commissioning status of the non-renewable energy generator units.

[0038] Among them, the medium- and long-term operating costs in step S22 include the medium- and long-term start-up costs of thermal power units. Step S22 is as follows: ; This represents the medium- to long-term startup cost of thermal power units in a multi-energy power generation system under the expected operational conditions.

[0039] S23. Construct upper-level constraints based on the predicted watershed operation data, and construct an upper-level model based on the upper-level constraints with the goal of minimizing the medium- and long-term operation costs.

[0040] The upper-level constraints include the water balance constraints of the reservoir, the water level constraints, and the planned output constraints of the wind and solar power stations.

[0041] (1) The specific water balance constraints of the reservoir are: ; in, This represents the reservoir water level during the medium- to long-term operating period t. This indicates the reservoir water level during a medium- to long-term operating cycle t-1. This represents the predicted inflow rate over a medium- to long-term operating cycle t. The average value of the estimated power generation flow during the medium-to-long-term operating period t is used for preliminary balancing at the upper level. The average value of the estimated wastewater discharge during the medium-to-long-term operating cycle t is used for preliminary balancing of the upper layer.

[0042] (2) The specific water level constraints of the reservoir are as follows: ; in, This indicates the minimum allowable water storage capacity of the reservoir; This indicates the maximum allowable water storage capacity of the reservoir.

[0043] (3) The specific planned output constraints for wind and solar power stations are as follows: ; ; in, This represents the planned total power output of the wind farm within a medium- to long-term operating cycle t. This represents the summation of the predicted theoretical power generation for all sub-periods k within the medium-to-long-term operating cycle t. It represents the maximum total power that the wind farm may generate under ideal conditions within the medium-to-long-term operating cycle t. This represents the planned total output of the photovoltaic power plant within a medium- to long-term operating cycle t. This represents the summation of the predicted theoretical power generation for all sub-periods k within the medium-to-long-term operating cycle t. It represents the maximum total power that the photovoltaic power plant can generate under ideal conditions within the medium-to-long-term operating cycle t.

[0044] S3. Construct a lower-level model based on the historical watershed operation data, with the objective of minimizing the short-term operating cost of the multi-energy power generation system under different water, wind, and solar resource scenarios.

[0045] Specifically, step S3 includes the following steps S31-S33: S31. Based on the historical watershed operation data, determine the fuel cost of non-renewable energy power generation in the target watershed, the energy curtailment penalty cost of the multi-energy power generation system under different water, wind and solar resource scenarios, and the predicted output data.

[0046] Fuel cost refers to the economic expenditure incurred during the generation of non-renewable energy power, which can be calculated by multiplying the unit fuel price by the actual consumption. Curtailment penalty cost refers to the economic penalty incurred due to insufficient renewable energy absorption resulting in wind and solar curtailment, which can be quantified by setting the product of the curtailed power and the penalty coefficient. Furthermore, predicted output data refers to the results of estimating the power generation capacity under future hydro-wind-solar resource scenarios based on historical watershed operation data, which can be generated using time series analysis or machine learning algorithms. The purpose of introducing these data is to ensure the objectivity and accuracy of the model input parameters, thereby improving the model's adaptability to real-world scenarios.

[0047] S32. Determine the short-term operating costs of the multi-energy power generation system under different water, wind, and solar resource scenarios based on the fuel cost and the energy curtailment penalty cost.

[0048] Specifically, step S32 includes: ; in, This represents the short-term operating cost of a multi-energy power generation system under different water, wind, and solar resource scenarios (s). This represents all decision variables of the lower-level model under different water, wind, and solar resource scenarios s; This represents the fuel cost coefficient for non-renewable energy generation. This represents the power output data of non-renewable energy generation on day k under the scenario of water, wind, and solar resources; This represents the penalty cost coefficient for wind and solar power curtailment; This represents the wind curtailment power of a multi-energy power generation system on day k under the scenario of water, wind, and solar resources; Let represent the curtailed solar power of a multi-energy power generation system on day k under the scenario of water, wind, and solar resources s; K represents the set of all short-term operating cycles under the medium-to-long-term operating cycle t.

[0049] S33. Construct lower-level constraints based on the predicted output data, and construct a lower-level model based on the lower-level constraints with the goal of minimizing the short-term operating cost.

[0050] The lower-level constraints include thermal power plant operation constraints and data non-negativity constraints. Specifically, the thermal power plant operation constraints are: ; This represents the thermal power output data for day k under the scenario of water, wind, and solar resources. This represents the minimum value of thermal power output data; This represents the maximum value of thermal power output data. The non-negativity constraint specifically requires that all curtailed energy and water resources variables be non-negative.

[0051] S4. Couple the medium- and long-term resource allocation strategy output by the upper-level model to the lower-level model, and couple the energy curtailment penalty cost output by the lower-level model to the upper-level model to obtain a two-layer optimization model.

[0052] The two-layer optimization model includes an upper-layer optimization model and a lower-layer optimization model.

[0053] Specifically, step S4 includes steps S41-S45: S41. Based on the medium- and long-term resource allocation strategy output by the upper-level model, construct the coupling constraints of the lower-level model to obtain the lower-level optimization model.

[0054] The medium- and long-term resource allocation strategy output by the upper-level model includes the target water level of the reservoir. The target water level refers to the reservoir water storage control target set during the medium- and long-term operation cycle to meet the economic and reliability requirements of the system, which is obtained through optimization by the upper-level model. Specifically, the medium- and long-term resource allocation strategy output by the upper-level optimization model also includes: V_i[t] - reservoir capacity (10,000 cubic meters), Q_release_i[t] - discharge flow (m³ / s), P_hydro_i[t] - power generation output (MW), and the planned output variable values ​​P_wind[t] and P_pv[t] for each wind farm w / photovoltaic power plant pv at each time step t.

[0055] Specifically, step S41 includes: ; in, represents the final water level of the reservoir on the last day of the medium-to-long-term operating cycle t under the water, wind and solar resources scenario s; K represents the set of all short-term operating cycles under the medium-to-long-term operating cycle t. This indicates the target water level of the reservoir during its medium- to long-term operating cycle t.

[0056] S42. The lower-level optimization model is used to conduct short-term operation simulations on multiple water, wind and solar resource scenarios to obtain the energy curtailment penalty cost of the multiple water, wind and solar resource scenarios under the coupling constraint conditions.

[0057] Based on the above description, the lower-level optimization model specifically works as follows: Based on the medium- to long-term resource allocation strategy (such as the target water level of a reservoir) given by the upper-level optimization model, it uses stochastic optimization or robust optimization methods on a short-term time scale (such as daily or hourly) to simulate the actual operation of the system under various representative future water, wind, and solar resource scenarios (such as abundant water, normal water, low water, strong wind, and light wind scenarios). The lower-level optimization model integrates the simulation results (such as actual wind and solar curtailment costs, load deficit costs, and frequent unit start-up and shutdown costs) into a curtailment penalty and feeds it back to the upper-level optimization model.

[0058] S43. Aggregate the energy curtailment penalty costs of the multiple water, wind and solar resource scenarios to obtain the aggregated penalty cost.

[0059] Specifically, step S43 includes: ; in, This represents the aggregation penalty cost during the medium-to-long-term operating cycle t; Let represent the probability of occurrence of water, scenery, and light resource scene s, and let S represent the set of all water, scenery, and light resource scenes; This represents the penalty cost coefficient for wind and solar power curtailment; This represents the wind curtailment power of a multi-energy power generation system on day k under the scenario of water, wind, and solar resources; Let represent the curtailed solar power of a multi-energy power generation system on day k under the scenario of water, wind, and solar resources s; K represents the set of all short-term operating cycles under the medium-to-long-term operating cycle t.

[0060] In this embodiment, = , This represents the optimal energy curtailment penalty cost of the lower-level optimization model given the decision variables U and the water, wind, and solar resource scenario s of the upper-level optimization model. Therefore, = .

[0061] S44. Update the medium- and long-term operating costs of the multi-energy power generation system based on the aggregated penalty cost to obtain the upper-level optimization model.

[0062] Specifically, step S44 includes: ; in, This represents the medium- to long-term operating cost of the updated multi-energy power generation system; U represents the decision variables of the upper-level optimization model. Indicates the cost of aggregation penalty; This indicates the start-up cost of non-renewable energy generation; This represents the expected operational status of a non-renewable energy generator unit under a medium- to long-term operating cycle t; T represents the set of all medium- to long-term operating cycles.

[0063] S45. The multi-energy power generation system is simulated for medium- and long-term operation using the upper-level optimization model to obtain the medium- and long-term resource allocation strategy of the multi-energy power generation system under the aggregation penalty cost.

[0064] Based on the above description, the upper-level optimization model specifically optimizes the target water level (or storage capacity) of the reservoir at the end of the period, the overall output curve of the wind and solar power plan, and the planned start-up mode of thermal power plants, using a medium- to long-term time scale (such as week, month, or year) and multiple objectives such as minimizing the total cost and carbon emissions of the multi-energy power generation system. The upper-level optimization model represents future uncertainties with predicted expected values, generating medium- to long-term guiding resource allocation strategies. In this embodiment, the upper-level optimization model takes minimizing the total operating cost of the multi-energy power generation system throughout the entire medium- to long-term operating cycle as its core objective.

[0065] Therefore, in this embodiment, the objective function of the upper-level optimization model is specifically: ; Where T is measured in months, T=1,2,...,12.

[0066] In this embodiment, the objective function of the lower-level optimization model is specifically: ; Where K is a day as the time scale, K=1,2,...,N, and N is the total number of days in the month.

[0067] like Figure 2 As shown, the interaction process between the upper-level optimization model and the lower-level optimization model is as follows: The decision variables, hydropower, wind and solar resources, and power generation load data obtained from the upper-level optimization model are input into the lower-level optimization model. All feasible solutions to the lower-level optimization model are defined based on the lower-level constraints and coupling constraints. The optimizer in the lower-level optimization model simulates different resource scheduling paths (such as increasing thermal power output, utilizing reservoir capacity, increasing hydropower, or curtailing wind and solar power) to minimize the short-term operating cost of the objective function, thus obtaining the optimal curtailment penalty cost under a specific hydropower, wind and solar resource scenario. The optimal curtailment penalty cost output by the lower-level optimization model is then fed back to the upper-level optimization model.

[0068] S5. Iteratively solve the two-layer optimization model to obtain the optimal resource allocation strategy for the target watershed in the medium to long term.

[0069] Specifically, step S5 includes steps S51-S53: S51. Iteratively solve the two-layer optimization model, and after each iteration, detect whether the two-layer optimization model meets the convergence condition based on the medium- and long-term resource allocation strategy output by the upper-layer optimization model in two consecutive iterations.

[0070] S52. If satisfied, the medium-to-long-term resource allocation strategy output by the last iteration of the upper-level optimization model is marked as the optimal resource allocation strategy for the target watershed in the medium-to-long-term time scale.

[0071] S53. If not satisfied, return to iteratively solve the two-layer optimization model.

[0072] In one optional implementation, the two-layer optimization model is iteratively solved using a decomposition coordination algorithm or an intelligent optimization algorithm. The upper-layer optimization model passes the optimization boundary (target water level, planned output) to the lower-layer optimization model, while the lower-layer optimization model feeds back the cost and constraints of the simulation operation to the upper-layer optimization model, until the results of the upper and lower-layer optimization models converge, ultimately yielding the optimal medium- to long-term resource allocation strategy.

[0073] S6. Schedule the multi-energy power generation system in the target basin based on the optimal resource allocation strategy.

[0074] Embodiment 2 of the present invention is as follows: The watershed water-wind-solar-hydro scheduling decision-making method based on a two-layer model differs from Example 1 in that it iteratively solves the two-layer optimization model using a decomposition and coordination algorithm, specifically including: Step 511: Initialize the iteration counter iter=0, and the maximum number of iterations iter max Convergence tolerance and the initial decision variable U of the upper-level model 0 .

[0075] Step 512: Solve the upper-level optimization model to obtain a new set of medium- and long-term resource allocation strategies, which include multiple decision variables obtained in this iteration. In this embodiment, the decision variables obtained by the upper-level optimization model in this iteration are... Specifically, the target water storage level of the reservoir. .

[0076] Step 513: Solve the upper-level optimization model to obtain... As a coupling constraint, it is passed to the lower-level optimization model, and is solved independently for each water, wind, and solar resource scenario to obtain the optimal short-term operating cost for each scenario. .

[0077] Step 514: Aggregate the penalty cost of all water, wind, and light resource scenarios obtained from the lower-level optimization model. This feedback is then sent to the upper-level optimization model. Among these, It will be used in the objective function of the next iteration iter+1 of the upper-level optimization model, thereby feeding back the short-term economic efficiency into medium- and long-term planning.

[0078] Step 515: Based on the upper-level decision variables from two consecutive iterations (decision variables obtained in this round of iteration) Decision variables obtained from the previous iteration This involves checking whether the two-level optimization model satisfies the convergence condition. The specific convergence condition is as follows: ; If the upper-level decision variables meet the convergence condition, proceed to step 517; otherwise, proceed to step 516.

[0079] Step 516: Set the iteration counter iter = iter + 1, and return to step 512 to perform a new round of iterative solution based on the aggregation penalty cost obtained in step 514.

[0080] Step 517: Output the optimal medium-to-long-term resource allocation strategy that satisfies the convergence condition of the upper-level optimization model. Simultaneously, output a detailed simulation report of the lower-level optimization model based on the optimal medium-to-long-term resource allocation strategy across all water, wind, and solar resource scenarios, providing multi-dimensional evaluations of reliability, economy, and other aspects.

[0081] In an optional implementation, step 517, the method for outputting the optimal medium- to long-term resource allocation strategy that satisfies the convergence condition of the upper-level optimization model, specifically includes: (1) Extract the variable values ​​of each reservoir i at each time step t (e.g., month, week) and the planned output variable values ​​of each wind farm w / photovoltaic power plant pv at each time step t from the optimal medium- and long-term resource allocation strategy. The variable values ​​of reservoir i include V_i[t] - reservoir capacity (10,000 cubic meters), Q_release_i[t] - discharge flow (m³ / s) and P_hydro_i[t] - power generation output (MW), and the output variable values ​​of wind farm / photovoltaic power plant are P_wind[t] and P_pv[t], respectively.

[0082] (2) Import the time series data V_i[t] into the MATLAB data processing tool to plot the reservoir capacity process line and the power output process line. The reservoir capacity process line is plotted with time on the horizontal axis and reservoir capacity on the vertical axis. The reservoir capacity process line is the optimal scheduling line, showing when the reservoir stores water and when it releases water. The power output process line shows the process of hydropower output changing over time.

[0083] (3) Plot the wind output P_wind[t], solar output P_pv[t] and load curve on the same graph to intuitively show the process of the optimization scheme using wind and solar energy to meet the demand.

[0084] (4) Directly read the optimal value of the objective function of the upper-level optimization model, that is, the minimum value of the medium- and long-term operating cost of the multi-energy power generation system, decompose the medium- and long-term operating cost, and determine the specific values ​​of fuel cost, start-up and shutdown cost, operation and maintenance cost and penalty cost.

[0085] Please refer to Figure 3 Embodiment 3 of the present invention is as follows: The basin water, wind and solar medium- and long-term scheduling decision-making terminal 100 based on a two-layer model includes a memory 101, a processor 102, and a computer program stored in the memory 101 and running on the processor 102. When the processor 102 executes the computer program, it implements each step in the basin water, wind and solar medium- and long-term scheduling decision-making method based on a two-layer model as described in Embodiment 1 or Embodiment 2.

[0086] In summary, this invention provides a watershed-based medium- and long-term scheduling decision-making method and terminal for water, wind, and solar power based on a two-layer model. By separating medium- and long-term resource allocation from short-term operational simulation, a dynamic feedback two-layer decision-making framework is constructed: the upper-layer model focuses on minimizing medium- and long-term operating costs, while the lower-layer model optimizes short-term operating costs for diverse water, wind, and solar resource scenarios, and feeds back the short-term energy curtailment penalty cost to the upper layer to correct the medium- and long-term strategy. This design not only ensures the synergy between medium- and long-term economic efficiency and short-term operational robustness, but also effectively addresses the uncertainty of wind and solar resources through scenario-based processing and probabilistic aggregation mechanisms, significantly improving the overall economic efficiency, reliability, and renewable energy absorption level of the system, and achieving adaptive optimization configuration across multiple time scales.

[0087] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent modifications made based on the content of the present invention specification and drawings, or direct or indirect applications in related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A watershed water, wind, and solar medium- to long-term scheduling decision-making method based on a two-layer model, characterized in that, Including the following steps: Obtain historical watershed operation data for the target watershed; Based on the historical watershed operation data, an upper-level model is constructed with the goal of minimizing the medium- and long-term operating costs of the multi-energy power generation system. Based on the historical watershed operation data, a lower-level model is constructed with the objective of minimizing the short-term operating costs of the multi-energy power generation system under different water, wind, and solar resource scenarios. The medium- and long-term resource allocation strategy output by the upper-level model is coupled to the lower-level model, and the energy curtailment penalty cost output by the lower-level model is coupled to the upper-level model to obtain a two-layer optimization model. The optimal resource allocation strategy for the target watershed in the medium to long term is obtained by iteratively solving the two-layer optimization model. The multi-energy power generation system in the target watershed is scheduled based on the optimal resource allocation strategy.

2. The method according to claim 1, characterized in that, The upper-level model constructed based on the historical watershed operation data, with the objective of minimizing the medium- and long-term operating costs of the multi-energy power generation system, includes: The startup cost of non-renewable energy generation in the target basin and the predicted basin operation data are determined based on the historical basin operation data. The medium- and long-term operating costs of the multi-energy power generation system are determined based on the startup costs and the expected commissioning status of the non-renewable energy generator units. Based on the predicted watershed operation data, upper-level constraints are constructed, and based on the upper-level constraints, an upper-level model is constructed with the goal of minimizing the medium- and long-term operation costs.

3. The method according to claim 1, characterized in that, The lower-level model constructed based on the historical watershed operation data, with the objective of minimizing short-term operating costs under different water, wind, and solar resource scenarios, includes: Based on the historical watershed operation data, the fuel cost of non-renewable energy power generation in the target watershed, the energy curtailment penalty cost of the multi-energy power generation system under different water, wind and solar resource scenarios, and the predicted output data are determined respectively. The short-term operating costs of the multi-energy power generation system under different water, wind and solar resource scenarios are determined based on the fuel cost and the energy curtailment penalty cost, respectively. Based on the predicted power output data, lower-level constraints are constructed, and based on the lower-level constraints, a lower-level model is constructed with the objective of minimizing the short-term operating cost.

4. The method according to claim 1, characterized in that, The two-layer optimization model includes an upper-layer optimization model and a lower-layer optimization model; The medium- to long-term resource allocation strategy output by the upper-level model is coupled to the lower-level model, and the energy curtailment penalty cost output by the lower-level model is coupled to the upper-level model, resulting in a two-layer optimization model including: Based on the medium- and long-term resource allocation strategy output by the upper-level model, the coupling constraints of the lower-level model are constructed to obtain the lower-level optimization model; The lower-level optimization model is used to conduct short-term operation simulations of multiple water, wind and solar resource scenarios to obtain the energy curtailment penalty cost of the multiple water, wind and solar resource scenarios under the coupling constraint conditions. The aggregated penalty cost is obtained by aggregating the energy curtailment penalty costs of the multiple water, wind and solar resource scenarios. The upper-level optimization model is obtained by updating the medium- and long-term operating costs of the multi-energy power generation system based on the aggregated penalty cost. The multi-energy power generation system is simulated in the medium and long term using the upper-level optimization model to obtain the medium and long-term resource allocation strategy of the multi-energy power generation system under the aggregation penalty cost.

5. The method according to claim 4, characterized in that, The medium- and long-term resource allocation strategy output by the upper-level model includes the target water storage level of the reservoir; The coupling constraints for constructing the lower-level model based on the medium- and long-term resource allocation strategy output by the upper-level model include: ； in, represents the final water level of the reservoir on the last day of the medium-to-long-term operating cycle t under the water, wind and solar resources scenario s; K represents the set of all short-term operating cycles under the medium-to-long-term operating cycle t. This indicates the target water level of the reservoir during its medium- to long-term operating cycle t.

6. The method according to claim 4, characterized in that, The aggregated penalty cost, obtained by aggregating the energy curtailment penalties of the multiple water, wind, and solar resource scenarios, includes: ； in, This represents the aggregation penalty cost during the medium-to-long-term operating cycle t; Let represent the probability of occurrence of water, scenery, and light resource scene s, and let S represent the set of all water, scenery, and light resource scenes; This represents the penalty cost coefficient for wind and solar power curtailment; This represents the wind curtailment power of a multi-energy power generation system on day k under the scenario of water, wind, and solar resources; Let represent the curtailed solar power of a multi-energy power generation system on day k under the scenario of water, wind, and solar resources s; K represents the set of all short-term operating cycles under the medium-to-long-term operating cycle t.

7. The method according to claim 6, characterized in that, Updating the medium- to long-term operating costs of the multi-energy power generation system based on the aggregated penalty cost includes: ； in, This represents the medium- to long-term operating cost of the updated multi-energy power generation system; U represents the decision variables of the upper-level optimization model. This indicates the start-up cost of non-renewable energy generation; This represents the expected operational status of a non-renewable energy generator unit under a medium- to long-term operating cycle t; T represents the set of all medium- to long-term operating cycles.

8. The method according to claim 4, characterized in that, The optimal resource allocation strategy for the target watershed in the medium to long term is obtained by iteratively solving the two-level optimization model. The two-layer optimization model is iteratively solved, and after each iteration, the convergence condition is checked based on the medium- and long-term resource allocation strategy output by the upper-layer optimization model in two consecutive iterations. If satisfied, the medium-to-long-term resource allocation strategy output by the last iteration of the upper-level optimization model is marked as the optimal resource allocation strategy for the target watershed in the medium-to-long-term time scale. If the conditions are not met, then return to iteratively solving the two-layer optimization model.

9. The method according to claim 3, characterized in that, The short-term operating costs of the multi-energy power generation system under different hydro, wind, and solar resource scenarios are determined based on the fuel cost and the energy curtailment penalty cost, respectively, including: ； in, This represents the short-term operating cost of a multi-energy power generation system under different water, wind, and solar resource scenarios (s). This represents all decision variables of the lower-level model under different water, wind, and solar resource scenarios s; This represents the fuel cost coefficient for non-renewable energy generation. This represents the power output data of non-renewable energy generation on day k under the scenario of water, wind, and solar resources; This represents the penalty cost coefficient for wind and solar power curtailment; This represents the wind curtailment power of a multi-energy power generation system on day k under the scenario of water, wind, and solar resources; Let represent the curtailed solar power of a multi-energy power generation system on day k under the scenario of water, wind, and solar resources s; K represents the set of all short-term operating cycles under the medium-to-long-term operating cycle t.

10. A watershed water, wind, and solar medium-to-long-term scheduling decision-making terminal based on a two-layer model, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements each step of the watershed water, wind, and solar medium- and long-term scheduling decision-making method based on a two-layer model as described in any one of claims 1-9.

Images