Strategy rolling-based hydropower dispatching correction method, system and equipment and medium

Abstract

The invention discloses a strategy rolling-based hydropower dispatching correction method, system and device and a medium. The method comprises the steps of obtaining cascade hydropower station group data and initializing a population of an improved moth fire suppression optimization algorithm; optimizing a global objective function by improving a moth fire suppression optimization algorithm in combination with an alternating direction multiplier algorithm, updating generating capacity parameters by applying Levy flight, and outputting an optimized day-level generating capacity and storage capacity plan; constructing a model prediction control objective function, performing rolling optimization on hour-level generating capacity and the updated storage capacity, and outputting an optimal generating capacity instruction of the next hour and an updated storage capacity state; solving a minute-level generating capacity instruction through a real-time optimization algorithm; the real-time storage capacity obtained after the minute-level generating capacity instruction is executed serves as the initial state of model prediction control at the next moment, and the population of the improved moth fire suppression optimization algorithm is synchronously updated; and obtaining a scheduling result report. According to the invention, the scheduling real-time performance and adaptability are obviously improved.

Description

Technical Field

[0001] This invention relates to the field of power system optimization scheduling technology, and in particular to a method, system, equipment and medium for hydropower scheduling correction based on rolling strategy. Background Technology

[0002] As global demand for renewable energy continues to grow, hydropower, as a clean and stable energy source, continues to play a crucial role in the power system. This is especially true for cascade hydropower stations, where upstream and downstream stations are interconnected by water. This connection enables a systematic scheduling approach, effectively improving water energy utilization efficiency and enhancing the economics of power generation.

[0003] However, most current scheduling optimization models still rely on the setting of some static parameters. This results in a slow response speed when faced with dynamic factors such as real-time changes in water flow and market electricity prices, making it difficult for scheduling schemes to adapt to rapidly changing operating environments. Moreover, some optimization algorithms are prone to getting trapped in local optima and lack sufficient global search capabilities, making it difficult to obtain an optimal scheduling scheme that balances economic benefits with system security.

[0004] Furthermore, traditional hierarchical scheduling methods often treat hourly and minute-level scheduling separately, ignoring the inherent coupling between different time scales. This can negatively impact the overall system's coordination. For cascade hydropower station groups, the hydraulic connections between upstream and downstream are inherently complex. Without an effective global coordination mechanism, not only will overall scheduling efficiency be further weakened, but it could also potentially lead to operational risks.

[0005] Another problem is that current scheduling results often lack a systematic data feedback and analysis mechanism, which limits the ability of scheduling schemes to iteratively optimize and revise strategies during long-term operation. In general, existing methods based on static programming, scheduling methods using only simple evolutionary algorithms, or cascade hydropower scheduling methods lacking multi-timescale fusion capabilities do indeed suffer from a series of prominent problems when dealing with real-time operating environments characterized by high real-time requirements, large hydrological fluctuations, and dynamically complex electricity prices, such as slow response, weak coordination capabilities, and low economic efficiency. Summary of the Invention

[0006] In view of the aforementioned existing problems, the present invention is proposed.

[0007] Therefore, this invention provides a strategy-based rolling hydropower scheduling correction method, system, equipment, and medium to solve the problems of slow response, weak coordination, and low economic efficiency in existing cascade hydropower scheduling methods.

[0008] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a hydropower scheduling correction method based on rolling strategy, comprising: Acquire data on a cascade hydropower station group and initialize the population for an improved moth-to-fire optimization algorithm; Based on the initial population and multiple constraints, the global objective function is optimized by combining the improved moth-to-fire optimization algorithm with the alternating direction multiplier algorithm, and the Lévy flight is applied to update the power generation parameters, outputting the optimized daily power generation and reservoir capacity plan. Based on the optimized daily power generation and reservoir capacity plan, a model predicts the control objective function, continuously optimizes the hourly power generation and updated reservoir capacity, and outputs the optimal power generation command and updated reservoir capacity status for the next hour. Based on the optimal power generation instruction for the next hour, the updated reservoir capacity status, real-time inflow, and real-time electricity price, the minute-level power generation instruction is solved using a real-time optimization algorithm. The real-time reservoir capacity obtained after executing the minute-level power generation command is used as the initial state of the model predictive control at the next moment, and the population of the improved moth-to-fire optimization algorithm is updated synchronously. Based on the optimized daily power generation and reservoir capacity plan, the optimal power generation instruction for the next hour and the updated reservoir capacity status, as well as the operational data generated after the execution of the minute-level power generation instruction, a scheduling result report is obtained.

[0009] As a preferred embodiment of the strategy-based rolling hydropower scheduling correction method of the present invention, the step of acquiring cascade hydropower station group data and initializing the population for the improved moth-to-a-flame optimization algorithm includes: Obtain the maximum reservoir capacity, minimum reservoir capacity, initial reservoir capacity, average power generation water consumption rate, and natural inflow data of each hydropower station to construct the physical operating boundary and initial state of the cascade hydropower station group; Based on the historical electricity price and hydrological data, as well as the constructed operating boundary, a time series model is used to predict the spot market electricity price and natural inflow of each power station in the future, providing dynamic input for scheduling optimization. Based on the predicted electricity price and natural inflow results, multiple initial moths containing multiple time-period power generation schemes are randomly generated using the average entropy strategy. High-quality initial solutions are then selected according to the set entropy threshold to obtain the initial population of the improved moth-to-fire optimization algorithm.

[0010] The beneficial effects of this preferred technical solution are that by integrating the physical boundary of the hydropower station, dynamic electricity price and inflow prediction, and population initialization based on the average entropy strategy, the diversity and feasibility of the initial solution are effectively improved, laying a high-quality starting point for subsequent global optimization.

[0011] As a preferred embodiment of the strategy-based rolling hydropower scheduling correction method of the present invention, the step of optimizing the global objective function by combining the improved moth-to-a-flame optimization algorithm with the alternating direction multiplier algorithm includes: Based on the distributed structure of the cascade hydropower station group, a local objective function is assigned to each hydropower station, and the improved moth-to-a-flame optimization algorithm is executed iteratively independently. Based on the local solutions generated by the independent optimization of each power station and the hydraulic connection between upstream and downstream, a hydraulic connection coefficient is introduced into the alternating direction multiplier method to construct a water quantity coupling penalty term. Based on the evolution of the moth's position during the iteration process, logistic chaotic mutation is applied, and long-step random perturbation is performed in conjunction with Levy flight.

[0012] The beneficial effects of this preferred technical solution are that by combining distributed local optimization with global coupling coordination, and introducing logistic chaotic mutation and Lévy flight to enhance population diversity, the algorithm's global search capability and the collaborative optimization accuracy of hydraulic coupling between cascade hydropower stations are effectively improved.

[0013] As a preferred embodiment of the strategy-based rolling hydropower scheduling correction method of the present invention, wherein: the construction of the model predictive control objective function includes: The negative value of the sum of the products of electricity generation and real-time electricity price is taken as the negative revenue item; The smoothing weight of power generation changes is dynamically adjusted based on the standard deviation of fluctuations in real-time inflow monitoring data. Set a storage capacity deviation penalty coefficient that increases linearly over time; Based on the negative return term, dynamic smoothing weights, and storage capacity deviation penalty coefficient, a model prediction control objective function is constructed.

[0014] As a preferred embodiment of the strategy-based rolling hydropower dispatch correction method of the present invention, wherein: the step of solving the minute-level power generation command through a real-time optimization algorithm includes: Use the real-time inflow monitoring data at the current moment as the inflow prediction value for the next moment; Introduce a smoothing penalty term for changes in power generation into the objective function; Based on the real-time inflow forecast and power plant operation constraints, the water balance constraints, reservoir capacity upper and lower limits constraints, and dynamic boundary constraints are linearized and transformed into a linear programming problem. Solving the linear programming problem yields minute-level power generation commands.

[0015] As a preferred embodiment of the strategy-based rolling hydropower dispatch correction method of the present invention, wherein: the step of using the real-time reservoir capacity obtained after executing the minute-level power generation command as the initial state of the model predictive control at the next moment, and synchronously updating the population of the improved moth-to-a-flame optimization algorithm, includes: The real-time storage capacity monitoring data is assigned as the initial state value for the next model prediction and control cycle. By combining real-time power generation with real-time reservoir capacity, a new individual moth location is formed; Based on the updated moth locations, the flames are recalculated and the population structure is reconstructed.

[0016] As a preferred embodiment of the strategy-based rolling hydropower scheduling correction method of the present invention, wherein obtaining the scheduling result report includes: The total revenue is calculated by multiplying the power generation in each time period by the corresponding real-time electricity price and then summing the results. The root mean square value of power generation fluctuation is obtained by averaging the squares of the differences in power generation between adjacent time steps and taking the square root. The total energy storage deviation is obtained by summing the absolute deviations between the actual storage capacity and the target storage capacity at each time step. Based on the total revenue, the root mean square value of power generation volatility, and the total energy storage deviation, a dispatch result report is obtained.

[0017] Secondly, the present invention provides a hydropower scheduling correction system based on rolling strategy, comprising: The population initialization module is used to acquire data from the cascade hydropower station group and initialize the population for the improved moth-to-fire optimization algorithm. The global collaborative optimization module is used to optimize the global objective function based on the initial population and multiple constraints by combining the improved moth-to-fire optimization algorithm with the alternating direction multiplier algorithm, and to apply the Levy flight to update the power generation parameters, and output the optimized daily power generation and reservoir capacity plan. The rolling optimization module is used to construct a model predictive control objective function based on the optimized daily power generation and reservoir capacity plan, continuously optimize the hourly power generation and updated reservoir capacity, and output the optimal power generation command and updated reservoir capacity status for the next hour. The minute-level real-time correction module is used to solve the minute-level power generation instruction based on the optimal power generation instruction for the next hour, the updated reservoir capacity status, real-time inflow, and real-time electricity price, through a real-time optimization algorithm. The state feedback and population update module is used to take the real-time reservoir capacity obtained after executing the minute-level power generation command as the initial state of the model predictive control at the next moment, and to synchronously update the population of the improved moth-to-fire optimization algorithm. The multi-scale scheduling evaluation module is used to obtain a scheduling result report based on the optimized daily power generation and reservoir capacity plan, the optimal power generation instruction for the next hour and the updated reservoir capacity status, and the operating data generated after the execution of the minute-level power generation instruction.

[0018] Thirdly, the present invention provides an electronic device, comprising: Memory, used to store programs; A processor for executing the computer-executable instructions, which, when executed by the processor, implement the steps of the policy-based rolling hydropower scheduling correction method.

[0019] Fourthly, the present invention provides a computer-readable storage medium, comprising: when the program is executed by a processor, the steps of implementing the policy-based rolling hydropower scheduling correction method.

[0020] The beneficial effects of this invention are as follows: By improving the Moth to a Flame (IMFO) optimization algorithm combined with the Alternating Directional Multiplier (ADMM) method for collaborative optimization, and introducing logistic chaotic mutation and Lévy flight to enhance population diversity, the invention achieves the technical effects of improving global search capability, avoiding premature convergence, and effectively coordinating upstream and downstream water volume coupling of cascade power stations. By constructing a Model Predictive Control (MPC) objective function that integrates negative revenue, power generation smoothing penalty, and time-increasing reservoir capacity deviation penalty, and based on real-time inflow dynamic correction prediction, the invention achieves the technical effects of rolling optimization of hourly scheduling instructions, balancing economic efficiency, operational stability, and energy storage targets. Through minute-level linear real-time optimization and closed-loop feedback mechanism, the real-time reservoir capacity after execution is synchronously updated to the initial state of MPC and the IMFO population is reconstructed, achieving the technical effects of tight coupling of multi-timescale scheduling, rapid response of scheduling schemes to hydrological and market dynamic changes, and significantly improving system real-time performance and robustness. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the 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. Wherein: Figure 1 This is a basic flowchart of a strategy-based rolling hydropower scheduling correction method provided in one embodiment of the present invention. Detailed Implementation

[0022] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0023] Example 1, referring to Figure 1 As an embodiment of the present invention, a hydropower scheduling correction method based on rolling strategy is provided, comprising: S100: Acquire data on the cascade hydropower station group and initialize the population for the improved moth-to-fire optimization algorithm; S200: Based on the initial population and multiple constraints, the global objective function is optimized by combining the improved moth-to-fire optimization algorithm with the alternating direction multiplier algorithm, and the Lévy flight is applied to update the power generation parameters, outputting the optimized daily power generation and reservoir capacity plan; S300: Based on the optimized daily power generation and reservoir capacity plan, construct a model predictive control objective function, continuously optimize hourly power generation and updated reservoir capacity, and output the optimal power generation command and updated reservoir capacity status for the next hour. S400: Based on the optimal power generation command for the next hour and the updated reservoir capacity status, real-time inflow and real-time electricity price, the minute-level power generation command is solved through a real-time optimization algorithm. S500: The real-time reservoir capacity obtained after executing the minute-level power generation command is used as the initial state of the model predictive control at the next moment, and the population of the improved moth-to-fire optimization algorithm is updated simultaneously. S600: Based on the optimized daily power generation and reservoir capacity plan, the optimal power generation command for the next hour and the updated reservoir capacity status, as well as the operational data generated after the execution of the minute-level power generation command, a scheduling result report is obtained.

[0024] It should be noted that existing cascade hydropower dispatching methods face a series of prominent challenges during operation. These include: most models rely on static or offline predicted inflow and electricity price data, lacking a rapid response mechanism to real-time hydrological fluctuations and spot market price dynamics, resulting in dispatching schemes lagging behind actual operating conditions; the optimization process often isolates different time scales (such as daily planning, hourly dispatching, and minute control), severing the coupling relationship between multiple scales and making it difficult to achieve refined coordination; traditional algorithms (such as classical heuristics or single-objective programming) are prone to getting trapped in local optima and do not effectively model the complex upstream and downstream hydraulic connections between cascade power stations, resulting in large water volume coordination errors and deviations from energy storage targets; after dispatching, there is a lack of a closed-loop feedback mechanism based on measured conditions (such as actual reservoir capacity), making it impossible to correct model biases and algorithm populations online, thus weakening the robustness of long-term operation; dispatching results usually only focus on power generation or revenue, lacking a comprehensive quantitative evaluation of volatility, energy storage deviations, and system coupling performance, limiting the scientific nature of decision-making and the ability to continuously optimize strategies.

[0025] Therefore, in response to the problems of slow response, weak coordination and low economic efficiency of the existing cascade hydropower dispatching methods, the following steps from S100 to S600 were taken: by integrating the improved moth-to-a-flame optimization algorithm and the alternating direction multiplier method, constructing a multi-time-scale rolling optimization framework and a closed-loop feedback mechanism, the power generation revenue, operational stability, upstream and downstream coordination and real-time response capability of cascade hydropower dispatching were significantly improved.

[0026] Example 2, this is an embodiment of the present invention, which provides a strategy-based rolling hydropower scheduling correction method based on the previous embodiment, including: In this embodiment of the application, step S100, which involves acquiring data on a cascade hydropower station group and initializing the population for the improved moth-to-fire optimization algorithm, includes: Obtain the maximum reservoir capacity, minimum reservoir capacity, initial reservoir capacity, average power generation water consumption rate, and natural inflow data of each hydropower station to construct the physical operating boundary and initial state of the cascade hydropower station group; Based on historical electricity prices and hydrological data, as well as the constructed operational boundary, a time series model is used to predict future spot market electricity prices and natural inflows to each power station, providing dynamic input for scheduling optimization. Based on the predicted electricity price and natural inflow results, multiple initial moths containing multiple time-period power generation schemes are randomly generated using the average entropy strategy. High-quality initial solutions are then selected according to the set entropy threshold to obtain the initial population of the improved moth-to-fire optimization algorithm.

[0027] In this embodiment of the application, step S100 involves collecting parameters of the hydropower station, including the maximum reservoir capacity. Minimum storage capacity Initial storage capacity Average water consumption rate for power generation Historical reality naturally flows into the stream Real-time monitoring of natural inflow and the target electricity price parameters used for optimization. ; In this embodiment, the electricity price and inflow prediction model collects historical electricity price and hydrological data from the past 24 months, uses the ARIMA time series model to analyze its trends and periodicity, and then predicts the spot market electricity price for the next 24 hours. and various hydropower stations Predicting natural inflow ; In an optional implementation, the electricity price and inflow prediction model in step S100 can also use LSTM (Long Short-Term Memory Neural Network) for multivariate time-series prediction. Historical electricity prices, inflow rates, rainfall, holidays, etc., are used as input features to train an end-to-end model that outputs future electricity price and inflow predictions.

[0028] In an optional implementation, the electricity price and inflow prediction model in step S100 can also construct a Prophet model based on historical electricity price and natural inflow time series data from the past 24 months to automatically fit trend, periodicity and holiday effects, thereby generating the predicted values ​​of spot market electricity price and natural inflow of each hydropower station for the next 24 hours.

[0029] Utilizing historical facts to naturally flow With prediction of natural inflow The difference is used to calculate the standard deviation of the inflow prediction error. .

[0030] In this embodiment of the application, step S200 optimizes the global objective function by combining the improved moth-to-a-flame optimization algorithm with the alternating direction multiplier algorithm, including: Based on the distributed structure of the cascade hydropower station group, a local objective function is assigned to each hydropower station, and the improved moth-to-a-flame optimization algorithm is executed iteratively independently. Based on the local solutions generated by the independent optimization of each power station and the hydraulic connection between upstream and downstream, a hydraulic connection coefficient is introduced into the alternating direction multiplier method to construct a water quantity coupling penalty term. Based on the evolution of the moth's position during the iteration process, logistic chaotic mutation is applied, and long-step random perturbation is performed in conjunction with Levy flight.

[0031] In this embodiment of the application, the diversity enhancement mechanism in the improved moth-to-fire optimization algorithm (IMFO) in step S200 improves population diversity, avoids premature convergence, and enhances global search capabilities by applying Logistic chaotic mutation to the moth position during the iteration process and combining it with Lévy flight for long-step random perturbation.

[0032] In an alternative implementation, the diversity enhancement mechanism in the improved Moth to Fire Optimization (IMFO) algorithm in step S200 can also employ a combination of Gaussian mutation and Cauchy perturbation. This involves applying small Gaussian perturbations to the vicinity of the current optimal solution to perform a local fine search, while simultaneously using Cauchy distribution to generate large jumps to explore new regions, thus balancing development and exploration.

[0033] In an optional implementation, the diversity enhancement mechanism in the improved moth-to-fire optimization algorithm (IMFO) in step S200 can also adopt a differential evolution (DE) mutation strategy. In each iteration, a differential vector is generated for each moth individual based on three different individuals in the population, and this mutation vector is used to replace the original chaotic and Lévy perturbation steps to update the moth position, thereby enhancing population diversity and search capability, while the rest of the IMFO framework and scheduling process remain unchanged.

[0034] In this embodiment of the application, step S200 randomly generates NP moths using an average entropy strategy. Each moth contains a 24-hour power generation plan, and the average entropy formula is: in, Population entropy, used to assess the diversity of initial solutions. The number of moths randomly generated when initializing the IMFO population. Let be the amount of electricity generated by the i-th moth at time t.

[0035] Set threshold ,like If the above is not the case, then the scheme will be retained, a set of high-quality initial power generation will be selected, and the total average power generation will be calculated based on the overall mean method. .

[0036] In this embodiment, a model predictive control objective function is constructed based on the negative revenue term, dynamic smoothing weights, and reservoir capacity deviation penalty coefficient. Specifically, the inflow forecast is corrected using an exponential moving average algorithm, updating the inflow forecast values ​​for all future time steps to ensure that the forecast values ​​reflect the latest real-time data trends. Furthermore, the reservoir capacity boundary is optimized by combining water balance, reservoir capacity, and daily power generation upper limit constraints with dynamically adjusted parameters to ensure operational safety and meet market load demands.

[0037] In this embodiment of the application, the objective function F is defined as follows: in, ; In this embodiment of the application, the power generation and storage capacity Applying physical and operational constraints to ensure the safety of hydropower station operation and meet market load demands, the water balance constraint formula is expressed as: in, and For the storage capacity of days t and t+1, The hydraulic connection coefficient is estimated by analyzing the flow transfer efficiency of upstream and downstream hydropower stations using a hydrological model (based on the water flow transmission characteristics of cascade hydropower stations) and combining historical operating data and hydrological monitoring. This refers to the power generation of the upstream hydropower station; Storage capacity constraints are expressed as: The daily power generation ceiling constraint is expressed as: The month-end energy storage constraint is expressed as: Local objective function Defined as: in, For hydroelectric power station During the period Electricity generation, To predict spot market electricity prices, For smoothing weights; The global objective function is expressed as: in, The global objective function is used to coordinate the optimization results of the cascade hydropower station group, while the local objective function of each hydropower station is represented. The sum of (power generation revenue and fluctuations), plus the upstream and downstream water volume coupling penalty term, ensures consistent hydraulic connections. As a coordination factor, different [factors / factors] were tested through multiple simulation experiments. The impact of values ​​(e.g., 0.1, 0.2, 0.3) on upstream and downstream water flow coordination and optimization convergence is investigated. The optimal value is selected as the empirical value for balancing coordination and computational efficiency. Here, the downstream discharge flow from the upstream power station s is determined by the hydraulic connection coefficient. Transmitted to downstream power plants This constitutes its controlled inflow component.

[0038] In this embodiment, the moth's position is updated through the core iterative step of IMFO, the Moth-to-Flame Algorithm (MFO). Represented as: in, Location of the moth (power generation) ), This represents the location of the flame (current optimal power generation). The parameters of the logarithmic spiral function control the moth (power generation scheme) to move closer to the flame (optimal solution). Amplitude, b is a constant of the logarithmic spiral function, MFO standard setting, t∈[−2+ [1] (Iteration count iter) Linear variation, maximum 100). To simulate the periodicity of the spiral trajectory of a moth; In this embodiment of the application, the Logistic chaotic mutation is represented as follows: in, For chaotic variables, ∈[0,1], initial values ​​are randomly generated; In this embodiment of the application, the disturbance applied by Lévy flight is represented as follows: in, The parameters controlling the step size distribution during Lévy's flight determine the stochastic step size characteristics of the moth's position update. The larger the stride length, the more it tends to favor long-distance jumps, according to Lévy's standard flight settings. This is a chaos factor used to control the step size of Lévy's flight. It is typically generated through chaotic maps (such as Logistic maps) to introduce randomness and enhance search diversity. and These are the maximum and minimum values ​​of the objective function at time t and state s, respectively, obtained through an adaptive boundary estimation method.

[0039] In this embodiment of the application, step S300, which involves constructing the model prediction control objective function, includes: The negative value of the sum of the products of electricity generation and real-time electricity price is taken as the negative revenue item; The smoothing weight of power generation changes is dynamically adjusted based on the standard deviation of fluctuations in real-time inflow monitoring data. Set a storage capacity deviation penalty coefficient that increases linearly over time; Based on the negative return term, dynamic smoothing weights, and storage capacity deviation penalty coefficient, a model predicts the control objective function.

[0040] In this embodiment of the application, the design of the storage capacity deviation penalty coefficient in MPC in step S300 is to set a storage capacity deviation penalty coefficient that increases linearly with time (the closer to the end of the scheduling period, the heavier the penalty), so as to guide the storage capacity at the end of the month to approach the target value.

[0041] In an optional implementation, the design of the storage capacity deviation penalty coefficient in MPC in step S300 can also be incorporated into the objective function of model predictive control, so that the penalty weight of storage capacity deviation gradually increases exponentially as the scheduling cycle progresses, especially significantly increasing the penalty intensity near the end of the scheduling period.

[0042] In an optional implementation, in step S300, the penalty weight in the objective function of model predictive control is adjusted in real time according to the magnitude of the deviation between the current storage capacity and the target storage capacity—the larger the deviation, the stronger the penalty, thereby forming a feedback adjustment mechanism to more flexibly guide the energy storage state to approach the target.

[0043] In this embodiment of the application, step S400, which involves solving the minute-level power generation command using a real-time optimization algorithm, includes: Use the real-time inflow monitoring data at the current moment as the inflow prediction value for the next moment; Introduce a smoothing penalty term for changes in power generation into the objective function; Based on the real-time inflow forecast and power plant operation constraints, the water balance constraints, reservoir capacity upper and lower limits constraints, and dynamic boundary constraints are linearized and transformed into a linear programming problem. The problem is solved using linear programming to obtain minute-level power generation commands. Specifically, the optimal power generation is quickly calculated using linear optimization methods, significantly improving the efficiency and accuracy of generating minute-level power generation commands and commands.

[0044] In an alternative implementation, the solution method for minute-level real-time optimization in step S400 can also adopt quadratic programming. If the quadratic form of the power generation smoothing term is retained (because it is a convex function), a QP model can be constructed to more accurately characterize the fluctuation penalty while ensuring optimality.

[0045] In an optional implementation, the solution method for minute-level real-time optimization in step S400 can also use rule-based heuristic control (such as a fuzzy logic controller), which can quickly output power generation instructions based on inputs such as real-time reservoir capacity deviation, electricity price level, and inflow change rate through preset rules.

[0046] In this embodiment of the application, step S500 uses the real-time reservoir capacity obtained after executing the minute-level power generation command as the initial state for the model predictive control at the next moment, and synchronously updates the population of the improved moth-to-a-flame optimization algorithm, including: The real-time storage capacity monitoring data is assigned as the initial state value for the next model prediction and control cycle. By combining real-time power generation with real-time reservoir capacity, a new individual moth location is formed; Based on the updated moth locations, the flame is recalculated and the population structure is reconstructed. Specifically, through a data feedback mechanism, real-time reservoir capacity monitoring data is directly used as the MPC state for the next moment, ensuring that the control state reflects the actual operating status of the hydropower station in real time. Combined with IMFO, the population is dynamically updated by using real-time power generation and reservoir capacity as the new moth locations and recalculating the flame, thereby enhancing the algorithm's adaptability to hydrological and electricity price fluctuations.

[0047] In this embodiment of the application, obtaining the scheduling result report in step S600 includes: The total revenue is calculated by multiplying the power generation in each time period by the corresponding real-time electricity price and then summing the results. The root mean square value of power generation fluctuation is obtained by averaging the squares of the differences in power generation between adjacent time steps and taking the square root. The total energy storage deviation is obtained by summing the absolute deviations between the actual storage capacity and the target storage capacity at each time step. Based on the total revenue, the root mean square value of power generation volatility, and the total energy storage deviation, a dispatch result report is obtained. Specifically, through data analysis, economic indicators, power generation volatility, energy storage deviation, and upstream and downstream water coupling errors are calculated to generate a dispatch result report containing multi-dimensional indicators for dispatchers' reference and decision-making.

[0048] Example 3 is an embodiment of the present invention. This embodiment differs from the first embodiment in that it provides a hydropower scheduling correction system based on rolling strategy.

[0049] It should be noted that the technical solution of the hydropower scheduling correction system based on rolling strategy is based on the same concept as the technical solution of the hydropower scheduling correction method based on rolling strategy described above. For details not described in detail in the technical solution of the hydropower scheduling correction system based on rolling strategy in this embodiment, please refer to the description of the technical solution of the hydropower scheduling correction method based on rolling strategy described above.

[0050] This embodiment provides a hydropower scheduling correction system based on rolling strategy, comprising: The population initialization module is used to acquire data from the cascade hydropower station group and initialize the population for the improved moth-to-fire optimization algorithm. The global collaborative optimization module is used to optimize the global objective function based on the initial population and multiple constraints by combining the improved moth-to-fire optimization algorithm with the alternating direction multiplier algorithm, and to apply the Levy flight to update the power generation parameters, and output the optimized daily power generation and reservoir capacity plan. The rolling optimization module is used to construct a model predictive control objective function based on the optimized daily power generation and reservoir capacity plan, continuously optimize the hourly power generation and updated reservoir capacity, and output the optimal power generation command and updated reservoir capacity status for the next hour. The minute-level real-time correction module is used to solve the minute-level power generation instruction based on the optimal power generation instruction for the next hour, the updated reservoir capacity status, real-time inflow, and real-time electricity price, through a real-time optimization algorithm. The state feedback and population update module is used to take the real-time reservoir capacity obtained after executing the minute-level power generation command as the initial state of the model predictive control at the next moment, and to synchronously update the population of the improved moth-to-fire optimization algorithm. The multi-scale scheduling evaluation module is used to obtain a scheduling result report based on the optimized daily power generation and reservoir capacity plan, the optimal power generation instruction for the next hour and the updated reservoir capacity status, and the operating data generated after the execution of the minute-level power generation instruction.

[0051] This embodiment also provides an electronic device applicable to a strategy-based rolling hydropower scheduling correction method, including: The system includes a memory and a processor. The memory stores computer-executable instructions, and the processor executes these instructions to implement a strategy-based rolling hydropower scheduling correction method as proposed in the above embodiments.

[0052] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements a strategy-based rolling hydropower scheduling correction method as proposed in the above embodiments.

[0053] The storage medium proposed in this embodiment and the method for implementing a strategy-based rolling hydropower scheduling correction proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0054] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0055] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A hydropower scheduling correction method based on rolling strategy, characterized in that, include: Acquire data on a cascade hydropower station group and initialize the population for an improved moth-to-fire optimization algorithm; Based on the initial population and multiple constraints, the global objective function is optimized by combining the improved moth-to-fire optimization algorithm with the alternating direction multiplier algorithm, and the Lévy flight is applied to update the power generation parameters, outputting the optimized daily power generation and reservoir capacity plan. Based on the optimized daily power generation and reservoir capacity plan, a model predicts the control objective function, continuously optimizes the hourly power generation and updated reservoir capacity, and outputs the optimal power generation command and updated reservoir capacity status for the next hour. Based on the optimal power generation instruction for the next hour, the updated reservoir capacity status, real-time inflow, and real-time electricity price, the minute-level power generation instruction is solved using a real-time optimization algorithm. The real-time reservoir capacity obtained after executing the minute-level power generation command is used as the initial state of the model predictive control at the next moment, and the population of the improved moth-to-fire optimization algorithm is updated synchronously. Based on the optimized daily power generation and reservoir capacity plan, the optimal power generation instruction for the next hour and the updated reservoir capacity status, as well as the operational data generated after the execution of the minute-level power generation instruction, a scheduling result report is obtained.

2. The hydropower scheduling correction method based on rolling strategy as described in claim 1, characterized in that: The process of acquiring data from the cascade hydropower station group and initializing the population for the improved moth-to-fire optimization algorithm includes: Obtain the maximum reservoir capacity, minimum reservoir capacity, initial reservoir capacity, average power generation water consumption rate, and natural inflow data of each hydropower station to construct the physical operating boundary and initial state of the cascade hydropower station group; Based on the historical electricity price and hydrological data, as well as the constructed operating boundary, a time series model is used to predict the spot market electricity price and natural inflow of each power station in the future, providing dynamic input for scheduling optimization. Based on the predicted electricity price and natural inflow results, multiple initial moths containing multiple time-period power generation schemes are randomly generated using the average entropy strategy. High-quality initial solutions are then selected according to the set entropy threshold to obtain the initial population of the improved moth-to-fire optimization algorithm.

3. The hydropower scheduling correction method based on strategy rolling as described in claim 1 or 2, characterized in that: The optimization of the global objective function by combining the improved moth-to-flame optimization algorithm with the alternating direction multiplier algorithm includes: Based on the distributed structure of the cascade hydropower station group, a local objective function is assigned to each hydropower station, and the improved moth-to-a-flame optimization algorithm is executed iteratively independently. Based on the local solutions generated by the independent optimization of each power station and the hydraulic connection between upstream and downstream, a hydraulic connection coefficient is introduced into the alternating direction multiplier method to construct a water quantity coupling penalty term. Based on the evolution of the moth's position during the iteration process, logistic chaotic mutation is applied, and long-step random perturbation is performed in conjunction with Levy flight.

4. The hydropower scheduling correction method based on rolling strategy as described in claim 3, characterized in that: The constructed model predicts the control objective function, including: The negative value of the sum of the products of electricity generation and real-time electricity price is taken as the negative revenue item; The smoothing weight of power generation changes is dynamically adjusted based on the standard deviation of fluctuations in real-time inflow monitoring data. Set a storage capacity deviation penalty coefficient that increases linearly over time; Based on the negative return term, dynamic smoothing weights, and storage capacity deviation penalty coefficient, a model prediction control objective function is constructed.

5. The hydropower scheduling correction method based on rolling strategy as described in claim 4, characterized in that: The process of solving minute-level power generation commands using a real-time optimization algorithm includes: Use the real-time inflow monitoring data at the current moment as the inflow prediction value for the next moment; Introduce a smoothing penalty term for changes in power generation into the objective function; Based on the real-time inflow forecast and power plant operation constraints, the water balance constraints, reservoir capacity upper and lower limits constraints, and dynamic boundary constraints are linearized and transformed into a linear programming problem. Solving the linear programming problem yields minute-level power generation commands.

6. The hydropower scheduling correction method based on rolling strategy as described in claim 5, characterized in that: The process of using the real-time reservoir capacity obtained after executing the minute-level power generation command as the initial state for model predictive control at the next moment, and synchronously updating the population of the improved moth-to-fire optimization algorithm, includes: The real-time storage capacity monitoring data is assigned as the initial state value for the next model prediction and control cycle. By combining real-time power generation with real-time reservoir capacity, a new individual moth location is formed; Based on the updated moth locations, the flames are recalculated and the population structure is reconstructed.

7. The hydropower scheduling correction method based on rolling strategy as described in claim 6, characterized in that: The obtained scheduling result report includes: The total revenue is calculated by multiplying the power generation in each time period by the corresponding real-time electricity price and then summing the results. The root mean square value of power generation fluctuation is obtained by averaging the squares of the differences in power generation between adjacent time steps and taking the square root. The total energy storage deviation is obtained by summing the absolute deviations between the actual storage capacity and the target storage capacity at each time step. Based on the total revenue, the root mean square value of power generation volatility, and the total energy storage deviation, a dispatch result report is obtained.

8. A hydropower scheduling correction system based on rolling strategy, using the method described in any one of claims 1-7, characterized in that, include: The population initialization module is used to acquire data from the cascade hydropower station group and initialize the population for the improved moth-to-fire optimization algorithm. The global collaborative optimization module is used to optimize the global objective function based on the initial population and multiple constraints by combining the improved moth-to-fire optimization algorithm with the alternating direction multiplier algorithm, and to apply the Levy flight to update the power generation parameters, and output the optimized daily power generation and reservoir capacity plan. The rolling optimization module is used to construct a model predictive control objective function based on the optimized daily power generation and reservoir capacity plan, continuously optimize the hourly power generation and updated reservoir capacity, and output the optimal power generation command and updated reservoir capacity status for the next hour. The minute-level real-time correction module is used to solve the minute-level power generation instruction based on the optimal power generation instruction for the next hour, the updated reservoir capacity status, real-time inflow, and real-time electricity price, through a real-time optimization algorithm. The state feedback and population update module is used to take the real-time reservoir capacity obtained after executing the minute-level power generation command as the initial state of the model predictive control at the next moment, and to synchronously update the population of the improved moth-to-fire optimization algorithm. The multi-scale scheduling evaluation module is used to obtain a scheduling result report based on the optimized daily power generation and reservoir capacity plan, the optimal power generation instruction for the next hour and the updated reservoir capacity status, and the operating data generated after the execution of the minute-level power generation instruction.

9. An electronic device, characterized in that, include: Memory, used to store programs; A processor for loading the program to perform the steps of the method as claimed in any one of claims 1-7.

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

Images