Cascade hydropower station optimization scheduling model construction and calculation method

Abstract

The invention discloses a cascade hydropower station optimal scheduling model construction and calculation method, and relates to the technical field of cascade hydropower station optimal scheduling model construction, and the method comprises the following steps: setting an initial scheduling line, and setting an initial scheduling line satisfying a constraint condition for each hydropower station according to analysis and judgment; through the dynamic optimization scheduling method, the cascade hydropower stations can optimize water resource distribution according to hydrological changes, power grid requirements and reservoir regulation capacity, the water energy utilization rate is increased, and abandoned water is reduced. Under the support of power grid load prediction, the output of a hydropower station is adjusted, the peak regulation capability is enhanced, and the stability of a power system is improved. A successive approximation dynamic programming algorithm is adopted, the calculation complexity of multi-reservoir joint optimization is reduced, the scheduling efficiency is improved, global scheduling is optimized in combination with an artificial intelligence algorithm, and the economical efficiency and reliability of cascade hydropower station operation are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of the construction of an optimal dispatching model for cascade hydropower stations, and particularly relates to a method for constructing and solving an optimal dispatching model for cascade hydropower stations. Background Art

[0002] The construction and solution of the optimal dispatching model for cascade hydropower stations refer to coordinating and optimizing among multiple hydropower stations in a basin to improve the utilization efficiency of water resources, reduce the water abandonment rate, enhance the stability of the power system, and improve the overall economic benefits. This model mainly formulates a reasonable dispatching plan by establishing a mathematical optimization model for the joint operation of hydropower stations and comprehensively considering factors such as the hydrological characteristics of the basin, the grid load demand, the reservoir regulation capacity, and the power generation cost. Specifically, this model usually includes an objective function (such as maximizing power generation, minimizing water abandonment, or optimizing grid adaptability), constraint conditions (such as water balance, unit output limit, etc.), and an optimization solution algorithm (such as dynamic programming, linear programming, artificial intelligence optimization, etc.). Through calculation and simulation, an optimal dispatching plan can be obtained to achieve the efficient development and reasonable allocation of water energy under different seasons and power demand conditions, thereby improving the overall benefits of the cascade hydropower station group.

[0003] The existing technology has the following deficiencies: In the process of constructing and solving the optimal dispatching model for cascade hydropower stations, an uncommon but potentially serious problem is the deviation of the global optimal solution caused by the dynamic mutual feedback effect between reservoirs. The DPSA (Successive Approximation Dynamic Programming Algorithm) is a dimensionality reduction method that can decompose a multi-dimensional problem into multiple one-dimensional problems for solution to reduce the computational complexity. However, the core assumption of this method is to optimize only a single reservoir each time and tentatively keep the operating states of other reservoirs unchanged. Although this strategy can gradually approach the optimal solution, due to the water compensation effect between reservoirs and the complex interaction of runoff in the basin, the updated runoff information after optimizing a single reservoir each time may accumulate errors in subsequent iterations, resulting in the global optimization of the overall dispatching strategy deviating from the true optimal solution. For example, if an upstream reservoir is determined to have a relatively conservative water release strategy in the early optimization stage, a downstream reservoir may be restricted by the flow rate during subsequent optimization and cannot fully utilize the incoming water, resulting in increased water abandonment or insufficient power supply. This dynamic mutual feedback effect is particularly obvious in the joint optimization of multiple reservoirs, which may lead to a reduction in the utilization efficiency of water resources in the entire basin, affect the safety of grid dispatching, and in extreme cases, cause local power supply shortages or flood dispatching errors. Therefore, how to enhance the dynamic information coupling between reservoirs and improve the global convergence based on the DPSA algorithm is a technical problem that urgently needs to be solved.

[0004] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure, and thus it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0005] The object of the present invention is to provide a method for constructing and solving an optimal scheduling model for cascade hydropower stations to solve the problems in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: A method for constructing and solving an optimal scheduling model for cascade hydropower stations, including the following steps:

[0007] Set the initial scheduling line. According to the analysis and judgment, an initial scheduling line that meets the constraint conditions is set for each hydropower station, specifically as follows:

[0008]

[0009] , where is the optimal scheduling value of hydropower station i in the jth optimization process at the current iteration d, d is the number of iterations, and n is the number of hydropower stations;

[0010] Fix the scheduling lines of n - 1 hydropower stations, and optimize the scheduling of the remaining one hydropower station. Optimize the output of the remaining reservoir according to the single - reservoir dynamic programming method to maximize the overall power generation benefit of the entire cascade hydropower station group. The optimized scheduling line is expressed as follows:

[0011]

[0012] Optimize the scheduling lines of all hydropower stations in turn. Then fix the scheduling lines of the remaining n - 1 hydropower stations and continue to optimize the scheduling strategy of the remaining hydropower stations to maximize the power generation benefit of the entire cascade hydropower station group. Repeat this process until all hydropower stations are optimized, and record the total benefits obtained from the calculations before and after. The formula is as follows;

[0013] E(d), E(d + 1)

[0014] , where E(d) is the overall benefit after the dth round of calculation, and E(d + 1) is the overall benefit after the next round of iterative optimization;

[0015] Convergence judgment. If the following formula condition is met: |E(d + 1)-E(d)|<∈, then terminate the calculation and continue iterative optimization until the optimal scheduling strategy is reached. Here, ∈ is the set iterative accuracy.

[0016] Preferably, based on the principle of reservoir water balance within the scheduling period, a mathematical optimization model is used for solution, with the optimization objective of maximizing the total power generation of the cascade hydropower stations. The formula is as follows:

[0017]

[0018] , where E is the maximum total power generation of the cascade hydropower stations, A iis the output coefficient of the i-th power station in a given time period, Q i，t is the outflow of the i-th power station in time period t, H i，t is the head of the i-th power station in time period t, M t is the total number of calculated hours in time period t, T is the total number of calculated time periods in a year, and N is the number of cascade hydropower stations.

[0019] Preferably, to satisfy the water balance constraint, the formula is as follows:

[0020]

[0021] , where V i，t+1 and V i，t are the reservoir volumes of the i-th power station at time periods t + 1 and t respectively, Q 入i,t is the inflow of the i-th power station in time period t, S i，t is the evaporation loss of the i-th power station in time period t, and Δt is the time step of the calculation time period.

[0022] Preferably, to satisfy the flow balance constraint, the formula is as follows:

[0023]

[0024] , where Q 入i+1,t is the inflow of the (i + 1)-th power station in time period t, q i，t ′ is the inter-basin inflow between the (i + 1)-th power station and the i-th power station in the t-th time period.

[0025] Preferably, to satisfy the reservoir water storage constraint, the formula is as follows:

[0026]

[0027] , where and are the minimum and maximum water storages of the i-th power station in time period t respectively.

[0028] Preferably, to satisfy the reservoir discharge flow constraint, the formula is as follows:

[0029]

[0030] , where and are the allowable minimum and maximum discharge flows of the i-th power station in time period t respectively, Q 出i，t is the discharge flow of the i-th power station in time period t.

[0031] Preferably, to satisfy the power station output constraint, the formula is as follows:

[0032]

[0033] , where and are respectively the minimum and maximum allowable unit outputs of the i-th power station during period t.

[0034] Preferably, the peak shaving optimization scheduling mode is adopted, and the optimization objectives are adjusted as follows:

[0035]

[0036] , where E′ is the maximized daily power generation of cascade hydropower stations, and N i，t is the output of the i-th reservoir during period t.

[0037] All variables are non-negative variables during the optimization calculation and satisfy the following formula: X≥0, where X represents all optimization variables.

[0038] In the above technical solution, the technical effects and advantages provided by the present invention are as follows:

[0039] Through the dynamic optimization scheduling method, the present invention enables cascade hydropower stations to accurately allocate water resources according to hydrological changes, grid demands, and reservoir regulation capabilities, thereby greatly improving the utilization efficiency of water energy. Traditional hydropower scheduling methods often adopt static rules or empirical rules, which may not be able to flexibly adapt to water resource changes during extreme weather, dry seasons, or flood seasons, resulting in waste of water energy or insufficient power supply. The present invention adopts dynamic programming (DP) and successive approximation dynamic programming algorithm (DPSA), which can coordinate and optimize the flow scheduling among different hydropower stations, dynamically adjust the output plan according to the current reservoir water storage and incoming water volume, and maximize the utilization rate of water energy. In addition, during the flood season, by optimizing the joint scheduling of the reservoir group, the water rejection rate can be reduced, and resource waste caused by reservoir overflow can be avoided; during the dry season, the water release plan of the reservoir can be reasonably arranged to ensure the continuity of power supply and improve the overall water energy utilization efficiency of the basin.

[0040] Combined with power grid load forecasting technology, the present invention can accurately predict the changes in power demand within the future scheduling period, and adjust the output plan of the hydropower station accordingly, improving the peak shaving capacity and operation stability of the power system. Since new energy sources (such as wind power and photovoltaic power) have large fluctuations, the power grid requires reliable peak shaving power sources to balance load fluctuations. The hydropower station has the characteristics of fast start-up and shutdown and is one of the ideal peak shaving power sources. However, the traditional hydropower dispatching method often lacks a real-time matching mechanism with the power grid load, which may lead to insufficient power supply during high load periods and serious water abandonment of the hydropower station during low load periods. The present invention uses a short-term load forecasting model and an intelligent dispatching algorithm to store water during low load periods and increase power generation during high load periods, ensuring the full play of the peak shaving capacity of the hydropower station. In addition, it can monitor the changes in power load in real time and dynamically adjust the reservoir dispatching plan, making the operation of the hydropower station more intelligent, ensuring the stability of the entire power system, and improving the safety and economy of power grid operation.

[0041] The present invention adopts the successive approximation dynamic programming algorithm (DPSA), which effectively reduces the computational complexity of the joint optimal dispatching of multiple reservoirs, making the optimal solution of large-scale cascade hydropower stations more efficient and feasible. In traditional global optimization methods, due to the large number of reservoirs and complex hydrological conditions, the computational amount is extremely large, resulting in a long time-consuming process for solving the dispatching problem and making it difficult to meet the real-time dispatching requirements of actual projects. The present invention decomposes the optimization task, optimizes only one reservoir in each iteration, and gradually adjusts the dispatching strategy of the entire system according to the optimization results, thus gradually approaching the optimal solution. This step-by-step optimization strategy not only reduces the calculation time but also improves the flexibility of the dispatching plan, enabling it to adapt to different operating environments. In addition, this method can be combined with artificial intelligence algorithms (such as genetic algorithms and particle swarm optimization) to improve the global optimization ability of the dispatching plan, avoid local optimal problems, and ensure the overall operating efficiency and economic benefits of cascade hydropower stations. Brief Description of the Drawings

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0043] Figure 1 It is a method flow chart of the method for constructing and solving the optimal dispatching model of cascade hydropower stations of the present invention. Detailed Embodiments

[0044] Example embodiments will now be described more fully with reference to the accompanying drawings. However, the example embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of the example embodiments to those skilled in the art.

[0045] The present invention provides a method for constructing and solving an optimized scheduling model of a cascade hydropower station, setting an initial scheduling line, and according to the analysis and judgment, setting an initial scheduling line that meets the constraint conditions for each hydropower station, specifically as follows: Figure 1 As shown, for the optimal scheduling value of hydropower station i in the jth optimization process at the current iteration d, where d is the number of iterations and n is the number of hydropower stations;

[0046]

[0047]

[0048] , where is the optimal scheduling value of hydropower station i in the jth optimization process at the current iteration d, d is the number of iterations, and n is the number of hydropower stations;

[0049] Fix the scheduling lines of n - 1 hydropower stations, optimize the scheduling of the remaining one hydropower station, and optimize the output of the remaining reservoir according to the single - reservoir dynamic programming method to maximize the overall power generation benefit of the entire cascade hydropower station group. The optimized scheduling line is expressed as follows:

[0050]

[0051] Successively optimize the scheduling lines of all hydropower stations in turn, then fix the scheduling lines of the remaining n - 1 hydropower stations, and continue to optimize the scheduling strategies of the remaining hydropower stations to maximize the power generation benefit of the overall cascade hydropower station group. Repeat this process until all hydropower stations are optimized, and record the total benefits obtained from the calculations before and after. The formula is as follows;

[0052] E(d), E(d + 1)

[0053] , where E(d) is the overall benefit after the dth round of calculation, and E(d + 1) is the overall benefit after the next round of iterative optimization;

[0054] Convergence judgment: If the following formula condition is met: |E(d + 1)-E(d)|<∈, then terminate the calculation and continue with iterative optimization until the optimal scheduling strategy is reached. Here, ∈ is the set iterative accuracy.

[0055] Based on the principle of reservoir water volume balance within the scheduling period, solve using a mathematical optimization model with the goal of maximizing the total power generation of the cascade hydropower station. The formula is as follows:

[0056]

[0057] , where E is the total maximum power generation of cascade hydropower stations, and A i is the output coefficient of the i-th power station in a given time period, and Q i，t is the outflow of the i-th power station in time period t, and H i，t is the water head of the i-th power station in time period t, and M t is the total number of calculated hours in time period t, T is the total number of calculated time periods in a year, and N is the number of cascade hydropower stations.

[0058] The water volume balance constraint is satisfied, and the formula is as follows:

[0059]

[0060] , where V i，t+1 and V i，t are the reservoir volume of the i-th power station at time t + 1 and time t respectively, and Q 入i,t is the inflow of the i-th power station in time period t, and S i，t is the evaporation loss of the i-th power station in time period t, and Δt is the time step of the calculation time period.

[0061] The flow balance constraint is satisfied, and the formula is as follows:

[0062]

[0063] , where Q 入i+1,t is the inflow of the (i + 1)-th power station in time period t, and q i，t ′ is the inter-basin inflow between the (i + 1)-th power station and the i-th power station in the t-th time period.

[0064] The reservoir water volume constraint is satisfied, and the formula is as follows:

[0065]

[0066] , where and are the minimum and maximum water storage of the i-th power station in time period t respectively.

[0067] The reservoir discharge constraint is satisfied, and the formula is as follows:

[0068]

[0069] , where and are the allowable minimum and maximum discharges of the i-th power station in time period t respectively, and Q 出i，t is the discharge of the i-th power station in time period t.

[0070] The power station output constraint is satisfied, and the formula is as follows:

[0071]

[0072] , where and are respectively the minimum and maximum allowable unit outputs of the i-th power station during period t.

[0073] Adopt the peak shaving optimization scheduling mode, and adjust the optimization objectives as follows:

[0074]

[0075] , where E′ is the maximized daily power generation of cascade hydropower stations, and N i，t is the output of the i-th reservoir during period t,

[0076] All variables are non-negative variables during the optimization calculation, and satisfy the following formula: X ≥ 0, where X represents all optimization variables, including output, flow rate, reservoir water storage, etc.

[0077] Embodiment 1: The core objective of the cascade hydropower station optimization scheduling is to maximize the water energy utilization rate of the entire basin by reasonably arranging the outputs of each reservoir, while ensuring the coordinated operation between reservoirs, reducing water abandonment, and improving the stability of power grid scheduling. In practical applications, the operation scheduling of cascade reservoirs is affected by various factors, such as incoming water conditions, power grid demands, environmental protection requirements, basin comprehensive management, etc. Therefore, it is particularly important to formulate a scientific, reasonable, and feasible scheduling plan. This embodiment adopts the dynamic programming (DP) algorithm for cascade hydropower station optimization scheduling with the goal of maximizing power generation. This method can find the optimal scheduling plan that meets the constraint conditions through recursive calculation considering the mutual influence of multiple reservoirs.

[0078] In the specific implementation process, it is first necessary to construct a scheduling optimization model for the cascade reservoir group. This model includes multiple constraint conditions, such as water balance constraints, output constraints, reservoir capacity limits, etc., and the objective function is set to achieve the maximum power generation benefit of the hydropower station group during the entire scheduling period. After the model is established, the dynamic programming method of bottom-up recursion is used for optimization calculation. Specifically, first fix the output of the downstream hydropower station, and then gradually calculate upstream according to the water regulation characteristics of the upstream reservoir, so that the outflow of the upstream reservoir matches the power grid demand and achieves the optimal power generation benefit throughout the cycle.

[0079] During the solution process, the dynamic programming algorithm uses a recursive approach to decompose complex multi-stage optimization problems into a series of sub-problems and gradually solve these sub-problems to finally obtain the global optimal solution. During the solution process of each sub-problem, the algorithm comprehensively considers the water volume changes in the current reservoir, power generation requirements, and the interactions between reservoirs to ensure the optimal scheduling plan for the entire cascade reservoir group. In each calculation step, the algorithm analyzes the current state of each reservoir (such as water level, output situation, etc.), and then based on historical data and prediction information, calculates the possible results under different scheduling strategies and selects the optimal strategy for execution.

[0080] To adapt to different scheduling cycles, the dynamic programming method can adjust the time step according to requirements. For example, for daily scheduling, an hourly time step is usually adopted to more finely control the output changes of hydropower stations; while for monthly or annual scheduling, a longer time step can be adopted for medium- and long-term optimization. In this way, it can be ensured that the scheduling plan can not only meet the short-term grid demand but also take into account the sustainability of long-term water resource management.

[0081] When applying this method, external factors such as meteorological conditions, inflow prediction, and grid load demand also need to be considered to ensure the practical feasibility of the scheduling plan. Since factors such as precipitation, temperature changes, and glacial meltwater affect the inflow of reservoirs, meteorological prediction models such as machine learning-based precipitation prediction and satellite remote sensing technology can be combined to improve the accuracy of the scheduling plan. For example, during the dry season, the system can predict the water resource shortage in advance and adjust the operation strategy of the reservoir to ensure the continuity of power supply; while during the wet season, the flood risk can be reduced and the water energy utilization rate can be improved by optimizing the scheduling strategy.

[0082] In addition, in actual projects, due to the limited regulation capacity of reservoirs, situations of water volume supply-demand mismatch may occur. For example, when the power demand is low, the reservoir may need to reduce the outflow to avoid wasting water resources; while during high-load periods, the reservoir needs to increase the output to meet the electricity demand of the grid. In this case, optimization scheduling software or artificial intelligence methods can be combined to further improve the solution efficiency and accuracy of the algorithm. For example, in the practice of hydropower station scheduling, the deep reinforcement learning (DRL) method can be used to enable the intelligent system to improve the prediction ability of future water conditions, grid demand and other factors based on continuous learning of historical data, so as to better optimize the scheduling decision.

[0083] The advantage of this method is that it can find a better solution in a complex cascade reservoir system, and the calculation speed is relatively fast, which is suitable for the optimal scheduling of large-scale hydropower systems. The characteristic of solving problems in stages of the dynamic programming algorithm makes it particularly suitable for solving optimization problems with strong time correlation such as cascade hydropower stations. Compared with the traditional linear programming method, dynamic programming can better capture the interaction between reservoirs, thus improving the overall optimization effect.

[0084] However, a major drawback of this method is that it requires a large amount of storage space during the calculation process. Since the dynamic programming algorithm needs to store a large number of intermediate results during the calculation, when the number of reservoirs is large, the computational complexity will increase rapidly. Especially when the scheduling period is long and the time step is small, the amount of calculation will further increase, which may lead to too long calculation time and affect practical applications. Therefore, in order to improve the calculation efficiency, segmented optimization or parallel computing technology can be used to reduce the calculation time. For example, using GPU parallel computing can perform calculations simultaneously on multiple computing cores, thus improving the execution efficiency of the algorithm. In addition, a hierarchical optimization strategy can also be adopted, that is, first conduct a rough optimization for a longer time scale, and then conduct a fine optimization for a short time scale to reduce the amount of calculation.

[0085] In practical applications, this method can be used for the scheduling of cascade hydropower stations of different scales. For example, for a hydropower system containing multiple large reservoirs, dynamic programming can help optimize the long-term scheduling strategy to ensure the coordinated operation between reservoirs; for small cascade power stations in a single basin, this method can also be used to optimize the short-term scheduling to improve the adaptability of the power grid. In addition, this method can also be combined with other optimization methods (such as intelligent optimization methods like genetic algorithms and particle swarm optimization) to further improve the calculation efficiency and scheduling quality. For example, in actual operation, the genetic algorithm can be used for preliminary optimization first to obtain a set of better initial scheduling schemes, and then the dynamic programming method can be used to refine and optimize them to obtain more accurate scheduling results.

[0086] To sum up, this embodiment uses the dynamic programming algorithm for the optimal scheduling of cascade hydropower stations, which can effectively improve the utilization rate of water energy, reduce water abandonment, and improve the stability of power grid scheduling. By reasonably constructing a scheduling optimization model, combining a recursive solution strategy, and combining external prediction factors for optimization, it can ensure the stable operation of hydropower stations in a complex environment. In addition, to solve the problem of high computational complexity, technologies such as parallel computing and hierarchical optimization can be adopted to further improve the calculation efficiency and applicability of the algorithm. In the future, this method can also be combined with advanced technologies such as artificial intelligence and big data analysis to further improve the intelligent level of scheduling optimization, making the operation of cascade hydropower stations more intelligent and efficient.

[0087] Embodiment 2: In the process of optimizing the operation of cascade hydropower stations, due to the complex interaction between reservoirs, directly adopting a global optimization method often involves a large amount of calculation and is difficult to meet the actual operation requirements. Especially when the number of reservoirs is large and the scheduling period is long, the traditional global optimization method will lead to an exponential increase in computational complexity. Therefore, in order to reduce the computational complexity, improve the computational efficiency, and at the same time ensure the rationality and feasibility of the cascade hydropower station scheduling scheme, this embodiment adopts the successive approximation dynamic programming algorithm (DPSA) for optimal scheduling.

[0088] The core idea of the DPSA method is to decompose a high-dimensional optimization problem into multiple one-dimensional problems for solution. That is, in each round of iteration, the scheduling lines of some reservoirs are fixed, and only one reservoir is optimized to make its scheduling strategy improve the power generation benefit of the entire cascade reservoir as much as possible. Then, in the next round of iteration, the scheduling line of the next reservoir is optimized, and the state of the entire system is updated according to the new calculation results. Such continuous cyclic iteration is carried out until the overall optimization goal converges, that is, when the change amount of the optimization results of two adjacent iterations is less than the set threshold, the calculation is stopped and the final scheduling scheme is output.

[0089] In practical applications, the first step of the DPSA method is to initialize the scheduling lines of each reservoir. This process is usually based on experience or historical data to set an initial operation plan for each reservoir to meet basic operation constraints such as water balance, minimum ecological flow, and unit output limit. The selection of the initialization scheme is crucial. A reasonable initial scheduling line can accelerate the algorithm convergence speed, reduce the number of iterations, and improve the computational efficiency. Common initialization methods include:

[0090] Initialization based on historical scheduling data: Adopt a relatively optimal scheduling scheme in past operation records as the initial scheduling line.

[0091] Initialization based on linear interpolation: Through the fitting of the reservoir scheduling curve, an approximately reasonable initial scheme is obtained.

[0092] Initialization based on randomized search: Generate multiple initial scheduling lines and select the one with the highest constraint satisfaction degree as the initial solution.

[0093] After initialization, enter the successive optimization stage. In each round of optimization, DPSA separately optimizes the scheduling strategy of a certain reservoir, adopts dynamic programming or other adaptive optimization algorithms to adjust the outflow of the reservoir, so as to improve the power generation benefit of the entire cascade reservoir group. After optimization, update the operation state of the reservoir, and optimize the scheduling scheme of the next reservoir based on the new state information, and so on, until the scheduling goal of the entire cascade reservoir system tends to be stable.

[0094] Advantages of the DPSA method

[0095] High computational efficiency: Compared with traditional global optimization methods, DPSA avoids the computational bottleneck of directly solving high-dimensional problems and reduces the amount of computation through step-by-step optimization, making it more suitable for the optimal scheduling of large-scale hydropower systems.

[0096] Strong flexibility: Since DPSA optimizes step by step rather than solving all at once, it can be dynamically adjusted based on actual operating conditions to improve the adaptability of the dispatching scheme. For example, when the water inflow to the reservoir changes significantly, local optimization can be performed in some key reservoirs without recalculating the dispatching scheme for the entire cascade reservoir.

[0097] Applicable to complex constraints: In the process of dispatching cascade hydropower stations, many constraints need to be considered, such as ecological flow constraints, unit minimum output constraints, reservoir maximum and minimum water level constraints, grid load demand constraints, etc. DPSA can gradually adjust the dispatching plan through step-by-step optimization to meet all constraints without falling into a situation where the solution cannot be solved.

[0098] Applicable to large-scale cascade hydropower systems: For complex dispatching systems with multiple hydropower stations in a basin, DPSA's step-by-step optimization strategy can effectively solve the problem of mutual influence between reservoirs and improve the overall water energy utilization rate.

[0099] Limitations and improvement strategies of DPSA method

[0100] Although the DPSA method has many advantages, it still has some problems in practical application, mainly including:

[0101] Possible trapping in local optimal solution: Since DPSA only adjusts the dispatching plan of one reservoir in each round of optimization, the final dispatching result may fail to reach the true global optimal solution. For example, an upstream reservoir may be set to a conservative discharge strategy in the first few rounds of optimization, while the downstream reservoir may not be able to generate electricity fully due to insufficient water in the subsequent optimization, resulting in a decrease in the overall power generation efficiency.

[0102] The convergence speed depends on the initial scheduling line: if the initial scheduling line is not set reasonably, DPSA may require more iterations to converge, or even fall into an infinite loop, resulting in reduced computational efficiency.

[0103] Not suitable for real-time scheduling: Since DPSA requires multiple rounds of iterations to obtain the final scheduling solution, its calculation time is still longer than that of simple heuristic algorithms. Therefore, in real-time scheduling applications, it may be necessary to combine it with fast optimization methods to increase the calculation speed.

[0104] In order to overcome these problems, in practical applications, intelligent optimization algorithms (such as genetic algorithms, particle swarm optimization, etc.) can be combined to improve the global optimization capability of DPSA. For example:

[0105] Genetic algorithm + DPSA: First, use the genetic algorithm to generate multiple possible scheduling plans, and select a better group as the initial scheduling line of DPSA. Then, use DPSA for fine optimization to improve the global search ability of the algorithm and avoid falling into local optima.

[0106] Particle swarm optimization + DPSA: Particle swarm optimization (PSO) can find better solutions within the search space. By combining it with DPSA, a balance can be achieved between global search and local optimization, and the convergence speed of the algorithm can be improved.

[0107] In addition, DPSA can also be combined with big data analysis and machine learning technologies. Use historical data to train deep learning models (such as LSTM) to predict future incoming water conditions, and adjust the optimization strategy of the reservoir according to the prediction results. For example, use machine learning to predict the water volume change in the next 7 days and combine it with DPSA for optimization, which can formulate a reasonable scheduling plan in advance and improve the adaptability of the scheduling plan.

[0108] Application of DPSA in the actual hydropower station scheduling

[0109] The DPSA method has been applied in the scheduling systems of multiple cascade hydropower stations. For example:

[0110] In the optimal scheduling of cascade reservoirs in the Yangtze River Basin, DPSA is used to optimize the joint scheduling of the Three Gorges, Gezhouba and downstream reservoirs to improve power generation efficiency and reduce the water rejection volume during the flood period.

[0111] In the hydropower station scheduling in the Yarlung Zangbo River Basin, due to the significant seasonal variation of water volume, DPSA combines with the meteorological prediction system to optimize the reservoir scheduling strategy in different seasons, making the power output of hydropower stations more stable during the dry season and reducing the water rejection loss.

[0112] This embodiment uses the successive approximation dynamic programming algorithm (DPSA) for the optimal scheduling of cascade hydropower stations. By gradually optimizing the scheduling line of the reservoir, the computational complexity is reduced and the computational efficiency is improved. This method has obvious advantages in the optimal scheduling of large-scale cascade hydropower systems, can effectively improve the water energy utilization rate, reduce water rejection, and improve the stability of power grid scheduling. However, DPSA may have the problem of local optima. Therefore, technologies such as genetic algorithm, particle swarm optimization, and big data analysis can be combined to further optimize the scheduling plan to ensure the global optimality of the scheduling results. In the future, this method can also be combined with artificial intelligence and cloud computing platforms to improve the computing speed and intelligent level, making the scheduling of cascade hydropower stations more accurate and efficient.

[0113] Embodiment 3: The optimal dispatching of a hydropower station not only needs to consider the rational utilization of water resources, but also needs to match the operation requirements of the power grid to ensure the stability of the power grid. Especially in the case of large fluctuations in the power grid load, as a clean energy source with strong regulation ability, the hydropower station undertakes the important responsibilities of peak shaving, valley filling, and stable power supply for the power grid. Therefore, in this embodiment, by combining power grid load forecasting with the optimal dispatching of the hydropower station, through accurate prediction of power demand, a reasonable dispatching plan is formulated to improve the peak shaving ability and power generation efficiency of the hydropower station, ensure the stable operation of the power grid, and avoid the imbalance between supply and demand.

[0114] The implementation steps of this method include three parts: power demand forecasting, formulation of an optimal dispatching plan, and real-time adjustment. During the optimization process, not only the water level regulation ability of the reservoir, the minimum ecological flow constraint, and the generator output constraint need to be considered, but also factors such as power grid demand, incoming water prediction, and extreme weather need to be comprehensively evaluated to ensure the scientificity and feasibility of the dispatching plan.

[0115] 1. Power demand forecasting

[0116] The first step in optimizing the dispatching of a hydropower station is to accurately predict the power grid load demand in order to reasonably arrange the output of the hydropower station. In practical applications, the power grid load is affected by multiple factors, such as meteorological conditions (temperature, humidity, wind speed), social and economic activities, and the electricity consumption habits of residents and industries. To improve the accuracy of load forecasting, the following methods are usually adopted:

[0117] Time series analysis: Using historical power load data, methods such as autoregressive (AR), moving average (MA), and ARIMA models are used for short-term or medium- to long-term forecasting.

[0118] Deep learning models: Such as long short-term memory networks (LSTM), which can improve the prediction ability of short-term fluctuations in power load through large-scale data training, and are especially suitable for processing load data with strong periodicity and high temporal correlation.

[0119] Regression analysis and machine learning methods: Such as random forest (RF), XGBoost, neural network (ANN), etc., by integrating multiple data sources (weather, holidays, industrial production conditions, etc.), a more comprehensive load forecasting model is established.

[0120] After obtaining the power grid load forecasting results, the dispatching system can formulate power generation plans for different time periods in advance. For example, during the high-temperature period in summer, due to the increase in air-conditioning load and high power demand, the hydropower station needs to make power generation plans in advance to cope with the peak load; while during the late-night low-load period, the hydropower station can reduce its output and store water to provide greater regulation ability for subsequent high-load periods.

[0121] 2. Formulation of an optimal dispatching plan

[0122] Based on the predicted load, the hydropower station scheduling system formulates an optimal power generation plan according to the grid demand to ensure that the output of the hydropower station matches the grid load curve as much as possible and reduces the problem of unstable power supply. Usually, the hydropower station scheduling adopts three modes: intraday scheduling, short-term scheduling, and medium- and long-term scheduling:

[0123] Intraday scheduling: In response to the load fluctuations within 24 hours, the output of the hydropower station is reasonably arranged so that the hydropower station generates electricity maximally during high-load periods and stores water during low-load periods.

[0124] Short-term scheduling (weekly scheduling, monthly scheduling): Combining information such as weather forecasts and industrial production arrangements, an optimized scheduling plan for 7 days or 30 days is formulated.

[0125] Medium- and long-term scheduling (seasonal scheduling): According to the basin hydrological conditions, the scheduling strategies for the wet season and the dry season are reasonably arranged to ensure the balance of annual water resource utilization.

[0126] When formulating the scheduling plan, the following key factors need to be comprehensively considered:

[0127] Reservoir regulation capacity: Different reservoirs have different regulation capacities. For example, large reservoirs have strong seasonal regulation capacities, while small reservoirs mainly conduct daily regulation. Therefore, during the optimization process, it is necessary to reasonably allocate the output of different reservoirs to operate them complementarily.

[0128] Grid peak shaving demand: During peak load periods, the hydropower station needs to provide sufficient power support; while during low load periods, the hydropower station can reduce power generation or even conduct pumped storage to balance the grid supply and demand.

[0129] Ecological flow constraint: During the operation of cascade hydropower stations, a certain amount of downstream discharge must be guaranteed to maintain the ecological balance of the basin. Therefore, the scheduling optimization plan needs to strictly comply with the ecological flow limit.

[0130] Unit output limit: The output of the hydropower station unit is restricted by factors such as the minimum stable output and the maximum installed capacity. The scheduling system needs to ensure that the unit operates within the economic output range.

[0131] 3. Real-time adjustment and emergency response

[0132] The actual change of the grid load may deviate from the predicted value. Therefore, during the operation of the hydropower station, the scheduling system needs to monitor the change of the grid load in real time and dynamically adjust the scheduling plan of the hydropower station according to the actual load demand to ensure the supply-demand balance. Usually, the real-time adjustment is mainly reflected in the following aspects:

[0133] Response to short-term load fluctuations: During peak electricity consumption periods or sudden load fluctuations, hydropower stations can quickly increase their output to provide short-term power supplementation and reduce the risk of grid frequency fluctuations.

[0134] Response to extreme weather and emergencies: When extreme weather (such as typhoons, heavy rains, cold snaps, etc.) or sudden accidents (such as unit failures, grid failures) occur, the dispatching system needs to quickly adjust the operation strategy of the hydropower station. For example, before heavy rain arrives, the hydropower station can pre-discharge water in advance to reduce the flood risk, and in the case of a sudden power outage, it can quickly increase the output of the hydropower station to ensure the stability of power supply.

[0135] Dispatch of standby power sources: In some cases, the grid may need to quickly start standby power sources to fill the power supply gap. Hydropower stations can serve as flexible standby power sources, starting and stopping quickly within a short time to improve the flexibility of grid dispatching.

[0136] To improve the response speed of real-time adjustment, an automated dispatching system can be introduced, adopting cloud computing, big data analysis, and artificial intelligence (AI) technologies to establish an intelligent hydropower station dispatching platform. This platform can achieve:

[0137] Real-time data acquisition: Through the SCADA system, real-time monitoring of data such as grid load, reservoir water level, flow rate, and unit operation status is carried out.

[0138] Intelligent decision support: Based on big data and AI algorithms, quickly analyze the optimal dispatching plan and automatically adjust the operation parameters of the hydropower station.

[0139] Remote control and coordinated dispatching: Realize remote intelligent coordinated dispatching of multiple hydropower stations to improve the overall operation efficiency.

[0140] The advantage of this method lies in its ability to accurately match the grid load demand, reduce waste of hydropower resources, and at the same time improve the economic benefits of hydropower stations and the security of the grid. Compared with traditional hydropower dispatching methods, this method combines grid load forecasting and hydropower station dispatching optimization, can better adapt to grid supply and demand fluctuations, improve the operation flexibility of hydropower stations and the grid peak shaving capacity. In addition, this method can also effectively respond to extreme weather, sudden accidents and other situations, and improve the emergency response ability of hydropower stations.

[0141] In the future, this method can also be combined with an artificial intelligence dispatching system, using deep learning, big data analysis, and cloud computing to further optimize the dispatching strategy of hydropower stations. For example, using an AI prediction model to automatically adjust dispatching parameters, combining with an intelligent sensor network to achieve real-time monitoring and control, and further improving the dispatching accuracy and system stability. With the increase in the penetration rate of new energy (wind power, photovoltaic), hydropower stations can also operate in coordination with new energy to improve the utilization rate of renewable energy in the grid and achieve more accurate and efficient optimized operation of cascade hydropower stations.

[0142] Through the dynamic optimization scheduling method, the cascade hydropower station of the present invention can perform precise water resource allocation according to hydrological changes, grid demand, and reservoir regulation capacity, thereby greatly improving the utilization efficiency of water energy. Traditional hydropower scheduling methods often adopt static rules or empirical rules, which may not be able to flexibly adapt to water resource changes during extreme weather, dry seasons, or wet seasons, resulting in wasted water energy or insufficient power supply. The present invention adopts dynamic programming (DP) and successive approximation dynamic programming algorithm (DPSA), which can coordinate and optimize the flow scheduling among different hydropower stations, dynamically adjust the output plan according to the current reservoir water storage and incoming water volume, and maximize the utilization rate of water energy. In addition, during the wet season, by optimizing the joint scheduling of the reservoir group, the water rejection rate can be reduced, and resource waste caused by reservoir overflow can be avoided; during the dry season, the water release plan of the reservoir can be reasonably arranged to ensure the continuity of power supply and improve the overall water energy utilization efficiency of the basin.

[0143] Combined with grid load forecasting technology, the present invention can accurately predict the change of power demand in the future scheduling period and adjust the output plan of the hydropower station accordingly, improving the peak shaving capacity and operation stability of the power system. Since new energy sources (such as wind power and photovoltaic power) have large fluctuations, the grid needs reliable peak shaving power sources to balance load fluctuations, and hydropower stations have the characteristics of fast start and stop, making them one of the ideal peak shaving power sources. However, traditional hydropower scheduling methods often lack a real-time matching mechanism with the grid load, which may lead to insufficient power supply during high load periods and serious water rejection of hydropower stations during low load periods. The present invention uses a short-term load forecasting model and an intelligent scheduling algorithm to store water during low load periods and increase power generation during high load periods, ensuring that the peak shaving capacity of the hydropower station is fully exerted. In addition, by real-time monitoring the change of power load and dynamically adjusting the reservoir scheduling plan, the operation of the hydropower station becomes more intelligent, ensuring the stability of the entire power system and improving the safety and economy of grid operation.

[0144] The present invention adopts the successive approximation dynamic programming algorithm (DPSA), which effectively reduces the computational complexity of the joint optimal operation of multi-reservoirs, making the optimal solution of large-scale cascade hydropower stations more efficient and feasible. In traditional global optimization methods, due to the large number of reservoirs and complex hydrological conditions, the computational amount is extremely large, resulting in a long time-consuming for the scheduling solution process and making it difficult to meet the real-time scheduling requirements of actual projects. The present invention decomposes the optimization task, optimizes only one reservoir in each round of iteration, and gradually adjusts the scheduling strategy of the whole system according to the optimization results, so as to gradually approach the optimal solution. This step-by-step optimization strategy not only reduces the computational time, but also improves the flexibility of the scheduling scheme, enabling it to adapt to different operating environments. In addition, this method can be combined with artificial intelligence algorithms (such as genetic algorithms, particle swarm optimization) to improve the global optimization ability of the scheduling scheme and avoid local optimal problems, thus ensuring the overall operating efficiency and economic benefits of cascade hydropower stations.

[0145] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0146] Only some exemplary embodiments of the present invention have been described by way of illustration above. Undoubtedly, for those of ordinary skill in the art, without departing from the spirit and scope of the present invention, the described embodiments can be modified in various different ways. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0147] It should be noted that in this article, if there are relational terms such as first and second, they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the existence of another identical element in the process, method, article or device comprising the element.

[0148] It should be understood that in various embodiments of the present application, the magnitude of the serial numbers of the above processes does not mean the order of execution is prior or posterior, and the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0149] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in connection with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Skilled artisans may use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of this application.

[0150] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0151] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0152] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit.

[0153] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all such changes or substitutions should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0154] Only some exemplary embodiments of the present invention have been described above by way of illustration. Undoubtedly, those of ordinary skill in the art can modify the described embodiments in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the protection scope of the claims of the present invention.

Claims

1. Construction and solution method of cascade hydropower station optimization dispatching model, characterized in that: The following steps are involved: Set the initial dispatch line. Based on the analysis and judgment, set an initial dispatch line that meets the constraints for each hydropower station, as follows: ， In the formula, is the optimal dispatch value of hydropower station i in the jth optimization process at the current iteration d, where d is the number of iterations and n is the number of hydropower stations; The dispatching lines of n-1 hydropower stations are fixed, and the remaining hydropower station is optimized. The output of the remaining reservoirs is optimized according to the single reservoir dynamic programming method to maximize the overall power generation efficiency of the entire cascade hydropower station group. The optimized dispatching line is expressed as follows: Optimize the dispatching lines of all hydropower stations in turn, then fix the dispatching lines of the remaining n-1 hydropower stations, and continue to optimize the dispatching strategies of the remaining hydropower stations to maximize the power generation efficiency of the entire cascade hydropower station group. Repeat this process until all hydropower stations have completed the optimization, and record the total benefits calculated twice before and after. The formula is as follows; E(d), E(d+1), Where E(d) is the overall benefit after the dth round of calculation, and E(d+1) is the overall benefit after the next round of iterative optimization; Convergence judgment: If the following formula conditions are met: |E(d+1)-E(d)|<∈, the calculation is terminated and iterative optimization continues until the optimal scheduling strategy is reached, where ∈ is the set iteration accuracy.

2. The method for constructing and solving the optimal dispatching model of cascade hydropower stations according to claim 1 is characterized in that: Based on the principle of reservoir water balance within the dispatching period, a mathematical optimization model is used to solve the problem, with maximizing the total power generation of cascade hydropower stations as the optimization goal. The formula is as follows: ， Where E is the maximum total power generation of cascade hydropower stations, A i is the output coefficient of the ith power station in a given period, Q i，t is the outflow of the ith power station in time period t, H i，t is the water head of the ith power station in time period t, M t is the total number of calculation hours in time period t, T is the total number of calculation time periods in a year, and N is the number of cascade hydropower stations.

3. The method for constructing and solving the optimal dispatching model of cascade hydropower stations according to claim 1 is characterized in that: Satisfy the water balance constraint, the formula is as follows: ， Where V i，t+1 and V i，t are the storage capacity of the i-th power station in time period t+1 and time period t, Q 入i,t is the inflow of the ith power station in time period t, S i，t is the evaporation loss of the ith power station in period t, and Δt is the time step of the calculation period.

4. The method for constructing and solving the optimal dispatching model of cascade hydropower stations according to claim 1 is characterized in that: Satisfying the flow balance constraint, the formula is as follows: ， In the formula, Q 入i+1,t is the inflow of the i+1th power station in time period t, q i，t ′ is the interval inflow between the i+1th power station and the ith power station in the tth period.

5. The method for constructing and solving the optimal dispatching model of cascade hydropower stations according to claim 1 is characterized in that: To meet the reservoir storage capacity constraint, the formula is as follows: ， In the formula, and are the minimum and maximum water storage capacities of the ith power station in time period t, respectively.

6. The method for constructing and solving the optimal dispatching model of cascade hydropower stations according to claim 1 is characterized in that: To meet the reservoir discharge flow constraint, the formula is as follows: ， In the formula, and are the minimum and maximum discharge flows allowed by the ith power station in time period t, Q 出i，t is the discharge flow of the ith power station in time period t.

7. The method for constructing and solving the optimal dispatching model of cascade hydropower stations according to claim 1 is characterized in that: To meet the power station output constraints, the formula is as follows: ， In the formula, and are the minimum and maximum allowable unit outputs of the ith power station in time period t.

8. The method for constructing and solving the optimal dispatching model of cascade hydropower stations according to claim 1 is characterized in that: The peak load optimization scheduling mode is adopted, and the optimization target is adjusted as follows: ， Where E′ is the maximum daily power generation of the cascade hydropower station, N i，t is the output of the ith reservoir in time period t, During the optimization calculation, all variables are non-negative variables and satisfy the following formula: X ≥ 0, where X represents all optimization variables.

Images