Reservoir optimization scheduling method based on uniform design
By combining uniform design and particle swarm optimization algorithm with real-time monitoring via sensor network, reservoir scheduling schemes are generated and optimized, solving the inaccuracy and inefficiency problems of traditional reservoir scheduling methods in complex environments, and realizing the scientific and efficient scheduling of reservoirs.
Patent Information
- Application Number
- CN202511631536.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-10
- Publication Date
- 2026-02-06
AI Technical Summary
Traditional reservoir scheduling methods may produce inaccurate or inefficient scheduling results when faced with complex hydrological environments and multiple objectives.
The uniform design theory is used to generate a set of uniformly distributed scheduling schemes, and the particle swarm optimization algorithm is used for global search. Through data cleaning and standardization, combined with real-time monitoring by sensor networks and a database of historical similar flood cases to assist decision-making, the scheduling schemes are dynamically adjusted.
It has improved the scientific nature and efficiency of reservoir scheduling, ensured the maximum utilization of water resources and the balance of multiple objectives, and enhanced the adaptability to complex scheduling scenarios and flood control safety.
Smart Images
Figure CN121481097A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of water conservancy engineering technology, and more specifically, to a method for optimal reservoir scheduling based on uniform design. Background Technology
[0002] With the increasing scarcity of global water resources and the impact of climate change, the scheduling and management of reservoirs are facing growing challenges.
[0003] Traditional reservoir scheduling methods often employ a single optimization strategy, which may lead to inaccurate or inefficient scheduling results when faced with complex hydrological environments and multi-objective requirements. To overcome this problem, modern reservoir scheduling requires a more scientific and systematic approach. Uniform design, an experimental design method based on optimization theory, can efficiently identify and evaluate the impact of different combinations of variables on the objective function. This method has been widely applied in engineering, especially demonstrating excellent performance in multi-objective optimization problems. Therefore, introducing uniform design into the field of reservoir optimization scheduling can provide more accurate decision support and solve scheduling problems under multiple objectives and complex constraints.
[0004] In view of this, we propose a reservoir optimization scheduling method based on uniform design. Summary of the Invention
[0005] The purpose of this invention is to provide a reservoir optimization scheduling method based on uniform design, which aims to solve the problem that existing reservoir scheduling methods may produce inaccurate or inefficient scheduling results when facing complex hydrological environments and multiple objective requirements.
[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a reservoir optimization scheduling method based on uniform design, comprising the following steps: S1. First, collect basic data of the reservoir, including historical hydrological data, real-time hydrological data, reservoir characteristic parameters, and flood control requirements. Then, remove outliers through data cleaning and use standardization methods to convert data of different dimensions into the [0,1] interval, providing a high-quality data foundation for subsequent analysis. S2. For the water release scheduling scenario, determine the decision variables such as water release flow rate, reservoir water level, and scheduling period. Within the feasible region of each variable, generate a uniformly distributed set of scheduling schemes using uniform design theory. Specifically, this is done through the formula... Calculate the solution points, where and Corresponding to the scheme number and variable dimension respectively, and For variable boundaries, To determine the number of possible solutions, a random perturbation term is added. Enhance the diversity of solutions and form an initial solution library; S3. Perform joint model evaluation on each scheme in the initial scheme library to form a scheme evaluation matrix containing multi-objective parameters; S4. The particle swarm optimization algorithm is used to optimize the evaluated scheme matrix, and finally the scheduling scheme with the best overall benefits is output. S5. Apply the optimized scheme to the actual scheduling of the reservoir.
[0007] Preferably, in step S1 above, the historical hydrological data includes multi-year rainfall, inflow and outflow data, the real-time hydrological data includes real-time rainfall, river level and inflow data, the reservoir characteristic parameters include reservoir capacity curve and gate size, and the flood control requirements include flood control limit level and downstream safe discharge standard.
[0008] Preferably, in step S2 above, the decision variables also include the safe discharge of the downstream river channel, which is determined by hydraulic calculation of the downstream river channel cross-section, and the set of decision variables satisfies the constraint that the discharge flow does not exceed the safe discharge of the downstream river channel.
[0009] Preferably, in step S2 above, the random disturbance term The value range is [−0.1, 0.1], and the normal distribution random number generation method is adopted. When the decision variable is a discrete variable, the disturbed value is mapped to the nearest feasible discrete value through rounding rules to ensure the engineering feasibility of the scheme. The feasible value of the discrete decision variable needs to be determined in advance according to the reservoir gate regulation accuracy and the actual engineering needs.
[0010] Preferably, step S5 above also includes real-time monitoring of rainfall, water level, and flow parameters through a sensor network. When the monitored values exceed the warning threshold, a rolling optimization strategy is adopted to shorten the scheduling cycle to 1 hour. At the same time, a database of historical similar flood cases is called to assist in decision-making, thereby realizing the dynamic adjustment and optimization of the scheduling scheme.
[0011] Preferably, in step S4 above, the particle swarm optimization algorithm uses the multi-objective evaluation result as the fitness function and updates the particle velocity and position using the following formula: Speed update formula:
[0012] Position update formula:
[0013] in, For the first The particle in the first The speed of each iteration For inertial weights, , As a learning factor, , It is a random number with a value in the interval [0,1]. For the first The optimal position of each individual particle. To be the globally optimal position The optimal solution in the current iteration is denoted as , and is the influence factor. Used to enhance the ability to learn from the current best solution.
[0014] Preferably, the inertial weight Using a linear decreasing strategy, the expression is:
[0015] in, =0.9, =0.4, This represents the current iteration number. This represents the maximum number of iterations.
[0016] Preferably, when the global optimal solution has not been updated for 10 consecutive iterations, an adaptive mutation operation is triggered, with a mutation probability of 1. This increases the probability of the particle swarm optimization algorithm escaping local optima, ultimately outputting a scheduling scheme with optimal overall benefits.
[0017] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention generates a uniformly distributed set of scheduling schemes within the feasible region of decision variables using uniform design theory. It combines this with particle swarm optimization algorithm for global search and introduces adaptive mutation operation to avoid the algorithm getting trapped in local optima. This allows for the exploration of potential optimal solutions over a larger scope, thereby ensuring that the final output scheduling scheme has better overall benefits. This improves the scheduling efficiency of reservoirs in different watersheds and at different times, maximizes the utilization of water resources, and ensures a balance between ecological water use, agricultural irrigation, hydrological conditions for fish spawning, power generation, and other social water needs, thus avoiding resource waste and scheduling conflicts in reservoir scheduling.
[0018] 2. This invention cleans and removes outliers from historical and real-time hydrological data, and uses a standardization method to convert data of different dimensions into the [0,1] interval, providing a high-quality data foundation for the water discharge evolution model and reservoir scheduling model, making the model evaluation results more reliable, and thus improving the scientific nature of the scheduling scheme.
[0019] 3. This invention targets discrete decision variables and maps the disturbed values to the nearest feasible discrete values through rounding rules. The feasible values of discrete variables are determined in advance based on the reservoir gate regulation accuracy and actual engineering needs. At the same time, during emergency dispatch, it combines real-time monitoring of sensor networks and a database of historical similar flood cases to assist decision-making, making the dispatch scheme more in line with the actual operation needs of the project and improving the adaptability to complex dispatch scenarios.
[0020] 4. This invention monitors rainfall, water level, and flow parameters in real time through a sensor network. When the monitored values exceed the warning threshold, a rolling optimization strategy is adopted to shorten the scheduling cycle to 1 hour. The scheduling plan can be dynamically adjusted according to the real-time changes in hydrological conditions, enhancing the reservoir's ability to cope with sudden floods and other abnormal situations, and improving the reservoir's flood control safety and scheduling flexibility. Attached Figure Description
[0021] Figure 1 This is a flowchart illustrating the optimized scheduling method in this invention. Detailed Implementation
[0022] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.
[0023] Example 1 A reservoir optimization scheduling method based on uniform design includes the following steps: S1. First, collect basic data about the reservoir, including historical hydrological data, real-time hydrological data, reservoir characteristic parameters, and flood control requirements. Then, remove outliers through data cleaning and use standardization methods to convert data of different dimensions into the [0,1] interval. This data cleaning and standardization process improves data quality and avoids interference from outliers and differences in dimensions on model analysis, laying the foundation for the accuracy of subsequent scheduling plans. S2. For the water release scheduling scenario, determine the decision variables such as water release flow rate, reservoir water level, and scheduling period. Within the feasible region of each variable, generate a uniformly distributed set of scheduling schemes using uniform design theory. Specifically, this is done through the formula... Calculate the solution points, where and Corresponding to the scheme number and variable dimension respectively, and For variable boundaries, To determine the number of possible solutions, a random perturbation term is added. To enhance the diversity of solutions, an initial solution library is formed. This library utilizes uniform design theory to generate a uniformly distributed set of solutions, and combines random perturbations to enhance diversity, ensuring that the solutions cover the feasible region and providing comprehensive candidate solutions for subsequent optimization. S3. Perform joint model evaluation on each scheme in the initial scheme library to form a scheme evaluation matrix containing multi-objective parameters. Through joint model evaluation, multi-objective parameters are comprehensively considered to provide a comprehensive quantitative basis for scheme optimization and avoid the limitations of single-objective evaluation. S4. The particle swarm optimization algorithm is used to optimize the evaluated scheme matrix and finally output the scheduling scheme with the best overall benefits. By leveraging the global search capability of the particle swarm optimization algorithm and combining the multi-objective evaluation results, the optimization efficiency of the scheme is improved, ensuring that the scheme with the best overall benefits is output, thus solving the problem that traditional algorithms are prone to getting trapped in local optima. S5. Apply the optimized plan to the actual operation of the reservoir, so as to realize the engineering implementation of the optimized plan, transform theoretical results into actual operation capabilities, and improve the scientific nature and efficiency of reservoir operation.
[0024] Furthermore, in step S1 above, historical hydrological data includes multi-year rainfall, inflow and outflow data, real-time hydrological data includes real-time rainfall, river level and inflow data, reservoir characteristic parameters include reservoir capacity curve and gate size, and flood control requirements include flood control limit level and downstream safe discharge standard, so as to clarify the scope of data collection, ensure the integrity and relevance of basic data, and provide accurate input conditions for subsequent data processing and model evaluation.
[0025] Furthermore, in step S2 above, the decision variables also include the safe discharge of the downstream river channel. This safe discharge is determined by hydraulic calculation of the downstream river channel cross-section, and the set of decision variables satisfies the constraint that the discharge flow does not exceed the safe discharge of the downstream river channel. By introducing the safe discharge of the downstream river channel as a decision variable and setting constraints, combined with hydraulic calculation parameters, it is ensured that the scheduling scheme meets the downstream flood control safety requirements, thereby enhancing the engineering feasibility and safety of the scheme.
[0026] Furthermore, in step S2 above, the random disturbance term The value range is [−0.1, 0.1], and a normal distribution random number generation method is used. When the decision variable is a discrete variable, the disturbed value is mapped to the nearest feasible discrete value through rounding rules to ensure the engineering feasibility of the scheme. The feasible value of the discrete decision variable needs to be determined in advance according to the regulation accuracy of the reservoir gate and the actual needs of the project to standardize the value and generation method of the random disturbance term. Combined with the discrete variable mapping rules, while ensuring the diversity of the scheme, it is ensured that the scheme meets the actual regulation accuracy of the project and avoids the disconnect between theoretical scheme and engineering practice.
[0027] Furthermore, in step S5 above, the method also includes real-time monitoring of rainfall, water level, and flow parameters through a sensor network. When the monitored values exceed the warning threshold, a rolling optimization strategy is adopted to shorten the scheduling cycle to 1 hour. At the same time, the historical similar flood case database is called to assist decision-making, so as to realize the dynamic adjustment and optimization of the scheduling plan. Through real-time monitoring and emergency scheduling mechanism, combined with rolling optimization and historical case-assisted decision-making, the reservoir's response capability to sudden floods and other abnormal situations is improved, the scheduling plan is dynamically optimized, and flood control safety and scheduling flexibility are enhanced.
[0028] Furthermore, in step S4 above, the particle swarm optimization algorithm uses the multi-objective evaluation result as the fitness function and updates the particle velocity and position using the following formula: Speed update formula:
[0029] Position update formula:
[0030] in, For the first The particle in the first The speed of each iteration For inertial weights, , As a learning factor, , It is a random number with a value in the interval [0,1]. For the first The optimal position of each individual particle. To be the globally optimal position The optimal solution in the current iteration is denoted as , and is the influence factor. This is used to enhance the learning ability of the current best solution, so as to clarify the optimization mechanism of the particle swarm algorithm through specific speed and position update formulas. By combining the learning of individual best, global best and current best solutions, the search efficiency and global optimization ability of the algorithm are improved, and the optimization effect of the solution is ensured.
[0031] Furthermore, inertia weight Using a linear decreasing strategy, the expression is:
[0032] in, =0.9, =0.4, This represents the current iteration number. To maximize the number of iterations, a linearly decreasing inertial weight strategy is adopted to balance the global search capability in the early stage and the local fine search capability in the later stage, thereby improving the convergence speed and optimization accuracy of the algorithm.
[0033] Furthermore, if the global optimal solution is not updated for 10 consecutive iterations, an adaptive mutation operation is triggered, with a mutation probability of... By using adaptive mutation operations, a mutation mechanism is triggered when the algorithm gets stuck in a local optimum, and the mutation probability is dynamically adjusted to enhance the ability of the particle swarm optimization algorithm to escape local optima, thus ensuring that the final output is the solution with the best overall benefits.
[0034] The embodiments disclosed in this invention are preferred embodiments, but are not limited thereto. Those skilled in the art can easily understand the spirit of this invention based on the above embodiments and make different extensions and variations, but as long as they do not depart from the spirit of this invention, they are all within the protection scope of this invention.
Claims
1. A reservoir optimization scheduling method based on uniform design, characterized in that, Includes the following steps: S1. First, collect basic data of the reservoir, including historical hydrological data, real-time hydrological data, reservoir characteristic parameters, and flood control requirements. Then, remove outliers through data cleaning and use standardization methods to convert data of different dimensions into the [0,1] interval, providing a high-quality data foundation for subsequent analysis. S2. For the water release scheduling scenario, determine the decision variables such as water release flow rate, reservoir water level, and scheduling period. Within the feasible region of each variable, generate a uniformly distributed set of scheduling schemes using uniform design theory. Specifically, this is done through the formula... Calculate the solution points, where and Corresponding to the scheme number and variable dimension respectively, and For variable boundaries, To determine the number of possible solutions, a random perturbation term is added. Enhance the diversity of solutions to form an initial solution library; S3. Perform joint model evaluation on each scheme in the initial scheme library to form a scheme evaluation matrix containing multi-objective parameters; S4. The particle swarm optimization algorithm is used to optimize the evaluated scheme matrix, and finally the scheduling scheme with the best overall benefits is output. S5. Apply the optimized scheme to the actual scheduling of the reservoir.
2. The reservoir optimization scheduling method based on uniform design according to claim 1, characterized in that, In step S1 above, the historical hydrological data includes multi-year rainfall, inflow and outflow data, the real-time hydrological data includes real-time rainfall, river level and inflow data, the reservoir characteristic parameters include reservoir capacity curve and gate size, and the flood control requirements include flood limit water level and downstream safe discharge standard.
3. The reservoir optimization scheduling method based on uniform design according to claim 1, characterized in that, In step S2 above, the decision variables also include the safe discharge of the downstream river channel. The safe discharge is determined by hydraulic calculation of the cross-section of the downstream river channel, and the set of decision variables satisfies the constraint that the discharge flow does not exceed the safe discharge of the downstream river channel.
4. The reservoir optimization scheduling method based on uniform design according to claim 1, characterized in that, In step S2 above, the random disturbance term The value range is [−0.1, 0.1], and the normal distribution random number generation method is adopted. When the decision variable is a discrete variable, the disturbed value is mapped to the nearest feasible discrete value through rounding rules to ensure the engineering feasibility of the scheme. The feasible value of the discrete decision variable needs to be determined in advance according to the reservoir gate regulation accuracy and the actual engineering needs.
5. The reservoir optimization scheduling method based on uniform design according to claim 1, characterized in that, In step S5 above, the method also includes real-time monitoring of rainfall, water level, and flow parameters through a sensor network. When the monitored values exceed the warning threshold, a rolling optimization strategy is adopted to shorten the scheduling cycle to 1 hour. At the same time, a database of historical similar flood cases is called to assist in decision-making, so as to realize the dynamic adjustment and optimization of the scheduling plan.
6. The reservoir optimization scheduling method based on uniform design according to claim 1, characterized in that, In step S4 above, the particle swarm optimization algorithm uses the multi-objective evaluation result as the fitness function and updates the particle velocity and position using the following formula: Speed update formula: ； Position update formula: ； in, For the first The particle in the first The speed of each iteration For inertial weights, , As a learning factor, , It is a random number with a value in the interval [0,1]. For the first The optimal position of each individual particle. To be the globally optimal position The optimal solution in the current iteration is denoted as , and is the influence factor. Used to enhance the ability to learn from the current best solution.
7. The reservoir optimization scheduling method based on uniform design according to claim 6, characterized in that, The inertial weight Using a linear decreasing strategy, the expression is: ； in, =0.9, =0.4, This represents the current iteration number. This represents the maximum number of iterations.
8. The reservoir optimization scheduling method based on uniform design according to claim 6, characterized in that, When the global optimal solution has not been updated for 10 consecutive iterations, an adaptive mutation operation is triggered, with a mutation probability of 1. This increases the probability of the particle swarm optimization algorithm escaping local optima, ultimately outputting a scheduling scheme with optimal overall benefits.