Research On Generalized State Search Algorithm And Scheduling Rules For Reservoirs

Optimal operation of reservoir systems has been a hot issue in the field of optimization of water resources. Lots of scholars have launch an in-depth study on it and made a series of achievements. With the continuous development of Chinese hydropower, the optimal scheduling problems of reservoir facing more and more challenges. Here are two main issues: with the expansion of the size of the reservoir systems, it becomes very difficult to solve the optimal scheduling model. The solving process is slow and time-consuming, even unable to get results. On the other hand, the uncertainty of reservoir runoff leads to the scheduling plan is unable to get the desired results. For these two problems, we launch a series of studies and propose corresponding solution strategy.Progressive optimality algorithm(POA) and dynamic programming with successive approximation(DPSA) are common algorithms to solve the optimization model of reservoir operation. They reduce the dimension from time and space dimension, and lower the computational scale. But the coordinate rotation strategy makes the algorithm easy to converge to the constraint boundaries and export pseudo-optimal solution. To solve this problem, this paper introduces the generalized time state and generalized space state(collectively called generalized states). This paper proposes a method to quickly generate the reference line of the reservoir regulation principles In addition, to obtain the initial trajectory of scheduling reservoir operation, according to the scheduling rules of reservoir, that is lifting head and reducing abandoned water may increase the power generation. This paper conducts simulation tests with the classic test case 4-reservoirs problem and 10-reservoirs problem and verifies the practicality of the proposed method. In addition to this, an example of Lishui river in Hunan Province is calculated to illustrate the effectiveness of the generalized state search algorithm.In order to solve reservoir operation runoff uncertainty, this paper thought implicit stochastic optimal scheduling to calculate the sample reservoir through a long series of historical data to optimize, and then fitted from the sample scheduling rules to guide the optimization operation of reservoir operation. Currently used to fit the reservoir operation rules are mainly a function of scheduling rules and fuzzy sets method. Wherein the scheduling function to fit the need for pre-functional form assumptions, and the current reservoir operation rules for specific forms is unknown scheduling function makes use of the presence of these certain limitations; commonly used form of linear scheduling function simply ignores the reservoir but because of scheduling Nonlinear itself makes scheduling rules does not fully reflect the objective laws of reservoir operation. Although fuzzy sets rules on these issues has been improving, but there is still the concept of "hard partition" and other issues. Therefore, this article summarizes the advantages and disadvantages of two algorithms based on the cloud model is introduced to extract the scheduling rules: cloud transform to extract from the sample data model concepts that best reflect the distribution of the sample cloud implements the concept of "soft" division; using qualitative concept of cloud will transform data into Boolean sample data, and the use of Apriori algorithm for data mining to obtain qualitative reservoir operation rules; use these rules to establish the rules of cloud model generator, to achieve a qualitative concept of quantitative data conversion, obtain reservoir operation rules. Scheduling rules extraction methods described herein can be introduced in a series of qualitative association rules provide reference for decision-making staff scheduling; can also cloud model generator generates a decision rule for interval scheduling personnel selected according to the actual situation and their own experience; also be able to provide a deterministic scheduling decisions, and therefore has a strong operability in actual schedule. In this study, the use of scheduling rules obtained guidance Jiangya hydropower plant simulation runs proved optimization and stability obtained by this method of scheduling rules.