Intelligent Methods And Their Applications To Optimal Selection Of Schemes For Floodwater Utilization By Reservoir Operation

Safe utilization of flood resources is an important strategic measure of the water conservancy work in modern china. Fully excavating the regulation and storage potentialities of existing reservoirs and fitly transforming some flood into available water resources, on the premise of not increasing the flood control risk, are more important for alleviating the water shortage problem in china, optimizing the spatial-temporal collocation of water resources and the comprehensive treatment of valley ecological environment. To coordinating the contradictions between flood and profiting to reservoirs, further exploiting and utilizing the water resources in flood season, based on the systematic exposition for modern conditions of flood resource utilization and the basic theory of reservoir operation by floodwater utilization, the research progress and application status of calculating intellectual methods applying in reservoir flood operation are summarized, besides, its feasibility and applicability are also discussed in this paper, and the major results are summarized as following:(1) The optimization of reservoir flood operation schemes can be solved as a complicated decision-making problem between many feasible operation schemes. Because of the large amount of calculation for projection pursuit method and the difficulty in determining the weights of index, the projection pursuit and fuzzy preference decision methods coupled with the maximum information entropy principle were proposed, and then the projection pursuit preference decision model and fuzzy preference decision method based on the maximum information entropy principle were established.(2) For the advantages of using set pair analysis method in characterizing the fuzziness and randomness of the system with certainty and uncertainty from the view of identical, discrepancy and contrary, the set pair analysis method was used to quantitative determine the difference situation of single index between feasible scheme and ideal optimal scheme, the connection number of single index was synthesized into comprehensive connection number by the weigh of index, then, the index value of practical operation schemes were used as the input term of the neural network, and the comprehensive connection number of practical scheme was used as the output term of the neural network, based on the training samples generated by the stochastic simulation method, the neural network model of optimal selection for reservoir flood control operation based on connection number (CN-ANN) was established.(3) The limited water level is the key point for increasing water storage and solving the contradictions between flood and profiting of reservoirs. After the dynamic adjustment of limited water level in reservoir flood season, the utilizable benefit was increased, but the probability for accident of reservoir was increased simultaneously. For this reason, the risk identification indexs in the operation system were fuzzed by using the triangular fuzzy number technique, the synthetic risk rate was transformed into a interval value from a certain value. Based on a large mount of flood hydrograph generated by the stochastic simulation method, the fuzzy synthetic risk rate interval-valued corresponding to different adjustment schemes of practical reservoir were calculated, and then, the optimizing model for dynamic limited water level of flood control operation in reservoir based on the analysis of fuzzy risk was established.(4) After applying these four intelligent optimal methods in floodwater utilization of practical reservoir operation, the results show that deeply researching the intelligent optimal methods in floodwater utilization of reservoir operation, which can further reveal the complex characteristics of the scheme optimal system for floodwater utilization of reservoir operation and be helpful for the development and perfection of the theory for floodwater utilization of reservoir operation, is a new research direction of water resource systems engineering.