Research Of Flood Control Forecast Operation Mode And Risk Analysis Of Multi-reservoirs

Not only do flood disasters happen frequently,but also shortage of water resources exists in China.The flood control and management has always been the focus of attention in many basins of.our country and even in the world.With the continuous development of society,science and technology,improving the flood resource utilization is the important measure to resolve water shortage.For multi-reservoirs,forecast operation mode with prediction information is the key technology to improve the flood resources utilization.Most reservoirs are in complex multi-reservoirs system in our country.Also,considering the uncertainty of hydrological forecast,this paper studies Qinghe reservior,Chaihe reservoir and their multi-reservoirs system.Flood forecasting accuracy and the uncertainty analysis method of hydrological forecast model based on optimal sampling are analyzed first.Then flood control forecast operation mode based on the prediction error analysis and flood control forecast operation of parallel multi-reservoirs considering the uncertainty of prediction error are studied.Finally,risk analysis of multi-reservoirs with joint flood control forecast operation mode is developed.The main research contents and achieved results are summarized as follows:(1)Flood forecast accuracy of Qinghe reservoir,Chaihe reservoir and Nan-Qing-Chai-Tie are analyzed,and the prediction error domains are determined.The runoff forecast accuracy of Qinghe reservoir is 94.1%.The runoff forecast accuracy of Chaihe reservoir and Nan-Qing-Chai-Tie are respectively 90%and 82%with the achievement.Large flood has the small absolute forecast error,which satisfy the requirement of forecast schedule.The maximum entropy model has been used to determine the probability density function and the prediction error domain.(2)The paper introduces the use of optimization algorithm as a sampling approach with an aim to improving sampling efficiency for multiple criteria uncertainty analysis using GLUE,and successfully presents a ?-NSGAII based sampling approach for multiple metrics GLUE analysis.In addition,the ?-NSGAII based sampling is much more efficient than the traditional method,i.e.?LHS.This approach is compared with the traditional LHS approach using a case study of flood forecasting uncertainty evaluation based on Xinanjiang model for Qinghe reservoir.All results that multiple metrics performance,sampling efficiency,parameters uncertainty and flood forecasting uncertainty show the performance of ?-NSGAII parameter sets is better.(3)The flood control forecast operation mode based on the prediction error analysis is studied.With two cases considering forecast error as 30 mm and 0 mm(accurate forecast)respectively,the flood routing results of flood control forecast operation mode considering forecast error are analyzed in Qinghe reservoir and Chaihe reservoir.The economical benefits have been significantly increased using forecast operation mode,with Qinghe reservoir water level from 127.00 m to 127.60 m and Chaihe reservoir water level raising from 104.00 m to 104.90 m.The flood control ability has also been improved with reducing the highest flood water level value obviously.In short,the flood control forecast operation mode considering the prediction error reduces flood control risk of the reservoir and downstream,also the benefit is remarkable.(4)This paper puts forward the flood control forecast operation optimization model of parallel multi-reservoirs considering the prediction error,which will enrich the theoretical basis of dynamic control of the limited water level.First of all,the general form of the forecast operation rules and the calculated steps are given,also the optimization model of the flood control forecast operation of parallel multi-reservoirs is set up.Taking Qinghe and Chaihe reservoirs for example,the flood control forecast operation rules of each reservoir are optimized.Finally,the comparison results of different operation rules are analyzed.The flood limited water levels of Qinghe and Chaihe reservoir can rise to 128.00 m and 105.20 m respectively,which can increase flood irrigation water 0.62 billion m3.The research fully realizes the potential advantages of the reservoir,also provide and verify the thought.(5)The method of determining upper bound of dynamic control of flood limited water level is proposed based on Bayes theorem and flood forecast error characteristics.Also,flood control risk of the downstream is studied using probability statistical analysis considering the uncertainty of interval flood forecast error.First,forecast error range is calculated by maximum entropy method and distribution probabilities of runoff forecast errors in each zone,namely the prior probability distributions of flood forecasting errors are obtained by integrating the density function.Then,the probabilities of the highest water levels considering different flood forecast error bounds being higher than corresponding design water levels are studied,and the risks of different flood forecast errors are inferred by Bayes theorem when the highest water level in flood operation meets with the design flood frequency.The results show that the method of forecast operation in Qinghe reservoir to determine the upper bound of dynamic control of flood limited water level 128.00 m is safe,also the water level 105.20 of Chaihe reservoir is safe.At last,the maximum combination flow distribution and occurrence probability of Tieling is obtained.The proposed risk analysis method has certain theoretical meaning and practical value.It can be the reference of the same type reservoirs.