Study On Flood Forecasting And Real-time Correction Method Of Shifosi Reservoir

The flood disaster is one of the most important factors that restrict the development of economy in our country, which brings serious threat to our life and production. Flood forecasting is an important non-engineering measure that can solve the problem of flood control and disaster reduction. The time that Shifosi reservoir was built is not long, so there is no flood forecasting scheme to guide flood operations. There are hydraulic engineerings in upstream of Shifosi basin, therefore, it’s urgent to study the flood forecasting scheme. So, this paper takes Shifosi basin as the research object to make the flood forecasting scheme and establish the real-time correction model. Then parameters optimization system of flood forecasting model is also proposed. These will provide great help in flood safety of Liao River. The main contents and conclusions are as follows:(1) The DHF model is chosen for flood forecasting of Shifosi reservoir based on the analysis of hydrological characteristics and land surface condition in Shifosi basin. The basin is divided into four sub-elements according to the characteristics in order to establish the flood forecasting scheme for Shifosi reservoir. The parameters are optimized with PSO and manual calibration.(2) Because a set of parameters can’t reflect all kinds of floods, the paper classifies all historical floods using ISODATA model taken sub-element of Songshu as the research object. Fuzzy Reasoning Method based.on comprehensive information is used to identify floods on-line. The ISODATA model and Fuzzy Reasoning Method are combined to study the real-time flood classification forecast.(3) This paper establishes the flood forecasting correction model based on knowledge and experience to correct the system error and flood hydrograph error of Shifosi reservoir. The regression analysis method is used to correct the attenuation action and the parabola method is used to eliminate the artificial flood peak phenomenon. Then, the correction model and autoregressive least square method are combined to correct the real-time flood hydrograph, which can improve the fitting degree of actual flood and simulated flood greatly.(4) Based on B/S Frame, this paper designs and develops the flood forecasting model parameters optimization system with JSP, Java language and SQLServer2000database. This system can optimize model parameters and simulate historical flood, which also could be used in the similar flood forecasting parameters optimization.