流域分类预报方法研究与应用-水文学与水资源专业论文.docxVIP

流域分类预报方法研究与应用The 流域分类预报方法研究与应用 The Research of Distribution of Forecast Water in Drainage basin or application Abstract nle forecasting programme for the large reservoirs in China at the moment has been completed,with a tagher accuracy ofhistorical floods check．However,the check for floods in these years is lower in preciNon．Those short-term floods，、Ⅳith a bigger e附in forecast value． and the historical ones are called non-standardized flood．It iS也e non-standardized flood that iS a real headache to the dispatchers on the forecast and control ofthe large reservoirs．Because the calculation precision is not easy to conlrul，and this will result in a bigger flood-forecasting error, and higher flood-dispatching difficulty．W1lat’S worse．it may even affect the safety ofthe reservoirs． There ale several reasoRs for the causes ofnon-standardized flood：the differences on the causes and intensity ofrainfall，on the requirements ofputting a cushion，and the continuous drought．For the hydrologic forecasting model ON concept that is with a complex structure and many parameters，there is not aⅡeffective way to solve the real-time correcting problems．As the non-linear technology such ag Fuzzy Sets Theory,Artificial NeuraI Networks and Genetic Algorithms being used widely and deeply in all walks oflife．it provides new ways and possibilities to solve all these problems． AcoordiIlg to the problems existing in the modem prediction,combing with the utility of the projects，using nowadays advanced classifying prediction method to take out the non． standardized floods and classify the non-standardized floods using the artificial nerve network．then according to the characters ofthe floods and levels ofthe factors，choosing different patterns ofpredicfiorts and optimize the predicted factors to predict the flood． Dispatching the folds using relevant rules to raise the accuracy ofthe prediction ofthe瑚ervoir in order to raise the level ofthe dispatching and increase the promoting benefits and SOcial