Font Size: a A A

Wavelet-BP Neural Network Bayesian Probabilistic Com-bination Forecasting Model Application In Forcast And Operation

Posted on:2016-12-04Degree:MasterType:Thesis
Country:ChinaCandidate:Phanthavong TulaxayFull Text:PDF
GTID:2272330470472202Subject:Hydrology and water resources
Abstract/Summary:PDF Full Text Request
Long-term hydrological forecast method has being research hot spots all the time all over the world, there are cause analysis method,the hydrological statistics, time series analysis method et al. to the modern theory of artificial neural network, wavelet theory, grey system et al. Each method has its advantages and disadvantages because of its mechanism and application on environment. In addition, with the role of hydropower stations in the grid system is increasing, as well as dispatch and operation of the grid system is also increasingly complex. So continuing research on forecast method, and reasonable and effective solution to the problem of hydrological system meaningful, an effective scheduling strategy has very important significance for guiding the reservoir operation. The main research results in this paper are listed as follows:(1) This paper presented a Wavelet-BP neural network Bayesian probabilistic combination forecasting model by simulates prior distribution and likelihood function with a linear regression. Its model input data was adjusted according and it was used to forecast the monthly runoff in the Namngum Reservoir. The results show that this method improves Wavelet-BP neural network model forecasting accuracy. Furthemore, compared with deterministic hydrologic forecasting method, Bayesian probabilistic forecasting can describe hydrologic forecasting uncertainties by distribution function and hence it is a feasible method that provides more meaningful information for decision making.(2) Based on predicted results, used Namngum Reservoir as study example to build a optimal dispatching mode. Compared conventional method and the POA algorithm respectively when adopted in long-term reservoirs regulation of Namngum hydropower station. The results show that Wavelet-BP neural network Bayesian probabilistic combination forecasting model can get a great deal power benefits of Namngum hydropower station. It also can provide reference for program of further generation schedule.
Keywords/Search Tags:Reservoir operation, Mid-long term runoff forecasts, BP artificial neural network, Wavelet analysis, Bayesian probabilistic, Namngum Hydropower reservoir
PDF Full Text Request
Related items