| Summary: | 碩士 === 國立臺灣大學 === 農業工程研究所 === 83 === The major control factors of the debris flow are rainfall
, pile , slope and watershed area . In present study , the
fuzzy theory advanced by Zadeh has been used for solving the
fuzziness of each factor . The fuzzy control system based on
fuzzy theory is an artificial intrllgence system which simula-
tes the logic process of human . The membership function
is applied to take the place of the sharp boundary that is
usually emploied in traditional set , and it makes the desc-
ription more reality.
Since hourly rainfall , the main elements of debris flow, varies tremendously in time and space , the traditional
method , such as ARMA model , can''t fit it effectively .
The artifical neural network （ANN） is capable of self-
organization and self - learning through modifying the
biological neural networks , and it can make a good desc-
ription to an non-linear system . A back-propagation of ANN
and GMDH （ group method of data handling ） have been
applied for investigating the hourly rainfall intensity with
different typhoon event in Hua-Lien area . At last ,
the method that combines of rainfall forecasting and fuzzy
control theory to forecast the degree of debris flow is
simulated . The results show that the method can make an
excellent judgement of the trend of debris flow.
|
|---|
