Application of Artificial Neural Networks to Classification,Velocity Distribution and Risk Assessment of Debris Flows
碩士 === 國立高雄應用科技大學 === 土木工程與防災科技研究所 === 104 === Typhoons and earthquakes usually lead to debris-flow disasters in Taiwan. Debris flows are special hyper concentration flow containing rocks, sands and water and usually classified into stony flows, muddy flows, and mixed flows. Traditionally approaches...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2016
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/hnz5u4 |
| Summary: | 碩士 === 國立高雄應用科技大學 === 土木工程與防災科技研究所 === 104 === Typhoons and earthquakes usually lead to debris-flow disasters in Taiwan. Debris flows are special hyper concentration flow containing rocks, sands and water and usually classified into stony flows, muddy flows, and mixed flows. Traditionally approaches employed for debris flow analysis are not appropriate to treat the uncertain influence factors which usually cause the problems to be nonlinear and coupled. However, artificial neural networks (ANN) have been proven to be beneficial in the treatment of nonlinear mapping problems. This research is aimed at application of back-propagation ANN (BPANN) and radial-basis function NN (RBFNN) to investigate the classification of debris-flows, the velocity - distribution modeling and risk assessment. Numerical examples of three types of debris-flows classification; velocity distributions are tested and verified by BPANN and/or RBFNN. Furthermore, case study of risk assessment using BPANN is conducted. Results show that BPANN can be effectively applied to classification debris flow and risk assessment and RBFNN is convenient for velocity distributions fitting. The accuracy obtained can reach above 85%.
|
|---|