PRELIMINARY AND FAST RISK ASSESSMENT OF DEBRIS-FLOW HAZARDS USING DIFFERENT ARTIFICIAL NEURAL NETWORKS

Abstract

Researches revealed that prediction of occurrence of debris-flow hazards can be achieved by artificial intelligence techniques. In the paper three different artificial neural networks (ANNs), i.e., back-propagation neural network (BPNN), radial basis function neural network (RBFNN) and probabilistic neural network (PNN), are employed for risk assessment of debris-flow hazards. Case of watershed area of Chen-Yu-Lan River in Nantou, Taiwan, studied by Hsiao (2003) who employed an integrated technique of digital terrain model (DTM), image processing schemes, and geographical information system (GIS) is considered. ANN models with 15 influence factors and with 7 influence factors obtained by principal component analysis (PCA) are tested, respectively. In the BPNN topology four different learning rules of MATLAB nntool are selected. It is shown that the percentage of accuracy of debris-flow hazard prediction can be up to 90.48% using BPNN models with totally 15 influence factors. Present study revealed that integration of DTM, GIS and ANN techniques can construct a preliminary and fast risk asse assessment of debris flow hazards for watershed areas.

Authors and Affiliations

LI-JENG HUANG, TZY-WEI CHEN

Keywords

Related Articles

Quantifying Impact on Water Environment by Using Fuzzy Set Theory

Assessment of the impact of developmental activities on water environment is difficult to measure as numbers of parameters are involved in the overall assessment of water quality. Efforts have been taken to develop Water...

The Work ability and Compressive Strength of Concrete using Textile Mill Sludge and Plasticizer

Growing population and life quality, both lead to higher consumption of resources. High rate of consumption of resources generate a lot of waste material. Improper management of waste material causes various environmenta...

NEW VADADE VILLAGE PROFILE & DIRECT RESEARCH FOR DETAILED STUDY ON WATER AND WASTE WATER MANAGEMENT: A REVIEW

This study focuses on understanding of the development sector of a village of Karvir Tehsil of Maharashtra state, named New Vadade. The research has been carried out through participatory approach by living in the villag...

Biomass Resource Assessment and Potential in India

The present study has been taken to access biomass potential in India and its resource assessment. Biomass energy through hybridization can prove be a boon to India. Uttar Pradesh has generated highest cumulative biomass...

Parametric Study On Polystyrene Sheet in Concrete

Now a days, we all are feeling higher degree of heat in concrete building. So, it is necessary to find solution to reduce temperature of house. The solution is, we can use polystyrene sheet in concrete block, which can h...

Download PDF file
  • EP ID EP220097
  • DOI -
  • Views 154
  • Downloads 0

How To Cite

LI-JENG HUANG, TZY-WEI CHEN (2017). PRELIMINARY AND FAST RISK ASSESSMENT OF DEBRIS-FLOW HAZARDS USING DIFFERENT ARTIFICIAL NEURAL NETWORKS. International Journal of Civil, Structural, Environmental and Infrastructure Engineering Research and Development (IJCSEIERD), 7(1), 27-36. https://europub.co.uk/articles/-A-220097