Applying ANFIS to Estimate Stream-Way Erosion

Abstract

The particular geological conditions of Taiwan and the onset of global climate change, have increased the intensity of precipitation, leading to greater turbulence within river channels. The erosion of riverbeds is an on-going concern, as the movement of sediment is at an all-time high. This study employs adaptive-network-based fuzzy inference to analyze and predict the erosion of river channels, and compares the results with those obtained through traditional linear regression. The Houfeng Bridge section of the Ta-Chia River in Taiwan was adopted as a case study. We investigated flow rate, gradient change, the quantity of sediment, the number of typhoons and floods, and geological conditions, as major causal factors of erosion. Results indicate that under heavy turbulence, the potential for erosion is very high in the flow areas of both Houfeung Bridge and the Chenglung Revetment. Our results are consistent with the actual conditions at the site.