Application of Artificial Neuro-Fuzzy Interference System in Rainfall-Runoff Modelling at Imus River, Cavite

Abstract

The study evaluates the performance of Artificial Neuro Fuzzy Interference System (ANFIS) and its applicability to rainfall runoff modelling considering Philippine setting specifically applied to Imus river basin located in Cavite. Rainfall-runoff modelling consists of complex approach and is a well-known to be a challenging field in hydrologic and hydraulic engineering. It can be suggested to use ANFIS approach in rainfall runoff modelling based on the results obtained for the selected basin. The input it requires to develop the model is less complex than of existing methods which uses a number of parameters. The structure of the runoff model was developed using MATLAB R2018a and it utilizes a Sugeno-Takagi Fuzzy Interference System to apply the ANFIS. To illustrate the flexibility of the approach a single input and dual input variable fuzzy model was calibrated and validated. On the basis of statistical parameter calculation using NSE, PBIAS, and RSR, the capability of the approach being used in rainfall runoff modelling is highly applicable. Calibrated rainfall-runoff model was developed using the Artificial Neuro Fuzzy and can readily be used for simulation. This model could be used for the simulation of flooding extent and can help government agencies to properly design the drainages and other flood control structures.