Uncertainty analysis of intelligent model of hybrid genetic algorithm and particle swarm optimization with ANFIS to predict threshold bank profile shape based on digital laser approach sensing

Abstract

Measuring the bank profile shape when bank sediment particles are undergoing incipient motion (threshold) is notable for researchers. In the present study, an advanced laser gauge sensor instrument is utilized to measure the coordinates of points located in a stable channel bank cross section. A robust Particle Swarm Optimization (PSO) algorithm combined with a Genetic Algorithm (GA) is developed in an Adaptive Neuro-Fuzzy Inference System (ANFIS) model to predict and compare the bank profile with measured data. The results show that the proposed ANFIS-PSO/GA optimization method had a determination coefficient (R2) and Mean Absolute Relative Error (MARE) of 0.9951 and 0.1575, respectively, and was thus highly accurate in stable bank shape prediction. With increasing flow discharge, the ANFIS-PSO/GA model accuracy augmented, and the model could simulate the widening of the water surface width quite well with an appropriate bank profile function that can be used in the design and implementation of stable channel geometry. Moreover, an uncertainty analysis was carried out to calculate the reliability of the proposed model based on the Monte-Carlo method. The results show that the ANFIS-PSO/GA model has less uncertainty in different hydraulic conditions of channels for predicting the vertical level of threshold bank profile and can predict satisfactorily the channel bank profiles.