Prediction of Critical Velocity in Pipeline Flow of Slurries Using TLBO Algorithm: A Comprehensive Study

Abstract

Proper estimation of the critical flow velocity of slurries (Vc) is one of the most important parameters to design slurry transport in pipeline systems. In this study, three standard soft computing data-driven models including artificial neural network (ANN), group method of data handling (GMDH), and neuro-fuzzy inference system (ANFIS) as well as their hybrid versions combined with the teaching–learning-based optimization (TLBO) meta-heuristic algorithm are developed to estimate the Vc through pipeline. The proposed models are built and tested for accuracy by evaluating the results of the models and the collected experimental data from the literature. The results are also compared with eight suggested empirical equations as well as the soft computing method of the gene-expression programming (GEP) model. The evaluation of the results indicates that the ANFIS-TLBO model surpasses the other models and suggested equations to determine the critical velocity of slurries. According to the finding of this study, using the TLBO algorithm improves the performance of ANN, GMDH, and ANFIS by over 15%, 21%, and 4% in terms of root mean squared error, respectively.