on the flow curve peak forecasting via artificial neural networks

Artificial neural networks (ANN) embody the wide family of various brain-inspired approaches finding their utilization at the solution of many up-to-date issues, e.g., voice, picture and video processing and recognition or regression analysis.The ANN-based regression analysis is able to provide a functional relationship between the high number of predictors and high number of outcomes.This ANN ability has been in the frame of the submitted research applied to offer an alternative methodism for the forecasting of hot flow curve global maximum coordinates under a wide range of thermomechanical circumstances.The research is aimed to evaluate three types of ANN -a Multi-Layer Perceptron (MLP), Radial Basis Neural Network (RBNN) and Generalized Regression Neural Network (GRNN).The results have showed that considering the description of both peak point coordinates simultaneously, the RBNN two-outputs model offered the best performance from the point of view of achieved accuracy, forecasting ability and even feasible computing time.