Prediction of laser solid freeform fabrication using neuro-fuzzy method

In this paper, a new application of a neuro-fuzzy method (ANFIS) to laser solid freeform fabrication (LSFF) is presented. The laser solid freeform fabrication process is a complex manufacturing technique that cannot be modeled analytically due to non-linear behaviours of the physical phenomena involved in the process. A neuro-fuzzy model is proposed to predict the clad height (coating thickness) as a function of laser pulse energy, laser pulse frequency, and traverse speed in a dynamic fashion. Four membership functions are assigned to be associated with each input of the model architecture. Experiments are performed to collect data for the training of the proposed network, and a set of unseen experimental data are also considered for the verification of the identified model. The effects of the assigned inputs on the clad height are discussed. The comparison between the experimental data and the model output shows promising results. The model can predict the process with an absolute error as low as 0.07%.