Abstract. The studies of tool life and formation of cutting burrs in roughing
machining field are core issues in high speed milling of compacted graphite
iron (CGI). Changing any one of the cutting parameters like cutting speed or
feed rate can result in varied tool life and different height of the cutting
burrs. In this work in order to study the relationship between cutting
parameters and tool life and height of the cutting burrs, a new differential
evolution algorithm based on adaptive neuro fuzzy inference system
(DE-ANFIS) as a multi-input and multi-output (MIMO) prediction model is
introduced to estimate the tool life and height of the cutting burrs. In
this model, the inputs are cutting speed, feed rate and exit angle, and the
outputs are tool life and height of the cutting burrs. There are 12 fuzzy
rules in DE-ANFIS architecture. Gaussian membership function is adopted
during the training process of the DE-ANFIS. The proposed DE-ANFIS model has
been compared with PSO-ANFIS, Artificial Neural Network (ANN) and Support
Vector Machines (SVM) models. To construct the predictive models, 25 cutting
data were obtained through the experiments. Compared with PSO-ANFIS, ANN and
SVM models, the results indicate that DE-ANFIS can provide a better
prediction accuracy of tool life and height of the cutting burrs, and
achieve the required product and productivity. Finally, the analysis of
variance (ANOVA) shows that the cutting speed and feed rate have the most
effects on the tool life and height of cutting burrs, respectively.