Dragan Rodić

Application of fuzzy logic and regression analysis for modeling surface roughness in face milliing

The objective of this study is to examine the influence of machining parameters on surface finish in face milling. A new approach in modeling surface roughness which uses artificial intelligence tools is described in this paper. This paper focuses on developing empirical models using fuzzy logic and regression analysis. The values of surface roughness predicted by these models are then compared. The results showed that the proposed system can significantly increase the accuracy of the product profile when compared to the conventional approaches, like regression analysis. The results indicate that the fuzzy logic modeling technique can be effectively used for the prediction of surface roughness in dry machining.

Application of Neuro-Fuzzy Systems for Modeling Surface Roughness Parameters for Difficult-to-Cut-Steel

The objective of this study is to examine the influence of machining parameters on surface finish in turning difficult-to-cut-steel. A new approach in modeling surface roughness which uses design of experiments is described in this paper. The values of surface roughness predicted by different models are then compared. Adaptive-neuro-fuzzy-inference system (ANFIS) was used. The results showed that the proposed system can significantly increase the accuracy of the product profile when compared to the conventional approaches. The results indicate that the design of experiments with central composition plan modeling technique can be effectively used for the prediction of the surface roughness for difficult-to-cut-steel.

RSM and Neural Network Modeling of Surface Roughness During Turning Hard Steel

In the paper examined was the influence of the cutting regime parameters on surface roughness parameters during turning of hard steel with cubic boron nitrite cutting insert. In this study for modeling of surface finish parameters was used central compositional design of experiment and artificial neural network. The values of surface roughness parameters Ra and Rt were predicted by this two-modeling methodology and determined models were then compared. The results showed that the proposed systems can significantly increase the accuracy of the product profile when compared to the conventional approaches. The results indicate that the design of experiments with central composition plan modeling technique and artificial neural network can be effectively used for the prediction of the surface roughness for hard steel and determined significand cutting regime parameters.

MONITORING OF THE DISCHARGE CURRENT BY HALL-EFFECT SENSOR

The paper describes the use of Hall-effect sensor to monitor the discharge current in electrical discharge machining (EDM). The discharge current across the gap between tool and workpiece is fed into developed acquisition system for the recording of impulses during processing. The data acquisition system consists of a sensor that works on the Hall element principle and microcontroller which collects and sends data on the PC that performs data acquisition. Experimental results have shown that discharge current and discharge duration can be clearly classified even with different machining conditions. The integration of the acquisition system can substantially improve the performance of the EDM process trought the analysis of discharge current.

ADAPTIVE NEURO-FUZZY MODELING OF THERMAL VOLTAGE PARAMETERS FOR TOOL LIFE ASSESSMENT IN FACE MILLING

The focus of this paper is to develop a reliable procedure to predict tool life during face milling process. This procedure involves a combination of Method of Least Squares and Neuro Fuzzy system. The factorial designs combined with the ANFIS techniques were applied to perform the prediction of thermal voltage. A least-squares linear regression is applied to perform the prediction of tool life from thermal-voltage signals. In this contribution we also discussed the construction of an ANFIS system that tends to provide a linguistic model for the estimation of thermal voltage obtained with different membership functions. This research focuses on developing ANFIS models using triangular and Gaussian membership functions. The work shows that the membership functions have the dominant effect among the on the accuracy model. The results indicate that the training of ANFIS with the Gaussian membership function obtains a higher accuracy rate in the prediction of thermal voltage, respectively tool life.

TOOL WEAR, CUTTING TEMPERATURE AND CUTTING FORCE DURING TURNING HARD STEEL

In this study, cutting tool`s wear, temperature and forces during turning process were investigated. Used were two types of inserts HM and CBN were taken as cutting tools and round bar of EN 90MnCrV8 hardened steel was used as the workpiece. Since the life of the cutting tool material strongly depends upon cutting temperature, it is important to predict wear and heat generation in the tool. Determination of temperature field in tool was by thermal camera. Determined was dependence of temperature tool wear parameter for two cutting tool materials as well.