İlhan Asiltürk

Modeling and prediction of surface roughness in turning operations using artificial neural network and multiple regression method

Research highlights? The surface roughness is measured during turning at different cutting parameters such as speed, feed, and depth of cut by full factorial experimental design. ? Artificial neural networks (ANN) and multiple regression approaches are used to model the surface roughness of AISI 1040 steel. ? The ANN model estimates the surface roughness with high accuracy compared to the multiple regression model. Machine parts during their useful life are significantly influenced by surface roughness quality. The machining process is more complex, and therefore, it is very hard to develop a comprehensive model involving all cutting parameters. In this study, the surface roughness is measured during turning at different cutting parameters such as speed, feed, and depth of cut. Full factorial experimental design is implemented to increase the confidence limit and reliability of the experimental data. Artificial neural networks (ANN) and multiple regression approaches are used to model the surface roughness of AISI 1040 steel. Multiple regression and neural network-based models are compared using statistical methods. It is clearly seen that the proposed models are capable of prediction of the surface roughness. The ANN model estimates the surface roughness with high accuracy compared to the multiple regression model.

An intelligent system approach for surface roughness and vibrations prediction in cylindrical grinding

This work aims to develop an adaptive network-based fuzzy inference system (ANFIS) for surface roughness and vibration prediction in cylindrical grinding. The system uses a piezoelectric accelerometer to generate a signal related to grinding features and surface roughness. To accomplish such a goal, an experimental study was carried out and consisted of 27 runs in a cylindrical grinding machine operating with an aluminium oxide grinding wheel and AISI 8620 steel workpiece. The workpiece speed, feed rate and depth of cut were used as an input to ANFIS, which in turn outputs surface roughness (Ra) and vibration (az ). Different neuro-fuzzy parameters were adopted during the training process of the system in order to improve online monitoring and prediction. Experimental validation runs were conducted to compare the measured surface roughness values with the values predicted online. The comparison shows that the gauss-shaped membership function achieved an online prediction accuracy of 99%.

Development of a neural network based surface roughness prediction system using cutting parameters and an accelerometer in turning

In this work, a technique is proposed to predict surface roughness by using neural network. Surface roughness could be predicted within a reasonable degree of accuracy by taking feed rate, cutting speed, depth of cut and three orthogonal axis (x, y, z) signals of vibrations of tool holder as input parameters. 27 experiments were performed by using a CNC lathe with a carbide cutting tool. Experimental data obtained from turning process were used for training and testing of neural network architecture based prediction system. When experimental and prediction results were compared, it has been seen that a mean accuracy of 91,17% was achieved.

Noncontact surface roughness measurement using a vision system

Surface roughness measurement is one of the basic measurement that determines the quality and performance of the final product. After machined operations, tracer end tools are commonly used in industry in order to measure the surface roughness that occurred on the surface. This measurement technique has disadvantages such as user errors because it requires calibration of the device occurring during measurement. In this study, measuring and evaluation techniques were conducted by using display devices over surface image which occurred on the processed surfaces. Surface measurement which performed by getting image makes easier measurement process because it is non-contact, and does not cause any damage. Measurement of surface roughness, and analysis was conducted more precise and accurate. Experimentally obtained results of the measurements on the parts in contact with the device is improved compared with the results of the non-contact image processing software, and satisfactory results were obtained.

Regression Modeling of Surface Roughness in Grinding

Grinding is a widely used manufacturing method in state of art industry. By realizing needs of manufacturers, grinding parameters must be carefully selected in order to maintain an optimum point for sustainable process. Surface roughness is generally accepted as an important indicator for grinding parameters. In this study, effects of grinding parameters to surface roughness were experimentally and statistically investigated. A complete factorial experimental flow was designed for three level and three variable. 62 HRC AISI 8620 cementation steel was used in grinding process with 95-96% Al2O3 grinding wheel. Surface roughness values (Ra, Rz) were measured at the end of process by using depth of cut, feed rate and workpiece speed as input parameters. Experimental results were used for modeling surface roughness values with linear, quadric and logarithmic regressions by the help of MINITAB 14 and SPSS 16 software. The best results according to comparison of models considering determination coefficient were achieved with quadric regression model (84.6% for Ra and 89% for Rz). As a result, a reliable model was developed in grinding process which is a highly complex machining operation and depth of cut was determined as the most effective parameter on grinding by variance analysis (ANOVA). Obtained theoretical and practical acquisitions can be used in various areas of manufacturing sector in the future.

Intelligent adaptive control and monitoring of band sawing

In this paper, we propose that a neuro-fuzzy based adaptive controller for control of band sawing process. The system composed of two different kinds of back propagation networks and a fuzzy logic controller. The first network accomplishes the reference forces that are compared with the measured cutting force values in the real time. The required feed rate and cutting speed are adjusted by the proposed controller. The cutting parameters are continuously updated by a secondary neural network model to compensate the disturbances. The system provides that possibility of identification of material to be cut. Material identification is determined by the measured cutting forces while cutting operation. Experimental results show that the performance of the proposed controller is very well as adaptation of the cutting speed and feed rate during band sawing in the real time.