Modeling tool wear in end-milling using enhanced GMDH learning networks

This paper presents an enhanced approach to predictive modeling for determining tool-wear in end-milling operations based on enhanced-group method of data handling (e-GMDH). Using milling input parameters (speed, feed, and depth-of-cut) and response (tool wear), the data for the model is partitioned into training and testing datasets, and the training dataset is used to realize a predictive model that is a function of the input parameters and the coefficients determined. In our approach, we first present a methodology for modeling, and then develop predictive model(s) of the problem being solved in the form of second-order equations based on the input data and coefficients realized. This approach leads to some generalization because it becomes possible to predict not only the test data obtained during experimentation, but other test data outside the experimental results can also be used. Moreover, this approach makes it easy to present the realized solution in a form that can be further optimized for the input parameters using some optimization techniques. The results realized using our e-GMDH method are promising, and the comparative study presented shows that the e-GMDH outperforms polynomial neural network (PNN); moreover, it is more flexible than the conventional GMDH, which tends to produce nonlinear solutions even for simple problems. In the investigation, the extended particle swarm optimization (PSO) technique was applied to obtain the optimal parameters. Consequently, the modeling approach is extremely useful in realizing a computer-aided process-planning system in an advanced manufacturing environment.     

Prediction of tool wear using regression and ANN models in end-milling operation

Tool wear prediction plays an important role in industry for higher productivity and product quality. Flank wear of cutting tools is often selected as the tool life criterion as it determines the diametric accuracy of machining, its stability and reliability. This paper focuses on two different models, namely, regression mathematical and artificial neural network (ANN) models for predicting tool wear. In the present work, flank wear is taken as the response (output) variable measured during milling, while cutting speed, feed and depth of cut are taken as input parameters. The Design of Experiments (DOE) technique is developed for three factors at five levels to conduct experiments. Experiments have been conducted for measuring tool wear based on the DOE technique in a universal milling machine on AISI 1020 steel using a carbide cutter. The experimental values are used in Six Sigma software for finding the coefficients to develop the regression model. The experimentally measured values are also used to train the feed forward back propagation artificial neural network (ANN) for prediction of tool wear. Predicted values of response by both models, i.e. regression and ANN are compared with the experimental values. The predictive neural network model was found to be capable of better predictions of tool flank wear within the trained range.     

Prediction of tool wear using regression and ANN models in end-milling operation a critical review

Palanisamy et al. (Int J Adv Manuf Technol Volume 37:29–41, 2008) focus on two different models in their research, namely, regression model (RM) artificial neural network (ANN) for predicting tool wear. The prediction values from RM and ANN models are compared with the experimental values conducted by them. The discussion shall focus on the following four main points that are not considered in the study.     

Cutting force and diamond tool wear in stone machining

Stone machining by diamond tool is a widespread process to manufacture both standard products, such as tiles, slabs, kerbs, and so on, and design shapes. Cutting force and energy may be used to monitor stone machining. Empirical models are required to guide the selection of cutting conditions. In this paper, the effects of cutting conditions on cutting force and cutting energy are related to the shape of the idealized chip thickness. These effects are put into relationship with the diamond tool wear too. The empirical models developed in this paper can be used to predict the variation of the cutting energy. Therefore these models can be used to guide the selection of cutting conditions and to predict when it is needed to change the tool. The chip generation and removal process has been quantified with the intention of assisting both the toolmaker and the stonemason in optimizing the tool composition and cutting process parameters, respectively.     

Cutting force in face milling with tool wear

The proposed mathematical model includes not only the cutting forces due to processes in the shear zone but also the component of the force due to processes at the tool’s real surface. The model takes account of the different configuration of the replaceable multifaceted plates and the forces at radial and linear sections of the mill tooth.     

Cutting forces,tool wear and surface finish in high speed diamond machining

We have investigated the cutting forces, the tool wear and the surface finish obtained in high speed diamond turning and milling of OFHC copper, brass CuZn39Pb3, aluminum AlMg5, and electroless nickel. In face turning experiments with constant material removal rate the cutting forces were recorded as a function of cutting speed between v_c = 150 m/min and 4500 m/min revealing a transition to adiabatic shearing which is supported by FEM simulations of the cutting process. Fly-cutting experiments carried out at low (v_c = 380 m/min) and at high cutting speed (v_c = 3800 m/min) showed that the rate of abrasive wear of the cutting edge is significantly higher at ordinary cutting speed than at high cutting speed in contrast to the experience made in conventional machining. Furthermore, it was found that the rate of chemically induced tool wear in diamond milling of steel is decreasing with decreasing tool engagement time per revolution. High speed diamond machining may also yield an improved surface roughness which was confirmed by comparing the step heights at grain boundaries obtained in diamond milling of OFHC copper and brass CuZn39Pb3 at low (v_c = 100 m/min) and high cutting speed (v_c = 2000 m/min). Thus, high speed diamond machining offers several advantages, let alone a major reduction of machining time.     

Cutting force and tool wear of single diamond-coated bead

The aim of this study is to optimize stone materials cutting by diamond wire cutter. Attention is focused on the study of the cutting process through sintered diamond-encrusted beads used to work granite. This study is aimed at understanding the interaction between the bead and the material; in detail, experimental equipment was designed to test individual diamond-coated beads, and it was installed on a numerical-control work center. This equipment made it possible to test an individual bead and, in particular, to determine its cutting power and its main force components on the basis of various process parameters, such as cutting velocity (V _t) and feed velocity (V _a). The test runs also made it possible to determine wear on each bead on the basis of process parameters. This is a necessary first step to be able, in the future, to optimize the tool and the cutting process.     

An artificial-neural-networks-based in-process tool wear prediction system in milling operations

An artificial-neural-networks-based in-process tool wear prediction (ANN-ITWP) system has been proposed and evaluated in this study. A total of 100 experimental data have been received for training through a back-propagation ANN model. The input variables for the proposed ANN-ITWP system were feed rate, depth of cut from the cutting parameters, and the average peak force in the y-direction collected online using a dynamometer. After the proposed ANN-ITWP system had been established, nine experimental testing cuts were conducted to evaluate the performance of the system. From the test results, it was evident that the system could predict the tool wear online with an average error of ±0.037 mm. Experiments have shown that the ANN-ITWP system is able to detect tool wear in 3-insert milling operations online, approaching a real-time basis .     

Cutting tool crater wear measurement with white light interferometry

In this paper we propose the use of white light interferometry for measurement of wear craters on cutting tool inserts. It was briefly discussed interferometry principles and tool wear during dry metal cutting, after which a discussion of the application of interferometry to cutting tool wear measurement was described. A comparison of the profiler data with the cutting forces measured during the metal cutting process in orthogonal mode under different conditions was undertaken. Lastly, a comparison between the tool wear to the evolution of the friction coefficient is given.     

Fuzzy-nets-based in-process surface roughness adaptive control system in end-milling operations

A fuzzy-nets-based in-process adaptive surface roughness control (FN-ASRC) system was developed to be able to adapt cutting parameters in-process and in a real time fashion to improve the surface roughness of machined parts when the surface roughness quality was not meeting customer requirements in the end-milling operations. The FN-ASRC system was comprised of two sub-systems: (1) fuzzy-nets in-process surface roughness recognition (FN-IPSRR); and (2) fuzzy-nets adaptive feed rate control (FN-AFRC) sub-system. To test the system, while the machining process was taking place, the FN-IPSRR system predicted the surface roughness, which was then compared to the desired surface roughness. If the desired surface roughness was not met, then, the FN-AFRC system proposed a new feed rate for the machining process. Once the feed rate was changed, and the cutting continued, the output of the surface roughness of the new feed rate was compared with the desired surface roughness. This proposed FN-ASRC system has been demonstrated to be successful using 25 experimental tests with 100% success rate.     

Analysis of force patterns and tool life in milling operations

The paper describes the practical effects of the operating parameters in the milling operation. Experiments have been conducted to measure cutting force and tool life under dry conditions. Based on the experimental results, three mathematical models have been developed: Force, TLife and Force/TLife. Further analyses have been conducted on the cutting force patterns: seasonal pattern and nonlinear trend. A process optimisation that is based on the minimum production cost has been applied to relate Force model, TLife model and machinability criteria, such as power consumption, cutting parameters and surface roughness.Nomenclature C _w cost of workpiece ($) - C _s set-up cost ($) - C _m machining cost ($) - C _o overhead cost ($) - C _r tool replacement cost ($) - C _t tool cost ($) - D diameter of the cutter (inch) - d depth of cut per pass (inch) - d ₀ required depth (inch) - e _t random error attth sample - F cutting force (N) - f feedrate (ipm) - L length of workpiece (inch) - N spindle speed (r.p.m.) - n number of teeth - P power of the motor (h.p.) - R surface roughness (µm) - R _e real part of a complex function - T tool life (min) - t sample number - t _m machining time (s) - t ₀ overhead time (s) - t _r tool replacement time (s) - t _s set-up time (s) - U _i unit cost of itemi ($/unit)v - v cutting speed (i.p.m.)     

Cutting tool wear detection using multiclass logical analysis of data

This article presents a new tool wear multiclass detection method. Based on the experimental data, tool wear classes are defined using the Douglas–Peucker algorithm. Logical analysis of data (LAD) is then used as machine learning, pattern recognition technique for double objectives of detecting the present tool wear class based on the recent sensors' readings of the time-dependent machining variables, and deriving new information about the intercorrelation between the tool wear and the machining variables, by doing pattern analysis. LAD is a data-driven technique which relies on combinatorial optimization and pattern recognition. The accuracy of LAD is compared to that of an artificial neural network (ANN) technique, since ANN is the most familiar machine learning technique. The proposed method is applied to experimental data those are gathered under various machining conditions. The results show that the proposed method detects the tool wear class correctly and with high accuracy.     

Cutting tool wear monitoring for reliability analysis using proportional hazards model

Electrochemical machining (ECM) is an important technology in machining difficult-to-cut materials and to shape free-form surfaces. In ECM, material is removed by electrochemical dissolution process, so part is machined without inducing residual stress and without tool wear. To improve technological factors in electrochemical machining, introduction of electrode tool ultrasonic vibration is justifiable. This method is called as ultrasonically assisted electrochemical machining (USAECM). In the first part of the paper, the analysis of electrolyte flow through the gap during USAECM has been presented. Based on computational fluid dynamic methods, multiphase, turbulent and unsteady electrolyte flow between anode and cathode (under assumption that cavitation phenomenon occurs) has been analysed. Discussion of the obtained solutions is the base to define optimal conditions of electrolyte flow in case of USAECM process. The second part of the paper is connected with experimental investigations of USAECM process. Classic experimental verification of obtained results in case of machining is extremely difficult, but influence of the ultrasonic vibration can be observed indirectly by changes in technological factors (in comparison to machining without ultrasonic intensification), whereas results of numerical simulation give possibility to understand reason and direction of technological factors changes. Investigations proved that ultrasonic vibrations change conditions of electrochemical dissolution and for optimal amplitude of vibration gives possibility to decrease the electrode polarisation.     

A weighted hidden Markov model approach for continuous-state tool wear monitoring and tool life prediction

Tool wear is one of the important indicators to reflect the health status of a machining system. In order to obtain tool’s wear status, tool condition monitoring (TCM) utilizes advanced sensor techniques, hoping to find out the wear status through those sensor signals. In this paper, a novel weighted hidden Markov model (HMM)-based approach is proposed for tool wear monitoring and tool life prediction, using the signals provided by TCM techniques. To describe the dynamic nature of wear evolution, a weighted HMM is first developed, which takes wear rate as the hidden state and formulates multiple HMMs in a weighted manner to include sufficient historical information. Explicit formulas to estimate the model parameters are also provided. Then, a particular probabilistic approach using the weighted HMM is proposed to estimate tool wear and predict tool’s remaining useful life during tool operation. The proposed weighted HMM-based approach is tested on a real dataset of a high-speed CNC milling machine cutters. The experimental results show that this approach is effective in estimating tool wear and predicting tool life, and it outperforms the conventional HMM approach.     

Prediction of lining wear life of bins and chutes in bulk solids handling operations

This paper is concerned with the prediction of lining wear life of bins and chutes in bulk solids handling plant. It focuses on abrasive wear and outlines the basic principles to be embodied in the development of a laboratory wear tester. Emphasis is given to a linear action wear tester developed jointly by the University of Twente, The Netherlands, and the University of Newcastle, Australia. The characteristics of abrasive wear in bins, hoppers and chutes are described and the application of test results to the prediction of wear life of lining materials is illustrated.     

The analysis of tool life and wear mechanisms in spindle speed variation machining

Regenerative chatter vibrations generally limit the achievable material removal rate in machining. The diffusion of spindle speed variation (SSV) as a chatter suppression strategy is mainly restricted to academy and research centers. A lack of knowledge concerning the effects of non-stationary machining is still limiting its use in real shop floors. This research is focused on the effects of spindle speed variation technique on tool duration and on wear mechanisms. No previous researches have been performed on this specific topic. Tool wear tests in turning were carried out following a factorial design: cutting speed and cutting speed modulation were the investigated factors. The carbide life was the observed process response. A statistical approach was used to analyze the effects of the factors on the tool life. Moreover, the analysis was extended to the wear mechanisms involved during both constant speed machining and SSV. The worn-out carbide surfaces were examined under a scanning electron microscope equipped with an energy dispersive X-ray spectrometer. Significant differences were appreciated. It was observed that SSV tends to detach the coatings of the inserts, entailing a mechanism that is quite unusual in wet steel turning and thus fostering the wear of the tool. The performed analysis allowed to deduce that the intensified tool wear (in SSV cutting) is mainly due to thermo-mechanical fatigue.