Ibrahim N. Tansel

Modeling micro-end-milling operations. Part I: analytical cutting force model

A new analytical cutting force model is proposed for micro-end-milling operations. The model calculates the chip thickness by considering the trajectory of the tool tip while the tool rotates and moves ahead continuously. The proposed approach allows the calculation of the cutting forces to be done accurately in typical micro-end-milling operations with very aggressively selected feed per tooth to tool radius ( ft/r) ratio. The difference of the simulated cutting forces between the proposed and conventional models can be experienced when ft/r is larger than 0.1. The estimated cutting force profile of the proposed model had good agreement with the experimental data.

Modeling micro-end-milling operations. Part II: tool run-out

The effect of run-out is clearly noticed in micro-end-milling operations, while the same run-out creates negligible change at the cutting force profile of conventional end-milling operations. In this paper, the cutting force characteristics of micro-end-milling operations with tool run-out are investigated. An analytical cutting force model is developed for micro-end-milling operations with tool run-out. The proposed model has a compact set of expressions to be able to estimate the cutting force characteristics very quickly compared to the numerical approaches. The cutting forces of micro-end-milling operations simulated by the proposed model had good agreement with the experimental data.

Tool wear estimation in micro-machining.: Part I: tool usage–cutting force relationship

The relationship between the cutting force characteristics and tool usage (wear) in a micro-end-milling operation was studied for two different metals. Neural-network-based usage estimation methods are proposed that use force-variation- and segmental-averaging-based encoding techniques.

Micro-end-milling-I. Wear and breakage

Unpredictable tool life and premature tool failure are major problems in micro-machining. In this study, the failure mechanisms of micro-end-mills were studied during the machining of aluminum, graphite electrodes and mild steel workpieces. Hundreds of machining operations were performed, and the pictures of cutting edges were taken with a scanning electron microscope to identify fatigue and extensive stress-related failure mechanisms. Also, the cutting force variation was monitored, i.e. the relationship between the utilization-related changes at the tool structure (wear), and the outcomes (increasing cutting force which means raising stress on the tiny shaft). Inspection of the cutting force variation patterns of large numbers of micro-end-mills indicated that tool failure occurs with chip clogging, fatigue and wear-related excessive stress depending on the characteristics of the workpiece. Two tool breakage prediction methods were developed by considering the variation of the static part of the feed direction cutting force. These methods used segmental averages and wavelet transformation coefficients. The accuracy of the proposed approaches were tested with experimental data and the agreement between the predictions and actual observations are reported.

DETECTION OF TOOL FAILURE IN END MILLING WITH WAVELET TRANSFORMATIONS AND NEURAL NETWORKS (WT-NN)

Abstract Detection of tool failure is very important in automated manufacturing. In this study, tool failure detection was conducted in two steps by using Wavelet Transformations and Neural Networks (WT-NN). In the first step, data were compressed by using wavelet transformations and unnecessary details were eliminated. In the second step, the estimated parameters of the wavelet transformations were classified by using Adaptive Resonance Theory (ART2)-type self-learning neural networks. Wavelet transformations represent transitionary data and complex patterns in a more compact form than time-series methods (frequency and time-domain) by using a family of the most suitable wave forms. Wavelet transformations can also be implemented on parallel processors and require less computations than Fast Fourier Transformation (FFT). The training of ART2-type neural networks is faster than backpropagation-type neural networks and ART2 is capable of updating its experience with the help of an operator while it is monitoring the sensory signals. The proposed approach was tested in over 171 cases and all the presented cases were accurately classified. The proposed system can be easily trained to inspect data during transition and/or any complex cutting conditions. The system will indicate failure instantaneously by creating a new category, thus alerting the operator.

Monitoring drill conditions with wavelet based encoding and neural networks

Abstract Encoding of thrust force signals of microdrilling operations with wavelet transformations and classification of estimated coefficients with adaptive resonance theory (ART2)-type neural networks are proposed for detection of severe tool damage just before complete tip breakage occurs. The coefficients of the wavelets were classified both directly and after a secondary encoding to reduce the humber of inputs. Direct classification of the wavelets was found to be more reliable in the sixty-one cases studied. The proposed approach was also tested with two sampling intervals. Large sampling intervals were used to inspect complete drilling cycles. Smaller sampling intervals were used to focus on thrust force variations during the motion of the machine tool table when it is driven by a stepping motor. It was found that the data collected at smaller sampling intervals were easier to classify to detect severe damage to the tool.

Modeling micro-end-milling operations. Part III: influence of tool wear

The characteristics of the cutting forces were studied at different usage levels and the analytical model of the micro-end-milling operations was modified to represent the tool wear. A new expression was derived from the model to estimate the remaining tool life from experimental data. The parameters of the model are estimated by using genetic algorithms. The difference between the simulated and experimental cutting force profiles for new and worn tools was less than 8%. The remaining tool life was estimated with typically 10% error from the experimental data. Maximum error was 20%. The introduced analytical model and genetic algorithm-based parameter estimation approach is very convenient for on-line tool wear monitoring without extensive experimental study.

Characterization of fouling kinetics in ultrafiltration systems by resistances in series model

Abstract Membrane processes are gaining importance for water treatment applications as a result of the advances in membrane technology and increasing requirements on water quality. Membrane fouling, which results in loss of productivity, is one of the most important operational concerns of membrane processes. A number of theoretical models have been developed to explain membrane fouling mechanisms. These models use the system parameters (i.e., viscosity, pore size, membrane thickness, pressure) and unsteady-state material and flow balance equations with specific boundary conditions. As a result, they are of scientific interest and have limited use for practical applications. The purpose of the study was to develop a simple flux model to estimate the characteristic fouling times for ultrafiltration systems based on operational data. A simple first-order model was developed based on the resistances-in-series approach to correlate flux to a characteristic clean membrane parameter, system parameters. The model was used to estimate the characteristic fouling times of a laboratory ultrafiltration system.

Micro-end-milling-III. Wear estimation and tool breakage detection using acoustic emission signals

Acoustic Emission (AE) signals have been used to monitor tool condition in conventional machining operations. In this paper, new procedures are proposed to detect tool breakage and to estimate tool condition (wear) by using AE. The proposed procedure filters the AE signals with a narrow band-width, band-pass filter and obtains the upper envelope of the harmonic signal by using analog hardware. The envelope is digitized, encoded and classified to monitor the machining operation. The characteristics of the envelope of the AE were evaluated to detect tool breakage. The encoded parameters of the envelope of the AE signals were classified by using the Adaptive Resonance Theory (ART2) and Abductory Induction Mechanism (AIM) to estimate wear. The proposed tool breakage and wear estimation techniques were tested on the experimental data. Both methods were found to be acceptable. However, the reliability of the tool breakage detection system was higher than the wear estimation method.

Detection of tool breakage in milling operations—I. The time series analysis approach

Abstract On-line monitoring of tool cutting conditions and tool breakage is very important for automated factories of the future. In this paper, the time series based tooth period modeling technique (TPMT) is proposed for detecting tool breakage by monitoring a cutting force or torque signal in any direction. TPMT uses the fast a posteriori error sequential technique (FAEST) for on-line modeling of cutting force or torque signals. Tool breakage is detected by evaluating variations of the characteristics of the monitored signal in each tooth period. TPMT was tested on simulated and experimental end milling data. The proposed technique detected tool breakage in all of the test cases without giving any false alarms in the transition cases.