Monitoring of material-removal mechanism in micro-electrical discharge machining by pulse classification and acoustic emission signals

Abstract

Micro-electrical discharge machining is a stochastic process where the interaction between the materials and the process parameters are difficult to understand. Monitoring of the process becomes necessary to achieve the dimensional accuracy of the micro-featured components. Although thermo-mechanical erosion is the most accepted material-removal mechanism, it fails to explain the material removal with very short pulse duration. Alternative postulate like electrostatic force-induced stress yielding provides a stronger argument, rising ambiguity over the material-removal process in the micro-electrical discharge machining regime. In this work, it was found that the stress waves released from the material during micro-electrical discharge-machining process indicate material removal by mechanical deformation and fracture mechanism. These stress waves were captured using the acoustic emission sensor. The discharge pulses were captured by voltage measurement and classified using voltage gradient and machining time duration into three major categories, open pulse, normal pulse and arc pulse. The acoustic emission signal features were extracted and identified by time–frequency–energy distribution analysis. A feed-forward back-propagation neural network mapping of the pulse instances was performed with the obtained acoustic emission signature. The time–frequency–energy distribution analysis of the acoustic emission and the scanning electron microscope images of the craters provide conclusive evidence that the material is removed by mechanical stress and fracture. The feed-forward back-propagation network model was trained to predict the discharge categories of the pulse instances with AE signal inputs which can be used for monitoring the material-removal mechanism in micro-electrical discharge machining operation.

Wire Tool Electrode Behavior and Wear under Discharge Pulses

“…The vibroacoustic monitoring was provided by piezoelectric accelerometers mounted on the elastic system of the machine [23,[25][26][27]51,52] (Figure 2). The data received by the accelerometers signals were forwarded to preamplifiers, amplifiers VShV003 (JSC Izmeritel, Taganrog, Russia), and an analog-to-digital converter E440 (L-card, Saint-Petersburg, Russia), and recorded with a personal computer (PC).…”

“…The vibroacoustic signal reflects the changes in the weight and structure of the workpiece, the main discharging factors that correlated with the force diagram in the working zone [27,55] that influence the amplitude of the signal in the wide range of spectra [23,25,26,51,52].…”

Electrical Discharge Machining of Oxide Nanocomposite: Nanomodification of Surface and Subsurface Layers

“…The tool electrode used in all the experiments was a brass wire d w of 0.25 mm in diameter made of CuZn35 (Cu-65%; Zn-35%). [23,33,35,69,72,73]. The signals received by the accelerometers were forwarded to preamplifiers, VShV003 amplifiers (Joint-stock company "Izmeritel", Taganrog, Russia), and an analog-to-digital converter E440 ("L-card", Saint-Petersburg, Russia), and were recorded at 1 min, 30 s, and 5 s before the end of processing with a personal computer (PC).…”

Wear state evaluation of inner-diameter saw blade based on vibration and noise signals during processing

“…For the vibration acceleration in experiment, the short-time Fourier transform 36,37 is used to analyze its time–frequency characteristics, and then changes of various quantities over time are analyzed. For machining noise signals, only time–frequency analysis is performed.…”

“…Vibration signal and noise signal are collected by a three-axis accelerometer and a sound signal, and the sampling frequency is 10 kHz. When measuring vibration acceleration and machining noise, the main test steps are as follows: For the vibration acceleration in experiment, the short-time Fourier transform 36,37 is used to analyze its time-frequency characteristics, and then changes of various quantities over time are analyzed. For machining noise signals, only time-frequency analysis is performed.…”