You are on page 1of 11

International Journal of Machine Tools & Manufacture 46 (2006) 10531063 www.elsevier.

com/locate/ijmactool

A study on ELID ultra precision grinding of optical glass with acoustic emission
D.J. Stephenson*, X. Sun, C. Zervos
School of Industrial and Manufacturing Science, Craneld University, Craneld, Bedford MK43 0AL, UK Received 16 June 2005; accepted 22 August 2005 Available online 12 October 2005

Abstract ELID grinding of BK7 glass and Zerodur was investigated using acoustic emission. Experiments showed that the contacting area between the wheel and workpiece in a grinding process was critical to inuence wheel loading for a ne grit size resin-bonded cup wheel. ELID can be used for efcient material removal when the wheel/workpiece contacting area is large. Correlations were observed between the dressing intensity on the ELID wheel and the detected AE signals. Aggressive ELID dressing parameters for grinding with ner grit size wheels corresponded to a lower AE level. With an increase in the processing time of an ELID wheel, low and stable AE amplitudes became large with uctuations due to the deterioration of the grinding wheel. Results indicate that the AE sensing technique has the potential to be adopted as an effective method for monitoring an ultra precision grinding process, identifying the condition of the grinding wheel and investigating the mechanism of ELID grinding. q 2005 Elsevier Ltd. All rights reserved.
Keywords: ELID; Glass grinding; Acoustic emission

1. Introduction In precision grinding, the most practical means of achieving a high quality ground surface is to use a wheel with ner abrasive grit size. However, as the grit size decreases, the space for storing the chips becomes smaller, and loading is easily encountered [1]. Wheel loading occurs when chips ll the pores between active grits in the surface of the wheel. When the removal rate exceeds the rate of chip storage available, chips will tend to become lodged in the chip storage space [2]. Grinding chips adhering to the wheel surface reduce the level of grit protrusion and space for storing new chips resulting in dull rubbing actions between the wheel and workpiece. Thus, poor surface nish and severe subsurface damage are expected in the loaded condition. Wheel loading can limit the machining efciency or even make grinding impossible. Resin- and metal-bond wheels are commonly used in precision grinding. They have

* Corresponding author. Tel.: C44 1234 754751; fax: C44 1234 751172. E-mail address: d.j.stephenson@craneld.ac.uk (D.J. Stephenson).

0890-6955/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijmachtools.2005.08.013

relatively few voids after dressing/trueing, and a freshly dressed surface will be too smooth and dense, with insufcient space between active grits to accommodate chips [2]. In-process self dressing may happen in precision grinding of optical glass for a ne grit size resin bond wheel while blunt abrasive grits and resin bond material are worn out to expose new abrasive grits. The effectiveness of self dressing using a resin bond wheel needs to be investigated. An electrolytic in-process dressing (ELID) technique was used to alleviate wheel loading for ne grit size metal bond wheels. ELID is an electrochemical technique that continuously dresses a metal bond wheel through in situ electrolysis [37]. Electrolysis chemically modies the grinding wheel surface and the modied layer of the grinding wheel is removed in the grinding process to provide the necessary grit protrusion and chip storage space. In precision grinding, the ability to maintain an optimum wheel topography is critical to achieve good quality ground surfaces. Real time process monitoring or sensing methods are needed to ensure the required grinding wheel condition and hence part quality [8]. Application of non-destructive evaluation (NDE) sensors can play a signicant role in real time monitoring of a grinding process. During ultra precision machining operations in optical glass, material is

1054

D.J. Stephenson et al. / International Journal of Machine Tools & Manufacture 46 (2006) 10531063

removed from the workpiece at very low material removal rate, the uncut chip thickness usually being at the nanometer level to achieve minimal surface/subsurface damage. Power consumption, vibration and force signals show very low sensitivity and signal-to-noise (ANR) for small depths of cut, due to the low-level forces involved in the cutting process. Monitoring precision cutting processes is difcult by means of some of the sensors that are commonly employed in conventional machining operations. However, the acoustic emissions (AE) signal has proven to be sensitive enough for precision grinding monitoring, and is better suited for monitoring very fast events than, e.g. force measurements [911]. Because AE waves propagate with frequencies from 100 kHz to 1 MHz, well above most structural natural frequencies, machine vibrations will not inuence the AE signal [10]. Acoustic emissions are, therefore, regarded as an ideal method for characterizing material removal activity, providing tool condition and part quality information. AE waves can be detected by an AE sensor (piezoelectric transducer), which is mounted at a position near the ground surface. The sources of acoustic emissions are the combination of elastic impact, friction, indentation cracks, bond fracture, chip fracture, grain fracture, and grit removal at the wheel/workpiece interface [8,9]. Previous research has indicated that worn grits, wheel loading, heavy friction, and hard bonded material can result in large AE energy [1113]. From the aspect of wheel loading, ploughing, and sliding contribute to the main source of acoustic emission energy. Ploughing, characterised as plastic deformation of the workpiece without material removal, consumes energy due to this deformation. Sliding energy is expended due to sliding friction between wear surfaces on the abrasive grains and the workpiece. The effect of wheel loading during extended grinding operations reduces the efciency of the abrasive cutting action resulting in large ploughing and sliding (rubbing) components due to the grit-work interaction. This is expected to increase the energy of the acoustic emission from the process. Much effort has been directed towards developing on-line condition monitoring systems that make use of features extracted from the AE signal. A relatively reliable method for industrial implementation is the detection of root mean square (RMS) value of the AE signal [11,14,15]. The root mean square (RMS) value of the AE signal, AERMS, is dened by, v u u T u1 AERMS Z t V 2 tdt (1) T
0

dE/dt f(AERMS)2 [15], (AERMS) has been used to assess AE energy by previous studies [11,14,15]. The ELID technique has been studied intensively during the past two decades. As to the principle of ELID, to the best knowledge of the authors, previous literature has not advanced signicantly from the description of Ohmori [4]. To understand and improve the ELID technique, the electrochemical behaviour of the metal matrix of the wheel in dressing should be investigated thoroughly. For the in-process monitoring of grinding, force was commonly used by previous research to evaluate a grinding process and investigate the mechanism of ELID [1618]. It was reported that ELID could provide reduced and almost constant grinding force at the start of grinding [18]. Lim investigated the effect of ELID parameters, indicating that grinding force decreased with the increase of the duty ratio of dressing current [17]. Fathima pointed out that a lower duty ratio for dressing was preferable for coarser grit size wheels that needed higher grindability and a higher duty ratio was recommended to be applied to ner grit size wheels in order to achieve high quality surfaces [16]. In this study, acoustic emission was adopted to assess the effectiveness of ELID in alleviating wheel loading and identify the condition of the grinding wheel. Results are based on tests performed using a stiff machine tool Tetraform C, grinding BK7 glass and Zerodur. The objective of this study was to evaluate the performance of ELID grinding with a cast iron bond wheel and compare this to grinding with a resin bond wheel without ELID. The level of AE corresponding to different electrical dressing parameters was investigated based on the in situ AE measurements. This study also investigated the mechanism of ELID, providing an indication of how optimal ELID grinding conditions can be achieved.

2. Experimental set up Tests of ELID and no-ELID grinding were conducted on a precision surface grinder Tetraform C [6]. Cast iron-bond (CIB) and resin-bond diamond cup wheels of 2 and 7 mm grit size with dimensions of 124 mm diameter and 4 mm face width were used. The workpiece materials were Zerodur and BK7 glass, either rectangular (16!10 mm) or round (50 mm diameter). The ELID system used stainless steel as the cathode electrode, covering 1/6 of the wheel surfaces with a 220 mm dressing gap. A synthetic water soluble grinding uid CEM, Fuji Die, Japan, was applied as coolant and electrolyte. The ELID power supply used was an ED-921 (Fuji Die, Japan). The owchart of the AE signal acquisition system is shown in Fig. 1. Acoustic emission signals were acquired using piezoelectric transducer sensors. Sensor1, shown in Fig. 1, is a model WDI with a broad-band of 1001000 kHz from Physical Acoustics Corp. The sensor was attached to the surface of the workpiece holder using petroleum jelly. The acoustic emission signals are converted into electrical signals by the sensor, amplied to usable voltage levels by the

where V(t) is the AE raw signal and T is the integration period. The AERMS value carries information of the AE raw signal power during each interval of time. AE energy has been commonly used as one of the most signicant methods to analyse AE signals [11,14,15]. As the energy rate, dE/dt, of the acoustic emission signal, can be expressed as

D.J. Stephenson et al. / International Journal of Machine Tools & Manufacture 46 (2006) 10531063

1055

Fig. 1. Flowchart of the AE monitoring system on Tetraform C.

preampliers and transferred to the AEDSP-32/16 card, which has 16-bit resolution for dada recording. The preamplier (1220 A) provides a gain of 100 (40 dB) and uses a bandwidth lter with 1001200 kHz bandwidth to eliminate the mechanical and acoustical background noise that prevails at low frequencies. A frequency of 2 Mega sample rate per second was selected for signal acquisition. The AE facility was used to acquire AE raw signals with short operating duration for AERMS and Fast Fourier Transformation (FFT) analysis. Another AE system, AE4000-1, Walter Dittel GmbH, with a S type sensor sensor 2 in Fig. 1, was used to collect rectied AE signals for monitoring the changes of AE in a complete grinding cycle.

3. Results and discussion 3.1. AE from resin bond and cast iron bond (ELID) wheels There are primary and secondary material removal zones for grinding with a cup wheel [19] as shown in Fig. 2. Generally, the primary material removal zone can be considered to perform the majority of material removal,

while the secondary material removal process removes the ground material at a very small rate and can be considered to act as a nishing zone. In ultra precision grinding, as the depth of cut is very small compared to the radius of the edge of the wheel, the role of primary and secondary removal zones as well as the boundary between primary and secondary material removal zones are hard to distinguish (Fig. 2). Therefore, the work in this paper does not attempt to distinguish the AE contributions coming from the different material removal zones. The AE signals generated by grinding with a resin-bond wheel (no ELID) and a CIB wheel (ELID) were studied. Preliminary tests were conducted on BK7 glass samples using 7 mm grit size wheels at a wheel speed of 39 m/s, 5 mm depth of cut and 6 mm/min feed rate. The contacting area between the wheel and workpiece during machining was 40 mm2. Fig. 3 plots the results of AERMS for a number of grinding passes, which indicates that the cast iron-bond wheel with ELID grinding generated higher AERMS than the resin-bond wheel and exhibited larger scatter in values. There were no rubbing marks or severe damage on the ground surfaces. Further tests were conducted on the BK7 samples with ground surfaces of 50 mm diameter at 39 m/s

Fig. 2. Primary and secondary material removal zones for ultra precision grinding with a cup wheel.

1056

D.J. Stephenson et al. / International Journal of Machine Tools & Manufacture 46 (2006) 10531063

Fig. 3. AERMS for grinding BK7 on 40 mm2 wheel/workpiece contacting area with cast iron bond wheel (ELID) and resin bond wheel (no-ELID).

wheel speed, 2 mm depth of cut, and 3 mm/min feed rate. The contacting area between the wheel and workpiece changed over the range 0200 mm2 during the grinding process. Fig. 4 shows the AE signals corresponding to the variation of the wheel/workpiece contacting area for the resin bond and metal bond wheels. The total amount of material removal was below 75 mm3 for each wheel. In Fig. 4, the level of AE for the resin bond wheel was generally lower than the metal bond wheel when the wheel/ workpiece contacting area was less than 150 mm 2 . However, the AE level increased at a considerable rate with the enlargement of the wheel/workpiece contacting area for the resin bond wheel. Fig. 4 shows that the amplitude of AE signals reached the peak values at point B before reaching the highest wheel/workpiece contacting area200 mm2. Apparently, the wheel/workpiece contacting area largely inuenced the AE amplitude during the grinding with the resin bond wheel. The positions of the peak value of AE signal are believed to correlate with

the generation of a poor surface quality. The AE signals generated by the ELID wheel in Fig. 4 exhibited lower AE levels under the same grinding parameters as the resin-bond wheel. The wheel/workpiece contacting area did not show a signicant inuence on the AE level for ELID grinding. Fig. 5 displays the AE signals in time and frequency domains when grinding with the resin-bond wheel and metal-bond wheel at a wheel/workpiece contacting area of 180 mm2. The AE signals generated by the resin bond wheel have much larger amplitudes than the signals generated by the metal bond wheel. The saw-tooth shape AE signals generated by the resin bond wheel may be attributed to the rubbing/sliding action between the blunt wheel and workpiece. The amplitudes of the frequency components for grinding with the resin bond wheel increase as a whole compared with the ELID grinding in Fig. 5(a) and (b). Noticeable differences between the frequency components generated by the two wheels can be observed in Fig. 5. Fig. 6 shows the ground surfaces generated by the two wheels. In both the cases of ELID grinding and resin bond wheel grinding, the samples underwent ten grinding passes, in order to observe the stability of the state of the wheel over prolonged periods of time, as well as increasing wheel/ workpiece contacting area. No rubbing marks were found on the surface ground by the metal bond wheel. On the contrary, severe rubbing damage was found on the ground surface generated by the resin-bond wheel, especially in the areas of increased contact between the wheel and the workpiece. The magnied micrograph of the rubbing mark in Fig. 6(a) shows deep cracks in parallel alignments. Dark coloured strips were visible on the surface of the resin bond wheel. This is an indication that the resin bond wheel has lost its cutting efciency, in direct contrast to the ELID wheel which is constantly dressed while grinding. Together with the polishing action of the oxide layer, the cutting action of the diamond grits remains efcient and results in a surface with high quality surface nish, without any marks, even in the areas of increased wheel/workpiece contact.

Fig. 4. AE signals generated by resin bonded wheel (no-ELID) and CIB wheel (ELID) with variable wheel/workpiece contacting area.

D.J. Stephenson et al. / International Journal of Machine Tools & Manufacture 46 (2006) 10531063

1057

Fig. 5. Time and spectrum display when grinding with a (a) ELID wheel, and (b) resin bonded wheel at 180 mm2 wheel/workpiece contacting area.

Fig. 7 shows SEM micrographs of the resin bond wheel. The micrographs are taken in two different areas, one away from and one near the leading edge of the wheel, which undergoes the most aggressive conditions while grinding. From the comparison of the two pictures, it is readily apparent that the wheel has been damaged during the process with extended cracking in the matrix material. The optical micrograph near the leading edge in Fig. 8(a) shows that the number of active diamond grits decreased signicantly compared to the unloaded wheel surface shown in Fig. 8(b). The effect of wheel loading during extended grinding operations reduces the efciency of the abrasive cutting action resulting in large ploughing and sliding (rubbing) components of the grit-workpiece interaction. With increasing wheel wear and loading the energy consumption due to ploughing and sliding components increases, thus accounting for the increasing acoustic emission. The results imply that

the area of wheel/workpiece contact is a critical factor affecting wheel loading for a resin-bond wheel. Severe wheel loading was developed for a ne grit size resin-bond wheel as the wheel/workpiece contacting area increased. It can be concluded that a ne grit size CIB cup wheel with ELID grinding performs better in overcoming wheel loading than a resin-bond wheel when the contacting area between the wheel and workpiece is large. For such conditions, the ELID method is expected to be more suitable for efcient material removal in precision grinding. 3.2. Detection of wheel state with AE An ELID wheel can quickly enter a stable cutting process after pre dressing. However, with an increase in the amount of material removal or processing time, an ELID wheel can eventually become unable to perform proper cutting due to

1058

D.J. Stephenson et al. / International Journal of Machine Tools & Manufacture 46 (2006) 10531063

Fig. 6. Ground surfaces generated by (a) resin bond wheel (no-ELID) and (b) CIB wheel (ELID).

Fig. 7. SEM micrographs of resin bond wheel after grinding: (a) away from the leading edge, (b) near the leading edge.

Fig. 8. Micrographs of (a) loaded resin bond wheel and (b) unloaded resin bond wheel.

D.J. Stephenson et al. / International Journal of Machine Tools & Manufacture 46 (2006) 10531063

1059

Fig. 9. (a) Current distributions in metal bond and electrolyte, and (b) zones of intensive dressing current.

poor wheel topography. As the grinding wheel has a rough surface and many non-conductive abrasive particles embedded inside, a non-uniform current distribution between the bulk of the electrolyte and the metal surface is expected as shown in Fig. 9(a). Current ow and distribution in the metal bond and electrolyte is represented by the contours in the gure. It can be seen that the abrasive grains and cavities disturb current ow. They cause a local increase in the current density around their periphery. The zones exposing the metal bond generated by the friction of chips removing surface oxide, shown in Fig. 9(b), are also the zones of intensive dressing current. This indicates that the uneven electrochemical reaction across the metal matrix surface of the grinding wheel will be generated by the non-uniform current distribution leading to different electrolytic action on the metal bond surface. Fig. 10 shows the variation of AE signals in a series of grinding cycles on BK7 glass. The AE signals are stable, exhibiting relatively small values when the total amount of material removal of the wheel was below 75 mm3. With the increase in the amount of material removal, AE amplitudes increase and become unstable. The optical micrograph of the surface of the CIB wheel shown in Fig. 11(a) shows the presence of cracks, voids and heavily corroded areas on the wheel surface after the large material removal. The long cracks may come from short cracks or aws, which were enlarged under the periodic forces from the workpiece at the wheel/work interface. Electrolytic corrosion can promote the cracking process by concentrating electrochemical action at these positions. To investigate the morphology of the cracks under the wheel surface, a Focus Ion Beam (FIB) technique was used for surface milling for monitoring the subsurface cross section of the wheel. Fig. 11(b) shows the trench generated by ion beam milling,

where both vertical and horizontal cracks were observed under the wheel surface. The horizontal cracks can expand to connect with the vertical cracks. With the increase of the number and severity of the cracks, breakage and removal of bond materials could happen, leading to poor wheel topography and, nally, the failure of the wheel. The stochastic distributed AE signals of large amplitude in Fig. 10 may correspond to the fracture of bond material. The gradually increasing AE level between grinding cycles may indicate a deterioration in the grinding wheel.

Fig. 10. AE signals from a CIB bond wheel after different amounts of material removal.

1060

D.J. Stephenson et al. / International Journal of Machine Tools & Manufacture 46 (2006) 10531063

Fig. 11. (a) Optical micrograph of the surface of CIB wheel, and (b) FIB micrograph of trench on the wheel surface generated by focus ion beam.

3.3. ELID parameters Electrolysis plays a vital role in the formation of the oxide layer on the surface of the ELID wheel. Faradays Law has been used to develop an expression for the

theoretical volumetric transformation of the bond material [20], which is vvol Z MIt zFrbond and dvvol MI Z zFrbond dt (2)

Fig. 12. Optical microscopy graphs of BK7 ground surfaces using ELID parameters (a) 10% duty ratio; 60 V peak voltage, (b) 70% duty ratio; 60 V peak voltage, and (c) 70% duty ratio; 90 V peak voltage. Grinding conditions: 39 m/s wheel speed, 12 rpm rotary speed of the rotary table, 1 mm depth of cut, and 6 mm/min feed rate.

D.J. Stephenson et al. / International Journal of Machine Tools & Manufacture 46 (2006) 10531063

1061

where M is the atomic weight of the reacting ions; I is the current; t is the reacting time; z is the valence of the reacting ions; F is Faradays constant; rbond is the density of the metal bond. According to Eq. (2), the electrolytic activity on the wheel surface is inuenced by the electrical current applied between the wheel and the cathode electrode. There are two parameters that can determine the dosage of dressing electricity for electrolysis from the used ELID power, which are duty ratio and peak voltage. Duty ratio is dened as the percentage of the on-time of the square waveform used in ELID. The peak voltage is the amplitude of the waveform output from the ELID power. Theoretically, the two parameters can inuence the speed of generation of the corrosive layer on the wheel surface. Experimental results show that the intensity of ELID is more likely to inuence the quality of a surface ground when grinding with a diamond wheel of very ne abrasive grit size. Fig. 12 shows optical micrographs of BK7 ground surfaces generated by a 2 mm grit size diamond CIB cup wheel using ELID with different combinations of duty ratio and peak voltage. The photograph

in Fig. 12(a) shows some serious rubbing damage on the ground surface at 10% duty ratio and 60 V peak voltage. The rubbing damage was believed to be generated by some blunt areas on the wheel surface. The optical micrograph at a larger magnication in Fig. 12(a) shows cracks running normal to the sliding direction. With the increase of duty ratio from 10 to 70%, the rubbing action is relieved to some extent as shown in Fig. 12(b), in which no cracks have been found on the rubbing damage. When 70% duty ratio and 90 V peak voltage is applied, rubbing marks diminished on the ground surface as shown in Fig. 12(c). These tests indicate that high duty ratio and/or peak voltage can provide sufcient wheel dressing for these grinding conditions. The cracks on the ground surface in Fig. 12(a) may be generated by thermal effects resulting from the rubbing between the wheel and the workpiece. As BK7 has a poor thermal conductivity, thermal cracks can occur when the in-process dressing for the ner grit size wheel is insufcient. The effect of the dressing parameters on AE was investigated. Tests were performed on 16!10 mm Zerodur samples with the 7 mm grit size wheel at 39 m/s wheel

Fig. 13. Raw AE signals and power spectrum graphs under ELID conditions (a) 10% duty ratio; 60 V peak voltage and (b) 70% duty ratio; 90 V peak voltage.

1062

D.J. Stephenson et al. / International Journal of Machine Tools & Manufacture 46 (2006) 10531063

surface and generates more chip storage space, minimising wheel loading and chatter. Gentle dressing parameters can result in insufcient dressing leading to large dull areas promoting less effective cutting between the wheel and workpiece. The less efcient grinding and larger contacting area between the hard metal bond and the workpiece contributes to a large AE level.

4. Conclusions AE detection can be used to identify wheel loading and assess the grinding state of a wheel. This investigation has shown that the acoustic emission energy increases as wheel loading occurs. ELID grinding with a ne grit size cup wheel is less likely to encounter wheel loading compared to a resin bonded wheel when there are long arcs of wheelwork contact. Therefore, ELID grinding is recommended for use in efcient precision grinding, when components are relatively large. The signicant increase in AE amplitude for the resin-bond wheel corresponded to severe rubbing on the ground surface. This indicated that a resin bond wheel could not perform effective self dressing when the machining arcs of the abrasive grits were long. However, resin bond wheels tend to generate lower AE amplitude when the wheel/workpiece contacting area is small. ELID grinding with more gentle dressing parameters can generate high AE energy for a 7 mm ne grinding wheel. Intensive inprocess dressing with more aggressive dressing parameters is recommended on ner grit size grinding wheels to diminish wheel loading and increase the cutting efciency. The application of dressing parameters should consider wheel conguration, grinding parameters and material properties of the workpiece and, therefore, depends on a complex set of interactions between many variables. The AE sensing technique has the potential to be adopted as an effective method for monitoring the complex ELID grinding process to ensure optimum grinding conditions are maintained.

Fig. 14. Inuence of duty ratio on AERMS voltage of AE signals.

speed, 5 mm depth of cut and 6 mm/min feed rate. Trueing and pre-dressing of the wheel were conducted before each test series. The AE recordings started from after the ground surface of each sample was levelled by a few passes. Fig. 13 shows the raw AE signals and power spectrum graphs using ELID parameters of 10%/60 V and 70%/90 V. The AE amplitude of the raw signals in the time domain was found to be lower when a more aggressive ELID parameter was used. The AE amplitudes of frequency components in the frequency domain shown in Fig. 13(a) and (b) also displayed a decrease when the dressing parameters became more aggressive. The rate of decrease of the frequency components is relatively large at frequencies of 240 and 300 kHz. Figs. 14 and 15 plot the inuence of duty ratio and peak voltage on AERMS, respectively. The results indicate that the acoustic emission energy increases with a decrease in duty ratio and peak voltage. The parameter of duty ratio has a more signicant effect on the AE energy than peak voltage. ELID grinding involves the removal and regeneration of the oxide layer on the wheel surface [35]. When a large dosage of electricity is applied in the electrolytic environment, the formation of the oxide lm on the wheel surface is rapid. The removal of the oxide layer in the grinding process can make new grits protrude at the wheel

Acknowledgements This work was supported in part by the EC project NanoGrind (GRD1-2001-40538).

References
[1] P. Shore, ELID for efcient grinding of super smooth surfaces, Ultraprecision grinding, Industrial Diamond Review 6 (1993) 316 323. [2] M. Shaw, Principles of Abrasive Processing, Oxford University Press, UK, 1996.

Fig. 15. Effect of peak voltage on AERMS voltage of AE signals.

D.J. Stephenson et al. / International Journal of Machine Tools & Manufacture 46 (2006) 10531063 [3] H. Ohmori, T. Nakagawa, Utilization of electrolysing non-linearity in precision grinding with ELID (Electrolytic in-process dressing) for fabrication of hard material components, Annals of the ICRP 46 (1) (1997) 261264. [4] H. Ohmori, T. Nakagawa, Mirror surface grinding of silicon wafer with electrolytic in-process dressing, Annals of the ICRP 39 (1) (1990) 329332. [5] H. Ohmori, I. Takahashi, B.P. Bandyopadhyay, Ultra-precision grinding of structural ceramics by electrolytic in-process dressing (ELID) grinding, Journal of Materials Processing Technology 57 (1996) 272277. [6] D.J. Stephenson, D. Veselovac, S. Manley, J. Corbett, Ultra-precision grinding of hard steels, Precision Engineering 25 (2001) 336345. [7] B.P. Bandyopadhyay, H. Ohmori, The effect of ELID grinding on the exural strength of silicon nitride, International Journal of Machine Tools & Manufacture 39 (1999) 839853. [8] T.W. Hwang, et al., Acoustic emission monitoring of high speed grinding of silicon nitride, Ultrasonics 38 (2000) 614619. [9] W. Hundt, et al., An approach to monitoring of the grinding process using acoustic emission (AE) technique, Annals of the ICRP 43 (1) (1994) 295298. [10] T.G. Bifano, Y. Yi, Acoustic emission as an indicator of materialremoval regime in glass micro-machining, Precision Engineering 14 (4) (1992) 219228. [11] D. Dornfeld, H. Cai, An investigation of grinding and wheel loading using acoustic emission, Transactions of ASME 106 (1984) 2833. [12] H.V. Ravindra, Y.G. Srinivasa, R. Krishnamurthy, Acoustic emission for tool condition monitoring in metal cutting, Wear 212 (1997) 7884.

1063

[13] A.A. Mokbel, T. Maksoud, Monitoring of the condition of diamond grinding wheels using acoustic emission technique, Journal of Materials processing Technology 101 (2000) 292297. [14] R.E. Williams, Acoustic emission characteristics of abrasive ow machining, Transactions of the ASME: Journal of Manufacturing Science and Engineering 120 (1998) 264271. [15] R. Teti, U. La Commare, Cutting conditions and work material state identication through acoustic emission methods, Annals of the ICRP 41 (1) (1992) 8992. [16] K. Fathima, A.S. Kumar, M. Rahman, H.S. Lim, A study on wear mechanism and wear reduction strategies in grinding wheels used for ELID grinding, Wear 254 (2003) 12471255. [17] H. Lim, K. Fathima, A. Senthil, M. Rahman, A fundamental study on the mechanism of electrolytic in-process dressing (ELID) grinding, International Journal of Machine Tools & Manufacture 42 (2002) 935943. [18] B.P. Bandyopadhyay, H. Ohmori, I. Takahashi, Efcient and stable grinding of ceramics by electrolytic in-process dressing (ELID), Journal of Materials Processing Technology 66 (1997) 1824. [19] K. Weinert, C. Willsch, in: Jahrbuch, Schleifen, Honen, Lappen, Polieren (Eds.), Beitrag zum Vergleich der Oberachenqualitat beim Seitenschleifen und Umfangsschleifen von Hartmetall 57, VulkanVerlag, 1993 pp. 162170. [20] E.S. Lee, J.D. Kim, A study on the analysis of grinding mechanism and development of dressing system by using optimum in-process electrolytic dressing, International Journal of Machine Tools & Manufacture 37 (12) (1997) 16731689.