Electrical discharge machining of ceramic nanocomposites: sublimation phenomena and adaptive control.

The productivity of electrical discharge machining (EDM) is relatively low owing to the natural laws of electrical erosion. Precise EDM demands uninterrupted control of the discharge gap and adjustment of process parameters. It is particularly critical for processing large workpieces with complex linear surfaces and for materials with threshold conductivities such as the new advanced ceramic nanocomposites Al2O3+TiC and Al2O3+SiCw+TiC(30-40%). In these cases, adequate flushing of erosion products is hampered by the geometry of the working space or by the small value of the required discharge gap, which does not exceed 2.2-2.5 μm. The methods of adaptive control in modern computer numerical control systems of EDM equipment based on measuring the electrical parameters in the working zone have been shown to be ineffective in the cases described above. This study aims to investigate the natural phenomena of material sublimation under discharge pulses for conductive ceramics and nanocomposites. The measured conductivities of the samples are higher than the percolation threshold. However, the question of machinability remains open owing to detected processing interruptions and poor quality of machined surfaces. New knowledge on EDM of conductive ceramics and nanocomposites can improve the final quality of the machined surfaces and productivity of the method by the introduction of advanced monitoring and control methods based on acoustic emissions. The manuscript presents an up-to-date overview and current state of the research on the subject area. The obtained morphology of the samples and discussion of the findings complete the experimental part of the study. The scientific basis for a new type of adaptive control system is provided. This can improve the effectiveness of parameter control for machining conductive ceramics and nanocomposites and contribute to an increase in the EDM performance for the most critical cases.

Introduction

Electrical discharge machining (EDM) is one of the priciest machining processes for producing metallic parts with complex linear surfaces and roughness Ra of up to 0.17–0.2 [1, 2, 3, 4, 5]. Modern computer numerical control (CNC) machines allow producing parts with a positioning accuracy of guiding bars of up to 80 nm [6, 7, 8, 9]. The technology application has been extended significantly by the addition of up to six rotating axes [10, 11]. At the same time, the use of new oil-based ecological dielectrics makes the process even more precise and safer for the environment [12, 13, 14, 15, 16]. However, EDM has low productivity and is applicable only for conductive materials [17, 18, 19, 20].

The low productivity of EDM is related to the natural laws of erosion under discharge current and the auxiliary time related to the work preparation process such as

time on CNC programming either manually or according to a three-dimensional (3D) model in a computer-aided manufacturing system;

time on workpiece fastening and locating of the relative coordinate system;

time on controlling tank flooding and refastening of the workpiece during processing, if necessary, according to the developed technological route; and

time on resolving manufacturing issues related to wire tool breakage, recharging, impossibility to restart the program at the place of its rupture, returning to the home point, restarting of CNC program, etc.

It is particularly critical in the case of machining an array of narrow slots for ejector pins of an injector mould [21, 22]. For example, up to 24 tiny slots are required in the form of 1.5 × 8 mm rounded rectangles at the workpiece of 15–25 mm thickness at 100–150 mm of remoteness from a nozzle for producing an injection mould of a car backlight body (Fig. 1). The wire can be blocked during EDM by falling out internal pieces, which can be stuck as well in the lower nozzle of the machine. Blocking of wire tool often leads to undesirable issues related to processing instability, short circuits, and wire breakage. Finally, it influences dramatically the quality of the machined surfaces, which cannot be finished by another machining method due to its small size and inaccessibility [23, 24, 25].

Fig. 1

The body of motor vehicle backlight: (a) A view of the injection mold die with the slots for push-type ejectors designed by JSC IMID (Moscow, Russian Federation); (b) A produced and assembled backlight (Lada Kalina Universal, AvtoVAZ, Tolyatti, Samara region, Russian Federation).

Another case is separating two geometrically mating profiles of an injection mould from a large workpiece with a weight of up to 300 kg used for manufacturing an injection mould of milk bins for transporting dairy products (Fig. 2). It is necessary to control the moment of the final separation of the two workpieces, which are held by a few bridges, to prevent uncontrolled falling into the tank of the machine [26, 27]. The practice of working at EDM machines allows detecting the moment of separation by a specific noise to pause the machine on time for refastening the parts [28, 29]. Similar problems exist in the case of machining a complex linear surface with a variable inclination [30]. During processing, the surface under variable angle flushing of the working zone is not often sufficient and the retained debris hampers the processing, reduces the processing speed, and leads to short circuits. Another big challenge for EDM can be cutting a workpiece with an uneven structure such as a set of hollow tubes or foam-like material or material with varying mesostructure even from the known conductive and easy-to-machine (by EDM) material [31, 32].

Fig. 2

Schematic diagram of the milk bin production by EDM (JSC IMID, Moscow, Russian Federation).

The excellent properties of ceramics such as high hardness, resistance to abrasion, erosion, and corrosion, absence of chemical interaction in the working medium and with the environment, and low thermal expansion make them unreplaceable for the aviation and aerospace industry and in other important applications [33, 34, 35, 36, 37, 38, 39, 40]. In medicine, ceramic implants show their reliability, high wear resistance, absence of metal ion accumulation, reduced risk of revision surgery, and negative consequences [41, 42, 43]. However, conventional machining technologies are very costly in the processing of ceramics. Studies [44, 45, 46] have shown and proved the possibility of improving the conductivity of ceramics by conductive additives with no losses on their physical and mechanical properties. The first advanced ceramics based on the commonly known and applicable Al2O3+30–40% system of TiC for the production of cutting tools have been available industrially since the early 80s (1981–1984) [47, 48, 49]. Over time, the production methods have been significantly improved by adding more fine-grained phases in the composite based on Al2O3 with 30% of TiC and using more advanced methods of pressing and sintering [50, 51]. Researchers have studied the percolation threshold and conductive properties of new advanced ceramics with respect to the proportion of conductive additives [52, 53]. The first studies on samples of composites and nanocomposites of Al2O3+TiC system produced by spark plasma sintering (SPS) were reported by the group headed by Prof. Zhang [54, 55, 56]. Then, a research group under the supervision of Prof. Torrecillas and Prof. Moya proposed an advanced SPS method assisted by electrical current to produce new nanocomposites of Al2O3+SiCw+TiC. In addition, the question of the percolation threshold and the achieved electrical and mechanical properties of samples were discussed [57, 58, 59]. Note that having a suitable conductivity for EDM and attaining precise machining are two different things. Since 1997, there have been a few publications related to EDM of a sample made of Al2O3+TiC composite [60, 61, 62]. The authors gave recommendations for EDM pulse generator parameters, but a satisfactory quality of machined surface that can be suitable for industrial use was not achieved. The issues on the machinability of Al2O3+SiCw+TiC ceramics by electrical discharge were studied as well by another known scientific group under the supervision of Prof. Pulak M. Pandey [63, 64, 65]. They showed optimised EDM parameters and analysed the surface after EDM. However, the research on EDM of ceramics with a suitable percolation threshold remains at the stage of general electrical parameter definition. Because the issue was already pre-studied experimentally [10, 66, 67, 68], that can be enough for the rough separation of parts. High-precision EDM of conductive ceramic composites and nanocomposites remains a challenging and science-intensive problem.

As it can be seen from the available references, Al2O3(+SiC)+TiC composites and nanocomposites obtained by production methods with 30–40% of additives are electrically suitable for EDM. The conductivity of these composites is often higher than the percolation threshold of materials that can be called conductors. Ceramic composites with 40% of conductive additives can have an electrical conductivity close to or even higher than that of hard alloys. Consequently, they should be suitable for EDM as shown in the many studies mentioned above. Nevertheless, in reality, a satisfactory quality of machined surface cannot be achieved.

The above problem can be related mostly to the high thermal resistance of ceramics and the small value of the required discharge gap, which is 15–20 times less than that of conductive materials [69]. The high thermal resistance of ceramics made them unsuitable for conventional approaches in EDM. They need another electrode tool material that is more thermally resistant for a proper combination with the workpiece. In this case, the conventional monitoring and adaptive control systems of EDM based on electrical parameters are unsuitable for the machining of ceramics [70, 71, 72, 73, 74, 75]. Studies [45, 46, 66, 71] that investigated composites with reinforced ferrous matrix and conductive composites based on nitride ceramics demonstrated the typical defects and difficulties in using conventional machining techniques.

As can be observed, being conductive is not sufficient for precision EDM. This research is aimed at revealing the nature of the erosion process of conductive ceramics under discharge current and developing adequate monitoring tools. It is planned to determine a suitable material for the wire tool and estimate the discharge gap.

Note that the EDM process takes place in conditions of full dielectric submersion. The working area is tiny because the discharge gap is approximately 0.045 mm for stainless steel and located at a distance of 1–2 mm from the dielectric level. Additional monitoring methods based on visual access to the working zone are not sufficient, but conventional monitoring methods based on electrical parameters proved to be ineffective [76, 77, 78].

In-situ monitoring based on acoustic emission is shown to be reliable for the requirements of mechanical processing [79, 80, 81, 82]. It can detect the wear rate of cutting tools or unstable conditions of a machine tool during processing. The significant advantage is the simplicity of installation and the possibility of collecting acoustic data remotely. The ability to use acoustic emission for an adaptive control system with higher reliability has been proved [74].

The scientific novelty of this study is the investigation of the natural phenomena of electrical erosion of the electrode materials under discharge current during the processing of conductive ceramic composites and nanocomposites (materials with suitable or threshold conductivities, but high thermal resistances) based on acoustic emission monitoring data and microscopy. The practical importance of the study is the analytical determination of the suitable electrode tool material and appropriate value of the discharge gap, as well as the recommendation on the developed in-situ monitoring method.

The main objectives of the study are as follows:

development of an in-situ monitoring method based on acoustic emission for the EDM requirements of samples with complex linear surfaces, uneven mesostructures, and threshold conductivities;

applicability of the developed system on samples of industrially obtained ceramic composite Al2O3+TiC and advanced ceramic nanocomposite Al2O3+SiCw+TiC;

analysis on the nature of oscillations during EDM and productivity of pulses, and investigation of the origin of instability during EDM of conductive ceramics; and

determination of a suitable material for the wire tool and appropriate value of the discharge gap, as well as an analytical investigation on the nature of the electrical erosion phenomenon of conductive ceramics.

The developed approach can be extended to newly created composites and nanocomposites with threshold conductivities, and the monitoring system and adaptive control system based on acoustic emission are suitable for any EDM, where the issue of adequate control of processing remains open and unsolved.

Materials & methods

Equipment and materials for development of in-situ monitoring method

The first experiments were conducted on workpieces produced from easy-to-process conductive materials with a low melting point. Workpieces with dimensions of 200 × 20 × 16 mm of sorted and gauged corrosion-resistant, heat-resistant, and high-temperature steel of 12KH18N10T grade (analogue of AISI 321, specifications by interstate standard GOST 5949-75) [83, 84, 85, 86, 87, 88] and 200 × 16 × 16 mm of D16 aluminium alloy (AlCu4Mg1, analogue of AISI 2024) [89, 90, 91, 92, 93, 94] were used.

These experiments were carried out on an industrial 4-axis EDM machine, Seibu M500S (Seibu Electric & Machinery Ltd., Fukuoka, Japan), using a 0.25 mm diameter CuZn35 brass wire with deionised water as medium. The workpieces were fastened at the machine worktable overhang to determine the correlation between the amplitude of the signal and working zone distance. Samples with 2 and 10 mm width were cut off alternately. One single cut was performed for at least five times on each workpiece.

The developed in-situ monitoring system was verified on hard and wear-resistant alloys capable of maintaining their properties at temperatures of 900–1150 °C. The 12 mm long workpieces of hard alloys of VK6O (analogue of M05 by ISO, Table 1) [95, 96, 97] and T15K6 (analogue of P10 by ISO, Table 2) grades [98, 99, 100, 101, 102] were processed on an industrial 4-axis EDM machine AgieCharmilles CUT 1000 (GF Machining Solutions, Bern, Switzerland) using a 0.25 mm diameter brass wire with deionised water as medium. To induce wire tool breakage, the pulse frequencies were changed manually and gradually during processing up to the moment of wire breakage. One single cut was performed for at least five times on each workpiece.

Table 1

The chemical composition of hard alloy VK6O in %.

Component	WC	TaC	Co
%	92	2	6

Table 2

The chemical composition of hard alloy T15K6 in %.

Component	WC	TiC	Co
%	79	15	6

The CNC programs were prepared manually, and the offsets of the path were not taken into consideration. The parameters of the EDM machines were chosen following the recommendations of the manual [103, 104, 105, 106, 107, 108, 109]. The parameters for the first group of experiments are presented in Table 3.

Table 3

EDM parameters for machining of stainless steel (by the example of SKD-11), which were chosen for experiments.

No.	Designation	Description	Value, equivalent units of machine code
1	Vо	Open voltage, as higher is Vо as faster is processing [code], the high value increase voltage in the discharge gap and instability of processing, in range 30÷150	55÷65
2	Vg	Voltage in the discharge gap for its regulation [code], the high value increase machining speed, decrease the discharge gap and reduce the stability of processing, in range 1÷99	32
3	Sg	Wire feed speed [mm/min]	1÷2
4	Cs	A parameter of high voltage and resistance [code], in absolute values high voltage Vh = 240V, resistance Ro = 8.5Ω, standard impulse voltage Vimp = 80V	101÷102
5	To	Time of connection of the current source [code], ∼0.1μs in absolute values, in range 1÷15	8
6	Toff	Time of disconnection of the current source [code], ∼0.1μs in absolute values, in range 6÷50	6
7	Ths	A parameter for To [code], in range 11÷99, where ‘3’ is Ths - the second value for To for unstable machining, where ‘4’ is Ths' the third value for To	43
8	Tad	A parameter for Toff [code], in range 104÷550, where “05” is Tad – the second value of Toff for unstable machining, “3” is 3·10 = 30 for Tad’, when Tad was repeated N times, but unstable machining continues	305
9	Sw	Wire rewinding speed [code], in absolute values ∼35 mm/s, 10 min of machining consumes about 20 m of wire; standard brass wire bobbin contains 4.5 kg of wire or about 10500 m of wire; the low value decrease wire vibrations, the high value decrease conical deviation, as wire became finer toward lower guide due to electrical erosion	35
10	Fw	A tension of the tool electrode [code], in absolute values ∼3÷4N; the low value leads to larger barrel-shaped deviation for the parts with the external contour and saddle-shaped deviation for the parts with the internal contour, the high value leads to frequent wire breakage	30÷40
11	Fd	A dielectric pressure in nozzles [code], where “2” is high pressure, “45” is a frequency of the pomp 45Hz, the parameter can be regulated on the machine (100÷160, 200÷260)	245
12	N	Switching method between direct and reverse polarity in an unstable cutting mode [code]: switching between To-Toff and Ths-Tad, the stable cutting is with To-Toff mode, for the unstable cutting system switches on Ths-Tad mode and then turn back to To-Toff mode; if it occurs N times, the CNC-system switches constantly for Ths'-Tad’ mode (00÷99)	68

Note that the workpieces were immersed in the dielectric medium for 10 min before processing to exclude the influence of thermal shrinkage of materials during processing [110]. The level of the dielectric medium in the working tank was set as 1–2 mm above the workpieces. The nozzle of the upper guide was set as 2–3 mm higher than the dielectric level.

The experiments were conducted with the use of a pulsed solid-state laser of diode pumping U-15 (RMI, Moscow, Russia) for comparison of the results with those of the monitoring of other impulse precision processes. The laser belongs to the one-mode vanadate laser type Nd:YVO4 with a wavelength l = 808 nm [111, 112, 113, 114, 115]. A plate of 200 × 60 × 4 mm of high-quality carbon steel 08 (analogue of AISI 1008, Table 4) was used as a workpiece for laser ablation.

Table 4

The chemical composition of carbon steel 08 in % (GOST R standards).

Element	C	Si	Mn	S	P	Cr	Ni	Cu	As	N
%	0.05-0.12	0.17-0.37	0.35-0.65	Up to 0.04	Up to 0.035	Up to 0.10	Up to 0.30	Up to 0.30	Up to 0.08	Up to 0.008

EDM of conductive ceramics

The 5 mm long × 40 mm diameter samples were produced by the SPS method on a machine for a hybrid current–assisted SPS [116, 117] KCE H-HP D 25-SD (FCT Electronic GmbH, Munich, Germany). The discs were sintered from alumina oxide powder Al2O3 with additions of silicon carbide in the form of whiskers SiCw and 30% of conductive titanium carbide TiC in the carbon dies [118, 119]. The control sample was made of tool ceramic VOK-6O (analogue of K01, ISO) [120, 121, 122], which consists of alumina oxide Al2O3 and conductive TiC.

The significant difference between the ceramics was in the conditions of production. The first sample was produced in the laboratories of MSTU STANKIN with the maximum achievable homogeneity of powders before hybrid sintering [59]. The main SPS parameters are presented in Table 5 [123,124]. The second sample was produced industrially and was taken from a batch of cutting inserts for lathe machines.

Table 5

Sustainable parameters of SPS assisted by electrical current for production of Al2O3-SiCw-TiC conductive ceramic nanocomposite.

Max temperature [°C]	Heating rate [°C·min−1]	Hold time [min]	Maximum pressure [MPa]
1780	100	5	40

The samples of conductive ceramics were processed by an EDM machine, AgieCharmilles CUT 30P (GF Machining Solutions, Bern, Switzerland). The acoustic parameters of processing were recorded.

A pre-analysis of the specific electrical conductivity of samples showed that their electrical resistance is 40–50 times higher than that of hard alloys. Moreover, the conductivity of ceramics is the minimum possible for materials to be conductors.

A material is considered a conductor if the specific electrical resistance ρ < 10−5 Ω m and a dielectric if ρ > 108 Ω m. It should be noted that the electrical resistance of good conductors can be as low as 10−8 Ω m, and for the best dielectrics, it can exceed 1016 Ω m [125, 126]. Thus, the ceramic samples of VOK-6O and sintered Al2O3+SiCw+TiC can be called conductors with threshold specific electrical resistances, which are suitable for EDM [127, 128, 129, 130].

The value of ρ for conductive ceramics is between the value for graphite of 8.0 × 10−6 Ω m and for nichrome (NiCr) of 1.12 × 10−6 Ω m [131]. A single-crystal graphite is anisotropic with electrical properties close to those of metals along the basic plane of the crystal lattice, when the conductivity in the perpendicular planes is a hundred times less [132]. However, graphite is more conventional in conditions of real manufacturing than NiCr. The EDM parameters for processing ceramics were chosen close to those for processing graphite with the least possible discharge gap. The linear speed of machining was accelerated to trigger short circuiting. The complexity of the work preparation procedures during the experiments is as mentioned above.

Characterisation of samples

The electrical resistance of samples was measured and controlled using a four-wire connection with a separate power source and a dual-channel nanovoltmeter, Keithley 2182A (Keithley Instruments, Solon, Ohio, USA), according to ASTM C611 standard. The specific electrical resistances ρ of the materials are presented in Table 6.

Table 6

Specific electrical resistance of the materials ρ with the environmental temperature +20 °C (* given for reference).

No.	Material	Grade	Short chemical description	Analogue (ISO/AISI)	Specific electrical resistance ρ [Ω⋅mm^2m]	Specific electrical resistance ρ [Ω·m]
1	Stainless steel	12Kh18N10T	X10CrNiTi18-10, X6CrNiTi18-10KT	AISI 321	0.746	7.46 × 10−7
2	Aluminum alloy	D16	AlCu4Mg1	AISI 2024	0.028	2.8 × 10−8
3	Brass alloy	-	CuZn35, 65/35 brass	ISO 426-1	0.065	6.5 × 10−8
4	Single-carbide hard alloy	VK6O	WC+Co	ISO M05	0.099	9.9 × 10−8
5	Two-carbide hard alloy	T15K6	WC+TiC+Co	ISO P10	0.079	7.9 × 10−8
6	Conductive ceramic of cutting insert	VOK-6O	Al2O3+TiC	ISO K01	3.510	3.51 × 10−6
7	Conductive ceramic nanocomposite sintered in MSTU Stankin [59]	-	Al2O3+SiCw+TiC(30%)	-	3.150	3.15 × 10−6
8	Nichrome*	-	NiCr	-	1.120	1.12 × 10−6
9	Graphite*	-	C, 1.CB.05a by Strunz	-	8.000	8.0 × 10−6
10	Carbide titanium*	-	TiC	-	0.600	6.0 × 10−7
11	Tungsten*	-	W	-	0.049	4.9 × 10−8

The microscopy of the samples were analysed by an Оlympus BX51M (Ryf AG, Grenchen, Switzerland) microscope with 20k× magnification and a Tescan VEGA3 LMH (Tescan Brno s.r.o., Brno, Czech Republic) scanning electron microscope (SEM) with 1M× magnification.

The spectrometry analysis of the samples was obtained by an X-ray diffractometer using CuKα radiation, PANalytical Empyrean Series 2 (Malvern Panalytical, Almelo, Netherlands). The phase composition was determined using PANalytical High Score Plus software and ICCD-2 database.

A high-precision profilometer, Hommel Tester T8000 (Jenoptik GmbH, Villingen-Schwenningen, Germany), was used to control the roughness of the machined surfaces.

Three-dimensional models and surface profile diagrams were obtained by a certified high-precision 3D microscope, MicroCAD Premium+ (GFMesstechnik GmbH, Teltow, Germany), with an error of ±1 μm.

The cross-sections of the samples for metallography and microscopy were prepared according to the standard probe preparation methods with the use of high-precision equipment: hydraulic press Opal 410, grinding machine Jade 700, and polishing machine Saphir 300 (ATM, Haan, Netherlands).

Results

Development of in-situ monitoring system

The in-situ monitoring system based on vibro-acoustic signals was developed and adopted for the needs of EDM according to the described features below [83, 133, 134, 135].

Investigations [136, 137, 138] showed that the oscillations of the wire tool during processing might be the main reason for instability when the forced oscillations of the wire tool progresses to self-oscillations.

The first experiments showed that the oscillation frequency of the tool in a groove with a substantial thickness of workpiece is close to the frequency established in [139, 140] for mechanical machining.

A diagram of applied forces in the working zone during EDM [141, 142] is depicted in Fig. 3, where

Fig. 3

A diagram of applied forces in the working zone during wire electrical discharge machining: (1) is a wire tool, (2) is a workpiece, (3) is a upper nozzle, (4) is a lower nozzle, (5) is rollers [74].

A is the amplitude of the wire in the inverse direction to the wire feed,

Δ is the discharge gap,

H is the thickness of workpiece,

H’ is the distance between nozzles,

F is the force of wire tool guiding,

ΣF is the summation of forces of discharge impact on wire tool,

ΣF is the summation of forces of wire electrode holding in the nozzles,

ΣF is the summation of forces of wire tension and rewinding,

F is the force of friction,

S is the wire feed speed,

S is the wire rewinding speed, and

W is the torsional moment of rewinding rollers.

If the mass of the wire sections between nozzles and the workpiece (h1 + h2), which is varied in the range of 3–5 mm (H’ - H), and the mass losses on tool-electrode wear [143, 144, 145, 146] are neglected, then the amplitude of oscillation of the wire tool may be expressed as a complex amplitude of the harmonic signal [147, 148, 149, 150, 151]:where A0 is the amplitude in the direction of wire feed, which is limited by the discharge gap Δ and is less than Δ, A ≤ A0 for stable EDM, is the damping coefficient expressed as a complex number in the form a + bi, and τ is the period of oscillation T.where is the index of excited oscillations (a complex number), and is the coefficient of dielectric medium resistance. In this case, iswhere h is the ratio between H and H:

H0 is the critical value of workpiece thickness, which is a characteristic of stable EDM to guarantee sufficient quality and geometric accuracy of the processed surfaces. H0 = 80–100 mm is the commonly known industrial recommendation for a wire diameter d = 0.25 mm and a water-based dielectric medium, whereas H0 = 8–10 mm for conductive ceramics. m is the mass of wire on the length H, while is a complex number of the ratio between K and K:

Here, K is the stiffness of the system, and K is the stiffness of the system for stable EDM. For a particular case, K can be expressed as

The coefficient of dielectric medium resistance [kg/s] is approximately correlated and inverse directed to F. Hence, depends on the density of the dielectric medium ς, strength area Π, and wire feed speed S. It can be expressed as

Thus, ®β depends on the ratio between the workpiece thickness and stiffness of the system, the process parameters, and viscosity of the working medium in the groove; it can vary in the limits of |1|.

A criterion for the self-oscillation process origin iswhere λ is the logarithmic decrement characterising the damping ratio of self-oscillation and, as a rule, is a resonant process describing the decrease in the amplitude of the oscillation process. It is equal to the natural logarithm of the ratio of two successive amplitudes of an oscillating quantity A at the same side:

The period of self-oscillations T depends on the mass of the wire on the length H and stiffness of the system, which depends on the wire tension:

If |q| < |μ|, then self-oscillations (including resonant) do not occur. If |q| > |μ|, then self-oscillations are generated, and the higher the intensity, the higher the excess of |q| over |μ|.

Note that μ is higher for an oil medium and lower for a dielectric medium based on water. Thus, EDM in an oil medium gives a better geometric accuracy in processing critical surfaces of parts with thickness of the workpiece larger than 80–100 mm for conventional conductive materials and larger than 8–10 mm for conductive ceramics.

The developed in-situ monitoring system [74] consists of accelerometers (which were placed at the upper guide and fastening system), preamplifiers and amplifiers (VShV003, Moscow, Russia), an analogue–digital converter (E440, Moscow, Russia), and a signal-recording device. The signals received from the accelerometers depend on the nature and intensity of the vibrations during processing, which depend on the summation of forces of discharges ΣF that rely on the EDM parameters.

The first experiments showed that the upper accelerometer gave a more informative symmetric signal than the lower one. Further, only the signal from the upper guide was considered. The vibro-acoustic signals were recorded with the following intervals: 1 min, 30 s, and 5 s before the end of processing. A spectral analysis was conducted for frequencies up to 10 kHz.

Validation of the developed system on conductive ceramics

The features of oscillation signals during conventional metalworking are relatively well-known [152, 153, 154, 155]. The recent studies completed this knowledge with the data concerning the control of electric discharge based on electromagnetic and vibro-acoustic emissions [156, 157, 158, 159, 160]. Further studies of the signal features during EDM of conductive materials showed a particular behaviour of the signal, which is suitable for the development of the monitoring method [75, 133, 134]. The developed adaptive control system based on acoustic emission could function along with a system based on electrical parameters, or even replace it totally. The obtained results demonstrated that vibration monitoring might provide relevant data about the state of processing. It may prove to be essential for controlling more complex machining processes for new classes of materials such as conductive ceramic nanocomposites. Thus, two types of conductive oxide ceramics were selected:

a 5 mm long × 40 mm diameter cylindrical sample of sintered ceramic nanocomposite Al2O3+SiCw+TiC with 30% of TiC for providing better electrical conductivity, and

a 5 mm long square cutting insert for a lathe machine tool made of a conductive tool ceramic Al2O3+TiC grade of VOK-6O (analogue of K01 by ISO).

Fig. 4 presents a portion of records of the root-mean-square (RMS) values of vibration and current signals during the EDM of Al2O3+SiCw+TiC ceramic. Note that the peaks and dips of the signals are opposed as marked by arrows.

Fig. 4

RMS diagrams during wire electrical discharge machining of the sample made of Al2O3+SiCw+TiC nanocomposite: (1) is the recorded vibroacoustic signal; (2) is the recorded operational current.

Short circuiting is associated with the dips of the vibration signal and sharp rises of the working current signal. The high electrical resistance of the conductive ceramics provided moderate alterations of the voltage and current during short circuiting. It did not permit determining these changes by the CNC system and applying timely corrections of the EDM parameters.

Studies [161, 162] showed that the growth of energy and efficiency of pulse acting on the surface to be machined is followed by a gradual increase in the RMS amplitude of vibration signal at a high-frequency band.

In the beginning, the segment of the wire probably was overheated and partly melted, and a subsequent breakage followed. At the same time, the wire rewinding rate was sufficient and adequate to continue processing. The observed quality of the machined surface was not sufficient because of the conducted experiments on optimization of EDM parameters during machining.

Fig. 5 presents the roughness of the machined surface, where R at a length of 1.0 mm is more than 65 μm.

Fig. 5

The machined surface of the sample made of Al2O3+SiCw+TiC nanocomposite after WEDM: (a) is microrelief of the surface; (b) is three-dimensional surface profile in the upper corner.

The process of EDM of the sample made of VOK-6O conductive ceramic was also forced to have frequent short circuits. The received record of VOK-6O sample processing shows similar slumps of the RMS amplitude of the signal at the moment of growth of the discharge current.

Fig. 6 shows the RMS values of vibro-acoustic signals in the octave band of 16 kHz and current signals. The lowest values of vibration signals are recorded at 0.35 s of short circuit when the highest peaks of the working current were observed. At the moment of short circuits, the vibration signal had a significantly different spectrum characteristic than during the stable phase of machining. This type of spectrum is associated with frictional contact [163, 164, 165, 166, 167].

Fig. 6

RMS diagrams at the moment of the short circuit during WEDM of the sample made of VOK-6O conductive ceramic composite: (1) is the recorded vibroacoustic signal; (2) is the recorded operational current.

Fig. 7 depicts the spectra of vibration signals recorded during the stable EDM and at the moment of short circuits. The amplitude of the vibro-acoustic signals during the stable EDM was many times higher than that during short circuits. The characteristic of the signal during short circuits was similar to that of the signal during stable EDM, which could be due to the electrodes' physical contact and friction. It is especially visible in the areas of spectral peaks associated with own frequencies of the ‘tool with guide nozzles–workpiece with fastening elements–machine’ system. During the detected friction contact, the vibro-acoustic amplitude was low and primarily defined by the speed of the relative movement of electrode surfaces and their hardness [137, 168, 169]. The wire rewinding speed was approximately 150 mm/s. The characteristic of the amplitude was different from that of the signal related to the work of the pump and EDM machine drivers. This can be helpful in detecting the moment of contact loss between electrodes during the reversal motion of the guides when the pulse generator is switched off. In other words, the CNC system should immediately dispatch a signal to switch the pulse generator off and switch the reverse mode on to break the contact between electrodes. Afterwards, the wire tool should be re-approached and the generator switched on.

Fig. 7

Spectra of the vibroacoustic signal during WEDM of the sample made of VOK-6O conductive ceramic composite: (a) is in the low frequency range; (b) is in the high frequency range; (1) is for regular processing; (2) is for short-circuit.

Discussion

Nature of oscillations during EDM

In the beginning, it was decided to focus on the moment of the part separation from the workpiece. Changes in the specific noise accompanied the process—from the monotonic buzzing sound to the higher-pitched sound gradually ascending to the moment of brittle fracture.

The experiments showed that the amplitude of the vibro-acoustic signals increases gradually when approaching the moment of part separation from the workpiece. Then, a series of characteristic peaks of the vibro-acoustic signals is observed. It provides an indication about the final bridge weakening between the part and workpiece. It slowly starts the motion under its mass accompanied by the discharge current pulses. The recorded signal demonstrates plastic and brittle stages of bridge rupture [10, 11, 29] at that moment. Additionally, the wire tool can be squeezed between the part and workpiece, and a series of short circuits, which causes the formation of burn defects at the machined surface [28], occurs.

During the experiments, it was established that even negligible changes in the part weight are reflected in the vibro-acoustic signals. For example, a 24.5 g weight and a 10 mm width sample of stainless steel show instability, which was recorded by a significant increase in the amplitude of the vibro-acoustic signals 5 s before separation. In addition, a 4.28 g weight and a 10 mm width sample of aluminium alloy show the beginning of separation moment, i.e. plastic deformation, which leads to wire tool clasping during the final bridge rupture 2 s before separation. This phenomenon was recorded in the broad frequency range of the signal and was also characterised by instability in the low frequencies. Thus, it was decided to use a frequency range higher than 4 kHz for monitoring and further analysis.

Note that the effective amplitude in the octave-frequency band is more stable than the high-frequency spectra of the vibro-acoustic signals because of averaging by the wider frequency range. As the moment of rupture approaches, the effective amplitude increases by approximately two times in the octave-frequency band of 4 kHz and, at the same time, the effective amplitude increases by 1.5 times in the octave-frequency band of 8 kHz [74]. According to these changes in amplitude, the processing state can be evaluated without access to the working area and by visual demonstration, which can be understood even for the inexperienced operator and used for adaptive control.

During the final rupture of the part from the workpiece when under the influence of its weight and discharge current pulses, the part starts its movement relative to the workpiece due to the natural process, which is similar to the stress–strain curve for construction materials [170, 171, 172], and the signal was recorded (Fig. 8). The movement disturbs the symmetry of the wire tool position in the formed groove, alters the discharge gap, and the subsequent contact of the wire tool and workpiece occurs inevitably.

Fig. 8

High-frequency spectra of the vibroacoustic signals during WEDM of the samples made of D16 aluminum alloy: (1) is at the moment of wire tool penetration; (2) at the moment when the discharge gap Δ is formed; the inset of RMS-t diagram presents the spectrum of the vibroacoustic signal at the moment of wire tool penetration.

It was detected that separate components of the spectrum on the last seconds of the records increase by ∼4 times in the frequency band of 8 kHz for aluminium alloy. The described changes in the vibro-acoustic signals are sufficient for developing the in-situ monitoring system and system of adaptive control to prevent undesirable short circuits and increase quality of the machined surfaces.

As it was determined, an excess in the amplitude of the vibro-acoustic signals shows the instability of the discharge gap Δ, which can be explained by the presence of erosion products in the discharge groove. It gives a basis for interactive changes in the EDM parameters during processing (adaptive control) especially at the moments of wire penetration and at the end of the operation to decrease the disturbing influences on the wire tool. The stability of the vibro-acoustic signals occurs when the discharge current pulses are uniformly distributed over time, the processing area of the workpiece is less than 70–100 mm thickness, and a sufficient, timely flushing of the erosion products from the discharge gap is ensured.

The influence of the wire tool tension force F on the vibro-acoustic signals during the experiments was evaluated by varying its range at ± 0.5 N (∼5 equivalent units of machine code). On the spectra of the vibro-acoustic signals' parameter records, on the length of stable machining of 12Kh18N10T stainless steel samples with a wire force tension F at a value of 3 and 4 N, there is an RMS signal dispersion with time (Fig. 9). However, with the decrease in tension, the tendency towards absolute values of parameters is traced for the range of high frequencies. The variation of F has a similar tendency for stainless-steel samples at 5 and 60 s before the end of processing (Fig. 9a). The characteristics of the diagrams for the two types of materials are different. This can be explained by the different elastic properties of the materials (Young’s modulus): the higher the elasticity of the system, the higher the sensitivity of the developed monitoring system (Fig. 9b). A decrease in the stiffness of the technological system components obstructs oscillation propagation [173, 174, 175, 176, 177]. It can be an advantage in the application of the developed system for processing of materials with advanced mechanical properties such as conductive ceramic composites and nanocomposites.

Fig. 9

RMS of octave spectra at the frequency band of 8 kHz with a variation of wire tension F ± 5N from the value of wire tension of stable WEDM: (a) for samples made of 12Kh18N10T stainless steel; (b) for samples made of D16 aluminum alloy.

Note that the RMS value of amplitude at 5 s before the end of processing exceeds the RMS value of amplitude at 60 s before the end of processing by at least 2 times at the frequency band of 8 kHz. The difference in RMS values decreases as the wire tension grows, demonstrating the lesser dependence on the weights of the samples, which differ by ∼6 times, as the stiffness of the system grows.

As it is known, there are three types of vibrations in technological systems (Fig. 10):

Fig. 10

The types of oscillation.

Free natural oscillations, which are characterised by a specific frequency ω determined by the properties of the technological system.

Forced oscillations, which are characterised by a periodically acting external force F, in the case of EDM. The frequency ω is determined by the frequency of discharge impulses, and the amplitude is related to the resonance phenomenon.

Self-oscillations, which are characterised by the internal energy of the system with a fixed frequency ω close to the frequency of natural oscillations ω and a fixed amplitude. They are associated with the low stiffness of the system and fluctuation of the acting force F. The frequency of vibrations increases with an increase in the system stiffness and decreases with a decrease in the workpiece thickness.

Because low-frequency vibrations (10–100 Hz) are associated with the appearance of low-frequency noise related to the work of the equipment and guiding bars and high-frequency oscillations (1500–8000 Hz) are associated with the wire tool vibrations and high-frequency noise (close to whistle) accompanying processing, it is recommended that EDM takes place outside the specified zone to exclude the phenomenon of resonance [178]:

Then, the process may be characterised as a process of stable cutting.

Three groups of forced oscillations can be distinguished during EDM:

Oscillations associated with the frequency characteristic of the electric current pulses discharged on the surface to be machined. The wire tool has larger oscillation amplitudes due to its lower stiffness and mass.

Oscillations caused by the design features of the EDM machine and drives of its guides.

Any other external oscillations.

Thus, based on the items mentioned above, we can conclude that

For the lower value of wire tension, flushing of the erosion products is hampered by wire bending and its higher amplitude due to the low frequency of vibrations, which have a complex characteristic, when ω is greater than ω.

For the higher value of wire tension, flushing of the erosion products is adequate, but the stiffness of the system is higher. Moreover, the wire amplitude is lower and the vibration frequency grows when ω is less than ω.

For the value of wire tension associated with stable cutting, the frequency of forced oscillation ω may be compensated by the frequency of self-oscillation ω.

During the disturbance of stable machining in the initial motion of the separating part, the influence of the vibro-acoustic signal parameters is less noticeable than that at the end of the operation. It may be explained by the phenomenon when the additional forced disturbances defined by part mobility mainly influence the increase in vibro-acoustic signal amplitude.

Productivity of pulses and origin of instability during EDM of conductive ceramics

The optimisation of EDM aims to ensure its maximum productivity, efficiency, and surface quality. This demands the maintenance of constant conditions or the regulation of process parameters, such as the discharge gap Δ, the concentration of erosion products in the gap, the temperature of the working fluid, and its flow rate. The gap Δ is the main parameter as it is mostly responsible for determining the quality of EDM, which occurs within the discharge gap [179, 180]. A small increase in Δ may change the breakdown conditions or even interrupt the discharge; a decrease in Δ impairs the yield of erosion products, reduces the productivity, promotes slag buildup on the electrodes, and facilitates short circuits. Without automatic regulation of the gap, EDM is ineffective [1, 12, 72]. For better regulation, Δ must be maintained above the value at which short circuits appear, and below the value at which breakdown is impossible and idling pulses appear. In maintaining an optimal gap, the rate of particle formation M in the discharge gap has to be equal to the rate M at which the particles (debris) leave the gap. The rate M is a function of the particle concentration γ:

When M and M are unequal, the change in concentration iswhere Q is the volume of the discharge gap and

For stable processing,

However, it is complicated to ensure a constant γ due to the many random factors that act on the working zone [181]. Any fluctuations that arise must be promptly eliminated by control signals that change the EDM parameters through diagnostic information. Fig. 11 shows the dependence of the rate M on the discharge gap Δ with different exit rates of the particles:

Fig. 11

Dependence of the rate of particle formation M on the discharge gap Δ at various exit rates M of the erosion products.

Analysis indicates that the maximum machining rate M decreases as M reduces because of the deterioration in the discharge of erosion products as the electrode is introduced in the material, and the number of working pulses is reduced [182, 183, 184].

Fig. 12 illustrates the dependences of the machining rate M, number of working pulses n, number of idling pulses n, and number of short circuit pulses n on the discharge gap Δ. Analysis shows that the gap corresponding to the maximum machining rate M is higher than the gap corresponding to the maximum working pulses (n). It may be explained by the excess of erosion products at n and the presence of short circuit pulses, which destabilise the process [133, 134, 185, 186].

Fig. 12

Dependence of the rate M and the number of pulses n on the discharge gap Δ [74].

Therefore, the in-situ monitoring and control systems aim to maintain the efficiency of pulse utilisation in the range of

It may be calculated as the ratio of the number of working pulses n that are involved in the process of material ablation to the total number n of pulses:

While ψ is useful for in-situ adaptive control, its use is hindered by the complexity and inertia of the measuring instruments required. Besides, it is difficult to assess the efficiency of EDM with respect to ψ [161, 187]. When the working fluid is contaminated with erosion products, the energy of the individual pulses is not entirely consumed in the process of metal ablation; some of the energy is consumed in the destruction of the erosion products [188, 189, 190]. Thus, the number of working pulses is assessed from the total width of the discharge pulses or their height at the front and back of their profiles. It leads to an imprecise assessment of the efficiency of EDM.

A more precise estimate may be obtained by relating the efficiency to the ratio of the useful energy consumed in the destruction of material and the total energy that enters the working zone in the form of discharge pulses. In the operation of independent generators, the energy of the discharges is proportional to the effective discharge current (I). A current sensor may be used to estimate this energy.

The parameters of the vibro-acoustic signals recorded by the accelerometer mounted on the workpiece were used to assess the useful energy. Monitoring based on acoustic emission is used for tool assessment in conventional machining processes [72, 73]. However, the widespread use of vibro-acoustic signals for in-situ monitoring of cutting and friction is difficult because a closed system with a nonlinear dependence of the vibro-acoustic signal on the contact load is formed with the direct contact of the tool and workpiece or contact of the indenter and counter body [152, 153, 191, 192, 193]. This nonlinear dependence is not observed when energy pulses act on the workpiece. Hence, the dynamic system resembles a linear model better, where the dynamic relation between the load source and the workpiece is permanent. In other words, the model of the dynamic system is significantly simplified, and its use in the in-situ monitoring and regulation of machining by high-energy fluxes is simplified [168, 189, 194, 195].

Let us consider the results of vibro-acoustic monitoring of another non-contact precise machining method, which allows the processing of materials for industrial manufacturing applications (micromachining, laser graving, and grinding). Laser treatment is close by its nature to the processes of material removal during EDM because in both cases, there is no mechanical contact between the ‘tool’ and workpiece [196]: the material removal occurs under thermal influences, initiated by pulses of concentrated energy flows. The mechanism of material removal during EDM is related to the processes of sublimation or evaporation [180, 197, 198, 199, 200, 201], similar to the mechanism of material removal during laser treatment, which is also called laser ablation [202].

The previously conducted experiments using a pulsed solid-state laser of diode pumping have shown that the vibro-acoustic signals during the processing of a workpiece made of high-quality carbon steel 08 (analogue of AISI 1008) steadily rise as the laser power rises and the volume of removed material increases [133]. The amplitude of vibro-acoustic signal A depends on the power of the impulses, and the experimental dependence of productivity of single pulses is directly proportional to their power. Therefore, the monitoring of vibration signals can be recorded or evaluated visibly and timely in contrast to the evaluation of the abstract concept of current pulses productivity.

Based on data published in [74, 133, 203], it can be concluded that there is a direct dependence between the discharge productivity of pulses and the usable energy proportion, and an inverse dependence with the duration of pulses. In the case of machining of materials with low or threshold conductivities such as conductive ceramic composites and nanocomposites, the duration of pulses will be so small such that it can be ignored in the calculations. Therefore, the main reason for EDM instability when the value of the workpiece thickness H is less than the critical value H0 (H < H0), which characterises stable EDM processing, can be the strong correlation between the usable energy proportion and concentration of erosion products in the channel of observation. When H > H0, there should be at least two main reasons of process instability: the concentration of particles in the volume of the discharge gap Q, which is dependent on the main technological parameters, and self-oscillations of the wire, when the stiffness of the system K is less than the critical value K0.

A dynamic model of the concentration of erosion products and its influence on the vibro-acoustic signal parameters was presented in previous works [74, 203]. It was shown that although EDM is a non-contact processing method, the developed technique of diagnostics based on vibro-acoustic signals may be used as a reliable in-situ monitoring system as in the case of many other processing methods based on mechanical contact between two components [152, 194, 204, 205, 206, 207, 208, 209].

Electrical erosion phenomenon of conductive ceramic composites and nanocomposites

A similar graph of vibro-acoustic signals was obtained during EDM of the samples made of VOK-6O. The vibro-acoustic signal related to the short circuits was accompanied by the indicated presence of frictional contact between the electrodes under voltage. As a result, the machined surfaces have traces and defects associated with the formation and development of brittle cracks related to excessive heating in the working zone. Fig. 13 shows a photograph of the sample made of VOK-6O conductive ceramic, where the mechanical destruction of the edges around the formed groove of the workpiece can be found. The formed cracks cause chipping of the ceramic sample, with the debris size exceeding the width of the groove of ∼0.26 mm in a few times. It was detected that the formed brass globules condensed on the machined surfaces of the groove during short circuits. This could be due to the inappropriate combination of materials for the wire tool and workpiece. The melting and boiling points of the components of these materials differ significantly. In this case, the melting and boiling points can be used as references for estimating the initiation temperature of the sublimation phenomena.

Fig. 13

A picture of the machined groove of the sample made of VOK-6O conductive ceramic composite after WEDM accompanied by series of short circuits.

According to study results [169, 210, 211], sublimation of brass under vacuum conditions starts with zinc as the most fusible material at 450 °C (the melting point of zinc is 419.5 °C, whereas the melting point of copper is 1085 °C). At the same time, a nanoporous structure is formed in the remaining copper [169]. The studies showed that when brass sublimates under normal conditions in the presence of an air medium, a whitish needle-shaped film forms for zinc oxides at a temperature of ∼300 °C, a single-layer film of CuO at a temperature of 300–375 °C, a bilayer film of black CuO, and a dark red Cu2O at 375–1100 °C.

The absence of traces of oxide films on the surface of the wire after EDM (Fig. 14) indicates that the conditions of the formed discharge channel are close to the conditions of brass sublimation in a vacuum. That is, an active medium of interaction with materials is absent. The presence of C and Si in Fig. 14b can be explained by the use of epoxy resin for filling the probe of the sample for the analyses.

Fig. 14

Characterization of wire tool after WEDM of the sample made of VOK-6O conductive ceramic composite: (a) distribution of all chemical elements along the diameter; (b) X-ray spectrometry analysis; (c) SEM of the wire tool break point with the presence of the erosion traces.

Therefore, with a discharge channel, the conditions in the inter-electrode gap are similar to the conditions of laser sublimation in a vacuum. However, it differs in the thermal source: it is the breakdown spark during EDM that initiates the occurrence of a conducting discharge channel. Hence, the sublimation, which begins with zinc, with increasing temperature progresses to the sublimation of copper. By the time the temperature reaches the critical value for the initiation of sublimation of alumina in the inter-electrode space, a relatively dense gas medium is formed. It leads to the formation of a plasma cloud (a rapidly expanding low-temperature plasma cloud) and prevents further sublimation. The surface of the electrodes partially adsorbs the gaseous material of the plasma, which also prevents continued sublimation. The results of the SEM of samples made of Al2O3+TiC(30%) conductive ceramic with similar conductive properties as the Al2O3+SiCw+TiC(30%) sample produced by the SPS method assisted by electric current (Fig. 15) show a picture of the distribution of the chemical elements before and after EDM. Fig. 15b illustrates the chaotically distributed components of the wire electrode and workpiece on the machined surface of the ceramic. The X-ray spectrometry analysis of VOK-6O conductive ceramic (Fig. 16) can also corroborate the above-mentioned assumption.

Fig. 15

SEM analysis of the chemical elements distribution of the samples made of Al2O3+TiC(30%) conductive ceramic nanocomposite: (a) cross-section of the sample before WEDM; (b) machined surface after WEDM, where (1) is nanoporous structure of the main components of the sample, (2) and (3) adsorbed easy-to-melt components of the wire tool.

Fig. 16

X-ray spectrometry analysis of the machined surface of the sample made of VOK-6O conductive ceramic composite after WEDM.

Thus, the temperature in the inter-electrode gap reaches that of TiC sublimation with the melting point of ∼3140 °C. However, the formed gas phase of alumina, zinc, and copper adsorbed by the surface of the nanoporous structure of titanium carbide prevents the removal of material in sufficient volume (efficient mass loss) at the formed groove. Then, the spark is re-initiated, and the cycle repeats.

The nanoporous structure (the recast layer + heat-affected zone) after sublimation occurs at a depth of 15–20 μm with a pore size of 2 ± 1 μm according to Mahdavinejad [170], Kazanskiy [210], and Ren [211]. In the conducted experiments with EDM, it can be said that the defects of the machined surface with a change in its chemical composition are determined to a depth of up to 4 μm [10, 11, 29].

The multiple cases of short circuits can be explained by the fact that the electrical resistance of the workpiece material is approximately 100 times higher than that of efficiently processed materials such as aluminium, steel, and hard alloys. These materials have different corresponding values of discharge gaps.

Thus, it can be assumed that during EDM of conductive ceramics, contact between the electrodes in the spark gap occurs due to a significant loss in the wire material with a sufficiently long pulse. The machining proceeds according to the scenario described above, when the easy-to-melt components of the wire electrode settle on the structure formed from the refractory titanium carbide. The wire continues rewinding and renews uninterruptedly the electrode material at the inter-electrode gap and feeds forward. At a particular moment, the wire material on the surface of the workpiece accumulates such an amount that provides partial adhesion of the electrodes in the spark gap. There is constant contact between the electrodes due to the filling of the inter-electrode space with the products of erosion, leading to short circuits without the possibility of having the wire off from the workpiece to re-establish the spark gap. Consequently, the wire electrode breaks due to the impossibility of further rewinding and the CNC system of the machine gives a signal on the interruption of processing and shows an error code in the display.

It should be noted that the current monitoring of the conditions in the inter-electrode space based on the electrical parameters remains blind to the processes happening between the two electrodes before wire breakage. This leads to defects in the machined surfaces and machining interruptions.

Discharge gap and wire tool material for EDM of conductive ceramics

The experimentally determined value of the spark gap for conductive ceramics was in the order of 0.005 mm, which is 15 times less than the recommended value for aluminium alloys of 0.075 mm. The value of 0.005 mm is more difficult to control by the drives of the EDM machine with a positioning accuracy of 2–3 μm; environmental conditions may also affect the accuracy of processing at rough modes. It should be noted that the difficulties in slot forming are related to the inappropriate combination of electrode materials and insufficient space for wire oscillations under pulses.

During the EDM of VOK-6O conductive ceramic, slender metallic fibres were formed on the globules of easy-to-melt components of the wire after short circuiting and at the moment of contact loss between the electrodes. These fibres continue to keep the contact between the electrodes at the reverse motion of the wire for re-establishing the discharge gap and prevent further processing.

Fig. 17 shows a picture of the formed groove, where solidified globules and fibres of 6–10 μm in diameter can be observed.

Fig. 17

Optical microscopy of metal fibers of easy-to-melt components formed in the machined groove of VOK-6O conductive ceramic composite after WEDM: (1) is fibers of 6–10 μm in plane.

The TiC content in the samples made them conductive because the electrical resistance of titanium carbide is 6.0 × 10−7 Ω m [212, 213], which is higher than that of CuZn35 brass. However, the temperature stability field of TiC reaches 3140 °C, which is highly resistant to molten low-melting metals and metals such as copper, aluminium, brass, cast iron, and steels [214, 215]. The primary component of the matrix is alumina, which is known as an excellent dielectric [216, 217], but some researchers consider it an n-type semiconductor with a permittivity of 9.5–10.0 and an electric strength of 10 kV/mm [218]. The melting point of aluminium oxide is 2044 °C, whereas the boiling point is 2977 °C [219, 220, 221, 222]. Note that the boiling point of any alloy depends on its components. The first component of the brass wire that reaches the boiling point of 907 °C is zinc; next is copper with the boiling point of 2590 °C, and thus, the boiling point of brass is 2590 °C [223]. Because the melting point of brass is 850–950 °C [224, 225] and the boiling point is 2590 °C, the most rational option would be the use of a 0.25 mm diameter commercially available tungsten wire. The melting point of tungsten is 3410 °C and its boiling point is 5660 °C; the electrical resistance is 4.9 × 10−8 Ω m for 20 °C [226, 227].

Conclusions

The presented research showed the functional opportunities of monitoring based on acoustic emission in searching for the optimal EDM factors for new classes of conductive ceramic nanocomposites. The developed method can provide a better quality of machined surface and geometric tolerance of product and prosper for the broad application of monitoring and control methods in real production. The use of the developed monitoring method appears reasonable for the improvement of machined surface quality and performance during conventional EDM. The method can be adopted for

workpieces with large thicknesses (more than 150–200 mm),

workpieces with a remotely placed working zone (more than 100 mm from a nozzle),

workpieces with uneven internal structures such as a set of hollow tubes, and

workpieces made of materials with threshold conductivities and high heat resistances (heat resistances higher than 1500 °C and melting points higher than 2500°С).

The developed monitoring system can be used for effective vertical positioning of the wire tool or the evaluation of wire tool amplitude during processing. It was proved that vibration monitoring allows controlling the moment of wire approach and its breakage.

The EDM experiments of conductive ceramics with brass wire tool showed repeated contacts between electrodes and helped to unveil the steps of material sublimation. CNC systems based on the detection of electrical pulses could not sense fast enough the electro-physical behaviour in the discharge gap due to the distinctive property of ceramics: high electrical resistance with a small value of required discharge gap. The prolonged physical contact between under-voltage electrodes results in frequent wire breakage and multiple cracks at the machined surfaces of a workpiece.

The control of the contact loss at the reversal motion is necessary as well. The loss of contact after short circuiting could be accompanied by the formation of tiny fibres of the easy-to-melt material components. That phenomenon extends the period of physical interaction between electrodes and can trigger subsequent repeated short circuits.

The developed method can supplement or even replace the existing monitoring methods of electrical parameters and contribute to the creation of a multi-factor adaptive control system. It can lead to the formation of a new vision and approach in the development of a new generation of CNC systems [74, 83, 228, 229].

The validation of the developed system on the samples of conductive ceramics and analytical investigation on the nature of electrical erosion promote better understanding on the phenomena of material sublimation under discharge impulses.

Declarations

Author contribution statement

Anna A Okunkova: Performed the experiments; Wrote the paper.

Sergey N. Grigoriev: Conceived and designed the experiments; Analyzed and interpreted the data.

Mikhail P. Kozochkin: Conceived and designed the experiments; Wrote the paper.

Artur N. Porvatov: Performed the experiments; Contributed reagents, materials, analysis tools or data.

Marina Volosova: Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data.

Funding statement

This work was supported by the Russian Science Foundation (project No. 18-19-00599).

Competing interest statement

The authors declare no conflict of interest.

Additional information

No additional information is available for this paper.