International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

Modeling of material removal in mechanical type advanced machining processes: a state-of-art review
Neelesh K. Jain, Vijay K. Jain
*

Mechanical Engineering Department, Indian Institute of Technology, Kanpur 208016, India Received 15 March 2000; received in revised form 28 November 2000; accepted 8 January 2001

Abstract Performance of any machining process is evaluated in terms of machining rate and surface finish produced. Higher machining rate and better surface finish are desirable for better performance of any machining process. Comprehensive qualitative and quantitative analysis of the material removal mechanism and subsequently the development of analytical model(s) of material removal (MR) are necessary for a better understanding and to achieve the optimum process performance. Analytical MR models are also necessary for simulation, optimization and planning (i.e. operation and process planning) of the process, prediction of process performance indicators, verification and improvements of experimental results, selection of appropriate models for specific type of work material and machining conditions, etc. Since the inception of different unconventional machining processes, various investigators have proposed different analytical models of material removal as functions of controllable process variables. A continual need for a comprehensive and exhaustive review of various analytical material removal models for different advanced machining processes is being felt. This paper is intended to fulfil this need in the area of advanced machining. Various analytical and some semi-empirical/empirical material removal models (approximately 40) for different mechanical type advanced machining processes have been comprehensively and exhaustively reviewed, and have been presented in a format suitable for quick reference.  2001 Elsevier Science Ltd. All rights reserved
Keywords: Advanced machining; Material removal rate (MRR); Modeling; USM; RUM; AJM; WJM; AWJM; AFM; MAF

* Corresponding author. Tel.: +91-512-597-916; fax: 91-512-590-260. E-mail address: vkjain@iitk.ac.in (V.K. Jain).

0890-6955/01/$ - see front matter  2001 Elsevier Science Ltd. All rights reserved. PII: S 0 8 9 0 - 6 9 5 5 ( 0 1 ) 0 0 0 1 0 - 4

1574

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

Nomenclature A d dm dmax dmin E G h hm H Hd Kc Ki l Patm rm Ra Ro a Rmax V Vl Vc Va n r sc se send si su sy Cross-sectional area (mm2) Diameter of an abrasive particle (mm) Mean diameter of abrasive grain (mm) Maximum diameter of abrasive grain within a particular mesh size (mm) Minimum diameter of abrasive grain within a particular mesh size (mm) Young’s modulus of elasticity (MPa or N/mm2) Shear modulus (MPa or N/mm2) Depth of indentation (mm) Maximum penetration depth (mm) Brinell hardness (MPa or N/mm2) Dynamic hardness, i.e. resistance with which target material resists the indentation (MPa or N/mm2) Fracture toughness or critical strain energy release rate (Nm 3/2) Constant of proportionality (for i=1,%,n) Length (mm) Atmospheric Pressure (MPa or N/mm2) Mean radius of abrasive grain (mm) Arithmetic average roughness value of the surface (µm) Initial surface roughness value (µm) Peak to valley roughness value (µm) Volumetric material removal rate (MRR) (mm3/s) Linear material removal rate (mm/s) Crater volume (mm3) Volume of impacting abrasive particle (mm3) Poisson’s ratio Density (kg/mm3 or ×106 g/cm3); Compressive yield strength (MPa or N/mm2) Elastic limit (MPa or N/mm2) Endurance limit (MPa or N/mm2) Stress induced (MPa or N/mm2) Ultimate strength (MPa or N/mm2) Tensile yield strength (MPa or N/mm2)

Relevant to ultra sonic machining (USM) A At b c C Amplitude of vibration (mm) Cross-sectional area of cutting tool (mm2) Burger vector (mm) Size of median or lateral crack (mm) Concentration of abrasive in the slurry by mass, i.e. ratio of mass of abrasive to mass of abrasive carrying medium

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1575

D f F Fm F t Lt N p P q R t T ta t Vwg xt x x ¯ m l w

Cutting tool diameter (mm) Cyclic frequency of vibrations (cycles/s) Mean static force over the time period T (N) Maximum static force (N) Mean static force over the time period t (N) Length of cutting tool (mm) Number of active abrasive grains in the working gap Pseudo pressure between tool and workpiece (=F/pR2) (MPa or N/mm2) Shocking force acting on a single abrasive particle (=F t/N) (N) Work hardening capacity (MPa or N/mm2) Tool radius (mm) Time (s) Time-period of vibrations (s) Time at which vibrating tool-tip comes in contact with abrasive particle(s) Fraction of time period during which vibrating tool-tip is in contact with the abrasive particles (s) Volume of tool-work gap (mm3) Distance between tool and work at time t (mm) Distance between tool and work at the end of stroke (mm) Average distance between tool and work during a stroke (mm) Lame’s constant [m=E/2(1+n)] (MPa or N/mm2) Indentation ratio [l=ht/hw=Hw/Ht] Angular frequency of vibration (rad/s)

Relevant to rotary ultrasonic machining (RUM) a c Cs l Lc P S Va d dw Mean half-diagonal length of an abrasive (mm) Radial extension of median or lateral crack (mm) Ratio of effective contact length to the mean diameter of abrasive particles Length of lateral crack (mm) Contact length i.e. the distance moved by a particle during penetration into the workpiece due to rotary motion of the tool (mm) Net contact load per abrasive particle (N) Rotational speed of cutting tool in RUM (rpm) Volume (conical shaped) removed under the action of a single abrasive particle (mm3) Depth at which lateral crack originates (mm) Maximum depth to which diamond particles penetrate the workpiece (mm)

Relevant to abrasive jet machining (AJM) e f Erosion rate, i.e. volume loss per unit impact (mm3/impact) Friction coefficient

1576

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

Ls m ma M ˙ Ma n na N p P Pu Sd Vb va vel vn vp vrp a ao dc e m f y h sb sfw

Sliding length (mm) An index Mass of single abrasive particle (kg) Total mass of all the abrasive particles (kg) Mass flow rate of abrasive particles (kg/s) Flaw parameter that characterizes flaw distribution Number of abrasive particles in a given volume of the carrier gas Effective number of abrasive particles impacting per unit time (N=nah) (per second) Nozzle pressure (MPa or N/mm2) Total load applied on the abrasive particles (N) Applied load per unit contact area (N/mm2) Standard deviation of natural logarithm of abrasive particle diameter Volume stressed in bending (mm3) Velocity of the impacting abrasive particle (mm/s) Maximum velocity of the abrasive particles at which collision is still purely elastic or at which elastic limit is just reached (mm/s) Component of particle velocity “va” normal to the work surface (mm/s) Component of particle velocity “va” parallel to the work surface (mm/s) Residual parallel component of the abrasive particle velocity with which it leaves the target surface (mm/s) Angle of impact, impingement or attack (angle between the direction of movement of an abrasive particle and the target surface) (degrees) Angle of attack at which vrp is zero (degrees) Critical plastic strain or erosion ductility Deformation wear factor (units of kinetic energy absorbed by the work surface to release one unit volume of material from the work surface by deformation wear) (J/mm3) Cutting wear factor (units of kinetic energy absorbed by the work surface to release one unit volume of material from the work surface by cutting wear) (J/mm3) Ratio of vertical to horizontal force components acting on a particle cutting face Ratio of contact length to depth of cut Proportion of abrasives effectively participating in erosion Fracture stress in bending (MPa or N/mm2) Flow stress of target or workpiece material (MPa or N/mm2)

Relevant to water jet machining/cutting (WJM/WJC) and cleaning (WJCI) C Cfw dj dn Jet spreading constant Total skin friction coefficient for the workpiece material Jet diameter (mm) Nozzle diameter (mm)

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1577

Feed rate or travel speed of the water jet in the horizontal direction (mm/s) Mass flow rate of water droplets per unit area at the center line of the water jet at a distance x from the nozzle exit (kg-mm 2s 1) mxr Mass flow rate of water droplets per unit area at a distance x from nozzle exit and ˙ at a distance r from center line of the water jet (kg-mm 2s 1) me Mass flow rate of water droplets per unit area at the nozzle exit (kg-mm 2s 1) ˙ P Water pressure from intensifier or pump (MPa or N/mm2) Pe Jet dynamic pressure at the nozzle exit (MPa or N/mm2) [Pe=(1/2000)rjv2] e r Distance of any point under consideration from the center line of the water jet (mm) Rx Radius of the jet in the water droplet zone at a distance x from the nozzle exit (mm) rn Nozzle radius (mm) ve Velocity of the water jet at the nozzle exit (mm/s) w Cleaning width (mm) wm Maximum cleaning width (mm) x Stand-off-distance (SOD) (mm) xc Critical stand-off-distance (SOD corresponding to zero cleaning width) (mm) xi Length of initial region of the jet (mm) xm Optimum stand-off-distance (SOD corresponding to maximum cleaning width) (mm) z Penetration depth (mm) z Penetration rate (mm/s) ˙ gw Velocity of sound in water (mm/s) hw Damping coefficient for the workpiece material (kg-mm 2s 1) l Stress coefficient dependent on the droplet size, coating thickness, and properties of jet, coating, and substrate materials rc Density of the coating material (kg/mm3 or ×106 g/cm3) rj Density of the water jet (=10−6 kg/mm3) (send)c Endurance strength of coating material (MPa or N/mm2) suc Endurance strength of coating material (MPa or N/mm2) Relevant to abrasive water jet machining (AWJM) aw Am C CD Cd Cf di dj Average grain size of wokpiece material (mm) Cross-sectional area of machined hole (mm2) A ratio indicating what portion of abrasive water jet (AWJ) is involved in cutting wear mode and C=1 (at/ao) if at ao Discharge coefficient Coefficient of drag or friction between AWJ and kerf wall An empirical coefficient accounting the effects of multiple impacts Diameter of nozzle insert (mm) Diameter abrasive water jet (mm)

f mxc ˙

1578

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

dwn f ff Hk ma ma ˙ mw ˙ P Pc Rf Rl Sf tf va vac vat vwe w wmin wt ao at e gw y m rj sfw ssw x

Diameter of water jet nozzle (mm) Feed rate or travel speed of water jet in the horizontal direction (mm/s) Fraction of stress wave energy responsible for inter-granular cracking Knoop hardness (MPa or N/mm2) Mass of single abrasive particle (kg) Mass flow rate of abrasive particles (kg/s) Mass flow rate of water (kg/s) [mw=CD(p/4)d 2 vwerj] ˙ wn Water pressure (MPa or N/mm2) Critical or threshold cutting pressure (MPa or N/mm2) Roundness factor of the abrasive particle, i.e. ratio between average diameter of particle corners to diameter of maximum inscribed circle Loading ratio, i.e. ratio of mass flow rate of abrasive particles to the mass flow rate of water [Rl=ma/mw] ˙ ˙ Sphericity factor of abrasive particle Thickness of fracture developed at backside of the material after passes of the jet (mm) Velocity of abrasive particle (mm/s) Critical or threshold velocity of abrasive particle (mm/s) Velocity of abrasive particle at the top of kerf (mm/s) {vat=x[vwe/(1+Rl)]} Velocity of water jet at the nozzle exit (mm/s) [vwe=√(2000P/rj)] Width of the kerf (mm) Minimum kerf width (mm) Width of the kerf at the top (mm) Angle of impingement at which maximum erosion occurs (degrees) Angle of impingement at the top of the surface (degrees) Deformation wear factor (units of kinetic energy absorbed by the work surface to remove one unit volume of the material from the work surface by deformation wear) (J/mm3 ) Energy required for inter-granular cracking per unit area (J/mm2) One-half of characteristic included angle of indentor (abrasive particle), (degrees) Ratio of tangential to normal force in micro-cutting (assumed as 0.5) Density of the water jet (=10−6 kg/mm3) Flow strength of the workpiece material (MPa or N/mm2) Specific energy of work material (J/mm3) Mixing efficiency

Relevant to abrasive flow machining (AFM) Ag C ls Cross-sectional area of the groove generated by a spherical abrasive grain (mm2) Percentage concentration of abrasives by weight in the AFM medium i.e. ratio of the weight of abrasives to the weight of abrasives and carrier medium expressed as % Length of stroke (mm)

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1579

lw Li M n N Ns Ria Rm Rw vf vp rm sr

Length of workpiece (mm) Actual contact length between workpiece and spherical abrasive grains during the i th stroke (mm) Abrasive mesh size (abrasive grain diameter d in mm=15.24/M) Number of AFM cycles, i.e. one upward (or forward) and one downward (or backward) stroke constitute one cycle Number of abrasive grains acting per unit area of contact of media and cylinder (abrasive grains per mm2) Total number of abrasive grains indenting the workpiece surface per stroke Surface roughness value after the ith stroke (m) Radius of media cylinder (mm) Radius of workpiece (mm) Velocity of media around the workpiece having constant radius (mm/min) Velocity of piston (mm/min) Density of media (kg/mm3 or ×106 g/cm3) Normal stress acting on the abrasive grain (MPa or N/mm2)

Relevant to magnetic abrasive finishing (MAF) Aair Am B F f HMm HMair I lair lm nc n N O(t) Oi On(t) Ra(t) t v V tl w Cross-sectional area of air-gap (mm2) Cross-sectional area of magnet (mm2) Magnetic flux density (weber/m2 or Tesla) Total force acting in the machining region (N) Force acting on a cutting edge of single magnetic abrasive particle ( f=F/Nn), (N) Magnetic field strength of magnetic abrasives (A/m) Magnetic field strength in the air-gap (A/m) Input current (A) Length of air-gap (mm) Length of magnet (mm) Number of turns in the coil Number of abrasive grains on a single magnetic abrasive particle Number of magnetic abrasive particles acting in the machining region simultaneously Out-of-roundness (OOR) at machining time t (OOR is the difference between the radii of maximum and minimum circles, which embrace surface irregularities, of a cylindrical part at a particular cross-section) (mm) Initial out-of-roundness (mm) Normalized out-of-roundness at time t (On(t)=O(t)/Oi) Surface roughness value after machining for time t (µ) Machining time (min) Velocity of magnetic abrasives (mm/min) Linear material removal rate in turning operation (mm/s) Volume ratio of ferromagnetic material in a magnetic abrasive

1580

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

V mo mem mrf 2q

Volume of the material removed by a cutting edge of single magnetic abrasive particle for finishing time t (mm3) Magnetic permeability in vacuum (Henry/m) Relative magnetic permeability of electromagnets Relative magnetic permeability of ferromagnetic material Mean angle of asperity of abrasive cutting edges (degrees)

Subscripts a air f m t w Abrasive Air-gap Abrasive carrying medium Magnet Tool Work surface

Abbreviations AFM Abrasive flow machining AJM Abrasive jet machining AWJM Abrasive water jet machining BEDMM Brush electro discharge mechanical machining CHM Chemical machining EBM Electron beam machining ECDM Electro chemical discharge machining ECG Electro chemical grinding ECM Electro chemical machining ECSM Electro chemical spark machining EDAG Electro discharge abrasive grinding EDM Electro discharge machining IBM Ion beam machining IJM Ice jet machining LBM Laser beam machining MAF Magnetic abrasive finishing PAM Plasma arc machining RUM Rotary ultrasonic machining US-EDM Ultra-sonic electro discharge machining USM Ultra-sonic machining WJC(M) Water jet cutting (machining) WJCl Water jet cleaning

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1581

1. Introduction Stiff market competition and continuously growing demand for improved product performance has led to the development of an ever-growing variety and quality of materials like carbides, ceramics, composites, semiconductors, variety of glass, diamond, etc. These materials possess properties like high strength and stiffness at elevated temperatures, extreme hardness, high brittleness, high strength to weight ratio, high corrosion and oxidation resistance, and chemical inertness. Although these properties result in superior product performance, their precise shaping and/or machining can be difficult. Conventional machining or shaping methods result in very high machining cost, and degradation of strength and some useful properties. Therefore, there is a need to develop processes capable of giving adequate material removal rate (MRR) as well as minimum damage to the material properties. In addition to the above requirements, complexity of work surface shape and size, surface integrity, precision and miniaturization (i.e. micro or nanomachining) requirements have also played a vital role in the development of different advanced (i.e. unconventional or nontraditional) machining processes. These processes are so called as they use new forms of energy and tools. As of now, different types of advanced machining processes have been developed including the basic processes, hybrid processes, and different versions of both. Table 1 presents a classification of the basic advanced machining processes according to the type of energy source. It also details the mechanism of material removal, energy source, cutting tool, and transfer media used in each process. By combining two or more than two basic (advanced and/or conventional) processes, some hybrid processes have also been developed, so that the potentials and capabilities of each individTable 1 Existing classification of advanced machining processes Type of energy Mechanical Main process USM AJM WJM AWJM IJM AFM MAF CHM Energy source Ultrasonic vibration Pneumatic pressure Hydraulic pressure Hydraulic pressure Hydraulic pressure Hydraulic pressure Magnetic field Corrosive agent Tool Sonotrode Abrasive jet Water jet Abrasive-water jet Ice jet Abrasives Magnetic abrasives Mask Transfer media Abrasive slurry Air Air Air Air Putty Air Etchant or reactive environment Electrolyte Mechanism of material removal Erosion or Abrasion

Chemical

Chemical dissolution or ablative relation Anode dissolution through ion displacement Melting and vaporization

Electrochemical

ECM

High current

Electrode

Thermal

EDM EBM IBM LBM PAM

High voltage Ionized material Ionized material Amplified light Ionized material

Electrode Electron beam Ion beam Laser beam Plasma jet

Dielectric Vacuum Atmosphere Air Plasma

1582

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

ual process can be exploited simultaneously. Table 2 depicts some of the hybrid advanced machining processes. Analytical models of machining processes require complete and in depth understanding of process mechanism, and thus are difficult to develop. But these models are very useful in parametric optimization, process simulation, operation and process planning, process parameter selection, parametric analysis (i.e. understanding the influence of various process parameters on the process performance measures), process performance prediction (such as surface finish, material removal rate, shape, etc.), verification of the experimental results, and improving the process performance by implementing/incorporating some of the theoretical findings. They are also helpful in developing an effective, economic, and efficient strategy for automation, optimization, and control of the process. A comprehensive survey of various analytical and empirical models is useful in comparing their limitations and applicability, and subsequently selecting an appropriate model for specific type of work material and machining conditions. Unfortunately, there is no literature available where most of the models for a particular type of process(es) are available in a single reference for examining and choosing a suitable one for the problem under consideration. Further, the assumptions, limitations, and applicability of these models are also not available in a format suitable for quick reference. Also, it is not feasible to include all types of advanced machining processes in a single article due to length constraints. Therefore, the objective of this study was to do a comprehensive, critical, and exhaustive review of various analytical material removal models for mechanical type advanced machining processes, highlighting assumptions made in their derivations, their limitations and applicability, and present it in a format suitable for quick reference. Sometimes, a good model proposed in the past, is ignored due to various reasons. This type of study can avoid such possibilities. This paper is organized as follows: Section 2 briefly mentions the mechanism of material removal in the mechanical type advanced machining processes. Different modeling techniques, and various analytical models of material removal for ultrasonic machining (USM), abrasive jet machining (AJM), water jet machining (WJM), and abrasive water jet machining (AWJM) proTable 2 Some hybrid advanced machining processes Name of hydbrid process ECDM or ECSM US-EDM EDAG ECG BEDMM Type of energy source Thermal and electrochemical Thermal and mechanical Thermal and grinding Electrochemical and grinding Thermal, mechanical and electrochemical Tool Electrode Sonotrode and electrode Grinding wheel Transfer media Electrolyte Dielectric Dielectric Mechanism of material removal Melting and vaporization Melting and vaporization

Melting, vaporization and chipping Grinding wheel Electrolyte Electrochemical dissolution of anode, and chipping Rotating metal brush Water-glass solution Electrochemical deposition, in water melting, vaporization, and mechanical rupture

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1583

cesses along with their working principles, applications, and parametric analysis, have been presented in Section 3. Section 4 describes the same for the advanced finishing processes like abrasive flow machining (AFM) and magnetic abrasive finishing (MAF). The last section concludes this review study. 2. Mechanical type advanced machining processes Most commonly used advanced machining processes employing mechanical types of energy are USM, AJM, WJM, AWJM, IJM (ice jet machining), AFM and MAF. The last two of these are advanced finishing processes. Basic schemes of machining using these processes have been depicted in Figs. 1–6. Most of these processes (except WJM and IJM) use abrasive particles as tool, and work material, which, in these processes, is invariably removed by the phenomenon of erosion. Erosion or erosive cutting is defined as removal of material from the work surface by a stream of impinging solid abrasive particles carried in a suitable carrier. Basically, the process of material removal by erosion is similar to that of grinding and single point tool cutting in the sense that in all these cases material is removed by an individual particle (or tool) by displacing (in ductile materials) or fracturing (in brittle materials). But erosion is the most difficult to describe among these processes due to uncertainties involved in determining number, shape, velocity and direction of the striking particles. According to J.G.A. Bitter, material removal through erosion may take place either due to cutting wear or deformation wear or both, depending on the work material characteristics and conditions of striking. Cutting wear occurs due to the cutting action of freely moving particles which causes plastic deformation, and is associated with the component of the force parallel to the target surface. It mainly occurs with ductile materials at low angles of attack. Deformation wear is caused due to repeated collisions, which results in cracking and spalling of the target material and is associated with the component of the force normal to the target surface. It mainly occurs with brittle materials and with ductile materials at the normal angle of attack. Practically, two forms of wear generally occur simultaneously where the angle of attack or impact can take any value.

3. Modeling of material removal Since the inception of the different advanced machining processes, different researchers from time to time have proposed different types of models to predict material removal. These models can be categorized into the following three categories [1]: 1. Analytical models: These models are developed based upon the mechanism of material removal through comprehensive theoretical analysis of the process. This analysis is based on certain assumptions and involves process parameters, properties of the tool, work, and operating medium. The results obtained from these models may not agree well with the experimental observations depending upon the assumptions made to simplify the analysis. 2. Phenomenological empirical models: These models are developed by conducting extensive experiments under various conditions, and correlating or fitting suitable relation to the experi-

1584

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

mental observations. These models can be used for prediction for those working or machining conditions only under which the experimental results were obtained. In other words, these models cannot be employed so commonly as analytical ones. 3. Numerical simulation models: These models predict material removal quite accurately and estimate their dependence on other variables numerically. However, in most cases, either analytical or empirical model(s) are used as base model(s) for calculating material removal. Numerical techniques are used to evaluate some basic parameters (say, force, temperature, etc.) required for analytical or empirical models. This paper describes a state-of-the-art review of various analytical and some empirical models of material removal for different mechanical type advanced machining processes in a tabular format (Tables 3–9) detailing names of the investigators, mechanism of material removal, analytical expression for material removal, assumptions made in the analysis, limitations of the model, remarks of the authors, and/or conclusions of the investigator(s). 3.1. Ultra-sonic machining (USM) The history of USM began with a patent granted to L. Balamuth in 1945. USM has also been referred to as ultrasonic drilling, ultrasonic cutting, ultrasonic grinding, ultrasonic dimensional machining, ultrasonic abrasive machining, slurry drilling, or ultrasonic impact grinding [2]. In USM, electrical energy of high frequency is converted into mechanical vibrations through a transducer. These vibrations are transmitted to the abrasive particles in the slurry via an energy-focusing device or horn/tool assembly (Fig. 1). The horn is also known as a sonotrode, concentrator, acoustic coupler, tool-holder, or velocity/mechanical transformer. Normally, transducers vibrate

Fig. 1. Basic scheme of ultra sonic machining (USM) [After Kaczmarck, 1976].

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1585

in longitudinal or compressive mode but can get lateral or transverse vibrations also. They are either magnetostrictive or piezoelectric type. Piezoelectric transducers are preferred due to higher energy efficiencies, no cooling requirements, adaptability to rotary operations, and no heat damage liability. But, machines using these transducers have low power capacity. Important process parameters of the USM are vibration amplitude and frequency, abrasive grain type, size and concentration, static force or pressure, and indentation ratio (l). From the different analytical models presented in Table 3, it can be concluded that a power relationship between MRR and amplitude of vibration exists. Different researchers have proposed different power indices. Neppiras (as mentioned in [6]), and Rosenberg [6] have suggested a square relationship; Goetze (as mentioned in [12]) suggested a linear relationship; while Wang and Rajurkar [12] have proposed an exponential index in the range of 1–2. A power relationship also seems to exist between MRR and frequency of vibrations since number of strikes and momentum increase with an increase in the frequency of vibrations, causing an increase in MRR. Shaw [4], Miller [5], and Cook [7] have suggested a linear relation; Neppiras (as mentioned in [6]) has proposed a square root relation; while Wang and Rajurkar [12] have suggested an exponential relationship with the exponent as 1.6. Optimum values of grain size and static force also exist. MRR initially increases with grain size but, when it approaches the dimension of the gap some larger grains cannot enter into the cutting zone and MRR starts decreasing. MRR increases with static force but after a certain limit of static force, MRR starts decreasing due to crushing and the impaired motion of abrasive particles. MRR improves with slurry concentration but saturation occurs at around 30– 40%. MRR decreases with indentation ratio (ratio of work to tool hardness) significantly. In addition to these, higher work material brittleness, slurry pressure, and decrease in carrier fluid viscosity increase MRR. USM is characterized by low MRR and no (or very little) surface damage to the work material. It can be used for machining both electrically conductive and non-conductive materials preferably with low ductility and high hardness (above 40 HRC). Its potential is not limited by electrical or chemical characteristics of the work material. Mechanisms of material removal in USM as suggested by Shaw [4], include: (1) mechanical abrasion by direct hammering of abrasive particles, (2) micro chipping due to impacting action of freely moving abrasive particles, (3) cavitation effect of abrasive slurry, and (4) chemical action associated with the carrier fluid. Various analytical MRR models for USM suggested thus far by various investigators [4–13], have been presented in Table 3, except that of Saha et al. [10], and Kamoun et al. [11]. Shaw [4] was the first researcher to propose a static and simple analytical model giving the relationship between MRR and vibration amplitude, frequency, abrasive grit size and concentration, and feed force. But, its predictions do not agree well with experimental observations. Miller [5] developed an MRR model based on plastic deformation restricting its application to ductile materials only, while Rosenberg and Kazantsev [6] included the statistical distribution of abrasive particle size in their computationally intensive model. Cook [7] proposed the simplest model to predict linear machining rate. Kainth et al. [8] proposed another static and complicated model using the abrasive particle size distribution as given in [6]. Nair and Ghosh [9] also proposed another computation intensive model simulating the principles of elastic wave propagation. Recently, Wang and Rajurkar [12] suggested a more realistic model taking into account the stochastic and dynamic nature of the process. But, this is applicable to perfectly brittle materials only. Lee and Chan [13] developed an analytical model to predict effects of vibration amplitude, grit size, and feed force on MRR and surface roughness for ceramic composites.

1586

Table 3 Different analytical material removal models for the ultra sonic machining (USM) process Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator and material removal material removal (where C is the year [Ref.] volumetric concentration of abrasives)

M.C. Shaw (1956) [4]

Direct MRR due to direct hammering hammering of 3/4 FA abrasive particles V K C1/4dmf Hw(1+l) (Primary) where K is a constant of proportionality and is given by K

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

2 (pK1)1/4K 3/4 2 3 Impacting by MRR due to particle impacting freely moving ra 3/4 5/2 3/2 f A Cdm V2=8.6K1 abrasive particles Hw (Secondary)

All abrasive particles are identical, rigid, spherical in shape, and having projections whose average diameter d1=K2d 2 m All impacts are identical. Material removal is proportional to volume of material removed per impact, number of particles impacting per cycle and frequency of impacting. Penetration depth is inversely proportional to flow stress of the (equal to Brinell hardness) work material. For a given area of tool face, the number of active abrasive grains is inversely proportional to the square of mean diameter of the abrasive grains i.e. C N=K1 2 dm

Analysis does not First theory to confirm qualitatively agree with the process conception. experimental results. Simple and easy to use. Does not predict correct effects of variation in A, F, f. Proportionality constant can only be determined experimentally. Predicts infinite increase in machining rate with static force while an optimum value exists due to grain crushing. No allowance for grain size variation and for replacing crushed grains.

(continued on next page)

Table 3 (continued) Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator and material removal material removal (where C is the year [Ref.] volumetric concentration of abrasives) MRR is given by V K Depth of cut per unit time is K where K is the constant of proportionality. AdmfpPatm C bwGwqwRVwgra C+1 AdmfFPatm C bwGwqwRVwgra C+1 Abrasive particles are of cubical size. Plastic deformation is directly proportional to the stress. Plastic flow stress equals Burger vector times Shear modulus. Cross-sectional area of cut does not change during machining. Viscosity effects in water slurry are negligible. Applicable to ductile materials only as MR is assumed to depend on plastic deformation. Some non-realistic assumptions like all abrasive particles are cubes, and all particles under the tool tip participate in the cutting action, have been made in its derivation. Number of active grains is derived assuming that slurry is drawn when tool recedes, it can happen only at low frequencies. No allowance for the grain size variation.

G.E. Miller (1957) [5]

By chipping plastically deformed and work hardened material. In ductile material, machining rate depends upon work hardening while in brittle materials it depends upon size and rate of chip formation.

Linear dependence of MR on A, d, f, F, Patm Inverse dependence of MR on b, G, q, R, ra, Vwg Optimum values of C and p (pseudo-pressure) exist.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635
(continued on next page)

1587

1588

Table 3 (continued) Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator and material removal material removal (where C is the year [Ref.] volumetric concentration of abrasives) Involves tedious computation and its solution requires numerical integration.

Rozenberg and Brittle fracture Kazantsev (1964) [6]

First researcher to incorporate size variability of abrasive particles. Number of chips is proportional to mean load. Penetration rate is proportional to (A2p)1/2 Damaged volume is related to the force: F V C(dm) N N
n(dm)

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635
Volume removal is related to stress or force by simple power law i.e. sm (A2p)1/3; and V s3 =A2p m

Total force acting on the particle is Abrasive particles are given by: incompressible and are of irregular shape but dmax can be considered as spheres having N fF(d x)j(d)dd projections whose radii x of curvature are Total material removed per cycle is proportional to the given by: mean dimension of particle. dmax Based on experimental V N fv(d x)j(d)dd; evidence the statistical distribution j(d) of x abrasive particle size d and number of chips is given by: i.e. number of abrasive dmax particles having diameter d is given by: N f(d)dd d N x 1 1.095 dm dm Distribution of force acting on 2 3 abrasive particles is given by: 1

(continued on next page)

Table 3 (continued) Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator and material removal material removal (where C is the year [Ref.] volumetric concentration of abrasives) Distribution of volume damage due to single particle:
0 for (d-x) <l1 for l1< (d-x) <l2 KOq1dm
3 3 3 3

f V (d-x) =
(1+ q1dm)
3 KOq1dm (l-q3l1) 3 q3 3 3

(dm-x+l1) for (d-x) >l2

where q1=4Ht/(aoHw); q3=4Ht/(coHa); l1=boy/(4Ht); and l2 Here Ko, ao, bo and co are constants independent of tool, abrasive, and workpiece properties, y is ratio of the diameter of a projection to mean diameter of the abrasive grain. Machining rate in mm/s is given by Vl 5.9(Arm)1/2 sit f Hw Abrasive grains are spheres of uniform radius. Tool and abrasives are rigid. Viscosity effects are negligible. A linear relationship between fraction of active grits and ratio of indentation depth to grit-radius has been assumed. Model predicts linear Static model which is relationship between simple and easy to use. static stress st and MRR, while MRR drops after a certain value of feed force. It predicts that MRR is proportional to square root of grain radius, while practically an optimum value of grain radius exists. (continued on next page) 1 q1 d qld q3 q3 m 1 1 m

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

N.H. Cook (1966) [7]

Hemispherical indentation fracture

1589

1590

Table 3 (continued) Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator and material removal material removal (where C is the year [Ref.] volumetric concentration of abrasives) Abrasive grains are spherical in shape and follow Rozenberg’s size distribution function to take into account particle size inhomogeneity. Average number of particles in the Motion of tool remains sinusoidal under loaded working gap is given by: conditions. Non-homogeneity in the abrasive particle size has been considered. Machining rate can be evaluated quantitatively using material properties of tool and workpiece, which is not possible with earlier theories.

G.S. Kainth, A. Nandy and K. Singh (1979) [8]

dmax Static MR model 2.29Nf based on V [(d x)d]1.5 indentation (1+l)1.5dm x fracture due to 2 3 direct hammering d 1 1 dd action. dm

N
d
max

CAtdmrf 0.57(ra+Crf) x ¯
2 3

d 3 1− dd
dmin

d −1 dm

And x is evaluated by solving
p/2dmax

Computationally expensive. Predicts linear relationship between MRR and static force F that is practically not true. Predicts linear increase in MRR with abrasive size, while an optimum value exists. Only predicted effect of amplitude on machining rate agrees qualitatively with experimental results. Theoretical machining rate is an order higher than practical values.

0.55NHw F (1+l)dm (d 2 dxt)
qs xt 2 3

1 where x=(x+A) ¯ Acosqs ; (p/2−qs) (x+A−dmax) ; and qs=sin−1 A xt=x+A Asinq

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

d 1 dm dddq

(continued on next page)

Table 3 (continued) Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator and material removal material removal (where C is the year [Ref.] volumetric concentration of abrasives) Volume of material removed Abrasive particles are per second V=VfNf; where rigid spheres. Vf=Volume fractured by single grit No consideration is per cycle, and given for MR due to: particle impacting, cavitation, or chemical action of abrasive carrying medium. 3 D2(dm+A sinqa) Tool tip motion is N C 2 ra SHM and it is described 3 d m +C rf by A(t)=A sin(2pft) Abrasive grit rests on a qa= at which vibrating tool makes brittle half-space and contact with abrasive at t=qa/2pf). receives only a single impact in the position considered. Volumetric MRR in mm3/s is given by V=K1(1000)−0.8rmrw4/5f 13/5A8/5; C3AtK52 where K1= ; and 6 Workpiece is assumed to be semi-infinite solid. Axis of the moving grit is perpendicular to the 2/5 5p3(1−nw) 2 free surface during the K5 cos (2pft f) ; 2mw machining process. Speed of abrasive grit here C3=Gap coefficient, and is same as that of the f=Phase lag used in the equation of vibrating cutting tool. amplitude i.e. A(t)=A sin(2pft f). This model can be used for perfectly brittle materials only like amorphous glass for which syield sfract. Its results are not expected to be true for those materials, which show some plastic behavior like carbides or polycrystalline glasses. It does not consider any action in the deformation zone. Derivation of the model is computationally extensive. Vf is to be calculated using fracture profile. Distribution of strain energy has been determined using principles of elasticwave propagation. This model explains the experimental observation about the existence of an optimum value of abrasive grain diameter for maximum MRR. Quantity of MRR can be predicted through purely theoretical analysis unlike previous theories.

E.V. Nair and A. Ghosh (1985) [9]

Brittle fracture due to periodic and direct hamering of abrasive particles.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

Z.Y. Wang and Combined effect K.P. Rajurkar of impact (1996) [12] indentation (i.e. penetration of grit with given speed) and fracture phenomenon.

1591
(continued on next page)

It accounts for both the dynamic and stochastic nature of the USM process. All previous theories have proposed static models based on indentation fracture mechanics, while USM is a dynamic impact dominated process, the static model is only a rough approximation. It predicts analytically an optimum grain size for a specified MRR.

1592

Table 3 (continued) Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator and material removal material removal (where C is the year [Ref.] volumetric concentration of abrasives) 2pk 2f 1 (aF+bA)2; Volumetric MRR= 3N T AtEt T 1+ cos(wta+f) a= ; b= t Lt p t f=Phase used in the equation of amplitude i.e. A(t)=A sin(wt f), and K1 is a constant as mentioned in the assumptions. Pre-existing flaws are assumed in the material for the initiation of median or lateral cracks. Size of median or lateral crack is related to P by cn=k1P; with n being assumed equal to 1.5. Cutting tool is assumed as a slender column.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

T.C. Lee and C.W. Chan (1997) [13]

Brittle fracture

Any increase in tool-tip vibration amplitude, static load, and abrasive particle size causes increase in MRR and surface roughness. Median crack size grows with an increase in load on the abrasive particle and decrease in fracture toughness of the workpiece material.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1593

3.1.1. Rotary ultrasonic machining (RUM) In 1964, Percy Legge (as mentioned in [14]) from the UK, detailed the concept of rotary ultrasonic machining (RUM), a hybrid process which combines mechanisms of USM and grinding process. RUM has potential for higher MRR and clean cuts while maintaining low cutting pressures and giving little surface damage and strength reduction. Therefore, it is especially suitable for ceramic machining. Also, the rotary motion of the tool in RUM enhances MRR, accuracy, and tool life, and reduces cutting forces. Various analytical models developed for the RUM process have been presented in Table 4. Prabhakar et al. [14] proposed a theoretical MRR model based on brittle fracture whose predictions do not agree with the experimental observations, while Pei et al. [15] reported a mechanistic model to predict MRR. Recently, Pei and Ferrira [16] also reported the modeling of material removal in RUM by ductile mode. 3.2. Abrasive jet machining (AJM) In the AJM process, a high velocity jet of abrasive particles and carrier gas coming out from a nozzle impinges on the target surface and erodes it, as shown in Fig. 2. The AJM process is characterized by relatively low power consumption and small capital cost. This process is suitable for hard and/or brittle metals/alloys, semiconductors, and non-metallic materials like glass, ceramics, etc. It is especially useful for the parts having thin sections but not suitable for the parts having sharp corners. Process performance of AJM, WJM, and AWJM can be evaluated in terms of MRR, geometry and finish of workpiece, and nozzle wear rate. The important process parameters of AJM are abrasive type, size and concentration, type of carrier gas, nozzle shape, size and wear characteristics, jet velocity, nozzle pressure, stand-off-distance (SOD) or nozzle-tip-distance (NTD). Commonly used abrasives in USM are alumina, silicon carbide, glass beads, dolomite, and sodium bicarbonate. Alumina and silicon carbide are used for cutting while glass beads, dolomite or sodium bicarbonate are used for cleaning, etching and deburring. Larger grains are used to get higher MRR, and smaller grains are used for fine surface finish but also cause more nozzle choking. A larger number of abrasive particles in a given volume of carrier gas gives higher MRR but an optimum value exists beyond which nozzle choking becomes frequent. The composition of the carrier gas affects MRR indirectly due to the dependence of the velocity–pressure relationship on it. Commonly used gases are CO2, nitrogen, and air. Air is mostly preferred due to universal availability and its non-toxic nature. The shape and size of the nozzle affect MRR by affecting jet velocity, while nozzle wear influences machining accuracy. Jet velocity depends on nozzle design, pressure, abrasive particle size, and their concentration. In AJM processes, the abrasive jet must impinge the work surface with a certain minimum (or critical) velocity, which is dependent on the type of abrasive and work material. Nozzle pressure affects MRR. Stand-offdistance affects not only MRR but also the shape and size of the cavity produced. As SOD increases, shape inaccuracy and jet velocity also increase giving higher MRR but beyond a certain limit jet velocity starts decreasing due to atmospheric drag. Thus an optimum value of SOD exists to achieve maximum MRR and shape accuracy. Material in AJM is removed due to erosive action caused by impingement of a high velocity abrasive jet on the work material surface. Different mechanisms of material removal, for both

1594

Table 4 Different analytical material removal models for the rotary ultrasonic machining (RUM) process Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Theoretical model based on brittle fracture gives 1 MRR=NfVaCs; whereVa= pc2d; 3 Cs Lc pDS ;L ; c a l; dm c 120f

D. Prabhakar, Brittle fracture Z.J. Pei, P.M. Ferreira and M. Haselkorn (1993) [14]

First attempt in predicting MRR in RUM. Theoretical model explains MRR process by applying indentation fracture theory.

2/3

l

0.0143

Ew Hva

P 2 a ; Kcw (l+a)3

d a [dw(2rm dw)]1/2;

Workpiece material is ideally brittle, i.e. no plastic deformation. Abrasive particles are rigid spheres of same size and all particles participate in cutting during each ultrasonic cycle. Material is removed by Hertz fracture, i.e. propagation and intersection of cracks ahead of and around the cutting tool or abrasive particles.

1/3

dw ; where Ka= 1−n2 1−n2 a w ; and Kw= . pEa pEw

9p2P2(Ka+Kw)2 16 rm

Theoretical model involves parameters, which are to be estimated using indentation fracture theory and not the work-material, tool and/or abrasive properties. All the abrasive particles have been assumed to participate in the cutting action, in reality only a fraction of them participate. This results in deviation between predicted and experimental MRR values.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635
(continued on next page)

Table 4 (continued) Assumptions Limitations Remarks and/or conclusions Mechanistic model can be used to predict MRR by observing a model parameter (proportionality constant), which is invariant for most of the machining conditions.

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Mechanistic model based on brittle Same as in [14] fracture gives Proportionality constant K remains invariant Lc dm dw 2 across a broad range of dw; V KpNf 1 dm 2 3 process parameters for a particular material. where K is the ratio of fractured volume to indentd volume by a single diamond particle.
2 1/3

Z.J. Pei, D. Brittle fracture Prabhakar, P.M. Ferreira and M. Haselkorn (1995) [15]

Mechanistic model can be used to predict the MRR only for a particular material and for a set of process parameters. Constant K has to be determined experimentally.

dw ;

9(Fm/N)2 1−n2 w 8 dm Ew

Fm dw p −sin−1 1− 2 A

pF ;

Lc

DS p dw sin−1 1 60f 2 A

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635
(continued on next page)

1595

1596

Table 4 (continued) Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.]

Z.J. Pei and P.M. Ferreira (1998) [16]

Ductile mode of Volumetric MRR=NfW; where W material removal is intersection volume between by plastic flow Abrasive-Particle-Swept-Envelope (APSE) and workpiece. APSE is described by: x=−r sinq cosf+ y=r cosq z=r sinq cosf+A sin(2pft) where tanf= 120fA cos(2pft) DS Relation between indentation depth and process parameters, and workpiece properties is given by: pDSt 60

Work material is rigidplastic. Abrasive particles are rigid spheres of same size. All abrasive particles participate in cutting action. Material volume removed by each abrasive particle in one cycle is equal to intersection volume of APSE and workpiece.

F Nscfwdw(2rm dw) dw p sin−1 1 2 A where scfw=Compressive flow strength of workpiece.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635
First and simplified model of RUM for ductile mode of material removal. MRR increases with increase in A, F, and S, while it decreases with an increase in number and diameter of abrasive particles.

This (plastic flow) model is applicable only in small static force range (i.e. when ductile mode dominates material removal) and fails to predict in large static force range, i.e. when brittle fracture is dominant mode of material removal. Computation of intersection volume between APSE and workpiece W is difficult, complicated, and requires numerical integration.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1597

Fig. 2. Basic scheme of abrasive jet machining (AJM) [4,17].

ductile and brittle materials, have been suggested by various investigators [18–31] as mentioned in Table 5. In the case of ductile materials, material is removed by plastic deformation and cutting wear, or plastic strain and deformation wear. In the case of brittle materials, it may take place due to indentation rupture, elastic–plastic deformation, critical plastic strain theory, radial cracking and propagation or surface energy criterion [2]. Table 5 presents various models of material removal in AJM processes proposed by different investigators [18–31] from time to time. A more comprehensive state-of-the-art review of mechanisms of material removal, MRR models, and experimentation on AJM until 1990, is available in [1]. Finnie [18] was the first researcher to analytically model the erosive wear of ductile materials by the impact of solid abrasive particles. But, predictions of this model do not agree well with the experimental results at higher impact angles and cannot be used for near-orthogonal impacts. Also, it does not take into account the effects of abrasive particle size and shape on the erosion and gives a velocity index of 2, while experimental results predict it to be in the range of 2 to 3 (as reported in [39]). Bitter [19] proposed a model, which also accounts for deformation wear (based on the plastic deformation) and predicts erosive wear of both brittle and ductile materials more accurately. Sheldon and Finnie [20] developed an analytical model for erosive cutting of brittle materials by normal impact of abrasives. But, the constants involved in this model require complex calculations. Neilson and Gilchrist [21] simplified Bitter’s model and detailed the procedure to correlate the erosion relationships with the experimental results. Sheldon and Kanhere [22] analyzed the erosion process of relatively soft and ductile materials (i.e. aluminum) considering the impact by a relatively large ( 2500 µm) single abrasive particle and developed a simple analytical material removal model, which is applicable at relatively low impact velocities (i.e. 450 m/s). This model predicts a velocity exponent of 3 rather than 2 as given

1598

Table 5 Different analytical material removal models for abrasive jet machining (AJM) process Assumptions Limitations Remarks and/or conclusions Erosion experiments were conducted first time to measure velocity of striking abrasive particle. Erosion is characterized by very high flow stresses, which is to be determined experimentally.

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Volume of material removed (mm3) Abrasive particle is due to cutting wear by M kg of harder than the work abrasives; surface. Ratio of vertical to 6 2 Mv2 a horizontal force sin2a sin a h 3 10 sfwyf f component acting on f particle face (f) is a if tana constant. 6 Ratio of contact length OR to depth of cut (y) is a constant. Mv2 cos2a f a h 3 if tana Particle cutting face is 10 sfwy 6 6 of uniform width and is For angular abrasive grains f 2; large compared to depth of cut. Constant and y 2. plastic flow stress is reached immediately after impact. Applicable for impact of ductile materials at smaller impact angles only because it underestimates the erosion at higher and normal impact angles. Value of f cannot be measured directly.

I. Finnie (1960) Ductile [18] materials: Plastic deformation and fracture. Brittle materials: Intersection of cracks radiating from point of impact.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635
(continued on next page)

Table 5 (continued)

Name of Mechanism of investigator(s) material and year [Ref.] removal If vasina vel, then volume of Abrasive particles are Complicated. material removed due to cutting spherical in shape. wear: Colliding abrasive particles deform 2MC(va sina−vel)2 Wc1 elastically while work va sina surface deforms elastically and plastically. C(va sina−vel)2 No deformation va cosa 106m (mm3); hardening va sina i.e. sc is constant. if a ao and vrp 0.
Wc2 1M[v2 cos2a−K1(va sina−vel)3/2] a (mm3); 2 106m

Proposed analytical model for material removal No deformation wear if v sina vno as collisions are purely elastic.

Assumptions

Limitations

Remarks and/or conclusions

J.G.A. Bitter (1963) [19]

By cutting wear in the case of ductile materials at smaller angles of impact. By deformation wear in the case of brittle materials and ductile materials at higher impact angles. if a ao and vrp=0. Volume of material removed by deformation wear: 1M(va sina−vel)2 (mm3); WD 2 106 e where
0.25

c

0.288 ra 103.75sew sew (kg−1mm1/2s5/2);

K1 4.62
2

1−na2 Ea and

s2.25 ew r0.25 a 1−nw2 Ew (mm/s)1/2;
2

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

sew2.5 1−na2 1−nw2 vel 5p2 0.5 ra Ea Ew

1599

(continued on next page)

1600

Table 5 (continued) Assumptions Limitations Remarks and/or conclusions Erosion of brittle materials is approx. inversely proportional to the square of flexural strength. For most of the material n is at least 5, hence resistance is approx. proportional to E0.8/s2. b

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] For both spherical (r=r ) and Abrasive particles are angular particles, volume removed rigid. per particle Impact is normal and dynamic effects may 0.6n 6 rm n−2 2.4n be neglected. v n−2 ; K4 a r Initial fracture starts when maximum tensile where stress equals the tensile 2n 1.2n strength of material K3 n−2 5pra n−2 K4 C2p ; 2 corresponding to size K2 3g and stress distribution in the region exposed to tensile stress. 1−2nw 3 1−n2 1/3 w Depth, up to which ; ;g K3 4p 4 Ew surface tensile stresses persist, is proportional and to the indentation depth 1/n (constant of n−1 ; K2 sbv1/n b proportionality 4p(n+1)K1 being K1). where K1 and C2 are constants of Zero probability failure proportionality stress and radial surface tensile stress vary in the same manner. Only for erosive cutting of brittle materials at normal impact angle. Difficult to predict the constant of proportionality as it involves many factors.

G.L. Sheldon and I. Finnie (1966) [20]

Brittle materials: propagation and intersection of cracks ahead and around cutting particle or tool.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635
(continued on next page)

Table 5 (continued) Assumptions Limitations Remarks and/or conclusions —

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Total volume of material removed due to cutting wear and deformation wear: M v2 cos2a sin(ma) a 2×106 m (va sina−vel)2 e if a ao OR M v2 cos2a (va sina−vel)2 a 2×106 m e if a ao, ao p/2m; where m is a constant. Volume of material removed per particle (mm3): Applicable for ductile materials up to impact velocities of 450 m/s. The relation is not completely fundamental in the nature as it is based on the empirical hardness relation of Meyer. Experimental value of hardness exponent is not well established and needs further investigation. Erosion process has been studied through impact of relatively large single particles using the simple indentation theory. It predicts velocity exponent as 3 while energy consideration predicts it as 2. Material removal is more for workhardened surfaces. Not explicitly mentioned. —

J.H. Neilson Brittle: Cracking and A. Gilchrist and spalling. (1968) [21] Ductile: Both cutting and deformation wear.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

G.L. Sheldon Flow of material and A. Kanhere around cavity by (1972) [22] advancing particle and breaking of the displaced material due to sufficient strain.

1601

Amount of material removed is related to the depth of penetration 3 3 1.5 d mva ra of the particle. K1 4.5 ; 10 Hvw Effects of strain hardening, and inertia where Hvw=Vickers diamond of the target material, pyramid hardness number of the target material, and K1 is a constant deformation of the abrasive particles have of proportionality. been neglected. For low velocities and normal impacts Meyer’s hardness relation is valid (i.e. F=ahn; where a is the load for unit diameter and can be replaced by Hvw.

(continued on next page)

1602

Table 5 (continued) Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Macroscopic wear rate (mm3/s): h Pva ; pag Hw where h=linear scaling factor and ag=factor determined by the geometry of the abrasive particle. Grit particles are ideally sharp. Geometrical similarity is preserved in the indentation fracture process. Residual stress field intensity about the deformation track is sufficiently high to drive the chip forming cracks to the surface. Abrasive particles are spherical in shape. Each impact of abrasive produces spherical indentation whose volume equals the volume of the material removed per impact. For brittle materials using fixed abrasive medium (i.e. particles are bonded to the tool).

B.R. Lawn (1975) [23]

Initiation and propagation of lateral cracks by residual stresses associated with incompatibility between deformation zone and surrounding elastic matrix.

Calculated wear rate is independent of the contact area between tool and work, number and size of indenting particles. It avoids arbitrariness and complications of statistical analysis. Under ideal chipping conditions, wear rate is uniquely determined by work material hardness. For brittle materials at Simple and easy to use. normal impact. Analytically MRR p1.75; Experimentally MRR p1.01 1.42

P.K. Sarkar and Rupture due to Volumetric material removal rate P.C. Pandey spherical (mm3/s): (1976) [24] indentation 3/4 ra caused by the k2nahd 3v3/2 a 3 12×10 Hw impact of abrasive particle. k2=Constant of proportionality.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635
(continued on next page)

Table 5 (continued) Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Material removal rate in kg/s: 0.1156 ˙ n raMav2 3 10 sfw For brittle materials at Nature of profile of normal impact. eroded surface with SOD has been explained qualitatively using compressible fluid flow conditions. Based on Hertz equation. For brittle materials. Comparison of predicted results of both the models with the experimental results shows considerable scatter. Plastic deformation is predominant when: load on abrasive is small, abrasive is blunt or becomes blunt, and Kwc/Hw is high. Indentation fracture is predominant when: load on abrasive is high, abrasive is sharp or remains sharp, and Kwc/Hw is low.

M.L. Neema and P.C. Pandey (1977) [25]

Indentation

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

M.A. Moore and F.S. King (1980) [26]

Plastic deformation and/or indentation fracture depending on the material and wear conditions. Practically material is removed by the combination of both mechanisms. Fracture mechanisms can cause 10 times wear than plastic deformation.

Volume wear per unit contact area Abrasive particles are 120° right cones and (mm3/mm2) due to plastic deformation: act like a scratch indentor with normal 0.2LsPu load being carried on Hw leading half-cone Volume wear per unit contact area surface. A contacting abrasive (mm3/mm2) due to indentation particle forms a groove fracture mechanism: by plastic deformation 2LsP1.25d 0.5K −0.75H−0.5 and a portion of this u m cw w groove volume is removed from the surface. Number of abrasive particles contacting the surface per unit area is proportional to d −2. Only a proportion of abrasive particles contacting the surface removes the material.

1603
(continued on next page)

1604

Table 5 (continued) Assumptions Limitations Remarks and/or conclusions For the erosion by spherical particles at normal incidence by the platelet mechanism. It predicts that erosion is dependent on the cube of the impact velocity.

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Mass loss from the target surface per unit mass of impinging particles: Eroded material is rigid perfectly plastic with no strain hardening, while eroding particles zrwr0.5v3 a a are rigid non-deforming 0.033 2 1.5 dcwHw spheres. where z=Fraction of indentation the Target material resists indentation with a volume which is plastically constant pressure i.e. deformed. dynamic hardness. Elastic forces are negligible. All initial KE of particle is available for indentation. Whole volume plastically deformed by each impact is subjected to the same amount of plastic strain increment, which is directed radially outward in the plane of target surface. A fraction of indentation volume is plastically deformed. For the erosion of ductile materials at normal impacts. Factor z/d2 cw cannot be measured directly. Effects of strain hardening and variations in strain rate with particle size and velocity have not been considered. No account of precise mechanism of material removal.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

I.M. Hutching (1981) [27]

Platelet mechanism in which flat disc shaped platelets are detached from the surface after many cycles of plastic deformation when plastic strain attains critical value.

(continued on next page)

Table 5 (continued) Assumptions Limitations Remarks and/or conclusions It assesses the effects of impacting particle size and velocity distribution on the erosion of brittle materials. Measured erosion rate is strongly dependent on Sd.

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] All particles of the same diameter impact the target with the s2 d same velocity. 2 e e(dm)exp {(n1 n2x) 9} Erosion rate is related 2 to impact velocity & Quantitative model based on quasi- particle diameter by static indentation predicts erosion e(d)=k d n1vn2 1 a being proportional to Impact velocity is related to particle v2.4d 3.7r1.2K −1.3H0.1 a m a cw w diameter by Quantitative model based on va=ad x; where a and x dynamic analysis of elastic–plastic are characteristics of stress field predicts erosion being erosion test conditions. proportional to Particle size has logarithmic–normal 19/6 11/3 1/4 −4/3 −1/4 va d m ra K cw Hw type of distribution. Interaction effects are negligible i.e. volume removed by multiple impacts is the summation of volumes removed by individual impacts. Empirical equation gives
0.75+0.5a

D.B. Marshall, A.G. Evans, M.E. Gulden, J.L. Routbort and R.O. Scattergood (1981) [28]

Semicircular radial median cracks normal to the target surface for strength degradation. Lateral cracks parallel to the target surface for material removal.

Measured erosion rate (volume loss per impact) is given by:

For brittle materials. Analysis is confined to elastic-plastic regime relevant to hard angular particles.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

G. Sundararajan Lateral and/or (1984) [29] radial cracking and their interlinkages.

Abrasive particles are spherical. ra Vc/Vb 0.8 v1.5+acosaasin1.625a Impacts are oblique. a 103Hdw Deceleration of the impacting particle in while the analytical model of horizontal direction is Ratner and Styller gives dependent on the 1.25 ra Vc friction force only, 2.5 (v sina) (cota f) while that in vertical Vb 103Hdw direction is dependent where a=0.5+0.3cos3a on normal load only [for the model of Ratner and Styller].

For ductile materials. Relation for volume removal is empirical.The equation is valid for oblique impacts, (i.e. a 60°) of ductile materials.

Proposed relation for crater volume accurately predicts it over the wide ranges of impact velocity (50– 350 m/s), impact angles ( 60°), and dynamic hardness of the target material (1–4 Gpa) unlike the mentioned analytical relation.

1605
(continued on next page)

1606

Table 5 (continued) Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Material removal rate in mm3/s is given by: ˙ a MRR K1Mav1.2 where K1 is constant and is given by
0.6 0.6

K.N. Murthy, D.C. Roy and P.K. Mishra (1986) [30]

Brittle fracture due to indentation at different impact sites.

K1

10−1.8 5p 2 4 (kg−1mm1.8s1.2).

1−n2 1−n2 a w + Ea Ew r0.4 a

Abrasive particles are blunt spheres of same average size. Normal impact of abrasive particle causes hemispherical shape of target material to be removed. Velocity of abrasive particles is one third of calculated velocity of the carrier gas. All particles participate equally in material removal and craters formed do not interact. KE of impacting particle is transformed in elastic strain energy of work material after impact.

There is discrepancy between theoretically predicted (=1.2) and experimentally found (=3) value of the velocity exponent. Velocity and angle of impact of all the particles will not be the same as assumed. All particles will not remove material.

For brittle materials at normal impact angle. Based on Hertz contact stress, which can be described by linear elastic fracture mechanics. Predicts existence of a threshold pressure below which no material removal takes place.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635
(continued on next page)

Table 5 (continued) Assumptions Limitations Remarks and/or conclusions For brittle materials at normal impact angle. Identical to the model of Sarkar and Pandey [24]. Abrasive grain velocity cannot be same as that of the carrier gas due to inertia. Whole plastically deformed volume will not be chipped out but only a fraction of it (may be 85–90%).

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Abrasive grains are identical spheres. Abrasive particles are assumed to be inelastic while workpiece is rigid perfectly plastic. Velocity of abrasive grains same as that of the carrier gas due to small grain size. Normally brittle materials have been assumed to behave in ductile manner for high velocity erosion. Hertz equations are valid for fine abrasive particles. Volume of material removed is approximately proportional to the plastically deformed material. Not able to explain the decrease in flow rate of air–abrasive mixture with increase in working pressure. Effect of pressure on MRR does NOT agree with the experimental results reported by Sarkar and Pandey [24]. Type of work, abrasive materials, and velocity of impact has NOT been mentioned so the assumption of brittle material behaving as ductile cannot be justified.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

P.K. Jain, A.K. Chitley and N.K. Nagar (1990) [31]

Plastic Material removal rate in mm3/s is deformation of given by: surface layer due 0.75 ra 2 to impact of N. MRR pd 3 v1.5 m a 3 3 12×10 syw abrasive particles.

1607

1608

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

by energy considerations therefore its predictions agree more closely with the experimental results as compared to previous models. Lawn [23] proposed a simple model for the wear of brittle materials by the abrasives bonded to the tool. Sarkar and Pandey [24], and Neema and Pandey [25] also proposed relatively simpler models for brittle materials. While the former considered the effect of nozzle pressure on MRR, the latter qualitatively explained the nature of the profile of the eroded surface with increasing SOD using compressible fluid flow conditions. Moore and King [26] proposed two models to predict wear volume of brittle materials by plastic deformation and indentation fracture. But there is considerable scatter between the prediction from these models and experimental results. Neglecting the effects of strain hardening and using critical plastic strain criterion, Hutching [27] proposed an easy-to-use model for the erosion of ductile materials by spherical particles at normal impacts. The only model incorporating the effects of statistical distribution of abrasive particle size and velocity on the erosion of brittle materials was proposed by Marshall et al. [28]. Sundararajan [29] proposed an empirical model to predict the volume of the crater formed during the oblique impacts of ductile materials. This model accurately predicts the crater volume over the wide range of abrasive velocity (50–360 m/s) and up to an impact angle of 60°. Based on Hertz’s contact stress, Murthy et al. [30] proposed another simple model to predict the MRR of brittle materials at normal impacts. It gives a velocity exponent equal to 1.2, while experimental observations reveal it to be approximately 3. Jain et al. [31] also developed a simple model identical to that of Sarkar and Pandey [24] for estimating the MRR of brittle materials by normal impacts. 3.3. Water jet machining (WJM) The kinetic energy of a high pressure and velocity water jet can be used either for destructive or precision cutting applications. Destructive applications of water jet include: hydraulic mining, tunnel boring, demolishing concrete structures, cutting anti-skid grooves in airways, making trenches and laying cables, etc. Precision applications include: hole making, cleaning, descaling, deburring, cutting of printed circuit boards (PCB), profile cutting of fiber-reinforced plastic (FRP) aircraft structures, etc. WJM is best suited for materials like kevlar, asbestos, glass epoxy, corrugated board, leather, paper products, graphite, boron, FRP, and some brittle materials. The published literature [3] indicates the applications for materials like steel and brass also, but the pressure required is extremely high. Advantages of this process are: clean and sharp cut, dustless environment, no heat generation, no tool wear, suitability to explosive environment, less maintenance, no pollution problems, and no thermal degradation of material properties [3]. Schematic diagram of the process is shown in Fig. 3(b). The main elements of the system are: (i) pressure source (hydraulic pump or intensifier) to generate very high pressure, (ii) cutting unit consisting of a nozzle and workpiece holding table, and (iii) filtration unit to collect water and to remove debris from it if it is to be reused. A suitable polymer is mixed with the water before it is sent to the intensifier to check the divergence of the stream coming out of nozzle, to prevent freezing, and to provide lubrication. The water jet becomes a stream of water droplets beyond a certain length of travel due to continuous interaction between water and the surrounding air. Three regions, namely initial, main, and final, that exist in the water jet are shown in Fig. 3(a). The initial region is close to the nozzle exit and a wedge shaped region known as a potential core exists inside this region, where the velocity is equal to the jet velocity at the nozzle exit. The

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1609

Fig. 3.

Water jet machining (WJM): (a) structure of water jet; (b) basic scheme of machining or cleaning [35].

water jet in the initial region is considered to be continuous with very little air entrained inside it. Eddy formation causes exchange of matter between water and the surrounding air and air gets entrained into the water stream breaking it into fine droplets. In the main region, mixing of the water jet with air continues to its full extent and it keeps disintegrating continuously into droplets. This results in the gradual expansion of cross section and a reduction in the velocity of the water jet. The main region of the water jet consists of two zones namely, the water droplet zone and water mist zone between the droplet zone and surrounding air. The final region is a diffusion region in which the water jet is totally broken into small droplets [35,36]. Important process parameters of WJM are stand-off-distance (SOD), water pressure, travel speed (or feed rate) of jet, and nozzle diameter. An optimum value of stand-off-distance exists for maximum MRR and it has been confirmed experimentally [34–36]. MRR increases with increasing SOD but beyond a certain value of SOD, MRR starts decreasing due to a reduction in jet velocity with increasing distance from the nozzle tip. SOD also affects accuracy and quality of cut. A divergent jet cuts less effectively and less accurately. Machined depth increases with an increase in water pressure for both metals and non-metals. Also, a threshold pressure exists below which no cutting takes place. The rate of cutting increases with an increase in nozzle size

1610

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

[35,36]. An optimum value of feed rate of jet also exists for maximum material removal. Cutting and piercing ability of the jet is most effective when it is directed normal to the work surface. Depth of groove also decreases with increasing feed rate. Geometry and finish of cut depend on nozzle design, jet velocity, feed rate, depth of cut, and properties of target material (especially hardness). As with the AJM process, nozzle wear depends on nozzle material, its hardness, nozzle design, and pressure (and hence velocity) of the water jet [32]. The mechanism of material removal in WJM is highly complex and involves four damage modes [35], namely: direct deformation, stress wave propagation, lateral outflow jetting and hydraulic penetration. Erosion is initiated by micro and macro cracks developed due to propagation of stress waves caused by repetitive impacts of water droplets. The last two modes propagate, extend and enlarge the existing cracks. When a high velocity water jet strikes a target surface, its velocity is virtually reduced to zero. Most of its kinetic energy is converted into very high pressure and induces stress waves in the target material. Material removal occurs due to fatigue after a certain number of cycles when equivalent dynamic stress generated by the water jet exceeds the endurance limit of the target material. Mechanisms of material removal in water jet cutting and water jet cleaning are different. Water jet cutting normally involves penetration of solid by a continuous jet while water jet cleaning involves erosion by discrete droplets. Analytical material removal models for WJM or WJC are presented in Table 6(a). Hashish and duPlessis [33] carried out theoretical investigations for the penetration of a wide range of solid materials (i.e. wood, coal, concrete, rocks, and plastics) by continuous high-velocity water jets, and developed a simple non-dimensional equation for the penetration depth. The same authors modified this equation by combining it with an empirical equation to take into account the jet spreading and velocity decay of the jet in the air, and used it to study the effects of SOD and number of passes on the penetration depth, volume removal, and specific energy consumption [34]. The analytical predictions of this model for penetration depth and specific energy are in good agreement with the published experimental results. Table 6(b) presents the analytical models for the water jet cleaning (WJCl) process. Leu et al. [35] did mathematical modeling and experimental verification of this process considering the water jet as stationary and striking at normal incidence angle. They developed a simple analytical expression for cleaning width as a function of SOD, water jet pressure, and nozzle radius. They also derived the relations for optimum and critical SOD and maximum cleaning width. Its predictions are in good agreement with experimental observations. Meng et al. [36] developed a computationally intensive and semi-empirical model for the WJCl process considering the water jet as moving and striking at normal incidence angle. 3.4. Abrasive water jet machining (AWJM) The best of the AJM and WJM processes are combined to create a process known as AWJM. This process relies on erosive action of an abrasive laden water jet for applications of cutting, drilling, and general cleaning and descaling of thick sections of very soft to very hard materials at high rates. A variety of materials can be machined by AWJM including copper and its alloys, aluminum, lead, steel, tungsten carbide, titanium, ceramics, composites, acrylic, concrete, rocks, graphite, silica glass, etc. Its most promising application includes machining of sandwiched honeycomb structural materials frequently used in the aerospace industries. The main advantages of the

Table 6 Different analytical material removal models for water jet processes Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.]

(a) Different analytical material removal models for the water jet machining (WJM) or water jet cutting (WJC) processes M. Hashish and Compressive Dimensionless equation expressing Bingham model has — Model is based on M.P. duPlessis failure of the penetration depth z been used to describe control volume (1978) [33] material time-dependent stressapproach to determine strain relationship. the hydrodynamic Area of cut is constant forces acting on 1000scw 1− and is equal to the boundaries of the rjv2 z e nozzle exit area. cutting slot. 1 dn Control volume is Its simplicity leads to stationary in horizontal closed form solution of 2Cfw plane. coupled fluid-solid Velocities are normal mechanics equation. p to the respective It allows convenient control surfaces. characterization of Body and pressure different materials based forces on a typical on sc, Cf and h control volume and density change due to 2Cfw rjve ve frictional heating have e− √p hw f been neglected.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635
(continued on next page)

1611

1612

Table 6 (continued) Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Dimensionless equations for penetration depth z zCfw dn K e−2.256
hwf CfwPef

M. Hashish and Compressive M.P. duPlessis failure of the (1979) [34] material 0.297 1 syw x 2/3 y 1 xi 2Pef

Dimensionless relation of the analytical model of Hashish and duPlessis [33] has been used. Width of cut is uniform along the kerf and is equal to the effective width of the jet at the start of cutting.

Predicted results have been compared with the published experimental results of wood, limestone and coal only. Volume removal predicted by it does not agree with the experimental results.

Dimensionless equation for volume cut per unit length along the cut V x syw VCfw 0.099 y4/3 1 d 2K xi 2Pef n 1 e−2.256
hwf CfwPef

Empirical equations for jet spreading and jet velocity decay in air have been incorporated. Optimum SOD exists in the range of 2 x/xi 5 and 2 CfPe/hf 10 only For cleaning and mining applications, the most efficient volume removal occurs at the optimum SOD and at hwf/CfwPc 10

xi where f=2 [0.5 0.57y+0.2y2]; x

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

y 1

xi scw x ; and K 2Pexi dn

(continued on next page)

Table 6 (continued) Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Analysis is applicable to normal incidence only and also all the experiments have been conducted using normal incidence of the water jet.

Maximum cleaning width is proportional to critical cleaning SOD (wm=0.525Cxc). Optimum SOD is 0.576 times critical SOD. Critical cleaning SOD is proportional to nozzle radius and onefourth power of water pressure xc2 rn2 P2 xc1 rn1 P1
1/4

(b) Different analytical material removal models for Water Jet Cleaning (WJCI) process M.C. Leu, P. When equivalent Cleaning width Jet spreading Meng, E.S. dynamic stress coefficient x 2/3 2/3 Geskin and L. generated by C and stress coefficient ; w 2Cx 1 xc Tismenesky water droplet l are independent of (1998) [35] flow exceeds the water pressure and Critical cleaning stand off distance (For stationary endurance limit nozzle radius. water jet) of the coating Jet radius in droplet lgwk 0.5 rn (Prj)0.25; xc 0.5 material under zone is related to (send)c C fluctuating stress. distance from the Mathematically it where k is flow resistance constant nozzle exit by canbeexpressedas in water jet system (=0.96 to 0.99), Rx=CX i.e. lmgw (send)c Mass flow rate in a ˙ water jet has the Ve k 2P/rj relationship
1.5 3

mxr ˙ mxc ˙ 1

r Rx

; and

2

mxc 5.62me ˙ ˙

rn Rx

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635
(continued on next page)

1613

1614

Table 6 (continued) Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Loss of coating material mass per unit area a
a lA r2n+2 0.5n+0.5 n (P ) f Rx

P. Meng, E.S. Geskin, M.C. Leu, F. Li and L. Tismenesky (1998) [36] (For moving water jet)
√ 2 2 Rx −y 1.5 3n+3

With n=2.875; optimal cleaning occurs at 0.588 times critical SOD. Critical SOD is proportional to f 0.148P0.287r1.148 n

y2+z2 1 Rx dz

When equivalent dynamic stress generated by water droplet flow exceeds the endurance limit of the coating material under fluctuating stress. Mathematically it canbeexpressedas lmgw (send)c ˙
√ 2 2 Rx −y

Analysis is applicable to normal incidence only. Computationally intensive. The relations are semiempirical and constant n is to be determined experimentally through regression analysis.

And critical stand-off-distance x2n+1 c where A=k1(7.95k)n+1(gw/suc)nrcr0.5n−0.5; j and B=2 (1 x1.5)3n+3dx; with x=r/Rx
0 1

lAB aof

r2n+2 0.5n+0.5 n P ; C2n+1

In addition to the assumptions made for stationary water jet following additional assumptions were also made: Springer’s semiempirical model for erosion by liquid droplets has been used. Incubation period has been neglected. There is complete cleaning for the concerned area if a ao, where ao is threshold value of a and is proportional to coating thickness.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

where k1 and n are empirical constants and l=la/lb=stress coef./strength coef.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1615

process include omni-directional cutting, minimal thermal damage, no burrs, high cutting speed and efficiency, suitability for automation, and cutting without delamination. The disadvantages include stray cutting, low nozzle life, high noise level and equipment cost, and hazards from rebounding abrasives. State of the art research and development that has taken place in the AWJM process has been mentioned in [37] covering aspects like physics of the process, operations, system and sensing, and new applications of it. The AWJM process uses a high velocity water jet in combination with abrasive particles for cutting different types of materials using a setup as shown in Fig. 4. A stream of small abrasive particles is introduced and entrained in the water jet in such a manner that water jet’s momentum is partly transferred to the abrasive particles. It has been concluded that the role of carrier fluid (water) is primarily to accelerate large quantities of abrasive particles to a high velocity and to produce a highly coherent jet [39]. The AWJM system includes the following elements: (1) pumping system to produce very high pressures, (2) abrasive feed system to deliver a precisely controlled stream of dry abrasives to the nozzle, (3) water jet nozzle to produce very high velocity water jet when pressurized water is passed through it, (4) abrasive jet nozzle to provide efficient mixing of abrasives and water jet to form a high velocity abrasive water jet, and (5) catcher to

Fig. 4. Basic scheme of abrasive water jet machining (AWJM) [38].

1616

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

contain the jet exiting from the workpiece and to minimize noise. Two major design concepts of abrasive jet nozzles are known as single-jet-side-feed nozzle and multiple-jet-central-feed nozzle. Important process parameters of AWJM can be categorized as hydraulic parameters: water pressure, and water flow rate (or water jet nozzle diameter); abrasive parameters: type, size, shape, and flow rate of abrasive particles; cutting parameters: traverse rate, stand-off-distance (SOD), number of passes, angle of attack, and target material; and mixing parameters: mixing method (forced or suction), abrasive condition (dry or slurry) and mixing chamber dimensions. The relationship between water pressure and depth of cut for different abrasive flow rates and nozzle diameters is almost linear. There is a minimum critical pressure below which no cutting occurs because a minimum abrasive velocity exists to cut a specific material. Water flow rate is directly proportional to the square root of water pressure and square of the water jet nozzle diameter. Depth of cut increases with an increase in water flow rate with decreasing slope as the saturation point is reached, while it varies linearly with water jet nozzle diameter for a given pressure [39]. Type of abrasive to be used depends on workpiece hardness. The harder the workpiece material the harder should be the abrasive. Particle fragmentation occurs irrespective of workpiece hardness. Commonly used abrasives are garnet, silica, and silicon carbide. Most commonly used abrasive mesh size is 100–150. An optimum abrasive particle size exists for a particular workpiece material and nozzle mixing chamber configuration. The range of particle size for optimum depth of cut is wider for brittle materials than for ductile materials. There seems to be no work available to quantify the effect of abrasive particles on depth of cut. Effect of particle size on the nozzle wear is also an important selection criterion [17]. The relationship between abrasive flow rate and depth of cut is linear up to a point. This linearity terminates at higher abrasive flow rates because particle velocity decreases more rapidly than the increase in number of impacts [38]. Depth of cut decreases with increasing traverse rate while an optimum traverse rate exists for maximum kerf area generation (=traverse rate times depth of cut). A minimum (or critical) traverse rate exists below which no further increase in depth of cut can be obtained. Cutting wear may often occur at low traverse rates and deformation wear at higher traverse rates [38]. An increase in stand-of-distance rapidly decreases machined depth, because with an increase the jet breaks into droplets resulting in free abrasive particles hence shallow penetration. A maximum value of SOD exists beyond which no cutting will take place This is used in cleaning and descaling applications [17]. Thus, a shorter SOD is desirable for higher productivity and better machining accuracy [42]. Depth of cut increases with an increase in number of passes. But an important issue is the rate of increase. The kerf cut during a pass, acts to focus the jet during the following pass and maintains the coherence of the jet [38]. Angle of impact determines the mode of material removal. For ductile materials, the highest material removal occurs at a small angle (i.e. 15–20°) of attack, while for brittle materials highest material removal is achieved at normal impact [42]. Properties of target material, particularly hardness and sensitivity to erosion by oblique and perpendicular impact, determine the mode and rate of material removal. Visual examination of the cutting process in AWJM suggests two dominant modes of material removal. First is erosion by the cutting wear mode due to particle impact at shallow impact angles on the top surface of the kerf. Second is deformation wear mode characterized by material removal due to excessive plastic deformation caused by particle impact at large impact angles, deeper into the kerf [38,39]. The analytical models for the AWJM process proposed by different investigators [38–42] to

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1617

predict the depth of cut and material removal rate are presented in Table 7. Hashish [38] used the erosion model of Finnie [18] to develop a model to predict combined depth of cut due to cutting and deformation wear for ductile materials only. The erosion model used in it gives a velocity index of 2 and does not incorporate the effects of abrasive particle size and shape. Hashish developed an improved erosion model [as reported in [39]] by expanding Finnie’s model [18] to include the effects of abrasive particle size and shape expressed by sphericity and roundness numbers. Hashish [39] used this model to predict depth of cut due to cutting wear, while the prediction of depth of cut due to deformation wear was based on Bitter’s model [19]. However, this model neglected the variation in kerf width along the depth of cut. Using the same improved erosion model, Paul et al. [40] developed an analytical model of generalized kerf shape for ductile materials considering variation in kerf width along its depth. The same authors [41] also developed a complex mathematical model for total depth of cut for polycrystalline brittle materials accounting for the effects of abrasive particle size and shape, but neglecting variation of kerf width along the depth of cut. Choi and Choi [42] developed an analytical model for AWJM of brittle materials. The expression developed by them to predict the volume of work material removed by a single abrasive particle is not in terms of process parameters, and moreover involves a constant of proportionality. 4. Advanced finishing processes High quality surface finish and accuracy are always desirable in a machined product as it gives improved product performance, minimum running-in-period and increased product life. Finishing operations represent the most crucial, expensive, and labor intensive area in the overall production process. Abrasive processes are invariably used for finishing operations as they provide a high level of surface finish and close geometric and size tolerances. Two distinct characteristics of the abrasive processes are very small size chips and self-sharpening of the cutting tools. The basic idea in these processes is to use a large number of cutting edges with random and indefinite orientation, and having variable geometry. The most commonly used abrasive fine finishing processes are lapping, honing, polishing, AFM, and MAF. There can be two extreme cases of contact between abrasive grains and work surface in the abrasive processes, i.e. (1) peaks of the grains enter into the peaks of the uneven surface of the workpiece. It is a salient feature of those abrasive finishing processes, which use fixed (or less strongly bonded) abrasives (i.e. grinding, lapping, honing, and polishing), or (2) peaks of the grains enter into the valley of the workpiece surface. It is characteristic of abrasive processes using free abrasives (AFM, MAP, etc.). Automation of the finishing process is required to improve process economy and consistency [44]. 4.1. Abrasive flow machining (AFM) AFM is one of the advanced finishing processes using abrasive particles mixed with carrier. This process was proposed around 1966 in the USA; initially slow developments took place with regards to the process performance parameters, physical models, surface characterization, abrasive grain wear, and process control. However, efforts in these directions picked up during the late eighties and early nineties. The ability of the finishing medium used in this process to reach

1618

Table 7 Different analytical material removal models for the abrasive water jet machining (AWJM) process Assumptions Limitations Remarks and/or conclusions It can be concluded that the role of carrier fluid is only to accelerate the abrasives to high velocities, as hydrodynamic forces acting on the abrasive particles are of less significance than inertial forces. Developed model uses the existing erosion theories. Its simplified equation requires more refinements.

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Total depth of cut due to combined Effects of cutting and deformation wear: hydrodynamic loading have been considered 2ma(P−Pc) C 2(1−C) ˙ according to the microh rjfdj(1+R1)2 4s 1000pe machining erosion model of Finnie (as sfw referred to in [38]). where s= at Ratio of vertical to horizontal resistance force is constant. Particle velocity remains unchanged over the shallow cuts produced by cutting wear. Angle of impact (a) changes linearly across the jet. Jet is two-dimensional and its width is equal to the jet diameter. This model is applicable for common ductile materials only. Role of abrasive particle size and its properties has not been considered. Effect of traverse rate is not hyperbolic as predicted by it.

M. Hashish (1984) [38]

Penetration process is cyclic and consists of two zones: cutting wear occurs when the particles impact the surface at shallow impact angles, and deformation wear occurs when particles impact at large impact angle.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635
(continued on next page)

Table 7 (continued) Assumptions Limitations Remarks and/or conclusions Developed model is of closed form. Analysis is based on an improved erosion model developed by expanding Finnie’s erosion model [18] Improved model includes particle shape and size expressed by sphericity and roundness numbers, and shows a velocity exponent of 2.5 sfw is related to the modulus of elasticity. Need is felt to characterize shape factors and breakage process of the abrasives.

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Based on improved erosion model, Width of cut is equal which gives volume removed as, to jet diameter and area of step cut is equal to 7ma va 2.5 cross-sectional area of V sin2a sina pra CK jet. Effects of Depth of cut due to cutting wear hydrodynamic loading is given by: on micro-cutting have been ignored as Cdj 14ma 2/5 vat ˙ ho concluded in [38]. 2.5 prafd 2 CK j Impact angle decays where CK is a characteristic linearly with distance velocity (in mm/s) and is given by from jet exit, i.e. CK=√3000sfwR0.6/ra f And based on Bitter’s model [19], a a 1 x t Cdj depth of cut due to deformation wear is given by: Particle velocity changes for small 1 hd depths of cut are 1000psfwfdj Cd vat + ignored while those for 2ma(vat−vac)2 dj (vat−vac) ˙ larger depths of cut are to be considered. Abrasive particles are uniformly distributed over cross-sectional area of jet and cutting is steady state. Jet spreading effects have been neglected. Volumetric MRR is linearly dependent on abrasive flow rate. Model is twodimensional, as jet penetration process has been considered as two-dimensional. Size distribution of abrasive particles is not considered. Effects of repeated impacts are not included in the improved erosion model.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

M. Hashish (1989) [39]

Penetration process is cyclic and consists of two regions: cutting wear occurs when the particles impact the surface at shallow impact angles, and deformation wear occurs when particles impact at large impact angle.

1619
(continued on next page)

1620

Table 7 (continued) Assumptions Limitations Remarks and/or conclusions Takes into account variation of kerf width along the depth of cut. Basic material property that determines total depth of cut is melting specific energy and not elastic modulus as concluded in [39]. Therefore, specific energy has been used instead of flow strength. No need to use threshold velocity concept in microcutting zone.

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Based on Hashish’s improved erosion model [39], depth of cut due to micro-cutting is given by: Kerf width is proportional to insert diamter, and gradually decreases to a 2.5 28maC vat ˙ minimum value which at1.5 hc 2.5praf(wt+wmin) CK occurs in the transition zone. where Velocity, mass flow rate, and energy of 3×106sswR0.6 f CK abrasives are uniformly ra distributed. Based on Bitter’s model [19], depth Cutting is steady state, of cut due to deformation wear is i.e. cutting front moves with a velocity equal to given by: the traverse speed of Q−hc(Qkdq−1) the jet. h; hd 1+Qkdq−Qk 2q2hc c d Impact angle decays linearly with the where distance from jet exit (1−C)ma(vat−vac)2 ¯ as assumed in [39]. ; Q Effective jet diameter 2wminfsfw is equal to the nozzle insert diameter. Only a part of jet vat ; q removes the material vat−vac by micro-cutting. Material removal and process is adiabatic as Cd in grinding without kd . dj ploughing and rubbing. Applicable for ductile materials only. Though present model takes care of kerf geometry, it is primarily twodimensional.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

S. Paul, A.M. Hoogstrate, C.A. Lutterwelt and H.J.J. Kals (1998) [40]

Ductile materials: At the top of kerf, MR is due to abrasive wear and microcutting caused by impacts of abrasive particles at shallow angles. At the bottom of kerf MR is due to plastic deformation caused by nearorthogonal impacts of the abrasive particles.

(continued on next page)

Table 7 (continued) Assumptions Limitations Remarks and/or conclusions Variation in particle shape and size has been considered. Threshold velocity has been assumed to be proportional to the specific energy or elastic modulus of work material.

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Depth of cut in the upper part of kerf (hu) due to erosion and cracking is given by: There is no particle fragmentation within the jet and upon the impact. Cma 14a1.5 va 2.5 ˙ t Mass flow rate and hu fdi 2.5pra CK energy of abrasive particles are uniformly 10p(1+nw)RfS 0.5Hkw(b3+b2m2)K f distributed. 18×106Ewgw Velocity of abrasive where K=v2a2a2Cfffaw; b3 and b2 a t particles does not are the constants mentioned in change due to kerf wall [41].Depth of cut in the lower part drag. of kerf (hl) due to plastic Velocities for water deformation and cracking, is and abrasive particles given by: are the same. Kerf width does not 4(1−C)ma (va−vac)2 ˙ hl vary and is equal to 106pfdi 2ssw nozzle insert diameter. 5ff rw 0.5 Hkw 1.5 1.5 −0.5 2 2 Cutting is steady state Cf Rf S f awva ; 3gw ra Ew i.e. cutting front moves with a velocity equal to the traverse speed of the jet. Impact angle decays linearly with the distance from jet exit as assumed in [39]. Only a part of the jet removes the material by micro-cutting. For brittle materials only. Essentially a twodimensional model. Variation in kerf width with depth of cut has been neglected. It can cause an error of up to 20% in predicting depth of cut.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

S. Paul, A.M. Hoogstrate, C.A. Lutterwelt and H.J.J. Kals (1998) [41]

Brittle materials: In the first zone, MR is due to micro-cutting and inter-granular fracture caused by impacts of the abrasive particles at shallow angles. In the second zone, MR is due to plastic deformation and inter-granular fracture caused by nearorthogonal impacts of the abrasive particles.

(continued on next page)

1621

1622

Table 7 (continued) Assumptions Limitations Remarks and/or conclusions It can quantitatively predict the large-scale fracture, which develops on the backside of the material as the jet passes through it. It also attempts to establish acoustic emissions (AE) as a basic tool for monitoring AWJM. It establishes a power function relationship between AE energy and process parameters of steady-state material removal process in AWJM. Material removal for brittle materials is hardly influenced by hardness of target material.

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Volume of material removed by single abrasive particle Both normal and tangential forces acting on the abrasive 1/9 11/9 22/9 2/9 22/9 Hvw m va tan ysin a 1021/9K1 particles are K 4/3 cw proportional to the where Hvw=Vicker diamond square of the hardness of the workpiece material penetration depth of the and K1 is constant. abrasive particle into Thickness of the fracture developed the target material. backside of the target material is Particle penetration into given by: the material is small as compared to its size. 2 2 1/3 (1−nw)PAm Material response is tf 0.05739 EwKcw similar to that of static indentation with a sharp indentor. Width of lateral crack can be approximated as the crack length under normal force. Applicable for brittle materials only. Dynamic effects of impact have been neglected.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

G.S. Choi and G.H. Choi (1997) [42]

Brittle materials: formation, propagation, and interaction of micro-cracks. Ductile materials: plastic flow, i.e. ploughing action of abrasive particles.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1623

conventionally inaccessible areas, to follow the complex contours, and to finish simultaneously on multiple edges and surfaces of the workpiece, make this process more versatile and unique than any other abrasive finishing process. The AFM process promises to provide accuracy, efficiency, economy, and the possibility of effective automation much needed by manufacturing industries. The AFM process can be used to deburr, polish, radius, remove recast layers, and induce residual compressive stresses. But it cannot repair surface defects and taper problems because it removes the material uniformly. Areas of application include aircraft valves, bodies and spools, turbine components, medical instruments’ components, automotive parts, semiconductor and pharmaceutical processing industries, electronic components, finishing of dies, finishing of phantom jets containing as many as 83 intersecting holes (peculiar application), etc. [43]. As depicted in Fig. 5, the AFM process involves the flow of an abrasive laden semisolid and self-deformable medium through (for internal surfaces) or around (for external surfaces) the workpiece surface(s). Abrasive flow machining action can be compared with that of a grinding or lapping operation. The AFM medium is extruded back and forth through the restrictive passage formed by the workpiece and/or tooling between two vertically opposed cylinders. The function of the tooling in the AFM process is to hold the part, direct the flow of the medium to the desired areas, and to form a restrictive passage. The AFM medium uniformly and gently removes the burrs, smoothes out the edges, and finishes the surfaces by removing surface peaks. Material removal is highest where restriction is greatest. The AFM medium is prepared by blending longchain polymer with gel like lubricant dilutent to give the desired viscosity. Medium exiting from

Fig. 5. Basic scheme of abrasive flow machining (AFM) [17].

1624

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

the workpiece is known as slug. Since abraded work material also becomes a part of the medium, the life of medium therefore depends on the amount of debris added to it, which in turn depends on a number of factors like batch quantity, type and size of abrasive particles, flow speed, etc. Energy supplied to the medium is mainly used in overcoming viscous resistance of medium to the flow, and in abrasion of workpiece surface, tooling, and medium containing cylinders [44]. Various parameters related to medium, machine setting, workpiece, and tooling affect the material removal and surface finish in AFM. Some of these are important controllable process parameters. Parameters related to the medium include type, size, and concentration of abrasives, and viscosity and rheological properties of carrier medium. Commonly used abrasives are alumina (Al2O3), silicon carbide (SiC), boron carbide (B4C), diamond, or a mixture of these. An increase in abrasive mesh size (or decrease in abrasive particle size), keeping other parameters unchanged, decreases MRR and surface roughness (or improves surface finish) due to decreased depth and width of indentation with an increase in abrasive mesh size. Therefore, coarse abrasives are recommended for higher MRR and fine abrasives for finer surface finish [48]. An increase in abrasive concentration in medium increases MRR and decreases surface roughness because more abrasive grains come into contact with the workpiece surface and cause more abrasion [47]. Variation of cumulative material removal against abrasive concentration shows the existence of an optimum value of abrasive concentration [48]. Viscosity of the medium is varied to control its flow pattern and to obtain the desired finishing results. High viscosity medium gives pure extrusion while low viscosity medium gives larger radii at passage openings [43]. Therefore, high viscosity medium is good for uniformly abrading walls of large passages, while low viscosity medium is better suited for finishing small passages and radiusing the edges. Increase in medium viscosity increases MRR and decreases surface roughness [43,44]. Parameters related to machine setting are number of cycles, extrusion pressure, speed and volume of flow of medium. MRR and surface roughness decrease with an increase in number of cycles. Initially the surface has sharp peaks, but once they are machined the surface becomes somewhat flatter than before and requires a higher tangential force, hence MRR decreases as the process continues [47]. An increase in extrusion pressure increases MRR and decreases surface roughness [43]. Viscosity of medium and extrusion pressure directly affect flow rate of medium, which in turn affects the amount of abrasion, uniformity of stock removal, and size of edge radius. Low flow rates of medium are good for uniform material removal while large flow rates produce larger radii [44]. An increase in flow speed of medium increases MRR and decreases the surface roughness value because it increases normal force on the abrasives and results in higher depth of indentation [47]. With an increase in flow volume of medium stock removal increases [44]. Workpiece related parameters are its initial surface roughness, hardness, stiffness, and shape. In AFM, higher peaks are abraded more than lower peaks and peaks are flattened into plateaus, therefore type of machining process used to prepare the workpiece prior to AFM, significantly affects the percentage improvement and improvement rate of surface finish in AFM. Amount of stock removed from milled surfaces and surfaces produced using wire electro-discharge machining (WEDM), is statistically different from that of turned and ground surfaces. WEDM surfaces were found better suited to AFM than turned and milled surfaces probably due to the weak junction between recast layer formed in WEDM and the parent material. But this needs further investigation [44]. Higher workpiece hardness gives relatively higher MRR and better surface finish. Parameters related to the AFM tooling are its material, shape, size, stiffness and stability, and medium flow passage. Commonly used tooling materials are steel,

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1625

nylon or urethane or their combination. Size, shape, and number of workpieces to be finished simultaneously, determine tooling size and shape. Tooling with larger passage area produces smaller slug length and lower stock removal. The nature of the AFM process is complex and stochastic due to non-newtonian flow behavior of the medium, variation in physical conditions of the medium due to changes in its viscosity, random location and movement of the abrasive grains, mixing of the workpiece material with the medium, and the random and complicated nature of material removal. This makes modeling of the AFM process complicated and difficult. Very few attempts have been made towards the modeling of material removal and surface finish obtained in AFM. A stochastic modeling and analysis technique known as the Data Dependent System (DDS) has been used to study the workpiece surface profile before and after AFM [43]. Wavelength decomposition of AFM generated surfaces indicated the presence of two wavelengths, the larger wavelength associated with the main path of the abrasive grains and a smaller wavelength associated with cutting edges. Information about frequencies and wavelengths of these waves has been used to estimate the number of dynamic active grains and abrasive grain wear. Both these methods can be useful in off-line optimization and on-line adaptive control or tool (or abrasive grain) replacement strategy in the AFM process [43]. Different analytical [45,46] and empirical [47] material removal models developed for the AFM have been presented in Table 8. Kumar [45] developed an MRR model for axi-symmetric shapes, which uses the finite element method (FEM) to calculate normal stress acting on the abrasive grains during machining. Jain et al. [46] also reported another FEM-based model to predict material removal and surface roughness generated during AFM of axi-symmetric shapes. The same authors used multi-variable regression analysis (MVRA) and neural networks for the empirical modeling of the AFM process [47]. 4.2. Magnetic abrasive finishing (MAF) MAF was mentioned for the first time in a patent by Henry P. Coats in 1938 [51]. Despite its origination in the USA, most of the developments in MAF took place in the former USSR, Bulgaria and Japan. MAF is not a sizing operation but it can be successfully used to eliminate the out-of-roundness (OOR) error. An important feature of MAF is its insensitivity towards runout, i.e. to eliminate OOR, it does not require that runout of the machine tool–workpiece system must be less than the specified OOR errors. It can also be used for finishing the workpieces machined even by rough turning operation, without using any intermediate operation like grinding, honing, lapping, and so forth [51]. Another merit is feasibility to control machining pressure by controlling the input current. Typical MAF configurations for flat surfaces [Fig. 6(a)], for internal finishing [Fig. 6(b)] and external finishing [Fig. 6(c)] of the rotary parts are detailed. In MAF, the workpiece to be finished is located between two magnetic poles. The gap between the poles and workpiece is filled with magnetic abrasive particles, which consist of abrasive grains and ferromagnetic particles. The necessary finishing pressure is applied by a magnetic field produced by electromagnets. As a result of this, magnetic abrasives join each other magnetically to form flexible magnetic abrasive brushes. When the workpiece and magnetic pole are given relative movement of rotary, vibratory and axial type, abrasive grains remove the peaks of workpiece surface unevenness and the work surface is finished [44]. In MAF, the cutting force attains maximum value when the cutting angle

1626

Table 8 Different analytical material removal models for the abrasive flow machining (AFM) process Assumptions Limitations Remarks and/or conclusions Computationally extensive as sr has to be calculated using the Finite Element Method (FEM). Analytical results do not agree well with the experimental results. Constant K is to be determined through the experiments.

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Total volume of material removed in n number of cycles: All abrasive particles are of same size, spherical in shape and 6 rm lsvf sr2 are evenly distributed K CnVm 16 ra lwvpH2 w in the media. where Vm is the volume of abrasive Normal force acting on each particle is the media between workpiece and same. −2 K=K1K 2 in which K1 is a constant Material removed by representing the % of grains participating in cutting action, and each particle in each K2 is a constant relating flow stress stroke is the same. Media behaves as a to BHN of workpiece material viscoplastic material. (K2=1 for brittle materials and Media flow is K2 1 ductile materials) incompressible, steady state and axisymmetric. Coulomb friction exists between the media and workpiece. Process is isothermal and temperature rise of media during machining is negligible. Applicable to axisymmetric components only as media flow has been assumed to be axi-symmetric.

T.R. Kumar (1998) [45]

Indentation of abrasive particles in the workpiece surface and susequently particle movement along the scratch length

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635
(continued on next page)

Table 8 (continued) Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Applicable to axisymmetric components only as media flow has been assumed to be axi-symmetric. Analysis is computationally extensive as sr has to be calculated using the Finite Element Method (FEM)

R.K. Jain, V.K. Jain and P.M. Dixit (1999) [46]

Ductile Total volume removed in n number materials: Both of cycles micro-ploughing 2n and microNsAg Li; cutting. i 1 Brittle materials: where Micro-cutting only. Ria R2 m Ns 2pNls ; Li 1 o lw; Rw Ra

Ag where a=√h(dm−h); and h Surface roughness after ith stroke (Ria) is given by:
i Ra Ri−1 a

d 2 −12a dm m a sin h 4 dm 2 ;

Analytical results agree well with the experimental results of Jain and Adsul [48]. Manner of computation of N has not been mentioned. Ratio of Rmax to Ra equal to 7 has been used while DDS result of Williams and Rajurkar [43] predicts it in the range of 1.4– 2.2. However, media compositions in both cases are very different.

dm 2

d 2 srd 2 m m − 4 4Hw

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

NlsR2 m A 7R2 g w

All abrasive grains are of the same size, irregular but spherical in shape with large negative rake angles. Each grain consists of single cutting edge only. Load on each particle is constant and is equal to average load. Every abrasive grain achieves same depth of penetration. Media flow is steady, incompressible and axisymmetric. Media behaves as Newtonian-fluid, and is isotropic and homogeoeous. Media properties remain constant with temperature, time, and space. Workpiece surface has uniform surface profile without statistical distribution. Abrasives move in a direction normal to scratches. Ratio of Rmax to Ra is 7 for the AFM generated surfaces.

1627
(continued on next page)

1628

Table 8 (continued) Assumptions Limitations Remarks and/or conclusions Results of this empirical model agree well with the experimental neural network results of the same authors. Unit of MRR used in regression analysis is grams per minute and derived empirical relation also gives MRR in the same units. Use of volumetric units would have been more appropriate as it eliminates dependence on material density.

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Experimental results Empirical relations are with workpiece as not of general nature. MRR 5.5285 aluminium have been used to find the 10−10n−0.1893C3.0776M−0.9371v1.6469 f relations. Surface roughness value in µm Multi Variable Regression Analysis Ra 2.8275 (MVRA) has been used 105n−0.2258C−1.3222M0.1368v−1.8221 for empirical modeling. f Method of least squares Here media flow velocity (vf) is in has been used to cm/min. calculate the regression coefficients. MRR in grams per minute

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

R.K. Jain, V.K. Jain and P.K. Kalra (1999) [47]

Ductile materials: Both micro-ploughing and microcutting. Brittle materials: Micro-cutting only.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1629

Fig. 6. Scheme of magnetic abrasive finishing (MAF): (a) plane finishing [44]; (b) internal finishing [49]; (c) external finishing [49].

1630

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

a is minimum (=tan−1l/2A). Smooth surfaces with a large ratio of wavelength to amplitude experience small cutting forces, while rough surfaces with a small ratio of wavelength to amplitude experience large cutting forces. Also, MRR at the higher peaks of unevenness is higher than the MRR at the lower peaks of unevenness. This reduces the distance between peaks and valleys of the machined surface, and consequently produces a smooth surface with low OOR. Important process parameters of MAF are: magnetic abrasive type and size, concentration of abrasive in ferromagnetic material, working gap between workpiece and magnetic poles, magnetic flux density, relative motion between magnetic abrasive particles and workpiece surface (or rotational speed), amplitude and frequency of vibrations, and axial movement or feed. Type of magnetic abrasives used in MAF can either be unbonded (a mixture of abrasives and ferromagnetic material) or bonded (abrasives held in a ferromagnetic matrix formed by sintering or other techniques). Experimental investigations have revealed that unbonded magnetic abrasives yield higher MRR while bonded magnetic abrasives produce a better surface finish [44]. Both the relative size and mixing proportion of abrasives and ferromagnetic materials influence the finishing characteristics of the surface produced. Larger particle size gives higher stock removal but rougher surface finish. Also, an optimum value of mixing ratio of abrasives and ferromagnetic particles exists to achieve the best surface finish and highest machined depth. An increase in magnetic flux density increases magnetic abrasive pressure and hence machined depth while surface finish also improves. Since magnetic flux concentrates on the edges of the ferromagnetic substance in the magnetic field, ground burrs can be removed more efficiently than the turned burrs. A decrease in the working gap increases the machined depth and also improves the surface finish. The literature review reveals that like AFM, not much work has been done on the modeling of material removal and surface finish in MAF. As of now, three models [49–51] are available and they are presented in Table 9. Kremen et al. [49] developed an empirical expression to estimate the machining time to produce a workpiece with specified roundness, while Kim and Choi [50] reported the modeling and simulation of surface roughness and MRR in MAF and used this to predict finishing time. Kremen et al. [51] proposed a theory, independent of size and material of workpiece, and design of MAF machine tool, to explain the out-of-roundness phenomenon based on the analysis of machining forces and material removal mechanisms. 5. Conclusions Use of advanced machining processes incurs high investment, operating, maintenance, tooling, and other costs. Optimal setting of process parameters is essential to achieve economic and efficient process performance. Setting of optimal process parameters depends on the selection of an appropriate analytical model for material removal and surface generation having wider applicability, which requires comparison of the applicability and limitations of the various models. In the absence of analytical models, optimum selection of process parameters requires extensive experimentation, which is time and money consuming. This paper has presented a comprehensive literature review of various analytical material removal models (and of surface generation for advanced finishing processes) developed until recently for different mechanical type advanced machining processes, in a format suitable for quick reference. In addition to this, brief process introduction, working principle, materials and

Table 9 Different analytical material removal models for the magnetic abrasive finishing (MAF) process Assumptions Limitations Remarks and/or conclusions Hyperbolic relation between normalized OOR errors and finishing time has been verified. Reduction rate of normalized OOR decreases with finishing time.

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Time required to reduce initial out- Based on Kremen’s Proposed relation to of-roundness (OOR) by the Law of Change of estimate finishing time specified amount Geometrical Form is empirical Solids i.e. Normalized 1 1 OOR errors in a crosst 1 ection of a cylindrical a On(t)1/b workpiece finished where a and b are experimentally using MAF, follow a determined constants, which depend hyperbolic decay, i.e. on the hardness and size of O(t) 1 workpiece, magnetic field density, On(t) and magnetic powder properties. Oi (1+at)b

G.Z. Kremen, Not mentioned E.A. Elsayed and J.L. Ribeiro (1994) [49]

J.D. Kim and M.S. Choi (1995) [50]

Mechanism of micro-cutting

Total stock removal by magnetic Abrasives move in a abrasive brush in the machining time direction normal to the t, is given by: scratch length. Workpiece surface has 2 Nn fvt uniform profile without 3 o −1 V 10 K1 (Ra ) ; statistical distribution. pHwlwtanq Cross-sectional area of and surface roughness value after magnetic brush is same machining for time t, is given by: as that of air-gap. Leakage of magnetic 1 Nn fvt field is negligible. Ra(t) Ro K1 a lw pHwlwtanq Magnetic core is saturated uniformly where K1 is a constant of throughout the crossproportionality. section.

Simulation results for surface roughness agree with experimental data more closely for lower magnetic flux density than they do for high magnetic flux density

Machining pressure between magnetic brush and workpiece has maximum value at about B=1.2 Tesla. Magnetic flux density in air-gap is greatly influenced in a reverse manner by the length of air-gap. Finishing efficiency is very sensitive to magnetic flux density. (continued on next page)

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635 1631

1632

Table 9 (continued) Assumptions Limitations Remarks and/or conclusions

Name of Mechanism of Proposed analytical model for investigator(s) material removal material removal and year [Ref.] Machining time to achieve specified OOR t Linear MRR Vl= egt(gOi+Vlt) ; Here K2=(Vtl/gOi); and g is a constant dependent on machining conditions. −1 On(t)+K2 ; loge g 1+K2 Magnetic force on the magnetic abrasive grains, coefficient of friction, and frictional force remain constant.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

G.Z. Kremen, E.A. Elsayed and V.I. Rafalovich (1996) [51]

Not mentioned

Theoretical results for OOR agree more closely with experimental results of [49] when OOR reaches its asymptotic value

MAF has been analyzed as sizing operation. Theory is independent of workpiece size and material, and MAF machine tool design. Limit of finishing accuracy is K2 →0

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1633

shape applications of the processes, parametric analysis, and state-of-the-art of MRR modeling for the concerned process has also been presented. Such a review can be of great help in understanding and selecting the most appropriate process model to perform the tasks of parametric optimization, process performance prediction, process automation and control, and devising a strategy to improve process performance.

Acknowledgements The authors acknowledge financial support by the Department of Science and Technology (DST), Government of India, New Delhi, for the project entitled “Automated Process Selection and Optimization of Advanced Machining Processes” under grant No. III.5 (199)/99-ET.

References
[1] P.M. Khodke, D.J. Tidke, Abrasive jet machining — a state-of-art review, Journal of Institution of Engineer (India) 77 (1996) 1–8. [2] T.B. Thoe, D.K. Aspinwall, M.L.H. Wise, Review on ultrasonic machining, Int. J. Mach. Tools Manufacture 38 (1998) 239–255. [3] P.C. Pandey, H.S. Shan, Modern Machining Processes, Tata McGraw-Hill Publishing Company Ltd, New Delhi, 1977. [4] A. Ghosh, A.K. Mallik, Manufacturing Science, Affiliated East-West Press Ltd, New Delhi, 1985. [5] G.E. Miller, Special theory of ultrasonic machining, J. Appl. Phys. 28 (1957) 149–156. [6] L.D. Rosenberg, V.F. Kazantsev, L.O. Makarov, D.F. Yakhimovich, Ultrasonic Cutting, Consultant Bureau, New York, 1964. [7] N.H. Cook, Manufacturing Analysis, Addison-Wesley, Reading, MA, 1966. [8] G.S. Kainth, A. Nandy, K. Singh, On the mechanics of material removal in ultrasonic machining, Int. J. Mach. Tool Des. Res. 19 (1979) 33–41. [9] E.V. Nair, A. Ghosh, A fundamental approach to the study of the mechanics of ultrasonic machining, Int. J. Prod. Res. 23 (1985) 731–753. [10] J. Saha, A. Bhattacharya, P.K. Mishra, Estimation of material removal in USM process — a theoretical and experimental study, in: 27th MTDR Conference, 1988, pp. 275–279. [11] H. Kamoun et al., Modeling the material removal in stationary mode for ultrasonic contour machining, in: ASME Winter Annual Meeting, 1993, pp. 759–770. [12] Z.Y. Wang, K.P. Rajurkar, Dynamic analysis of the ultrasonic machining process, Trans. ASME, J. Mfg. Sci. Engng. 118 (1996) 376–381. [13] T.C. Lee, C.W. Chan, Mechanism of ultrasonic machining of ceramic composites, J. Mater. Process. Technol. 71 (1997) 195–201. [14] D. Prabhakar, Z.J. Pei, P.M. Ferreira, M. Haselkorn, A theoretical model for predicting material removal rates in rotary ultrasonic machining of ceramics, Trans. NAMRI/SME XXI (1993) 167–172. [15] Z.J. Pei, D. Prabhakar, P.M. Ferreira, M. Haselkorn, A mechanistic approach to the prediction of material removal rates in rotary ultrasonic machining, Trans. ASME, J. Engng. Ind. 117 (1995) 142–151. [16] Z.J. Pei, P.M. Ferreira, Modeling of ductile mode material removal in rotary ultrasonic machining, Int. J. Mach. Tool Manufacture 38 (1998) 1399–1418. [17] V.K. Jain, Advanced Machining Processes, Allied Publishers, Mumbai, 2001 (in press). [18] I. Finnie, Erosion of surfaces by solid particles, Wear 3 (1960) 87–103. [19] J.G.A. Bitter, A study of erosion phenomena — I and II, Wear 6 (1963) 5–21.

1634

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

[20] G.L. Sheldon, I. Finnie, The mechanism of material removal in the erosive cutting of brittle materials, Trans. ASME, J. Engng. Ind. 88 (1966) 393–400. [21] J.H. Neilson, A. Gilchrist, Erosion by a stream of solid particles, Wear 11 (1968) 111–122. [22] G.L. Sheldon, A. Kanhere, An investigation of impingement erosion using single particles, Wear 21 (1972) 195–209. [23] B.R. Lawn, A model for brittle solids under fixed abrasive conditions, Wear 33 (1975) 369–372. [24] P.K. Sarkar, P.C. Pandey, Some investigations on abrasive jet machining, J. Inst. Eng. (India) 56 (1976) 284–287. [25] M.L. Neema, P.C. Pandey, Erosion of glass when acted upon by abrasive jet, in: Proceedings of International Conference on Wear of Materials, 1977, pp. 387–392. [26] M.A. Moore, F.S. King, Abrasive wear of brittle materials, Wear 60 (1980) 123–140. [27] M. Hutching, A model for the erosion of metals by solid particles at normal incidence, Wear 70 (1981) 269–281. [28] D.B. Marshall, A.G. Evans, M.E. Gulden, J.L. Roubort, R.O. Scatergood, Particle size distribution effects on the solid particle erosion of brittle materials, Wear 71 (1981) 363–373. [29] G. Sundararajan, An empirical relation for the volume of crater formed during high velocity oblique impact tests, Wear 97 (1984) 9–16. [30] K.N. Murthy, D.C. Roy, P.K. Mishra, Material removal of brittle materials in abrasive jet machining, in: Proceedings of 12th AIMTDR Conference, 1986, pp. 321–324. [31] P.K. Jain, A.K. Chitley, N.K. Nagar, Theoretical and experimental investigation for MRR in abrasive jet machining, in: Proceedings of the 14th AIMTDR Conference, I.I.T., Delhi, 1990, pp. 273–278. [32] P.K. Mishra, Nonconventional Machining, Narosa Publishing House, New Delhi, 1997. [33] M. Hashish, M.P. DuPlessis, Theoretical and experimental investigation of continuous jet penetration of solids, Trans. ASME, J. Engng. Ind. 100 (1978) 88–94. [34] M. Hashish, M.P. Du Plessis, Prediction equations relating high velocity jet cutting performance to stand-offdistance and multipasses, Trans. ASME, J. Engng. Ind. 101 (1979) 311–318. [35] M.C. Leu, P. Meng, E.S. Geskinand, L. Tismeneskiy, Mathematical modeling and experimental verification of stationary waterjet cleaning process, Trans. ASME, J. Mfg. Sci. Engng. 120 (1998) 571–579. [36] P. Meng, E.S. Geskin, M.C. Leuand, L. Tismeneskiy, An analytical and experimental study of cleaning with moving waterjets, Trans. ASME, J. Mfg. Sci. Engng. 120 (1998) 580–589. [37] R. Kovacevic, M. Hashish, R. Mohan, M. Ramulu, T.J. Kim, E.S. Geskin, State of the art of research and development in abrasive waterjet machining, Trans. ASME, J. Mfg. Sci. Engng. 119 (1997) 776–785. [38] M. Hashish, A modeling study of metal cutting with abrasive water jets, Trans. ASME, J. Engng. Mater. Tech. 106 (1984) 88–100. [39] M. Hashish, A model for abrasive water jet (AWJ) machining, Trans. ASME, J. Engng. Mater. Tech. 111 (1989) 154–162. [40] S. Paul, A.M. Hoogstrate, C.A. van Luttervelt, H.J.J. Kals, Analytical and experimental modeling of abrasive water jet cutting of ductile materials, J. Mater. Process. Technol. 73 (1998) 189–199. [41] S. Paul, A.M. Hoogstrate, C.A. van Luttervelt, H.J.J. Kals, Analytical modeling of the total depth of cut in abrasive water jet machining of polycrystalline brittle materials, J. Mater. Process. Technol. 73 (1998) 206–212. [42] G.S. Choi, G.H. Choi, Process analysis and monitoring in abrasive water jet machining of alumina ceramics, Int. J. Mach. Tools Manufacture 37 (1997) 295–307. [43] R.E. Williams, K.P. Rajurkar, Stochastic modeling and analysis of abrasive flow machining, Trans. ASME, J. Engng. Ind. 114 (1992) 74–81. [44] R.K. Jain, V.K. Jain, Abrasive fine finishing processes — a review, Int. J. Mfg. Sci. Prod. 2 (1999) 55–68. [45] T.R. Kumar, Theoretical and experimental investigations into abrasive flow machining process, M. Tech. Thesis, I.I.T. Kanpur, July 1998. [46] R.K. Jain, V.K. Jain, P.M. Dixit, Modeling of material removal and surface roughness in abrasive flow machining process, Int. J. Mach. Tools Manufacture 39 (1999) 1903–1923. [47] R.K. Jain, V.K. Jain, P.K. Kalra, Modeling of abrasive flow machining process: a neural network approach, Wear 231 (1999) 242–248. [48] V.K. Jain, S.G. Adsul, Experimental investigation into abrasive flow machining, Int. J. Mach. Tools Manufacture 40 (2000) 1003–1021.

N.K. Jain, V.K. Jain / International Journal of Machine Tools & Manufacture 41 (2001) 1573–1635

1635

[49] G.Z. Kremen, E.A. Elsayed, J.L. Ribeiro, Machining time estimation for magnetic abrasive processes, Int. J. Prod. Res. 32 (1994) 2817–2825. [50] J.D. Kim, M.S. Choi, Simulation for the prediction of surface accuracy in magnetic abrasive machining, J. Mater. Process. Technol. 53 (1995) 630–642. [51] G.Z. Kremen, E.A. Elsayed, V.I. Rafalovich, Mechanism of material removal in the magnetic abrasive process and the accuracy of machining, Int. J. Prod. Res. 34 (1996) 2629–2638.

Sign up to vote on this title
UsefulNot useful