prediction of surface roughness in machining of homogenized sicp reinforced aluminium metal matrix...

International Journal of Manufacturing Science and Technology4(2) June 2010; pp. 81-91

© Serials Publications

* Assistant Professor, Department of Mechanical Engineering, GRRIET, Hyderabad-500090, India, (E-mail:[email protected])

** Associate Professor, Department of Mechanical Engineering, JNT University, Kukatpally, Hyderabad- 500 85India.

*** Professor, Department of Mechanical Engineering, JNT University, Kukatpally, Hyderabad-50085, India

Prediction of Surface Roughness in Machining of HomogenizedSiCP Reinforced Aluminium Metal Matrix Composites UsingMultiple Regression Analysis

Dareddy Ramana Reddy*, Banoth Balunaik** and T. Kishen Kumar Reddy***

Abstract: Metal-matrix composites are a relatively new range of materials possessing severalcharacteristics that make them useful in situations where low weight, high strength, high stiffness,and an ability to operate at elevated temperatures are required. The machining of aluminium matrixcomposites reinforced with particulates (MMCp) causes problems because of rapid tool wear dueto the extremely high hardness of particles such as silicon carbide and aluminium oxide. Practicalconsiderations for machining MMCp include selection of the cutting speeds, type of grinding wheeland surface roughness. These materials are difficult to machine. This paper presents results froman ongoing investigation into the factors affecting the finish machining of an Al/SiC MMC. Diamondgrinding wheels were used to machine an Al-SiCp (30 vol. %) metal-matrix composite. The influenceof the cutting speed, feed and depth of cut on the tool wear, the surface finish, and the cutting forceswas established. It was found that good grinding performance can be achieved in terms of lowerorder grinding force components and low AE energy release. The underlying machining mechanismappeared to be quite different from that of an aluminium alloy similar in composition and mechanicalproperties to that of the aluminium matrix material of the MMC. The automation and optimizationof manufacturing process play an important role in improving productivity. Using the data obtainedduring the grinding of MMCs, the process was modeled for estimation of surface roughness and toidentify the process status using regression analysis.

Keywords: A. Composites; B. Grinding; C. Statistics, Multiple regression analysis

1. INTRODUCTION

Today’s technological progress is due to sophistication in materials. As the technology becomesmore and more advanced, the materials used have to be reliable and more efficient inperformance. Materials should be light in weight, strong, tough, wear resistant and capable ofwithstanding extreme operating environments such as high temperature, pressure, cryogenicconditions, high vacuum, highly corrosive and in some cases electric, magnetic or irradiationfields. No individual material is capable of meeting these demands. For example, steel hasbeen the most widely used engineering material due to its many advantages like high stiffness,high strength, ease of joining, high toughness and low cost; but has the disadvantage that it isheavy (7800 Kg/m3) and is prone to corrosion. Aluminum is lighter (2700 Kg/m3) and can bemade as strong as steel but cannot be welded easily. Thus it is seen that the material system

82 / International Journal of Manufacturing Science and Technology

should preferably be designed for specific end use (tailor-made) necessitating a combinationof desired properties. This has resulted in the development of advanced materials and compositesform one class of such materials developed using modern innovative techniques. Compositesrepresent a combination of at least two chemically distinct materials with a distinct interfaceseparating the constituents. These are three kinds of them: Polymeric matrix (PMC such asFRP), Metal matrix (MMC) and Ceramic matrix (CMC) materials. Their high strength to weightratio, enhanced resistance to environmental hazards, lower density, high fatigue resistance,wear resistance and related properties has widened the range of application leading to largescale substitution of conventional engineering materials for Figure 1 Comparison betweenmonolithic materials and composite materials aerospace to consumer goods. Figure 1 representsa comparison of properties between monolithic materials and composite materials. The specificmodulus (the ratio between the Young’s modulus (E) and the density (ρ) or specific weight(w) of the material) and specific strength (the ratio between the strength (σ

ult) and the density

(ρ) or specific weight (w) of the material) are high for composite materials compared to metals.This results in reduced space requirements and lower material and energy costs.

Figure 1: Comparison Between Monolithic Materials and Composite Materials

Metal matrix composites are either in use or being prototyped for the space shuttle,commercial airlines, electronic substrates, bicycles, automobiles, golf clubs and many otherapplications. The majority are aluminum matrix composites, a growing number of applicationsrequire the matrix properties of super alloys, Titanium, copper, magnesium. The term metalmatrix composite (MMC) covers a wide range of scales and micro structures. Common tothem all is a metallic matrix, which is normally contiguous. The reinforcing constituents are inmost cases a ceramic, intermetallics or semi-conductors [1]. Particulate reinforced Al-metalmatrix composite (PRAlMMC) is one of the important composites among the metal matrixcomposites, which have SiC particles with aluminium matrix is harder than tungsten carbide(WC), which pose many problems in machining. The aluminium alloy reinforced withdiscontinuous ceramic reinforcements is rapidly replacing conventional materials in variousautomotive and aerospace industries. Al-SiC-MMCs machining is one of the major problems,while resist its wide spread engineering application. It is found that surface finish is very poorwhile carbide tip tools are used for machining [2].Hence effective machining with generation

Prediction of Surface Roughness in Machining of Homogenized SiCP Reinforced… / 83

good surface finish on the Al-SiC-MMC jobs during grinding operations is a challenge to themanufacturing engineers.

The hard reinforcement in a MMC provides them, the preferred higher wear resistance.However it is detrimental to cutting tools and forming dies. Heterogeneity and anisotropy ofMMCs make their machining significantly different from that of conventional metals and alloys.The tool continuously encounters alternate matrix and reinforcement whose response tomachining is entirely different and depends on diverse reinforcement and matrix properties.Fiber orientation/particle distribution, relative volume and size of reinforcement in MMCscause severe force fluctuations posing fracture toughness and fatigue related machiningproblems. Accordingly the requirement on the part of the cutting tool changes continuouslyand it is this variation in the requirement of the tool that makes MMCs difficult to machine.Also conventional machining of MMCs is difficult due to the presence of comparatively highvolume fraction of hard ceramic reinforcement which causes rapid abrasive tool wear duringmachining resulting in very short tool lives. Although the latest innovative manufacturingprocesses such as in-situ MMC development and Thixotropic processing can produce near netshape components, final machining and finishing processes like grinding are still required tofabricate MMC component to the required dimensional tolerance/surface texture. For effectivemachining of MMCs ultra hard tool materials like PCD and PCBN are necessary. PCD andPCBN tools are used for grinding MMCs. Trials were carried out with conventional siliconcarbide grinding wheel also to explore the cutting requirements for the same while machiningMMCs [4].

A burning phenomenon of a workpiece is one of grinding faults that happen many times tothe ground surface. The grinding burn is a discoloration phenomenon according to thicknessof oxide layer on the ground surface. At the onset of grinding burns, a surface roughnessdeteriorates [5]. In view of these above mentioned machining problems, main objectives of thepaper is to study the influence of different cutting parameters like cutting speed, feed rate,depth of cut o the machinability characteristics like surface finish during grinding of AL/SiCp-MMC. The surface finish for different sets of experiments were examined and compared forsearching out the suitable parameter combination through highlighting the drawbacks andsuggesting proper measures to be undertaken during machining performance which mayovercome the machining barriers from Al-SiC-MMC. Since grinding is a complexmanufacturing process with a lot of parameters which influence each other, modeling can beuseful tool to the comprehension of the process itself. In this work some empirical relationshipsfor predicting how the grinding parameters affect the workpiece surface roughness are presented.

2. EXPERIMENTAL PROCEDURE

2.1. Workpiece Material

Metal matrix composites are tailor made to suit the requirements by changing the type, shape,size and volume fraction of reinforcement. Thus a large number of varieties exist. For thepresent study on the characterization and monitoring of grinding of MMCs, MMC with 2124Alas the matrix reinforced with 30% volume fraction of silicon carbide particles processed throughpowder metallurgy route is used. The average size of the reinforcement particle is 5 mm. Somedetails of Al2124 alloy are given below in Table 1.


Table 1Some Details of Al-2124 Aluminum Alloy

Composition Al-4.0%Cu-0.5%Mn-0.3%Fe-1.5%Mg

Density 2.78 g/cc

Modulus of Elasticity 73 GPa

Coefficient of thermal expansion 22.9 mm/m-oC

Thermal conductivity 193 W/m-K

2.2. Grinding Wheels

Resin bonded diamond wheels of 80/100, 120/140, 170/200 grit sizes.

All wheels were of 150mm outer diameter. Single point diamond dressing tool isused to dress the silicon carbide wheel and a ceramic stick to dress the super-abrasivewheels.

2.3. Experimental Setup Details

Experiments were carried out on a SHUTTE WU3mS type tool and cutter surface-grindingmachine fitted with an automatic table traverse unit. The grinding conditions in whichexperiments are carried out were listed in Table 2. All the experiments were carried out in dryconditions.

Table 2Grinding Conditions

Parameter Range

Wheel speeds (Speed) 1000 mpm, 1400 mpm, 2800 mpm,

Table Speed (Feed) 0.2 mpm, 0.6 mpm, 0.9 mpm

Depth of Cut (DoC) 10 mm, 20 mm, 30 mm

2.4. Experimental Procedure

Initial grinding trials are carried out using Diamond wheel of 120/140 grit to assess the grindingperformance while grinding Al2124/SiCp metal matrix composite in dry conditions withconditions listed in table 2.2.Data on surface roughness is collected, analyzed and performancefor Diamond wheel of 120/140 grit was evaluated.

3. RESULTS AND DISCUSSION

The performance index in any manufacturing system can be broadly classified into two classes:system performance index and process performance index. In order to maintain the desiredperformance level, it is necessary to monitor performance indicators such as cutting forcecomponents, acoustic emission signal emanating from cutting tool, cutting temperature. Theperformance of cutting tool can be evaluated by monitoring any of these indicators apart fromsurface quality of the machined part. Depending on the requirement, either one or more ofthese indicators are monitored.


3.1. Performance of Diamond 120 Grinding Wheel

Grinding trials on Al2124/SiCp metal matrix composite were carried out using PCD wheels.The performance of the grinding wheel was assessed in terms of grinding force components,acoustic emission signal characteristics and surface finish of ground surfaces.

3.1.1. Observations on Grinding Forces

Typical observed parametric influence on grinding force components when grinding withDiamond 120 wheel is shown in Figure 2. It is seen that with table traverse mostly the grindingforce components show an increasing trend. With low table traverse, a visible reduction ingrinding force can be seen with higher speeds of grinding. At lower speed of 500mpm, theforces are high compared to a higher speed of 1000mpm. This is due to the reduced penetrativeeffect of the abrasive at lower speeds. At higher speed of 2800mpm, the grinding forces reduceddue to thermal softening of the material.

(a) Tangential Component (b) Normal Component

Figure 2: Parametric Influence of Grinding Conditions on Grinding Force ComponentsWhen Grinding with Diamond 120/140 Wheel


3.1.2. Observations on AE Monitored

Acoustic emission is high frequency (transient) elastic stress wave emitted by a material understate of stress; hence it indicates clearly the status of the material under stress. The response ofthe Al/SiC

p composite to grinding environment has been assessed by monitoring the AE signal

from the work piece. The raw signal was acquired by a suitable AE sensor and was characterizedin terms of rms values.

Grinding is usually carried out on hard ferrous materials. Relatively softer non-ferrousmaterial usually is not an ideal material for abrasion. However composite materials of a non-ferrous material matrix (like aluminium) pose problems in turning and related processing dueto the presence of harder reinforcement particles. Hence grinding is resorted to. During grindingof Al/SiC

p composites one can anticipate ploughing and smearing of softer aluminium matrix,

loading of the wheel, pull-out/dislodgement of reinforcement (SiCp) and spalling of loaded

material. This results in varied/ inconsistent material response.Typical observed AE rms values are monitored at different grinding conditions are shown

in Figure 3. At lower wheel speeds only marginal change in AE power. But at higher wheelspeed, higher power of emission due to chipping of diamond. There is a marginal increase inAE rms with increase in table traverse. At 0.9mpm table traverse and higher wheel speed of2800mpm, 20ìm depth of cut, the AE rms is high due to self sharpening of the abrasive grit.This can be attributed from the higher force values.

Figure 3: Parametric Influence of Grinding Conditions on AE rms when Grinding withDiamond 120-140 Wheel


3.1.3. Observations on Temperature and Surface Finish

Typical observed parametric influence of grinding zone temperature monitored during grindingis shown in Figure 4. At low grinding depths and higher table traverse rate, there is a progressiverise in temperature but with increase in grinding depth, there is a gradual rise in temperature.With increase in wheel speed, there is an increase in grinding temperature which consequentlyaffects the surface finish. At higher speeds, the surface finish is poor owing to higher temperaturewhich leads to thermal softening of material. This is attributed to loading.

(a) Temperature (ºC) (b) Surface Roughness (µm)

Figure 3: Parametric Influence of Grinding Conditions on Temperature and Surface Roughnesswhen Grinding with Diamond 120/140 Wheel


3.2. Multiple Regression Analysis

3.2.1. Introduction

Regression analysis provides an insight into the relationship between dependant and independentfeatures and indicates how to plan the collection of data. There may be a functional relationshipbetween the variables which we can approximate by some simple mathematical function, suchas a linear or polynomial, which contains the appropriate variables. Where no sensible physicalrelationship exists between variables you can relate them by some sort of mathematicalequations. While the equation may be physically meaningless, it may be extremely valuablefor predicting the value of some variables from knowledge of other variables.

The variables can be distinguished two main types in regression analysis, the predictorvariables (Independent Variables) and the response variables (Dependent variables). Thedistinction between the predictor and response variables is not always clear and some timesdepends on our objective. If the true relationship between a response and the predictor variablesis known, then it is easy to understand, control and predict the response. The principle advantageis that it allows us to use more of the information available with us to estimate the dependantvariable. Multiple regression is useful when the intent of the analysis for prediction. A linearcombination of the independent variables that is maximally correlated with dependant variableis sought. Multiple regressions are an extension of correlation. In regression minimizing thesum of squared errors is equivalent to maximizing the correlation between the observed andpredicted values. Regression looks at variability about the mean. Regression breaks thecomparison down into variability due to regression and variability about the regression. Squaredmultiple correlation (R2) which measures the properties of total variance. Due to the discretenature of data recording during the machining operation and inevitable minute changes in theoperational conditions during data measurement and recording, proper data processing cansignificantly improve the prediction ability of the model.

3.2.2. Assumptions Made

• Responses are independent.

• Normality (Multivariate residuals are normally distributed about the mean).

• Linearity (Assumed prediction of linear relationship between predicted dependant variablesand error of prediction).

• Independence of error.

3.2.3. Multiple Regression Predictions

The surface roughness model has been developed by using grinding and process parameters.In this model grinding parameters (speed, feed & depth of cut) and process parameters(normal, tangential forces, acoustic emission signals) are taken as independent parameters.The output parameter is the surface roughness. This model was used to predict surfaceroughness at particular desired points. The numeric values of predicted responses are alsoshown in Table 1. The difference between measured and predicted responses is shown inFigure 5 and Figure 6.


Table 1Measured and Predicted Values from Regression Model

Ra (measured) Ra (predicted) % error

1.424 1.65 -15.87078652

1.837 1.99 -8.328796952

2.212 2.04 7.775768535

2.263 2 11.62174105

1.843 2.05 -11.23168747

2.182 2.13 2.383134739

2.135 2.03 4.918032787

2.484 2.21 11.03059581

1.44 1.44 0

2.42 2.52 -4.132231405

2.369 2.46 -3.841283242

2.124 2.21 -4.048964218

1.623 1.66 -2.279728897

2.036 1.95 4.223968566

1.619 1.57 3.026559605

2.174 2.02 7.083716651

1.632 1.7 -4.166666667

1.552 1.57 -1.159793814

1.86 1.97 -5.913978495

1.492 1.6 -7.238605898

2.24 2 10.71428571

2.32 2.29 1.293103448

2.4 2.32 3.333333333

1.972 2.16 -9.53346856

1.87 1.8 3.743315508

1.692 1.8 -6.382978723

1.849 1.85 -0.054083288

2.083 1.91 8.305328853

1.432 1.85 -29.18994413

1.721 1.49 13.42242882

2.265 2.18 3.752759382

1.954 2.01 -2.86591607

2.384 2.34 1.845637584

1.765 1.68 4.815864023

1.942 2.14 -10.19567456

2.32 2.15 7.327586207


Figure 5: Measured Vs Predicted Surface Roughness

Figure 5: Performance of Regression Model for Surface Roughness Prediction

4. CONCLUSIONS

The performance of the grinding wheel was assessed in terms of grinding force components,acoustic emission signal characteristics, grinding temperatures and surface finish of ground surface.Referring to the illustrations on performance of diamond grinding wheel, it is concluded that:

• Study on grinding of Al/SiCp composites with diamond grinding wheel has shown

that good grinding performance can be achieved in terms of lower order grindingforce components and low AE energy release


• At higher speeds, the grinding forces reduced due to thermal softening of the material

• With higher speed and depth of grinding, diamond can deteriorate leading tographitization and sliding dominant grinding.

• At higher speeds, the surface finish is poor owing to higher temperature which leadsto thermal softening of material. This is attributed to loading.

• Modeling of grinding process can be effectively done using multiple regression analysis.

References

[1] A.Dillio, A.paoletti, (2009), “Characterization and Modeling of the Grinding Process of Metal MatrixComposites”, CIRP Annals-Materials Technology, 58, pp. 291-294.

[2] Y. Zhu, H. A. Kishawy (2005), “Influence of Alumina Particles on the Mechanics of MachiningMetal Matrix Composites”, I.J.of Machine Tolls and Manufacture, 45, pp. 389-398.

[3] A. Manna, B. Bhattacharayya (2003), “A Study on Machinability of Al/SiC-MMC”, Journal ofMaterial Processing Technology, 140, pp. 711-716.

[4] Jae-Seob Kwak, Man-Kyung Ha (2004), “Neural Network Approach for Diagnosis of GrindingOperation by Acoustic Emission and Power Signals”, Journal of Material Processing Technology,147, pp. 65-71.

[5] J. T. Lin, D. Bhattacharyya (2003), “Multiple Regression and Neural Network Analyses in CompositeMachining”, Composite Science and Technology, 63, pp. 539-548.

[6] E. Kilickap, O. cakir, M. Aksoy, A. Inan (2005), “Study of Tool Wear and Surface Roughness inMachining of Homogenized SiCp Reinforced Aluminium Metal Matrix Composite”. Journal ofMaterial processing Technology, 164-165, pp. 862-867.

[7] J. Paulo Davim, C. A. Conceicao Antonio (2001), “Optimization of Cutting Conditions in Machiningof Aluminium Matrix Composites Using a Numerical and Experimental Model”, Journal of MaterialProcessing Technology, 112, pp. 78-82.

[8] A. Jawaid, S. Barnes, and S.R. Ghadimzadeh (1999), “Drilling of Particulate Aluminium SiliconCarbide Metal Matrix Composites”, in Conference Proceedings Machining of Composite MaterialsSymposium, Chicago, USA, 35.

[9] C. Lane (1983), “Machining Discontinously-reinforced Aluminum Composites”, in: Tool andManufacturing Engineers Handbook, Vol. 7, Society of Manufacturing Engineers.

prediction of surface roughness in machining of homogenized sicp reinforced aluminium metal matrix...

Documents