Abstract

Titanium alloys have superior mechanical properties such as high strength to weight ratio, low elastic modulus and high corrosion resistance. The importance of surface finish has increased in specific applications such as aerospace components, biomedical instruments and cryogenic equipment. The lower amount of oxygen, carbon and iron content has made Ti-6Al-4 V Extra Low Interstitial (ELI) perfect for such applications. In this paper, a second order model is developed to correlate the effect of four parameters namely cutting speed, feed, depth of cut and tool nose radius on surface roughness of Ti-6Al-4 V ELI. Three levels were chosen for all four parameters. Three coated cutting tool inserts with three different nose radii of 0.4, 0.8 and 1.2 mm were used. The Box–Behnken method of response surface methodology was adopted to minimize the number of experiments. Data of 27 experiments were used to build the model and predict the surface roughness using the response surface methodology. The model was tested by ANOVA. The predicted value of surface roughness varies only 1.5% with the experimental result, which confirms acceptability of the developed model.

Keywords: titanium alloy, RSM, ANOVA, tool nose radius, surface roughness

#### 1. Introduction

Surface roughness is a key feature for selection of material required to make any component, especially when it is used in applications such as biomedical implants [1], orthopedic implants, cryogenic and marine equipment. Ti-6Al-4 V Extra Low Interstitials (ELI) has a low oxygen, carbon and iron content. It exhibits properties such as a high strength to weight ratio, relatively low elastic modulus and remarkable corrosion resistance, which makes it perfectly suitable for application in aerospace components [2]. With many superior qualities, the material also has disadvantages such as difficult to machine, high chemical reactivity and low thermal conductivity. Titanium alloys are suitable for components having high reliability and hence the surface roughness must be maintained to a desired level [3]. Investigation of cutting parameters on finished products has been undertaken by many researchers. Parameters such as cutting speed, feed, depth of cut [3–6], cutting tool inserts [5] and cutting time [6] are investigated by many researchers. The effect of cutting parameters on

Proceedings of the 4th International Conference on Innovations in Automation...

power consumption and surface roughness while turning EN-31 was investigated by [7]. The effect of cutting speed, feed and depth of cut on cutting temperature while turning EN-36 was also investigated by [8]. A larger nose radius created a finer surface finish [3, 8]. Chou and Song [9] investigated the effect of tool nose radius on finish hard turning and reported that a large nose radius gave a better surface finish but at the same time induces higher tool wear. In the present paper, the effect of four input parameters namely cutting speed, feed, depth of cut and tool nose radius on surface roughness are investigated. Experiments were randomized using the Box– Behnken method, which is an effective method of response surface methodology. Using these experimental data, a second order model was developed to predict the response. The model was tested by the popular method of Analysis of Variance (ANOVA). Confirmation of the experiment was carried out in order to check the acceptability of the developed model.

### 2. Experimental procedure

#### 2.1 Workpiece and cutting tool

The material used for the experiment was Ti-6Al-4 V ELI and this was available in the form of a round bar with a 70 mm diameter and 250 mm length. The properties of the material is shown in Table 1.

The cutting insert used was coated cemented carbide inserts with ISO designation as TNMG 160404, TNMG 160408 and TNMG 160412 with nose radius 0.4, 0.8 and 1.2 mm respectively.

#### 2.2 Machining tests

All experiments were performed in a dry environment using CNC turning center STC-200. The surface roughness was measured at three different positions and then the average was noted. The measurement was done using a Mitutoyo SJ-210 surface roughness tester. Four different cutting parameters were chosen. Cutting speed, feed and depth of cut are process parameters and nose radius is the parameter of the cutting tool geometry. Table 2 indicates cutting parameters and their levels.


Table 1.

Properties of Ti-6Al-4 V ELI.


Table 2. Cutting parameters and levels. Experimental Investigation of Effect of Tool Nose Radius and Cutting Parameters on Surface… DOI: http://dx.doi.org/10.5772/intechopen.81083

#### Figure 1. CNC turning centre STC-200.

In this study, the Box–Behnken method of response surface methodology was adopted. It is a highly suitable method for minimizing the number of experiments required for investigation. As in this case, four different parameters with three levels were chosen, a total of 27 experiments were performed in an order sequence determined by the Box–Behnken method. Experimental set up is shown in Figure 1 and the corresponding cutting parameters with measured responses are listed in Table 3.


Proceedings of the 4th International Conference on Innovations in Automation...


#### Table 3. Design of experiments and measured responses.

#### 3. Result and discussion

A total of 27 experiments were performed and results were used to investigate the effect of parameters on surface roughness. It was found that the range of surface roughness varied from 2.035 to 8.961 μm. The relationship between the cutting parameters and surface roughness was developed using Minitab-17, which can be expressed as

$$\begin{array}{l} \text{Ra} = 20.73 - 0.0482 \text{ cs} - 66.4 \text{ f} - 4.58 \text{ doc} - 12.83 \text{ nr} \\ + 0.000232 \text{ cs}^\* \text{cs} + 320 \text{ f}^\* \text{f} + 2.26 \text{ doc}^\* \text{doc} + 4.37 \text{ nr}^\* \text{nr} \\ - 0.0007 \text{ cs}^\* \text{doc} - 0.0057 \text{ cs}^\* \text{nr} - 0.45 \text{ doc}^\* \text{nr} \end{array}$$

In order to check the adequacy of the model, an Analysis of variance test was performed. It consists of three tests i.e. significance of regression model, significance of coefficients and lack of fit. It showed that the model has a 93.62% of coefficient of correlation (R2 ). Table 4 shows the ANOVA for the current model for Ra, from which it can be said that the nose radius has the maximum effect in


Table 4. ANOVA table for Ra. Experimental Investigation of Effect of Tool Nose Radius and Cutting Parameters on Surface… DOI: http://dx.doi.org/10.5772/intechopen.81083

Figure 2. Main and interaction effects.

reduction of surface roughness followed by depth of cut and feed rate. And among the interactions, the depth of cut and nose radius have greatest effect on Ra.

The effects of the main factors and interaction of the factors on surface roughness are shown in Figure 2. It can be observed from the main effect graph that surface roughness improves with higher cutting speed. This is due to less dissipation of heat from the workpiece at higher cutting speeds and a reduction in cutting force. The increase in surface roughness at a higher feed rate is the result of less available time to carry out the heat from the cutting zone as more material is removed, which causes accumulation of chips between the tool and workpiece. The increase in cutting force at a higher feed rate is also responsible for an increase in surface roughness.

#### 3.1 Optimization plot

The basic aim of this study was to investigate the relationship between the desired surface roughness and optimized cutting parameters. And to achieve it, RSM is a helpful technique. Here the goal is to minimize surface roughness. After the optimization process, the optimum values of the process parameters were obtained as shown in Table 5. The predicted and experimental value of response are shown in Table 5. Figure 3 shows the optimization plots. In order to check the acceptability of the model, a confirmation test was also carried out. It can be clearly observed that the predicted value of surface roughness varies only 1.5% with the experimental result and hence the model seems to be acceptable.

