Machinability investigation in dry turning of Ti–6Al–4V with a novel TiB 2 –ZrC cermet tool using RSM

To ensure accuracy and improve the processing efficiency of Ti–6Al–4V alloys, dry turning experiment of Ti–6Al–4V was carried out using a novel TiB 2 –ZrC cermet tool. The tool was reinforced by nanoscale VC additive and exhibited excellent hardness and fracture toughness. Response surface methodology (RSM) was used in the experiment to verify and evaluate the cutting performance of TiB 2 –ZrC cermet tool. The cutting forces and surface roughness ( R a ) were selected as the optimization objective. Then the analysis of variance (ANOVA) was employed to ascertain the effective cutting parameters on response factors and demonstrate accuracy of the models. It was found that the effective cutting parameters on surface roughness was feed rate, while cutting depth significantly affected cutting forces. And the confirmation experiments showed that the predicted values coincide with experimental values nearly. Based on the optimized cutting parameters, the tool life and tool wear mechanism were investigated. When the v c , a p and f were 100 mm/min, 0.16 mm, 0.1 mm/rev, respectively, the cutting length and tool life could reach to 3233 m and 29.4 min, respectively, due to the excellent wear resistance and stability of TiB 2 –ZrC cermet tool at high cutting temperature. In this case, the main wear mechanism was adhesive wear and diffusion wear.


Introduction
Ti-6Al-4V alloy has been used in a wide range of various fields for many superior physicochemical properties. Dai et al. (2016) [1] mention that Ti-6Al-4V shows a special attraction in improving the aircraft performance. Then, Ochonogor et al. (2017) [2] have reviewed that Ti-6Al-4V is always utilized to produce aircraft engine blade due to superior strength weight ratio and high-temperature strength. Because of high corrosion resistance and light weight, Kaur and Singh (2019) [3] report that Ti-6Al-4V is suitable for total joint replacement and fracture fixation elements in surgical. However, Arrazola et al. (2009) [4] point out that the strong friction in cutting area induces massive process heat and rapid wear in high-efficiency machining process when machining Ti-6Al-4V alloy. Hayat et al. (2019) [5] summarize the poor machinability of Ti-6Al-4V such as the low thermal conductivity, low elastic modulus, and the high chemical activity. These difficulties in processing limits the application of Ti-6Al-4V. Therefore, complex tool geometry, different cutting conditions and cutting fluids, new tool materials have been studied by many researchers to enhance the machining efficiency of Ti-6Al-4V. Xie et al. (2013) [6] employ a new micro-grooved geometry tool to dry turn titanium alloy and study the influence of micro-groove shape and size on cutting temperature and cutting force. Da Silva et al. (2013) [7] investigate the effects of different coolant pressures on tool life and wear mechanisms when machining Ti-6Al-4V at high speed. However, the performance of the tool material itself is the most important factor for efficient machining of Ti-6Al-4V. Therefore, super hard tool materials like ceramic and PCBN (Polycrystalline Cubic Boron Nitride) are prepared to machining Ti-6Al-4V for interest of their superhigh hardness and stability at high temperature. Andriya N (2012) [8] find that more built-up edges are formed and the coating is peeled off from the rake face resulting from high cutting temperature and strong adhesion wear in cutting zone when using PVD-coated TiAlN tools to dry machine Ti-6Al-4V. Due to the high hardness and strength, and oxidation resistance, TiB2 ceramic is coated on tungsten carbide or other high toughness materials to obtain a high hardness of surface layer and tough substrate. Corduan et al. (2003) [9] demonstrate that there are no problems of adhesion when use TiB2 monolayer to cut Ti-6Al-4V at a relative low cutting speed. Recently, TiB2 ceramic have been directly fabricated to tool material to meet the performance requirements of high-speed cutting of Ti-6Al-4V alloy. However, homogeneous TiB2 cermet tools still have the disadvantage that they cannot have both high toughness and high hardness. Zou et al. (2012) [10] adjust and control the sintering processes to prepare TiB2-TiC+8wt% nano-Ni cermet tool material to enhance the fracture toughness of TiB2 cermet. TiC and melt phase are introduced by Song et al. (2014) [11] to reinforce TiB2-based ceramic tool materials. Using reactive hot pressing process, TiC and SiC are added in TiB2 ceramic with different amounts by Zhao et al. (2016) [12] to fabricate TiB2-based ceramic tool materials. Yuan et al. (2019) [13] design and fabricate a new cermet tool material to enhance the surface hardness and fracture toughness of cutting tools. However, these tool materials are still in the research and development stage of the laboratory without actually verifying the cutting performance of TiB2 ceramic tool to cut Ti-6Al-4V. Tan et al. (2018) [14] manufacture TiB2-20 vol%B4C (TB20) and TiB2-80 vol%B4C (TB80) tools to test in turning Ti-6Al-4V. The results show that the tool life of TB20 is about 1.3 times longer than that of TB80 and tungsten carbide tool, and the cutting temperature increment of TB20 is the smallest over the entire life cycle of the three tools.
However, the tool life of TiB2 ceramic tool is not determined at optimum cutting parameters.
The optimization method of metal cutting process parameters is an important tool to continuously improve machining accuracy and efficiency. Taguchi method and iterative mathematical search technique are often used to modelling the relationships between output and input and determining the optimal cutting parameters. Nalbant et al. (2007) [15] use Taguchi method to obtain optimal surface roughness by optimizing cutting parameters for in turning AISI 1030 steel. Mukherjee and Ray (2006) [16] review that iterative mathematical search technique can find the optimal cutting parameter in metal cutting processes. However, none of these two methods can continuously analyze all levels of the experiment in the process of optimizing experimental conditions. Karim et al. (2011) [17] report that response surface methodology (RSM) is a dynamic method of design of experiment (DOE) for establishing relationships between output responses and input variables to achieve the optimizing and regression modeling in various engineering fields. Neşeli et al. (2011) [18] employ RSM to modelling the relationships between surface roughness and tool geometry and determining the optimal cutting parameters. Gupta et al. (2016) [19] optimize the cutting parameters in turning titanium alloy using CBN insert based on RSM. Hashmi et al. (2016) [20] use RSM to establish and optimize a surface roughness (Ra) model in milling Ti-6Al-4V with carbide insert. And they all find that compared with cutting speed and cutting depth, feed speed significantly affects surface roughness. Aouici et al. (2012) [21] find that the cutting forces are influenced principally by the cutting depth and workpiece hardness by analyzing cutting forces in hard turning with CBN tool using RSM.
However, they only point out the significant factors influencing the cutting force and give the optimal cutting parameter range, but do not further discuss and analyze why and how the significant factors influence the output response.
In this study, a novel TiB2-ZrC cermet tools were designed and fabricated to machine Ti-6Al-4V alloy. And RSM was used to study the significance of cutting speed (vc), cutting depth (ap), and feed rate (f) on cutting forces and surface roughness. And then, cutting parameters were optimized by RSM model to realize a small cutting force and fine surface roughness. Meanwhile, the effects of cutting forces and cutting temperature on tool life and wear mechanism were discussed.

Means and materials
Ti-6Al-4V alloy bar (Baoti group co., LTD), 100 mm diameter, was used in the dry turning test. Tables 1 and 2 show the chemical composition and physical properties of the Ti-6Al-4V, respectively. XRD analysis results in Fig. 1 show that the alloy bar is consisting of α and β phases, which is consistent with standard Ti-6Al-4V alloy [21]. The TiB2-ZrC cermet tool materials (59 mass% TiB2, 25 mass% ZrC, 8 mass% Ni, 4 mass% Mo, 4 mass% VC) were sintered at 1650 °C for 30 min under 30 MPa in vacuum by hot-pressing. And then, TiB2-ZrC cermet materials were processed into standard inserts (12.7 mm × 12.7 mm × 4.76 mm, 0.8 mm nose radius) according to ISO [23]. The mechanical properties, density and grain size of the TiB2-ZrC cermet tools were shown in Table 3.

Experimental design
In this work, the relationships between the input parameters and the output response are given as: where φ is the response function, Y called the response factor is cutting force component or surface roughness. Input parameters vc, f, and ap are cutting speed, cutting depth, and feed rate, respectively. In this work, the quadratic mathematical model of RMS is given by Neşeli et al. [18]: where a0 is the constant term, the coefficients b1, b2, b3 and b11, b22, b33 are the linear and the product terms, respectively, while b12, b13, b23 are the interactive terms. Xi represents input parameters.  Based on Box-Behnken Designs (BBDs), the experimental data needed for the analysis and optimization was collected by encoding each numerical factor as -1, 0, and +1. The assignment of the factor levels was shown in Table 5. A total of 15 experiments were performed, and the experimental results were shown in Table 6.

Results and discussion
All the values of output response in Table 6 show that the feed force (Fa), thrust force (Fr) and tangential force (Fv) are obtained in the range of (22.78~55.48) N, (56.37~171.84) N and (58.07~131.55) N, respectively. The range of surface roughness (Ra) was 0.44 μm to 1.43 μm.

Statistical analysis
To analyze the effects of cutting speed, cutting depth and feed rate on cutting force components and surface roughness, the ANOVAs for Fa, Fr, Fv and Ra are calculated and showed in Tables 7-10, respectively. This analysis was out for a 5% significance level, i.e., for a 95% confidence level. Table 7 shows ANOVA results of feed force (Fa). It can be seen that ap and f, interaction term ap×f, and products ap 2 all significantly affect feed force (Fa). But the effect of cutting depth has the greatest influence on Fa with 28.79% contribution to the model. Cutting speed and its interactions have no statistical significance on Fa.   Concerning now surface roughness, the ANOVA results of Ra is showed in Table 10. The feed rate and product f 2 show a very significant effect on surface roughness, in particular the feed rate. The reason of the highest contribution (36.980%) of feed rate on surface roughness is that the increase of feed rate heightens the helicoid movement between tool and workpiece. The intensification of helicoid movement makes furrows on surface of workpiece deeper and broader.
The second significant factor influence surface roughness is product f 2 with a 16.700% contribution. Other model terms do not show any statistical significance on surface roughness.   F where R 2 is the determination coefficient. These models can provide predicted value of cutting force components and surface roughness before cutting Ti-6Al-4V. The measured and predicted cutting force components and surface roughness are showed in Figs. 3 and 4, respectively.  which indicates that cutting depth (ap) play a greater influence on cutting forces, especially on the thrust force (Fig. 5a). The reason is that cutting thickness does not change but cutting width increases, accompanied by the cutting load on the cutting edge also increases when cutting depth (ap) increased, resulting in a doubling of the deformation force and friction force. When cutting depth (ap) does not change, the three-dimensional surface of feed rate (f) is more precipitous than that of cutting speed (vc), which indicates that feed rate (f) poses a significant effect on the cutting force. Furthermore, with feed rate (f) increasing, although the cutting width does not change, the increase of cutting thickness causes deformation force growth resulting in feed rate (f) play a greater influence on cutting forces when cutting speed (vc) is at a high level (Fig. 5b). When cutting speed (vc) remains at an intermediate level, the cutting forces increase simultaneously with the increase in cutting depth (ap) and feed rate (f), but the three-dimensional surface of cutting depth (ap) is more precipitous than that of feed rate (f), which indicates that cutting depth (ap) shows a significant effect on cutting forces (Fig. 5c) Therefore, through quantitatively analyzing the effect of cutting parameters on cutting forces, it can be found that cutting depth and feed rate have great effect on cutting forces. Andriya N [8] also found that the cutting forces were easily affected by feed rate and cutting depth when turning titanium alloy. The reason is that the increment of cutting depth and feed rate will cause the increase in deformation force of workpiece Ti-6Al-4V and friction force in cutting area, thereby significantly affecting the cutting forces. However, the three-dimensional surface of feed rate (f) is more precipitous than those of cutting speed (vc) and cutting depth (ap) when cutting depth (ap) and cutting speed (vc) are constant, respectively, so feed rate (f) has the greatest influence on surface roughness. The reason is that feed rate can control the width of helicoid furrows under the condition of helicoid movement, combining with the tool shape and relative displacement of tool-workpiece on the surface of workpiece. These furrows become deeper and wider with the increase of feed rate (f). And the feed rate has a more significant effect on surface roughness at low cutting speed than at high cutting speed. It is suggested that lower feed rate (f) and higher cutting speed (vc) should be selected when turning Ti-6Al-4V alloy.

Optimization of cutting parameters
In order to obtain the lowest cutting forces and ideal surface roughness during turning Ti-6Al-4V with TiB2-ZrC cermet tool, the optimal cutting parameters are investigated. The goals and the ranges of cutting parameters are summarized in Table 11. The optimal cutting parameters are cutting speed of 120 m/min, cutting depth of 0. 16  Obviously, the quadratic models are excellently accurate to predict the cutting forces and surface roughness after cutting parameters are given.

Tool life and wear mechanism
The tool life and wear mechanism are studied at the cutting condition of optimal cutting parameters. Fig. 7 shows flank wear of TiB2-ZrC cermet tool at two optimal cutting parameters of cutting forces and surface roughness, respectively. Evidently the wear rate of VB is high at the   When VB reaches about 600 μm, the cutting temperatures of two optimal cutting parameters increase to about 872 °C and 823 °C, respectively. The cutting temperature of optimal cutting parameters for cutting forces is generally higher than that of optimal cutting parameters for surface roughness, especially at the later stage of tool wear. This is an important reason why the tool life of optimal cutting parameters for surface roughness is higher than that of optimal cutting parameters for cutting forces, because higher cutting temperature not only accelerates the softening of Ti-6Al-4V and exacerbates adhesion wear of TiB2-ZrC cermet tool, but also cause TiB2-ZrC cermet tool to diffuse wear to worsens tool life. Wear morphology of TiB2-ZrC cermet tool when optimal cutting parameters for cutting forces are employed is observed by SEM, as shown in Fig. 9. The mechanical damage such as the built-up edge on rake face and the crater on main cutting edge can be observed intuitively. By comparing the composition of elements on flank face, it was found that Al element was detected at point 001, 002, 004 in the wear area compared with point 003 in the non-wear area. Therefore, it could be judged that there was adhesive wear phenomenon on TiB2-ZrC cermet tool. The adherent workpiece materials are most likely to concentrate on small chippings and the boundary between crater and rake face, and the adhesive layer is bonded to the flank face with a very thin layer. Therefore, the main wear mechanism of TiB2-ZrC cermet tool is adhesive wear.  The red line in Fig. 10(a) corresponding to the EDS lines scan plot shown in Fig. 10(b). There is a obvious change in composition of elements from the rake face towards tool substrate on the flank face. The few element Al diffuse into the wear surface, and the detection of element O shows that the tool surface slightly reacts with oxygen in the air. Therefore, the main wear mechanism of TiB2-ZrC cermet tool can be summarized as adhesive wear, diffusion wear and slight oxidation wear in turning Ti-6Al-4V.

Conclusion
(1) The cutting depth and feed rate significantly influence the cutting forces in turning Ti-6Al-4V with TiB2-ZrC cermet tool. The effects of interaction terms ap×f, vc×ap, vc×f, and product ap 2 on the cutting forces are also significant. Additionally, the surface roughness is deeply affected by the feed rate and product f 2 .