Experimental and Statistical Analysis on Machinability of Nimonic80a Superalloy with PVD Coated Carb

2.1. Experimental Study

In experimental work, Nimonic 80A superalloy was used as workpiece material, as a result of the literature survey, on which no machining process was performed before. Chemical composition of workpiece material is given in Table 1. The workpiece dimensions are Ø40x250 mm and the hardness of the material used in the tests is 290 HV. In the experiments, CNGN 120408-MS coated carbide cutting tools, by Kyocera, were used as inserts having tool nose radius (rε) of 0.8 mm. These inserts with MS (γ; chip breaker angle 19°) chip breaker geometry (Fig. 1a) are rigidly connected to the tool holder in the form of PCLNR 2525M12 with an approaching angle (χr) of 95° on the CNC lathe. Table 1. Chemical composition of Nimonic 80A. C 0.052

Figure 1. Ch ip breaker geometry on the inserts The levels of cutting parameters were determined based on the cutting tool manufacturer's recommendation and literature reviews. Three different levels of cutting depth (a), cutting velocity (V) and feed rate (f) were selected and given in Table 2. The tests were carried out on a TC 35 Johnford CNC lathe with Fanuc control unit according to the parameters. Table 2. The parameters and levels Turning Parameters a (mm) V (m/min) f (mm/rev)

Firstly, measuring of cutting forces were provided by a 9257B type dynamometer. The cutting force data perceived by the dynamometer was taken to computer using Kistler Type 5019B130 Multichannel Charge Amplifier using Type 2855A3A / D Board CIO-DAS 1602/12 data acquisition card and Kistler Type 2825A1-2 Dynoware software. The cutting forces components (Fig.1b) were obtained by dynamometer and equipment mentioned above. As a second part of the study, surface quality was investigated by measuring surface roughness values (Ra) via Mahr Perthometer M1 device. The Ra values was evaluated by taking average of three values measured from each 120° of cylindrical part. Finally, regression models were performed for the cutting force components and the surface roughness. The stages of the experimental setup and the regression analysis are schematically shown in Fig. 2.

2.2. Regression Analysis

Regression analysis is used to determine the relationship between two or more variables that have causal relation between them and to make predictions or estimations about that topic using this relation. In this analysis technique, a mathematical model is used to describe the relationship between two (simple regression) or more variables (multiple regression), and this model is called the regression model. In this study, the multiple regression is performed due to having three independent variables (cutting velocity, feed rate and cutting depth) and four dependent variables (cutting force components and surface roughness) [2]. The cutting parameters directly affecting the cutting forces and surface roughness have a great importance in terms of the validity and reliability of the models to be developed for force components (Fc, Ff and Fr) and surface roughness (Ra). For this reason, cutting parameters, namely cutting velocity, feed rate and cutting depth are taken into consideration in the developed model of cutting forces and surface roughness separately. Therefore, it is aimed to develop the models of cutting forces and roughness which is easy to apply to real cutting conditions that is economical, understandable, high reliable by using basic cutting parameters.

3.1. Main Cutting Force (Fc)

The main cutting force (Fc) values which are important at the primary level in terms of energy consumption in turning are firstly taken into account in the analysis of the cutting forces. The Fc values have about 50% increase with increasing feed rate from 0.1 to 0.2 mm/rev while this increase rate is about 30% by the feed rate from 0.2 to 0.3 mm/rev as shown in Fig. 3. This variation in the main cutting force is similar for each cutting depth. The Fc values have about %65 increase with increasing cutting depth from 0.5 to 1 mm while %35 increase by cutting depth from 1 to 1.5 mm (Fig. 3). This increasing trend in the Fc is also obtained for each feed rate. On

the other hand, there is decreasing trend by increasing cutting velocity as expected but the decreasing ratio is inconsiderable as also seen from the result of variance analysis. The minimum Fc value is measured as 242 N in cutting depth of 0.5 mm, feed rate of 0.1 mm/rev and cutting velocity of 45 m/min.

Figure 3. The distribution of Fc values The experimental results are also evaluated with analysis of variance (ANOVA) with 95% confidence level to determine the effects of parameters on machining outputs, namely Fc, Fr, Ff and Ra. According to ANOVA results, P values must be less than 0.05 to understand that the parameter is effective on the machining outputs. Table 3 shows that the cutting depth is the most important factor on Fc with 47.8% PCR (Percentage contribution ratio). It can be said that the feed rate is the secondary important parameter on cutting forces (45.88% PCR), nevertheless the cutting velocity has inconsiderable effects on the Fc. Table 3. ANOVA results for Fc. Source V f a Error Total

3.2. Radial Force (Fr)

The radial force (Fr) values which are generally the lowest one of the cutting force components in conventional cylindrical turning operation. The Fr values is lower than the Ff values for the cutting depth of 1 and 1.5 mm as expected, but higher for the cutting depth of 0.5 mm. Since the tool nose radius (0.8 mm) is higher than cutting depth, the radial force is higher due to chip formation occurs depending on ploughing effect. It is valid only for the cutting depth of 0.5 mm that the Fr values have about 25% increase with increasing feed rate both from 0.1 to 0.2 mm/rev and from 0.2 to 0.3 mm/rev. This increase ratio is 50% by increasing the feed rate in the range 0.1-0.3 mm/rev for the cutting depth of 1 and 1.5 mm as shown in Fig. 4. The Fr values have a little change by increasing the cutting depth (Fig. 4) as indicated Ref [21-22]. The tool-chip contact area and so the force distribution is smaller than that of other forces depending on approaching angle of cutting tool. Therefore, the Fr component has little increment by increasing cutting depth due to less change of the tool-chip contact area on Fr direction. Moreover, the Fr values are generally decrease by increasing cutting velocity as seen in Fig.4

[21]. The minimum Fr value is obtained as 142.3 N in cutting depth of 0.5 mm, feed rate of 0.1 mm/rev and cutting velocity of 75 m/min.

Figure 4. The distribution of Fr values According to ANOVA results, Table 4 shows that all cutting parameters is important for Fr, however the feed rate is the most important factor on the Fr with 87.63% PCR. Moreover, the cutting velocity is the secondary important parameter with 5.14% PCR while the cutting depth is the less important parameter with 4.47% PCR. Table 4. ANOVA results for Fr. Source V f a Error Total

3.3. Feed Force (Ff)

Feed force which is parallel to the feed direction, can generally be up to about 55% of the force Fc. The relationship between the feed force (Ff) and the radial force (Fr) was explained in detail the part of the radial force (Fr). It is valid for all the cutting depth that the Ff values have about 18% increase with increasing feed rate both from 0.1 to 0.2 mm/rev and 24% increase by the feed rate from 0.2 to 0.3 mm/rev as seen from Fig. 5. The Ff values have 90% increase by increasing the cutting depth from 0.5 to 1 mm while 35% increase from 1 to 1.5 mm. On the contrary, the cutting velocity has reducing effect on the Ff values by 13% and %10 when it increases from 45 to 60 m/min and from 60 to 75 m/min, respectively (Fig. 5). The minimum Ff value is obtained as 123.7 N in cutting depth of 0.5 mm, feed rate of 0.1 mm/rev and cutting velocity of 75 m/min.

Figure 5. The distribution of Ff values The most significant parameter on Ff is the cutting depth with 80.17% PCR (Table 5). It can be seen that the feed rate is the secondary important parameter with 12.52% PCR while the cutting velocity has 3.49% PCR. Table 5. ANOVA results for Ff. Source V f a Error Total

As a whole, the increase in cutting forces with increasing feed rate (f) and cutting depth (a) is expected situation in machining processes [23,24]. As a reason of this result, it is possible to show the increase in the tool-to-chip contact area and so the energy consumption increases in chip formation by the increase in f and a value. The cutting forces display regular and expected tendency, especially for the cutting depth of 1 and 1.5 mm. However, irregular tendency in cutting depth of 0.5 mm (smaller than tool nose radius of 0.8 mm) can be referred to ploughing effect. The ploughing process are shown by the effect of minimum cutting depth, especially in Fr and Ff cutting force components. The increase in ploughing force is resulted from uncut chip thickness or cutting depth which is less than a critical value (a≥rε) [25-28], that is micromachining process. Thus, unexpected tendencies for Fr and Ff are observed only in the minimum cutting depth.

3.4. Surface Roughness (Ra)

The improved surface quality increases fatigue strength, wear resistance and decreases the corrosion behavior of the finished parts. The surface roughness depends on various factors like material hardness, tool material and geometry, coating material, heat transfer and cutting parameters. Hence the optimized surface roughness is important, especially for good surface quality [29-30] owing to requirement of high strength in assembly operations in defense and aerospace industry where nickel-based superalloys are frequently used in recent times. The surface roughness values increase with increasing feed rate as expected for all cutting velocity and cutting depth. The Ra values have about 160% increase with increasing feed rate from 0.1 to 0.2 mm/rev while this increase rate is about 135% by the feed rate from 0.2 to 0.3 mm/rev for all cutting velocity and cutting depth values as shown in Fig. 6.

Figure 6. The distribution of Ra values The cutting velocity has additive effect on Ra values for the cutting depth of 0.5 mm due to ploughing effect mentioned above. The surface roughness increases considerably when the cutting velocity is 75 m/min in the minimum cutting depth. On the other hand, in higher cutting depths surface quality improves with increasing cutting velocity. This shows that the nickel-based superalloy should be machined in medium cutting velocity (60 m/min) and low feed rate in order to obtain high surface quality. The Fig. 6 also demonstrates that the cutting depth has very low effect on the surface roughness once tool nose radius is bigger than cutting depth. The minimum Ra value is obtained as 0.6 µm in cutting depth of 1 mm, feed rate of 0.1 mm/rev and cutting velocity of 60 m/min. Table 6 shows that the feed rate is the most and only significant parameter on Ra with 97.39% PCR. Table 6. ANOVA results for Ra. Source V f a Error Total

3.5. Regression Model

The multiple regression model is developed to predict the cutting forces and surface roughness without needs for the experiments in these cutting conditions. The models are the second-degree regression models for both cutting forces (Fc, Fr and Ff) and surface roughness shown in Eqn. 2-5, respectively. 𝐹𝑐 = (7.33 − 0.0301 ∗ 𝑉 + 45.64 ∗ 𝑓 + 9.377 ∗ 𝑎)2 𝐹𝑟

The coefficients of determination (R ) show the relation between independent (V, f, a) and dependent variables (Fc, Fr, Ff, Ra) and are found as 0.9778, 0.9702, 0.9638, 0.9707 for main cutting force, radial force, feed force and surface roughness, respectively (Fig. 7-10). Furthermore, the deviation between experimental and modeling results for Fc, Fr, Ff and Ra are calculated as 4.53%, 3.21%, 7% and 9.12%, respectively. The coefficients of determination and

deviation values prove that these models are accurate and can be reliably used in related industries working on the material.

Figure 7. The comparison of experimental and modeling results for Fc

Figure 8. The comparison of experimental and modeling results for Fr

Figure 9. The comparison of experimental and modeling results for Ff

Figure 10. The comparison of experimental and modeling results for Ra