Influence of Cutting Conditions on Tool Life, Tool Wear and Surface Finish in the Face Milling Process

Influence of Cutting Conditions on Tool Life, Tool Wear and Surface Finish in the Face Milling Process

J. Caldeirani Filho

UFU-Universidade Federal de Uberlândia

Faculdade de Engenharia Mecânica

Laboratório de Ensino e Pesquisa em Usinagem

38400-089 Uberlândia, MG. Brazil

A. E. Diniz

UNICAMP

Universidade Estadual de Campinas

Faculdade de Engenharia Mecânica

Departamento de Engenharia de Fabricação

13083-970 Campinas, SP. Brazil

[e-mail]

The main goal of this work is to study the influence of cutting conditions – cutting speed, feed velocity and feed per tooth - on tool life and surface finish of the workpiece in the face milling of flat surfaces. Aiming to achieve this goal, several milling experiments were carried out with different cutting speeds, feed velocities and feeds per tooth. In the first phase of the experiments, cutting speed was varied without varying feed velocity, which caused a variation in feed per tooth. In the second phase of the experiments, cutting speed and feed velocity were varied in such a way that feed per tooth was kept constant. Tool flank wear and surface roughness of the workpiece were measured as cutting time elapsed. The main conclusions of this work are that a) cutting speed has a strong influence on tool life, regardless of whether feed velocity or feed per tooth varies and b) an increase in surface roughness of the workpiece is not closely related to an increase in wear of the primary cutting edge.

Keywords: Face milling, tool wear, tool chipping, tool life

Introduction

The face milling process is frequently used in industrial processes to machine large flat surfaces in a very fast and precise way. As can be seen in the articles cited in this work, as well as in others, much has been done to achieve a better understanding of the phenomena that occur in this process caused by the interrupted cutting that each cutting edge performs. Besides, much effort has been done to establish the relationship between cutting speed and tool wear and tool life. When cutting speed varies, two different conditions can occur. If feed velocity varies proportionately, feed per tooth is kept constant. If feed velocity is not varied simultaneously, then feed per tooth decreases as cutting speed increases. Therefore, it is necessary to verify if the influence of cutting speed on tool wear and tool life is independent of feed velocity. Besides, it is also important to verify how all these parameters influence the surface roughness of the workpiece. The main goal of this work was to answer these questions. Several milling experiments were carried out under different cutting conditions. In the first part of the experiments, cutting speed was varied without varying feed velocity, which caused a variation in feed per tooth. In the second phase of the experiments, cutting speed and feed velocity were varied in such a way that feed per tooth was kept constant. Tool flank wear and surface roughness of the workpiece were measured as cutting time elapsed.

Tool Wear and Tool Life in the Milling Process

In the milling process, the end of tool life is more frequently caused by chipping, cracks and breakage of the edge (rather than regular tool wear) than in other machining processes, such as turning and drilling. This occurs because milling is an interrupted operation, where tool cutting edge enters and exits the workpiece several times per second. In addition, chip thickness varies as the edge penetrates the workpiece. Regular tool wear mechanisms will be predominant only if the tool is tough enough to resist the mechanical and thermal shocks of the process.

Bhatia et al. (1978) and Chandrasekaram (1985) concluded that the major cause of tool failure at high cutting speeds is cracking of a thermal origin. This occurs because the edges are exposed to a high level of thermal shock due to the high temperatures caused by high speeds and high degree of temperature variation typical of the process. At low cutting speeds, cracks of a mechanical origin are mainly responsible for tool failure, as, in this situation, cutting forces are higher and temperatures are lower. Cracks of a mechanical origin may occur due to shocks either at the entrance of the cutting edge (The,1977) or during the exit of the edge from the workpiece (Pekelharing, 1978 and 1984; Van Luttervelt and Willemse, 1984). Problems due to shocks at the entrance of the cutting edge can be worsened by the tendency of the chip to adhere to the tool rake face (Kabaldin, 1980).

Pekelharing (1978) has stated that one of the causes of the excessive chipping of the carbide tools used in milling operations is a phenomenon he called "foot forming". When the tool edge is ready to exit the workpiece, it causes a rotation of the primary shear plane, making its angle negative and instantaneously increasing the force on the edge. On the other hand, Caldeirani (1998) showed that in face milling of steel with tools of carbide indexable inserts, the entrance of the cutting edge into the workpiece is more critical for the chipping of the cutting edge than its exit.

As in other cutting processes, cutting speed is the most influential cutting condition on tool life, followed by feed and then by depth of cut (Ferraresi, 1972). In milling process, due to the fact that the rotation of the tool is usually independent of the speed of the machine table, an alteration of cutting speed (v_c) means a change of feed per tooth (f_z), when the speed of the table or feed velocity (v_f) is kept constant, as shown in the following equation:

where:

z = number of cutting edges (teeth) of the tool

D = mill diameter

n = revolutions per minute of the tool

This work shall attempt to separate the effects of cutting speed, feed per tooth and feed velocity on tool wear and tool life.

An increase in cutting speed in all machining operations increases temperature, which decreases the hardness of the tool material and facilitates the occurrence of phenomena like abrasion and diffusion (Sandvik Coromant, 1994). Besides these effects, in milling operations, an increase in cutting speed increases the frequency of tool edge entrance into the workpiece (increasing the number of shocks per minute) and also the energy of the shock between the cutting edge and the workpiece. This makes cutting speed even more important to the end of tool life.

The distance between two peaks of roughness of a surface obtained by face milling coincides with the feed per tooth. Besides feed marks, the roughness of this kind of surface is also dependent on other factors, such as (Sandvik Coromant, 1989)

• irregular positioning of the inserts in an axial direction

• tool wear – mainly of the secondary cutting edge

• nonuniform tool wear of the cutting edges

• irregular chip flow

• conditions of the machine tool

• stiffness of the workpiece and fixture device

• insert geometry, mainly the size of the secondary cutting edge (parallel land)

Most of the carbide inserts used in face milling present what is called parallel land, i.e., the secondary cutting edge is a straight edge, parallel to the plane of tool rotation, as can be seen in figure 1. The size of this parallel land (b_s) is very important for surface roughness. As the b_s/ f_z ratio increases, more roughness peaks of the surface being cut are smoothened by the cutting edges, decreasing the surface roughness. Sandvik (1989) suggests that this ratio should be between 3 and 10, depending on the position of the inserts in the axial direction. The inserts should be clamped leaving the smaller axial desviation as possible in order to minimize surface roughness and therefore a lower b_s/f_z ratio can be used.

Materials, Equipment and Experimental Procedures

The experiments were carried out in a very rigid CNC milling machine with 22 CV of power in the main motor. The ISO codes of the tool used were SEKR 1204AZ-WM P25 (Sandvik GC-A carbide class) for the inserts, and R260 22-125-157 for the mill. The tool was coated with TiCN (1,5 mm) and TiN (2,0 mm). The tool had eight edges and its diameter was 125 mm. The parallel land of the cutting edges (b_s in figure 1) was 2 mm.

Tool flank wear was measured by an optical microscope. The surface roughness of the workpiece was measured by a portable Mitutoyo instrument.

The workpieces made from AISI 1045 steel were rectangular in shape with a length of 520 mm and a width of a_e = 87.5 mm, such that the ratio mill diameter (D) / workpiece width (a_e) was 1.43 (125 / 87.5), thus within the range recommended by Sandvik (1994). The conducted operation was an asymmetric milling with a distance ( j ) of 4.75 mm between the end of the mill diameter and the beginning of the workpiece, as shown on figure 2.

Table 1 shows the cutting speeds, feeds per tooth and feed velocities used in each experiment.

Each experiments ended when flank wear V_Bmax reached 0.7 mm, which was considered the end of tool life. Several times the experiments were interrupted in order to measure the flank wear of the tool and the surface roughness of the workpiece.

Experiments 1 and 2 were done three times each in order to test the reliability of the results.

Results and Discussion

Figure 3 shows the behavior of flank wear (V_Bmax) of the mill edges versus the machined length (length in the feed direction - L_f) for different cutting speeds when the feed velocity was kept constant. Therefore, the feed per tooth decreased as cutting speed increased. It can be seen in this figure that as cutting speed increases, tool wear increases strongly, even with a decrease in feed per tooth.

Figure 4 shows the behavior of V_Bmax versus the length machined for different cutting speeds when feed velocity was varied in the same proportion in order to keep feed per tooth constant. It can be seen again in this figure that cutting speed strongly influences tool wear and, consequently, tool life. It must be discovered, however, how much of this influence was caused by the increase in cutting speed and how much was caused by the increase in the volume of chip removed due to the increase in feed velocity.

To answer this question figure 5 was made. This figure shows tool life (machined length versus cutting speed), using the tool life criterion of V_Bmax = 0.7 mm when feed velocity is constant and when feed velocity is varied in order to keep feed per tooth constant. It can be seen how strong the influence of cutting speed on the tool life is for both conditions. For the curve where feed per tooth varies, it was expected that the decrease in feed would cause a positive influence on tool life and prevent the influence of cutting speed from being very strong. But this did not occur, on the contrary, the influence of cutting speed in this curve was shown to be even greater than it was for the other curve. Therefore, it can be concluded that the influence of either feed per tooth or feed velocity (which in this case is proportional to the volume of chip removed per minute) is much lower than the influence of cutting speed on tool wear and tool life.

When cutting speed increased without an increase in feed velocity, the frequency of entrance of each cutting edge into the workpiece increased and also increased the speed of friction between tool and workpiece and between chip and tool. On the other hand, the chip cross section and the volume of chip cut for each tooth per revolution decreased. When cutting speed and feed velocity increased proportionally, similarly to the previous condition, the frequency of entrance of each cutting edge into the workpiece and the speed of friction between tool and workpiece and between chip and tool increased, but the volume of chip cut for each tooth per revolution was kept constant. The results in figures 3, 4 and 5 show that either the size of chip cross section or the volume of chip cut for each tooth per revolution is not very important for tool life. What is very important is the number of times each cutting edge enters into the workpiece and the heat generated by the friction between workpiece and tool and between chip and tool, which are caused by the increase in cutting speed. When the cutting edges used in the experiments were observed under a microscope, it could be seen that they were much more chipped than properly worn, which means that the shock between the edge and workpiece at the entrance of cutting was more important than friction in causing the end of tool life.

After this discussion, it is interesting to go back to figures 3 and 4 and verify that tool wear increases slowly (at least for the curves with lower cutting speeds) at the beginning of tool life and increases its rate of growth after a specific machined length. This happens because the tool was coated. While the coating is still intact, the tool wear rate is very low, due to the high wear resistance of the coating. After a specific machined length and wear, the coating is worn off and the workpiece comes into direct contact with the substrate of the tool, which is much less resistant. Therefore, the tool wear rate increases sharply and so the slope of the curves. For the curves with higher cutting speeds the behavior may be the same, but we were unable to detect this because tool life was very short and we were able to measure tool wear at only two moments of tool life.

Figure 6 shows the behavior of the average surface roughness of the workpiece (Ra) versus the machined feed length (L_f) for three different cutting speed values, when feed velocity varied proportionally and, therefore, feed per tooth was kept constant. The main goal of these experiments was to verify whether the influence of feed per tooth on roughness is predominant or whether an increase in the volume of chip removed per minute (caused by the increase in feed velocity) would generate any change in the surface roughness of the workpiece.

It can be seen in this figure that when v_c = 331.4 m/min the initial Ra value was high. This occurred, not because under this condition the largest amount of material was removed per minute (highest v_f), but due to the fact that tool wear was already high when Ra was measured for the first time, under the high cutting speed used (see figure 4 – V_Bmax was around 0.5 mm at this moment).

The other two curves in figure 6 are related to lower cutting speeds and, thus, tool wear was not very high when the first surface roughness measurements were done. These two curves show very close initial values of Ra. This fact shows that neither cutting speed nor volume of material removed per minute (or feed velocity) strongly influences the surface roughness of the workpiece. Therefore, factors like cutting force or power consumed, or vibration of the system, which are influenced by cutting speed and feed velocity, did not influence surface roughness in this case.

As the machined feed length increases, the surface roughness of the workpiece also increases due to tool wear. However, some observations must be made with respect to this topic. The first is that the roughness values obtained were very low. Initially, when the tool had showed almost no wear, either on the primary or on the secondary cutting edge, the Ra values of around 0.4 mm were much lower than the typical values for milling operations (Agostinho et al, 1977). This value is usually obtained in grinding operations. This fact occurred mainly due to the action of the secondary cutting edge (called parallel land in the milling inserts). This is a straight edge, parallel to the plane of tool rotation (Sandvik Coromant, 1994). This edge was 2.0 mm long in the kind of insert used in this work, while the feed per tooth used was 0.12 mm. This means that the same portion of the workpiece touched these tool edges sixteen times in two successive rotations of the mill (the mill used had eight edges) after chip removal. Therefore, each portion of the workpiece was smoothed thoroughly by the parallel lands after the cutting, resulting in the low Ra values obtained. As machined length increased, roughness also increased, but it never exceeded 1.6 mm, which was the value expected for a milling operation, according to Agostinho et al. (1977). This occurred because the primary cutting edge presented much more wear than the secondary edge which, as stated before, is the one responsible for the surface finish of the workpiece.

The second observation is that the experiments with v_c = 240 and 288.6 m/min presented similar Ra values at the beginning of tool life (when the tool presented almost no wear), but they had very different values for surface roughness at the end of tool life, when V_Bmax = 0.7 mm. As can be seen in figure 3, this value of tool wear was reached when L_f = 7100 mm for v_c = 288.6 m/min and L_f = 8600 mm for v_c = 240 m/min. At these moments, the Ra values were 1.45 mm, and 0.9 mm, respectively. This fact reinforces the idea that wear on the primary cutting edge was not responsible for the worsening of the workpiece surface finish. Moreover, it also points out that the amount of flank wear on the primary cutting edge is not related to the wear on the secondary cutting edge, which is responsible for surface roughness. In other words, the surface roughness of the workpiece did not increase in relation to an increase in wear on the primary cutting edge because the wear on the secondary cutting edge did not increase either. Thus, surface roughness is proportional to the cutting speed.

The third observation with respect to this topic is that the higher the cutting speed and feed velocity the greater the slope of curve L_f X Ra and, therefore, the quicker the end of tool life is reached. The next question to be answered is which is more responsible for the slope of this curve: cutting speed or feed velocity.

Figure 7 shows the average surface roughness of the workpiece versus machined length for three different cutting speeds, when feed velocity was kept constant and, therefore, feed per tooth decreased as cutting speed increased.

Again the curve with v_c = 288.6 m/min has a very high initial value of Ra due to the high value of tool wear when the first measurement of surface roughness was done. This happened due to the high cutting speed used. Several occurrences already shown in figure 6 happened again in the experiments related to figure 7. The Ra values were again very low and their increase was not directly connected to the increase in flank wear of the primary cutting edge. The unexpected fact shown in this figure is that the curve related to f_z = 0.12 mm has an initial Ra value very close to that for the curve related to f_z = 0.15 mm. Therefore, the claim in the literature (Sandvik, 1989) was not confirmed, i.e., an increase in f_z would cause an increase in surface roughness. The best explanation for this occurrence is that the ratio between width of the parallel land and feed per tooth is very large (between 13.3 and 16.7, depending on the feed per tooth). Therefore, the workpiece is largely smoothed by the parallel land after chip removal, which removes all the roughness peaks and nullifies all the influence of feed per tooth on the surface roughness of the workpiece. In other words, the fact that a portion of the workpiece receives the action of the cutting edge 16 times (for f_z = 0.12 mm) instead of 13 times (for f_z = 0.15 mm) does not improve surface roughness, since 13 times are already enough to remove all the roughness peaks that can be removed.

At the end of this discussion the following question can be answered:

What are the most suitable cutting conditions for a finishing milling operation, when a mill with a large parallel land is used?

Since under this condition the feed per tooth does not influence the surface roughness up to a certain point, it must be high close to this point. To obtain this feed per tooth without changing the number of cutting edges of the mill, cutting speed must be low and feed velocity high. A low cutting speed will cause a long tool life and a high feed velocity will make the cutting time per workpiece short. Therefore, the number of workpieces machined per tool life will be maximized. A limitation of this procedure is the high values of cutting force and power consumed by cutting, caused by the high volume of material removed per minute (due to the high feed velocity).

Conclusions

Based on the results obtained in this work, the following conclusions can be drawn for the face milling process with conditions similar to those used here:

• Variation in cutting speed has a predominant influence on tool life, regardless of whether there is variation in either feed velocity or feed per tooth;

• The frequency of entrance of cutting edges into the workpiece is the most important factor influencing tool wear and tool life;

• Values obtained for average surface roughness were always below the expected values;

• Wear on the primary cutting edge does not have any relationship with the surface roughness of the workpiece. Surface roughness increased during tool life due to wear on the secondary cutting edge, which, however, was always mild. This fact made the surface roughness of the workpiece remain low, even at the end of tool life;

• When the parallel land of the cutting edge is much larger than the feed per tooth, this one has little influence on the surface roughness of the workpiece.

Manuscript received: June 1999, Technical Editor: Álisson Rocha Machado

• Agostinho, O. L.; Rodrigues, A. C. S. and Lirani, J., 1977, "Tolerâncias, Ajustes, Desvios e Análise de Dimensões", Ed. Edgard Blücher Ltda, São Paulo, Brazil, 215 p. (in Portuguese).
• Bhatia, S. M.; Pandey, P. C. and Shaw, H. S., 1978, "Thermal Cracking of Carbide Tools During Intermittent Cutting, Wear", 51, 201-211.
• Caldeirani Filho, J., 1998, "Study and Monitoring of Face Milling Process with Indexable Insert Mills", Doctoral Thesis, State University of Campinas, Brazil, 147 p. (in Portuguese).
• Chandrasekaran, H.,1985, "Thermal Fatigue on Tool Carbides and its Relevance of Milling Cutters", Annals of the CIRP, vol 34, 125-128.
• Ferraresi, D., 1972, "Metal Cutting", Brazilian Metal Society, São Paulo, Brazil, 71 p. (in Portuguese).
• Kabaldin, Y. G., 1980, "Temperature and Adhesion in Continuous and Interrupted Machining", Machines and Tooling, 51, 33-36.
• Pekelharing, A. J., 1978, "The Exit Failure in Interrupted Cutting", Annals of the CIRP, vol. 27, 5-10.
• Pekelharing, A. J., 1984, "The Exit Failure of Cemented Carbide Face-Milling Cutters. Part I - Fundamentals and Phenomena", Annals of the CIRP, vol. 33, 47-50.
• Sandvik Coromant., 1994, "Modern Metal Cutting", Sandvik Coromant Technical Editorial Dept, Tofters Tryckeri AB, 1º Edition, Sweden, 159 p.
• Sandvik Coromant., 1989, "Milling Handbook 3", Sandvik do Brasil S. A. , Brazil, 76 p. (in Portuguese).
• The, J. H. L., 1977, "High-Speed Films of the Incipient Cutting Process in Machining at Conventional Speeds", International Journal of Engineering for Industry - ASME 99, 263-268.
• Van Luttervelt, C. A. and Willemse, H. G., 1984, "The Exit Failure of Cemented Carbide Face Milling Cutter. Part II - Testing of Commercial Cutters", Annals of the CIRP, vol. 33, 51-54.

Publication Dates

• Publication in this collection
  18 Sept 2002
• Date of issue
  Mar 2002

History

• Received
  June 1999