Sage Journals: Discover world-class research

Abstract

Ultrasonic drilling is a suitable process to enhance the generated surfaces by additive manufacturing. In this study, polylactic acid was selected as the workpiece. The examination parameters were thrust force, delamination, geometrical tolerance, chip adhesion, hole wall morphology and surface roughness. It was explained that the harmonic movement of drill bit in ultrasonic drilling reduced thrust force, delamination, circularity, cylindricality and surface roughness up to 14.5%, 3.7%, 44%, 38%, 20% respectively and removed chip adhesion. Furthermore, number of end-mill flutes was examined and observed that 4-flutes compared to 2-flutes induced reduction in thrust force, delamination, circularity, cylindricality and surface roughness up to 15.2%, 2%, 7.5%, 18.9%, 12.5% respectively. Besides, analysis of variance was established to determine the significant parameters. Finally, non-dominated sorting genetic algorithm-II technique was implemented in order to carry out multi-response optimization.

Keywords

Additive manufacturing ultrasonic poly lactic acid delamination geometrical tolerance surface roughness morphology non-dominated sorting genetic algorithm-ii

Introduction

Additive manufacturing (AM) refers to one of the manufacturing methods in which the final product is fabricated by adding a source of material based on the desired geometry. The AM approach differs from the reducing methods, such as machining in which the material is removed from the original size. Production of more complicated components could be achieved by using AM method.¹ Poly Lactic Acid (PLA) is one of the materials produced by AM process. This material is also called plant-based thermoplastic. One more important thing is that PLA is a biodegradable and recyclable material prepared from renewable raw materials.^2–4 Its applications are as follows: packing, covering, industrial clothing with optimal UV resistance, good resistance to staining, and medical usage such as PLA screws in orthopedic surgery.⁵ Fused deposition modeling (FDM) is a rapidly emerging and low-cost AM method that could be used to manufacture PLA specimens.^6,7

Apart from the AM process, the post-processing activities on AM parts are another important issue. As an example, drilling these parts is a common operation to generate the holes. In latter studies, Ming et al.⁸ investigated chip formation and hole quality in the drilling of additively manufactured Ti6Al4V alloy. In another similar work, Rysava et al.,⁹ reported that the best quality of the hole was attained when the cutting speed and feed value were at the lowest value. Different special drills have been utilized in this regard by various researchers.^10,11 These works were implemented in conventional drilling (CD) while utilizing ultrasonic drilling (UD) method could improve the results to some extent. In UD approach, a high-frequency vibration with a small amplitude is superimposed on the drill bit.¹² Figure 1 shows a schematic of UD process. With respect to the previous studies,^13,14 it can be observed that UD could improve some machinability factors where non-AM materials were used as a case study. Drilling tool can have a coupled rotary-vibratory movement. UD process was applied to different materials by some researchers. Debnath et al. ¹⁵ exerted the ultrasonic vibration in drilling of glass fiber-reinforced epoxy laminates where hole circumferential quality has been improved compared to CD. Feng et al.¹⁶ examined UD process when the workpiece material was carbon reinforced polymers. They concluded lower tool wear in UD compared to CD. Tabatabaeian et al.¹⁷ showed that the presence of delamination could effectively influence the mechanical properties of glass fiber composites during the drilling process. It has been claimed that the delamination phenomenon was more considerable at higher rotary speeds. Wu et al.¹⁸ analyzed the drilling process of carbon fiber–reinforced plastics (CFRP) when ultrasonic vibration was applied to the cutting process. They focused on the delamination factor which is one of the problems occurred in drilling of layered materials by considering the effect of cutting parameters on this factor. As a result, it was indicated that by feed value increase, larger delamination is generated. However, the result of UD process was better than CD one.

Figure 1.

A schematic of vibratory drilling. (U is displacement, a is amplitude, $ω$ is the angular frequency, and t is time).

In recent decades, machining parameters’ optimization is one of worthwhile ways aimed at minimizing machining cost and time and enhancing machining performance.¹⁹ Earlier researchers have used different optimization methods such as desirability function,²⁰ Grey relational analysis,²¹ etc. Nowadays, scholars are focusing on applying artificial intelligence-based methods which include Genetic Algorithm (GA),²² Artificial Neural Networks (ANN),²³ Particle Swarm Optimization (PSO),²⁴ Harmony Search Algorithm (HSA),²⁵ etc. which are expected to conquer some of limitations in traditional methods. In addition, because of the ability to find a set of trade-off solutions in a single optimization run, an evolutionary multi-objective optimization algorithms, especially the Non-dominated Sorting Genetic Algorithm-II (NSGA-II) method, have been indicated to be suitable.²⁶ GA is an appropriate method for regression-based optimization methods. Improved type of GA is called NSGA-II which is used for multi-objective optimization. In NSGA-II, the solution will be organized and according to crowding distance, a new solution is produced to be fitted with all of the objective functions. Detailed description about mechanism of NSGA-II optimization can be found in different references.^27,28

In this regard, various researchers have used this method to predict and find the optimal solutions. Wang et al.²⁹ conducted a multi-objective optimization of CFRP drilling parameters by using ANN and NSGA-II methods. ANN was developed to establish the thrust force and delamination functions. Afterwards, NSGA-II was used to accomplish the optimization. Nandi et al.³⁰ also used NSGA-II to optimize the laser micro drilling of alumina. Daniel et al.³¹ used NSGA-II in order to find the optimized parameters for drilling aluminum metal matrix composites. By NSGA-II, the optimal settings found which simultaneously material removal rate was maximized and process temperature was minimized.

That is obtained from the literature review, using AM segments in different areas are developing. On the other hand, after producing AM parts, that is required to conduct further processes to achieve the desired surface roughness or geometrical tolerances. In addition, sometimes some modifications are required in the AM produced segments. In the mentioned situations, machining processes are conducted. One of the suitable processes is ultrasonic assisted machining. Up to now, no comprehensive investigation about ultrasonic drilling of AM parts has been published. Therefore, that seems to pay attention more deeply in this area and present performance of vibratory drilling and suggest parameters’ set for conducting the drilling operation in order to achieve the best surface quality and geometrical tolerance. The surface quality, delamination and geometrical tolerances have been discussed deeply by SEM (Scanning Electron Microscopy) and 3D topography images. Besides, analysis of variance was conducted in order to examine the effect of process parameters on hole quality and establish mathematical correlations. Eventually, the optimum parameters combination that gives an appropriate multi-objective efficiency in the drilling operation was determined via NSGA-II optimization technique.

Experimental procedures

The experimental examinations were conducted on a three-axis milling machine. An additively manufactured PLA (Poly Lactic Acid) was selected as workpiece material produced by Prusa i3 device. The used PLA filament is “PLA+” grade from eSUN company in gray color. Physical properties of the used PLA material are listed in Table 1 which are according to the manufacturer datasheet. Based on the manufacturer recommendations, the printing conditions were adjusted. Also infill density was set to 100% to prevent void generation which may cause damage while drilling operation. The workpiece dimension is as follows: 100 × 50 × 5 mm. The printing parameters of the PLA workpiece are presented in Table 2. Changing the printing condition can be effective on the results because mechanical properties of the printed material will be changed. This needs other investigation to discuss deeply about the printing parameters’ influence. However, the present investigation examines the drilling condition effects. Owing to the sharp edges producing better surface quality, standard HSS tools with 4-flutes and 2-flutes were selected as drill bits in which their diameter was 6 mm (Figure 2). HSS tools are made of carbon steel with the addition of additional elements such as chrome and vanadium. As a result, they are incredibly durable. Moreover, these tools are inexpensive and suitable for industry. The drilling parameters are listed in Table 3 which were selected based on the manufacturer’s recommendation, workpiece material and other researchers’ experience.^32,33 Three levels were defined for each particular cutting parameters during CD and UD. Each test was conducted for three times and the average values were reported.

Table 1.

Physical properties of PLA.

Parameter	Unit	Value
Density	g/cm³	1.23
Tensile strength	MPa	60
Elongation at break	%	20
Bending strength	MPa	74
Flexural modulus	MPa	1973
Impact strength	KJ/m²	6

Table 2.

The printing parameters of PLA.

Parameter	Unit	Value
Extrusion temperature	°C	215
Bed temperature	°C	60
Infill density	%	100
Number of layers	-	12
Raster angle	°	90

Figure 2.

The used tool: (a) HSS tool with 4 flutes, (b) HSS tool with 2 flutes.

Table 3.

The experimental drilling parameters.

Parameters	Unit	Values
Rotational speed	rpm	565, 955, 1500
Feed value	mm/rev	0.08, 0.15, 0.25
Ultrasonic vibration	-	On, off
Number of end-mill flutes	-	2, 4

Figure 3 shows the experimental setup where UD tool was used (the frequency was 19.5 kHz). Furthermore, other equipment is illustrated in Figure 4. In this study, the used equipment are as follows:

• A Visual measurement machine (VMM) was used to take the images from the machined holes.

• A Kistler 9257B dynamometer was used to measure thrust force during drilling operation.

• A 3 kW ultrasonic generator (MPI Company from Switzerland) was used to generate ultrasonic vibrations.

• A Coordinate measurement machine (CMM) (Sky 8 from Italy) was used to measure the diameter, circularity, and cylindricality of the machined holes.

• A scanning electron microscope (SEM) (MIRA3 FE-SEM manufactured by Tescan company in Czech Republic) was used to investigate morphology of the generated holes’ wall. The resolution of this device is up to 1 nm and its magnification power is up to 1 million times when applying 30 kV.

• An Atomic Force Microscopy (AFM) (ENTEGRA AFMNT – MDT model manufactured by NT-MDT company) device was used to examine 3D topography of the specimens.

• A Mahr roughness tester (Mar Surf PS1) was used to measure the surface roughness.

Figure 3.

The experimental setup.

Figure 4.

(a) The dynamometer, (b) VMM, and (c) related to the force measurement.

Results and discussions

In this work, two types of drilling processes have been implemented: CD and UD. Thrust forces were measured during the process. After machining, the specimens were evaluated by considering the delamination, circularity, cylindricality and the surface roughness of the machined holes. Besides, effects of chip adhesion on quality of the drilled holes were also considered. Furthermore, Morphology of the generated holes were evaluated to better understand the surface quality.

Thrust force

Effects of machining parameters

The average thrust force measurement results are compared in Figure 5. In general, thrust forces were decreased by cutting velocity increase and increased by feed value increase in both CD and UD. It could be due to thinner and thicker chip formation resulting from high cutting velocity and high feed value, respectively.³⁴ Besides, thrust force values obtained in UD are almost lower than CD ones in all cutting conditions. Two reasons caused this outcome to be extracted: engagement time and the positive tool rake angle.

Figure 5.

Thrust force results at different cutting conditions.

In CD, the thrust force is stabilized after running the operation, while it is oscillated harmonically in UD. In other words, in the engagement time, the thrust force goes to the peak value and in the disengagement time, it goes back to the zero value. Note that the peak value in UD is lower than that of CD. Therefore, the average values in UD are lower compared to CD.

Besides, standard deviation (STD) of the thrust force results is shown in Figure 5. The STD is a statistic measuring the dispersion of the dataset relative to its average value and is computed as the square root of the variance. That is observed maximum and average value of STD is 9.04 N and 5.19 N, respectively. Relatively high amount of STD is because of different interlaminar adhesion of the layers in various places of the workpiece which causes difference in the thrust force values even with the same drilling parameters.

The second reason for force reduction might be more positive tool rake angle generated in UD. In CD, the rake angle ( $α$ ) is not constant and it changes along the drill tip length.³⁵ It is attained from equation (1).

\tan α = \frac{r_{x}}{R} \times \frac{\tan ω}{\sin φ}

(1)

In this equation,

r_{x}

, (R),

ω

, and

φ

are the selected radius at each particular point along the lip-length, the radius of drill bit, the helix angle, and the half angle of the drill point. Furthermore, the rake angle has some differences by the cutting rake angle generated during the operation (

α_{w}

). The cutting rake angle is calculated based on equation (2) where

β

is the helical angle. It is generated by feed motion (f) which is also calculated concerning equation (3). In this equation, D is the drill diameter.³⁶

α_{w} = α + β

(2)

\tan β = \frac{f}{π D}

(3)

In accordance with above equations, more positive rake angle is generated in UD. In fact, there are two harmonic steps (downward and upward) in UD. As the vibratory drill bit goes down, the momentary feed value (f) increases, resulting more positive rake angle. It was proved that an increase in tool rake angle causes a decrease in the length of tool-chip contact resulting in lower friction and heat in this zone.³⁶ As a result, the thrust force is reduced. Although in the upward motion the momentary feed value (f) decreases, it could not negatively affect the thrust force when the engagement of tool and chip is insignificant.

Effect of end-mill flutes

The effect of end-mill tool flutes number on thrust force is shown in Figure 6 that demonstrates changing the number of flutes from 2 to 4 causes thrust force decrement. This is related to participating more flutes in cutting process and dividing the thrust force in more tool flutes. The maximum STD value in Figure 6 is 7.7 which is due to inherent structure difference in AM workpieces. The interlaminar adhesion is a little different in the workpiece structure and the required thrust force to create a hole will be different even by the same drilling condition.

Figure 6.

Number of end-mill flutes influence on thrust force results.

Statistical analysis for thrust force

In this section, statistical analysis was carried out for thrust force. Initially, regression correlations were established for thrust force (Table 4). In order to examine the fitness of the calculated correlations, R-squared parameter was used. The resulted R-squared number (98.92%) indicates higher fitting ability of the presented regression correlations.

Table 4.

Regression correlations for thrust force.

Process	Flutes	Regression model
CD	2	Thrust force = 71.91+43.1Feed+0.04693Velocity-17.1FeedFeed-0.000036 VelocityVelocity+0.00231FeedVelocity*
CD	4	Thrust force = 63.47+44.4Feed+0.04761Velocity-17.1FeedFeed-0.000036 VelocityVelocity+0.00231FeedVelocity*
UD	2	Thrust force = 61.90+26.0Feed+0.05096Velocity-17.1FeedFeed-0.000036 VelocityVelocity+0.00231FeedVelocity*
UD	4	Thrust force = 53.88+27.2Feed+0.05164Velocity-17.1FeedFeed-0.000036 VelocityVelocity+0.00231FeedVelocity*

In the following, an Analysis of Variance (ANOVA) was conducted for thrust force (Table 5). ANOVA is an excellent statistical tool for determining if the experimental results are significant or not. From Table 5, that is indicated cutting velocity has the most influence on thrust force results (64.53%). Also, p-Value is less than 0.05 for cutting velocity which shows the parameter significance. The number of flutes, ultrasonic vibration, and feed value are in the next orders. Besides, interaction of input variables should be considered. That is obtained from Table 5, the interaction of cutting velocity and ultrasonic vibration has the most influence (p-Value = .007) over other interactions. Therefore, in order to reduce thrust force in drilling operation of 3D-printed PLA, increasing cutting velocity with adding ultrasonic vibration simultaneously is a good suggestion.

Table 5.

Analysis of variance for thrust force.

Source	DF	Influence (%)	Adj SS	Adj MS	F-value	p-value
Regression	12	98.92	5199.38	433.281	174.79	0.000
Velocity	1	64.53	140.08	140.082	56.51	0.000
Feed	1	3.37	4.48	4.476	1.81	0.192
Flute	1	9.24	45.35	45.353	18.30	0.000
Ultrasonic	1	12.34	44.24	44.241	17.85	0.000
VelocityVelocity*	1	8.77	460.79	460.790	185.89	0.000
FeedFeed*	1	0.00	0.11	0.113	0.05	0.833
VelocityFeed*	1	0.00	0.14	0.137	0.06	0.816
VelocityFlute*	1	0.01	0.61	0.612	0.25	0.624
VelocityUltrasonic*	1	0.41	21.49	21.486	8.67	0.007
FeedFlute*	1	0.00	0.07	0.072	0.03	0.867
FeedUltrasonic*	1	0.24	12.88	12.877	5.19	0.032
FluteUltrasonic*	1	0.01	0.40	0.401	0.16	0.691
Error	23	1.08	57.01	2.479
Total	35	100.00

Delamination

Effects of machining parameters

Different structural deformations can be formed while machining composites such as delamination and micro-cracks.³⁷ Delamination is chiefly an interlayer failure phenomenon caused by external force such as drilling operation.³⁸ Delamination is one of the phenomena that can occur during the post-processing of additively manufactured parts considering these kinds of parts are produced layer by layer.³⁹ With respect to that, the conventional machining of these materials is usually faced with some difficulties.^40,41 Therefore, it is imperative that efficient machining methods such as hybrid ones (e.g., UD) be used to improve the conditions. Figure 7 shows schematic of delamination. Different methods have been proposed to calculate the delamination factor.⁴² In this study, the delamination factor ( $F_{d}$ ) is calculated based on equation (4).⁴³

F_{d} = \frac{D_{2}}{D_{1}}

(4)

Figure 7.

A schematic of delamination.

In this equation, $D_{1}$ and $D_{2}$ are the drilled hole diameter and the maximum hole damage diameter, respectively. $D_{1}$ is obtained by using a CMM device and $D_{2}$ is obtained by using a VMM device.⁴⁴ The VMM results of the delamination are given in Figure 8(a) where they are at different cutting velocities when the feed value is at its high level (0.25 mm/rev). It was clearly seen that delamination was remarkably eliminated when ultrasonic vibration has been added to drilling process, while it was severely happened during CD process. For better understanding, the total results of delamination factor are collected in Figure 8(b).

Figure 8.

(a) The VMM results of delamination during CD and UD at different velocities (rpm) (feed value = 0.25 mm/rev), (b) Delamination results at different cutting conditions.

Figure 8(b) shows that an increment in feed value causes delamination increase. Increasing feed value from 0.08 mm/rev to 0.25 mm/rev resulted in 16% increase in delamination. In general, higher feed value results in more volume of uncut chip thickness causing larger thrust force requirement. Consequently, an increase in thrust forces increases the delamination. Jinyang Xu et al. also have declared the same trend in drilling of carbon fiber reinforced Polyether ether ketone (PEEK) and polyimide.⁴⁵ Conversely, the effect of cutting velocities was not in a specific trend, and in some levels, reduced the delamination factor and in some others vice versa.⁴⁶ That should be noted cutting velocity increment results in drilling temperature increment.^47,48 One more parameter which could be effective on the delamination is the temperature. In fact, higher thrust forces cause tool-chip contact to be increased resulting in higher temperatures. As a result, the softening is happened for PLA where the chip sticks to the drill bit which induces larger delamination in the workpiece.⁴⁹

In comparison, it is ascertained from the Figure 8(b) that the delamination values in UD have lower variations compared to CD. UD process can reduce delamination value up to 4%. It can be explained by more UD process stability resulting from lower thrust forces, temperature and tool-chip contact length.^50,51

Besides, the average and maximum values of STD in Figure 8(b) are 0.022 and 0.049 which can be related to the delamination measuring method. In the used method, selecting the points to calculate the drilled hole diameter (D₁) and the maximum hole damage diameter (D₂) should be conducted by the operator that induces measuring difference. However, to avoid the mentioned measuring error, delamination of each hole was computed three times by different points’ selection.

Effect of end-mill flutes

The effects of the end-mill tool flutes number on delamination is observed in Figure 9. Increasing tool flutes has resulted in delamination decrement about 3% which is not very considerable. This can be due to lower thrust force in 4-flute end-mill described in the last section. Also, that is attributed to the participation of more cutting tool edges in a revolution which helps in delamination decrement. The maximum STD value in Figure 9 is 0.03 which has been discussed at section 3.2.1.

Figure 9.

Number of end-mill flutes influence on delamination results.

Statistical Analysis for Delamination

Regression correlations were also established for determining delamination factor (Table 6). Higher value of R-Squared indicates high strength of the calculated relationship.

Table 6.

Regression correlations for delamination factor.

Process	Flutes	Regression model
CD	2	Delamination = 1.0922-0.676Feed-0.000040 Velocity+3.753FeedFeed+0.000127FeedVelocity
CD	4	Delamination = 1.0720-0.689Feed-0.000040 Velocity+3.753FeedFeed+0.000127FeedVelocity
UD	2	Delamination = 1.0937-0.740Feed-0.000046 Velocity+3.753FeedFeed+0.000127FeedVelocity
UD	4	Delamination = 1.0748-0.753Feed-0.000046 Velocity+3.753FeedFeed+0.000127FeedVelocity

In the following, the calculated regression correlations were analyzed with ANOVA. The results are in Table 7. According to the calculated p-Value in Table 7, feed value is a significant parameter with the most influence on delamination factor (82.93%).

Table 7.

Analysis of variance for delamination factor.

Source	DF	Influence (%)	Adj SS	Adj MS	F-Value	p-Value
Regression	12	95.90	0.090719	0.007560	44.80	0.000
Velocity	1	0.03	0.000103	0.000103	0.61	0.442
Feed	1	82.93	0.001122	0.001122	6.65	0.017
Flute	1	4.52	0.000261	0.000261	1.55	0.226
Ultrasonic	1	1.90	0.000000	0.000000	0.00	0.996
VelocityVelocity*	1	0.05	0.000050	0.000050	0.30	0.591
FeedFeed*	1	5.78	0.005463	0.005463	32.38	0.000
VelocityFeed*	1	0.44	0.000413	0.000413	2.45	0.132
VelocityFlute*	1	0.00	0.000000	0.000000	0.00	0.997
VelocityUltrasonic*	1	0.05	0.000047	0.000047	0.28	0.601
FeedFlute*	1	0.01	0.000008	0.000008	0.05	0.829
FeedUltrasonic*	1	0.19	0.000179	0.000179	1.06	0.314
FluteUltrasonic*	1	0.00	0.000004	0.000004	0.02	0.878
Error	23	4.10	0.003881	0.000169
Total	35	100.00

Geometrical tolerances

Effect of machining parameters

Two other parameters, which were taken into account in drilling process, are circularity and cylindricality. To evaluate them, a CMM device was used. The final evaluation results are listed in two graphs in Figure 10. In both CD and UD, the effect of feed variations was more than cutting velocity. It is seen that an increase in feed value from 0.08 mm/rev to 0.25 mm/rev caused deterioration in both circularity and cylindricality up to 172% and 125%. However, this effect was reduced at high level of cutting velocity for cylindricality. In this condition, thrust force and temperature increases in the cutting zone and the chip tends to stick to the drill bit.⁵² Apart from the feed value, the graphs show that the quality of these parameters were improved by an increase in cutting velocity where the velocity could somehow decrease the negative effect of feed value.

Figure 10.

(a) Circularity and (b) Cylindricality at different cutting conditions.

Furthermore, the average and maximum values of STD of circularity in Figure 10(a) are 0.002 and 0.006 mm, respectively which can be due to the point selection by the probe in CMM device. This leads to different circularity values even with the same drilling parameters. Moreover, the average and maximum values of STD of cylindricality in Figure 10(b) are 0.004 and 0.011 mm, respectively which is related to the point selection by CMM probe from the hole wall. This induces various cylindricality values with the same drilling parameters.

As the UD results are compared to CD, it was revealed that the vibratory motion improved the cutting process. Almost all holes in UD were produced with more precise circularity (up to 38%) and cylindricality (up to 16.5%). It can be explained by lower thrust forces in UD. Moreover, the intermittent movement of drill bit in UD caused the temperature on the cutting tool to be reduced and lower adhesion of chip to drill to be happened.⁵³ To clarify the above explanations, Figure 11 is prepared. It is clearly seen that lower adhesion has been occurred in UD compared to CD. As the PLA is a soft material, each particular variation can significantly affect the geometrical accuracy of the drilled hole. In particular, it is predictable that as long as the chip sticks on the body of the drill bit or sticks as a built-up edge in CD, the drilled holes are not to be generated, precisely. That being the case, the vibro-impacts, which are harmonically generated in UD, trigger less adhesion of chip to the body of the drill bit.⁵⁴

Figure 11.

Chip adhesion on the drill bit.

Effect of End-Mill Flutes

Number of end-mill tool flutes influence on circularity and cylindricality is depicted in Figure 12 That is shown increasing tool flutes results in circularity and cylindricality error reduction up to 7.5% and 19%, respectively which indicates better geometrical accuracy. That is due to lower thrust force and temperature due to less participation of one cutting edge in a revolution. The maximum STD value in Figure 12(a) and (b) are 0.002 mm and 0.005 mm which were discussed at the previous section (section 3.3.1).

Figure 12.

Number of end-mill flutes effect on (a) circularity and (b) cylindricality.

Statistical analysis for circularity and cylindricality

Regression correlations for circularity and cylindricality are presented in Tables 8 and 9. Both the regression correlations have acceptable R-squared values which indicate high trustiness.

Table 8.

Regression correlations for circularity.

Process	Flutes	Regression model
CD	2	Circularity = 0.01229+0.3338Feed-0.000020Velocity-0.432FeedFeed-0.000036FeedVelocity
CD	4	Circularity = 0.01118+0.3308Feed-0.000020Velocity-0.432FeedFeed-0.000036FeedVelocity
UD	2	Circularity = 0.01395+0.2696Feed-0.000019Velocity-0.432FeedFeed-0.000036FeedVelocity
UD	4	Circularity = 0.01297+0.2666Feed-0.000019Velocity-0.432FeedFeed-0.000036FeedVelocity

Table 9.

Regression correlations for cylindricality.

Process	Flutes	Regression model
CD	2	Cylindricality = 0.0758+0.453Feed-0.000102Velocity– 0.1FeedFeed-0.000202FeedVelocity
CD	4	Cylindricality = 0.0733+0.456Feed-0.000101Velocity– 0.1FeedFeed-0.000202FeedVelocity
UD	2	Cylindricality = 0.0548+0.362Feed-0.000081Velocity– 0.1FeedFeed-0.000202FeedVelocity
UD	4	Cylindricality = 0.0519+0.364Feed-0.000080Velocity– 0.1FeedFeed-0.000202FeedVelocity

In the following, ANOVA results are shown in Tables 10 and 11. The most influential variable on circularity and cylindricality is feed value (63.26%) and cutting velocity (40.58%), respectively. In higher feed values, the input part of the drilled hole is destroyed (circularity measuring zone). Therefore, feed value parameter has the most effect on circularity error. However, cylindricality of the hole was measured in three circular sections of the drilled hole (with 12 points). Therefore, based on the ANOVA calculations (Table 11), cutting velocity is more significant compared to feed value.

Table 10.

Analysis of variance for circularity.

Source	DF	Influence (%)	Adj SS	Adj MS	F-Value	p-Value
Regression	12	95.83	0.003791	0.000316	44.00	0.000
Velocity	1	11.15	0.000028	0.000028	3.91	0.060
Feed	1	63.26	0.000290	0.000290	40.40	0.000
Flute	1	0.26	0.000001	0.000001	0.11	0.743
Ultrasonic	1	13.34	0.000001	0.000001	0.13	0.721
VelocityVelocity*	1	0.54	0.000021	0.000021	2.97	0.098
FeedFeed*	1	1.83	0.000073	0.000073	10.11	0.004
VelocityFeed*	1	0.83	0.000033	0.000033	4.56	0.044
VelocityFlute*	1	0.01	0.000000	0.000000	0.04	0.845
VelocityUltrasonic*	1	0.03	0.000001	0.000001	0.15	0.706
FeedFlute*	1	0.01	0.000000	0.000000	0.05	0.818
FeedUltrasonic*	1	4.57	0.000181	0.000181	25.18	0.000
FluteUltrasonic*	1	0.00	0.000000	0.000000	0.01	0.939
Error	23	4.17	0.000165	0.000007
Total	35	100.00

Table 11.

Analysis of variance for cylindricality.

Source	DF	Influence (%)	Adj SS	Adj MS	F-Value	p-Value
Regression	12	91.91	0.017467	0.001456	21.79	0.000
Velocity	1	40.58	0.000688	0.000688	10.30	0.004
Feed	1	27.18	0.000520	0.000520	7.79	0.010
Flute	1	0.14	0.000004	0.000004	0.06	0.811
Ultrasonic	1	10.32	0.000172	0.000172	2.57	0.123
VelocityVelocity*	1	3.14	0.000597	0.000597	8.94	0.007
FeedFeed*	1	0.02	0.000004	0.000004	0.06	0.813
VelocityFeed*	1	5.55	0.001054	0.001054	15.78	0.001
VelocityFlute*	1	0.00	0.000000	0.000000	0.01	0.932
VelocityUltrasonic*	1	3.05	0.000580	0.000580	8.69	0.007
FeedFlute*	1	0.00	0.000000	0.000000	0.00	0.955
FeedUltrasonic*	1	1.93	0.000367	0.000367	5.49	0.028
FluteUltrasonic*	1	0.00	0.000000	0.000000	0.01	0.937
Error	23	8.09	0.001537	0.000067
Total	35	100.00

Surface roughness and morphology

For this purpose, the parts were cut and the surface roughness of the inner walls of the generated holes was acquired by a roughness meter.

Effect of machining parameters

As it is observed in Figure 13, feed value increase from 0.08 mm/rev to 0.25 mm/rev has induced surface roughness increment up to 46%, while surface roughness decrement has been obtained up to 44% by cutting velocity increase from 565 rpm to 1500 rpm in both UD and CD processes.

Figure 13.

Surface roughness results at different cutting conditions in CD and UD.

Besides, the average and maximum STD values for surface roughness in Figure 13 are 0.05 and 0.11 µm, respectively. That can be due to interlaminar structure difference in various places of the AM workpiece.

Figures 14 and 15 show the SEM images of the hole surface in different feed values and cutting velocities, respectively. Based on Figure 14, at higher feed values, that is shown the surface is deteriorated and the resulted surface is rough and many ups and downs are observed. At lower feed values, the ups and down have been reduced and the resulted surface is smoother comparing to higher feed value.

Figure 14.

SEM image of the hole surface in fee value of (a) 0.08 mm/rev, (b) 0.25 mm/rev.

Figure 15.

SEM image of the hole surface in rotational speed of (a) 525 rpm, (b) 1500 rpm.

According to Figure 15, higher rotational speed makes the surface smoother, intensely. Figure 15(b) shows that increasing rotational speed from 525 rpm to 1500 rpm removes the chips on the hole surface and generates the clean one. Higher temperature at higher rotational speeds causes the surface to be soften (Figure 15(b)).

Furthermore, it was clear from Figure 13 that the average values are lower in UD than in CD under all cutting conditions. In total, a 15-20% reduction was observed when the drilling process was regulated using ultrasonic vibration. Based on the thrust force results, it can be said that lower thrust force in UD indicates easier chip formation in UD, leading to better surface quality. On the other hand, less variation in thrust force generation in UD means that the process is more stable than CD. Therefore, higher stability and easier chip shearing in UD results in lower surface roughness comparing to CD. For better understanding, Figure 16 is also shown. Based on Figure 16, the generated feed marks in CD were somehow decreased in UD. The UD results in smoother surface. However, in CD, the machining marks are observed on the surface which causes surface roughness increment. Furthermore, 3D images taken from the generated hole surfaces for CD and UD support this phenomenon (Figure 17). The differences in peak-valley formations on the drilled hole surfaces are also evidence. According to Figure 17(a), the peak-valley difference in CD was obtained as 307 nm which has been reduced in UD to 140.6 nm.

Figure 16.

SEM image of the hole surface in (a) CD and (b) UD.

Figure 17.

3D topography of the drilled hole surface in (a) CD and (b) UD.

In addition to the above reasons, chip type can also be effective on roughness results. Chips from the CD are continuous helical coil³³ and fill the drill grooves where escaping the chips are difficult. This can be aggravated in case of high feed values. In such conditions, chip friction and cutting zone temperature increase. These lead to a decrease in surface quality.³⁴ Therefore, chip breakage that can be seen in UD, can improve the machinability.

Effect of end-mill flutes

Number of End-mill tool flutes influence on surface roughness is depicted in Figure 18. That is shown increasing tool flutes results in surface roughness decrement up to 13%. That can be due to lower thrust force and temperature because of less participation of one cutting edge in a revolution. In addition, the hole surface morphology is demonstrated in Figure 19. That is shown, by using the tool with 4 flutes, voids have been reduced and the surface is more homogeneous. The maximum STD value in Figure 18 is 0.1 µm which was discussed at previous section (section 3.4.1).

Figure 18.

Number of end-mill flutes influence on surface roughness results.

Figure 19.

SEM image of the hole surface in machining by endmill tool with (a) 2 flutes, (b) 4 flutes.

Statistical Analysis for Surface Roughness

Regression correlations were provided to indicate the surface roughness values (Table 12). The correspondent R-Squared value was high enough (99.66%) which indicates a high strength relationship.

Table 12.

Regression correlations for surface roughness.

Process	Flutes	Regression model R-squared = 99.66%
CD	2	Surface roughness = 1.1221 + 2.019 Feed - 0.000931 Velocity - 3.246 FeedFeed - 0.000289 FeedVelocity
CD	4	Surface roughness = 0.9985 + 1.908 Feed - 0.000884 Velocity - 3.246 FeedFeed - 0.000289 FeedVelocity
UD	2	Surface roughness = 0.9298 + 1.901 Feed - 0.000856 Velocity - 3.246 FeedFeed - 0.000289 FeedVelocity
UD	4	Surface roughness = 0.8287 + 1.790 Feed - 0.000809 Velocity - 3.246 FeedFeed - 0.000289 FeedVelocity

Moreover, ANOVA is presented in Table 13. According to the calculated p-Values in Table 13, surface roughness is significantly affected by all the input factors due to lower p-values (<0.05). However, cutting velocity is the most influential factor on the surface roughness (67.29%).

Table 13.

Analysis of variance for surface roughness.

Source	DF	Influence (%)	Adj SS	Adj MS	F-value	p-value
Regression	12	99.66	0.913525	0.076127	567.62	0.000
Feed	1	5.30	0.011799	0.011799	87.98	0.000
Velocity	1	67.29	0.062986	0.062986	469.64	0.000
Flutes	1	6.71	0.009718	0.009718	72.46	0.000
Ultrasonic	1	15.35	0.023517	0.023517	175.35	0.000
Feed*Feed	1	0.45	0.004087	0.004087	30.48	0.000
Velocity*Velocity	1	2.96	0.027105	0.027105	202.10	0.000
Feed*Velocity	1	0.23	0.002150	0.002150	16.03	0.001
Feed*Flutes	1	0.06	0.000540	0.000540	4.03	0.057
Feed*Ultrasonic	1	0.07	0.000610	0.000610	4.54	0.044
Velocity*Flutes	1	0.32	0.002940	0.002940	21.92	0.000
Velocity*Ultrasonic	1	0.80	0.007323	0.007323	54.60	0.000
Flutes*Ultrasonic	1	0.12	0.001138	0.001138	8.49	0.008
Error	23	0.34	0.003085	0.000134
Total	35	100.00

NSGA-II Optimization and Verification

The regression models were established in the previous sections in order to analyze the effects of each parameter and also predict the responses. However, in order to find out the optimum condition, multi-objective optimization is required. In this regard, NSGA-II algorithm was used to obtain the optimal parametric combination. Regression models that calculated for each parameter in last sections, were used as a correlating model for each object. Hence, the objectives of the NSGA-II optimization are minimization of thrust force, delamination, circularity, cylindricality and surface roughness.

Accuracy and time of the NSGA-II optimization procedure depends upon the controlling factors that should be adjusted precisely. The required controlling factors are population size, maximum iterations number, crossover and mutation which were adjusted to 150, 100, 0.7, 0.55, respectively.

NSGA-II was conducted in MATLAB software environment. Since, in this paper, thrust force and delamination are the most significant problems and should be considered, the Pareto-optimal solution set is depicted in Figure 20(a), which Objective1 and Objective2 are thrust force and delamination, respectively.

Figure .20.

(a) Pareto front for thrust force (Objective (1) and delamination (Objective 2). (b) Score histogram (fun1: thrust force, fun2: delamination, fun3: cylindricality, fun4: circularity, fun5: surface roughness).

The 53 solutions in Pareto-optimal solution set were acquired. The first 20 solutions are presented in Table 14. That should be noted due to the heuristic nature of NSGA-II, the results of each solution will be slightly distinct. The resultant values range for objective functions which is called scores’ range in NSGA-II, is demonstrated in Figure 20(b).

Table 14.

NSGA-II optimization results (the first 20 solutions).

No.	Input parameters				Output results
No.	Process	Flute	Feed value	Rotational speed	Thrust force	Delamination	Cylindricality	Circularity	Surface roughness
1	CD	4	0.156	907.731	83.839	1.037	0.022	0.029	0.373
2	UD	4	0.141	734.525	76.1218	1.0226	0.0215	0.0243	0.3920
3	UD	4	0.082	617.197	74.2787	1.0163	0.0216	0.0184	0.4403
4	UD	4	0.097	650.539	74.8727	1.0152	0.0215	0.0202	0.4272
5	UD	4	0.109	670.442	75.2427	1.0157	0.0220	0.0215	0.4219
6	CD	4	0.203	987.454	84.148	1.072	0.022	0.034	0.322
7	UD	4	0.160	766.963	76.5184	1.0310	0.0215	0.0256	0.3763
8	UD	2	0.089	653.254	82.128	1.035	0.022	0.020	0.497
9	CD	4	0.164	923.587	83.911	1.042	0.022	0.030	0.364
10	UD	4	0.156	757.605	76.4373	1.0288	0.0217	0.0254	0.3815
11	UD	4	0.109	674.914	75.2688	1.0156	0.0215	0.0214	0.4180
12	UD	4	0.181	798.201	76.8648	1.0432	0.0215	0.0267	0.3591
13	UD	4	0.103	662.120	75.0583	1.0152	0.0215	0.0208	0.4231
14	UD	4	0.195	817.446	77.0648	1.0534	0.0216	0.0273	0.3474
15	UD	4	0.081	598.790	74.1063	1.0170	0.0231	0.0187	0.4541
16	UD	4	0.121	698.364	75.6189	1.0172	0.0216	0.0226	0.4086
17	UD	4	0.093	639.007	74.6850	1.0154	0.0217	0.0197	0.4325
18	UD	4	0.169	778.696	76.6691	1.0355	0.0216	0.0261	0.3701
19	CD	2	0.185	954.068	84.159	1.057	0.022	0.032	0.346
20	UD	4	0.146	743.425	76.2386	1.0246	0.0216	0.0247	0.3886

In the following, based on the requirements, the suitable combination of parameters was selected as an optimal set (Table 15) and compared with the experimental results. As obtained from optimization results, the optimal parameters’ combination is in UD process with feed value of 0.08 mm/rev, rotational speed of 565 rpm with 4-flutes tool (No. 3 in Table 14). The comparison results are shown in Table 15. That is indicated the largest error amount is related to cylindricality parameter (47.95%) which should be enhanced by improving the regression correlation. Other error values show that the NSGA-II is an appropriate tool to optimize and predict the process responses.

Table 15.

Verification tests results.

Parameter	Predicted value	Experimental result	Error value (%)
Thrust force	74.2787	74	0.37
Delamination	1.0163	1.0213	0.48
Cylindricality	0.0216	0.0415	47.95
Circularity	0.0184	0.0204	9.8
Surface roughness	0.4403	0.56	21.37

Conclusions

In the present study, quality of the generated holes during UD of 3days-printed polylactic acid thermoplastic was evaluated. The examinations were carried out on the thrust force, delamination, circularity, cylindricality, chip adhesion, morphology and surface roughness. The main results can be summarized as follows:

• The thrust force results showed that the obtained values in UD were almost lower than CD ones. Drill-chip disengagement time and more positive tool rake angle have been introduced as two main reasons.

• An increment in feed value caused the delamination factor to be increased. In general, higher feed value results in more volume of uncut chip thickness which induces larger delamination. The effect of cutting velocities has not been in a specific trend, and in some levels, it had a negative effect and some others vice versa.

• It was reported that delamination was remarkably reduced while UD process. This event has been explained by the lower softening phenomenon and lower thrust forces generated during UD compared to CD.

• It was seen that an increase in feed value caused both circularity and cylindricality were worsen. Because higher thrust forces were generated in high level of feed values. In comparison, all holes in UD were produced with more precise circularity and cylindricality. It has been clarified by lower adhesion of chip to drill which improves the surface accuracy.

• Drilling with a 4-flute end-mill tool compared to a 2-flute tool resulted in lower thrust force which is due to participating more tool edges per one revolution in cutting process. Therefore, delamination damage was also reduced. Besides, geometrical accuracy was improved.

• Surface quality improvement was seen while UD process comparing to CD. The morphology and 3D topography of the hole wall helped to prove this fact. That was also obtained from morphology images, increasing cutting velocity has the most influence on the hole wall surface quality.

• Based on ANOVA, the most influential parameter on hole quality were feed value and cutting velocity, respectively. Therefore, feed value decrement and cutting velocity increment result in lower damage in the produced hole.

• NSGA-II optimization based on the established regression correlations was conducted and the optimal machining condition was found out. The NSGA-II controlling factors examined and the appropriate factors were used which include: population size of 150, maximum iterations number of 100, crossover of 0.7 and mutation of 0.55.

• Verification tests were conducted that indicated the largest error value (47.95%) is related to cylindricality. In addition, thrust force, delamination, cylindricality, circularity and surface roughness error values are 0.37%, 0.48, 9.8% and 21.37%, respectively.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Mohammad Baraheni

References

Bordin

Bruschi

Ghiotti

, et al. Analysis of tool wear in cryogenic machining of additive manufactured Ti6Al4V alloy. Wear 2015; 328-329: 89–99.

Bajpai

Debnath

Singh

. Hole making in natural fiber-reinforced polylactic acid laminates: an experimental investigation. J Thermoplast Compos Mater 2017; 30(1): 30–46.

Bajpai

Singh

Madaan

. Development and characterization of PLA-based green composites: a review. J Thermoplast Compos Mater 2014; 27(1): 52–81.

Kaiser

Anuar

Razak

. Ductile–brittle transition temperature of polylactic acid-based biocomposite. J Thermoplast Compos Mater 2013; 26(2): 216–226.

Castro-Aguirre

Iñiguez-Franco

Samsudin

, et al. Poly (lactic acid)—mass production, processing, industrial applications, and end of life. Adv Drug Deliv Rev 2016; 107: 333–366.

Chaudhry

Butt

Mubashar

, et al. Effect of carbon fibre on reinforcement of thermoplastics using FDM and RSM. J Thermoplast Compos Mater 2022; 35(3): 352–374.

Naveed

Butt

Mubashar

, et al. Design and verification of enhanced CFRTPCs fabrication technique using fused deposition modeling. J Thermoplast Compos Mater 2022; 35(11): 1957–1980.

Ming

Dang

, et al. Chip formation and hole quality in dry drilling additive manufactured Ti6Al4V. Mater Manuf Process 2020; 35(1): 43–51.

Rysava

Bruschi

Carmignato

, et al. Micro-drilling and threading of the Ti6Al4 v titanium alloy produced through additive manufacturing. Procedia CIRP 2016; 46: 583–586.

10.

Geier

, et al. Experimental study of drilling behaviors and damage issues for woven GFRP composites using special drills. J Mater Res Technol 2022; 21: 1256–1273.

11.

Lin

Davim

, et al. Wear behavior of special tools in the drilling of CFRP composite laminates. Wear 2021; 476: 203738.

12.

Teimouri

Amini

Bami

. Evaluation of optimized surface properties and residual stress in ultrasonic assisted ball burnishing of AA6061-T6. Measurement 2018; 116: 129–139.

13.

Wang

Zhang

Feng

, et al. Experimental and theoretical investigation on critical cutting force in rotary ultrasonic drilling of brittle materials and composites. Int J Mech Sci 2018; 135: 555–564.

14.

Kumar

Dvivedi

. On machining of hard and brittle materials using rotary tool micro-ultrasonic drilling process. Mater Manuf Process 2019; 34(7): 736–748.

15.

Debnath

Singh

Dvivedi

. Rotary mode ultrasonic drilling of glass fiber-reinforced epoxy laminates. J Compos Mater 2015; 49(8): 949–963.

16.

Feng

Cong

Pei

, et al. Rotary ultrasonic machining of carbon fiber-reinforced polymer: feasibility study. Mach Sci Technol 2012; 16(3): 380–398.

17.

Tabatabaeian

Baraheni

Amini

, et al. Environmental, mechanical and materialistic effects on delamination damage of glass fiber composites: analysis and optimization. J Compos Mater 2019; 53(26-27): 3671–3680.

18.

Gao

, et al. Effects of machining conditions on the hole wall delamination in both conventional and ultrasonic-assisted CFRP drilling. Int J Adv Manuf Technol 2019; 104: 2301–2315.

19.

Rajmohan

Vinayagamoorthy

Mohan

. Review on effect machining parameters on performance of natural fibre–reinforced composites (NFRCs). J Thermoplast Compos Mater 2019; 32(9): 1282–1302.

20.

Baraheni

Tabatabaeian

Amini

, et al. Parametric analysis of delamination in GFRP composite profiles by performing rotary ultrasonic drilling approach: experimental and statistical study. Compos B Eng 2019; 172: 612–620.

21.

Yaşar

Günay

Kılık

, et al. Multiresponse optimization of drillability factors and mechanical properties of chitosan-reinforced polypropylene composite. J Thermoplast Compos Mater 2022; 35(10): 1660–1682.

22.

Sharma

Pahuja

Jha

. Application of Taguchi coupled with genetic algorithm (GA) for optimizing surface quality in drilling of duplex stainless steel 2205. Proc IME E J Process Mech Eng. 2023; 0(0): 095440892311518.

23.

Kharwar

Verma

Singh

. Neural network modeling and combined compromise solution (CoCoSo) method for optimization of drilling performances in polymer nanocomposites. J Thermoplast Compos Mater 2022; 35(10): 1604–1631.

24.

Badiger

Desai

Ramesh

, et al. Cutting forces, surface roughness and tool wear quality assessment using ANN and PSO approach during machining of MDN431 with TiN/AlN-coated cutting tool. Arabian J Sci Eng 2019; 44(9): 7465–7477.

25.

Kumar

Verma

Mondal

. Taguchi-grey theory based harmony search algorithm (GR-HSA) for predictive modeling and multi-objective optimization in drilling of polymer composites. Exp Tech 2021; 45: 531–548.

26.

Deb

Datta

. Hybrid evolutionary multi-objective optimization and analysis of machining operations. Eng Optim 2012; 44(6): 685–706.

27.

Baraheni

Azarhoushang

Daneshi

, et al. Development of an expert system for optimal design of the grinding process. Int J Adv Manuf Technol 2021; 116: 2823–2833.

28.

Yildiz

Ozturk

. Hybrid enhanced genetic algorithm to select optimal machining parameters in turning operation. Proc IME B J Eng Manufact 2006; 220(12): 2041–2053.

29.

Wang

Jia

. Multi-objective optimization of CFRP drilling parameters with a hybrid method integrating the ANN, NSGA-II and fuzzy C-means. Compos Struct 2020; 235: 111803.

30.

Nandi

Kuar

. Parametric optimisation of Nd: YAG laser micro-drilling of alumina using NSGA II. Int J Mach Mach Mater 2015; 17(1): 1–21.

31.

Ajith Arul Daniel

Kumar

VijayAnanth

, et al. Multi-objective optimization of drilling of Al5059-SiC-2% MoS2 composites using NSGA-II. Mater Today Proc 2020; 22: 853–857.

32.

Baraheni

Soudmand

Amini

, et al. Burr constitution analysis in ultrasonic-assisted drilling of CFRP/nano-graphene via experimental and data-driven methodologies. J Reinforc Plast Compos 2024; ■■■: 07316844231225593.

33.

Geier

Shen

, et al. A review on CFRP drilling: fundamental mechanisms, damage issues, and approaches toward high-quality drilling. J Mater Res Technol 2023; 24: 9677–9709.

34.

Paktinat

Amini

. Experiments and finite element simulation of ultrasonic assisted drilling. J Manuf Sci Eng 2018; 140(10): 101002.

35.

Altintas

Ber

. Manufacturing automation: metal cutting mechanics, machine tool vibrations, and CNC design. Appl Mech Rev 2001; 54(5): B84.

36.

Lotfi

Amini

. Effect of longitudinally intermittent movement of cutting tool in drilling of AISI 1045 steel: a three-dimensional numerical simulation. Proc IME C J Mech Eng Sci 2019; 233(12): 4081–4090.

37.

Poór

Geier

Pereszlai

, et al. A critical review of the drilling of CFRP composites: burr formation, characterisation and challenges. Compos B Eng 2021; 223: 109155.

38.

Lotfi

Dao

, et al. Natural fiber–reinforced composites: a review on material, manufacturing, and machinability. J Thermoplast Compos Mater 2021; 34(2): 238–284.

39.

Baraheni

Amini

. Feasibility study of delamination in rotary ultrasonic-assisted drilling of glass fiber reinforced plastics. J Reinforc Plast Compos 2018; 37(1): 3–12.

40.

Marathe

Javali

. Effect of drilling parameters on specific cutting energy and delamination of a composite made of unsaturated polyester resin and chopped glass fibres. J Thermoplast Compos Mater 2016; 29(9): 1261–1281.

41.

Basavarajappa

Venkatesh

Gaitonde

, et al. Experimental investigations on some aspects of machinability in drilling of glass epoxy polymer composites. J Thermoplast Compos Mater 2012; 25(3): 363–387.

42.

, et al. Study of drilling-induced defects for CFRP composites using new criteria. Compos Struct 2018; 201: 1076–1087.

43.

Baraheni

Amini

. Comprehensive optimization of process parameters in rotary ultrasonic drilling of CFRP aimed at minimizing delamination. International Journal of Lightweight Materials and Manufacture 2019; 2(4): 379–387.

44.

Kolesnyk

, et al. A critical review addressing conventional twist drilling mechanisms and quality of CFRP/Ti stacks. Journal of Materials Research and Technology 2023; 24: 6614–6651.

45.

Huang

Davim

, et al. On the machining behavior of carbon fiber reinforced polyimide and PEEK thermoplastic composites. Polym Compos 2020; 41(9): 3649–3663.

46.

Liu

Cong

Pei

, et al. A cutting force model for rotary ultrasonic machining of brittle materials. Int J Mach Tool Manufact 2012; 52(1): 77–84.

47.

Dang

, et al. A study on drilling high-strength CFRP laminates: frictional heat and cutting temperature. Materials 2018; 11(12): 2366.

48.

Mansori

, et al. Study on the frictional heat at tool-work interface when drilling CFRP composites. Procedia Manuf 2018; 26: 415–423.

49.

Amini

Soleimani

Paktinat

, et al. Effect of longitudinal− torsional vibration in ultrasonic-assisted drilling. Mater Manuf Process 2017; 32(6): 616–622.

50.

Chu

N-H

Nguyen

V-D

Ngo

Q-H

. Machinability enhancements of ultrasonic-assisted deep drilling of aluminum alloys. Mach Sci Technol 2020; 24(1): 112–135.

51.

Shard

Agarwal

Garg

, et al. In-situ temperature monitoring during rotary ultrasonic-assisted drilling of fiber-reinforced composites. J Thermoplast Compos Mater 2023; 36(10): 3919–3942.

52.

Amini

Paktinat

Barani

, et al. Vibration drilling of Al2024-T6. Mater Manuf Process 2013; 28(4): 476–480.

53.

Moghaddas

Graff

. Temperature measurement in the ultrasonic-assisted drilling process. Int J Adv Manuf Technol 2019; 103: 187–199.

54.

Lotfi

Amini

. Experimental and numerical study of ultrasonically-assisted drilling. Ultrasonics 2017; 75: 185–193.

Experimental evaluation and optimization of parameters affecting delamination,geometrical tolerance and surface roughness in ultrasonic drilling of 3D-Printed PLA thermoplastic

Abstract

Keywords

Introduction

Experimental procedures

Results and discussions

Thrust force

Effects of machining parameters

Effect of end-mill flutes

Statistical analysis for thrust force

Delamination

Effects of machining parameters

Effect of end-mill flutes

Statistical Analysis for Delamination

Geometrical tolerances

Effect of machining parameters

Effect of End-Mill Flutes

Statistical analysis for circularity and cylindricality

Surface roughness and morphology

Effect of machining parameters

Effect of end-mill flutes

Statistical Analysis for Surface Roughness

NSGA-II Optimization and Verification

Conclusions

Footnotes

Declaration of conflicting interests

Funding

ORCID iD

References