Correlation Between Acoustic Emission, Thrust and Tool Wear in Drilling

Correlation Between Acoustic Emission, Thrust and Tool Wear in Drilling

Available online at www.sciencedirect.com ScienceDirect Procedia Materials Science 8 (2015) 693 – 701 International Congress of Science and Technolo...

968KB Sizes 0 Downloads 65 Views

Available online at www.sciencedirect.com

ScienceDirect Procedia Materials Science 8 (2015) 693 – 701

International Congress of Science and Technology of Metallurgy and Materials, SAM CONAMET 2013

Correlation Between Acoustic Emission, Thrust and Tool Wear in Drilling Guido Ferrari (a,b,c)*, Martín P. Gómez (b,c) (a)

Instituto Sábato -Universidad Nacional de Gral. San Martín–Comisión Nacional de Energía Atómica-Av. Gral. Paz 1499, San Martín (1650), Buenos Aires, Argentina. (b) Grupo de Ondas Elásticas – Proyecto ICES, Comisión Nacional de Energía Atómica -Av. Gral. Paz 1499, San Martín (1650), Buenos Aires, Argentina (c) Universidad Tecnológica Nacional, Facultad Regional Delta -San Martín 1171, Campana, Buenos Aires Argentina.

Abstract The Acoustic Emission (AE) generated during a drilling process can be used to control the tool wear and therefore the quality of finishing of drilling process at real time, as well as the reduction of time loss due to tool exchange. In this work, steel workpieces were drilled and the AE was evaluated for different states of drill wear. AE signals were correlated with the thrust force. This searching is an attempt to find parameters of AE to assess the degree of tool wear and consequently the quality of finishing of the product. Nine workpieces were drilled with five HSS twist drills containing artificial failures; the thrust and the feed velocity were constant. The specimens were drilled with an industrial milling machine mounted on a mechanical device for measuring the Th. AE and thrust force were recorded simultaneously. © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license © 2014 The Authors. Published by Elsevier Ltd. (http://creativecommons.org/licenses/by-nc-nd/4.0/). Selection and peer-review under responsibility of the scientific committee of SAM - CONAMET 2013. Selection and peer-review under responsibility of the scientific committee of SAM - CONAMET 2013 Keywords: Acoustic emission, drilling, manufacturing process, thrust force, process monitoring.

* Corresponding author. Tel.: +54-11 6772-7990. E-mail address:[email protected]

2211-8128 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Selection and peer-review under responsibility of the scientific committee of SAM - CONAMET 2013 doi:10.1016/j.mspro.2015.04.126

694

1.

Guido Ferrari and Martín P. Gómez / Procedia Materials Science 8 (2015) 693 – 701

Introduction

At present, the performance of machining processes is developed at elevated cutting speed, demanding higher tolerances in geometrical and surface finishing aspects. For that reason, online monitoring systems would be very useful in order to assess the progressive tool wear during machining and to evaluate warnings. For example, in machining process the assessment of the tool wear would be a good indicator to avoid the manufacturing of defective workpieces, reducing costs and rework operations. With this aim, some direct and indirect methods are designed to evaluate the tool wear, Byrne et al. (2001). Direct methods looks for geometrical modifications produced by wear of the cutting tool, making a comparison between new and worn tools. This assessment could be done using optical methods with artificial vision, or direct evaluation of the machine operator among others. The main advantage of this method is the direct measurement of the wear. The disadvantage is the necessity to find a way to make the inspection of the tool surface during the cutting process (difficult of implementation), typically in stops previously defined, extracting the tool and making the wear evaluation process, Dornfeld and Kannatey Asibu (1980), Teti et al. (2010). Indirect methods are based on the measurements of auxiliary parameters like thrust force, torque, AE, vibrations, temperature and other, correlating these parameters with the efficiency of the processes and the level of wear of the drill, Sick (2002). In general, the need to identify particular sets of variable that changes with the tool condition becomes the indirect methods less effective than direct. In this sense, in literature much works have been done for turning operations where the tool is static regarding the turning piece, being much more complex in milling and drilling operations due to the turn of the tool, Byrne et al. (2001). During the machining of metallic materials there are many mechanisms that act as sources of AE such us cut, plastic deformation, friction of the chip with the tool and with the edge of the drilling hole, microcraking, and others, Jantunen (2002), Li (2002). Some sources such as plastic deformation or friction generate continuous AE signals. Other sources as chip and tool breakage produce discrete (or burst type) AE. Then, the AE signals of machining are a summation of those different patterns activated by different mechanisms, Gómez (2012). The present work is an extension of a previous done by Gómez et al. (2010, 2012) in which the correlation between AE, tool wear and torque in a drilling process was determined. In that work, a linear relationship between the Mean Power (MP) of EA and Torque (Tq) was found. Also, both AE parameters provided information about the tool condition and cutting process. In this work, AE measurements were done in similar condition that the previous including a new condition (drill bit with spark-eroded crater). In this case, an analogous analysis was done for the Thrust Force (Th) (that indicates the force made by the drill bit in the direction of the drilling as it makes the hole). In the drilling process this force is complementary to the torque (the torque acts in the plane of rotation of the drill, perpendicular to the thrust force) and both could be used to make a characterization of the tool condition, Totis (2010), Chen and Liao (1996), Lin and Ting (2003).For that purpose a comparison was made between AE signals and others parameters with drills in new state and drills with artificially worn edges simulating normal defects that appear due to use in the industrial process. Nomenclature AE CR SC FE MARSE MC MP N RMS Th Tq

Acoustic Emission Drill bit with external Corner Removed Drill bit with Spark-eroded Crater Drill bit with Flattened Edges Measured Area under Rectified Signal Envelope Drill bit with Mechanical Crater Mean Power Drill bit in New state Root Mean Square Value Thrust Force Torque

Guido Ferrari and Martín P. Gómez / Procedia Materials Science 8 (2015) 693 – 701

695

2. Methodology 2.1. Materials 2.1.1. Specimen Nine drilling tests were performed, one by each drilled workpiece. Each workpiece was extracted from a hexagonal rod of SAE 1040 steel. For each specimen, a pilot hole was made with the aim to eliminate the action of the chisel (center edge of the drill) and make the cutting process more stable (Table 1). Also, a micro structural control was performed to evaluate the homogeneity of the material; the results are shown in Table 2. The analysis showed the homogeneity of the measured parameters (grain size, inclusions, hardness and micro hardness). Table 1. Features of the hexagonal workpiece.

Feature Length Distance between faces Length of pilot hole Steel chemical composition

Results 95 mm±0.05 mm 14.30 ±0.03 mm 1.5 ±0.02 mm of diameter 10 mm of deep. %W C:0,4 Mn:0,72 Si:0,29 P:0.011 S:0,012

Table 2. Microstructural result of the specimen.

Analysis Grain size (ASTM E112)

Results 7-8

Inclusion Inspection

Very few, evenly distributed

Average micro hardness in transversal section

298.6 HV (Perpendicular to extrusion direction)

Average micro hardness in longitudinal section

282,4 HV (Parallel to extrusion direction)

a.

Brinnell hardness average value

58.6HBr

(ball 2,5 mm and 187,5 Kg load)

(Perpendicular to extrusion direction)

b. Fig. 1. (a)Test Specimen; (b) Micrography of structural grain in extrusion direction.

696

Guido Ferrari and Martín P. Gómez / Procedia Materials Science 8 (2015) 693 – 701

ƒ 2.1.2 Drill bits Five HSS twist drill bits with standard geometry (DIN 1412), point angle of 118 degrees and 5 mm of diameter were chosen for drilling tests. Each drill bit represented a different condition of wear. The first, labeled as “new”(N) drill bit, represented edges in regular (or reference) condition. Two drill bits were worn by spark-erosion in their cutting edges representing the condition SC with craters in their cutting edges and FE with flattened cutting edges. Repeatability and accuracy were two reasons to choose spark-erosion as wearing method. Also, two bits were mechanically worn, one of them was cratered in extensive drilling (MC) and the other was modified removing the outer corners (CR). Fig. 2 shows SEM images of the cutting edges in different conditions.

a.

. b.

d.

c.

e. Fig. 2. SEM image: (a)N; (b)MC; (c) SC; (d)FE; (e)CR.

2.2. Equipment description and operational parameters The drilling operation was carried out by a FIRST LC-50 milling machine with automatic feed and selectable speed. The spindle speed was fixed at 499±7 RPM and the feed rate at 0.285 ± 0.005 mm/s. AE parameters were measured with a PAC PCI-2 two channel AE system with 18 bit resolution. A PAC AE-60 broadband sensor with a 40 dB preamplifier was used. Regarding the configuration of the AE acquisition board, the sample rate was 5.106 samples/s, the threshold was fixed at 27 dB and the waveform size was 15360 samples per hit (3ms). The load was measured with a FLEXAR CZA-200, S type load cell, with a maximum capacity of 200 Kg and 3 ± 10% mV/V of sensibility. 2.2.1. Thrust force measurement The thrust force generated during drilling was measured with a load cell mounted in a cylindrical hold, sustaining a sliding plate assembled as is shown in Fig.3. Over the sliding plate was mounted the workpiece (test specimen). The whole system was mounted on the mobile table of the milling machine.

697

Guido Ferrari and Martín P. Gómez / Procedia Materials Science 8 (2015) 693 – 701

Helicoidal Bit

Test Specimen Sliding Plate Load Cell.

Specimen/Load Cell Mounting Device Test. Fig. 3. Thrust measurement device.

2.2.3. AE sensor mounting The AE sensor was mounted on the specimen surface, fixed at 40 mm from the top of the specimen to prevent an overheat caused by drilling, with the aim to preserve the piezoelectric sensor. A couplant gel linked both surfaces to minimize a jump of acoustic impedance.

2.3. Performance of the Test The drilling process was carried out in automatic feed mode. The drill bit descend until reach contact to the surface of the specimen, enter into the pilot hole, and start the cutting process during approximately 20 seconds until the end of the pilot hole. At this point, the action of the chisel was activated during additional 20 second, up to the stop of the test. Nine tests were performed in which different drill bit condition was tested as shown on Table 3. The duration of each test was approximately 40 seconds, and two thousand samples of each AE parameters were recorded in each trial. Table 3.Number of test and drill bit used

Test Drill Bit

1 N1

2 N2

3 MC1

4 MC2

5 SC1

6 SC2

7 FE1

8 FE2

9 CR

2.4. Data Analysis. For this study, the selected AE parameters were: RMS, MARSE Energy, Amplitude and Mean Power (ratio between MARSE and duration), Gómez (2012), simultaneously the thrust force was measured and recorded like an external parameter by the AE board. The study of all parameters was made for the drill stage with pilot hole (with no chisel action) except in the Th graph when both stages of the drilling processes were considered. Given the large number data produced, the analysis implemented was statistical.

698

Guido Ferrari and Martín P. Gómez / Procedia Materials Science 8 (2015) 693 – 701

3. Results and discussion 3.1. Measurement of AE features and Thrust Force. From the data gathered between the MP and de Th (Fig.4), the same lineal relationship evidenced among MP and Torque in previous work can’t be observed, Gómez et al. (2010, 2012). However, it was verified the use of the MP as parameter of EA for tool wear characterization.

Fig. 4. Mean Power and Th vs. drill bit condition.

Typically, AE processes are stochastic due to multiple factor involved in the elastic waves sources. For instance, in drilling process, chip fracture or collision, microcracking and tool breakage produce burst type sources overlapped with continuous emission caused from plastic deformation and friction. For this reason, some average values of AE parameters have a wide dispersion and in other cases the dispersion is smaller, this is related to the definite drill bit wear condition. Fig. 5 a-b-c-d shows the average and the variance of the AE parameters (RMS, Amplitude, MP, Absolute Energy) for different drill bit conditions.

a

b Fig. 5. (a) Amplitude and Variance vs. drill bit condition; (b) MP and Variance vs. drill bit condition.

Guido Ferrari and Martín P. Gómez / Procedia Materials Science 8 (2015) 693 – 701

c

699

d Fig. 5. (c) RMS and Variance vs. drill bit condition; (d) Abs. Energy and Variance vs. drill bit condition.

The dispersion of the data of the AE parameters contains valuable information about the cutting process and could be useful in the moment of the tool characterization. This was previously proposed by Gómez et al. (2010, 2012) and the MP was selected as the best parameter for clustering categorization of the different wear condition for drilling. That behavior was verified in this work including the condition of cratered drill bit by spark-erosion. Three well defined behaviors are shown in Fig. 6, one for new drill bits, other for drill bits with damage in the cutting edges and the latest, for drill bits with the external corners removed. For bits in condition labeled “new”, the friction and the chip breakage are low, but in the cases with edge damage the chip presents different rupture pattern with a high frequency breakage, and in the case with wear in the external corner the main distinguish factor are the friction processes. Fig. 6 represents the variance of the MP vs. average MP. Tests made with drill bit condition “new” are located near the origin of the axis. All the tests made with drills with damaged edges were aligned in a growing potential function and the condition of CR showed a different behavior than other. -4

6x10

Variance MP (AE Counts/ms)

2

SC2 SC1 -4

4x10

MC1 MC2 -4

2x10

FE1 CR1

FE2 N1 0

N2 -2

2x10

-2

4x10

Average MP (AE Counts/ms)

Fig.6. Variance MP vs. Average MP for every drill bit condition.

-2

6x10

700

Guido Ferrari and Martín P. Gómez / Procedia Materials Science 8 (2015) 693 – 701

Graphing other AE parameters, areas with different level of wear can be observed, but based on those AE features is not possible to separate different drill bit conditions, as was mentioned above. As an example, two AE parameters were plotted, Fig. 7 a–b. -8

-23

SC2

SC2

1x10

SC1 2

Variance Abs. Energy (J )

2

Variance Amplitude (V )

6x10

-8

N1

4x10

MC2

FE1

MC1

FE2 N2 CR1 -8

2x10

SC1 -24

MC1

9x10

-24

6x10

MC2

-24

3x10

FE1

N1

0

0,5

1,0

1,5

a

-12

0

Average Amplitude (V)

CR1

FE2

N2

0 0,0

-12

2x10

1x10

Average Abs. Energy (J)

b

Fig. 7. (a) Variance Amplitude vs. Average Amplitude; (b)Variance Abs. Energy vs. Average Abs. Energy.

3.2.Process monitoring by Thrust As seen in literature, the torque could be a good indicator of the drilling processes representing the different stages as the bit passes to the pilot hole into the drilling of the full surface of the specimen. In this case the Th can also be used with greater sensitivity than the torque to differentiate those drilling stages. The following description shows the Th as a function of time as the drill bit thrust forwarded into the material and four stages can be observed: Stage 1: The drill bit approaches to the specimen surface until start the cutting processes (0-5 sec). Stage 2: The cutting process is developed in stable condition due to the pilot hole (5-24 sec). Stage 3: The drill bit reaches the end of the pilot hole and starts the chisel action increasing the Th (24-29 sec). Stage 4: The drilling process is developed cutting and chiseling by absence of pilot hole until the end of the test (29-50 sec). Those stages can be identified in all test in all drilling condition, Fig. 8 shows a comparison between two different conditions. Test 2. Drill Bit in NEW Condition.

50

Stage 3

Stage 3

Stage 4

90

Thrust Force (Kg)

Trust Force (Kg)

Stage 1

80

40

30

Test 3. Drill Bit Mechanical Cratered.

100

Stage 1

20

Stage 2

10

70

Stage 4

60 50 40

Stage 2

30 20 10

0

0 0

10

20

30

40

50

60

0

a

10

20

30

40

Time (s)

Time (s)

b Fig. 8. (a) Th for drill bit in “new” condition; (b) Th for drill bit mechanical cratered.

50

60

Guido Ferrari and Martín P. Gómez / Procedia Materials Science 8 (2015) 693 – 701

701

4. Conclusions With the aim of distinguish different conditions of wear of drill bits during machining, Th and AE features were assessed and correlated. In regard to the Th, a linear correlation with the MP was not obtained, as that showed for Tq in former works. However, the use of the Th as a sensible parameter to corroborate the stage of the cutting process was verified. Also, the use of the AE MP as a good parameter for tool wear characterization was established. The wide dispersion of average values of typical AE features could be used as valuable information at the moment of tool characterization. The states representing different drill bit conditions were grouped in a graph that relates average and variance of the studied parameter. In agreement with previous works of the authors, for the representation of variance of MP vs. average MP three behaviors were distinguished, one for new drill bits, other for damaged cutting edges and the third for condition of removed external corner. These results could be associated with the type of AE source involved in each drill bit condition. Drill condition with damage in cutting edges exposed a potential increase; meanwhile the condition related with higher friction (CR) revealed a separated position, in the variance MP vs. average MP graph. Acknowledgments To Julio Migliori, Edgardo Cabanillas, Adriana Domínguez and Guillermo Arnaldo for technical cooperation. References Byrne, G., D. Dornfeld, I. Inasaki, G. Kettler, W. Koenig and R. Teti,2001. Tool Condition Monitoring (TCM). The Status of Research and Industrial Application. Annals of CIRP 44, 541-567. Chen Y. and Y. Liao, 2003. Study on wear mechanisms in drilling of Inconel 718 superalloy. Journal of Materials Processing Technology 140, 269-273 Dornfeld, D. and E. KannateyAsibu Jr, 1980. Acoustic emission during orthogonal metal cutting.Int. J. Mech. Sci. 22, 285-296. Gómez M.P., 2012. “Estudio de la señales de EA generadas en el proceso de corte de metales”. Aplicaciones a procesos de taladrado.” Tesis de Doctorado, UNSAM . Gómez M.P. A.M. Hey, C.E. D’Attelis and J.E. Ruzzante, 2012. Assessment of cutting tool condition by acoustic emission. Procedia Materials Science1, 321–328. Gómez M.P, A.M. Hey, J.E. Ruzzante and C.E. D´Attellis, 2010. Tool wear evaluation in drilling by acoustic emission. Physics Procedia 3, 819– 825. Jantunen E., 2002. A summary of methods applied to tool condition monitoring indrilling. Int. J. Mach. Tools Manuf., 4, 997-1010. Li X.,2002. A brief review: Acoustic emission for tool wear monitoring during turning.Intl. J. of Mach. Tools and Manufacture, 42, 157-165. Lin S. and C. Ting, 1996. Drill wear monitoring using neural networks. Int. J. Mach. Tools Manufact. 36, 465-475. Teti R, K. Jemielniak, G. O’Donnell and D. Dornfeld, 2010.Advanced monitoring of machining operations. CIRP Annals-Manufacturing Technology 59, 717-739. Totis G., G. Wirtz, M. Sortino, D. Veselovac, E. Kuljanic and F. Klocke,2010. Development of a dynamometer for measuring individual cutting edge forces in face milling, Mechanical Systems and Signal Processing 24, 1844-1857. Sick B., 2002. Review on-line and indirect tool wear monitoring in turning with artificial neural networks: a review of more than a decade of research. Mechanical Systems and Signal Processing 16, 487-546.