Experimental and Finite Element Study of Laser Cutting Induced ...

Proceedings of International Conference on Computational Engineering & Science, Edited by S. N. Atluri, Reno, Nevada, July 31-August 2, 2002

Experimental and Finite Element Study of Laser Cutting Induced Distortion in A Production Environment Y. P. Yang, et al

Summary Laser cutting induced distortion was systematically investigated with experimental methods and finite element analyses for a Caterpillar front linkage structure. The final distortions in plates after cutting were measured with CMM system and the distortion histories were recorded with laser sensors and data acquisition system. To develop laser cuttinginduced distortion prediction methodologies, the cutting process was simulated with finite element methods. Both experimental and simulation results consistently showed that two dominant distortion modes are present, even under laser cutting conditions: (a) in-plane movement due to release of the internal stresses resulted from steel processing; (b) to a lesser degree, out-of-plane distortions. Both modes of distortions can be mitigated by appropriate considerations of cutting sequence

Introduction Recently, laser beam cutting (LBC) has received an increasing attention for a wide variety of applications in industries due to its excellent cut quality with high productivity and flexibility. Compared with oxyfuel gas cutting and plasma cutting, laser beam cutting applies low heat input to a workpiece. However, experiences have showed that distortion may still occur. Little research has been devoted to understand distortion mechanisms associated with thermal cutting processes [1]. Most researches have mainly focused on modeling heat flow and fluid flow and improving cutting quality by optimizing cutting shield gas and parameters in laser cutting process [2-5]. It is important to understand the distortion mechanism because the impact of the distortions on various down-stream manufacturing processes, such as welding fabrication, can be enormous. In this paper, laser cutting-induced distortions were investigated by detailed experiments and finite element analyses in a real production environment.

Cutting Distortion Measurement A CO2 Tanaka laser-cutting machine was used to perform the cutting with 6kW power and 51inch/min travel speed on a large cutting table as shown in Fig. 1. Laser sensors were mounted on Laser torch critical locations to monitor the in-plane and out-of-plane cutting distortions as a function time. During cutting, real time cutting distortions were recorded and Laser Sensor shown on the screen of data acquisition system. Two parts from a boomFig. 1 Laser Cutting Setup structure (No. 393 and 372) were cut from a 4' by 10' half-inch plate. The total cut length is about 9 meters. The cutting sequences are shown in Fig. 2. The plate surface was initially etched with 4'' by 4'' grids by low power laser for initial and final distortion


measurement. Before cutting, the initial distortion was measured at the intersection points of the grids by CMM system. After cutting, the final distortions were measured again at the intersection points. By subtracting the initial distortion from the final distortions, the cuttinginduced distortion could be obtained. Real time measurement: Nine laser sensors (LS) were amounted on critical locations shown in Fig. 2 to monitor the in-plane and out-of-plane cutting distortion. Fig. 3a shows the in-plane movement during cutting. The distortions at LS35 (about 1 mm) represent the typical Reference point A1

Square: 4’’ by 4’’

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 A B

LS34

In-Plane Out-of-Plane Cutting Sequence

Y

C D E F G H I J K L

4’

X

LS28

LS30

LS32

Part 372

Part 393 LS33

LS29

LS35

LS37 1

LS36

LAS ER 31

10’ 1

Start/Stop

LAS ER 28

LAS ER 34

0.5

LAS ER 29

0.5

Fig. LAS 2 ER Cutting Sequences and Laser Sensor Position on aLASEtched Plate 35 ER 30 LAS ER 36

0 -0.5 -1

-1.5

Dis place me nt (mm)

Displace me nt (mm)

LS31

LAS ER 32

0

LAS ER 33 -0.5 -1 -1.5

-2

-2 0

200 4 00

60 0

80 0 1 00 0 1 200 14 00 16 00 180 0 2 00 0 2 200 24 00

Time (s e c)

0

200 4 00

60 0

80 0 1 00 0 1 200 14 00 16 00 180 0 2 00 0 2 200 24 00

Time (s e c)

(b) Out-of-Plane Distortion (a) In-plane Distortion Fig. 3 Cutting Distortion History

local movement of the plate skeleton. The displacements at LS31 and LS36 represent the plate movement during cutting which is less than 1mm. Fig. 3b shows the out-of-plane displacement of cut part during cutting. The displacements at LS30 are almost zero and the displacements near both ends (LS32, 33 and 28) are about 1 to 1.25mm, which shows that the cut part bows down. Note that the distortions at LS37 will be plotted later to compare with the FE prediction. The final cutting-induced distortion will be measured on the cut part.


Final cutting distortion: Real-time measurements provide the distortion evolution process at several points. To establish a distortion pattern on the entire part after cutting, we used CMM system to measure the final distortions on the itched lines of the cut part. Fig. 4 shows the plate initial distortions, which indicates that the plate was already deformed before cutting. The initial distortion at Line K was plotted in Fig. 5 as marked with "Before". After cutting, the distortion at Line K was measured again using the same reference plane and plotted in Fig. 5 as marked with "After". By subtracting "before" from "After", the actual G H I J K L

U(mm) -1.1 -0.9 -0.8 -0.6 -0.4 -0.3 -0.1 0.1 0.2 0.4 0.5 0.7 0.9 1.0 1.2

Part 393

Start/Stop

0 500

5

U (mm)

1000 0

1500

-5 0

2000

X ( 500 mm )

Y(

) mm

2500 1000 3000

Distortion (mm)

1.4 1 0.6

Before

0.2 -0.2 -0.6

Distortion

-1

After

-1.4 0

Fig. 4 Plate Initial Distortion

200

400

600

800

1000

1200

1400

1600

Distance (mm) Fig. 5 Out-of -Plane Distortion along Line K on Part 393

Laser torch

Cut line Laser sensor Part shifted Direction Cut line Cut line Laser sensor

Laser sensor

C-clamp

Fig. 6 Cut Part In-Plane Movement

cutting-induced distortion was obtained as marked with "distortion" in Fig. 5. The magnitude differences in the "Distortion" line are the relative distortion due to cutting that is about 1.3 mm. Using the same method, we obtained the distortion at all etched lines.


Cut part movement Fig. 6 shows that the cut part is shifted to one side and a large gap is formed along the final cutting line. If cutting continues in the cut part, the dimensional error will become accumulative due to the cut part movement. Therefore, it is necessary to reposition the laser torch based on the new reference during cutting to keep the laser cutting accuracy.

Laser Cutting Simulation The same laser cutting process was simulated with finite element method (FEM) in this investigation. Both cutting and welding can be considered a thermo-elastic-plastic process, but the ways in which the material is filled or removed are exactly opposite. Welding is used to join two separate parts together by filling weld metal into a groove; cutting is used for separating a plate by removing the melted metal along the cutting line. During cutting, the metal along the cutting line is heated to the melting point, and then blown away by the pressure of the gas flow, and the cutting kerf is formed. Although most heat introduced into the plate goes away with the dropped molten slag, some of the heat dissipates into plate through conduction. The rapid heating and cooling process generate both localized residual stresses and distortions. 4’ x 10’ Plate Simulated Contact Points

Cutting Table (Modeled as Rigid Surface)

z y

x Fig. 7 Finite Element Model -- Part 393

The finite element method used to simulate the cutting process was similar to that used for the simulation of the welding process. The transient temperature fields were obtained through thermal solution performed by CTSP (a comprehensive thermal solution package) developed by Battelle and Caterpillar [6]. The stress and deformation were calculated by ABAQUS through thermal-elastic-plastic analysis based on the computed temperature fields. The forming of the cutting kerf was simulated by element removing techniques. In both thermal and structural solution, the elements located in the cutting line were removed from the mesh as soon as the heat source passed. The contact between the cut plate and the cutting table could be simulated with a node-based rigid surface. A 3D shell finite element model shown in Fig. 7 was used to simulate the laser cutting of front linkage-structure parts. In the model the contacts between the plate and the cutting table


were simulated by a node-based rigid surface. The contact locations were represented with the many triangular symbols. Fig. 8 shows transient temperature distributions during cutting. The elements behind the heat source were removed as shown in white color. Fig. 9 shows the comparisons of real time in-plane distortion and cut part movement between experiment and FEM. The plate movement during cutting was predicted by FEM and a good agreement was achieved in comparing distortion history. Fig. 10 shows the typical skeleton movement during thermal cutting. This movement caused by the prior stress released which inherited from plate mill process.

Fig. 8 Temperature Distribution during Cutting

(b) Predicted

Laser Sensor

0.0

(a) Experiment

LS37

Distortion, mm

-0.2

-0.4

-0.6

Predicted Measurement -0.8

-1.0 0

25

50

75

100

125

150

175

Time, second

(c) Real Time Distortion Fig. 9 In-Plane Distortion/Movement

200


Conclusion The laser cutting induced distortion was systematically investigated with experimental methods and finite element analyses in a production environment to understand distortion mechanisms and develop effective modeling procedures. Two types of distortions have been identified. One is dominated by in-plane movement. The other is dominated by out-of-plane distortion. The latter can exhibit buckling for skeletons under certain conditions. Three types of distortion were observed in laser cutting process: cut part in-plane movement, out-of-plane distortion and skeleton movement. The plate in-plane movement is about 1mm and the out-ofplane distortion is about 1.3mm which is relative small for the large cut plate and cut part. But it could be important for the following cutting as well as the down-stream fabrication process. The comparison between experimental measurements and predictions shows that the finite element based cutting modeling procedures can predict cutting distortion with reasonable accuracy for cutting distortion mitigations in practical applications.

Magnification Factor: 20

Laser here

Clamped Fig. 10 In-Plane Distortion during Cutting

Reference 1 Ueda Y., Murakawa H., Gu S. M., Okumoto Y. and Ishiyama M. (1994), “FEM Simulation of Gas and Plasma Cutting with Emphasis on Precision of Cutting”, Trans. JWRI, Vol. 23, No. 1, pp. 93-102 2 Hung C. I., and Tsai J. S. (1996), “Magnetic Effects on the Linear Behavior of Molten Flow in a Laser Cutting Process”, J. Phys. D: Appl. Phys., Vol. 29, pp. 3022-3031. 3 Matsuyama K. I. (1997), Mathematical Modeling of Kerf Formation Phenomena in Thermal Cutting”, Welding in the World, Vol. 39, No. 1, pp. 28-34. 4 Na S. J., Yang, Y. S., Koo H. M. and Kim T. K. (1989), “Effect of Shielding Gas Pressure in Laser Cutting of Sheet Metals”, Journal of Engineering Materials and Technology, Vol. 111, pp. 314-318. 5 Masumoto I., Kutsuna M. and Ichikawa K. (1992), “Relation between Process Parameters and Cut Quality in Laser Cutting of Aluminum of Aluminum Alloys”, Transaction of Japan Welding Society, Vol. 23, No. 23, No. 2, pp. 7-14 6 Cao Z., Dong P., and Brust F. W. (1998), "A Highly Efficient Heat-Flow Solution Procedure", Proceedings of ICES'98, October 7 - 9, Atlanta.