Simulation concept for machined surface roughness and shape deviations prediction.

Polakovic, Milos ; Buransky, Ivan ; Peterka, Jozef 等

1. INTRODUCTION

By machining process with geometrically defined cutting tools the deviations from designed CAD surfaces as well as dimension's deviations always arise. There are various factors such as machine-tool-clamping system stiffness, cutting force intensity and orientation, cutting edge geometry, tool wear, vibrations, used finishing strategy, CNC positioning precision, etc., which have different level of effect on the arising deviations.

Our goal is to develop a simulation algorithm, which would be able to simulate finishing process, specifically machining with ball end mill. Designed algorithm will take into account inputs such as tool geometry, stiffness, material properties, deflection, cutting forces and cutting parameters. The simulation output data will be used for graphical display of machined surfaces and for further analyses of shape deviations and surface roughness as well. Timeline of force intensity and tool load will also be analyzed and used for toolpath or cutting parameters optimization.

2. MACHINED PART VS. CAD MODEL

A low cost experiment for evaluation of machined surface deviations and roughness was prepared. According to information's from resources (Choi & Jerard, 1998), (Fornusek & Rybin, 2000) the appropriate model features and surface shapes were designed using the CAD software Delcam PowerSHAPE 5.8.21. The roughing and finishing strategies in CAM software PowerMILL 6.0.0.8 were applied for generating various roughing toolpaths as well as finishing toolpaths for ball end mill. In order to keep the machining costs as low as possible, the surfaces were machined with 3axis continuous milling strategy and only the shape deviations and surface roughness were evaluated. The test workpiece was machined by CNC milling machine VMC EAGLE1000 (positioning accuracy [+ or -] 0,005 mm, repeatable accuracy [+ or -] 0,003 mm).

[FIGURE 1 OMITTED]

As a workpiece material the steel with hardness HB 264 was selected. For machining purposes the tool with PVD coated inserts and PVD coated cemented carbide ball end mills have been used.

2.1 Evaluation of machined workpiece

Machined workpiece has been digitized with non-contact optical 3D scanner GOM ATOS I 350 (scanning accuracy 0,02mm). Acquired digital data has been processed in software GOM ATOS v 6.0.2-5 and compared to original CAD model using the feature "CAD Comparison". After this procedure the deviation map has been obtained. This map indicates various areas on machined surface, where the tool deflection created significant deviations up to 0,2 mm.

[FIGURE 2 OMITTED]

2.2 Roughness evaluation

In case of ball end mill finishing, the lowest roughness values (Ra = 0.8 [micro]m) has been obtained on workpiece areas, where the cutting effective radius converged at nominal radius of used ball end mill. The highest values (Ra = 5 [micro]m) has been detected on areas, where the effective radius was lowest, therefore the cutting velocity was insufficient. Various finishing strategies used by machining of this part show a different roughness pattern as well. For roughness evaluation the contact roughness measuring device Surtronick 3+ has been applied.

[FIGURE 3 OMITTED]

3. SIMULATION CONCEPT

The main aim is to simulate finishing process with ball end mill and determine suitability of applied strategy for toolpath generation as well as suitability of selected tool and cutting parameters, considering the arising shape deviations and surface roughness. Computational algorithm will simulate cutting movement of tool edge elements. This movement is a combination of translation and rotation of edge elements on the loaded tool path. The path is generated and saved into a CL data form in the CAM software. The interactions between cutting edge elements and workpiece elements will be detected and the corresponding chip cross sections, cutting forces and tool deflections will be calculated. The core of the algorithm development will be an appropriate implementation of tool, workpiece, cutting kinematics, cutting forces and tool deflection sub-models.

3.1 The smallest building element "The Voxel"

The Voxel will be used as a building element for physical representation of tool, cutting edge and workpiece geometry (Cohen-Or & Kaufman, 1995). Voxel is the smallest volume element, which can contain the information's about local volume (density, material, color, normal vector, forces, etc.). Voxels are usually distributed along the Cartesian's grid using the USD (Uniform Space Decomposition). Because the resolution needed for roughness simulation is very fine, the amount of needed Voxels would exceed any computational capability of common desktop computers. Therefore the algorithm called "Cube-Marching" (Lorensen & Cline, 1987) will be implemented. This algorithm is capable to convert Voxels to triangle mesh on the fly. Using the GPU hardware acceleration the triangle mesh will be recalculated and displayed much faster. Therefore the Voxel engine will be used only for physical representation of models and the triangles will be used for graphical representation of machined surfaces.

[FIGURE 4 OMITTED]

3.2 Tool, cutting edge and workpiece models

Geometrical model of tool will be created in the CAD software, then converted to STL triangle mesh format and finally, using the USD decomposition, converted to Voxel model. The same procedure for acquiring workpiece blank voxel model will be applied as well.

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

3.3 Tool deflection model

Model for tool deflection will be based on results from work (Kim et al., 2003). Approximated equation of deviation between deflected and non deflected tool can be expressed as follows:

[delta] = [[delta].sub.s] + [[delta].sub.f] + [[phi].subs] (Lf - z) (1)

-where [[delta].sub.s] is deflection of the shank, [[delta].sub.f] is deflection of the flute, [[phi].sub.s] is deflection angle of the shank, Lf is length of the flute and z is the coordinate where the deflection is being calculated.

[FIGURE 7 OMITTED]

4. CONCLUSION

The CAM software nowadays provides possibilities of graphical cutting simulation. This type of simulation gives only an idea of cutter movement and material removal, other information's such as cutting forces, tool deflection, vibrations, surface quality after machining is unknown, respectively can not be determined. This paper tries to show a concept how to expand graphical simulation in a combination with physical cutting simulation, where the Voxel elements and triangle mesh are combined into a one cutting simulation packet.

5. REFERENCES

Choi, B.K. & Jerard, R.B. (1998). Sculptured surface machining, Kluwer Academic Publishers, ISBN 0 412 78020 8, U.S.A.

Cohen-Or, D. & Kaufman, A. (1995). Fundamentals of Surface Voxelization, Available from: http://www.sciencedirect.com/ Accessed: 2008-02-02

Fornusek, T. & Rybin, J. (2000). Accuracy determining of multiaxis controlled milling machine, Available from: http://www.fs.vsb.cz/ Accessed: 2007-12-10

Kim, G.M.; Kim, B.H. & Chu, C.N. (2003). Estimation of cutter deflection and form error in ball-end milling processes, Available from: http://www.sciencedirect.com/ Accessed: 2007-12-10

Lorensen, W.E. & Cline, H.E. (1987). Marching cubes: A high resolution 3D surface construction algorithm, Available from: http://portal.acm.org/citation.cfm?id=37401.37422 Accessed: 2007-10-10

POLAKOVIC, M[ilos]; BURANSKY, I[van] & PETERKA, J[ozef] *

* Supervisor, Mentor