1 Introduction
2 State of the art
3 Approach of a self-optimizing process planning
3.1 Simulation of the polishing process
3.1.1 Contact width model of the polishing tool
3.1.2 Initial tool path
3.1.3 Representation and processing of the local workpiece information
3.2 Adaptation of the polishing process to reflect individual processing requirements
4 Experimental investigations
Parameter | Setting |
---|---|
Cut-off DIN EN ISO 4288 | 0.8 mm |
Roughness parameter DIN EN ISO 4287 | Ra in μm |
Measuring direction | Polishing direction |
Measuring position | Center of polished path |
Vertical resolution | 0.77 nm |
4.1 Generating training data sets
4.2 Extension of the planning algorithm with a self-optimizing roughness model
Parameter | Setting |
---|---|
App | Regression learner |
Validation | Cross-validation, 50 folds |
Model type | Cubic SVM |
Kernel function | Cubic |
Kernel scale | Automatic |
Box constrait | Automatic |
Epsilon | Automatic |
Standardize data | True |
R2 | 83% |
4.3 Validation of the self-optimizing planning algorithm
Workpiece | Iteration | |||||
---|---|---|---|---|---|---|
Initial | 1st | 2nd | 3rd | 4th | 5th | |
P1 | 0.07 | 0.22 | 0.15 | 0.31 | 0.08 | 0.06 |
P2 | 0.07 | 0.10 | 0.05 | 0.04 | Done | Done |
P3 | 0.07 | 0.13 | 0.02 | 0.02 | Done | Done |
R60 | 0.05 | 0.28 | 0.00 | 0.04 | Done | Done |
R90 | 0.05 | 0.15 | 0.07 | 0.25 | 0.03 | 0.07 |