Fuzzy rule-based model to estimate surface roughness and wear in hard coatings
n this paper, a new approach in predicting the surface roughness and flank wear of hard coatings using fuzzy rule-based model is implemented. Hard coatings is important for cutting tool due to its excellent performances in 800°C temperature during high speed machining. The coating process were run u...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Published: |
2012
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/47017/ http://dx.doi.org/10.1109/ICSMC.2012.6377873 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | n this paper, a new approach in predicting the surface roughness and flank wear of hard coatings using fuzzy rule-based model is implemented. Hard coatings is important for cutting tool due to its excellent performances in 800°C temperature during high speed machining. The coating process were run using Physical Vapor Deposition (PVD) magnetron sputtering process. An experiment matrix called Response Surface Methodology (RSM) was used to collect data based on optimized data point. Sputtering power, substrate bias voltage and substrate temperature were used as the variables, and coating roughness and flank wear as the output responses of the coating process. The collected experimental data were used to develop fuzzy rules. Five triangular membership functions (MFs) for input variables and nine MFs for output responses were used in constructing the models. The results of fuzzy rule-based models were compared against the experimental result based on the percentage error, co-efficient determination (R2) and model accuracy. The rule-based model for coating roughness showed an excellent result with respective smallest percentage error, R2 and model accuracy were 0.85%, 0.953 and 89.20% respectively. Meanwhile, the fuzzy flank wear model indicated 6.38%, 0.91 and 81.79% for smallest percentage error, R2 and model accuracy. Thus, fuzzy logic can be a good alternative in predicting coating roughness and flank wear in hard coatings. |
|---|