Application of ANFIS in Predicting of TiAlN Coatings Hardness
In this paper, a new approach in predicting the hardness of Titanium Aluminum Nitrite (TiAlN) coatings using Adaptive Neuro-Fuzzy Inference System (ANFIS) is implemented. TiAlN coated cutting tool is widely used in machining due to its excellent properties. The TiAlN coatings were formed using Phys...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
2011
|
| Online Access: | http://eprints.utem.edu.my/id/eprint/197/1/Published-_Application_of_ANFIS_in_Predicting_of_TiAlN_Coatings_Hardness-inproceeding.pdf http://eprints.utem.edu.my/id/eprint/197/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | In this paper, a new approach in predicting the hardness of Titanium Aluminum Nitrite (TiAlN) coatings using Adaptive Neuro-Fuzzy Inference System (ANFIS) is implemented. TiAlN
coated cutting tool is widely used in machining due to its excellent properties. The TiAlN coatings were formed using Physical Vapor Deposition (PVD) magnetron sputtering process. The substrate sputtering power, bias voltage and temperature were selected as the input parameters and the hardness as an output of the process. A statistical design of experiment called Response Surface Methodology (RSM) was used in collecting optimized data. The ANFIS model was trained using the limited experimental data. The triangular, trapezoidal, bell and Gaussian shapes of membership functions were used for inputs as well as output. The results of ANFIS model were validated with the testing data and compared with fuzzy and nonlinear RSM hardness models in terms of the root mean square error (RMSE) and model prediction accuracy. The result indicated that the ANFIS model using 3-3-3 triangular shapes membership function obtained better result compared to the fuzzy and nonlinear
RSM hardness models. The result also indicated that the ANFIS model could predict the output response in high prediction accuracy even using limited training data. |
|---|