CORROSION RATE PREDICTION OF MAGNESIUM ALLOY USING MACHINE LEARNING
Magnesium as degradable biomaterial is a suitable choice for bone implant because it has low density and has stiffness modulus that similar to human bone. However, the use of magnesium is still limited because it degrades rapidly, especially in environments with high Cl- levels such as in the human...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/51952 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Magnesium as degradable biomaterial is a suitable choice for bone implant because it has low density and has stiffness modulus that similar to human bone. However, the use of magnesium is still limited because it degrades rapidly, especially in environments with high Cl- levels such as in the human body. The addition of Zr to Mg is considered to refine Mg matrix grains and increase the corrosion resistance of Mg alloys. Methods to determine optimal composition as an initial reference before experiments needs to be researched. In this study, the Bagging Regressor (BR), Random Forest (RF) and Gradient Tree Boosting (GB) are used as machine learning. The results of this study is generate a model with the accuracy of the RF is RMSE: 1.445; MAE: 1.035, R2: 0.728, BR is RMSE: 1.412; MAE: 1.012, R2: 0.74, while the GB is RMSE: 1.345; MAE: 0.894; R2: 0.764. The highest error difference for the RF model is 1.33 mm/year, BR model is 1.004 mm/year while the GB model is 1.224 mm/year. The recommendation of wt% Zr based on the lowest corrosion rate prediction is Mg-0.7Zr with a corrosion rate of 3.222 mm/ year. |
|---|