APPLICATION OF ARTIFICIAL INTELLIGENCE IN PERMEABILITY ESTIMATION FROM LOG DATA
Certain methods have been developed to estimate permeability value from log data. a several methods have been developed, from empirical correlation through artificial intelligemce. In this case study permeability estimation using fuzzy logic, which tuned by Genetic Algorithm (GA optimized fuzzy logi...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/21403 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Certain methods have been developed to estimate permeability value from log data. a several methods have been developed, from empirical correlation through artificial intelligemce. In this case study permeability estimation using fuzzy logic, which tuned by Genetic Algorithm (GA optimized fuzzy logic) and Artificial Neural Network (Adaptive Neuro Fuzzy Inference Sistem) is compared and derived from the same raw log data as inputs are: density, neutron, resistivity, and gamma ray. Those inputs are then correlated to laboratorium measured permeability. The output is estimated permeability; this estimated permeability is compared to permeability calculated from core data. Two method are proposed to be compared: ANFIS and GA aided Fuzzy logic <br />
<br />
<br />
Result is quite astonishing with R2 value is ANFIS is able to characterize the correlation between log data and core permability with 0.99999998, while GA aided fuzzy logic only has 0.851457. Further analysis is conducted to ANFIS using sensitivity analysis, by changing the number of data used for training. It is stated that even by only using 20% of training data, ANFIS able to correlate with R2 0.658251. This thing improves the level of confidence in generating permeability value just from only a few cored points. |
|---|