Table 3 ACOFIS setup for predicting pressure and temperature in the same setup.
ACOFIS for predicting pressure | ACOFIS for predicting temperature | |
|---|---|---|
CFD case study | Nanofluid turbulent flow in heated porous pipe | Nanofluid turbulent flow in heated porous pipe |
AI method | Combination of ACO with FIS | Combination of ACO with FIS |
Material of case study | Nanofluid (Al2O3) | Nanofluid (Al2O3) |
Number of input in the best intelligence | 3 | 3 |
Pheromone effect in the best intelligence (ACO parameter) | 0.1 | 0.4 |
Changes in number of inputs was evaluated(FIS parameter) | 2,3 | 3 |
Changes in pheromone effect was evaluated(ACO parameter) | 0.1, 0.2, 0.3, 0.4, 0.5, 0.6 | 0.1, 0.2, 0.3, 0.4, 0.5, 0.6 |
The highest of correlation coefficient in testing process with 100% of data | 0.994 | 0.968 |
P(%) percentage of used data in training process | 77% | 77% |
Number of data | 2 inputs (537) and 3 inputs (2685) | 3 inputs (2685) |
Number of iteration | 115 | 115 |
Type of data clustering | FCM clustering | FCM clustering |
Type of membership function | Guassmf | Guassmf |
Number of MFs for each input | 16 | 16 |
Number of rules (which is for hidden layer of FIS) | 16 | 16 |
Number of membership functions (MFs) for output | 16 | 16 |
ACOFIS input1 | x-direction | x-direction |
ACOFIS input2 | y-direction | y-direction |
ACOFIS input3 | Nano particle Fraction = 0.5,0.8,1,1.5,2% | Nano particle Fraction = 0.5,0.8,1,1.5,2% |
ACOFIS output | Pressure | Temperature |
