I am Saleh, a nanofluid researcher seeking help with ANN implementation for predicting the thermal conductivity of Al2O3/water-ethylene glycol hybrid nanofluid.
I have prepared a small dataset of 30 data points which I will divide into 15% for validation, 70% for training, and 15% for testing. I have selected the following specifications for the ANN based on other research trends :
-
ANN morphology: Multilayer perceptron (MLP) network
-
Network type: Feed-forward Artificial Neural Network (ANN)
-
Training Method: Levenberg–Marquardt algorithm
-
Error Criteria: Mean Square Error (MSE)
-
Number of Hidden Layers: 1
-
Number of neurons in the hidden layer: Increase from 7 to 21 with a step size of 2
-
Hidden Layer function: Tangent-sigmoid function
-
Output Layer function: Linear (Purelin activation function)
-
Inputs: Volume fraction (φ) and temperature (T)
-
Outputs: Thermal conductivity of the nanofluid (knf)
With my knowledge in nanofluids but limited coding skills, I kindly request assistance from the community in implementing the ANN for my project.
for anybody that is willing to help
contact info below