A new approch to estimate kerosene properties by artificial neural networks and crude oil properties

In this study, estimation of properties of kerosene from crude oil properties is considered. There is no model for this purpose. We use neural network simulation to correlate these properties. One hidden layer network is selected for this simulation and number of hidden layer is calculated by trial and error. The network with 4 hidden layers has the best agreement with experimental data.

Keywords:

kerosene , crude oil , artificial neural networks , property