Browsing by Author "Al Obaidi, Aymen Abdulqader Abbas"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
thesis.listelement.badge PREDICTION OF KINEMATIC VISCOSITY AND DENSITY OF DIESEL FUEL FROM PHYSICAL PROPERTIES BY ARTIFICAL NEURAL NETWORKS(2022-01-20) Al Obaidi, Aymen Abdulqader Abbas; KAYI, Hakan; E. MACHIN, NesrinIn this study, the first purpose is to predict the kinematic viscosity and density of diesel fuel from the physical properties of it by employing artificial neural network modeling approach to be able to save time and money spent on experimental studies, and the next one is to ascertain a simple formula to kinematic viscosity and density. The experimental data for diesel fuel utilized throughout this study is taken directly from the North Oil Company in Iraq, hence the original and unique source of data has been used. In daily routine, every truck holding diesel needs to be monitored, and density and viscosity measurements need to be performed. Nonlinear autoregressive neural networks with external input (NARX) and a typical feedforward neural network with backpropagation algorithm (NN) are used in the prediction of kinematic viscosity and density. Evaluation of the obtained results indicated promising performance for these networks. In the prediction of the density values, NARX type neural network performed better than NN, and for the prediction of kinematic viscosity values NN performed superior to NARX. In addition, multiple linear regression (MLR) methodology is employed and the coefficients of the five independent variables for the first order linear equations are obtained for kinematic viscosity and density.