An approach based on gamma backscattering technique to measuring the density of liquid using the low-intensity radioactive source.

This work aims to develop a practical solution to measure the density of a liquid. Two purposes of this study: (1) using a low-activity source to measure the density of a liquid, and (2) simplifying the experimental arrangement to reduce the size and weight of the measuring system. The proposed solution is to develop a measurement technique without both detector and source collimators, while it considers an appropriate technique for analyzing the backscattering spectrum. To validate the proposed method, we used two groups of liquid: one group of liquids with a certified density and one group of liquids collected from the market. For the first group, the obtained results showed that the relative errors between the measured density and the reference one are below 6.8% and the uncertainties in density are below 4%, which confirms the feasibility of the proposed approach. For the second group, the liquids collected from the market include 70 percent alcohol, cooking oil, saltwater, fresh milk, diesel oil, dishwashing liquid, machine oil, and wine. The results obtained show that the relative errors between the densities determined by the proposed method and densities determined by the traditional method using density kit are less than 4.3%, the uncertainties in density when using the proposed method are below 3.2%. These results initially confirm that the proposed solution is completely applicable in measuring the density of a liquid.

ANN coupled with Monte Carlo simulation for predicting the concentration of acids.

The present study proposes a new approach for determining the concentration of acids. The method is based on the combination of Monte Carlo simulation and artificial neural network (ANN) technique for predicting the concentration of acids. Firstly, a Monte Carlo simulation model is validated based on the comparison of simulation data with experimental data. Then, the whole data derived from the Monte Carlo simulation using the MCNP code was used to train the ANN model. The trained ANN model was used to predict the percentage concentrations of 14 acid samples, which yields the maximum relative deviation between the predicted and the reference concentrations is less than 3.5%.