NETWORK PARAMETERS ANALYSIS OF THE BRAZIL NUT EFFECT PHENOMENON ON BINARY MIXTURE TWO-DIMENSIONAL GRANULAR SYSTEMS

• Main Author: Iqbal Rahmadhan Putra, Muhammad
• Format: Final Project
• Language: Indonesia
• Subjects: Fisika
• Online Access: https://digilib.itb.ac.id/gdl/view/41274
• Institution: Institut Teknologi Bandung

The Brazil Nut Effect is one of the phenomena that could be observed in a binary mixture of granular systems when two types of granules are vertically vibrated the two types of granules tend to be partially or completely segregated, in previous studies it was observed that granules larger in size (volume or mass) goes up to the top. The method used to characterize conditions at all times during the segregation process has been carried out globally by determining the segregation coefficient and calculating the center of mass of the system. Both methods can only describe the condition of the system globally, but not to arrive at the mesoscale. To be able to characterize the system for global scale and mesoscale at each time, network analysis is used which has been used in the analysis of grain systems with compression. In this research, characterization of the condition of the grain system with various initial conditions at each time during the segregation process due to the Brazil Nut Effect occurs. The methodology of the research is divided into three; (1) experiments for image capture of granular system conditions, (2) image processing for networks extraction, and (3) network analysis to obtain network parameters. Experiments were carried out using to types of initial condition, namely dense and sparse configuration. Variations were also made on the value of the frequency and amplitude of vibrations that form normalized accelerations. The calculated network parameters are network density, betweenness centrality, and number of communities which are the results of modularity optimization. The calculation results show different characteristics of the two initial configuration types for each network parameter. Furthermore, the correlation coefficient value of each network parameter is calculated against the segregation coefficient which shows that network density has a strong linear correlation with the segregation coefficient compared to other network parameters.