SENSING MECHANISM AND REAL TIME (SMART) COMPUTING FOR GRANULAR MATERIALS

Restricted (Penn State Only)
Author:
Liu, Shushu
Graduate Program:
Civil Engineering
Degree:
Doctor of Philosophy
Document Type:
Dissertation
Date of Defense:
August 22, 2018
Committee Members:
  • Hai Huang, Dissertation Advisor
  • Hai Huang, Committee Chair
  • Ming Xiao, Committee Member
  • Tong Qiu, Committee Member
  • Shimin Liu, Outside Member
  • Tong Qiu, Committee Chair
  • Tong Qiu, Dissertation Advisor
Keywords:
  • Railroad
  • Geotechnical Engineering
  • Ballast
  • SmartRock
Abstract:
Granular materials play significant roles in many civil engineering infrastructures, such as highway, railroad, etc. Characteristics of granular materials have played a critical role in the infrastructure performance and often been evaluated in the laboratory test, field investigation, and numerical simulations. The granular materials are usually embedded below the surface of highway and railroad, which makes it difficult to diagnose distresses by visual inspection from the surface if unfavorable conditions exist. Additionally, the behavior of granular material is hard to predict accurately using numerical simulations due to its irregular shape and anisotropic property. In order to overcome the limitations mentioned above, a wireless device, “SmartRock” was developed in this study. The SmartRock is a 3D printed device to be similar to the realistic shape of a granular particle. It is composed of a tri-accelerometer, tri-gyroscope, and tri-magnetometer that can record its translation, rotation and orientation, respectively. It can send the recorded data to a base station through Bluetooth in real-time. A program developed in MATLAB is able to analyze the recorded movement data and visualize the real-time SmartRock particle movement. A sleep mode was built in the SmartRock to save battery consumption, which enables long-term monitoring. The developed SmartRock was applied in the laboratory to investigate the effect of geogrid on railroad ballast particle movement, and in the field to characterize the railroad ballast particle movement in the revenue service lines where mud spots exist. The results show that the particle movement pattern is closely related to the performance of the granular materials; particle appears to experience more intensive translational acceleration and rotation when the condition of the granular material is not stable. The results also show that the SmartRock is capable of recording and visualizing real-time particle movement, thus can be used as a field monitoring tool to evaluate the performance of granular materials. Further, the SmartRock measurements were integrated into numerical simulations. A data-fusion based computing framework, sensing mechanism and real time (SMART) computing for granular materials, was developed and validated. The SMART computing consists of: 1) real-time data acquisition of particle kinematics through the SmartRocks that are embedded at discrete locations in a granular assemblage, and 2) a built-in data-fusion-based algorithm using the Kalman filter to integrate the prediction generated by DEM and the measurements reported by “SmartRocks”. The SMART computing algorithm was developed and its performance was investigated through a series of ball collision experiments – two-ball center-to-center, two-ball off-center, and multi-ball collisions. The SMART computing was then applied to simulate large-scale triaxial shear test. According to the comparison among the laboratory tests, DEM-only and SMART computing simulations, it is concluded that the SMART computing is able to improve the computation accuracy of the triaxial shear test, and tackle the problems beyond what the DEM-only simulations could do so far.