Title:
Real Time Plastic Viscosity Prediction Through Video Recognition
Author(s):
Guo
Publication:
Web Session
Volume:
ws_S22_Guo.pdf
Issue:
Appears on pages(s):
Keywords:
DOI:
Date:
3/28/2022
Abstract:
Plastic viscosity is a key property of ultra-high-performance concrete (UHPC), which needs to be well controlled to achieve superior fresh and mechanical properties as well as durability. A low viscosity can cause segregation of steel fibers, thus affecting the flexural properties, while a high viscosity may result in a high air content and adversely influence the mechanical properties and durability of UHPC. Current tools for measuring plastic viscosity are mainly sophisticated rheometers, which are expensive and hardly used in field applications. This research presents an innovative method to assess the viscosity of UHPC using a video taken during mixing of UHPC. The unique flow features of UHPC are extracted from the video and analyzed using a long-term recurrent convolutional network. The effects of viscosity on fiber distribution, air content, and flexural behaviors of UHPC are investigated. The results show that the presented method enables real-time assessment of the viscosity and reasonable prediction of the flexural properties of UHPC. This study is expected to greatly facilitate quality control of production of UHPC.