VINIF.2021.DA00192 – The application of Machine learning and digital twin model to structural health monitoring based on big data collected from a smart sensor network

Principle Investigator
Assoc.Prof. Bui Tien Thanh
Host Organization
University of Transport and Communications

The project is implemented with 3 main goals:

  • Develop and deploy the installation of smart sensors to measure and analyze vibration characteristics to monitor the health of large bridge structures.
  • Build a program (toolbox) to analyze big data using AI (Artificial intelligence) and stochastic system identification, automatically identify and obtain structural vibration characteristics.
  • Build a digital twin model to monitor the structural health of key large-span bridges in the transportation system.

Main contents of the project

  • Conduct laboratory and field experiments, especially focusing on smart sensor devices to collect data reflecting the behavior of building structures and renewable energy capture.
  • Build a digital twin model that can reflect the states and characteristics of real structures that change over time.
  • Collect and process data sets obtained from smart sensors.
  • Research and build artificial intelligence models combined with established digital twin to analyze large data sets obtained from sensors, health monitoring, and construction damage detection.
  • Building a synchronous health monitoring system for key transportation projects in Vietnam.

Project impact

Technology using smart sensors consumes little energy, reduces costs, and is environmentally friendly.

  • Smart sensors can obtain large amounts of data during structural health monitoring, thereby accurately reflecting the structural status and condition. Enhance the accuracy of bridge inspection results.
  • Does not obstruct the circulation of vehicles on and under the bridge (no need to stop traffic like traditional inspection methods).
  • There is no cost for hiring workers to install bridge inspection equipment (beams or conductor equipment…) and no cost for hiring workers to ensure traffic during the bridge inspection process.
  • Reduce construction maintenance costs and bridge incidents to the lowest level. The proposed technology uses permanently installed sensors, the data collected is real-time data combined with artificial intelligence that can analyze and identify occurring problems with bridge construction quickly and accurately.
Principle Investigator
Assoc.Prof. Bui Tien Thanh
Host Organization
University of Transport and Communications

Tags

Expect Progress
15/11/2021
15/06/2022
Phase 1

– Submit article number 1 for publication in a domestic journal.
– Dataset collected from sensors.
– International Conference Paper number 1 accepted for publication.
– Scientific paper submitted for publication in a Q1 journal.
– Q1 paper number 2 submitted for publication.
– 01 student defending master’s thesis.

15/02/2023
Phase 2

– Finite Element Model and presentation report of the model;
– International Conference Paper number 2 accepted for publication.
– Q1 paper number 3 submitted for publication.
– Report;
– Article number 2 published in a domestic journal;
– Support NCS in defending thesis;
– Q1 paper number 1 accepted for publication;
– Q1 paper number 4 submitted for publication.

15/11/2023
Phase 3

– Program code for detecting defects;
– Q1 paper number 3 accepted for publication;
– Q1 paper number 4 accepted for publication;
– Q1 paper number 5 accepted for publication;
– Summary report;
– Registration of solution accepted.

Tags