Early stage damage detection in concrete using coda signals and AI

Principal Investigators 

Prof. Dr.-Ing. Wolfgang Utschick
Professorship for Methods in Signal Processing, Department of Electrical and Computer Engineering, TUM School of Computation, Information and Technology

Prof. Dr.-Ing. Prof. h. c. Christoph Gehlen
Chair for Materials Science and Testing, TUM School of Engineering and Design

Project summary

The economic performance and stability of modern society depend on the reliability and durability of reinforced concrete structures, a fundamental component of our built world. Therefore, maintenance of infrastructure remain one of the primary core tasks of the government. If damage at an early stage is detected and precautionary measures are applied, maintenance costs can be significantly reduced and lives can be saved by preempting failure. Concrete damage at an early stage is characterized by microcracks, much smaller than the aggregate size whose detection is not possible using conventional ultrasonic (US) monitoring. However, the multiple-scattered late arriving US signals (i.e., Coda Signals) contain rich information that detects weak changes. While the high precision and sensitivity of the Coda Signals can be used to identify precursor damage events that precede catastrophic failure, extracting this information is challenging. In this project (as a spin-off of DFG FOR 2825: a method for detecting damage precursors leveraging Coda Signals and Machine Learning (ML) approaches will be developed. Lab measurements of Coda Signals serve as a basis for generating a feature database using ML. Using this feature database, an AI-based damage precursor identifier and classifier is developed and finally applied to the ongoing monitoring of the underground station Scheidplatz in Munich. Detecting such precursor damage will be a breakthrough for implementing a robust and cost-effective early warning system for reinforced concrete structures.