Damage detection by applying statistical methods to PZT impedance measurements

Abstract

An effective integrated structural health monitoring system must include a method of sensing and a process of damage identification that are optimized to work together. The result is a system that provides an automated and quantified assessment of damage present in a structure. Two candidates for such a symbiosis of sensing and damage identification are impedance-based measurement and statistical process control. The impedance-based structural health monitoring method uses a high frequency signal to excite a structure through a bonded piezoelectric patch and measures the impedance response of the excited structure across a frequency spectrum. In structural damage cases such as threads loosening or a crack developing, the structure in question will begin to show a change in impedance. Once measured, a damage sensitive feature from this impedance change can be statistically quantified into different damage cases by statistical process control. This paper addresses impedance measurements from experimental structures and a subsequent statistical method for quantitatively determining when the impedance signature of the structures has changed significantly enough to warrant the classification of "damaged". Simple features and hypothesis testing algorithms are explored in an effort to create real-time solutions and reduce the complexity of damage identification for future use in low resource integrated structural health monitoring systems.

Convolutional neural network and impedance-based SHM applied to damage detection

“…In addition, with the current technological advances on data acquisition systems, SHM methods may consider much larger files of monitoring data, which provide greater reliability to structural prognosis [15]. In contrast, the greater the amount of storage data, the more complex the SHM methods become, requiring the use of improved techniques that operate in the field of Big Data [16].…”

“…Because of this complexity, the most part of the SHM methods use high frequency sampling and abstraction and pattern classification techniques [3,15]. These techniques map the unknown relationship between the current state of pristine and damaged structural patterns observed.…”

Early Age Concrete Strength Monitoring Using Power Spectral Density and Wavelet Packet Analysis

“…e results show that the EMI method is very sensitive to the early development of concrete strength, so it is considered suitable for the monitoring of concrete strength development. e research from professor Inman [15][16][17] focuses on the monitoring of structural mechanical properties of PZT sensors. e research team of Zhu [11,[18][19][20] used PZT sensors to conduct a series of experimental studies on bolt loosening of steel frame structures, solidification process of concrete cubes, and temperature compensation of piezoelectric impedance technology, which promoted the application of EMI method in civil engineering.…”