Temperature-independent localization algorithm using guided wave interrogation methods

A novel probabilistic approach for damage localization and prognosis including temperature compensation

“…Therefore, a truncated multivariate Gaussian prior probability was applied to the lug joint near the shoulder at which the crack originated with a mean location of (211, 80) mm (see Figure 4) and covariance parameters of s xx , s yy , s xy , and s yx as 15, 15, 0, and 0 mm, respectively. A brief discussion of the performance using the localization method with the fixed prior is given here; a more complete discussion can be found in previous works (Hensberry, 2013;Hensberry et al, 2013). Figure 8 For the 4 temperatures and 13 crack lengths tested, the average crack tip localization error was found to be approximately 9 mm.…”

“…Time-of-flight extraction using grouped MPD A complete description of the grouped MPD algorithm used in this investigation for time-of-flight extraction can be found in the literature (Hensberry, 2013;Hensberry et al, 2013). One of the main difficulties in applying signal processing techniques to SHM wave signals is that the portion of the signal that contains information about the damage often has a small amount of energy compared to the dominant modes in the signal.…”

“…When attempting to localize damage at unknown temperatures, the largest factor contributing to error is uncertainty in the speed of the wave that is interrogating the damage (Raghavan and Cesnik, 2008). In order to achieve damage localization capability that is robust to unknown temperature variation effects, the temperature estimation and velocity compensation algorithm presented in Hensberry et al (2013) and Hensberry (2013) are utilized. As described below, this approach estimates the temperature to determine the actual velocity of the waves in the structure and utilizes this temperature-compensated velocity in the damage localization algorithm.…”

“…One of the primary research areas in structural health monitoring (SHM) (Farrar et al, 2004; Farrar and Worden, 2007) is that of damage localization and prognosis (Kim et al, 2011; Soni et al, 2010). The goal is to reliably locate and predict damage in complex structures, while accounting for the uncertainties in sensor measurements and maintaining robustness to variation in environmental parameters such as temperature.…”

A hybrid prognosis model for predicting fatigue crack propagation under biaxial in-phase and out-of-phase loading

“…For instance, the specimen can be instrumented with multiple piezoelectric transducers and damage localization and quantification can be performed using guided Lamb wave based methods. 29,30 The crack length data obtained from these sensors can be used as an input to the prognosis model for predicting crack growth. The novelty of the proposed approach is the development of a generalized hybrid prognosis model by fusing physics-based modeling with data-driven approaches to accurately predict fatigue crack growth propagation and fatigue life under a variety of complex biaxial loading conditions using very less training data.…”