Detection of seepage-induced internal instability using acoustic emission

<div>Seepage-induced internal erosion is and has for a long time been a matter of concern regarding water-retaining earth structures. Uncertainties about which conditions can be considered hazardous and incomplete knowledge about the involved physical dynamics and the structures themselves illustrate why this is such a difficult problem to solve. As an implication, predicting the occurrence of internal erosion through modelling and implementation of theoretical frameworks, despite being highly sensible and desirable, are still insufficient approaches. Conversely, the ability to directly detect the occurrence of internal erosion in its early stages is a way to hugely minimise the effects of structural damage by providing early warnings and possibly avoiding disaster – as well as producing information for better facing this matter in the future.</div><div>Current detection approaches are limited by either inferring its occurrence in already advanced stages (too late for intervention without calamity), implying only fluid seepage (being blind to particle transport, which is critical) or simple susceptibility assessments based on material particle size distributions (which ignores random non-homogeneities and discounts differences between design and construction). A method that specifically detects particle transport by fluid seepage (or internal instability) is lacking but would enable timely interventions.</div><div>In this research, acoustic emission (AE) has been investigated for this application. Results from laboratory experiments with a bespoke, purpose designed and built permeameter show that seepage-induced internal erosion processes can detected and monitored using AE. The experimental programme included 22 tests (in two rounds – pre- and post-commissioning of bespoke apparatus) and employed materials used in the construction of earth dams and with varying degrees of estimated internal instability (which varied depending on the different criteria used). These soils were subject to permeating fluid flow and monitored for changes to hydromechanical parameters and the development of internal erosion, from the start of seepage-induced particle movement to piping.</div><div>The measurement and interpretation of AE in this context was based on filtering unwanted environmental noise and registering when the signal exceeds a predefined/calibrated threshold (i.e. employing an approach to minimise false alarms). A strong correlation between the occurrence of internal erosion and detectable AE has been found. It was possible to use AE to differentiate between fluid flow with and without particle transport – especially the transition from one to the other, or the onset of internal erosion – as well as observing the evolution of the erosion processes. AE rates tended to increase proportionally to the transport soil particles, with elevated AE activity occurring during the formation of preferential flow pathways through the soil.</div><div>The observations produced in this study show that the use of AE for monitoring the occurrence of seepage-induced internal erosion is feasible, with the necessity of a trained professional to analyse the produce data and account for particularities of individual circumstances. However, the datasets and new understanding produced in this study also indicate that the development of automated interpretation algorithms can be done (e.g. by following the recommendations made in this thesis).</div>