Bridges, as well as off-shore wind turbines and other marine structures, are susceptible to failures due to local scour, which is a dynamic phenomenon that is caused by flowing water removing the bed material from around piles, piers, and abutments. If extended over a critical depth, scour can jeopardize the stability and safety of overwater bridges. In fact, scour is the predominant cause of overwater bridge failures in North America and around the world. Monitoring, as part of bridge maintenance, can prevent scour-induced damage and failure by continuously measuring the extent of scour so that preventative measures can be taken in a timely manner.

Over the years, numerous sensing systems have been developed for monitoring bridge scour by measuring scour depth at locations near bridge piers and abutments. Due to the limitations of periodic inspections conducted by trained divers and by using portable instruments, fixed monitoring systems have become the viable solution. Existing fixed scour sensors include sonar systems, float-out devices, and tilt meters, to name a few. These systems each offer unique advantages, but have limitations (e.g., high costs, low reliability, limited accuracy, etc.) that have restricted their implementation in practice. Therefore, attempts to develop more efficient monitoring schemes continue. In this study two novel scour sensing schemes were evaluated. The first uses driven piezoelectric rods to continuously measure scour depth;​ and in the second, buried dissolved oxygen (DO) optodes detect scour at discrete depths. Laboratory flume experiments were conducted to validate the proposed sensing systems.

In the first sensing scheme, piezoelectric rods are driven into the stream bed at a location where scour depths are wanted. As the scour hole extends, the exposed length of the rod changes, causing the flow-induced voltage signal acquired from the sensor to also vary. Scour depth at the sensor location is determined based on the fact that the natural frequency of the cantilevered sensing rod is inversely related to its length. Prototype piezoelectric rods, in which a poly(vinylidene fluoride) (PVDF) polymer strip forms the main sensing component, were designed and developed. Following various preliminary validation tests, extensive laboratory experiments were performed in which the in-house piezoelectric sensing rods were driven into the soil surrounding a mock bridge pier inside a flume simulating scour conditions. The piezo-sensor was calibrated through eigenfrequency analyses. The second sensing system utilized commercially available miniature DO probes. DO

levels are very low in streambed sediments, as compared to the standard level of oxygen in flowing water. Therefore, scour depths can be determined by installing sensors to monitor DO levels at various depths along the buried length of a bridge pier or abutment. The measured DO is negligible when a sensor is buried but would increase significantly once scour occurs and exposes the sensor to flowing water. A set of experiments was conducted in which four dissolved oxygen probes were embedded at different soil depths in the vicinity of a mock bridge pier inside a laboratory flume simulating scour conditions. The measured DO jumped to water DO levels once scour exposed the sensing tip of the probes to flowing water, thereby providing discrete measurements of the maximum scour depth.

The sensing concepts behind both scour monitoring schemes were confirmed through comparing the detected and observed scour depths. The PVDF-based sensors provide continuous scour depth measurements, as opposed to discrete ones offered by the DO sensing system. Both sensing schemes were also able to detect any subsequent refilling of the scour hole through the deposition of sediments. Following separate analyses of the results, future research is suggested for the two sensing techniques to gain a better understanding of their advantages, shortcomings, and potential applications.

In addition to developing and validating the aforementioned scour sensing schemes, research was conducted aimed at creating a practical warning-time based framework for scour sensor response interpretation. First, the general form of the framework, applicable to a wide range of damage detection operations, was developed. The purpose of structural health monitoring (SHM) is to diagnose any damage or malfunction in an engineering system in a timely manner. Timely detection implies that sufficient warning time is given to perform required maintenance to prevent structural failure. Warning time information is therefore very useful in the design and planning of maintenance procedures. The framework developed as part of this research, is a simple and practical tool for predicting warning times given detected damage (i.e. sensor outputs). The framework incorporates a probabilistic analysis of damage progression such that the uncertainty in warning times can also be determined and used for risk-based decision making. To demonstrate the framework’s applicability to scour monitoring, a detailed example was considered, where the progression of bridge scour was obtained through computational fluid dynamics (CFD) simulations using the software Flow-3D. The resulting diagrams from the framework can be used as an effective tool in estimating the warning time and the uncertainty in the warning time given a detected scour depth. The warning information is extremely useful in identifying and planning the required maintenance procedures based on the available resources.