A Data Processing and Control System to Support Remote Infrastructure Monitoring

1. Background and Motivation As the nation's transportation infrastructure ages, the need for structural health monitoring has become increasingly important from economic and safety viewpoints. Events, such as the catastrophic failure of the St. Anthony Falls Bridge in August of 2007, have raised public awareness, and motivated engineers to adapt and develop innovative technologies and techniques to complement visual inspections. These efforts coupled with advances in computational power, sensing and communications technologies have led to the formulation of the concept of smart infrastructure. The idea is that sensor networks, monitoring sets of critical elements, can continuously generate structural and functional data that could uncover potential problems (long before they might be spotted by human inspections), and support managerial decisions, e.g., repairs, intended to ensure a facility's ability to support the functions for which it was designed and built, in a reliable, safe, cost-effective, and environmentally-sustainable fashion. The Infrastructure Technology Institute (ITI) at Northwestern University has been at the forefront of the aforementioned efforts, having deployed numerous advanced health monitoring systems in the last 15-20 years. While tremendous strides have been made in addressing technological challenges, e.g., delivering power to sensors, protecting sensors from damage caused by the environment, ensuring reliable communications, much is needed in terms data processing, analysis and control capabilities, in order to achieve real-time health monitoring and management capabilities. The work proposed herein is a significant step in this direction, and complements our research previously funded by ITI and by the National Science Foundation. Specifically, our objective is to develop, implement, and validate a statistical process control framework with capabilities to:  Process measurements related to structural health, traffic loads, and environmental conditions in an integrated fashion, thereby yielding comprehensive, facility-level condition assessment and forecasting capabilities;  Provide real-time, reliable alerts when potential damage or risk of structural change in the facilities is detected; and  Determine the nature of the detected changes, i.e., infer underlying structural properties, and identify possible assignable causes. In terms of connections to the USDOT's research goals, the proposed work involves development and implementation of cutting-edge, transformative research tools to support information management, and decisions related to the management/renewal of surface transportation infrastructure to ensure that it operates in a "state of good repair". Again, we also emphasize that the work is complementary to ITI's work/expertise in deployment of advanced remote monitoring systems. In addition to complying with ITI's reporting requirements, and to efforts to disseminate the work through publications and presentations, the main deliverable from this project will be a demonstration software system (implemented in Matlab). The system will automate the process of analyzing the data streams collected for the Hurley Bridge (Wisconsin Structure B-26-7). This system will be installed on the Institute's server and will support/contribute to the monitoring work of the Research Engineering Group at ITI. The primary goal of structural monitoring is to ensure the longevity and safety of transportation infrastructure (bridges, highways, railways, tunnels, etc.) as well as to optimize their management. It plays a fundamental role during the construction phase to verify the design hypothesis and the overall quality of services. It also allows performance evaluation of new techniques and materials used in construction as well as rehabilitation of infrastructure. More importantly, effective real time monitoring of critical components on infrastructure approaching their life span would help to ensure that they operate safely while allowing the postponement of major investments in rehabilitation. In accordance, there is a booming implementation of Structural Health Monitoring (SHM) systems facilitated with these advanced instrumentation technologies in the U.S. Recent cases of bridges that have been instrumented by the ITI at Northwestern University include the John F. Kennedy Memorial Bridge in Louisville, KY, the Michigan Street Lift Bridge in Sturgeon Bay, WC, as well as the Wisconsin Structure B-26-7 that carries westbound traffic over the Montreal River from Michigan to Wisconsin. Field measurements of structural conditions including strain, acceleration, displacement, etc. as well as exogenous factors including traffic loadings, weather, wind, etc. are collected on critical components of the bridges in a timely fashion. In a previous work, we have been working on developing a process control system for the JFK Memorial Bridge mentioned above. A failed anchor bolt was previously identified in an uplifting bearing assembly, and a retrofit consisting of a replacement anchor was installed coupled with remote monitoring instrumentations. Based on the measurements collected, we incorporated statistical methods to automate the control process, and were able to infer a possible degradation of the bolt approximately 6 weeks before it failed. In preliminary analysis of Wisconsin Structure B-26-7 (shown in the appendix) using tools we have developed, we were able to identify a subtle transversal displacement, consistent with anecdotal reports from the Wisconsin DOT that the bridge is "walking off" the bearings in the transverse direction. Inspired by the successful progress on these demonstration projects, we propose to adapt and extend the statistical tools we are developing to address various managerial concerns. In particular, the objectives of proposed research are to: * Develop and validate a flexible and statistically rigorous framework to monitor structural and functional characteristics of (critical components within) infrastructure * Exploit novel applications that allow for (indirect) inferences of structural integrity (e.g. estimate settlement conditions from energy use during bridge lifting) * Incorporate and evaluate the interventional effect of dynamic exogenous factors (e.g. traffic loading) on both the transit and long-term structural behavior of infrastructure With the purpose of preserving and improving transportation infrastructure systems, the process of infrastructure asset management requires evaluating the effect of managerial decisions on the performance of these systems. This involves measuring and assessing the condition of the system structures as well as generating reliable forecasts. Nowadays, much of the U.S. transportation infrastructure is mature and portions are nearing the end of their service lives and need to be replaced. Hence the aforementioned managerial decisions are increasingly important due to both the far-reaching and serious negative impacts of the deficient infrastructure, as well as the scale of expenditures. The proposed work is motivated in this scenario and involves the following aspects in accordance with USDOT's research goals. First, the health of transportation infrastructure eliminates property damages and, more importantly, reduces transit-related accidents and casualties. Second, it develops and implements cutting-edge and transformative research tool in support of transportation decisions. Finally, it serves necessarily as a complementary work to ITI's expertise in developing advanced remote monitoring technologies.