The Origins of Measurement Uncertainty in SHM - NPL Footbridge Case Study

This paper explores the effects of measurement uncertainty in structural health monitoring (SHM) applications. Uncertainty quantification is an essential part of effective design of next generation SHM systems for smart asset monitoring. Conventional measurement uncertainty analysis which is based on the GUM (Guide to the expression of uncertainty in measurement [1]) cannot straightforwardly be applied to long-duration time series data such as arise in SHM applications. We therefore discuss some alternative approaches and make recommendations for best practice in interpreting SHM datasets. This work is based on experimental data from a well-established monitoring system installed on a concrete reinforced footbridge at the UK's National Physical Laboratory (NPL). Some challenges in the development of a methodology for quantifying measurement uncertainty for civil engineering applications at sensor level and system level are described. The footbridge dates from 1960 but is no longer in use and has undergone deliberate damage and repair cycles over the period of two years, 2010-2011. The data obtained during different stages of progressively increasing damage give a unique opportunity to explore the question of the minimum level of damage that can be reliably detected for a specified degree of accuracy. Our investigation is at the early stages. Therefore only initial findings will be presented. INTRODUCTION: OVERVIEW OF EXPERIMENTAL WORK At NPL we undertook an experimental study by converting a reinforced concrete footbridge to a structural health monitoring demonstrator to provide an independent assessment of sensor capabilities and sensor performance for short and long-term monitoring. The extensive experimental programme was conducted over three years from 2009 to 2011 to investigate the reliability of methods for damage assessment. _____________ E. Barton, T. Esward National Physical Laboratory, Hampton Road, Teddington, TW11 0LW. 6th European Workshop on Structural Health Monitoring We.4.E.2 License: http://creativecommons.org/licenses/by/3.0/ 1 During the first year of this project a 1960’s footbridge was used to create a fullscale test bed designed to investigate the capability of sensing systems for long-term monitoring in outdoor conditions. Sensors necessary for monitoring large civil engineering structures were identified and installed in locations based on the results of modeling and surveying. The performance of wireless and wired sensor and acquisitions systems was evaluated during static loading of the bridge similar to the loading tests used for bridge assessment. The results showed that the sensors were capable of real-time monitoring of the conditions of the structure and have been presented previously [2]. An extensive experimental program of damage / repair cycles was conducted during a second and third year. After detailed discussions with project partners including experts from Sinclair Knight Merz (SKM) and the UK’s Highways Agency the various types of damage of civil engineering structures were narrowed down to two key areas: deterioration of reinforced concrete and integrity of repair / strengthening patches. The first stage involved removal of the concrete followed by repairs using the best industrial practices and concrete repair materials. Damage locations and extent of removal reasonably simulated spalling of concrete owing to reinforcement corrosion the deterioration of concrete structures, as is observed in the field. Embedded and surface sensors gave a unique insight in to the load distribution though the cross section of structural members. The second stage started in September 2011 with the cantilever structurally weakened by cutting through the reinforcement, and then strengthened using 4m long carbon fibre reinforced polymer (CFRP) panels installed on the top deck. Various design assumptions are made for current repair and strengthening methods and field data on this topic are scarce. Figure 1a shows the 4m carbon fibre repair panels on top of the cantilever installed to the specifications of the Highway Agency by Concrete Repairs Ltd, a company with 60 years experience in repairs. Fatigue tests used a unique dynamic testing facility, specially designed for this demonstrator by adapting an Instron hydraulic system. a b Figure 1a. Carbon fiber repair patches on top of the cantilever. Figure 1b. Static test set up. The static and forced vibration tests were performed regularly to assess the structure and provide useful full-scale experimental data on the design and performance of concrete repair and CFRP strengthening methods. The test set up is