Open Access Open Access  Restricted Access Subscription or Fee Access

Vehicle-Bourne Autonomous Railroad Bridge Impairment Detection Systems



Timber bridges constitute approximately 30% of current railroad bridge inventories in North America. Inspections of these structures involve visual assessments of the observable condition of individual components that comprise the bridge system. Special inspections may call for field personnel to observe a bridge under load: that is, while a train is crossing. Individually monitoring large numbers of timber bridges can be resource intensive, and the use of electro-mechanical sensors such as strain gages, displacement transducers, load cells, accelerometers, and related instrumentation is limited to rare exceptions to enhance visual inspections; it is done “as deemed necessary” on a “bridge-by-bridge” basis. The fitness of a timber bridge strongly correlates to its structural response under load. Specifically, strength deficiencies and other impairments in a timber bridge often result in degradation in member and connection stiffness, and such degradation can be detected under laboratory conditions. One specific impairment scenario common to timber bridge systems is timber stringer fatigue degradation. Models of timber bridges subject to varying levels of impairment are presented. Preliminary quasi-static MATLAB models, SAP2000 models, and refined LS-DYNA models were developed to investigate the effects on structural behavior resulting from typical timber bridge impairments such as stringer fatigue. Analytical results indicate significantly increased deflections for impaired stringers. For example, impaired bridges with an effective bending stiffness approximately 1/3 that of unimpaired bridges may experience nearly 100% increase in deflection. This paper presents a Vehicle-Bourne Autonomous Impairment Detection System that will facilitate detecting structural impairments in timber railroad bridges using data gathered from rail vehicles that cross the bridges. Deflection profiles are related to wheel path accelerations which serve as the primary diagnostic data stream with which to evaluate a timber bridge. Accelerations can be measured via instrumented rolling stock, and the data streams can be processed by a Structural Impairment Detection System (SIDS) comprising arrays of competitive artificial neural networks.

Full Text: