A Predicting Fatigue Life of Miter Gate Anchorages with Stochastic Modeling and Limited Sensor Data

Abstract

The US Army Corps of Engineers (USACE) maintains a large inventory of lock and dam structures along inland waterways in the US. Miter gates are key components of lock systems and therefore the ability to assess the fatigue life of miter gates is critical to direct maintenance operations and prevent costly disruptions to commercial traffic. Miter gates are supported by steel anchorages, which transfer the gate overturning reaction to the concrete lock wall. Many anchorages have been in service for 80 to 90 years and may be approaching the end of their fatigue life. However, their condition is unknown because they are embedded in concrete. We present a new method that addresses four major challenges for predicting fatigue life of miter gate anchorages: (1) estimating loads on the anchorage; (2) computing stress ranges in the anchorage based on those loads, accounting for steel-concrete interaction; (3) estimating the number of load cycles on the anchorage; and (4) incorporating uncertainty and the ability to update estimates based on inspection data. The goal of this method is to predict anchorage fatigue life without strain gage data. It is prohibitively expensive to install instrumentation on every anchor in the USACE inventory, and therefore the limited available instrumentation is used to verify the proposed methodology. Using this framework, fatigue life estimates are made using S-N curves. Ultimately, the deterministic fatigue life predictions will be incorporated in a stochastic model that accounts for uncertainties in loading and modeling assumptions.