Multi-modal model updating of miter gates on navigational locks

Navigational locks and dams enable barge traffic, which is critical to the U.S. economy. Unscheduled lock outages are often caused by miter gate failures, which can be prevented with up-to-date structural health information. While strain gages have been deployed on several miter gates for monitoring, intuitively interpreting structural health information from the measurements is difficult. Consequently, some researchers proposed using strain measurements to update a numerical model of the miter gate and employed this updated model to estimate the structural state of the miter gate. However, strain gages alone are often inadequate for effective model updating because strain gages only provide information in the vicinity of the gage; thus, effective model updating would require a dense installation of strain gages, which is prohibitively expensive, both in installation and maintenance costs. Recently, researchers proposed computer vision techniques to collect additional information for model updating. Particularly, an optical flow technique using a sequence of camera images of the miter gate was employed to produce vision-based global displacements for broad regions of the structure. In contrast to a strain gage system, installation and maintenance of a camera system has potentially low costs. While promising, researchers found that the motion of the camera, often experienced during field measurements, limited the accuracy of measurements. Moreover, they did not take advantage of the existing measured strain data. This research proposes a multi-modal model updating strategy that handles camera ego-motion through measurement parameterization. The measured strains are combined with vision-based displacements to provide comprehensive coverage of the structure. The proposed approach is then applied to the miter gate at The Dalles Lock and Dam in Oregon. The results demonstrate the proposed multi-modal model updating strategy provides an effective tool for monitoring the state of miter gates and has the potential to significantly improve the management of navigational infrastructure.