Search for: All records

This paper introduces a novel approach to enhance the docking mechanism of sensor packages deployed on bridges using unmanned aerial vehicles (UAVs). The current electropermanent magnet (EPM) system faces challenges in achieving efficient and stable docking due to factors such as airflow, GPS stabilization, and the time required for EPM activation. To address these issues, a biased EPM design is proposed, utilizing additional permanent magnets to achieve neutral buoyancy during UAV deployment. This design optimally balances the weight of the drone and sensor package, providing advantages such as improved stability against external factors and reduced pilot fatigue. Experimental results demonstrate the feasibility of the proposed design, indicating enhanced hold force and an extended range for efficient docking. 

Real-time model updating for active structures experiencing high-rate dynamic events such as; hypersonic vehicles, active blast mitigation, and ballistic packages require that continuous changes in the structure’s state be updated on a timescale of 1 ms or less. This requires the development of real-time model updating techniques capable of tracking the structure’s state. The Local Eigenvalue Modification Procedure (LEMP) is a structural dynamic modification procedure that converts the computationally intensive global eigenvalue problem used in modal analysis into a set of second-order equations that are more readily handled. Implementation of LEMP for tracking a structure’s state results in secular equations that must be solved to obtain the modified eigenvalues of the structure’s state. In this work, the roots of the secular equations are solved iteratively using a divide and conquer approach, leading to faster root convergence. The present study reports on developing a real-time computing module to perform LEMP in the context of real-time model updating with a stringent timing constraint of 1 ms or less. In this preliminary work, LEMP is applied to tracking the condition of a numerical cantilever beam structure, which depicts changes in a structure’s state as a change in the roller position. A discussion of variations in timing results and accuracy are discussed. 

Abstract Structures operating in high-rate dynamic environments, such as hypersonic vehicles, orbital space infrastructure, and blast mitigation systems, require microsecond (μs) decision-making. Advances in real-time sensing, edge-computing, and high-bandwidth computer memory are enabling emerging technologies such as High-rate structural health monitoring (HR-SHM) to become more feasible. Due to the time restrictions such systems operate under, a target of 1 millisecond (ms) from event detection to decision-making is set at the goal to enable HR-SHM. With minimizing latency in mind, a data-driven method that relies on time-series measurements processed in real-time to infer the state of the structure is investigated in this preliminary work. A methodology for deploying LSTM-based state estimators for structures using subsampled time-series vibration data is presented. The proposed estimator is deployed to an embedded real-time device and the achieved accuracy along with system timing are discussed. The proposed approach has shown potential for high-rate state estimation as it provides sufficient accuracy for the considered structure while a time-step of 2.5 ms is achieved. The Contributions of this work are twofold: 1) a framework for deploying LSTM models in real-time for high-rate state estimation, 2) an experimental validation of LSTMs running on a real-time computing system. 

Validation of state observers for high-rate structural health monitoring requires the testing of state observers on a large library of pre-recorded signals, both uni- and multi-variate. However, experimental testing of high-value structures can be cost and time prohibitive. While finite element modeling can generate additional datasets, it lacks the fidelity to reproduce the non-stationarities present in the signal, particularly at the higher end of the digitized signal's frequency band. In this preliminary work, generative adversarial networks are investigated for the synthesis of uni- and multi-variate acceleration signals for an electronics package under shock. Generative adversarial networks are a class of deep learning approach that learns to generate new data that is statistically similar to the original data but not identical and thus augmenting the data diversity and balance. This chapter presents a methodology for synthesizing statistically indistinguishable time-series data for a structure under shock. Results show that generative adversarial networks are capable of producing material reminiscent of that obtained through experimental testing. The generated data is compared statistically to experimental data, and the accuracy, diversity, and limitations of the method are discussed.