Current maintenance intervals of mechanical systems are scheduled a priori based on the life of the system, resulting in expensive maintenance scheduling, and often undermining the safety of passengers. Going forward, the actual usage of a vehicle will be used to predict stresses in its structure, and therefore, to define a specific maintenance scheduling. Machine learning (ML) algorithms can be used to map a reduced set of data coming from real-time measurements of a structure into a detailed/high-fidelity finite element analysis (FEA) model of the same system. As a result, the FEA-based ML approach will directly estimate the stress distribution over the entire system during operations, thus improving the ability to define ad-hoc, safe, and efficient maintenance procedures. The paper initially presents a review of the current state-of-the-art of ML methods applied to finite elements. A surrogate finite element approach based on ML algorithms is also proposed to estimate the time-varying response of a one-dimensional beam. Several ML regression models, such as decision trees and artificial neural networks, have been developed, and their performance is compared for direct estimation of the stress distribution over a beam structure. The surrogate finite element models based on ML algorithms are able to estimate the response of the beam accurately, with artificial neural networks providing more accurate results.

In Spring 2020, the College of Engineering at San José State University (SJSU) conducted a comprehensive analysis of the impact of COVID-19 on faculty who were forced to transition to an online learning environment. The purpose of this study is to assess the impact of COVID-19 on faculty teaching methods, assessment methods, and personal well-being. The study was a combination of a quantitative survey and a qualitative study using interviews of engineering faculty teaching in Spring 2020. In the first part, we surveyed all faculty teaching during Spring 2020 in the SJSU College of Engineering about their experiences after the move to 100% online instruction in March 2020. In the second part of the research, we interviewed 23 faculty members to obtain a more in-depth understanding of their experiences during the move online in Spring 2020. Overall, 98 faculty participated in the survey: lecturers (58), tenure-track (18), tenured (13), adjunct (1), and Teaching Associates (1). The faculty reported being worried about their family and their students’ well-being. In addition, 65% of faculty members reported either a moderate or a great deal of stress related to the shelter in place, and this percentage was higher for female faculty (74%) and for tenure-track faculty (83%). Overall, faculty members felt that they had their classes under control most of the time and that the transition to online teaching was positive, even if they felt they had too much work to do and felt always in a hurry and under pressure. From a teaching perspective, the interviews highlight that faculty members’ main concerns focus on testing and assessment and students’ engagement. Overall, SJSU College of Engineering faculty members felt under stress in the transition to online teaching, especially the tenure-track faculty members, but were able to transition their classes with ease.

Modal testing is a common step in aerostructure design, serving to validate the predicted natural frequencies and mode shapes obtained through computational methods. The strategic placement of sensors during testing is crucial for accurately measuring the intended natural frequencies. However, conventional methodologies for sensor placement are often time-consuming and involve iterative processes. This study explores the potential of machine learning techniques to enhance sensor selection methodologies. Three machine learning-based approaches are introduced and assessed, and their efficiencies are compared with established techniques. The evaluation of these methodologies is conducted using a numerical model of a beam to simulate real-world scenarios. The results offer insights into the efficacy of machine learning in optimizing sensor placement, presenting an innovative perspective on enhancing the efficiency and precision of modal testing procedures in aerostructure design.

Experimental monitoring of dynamic response is generally limited to few locations in the system. However, the analysis of structural performance and design of control systems would benefit from a complete knowledge of the system dynamic during service. A numerical approach is developed to numerically reconstruct the load and response of a complete structure from few reference points, based on a modal approach for projecting the response at few points on the domain of the structure. This methodology is particularly advantageous when full-field monitoring of a structure is not a possible solution. An assembly of two beams joined by a nonlinear torsional spring is analyzed in case of different load distributions acting on its span. The approach is shown to be robust and reliable.

In the monitoring of structural systems, the use of multiple high-end sensors may prove to be economically prohibitive. The alternative approach would be to use fewer devices capable of moving across the span of the structural system. In the proposed approach, a velocity sensor that is able to move across the spatial domain and obtain point-wise velocity measurements is combined to a novel dynamic observer. Based on the measured velocities, a state estimator is developed, the gain of which depends on the motion of the sensor. The motion of the sensor is defined using Lyapunov redesign methods and depends only on the estimation error at the current sensor position. The guidance policy is performance based and steers the sensor to spatial regions of the structure with larger estimation errors. The proposed approach is validated with a one-dimensional flexible structure, mathematically described by an Euler–Bernoulli partial differential equation. The moving sensor is realized through the use of a laser scanning vibrometer that provides both the moving measurements and additional measurements against which the proposed approach will be validated. Once measurements over a large number of locations are acquired, the experimental results are fed to the algorithm that selects the instantaneous sensor location. Experimental results for linear and nonlinear beam cases are presented to show the feasibility and robustness of the proposed approach.

Modal testing is a common step in aerostructure design, serving to validate the predicted natural frequencies and mode shapes obtained through computational methods. The strategic placement of sensors during testing is crucial to accurately measuring the intended natural frequencies. However, conventional methodologies for sensor placement are often time-consuming and involve iterative processes. This study explores the potential of machine learning techniques to enhance sensor selection methodologies. Three machine learning-based approaches are introduced and assessed, comparing their efficiency with established techniques. The evaluation of these methodologies is conducted using a numerical model of a beam to simulate real-world scenarios. The results offer insights into the efficacy of machine learning in optimizing sensor placement, presenting an innovative perspective on enhancing the efficiency and precision of modal testing procedures in aerostructure design.