Browsing by Author "Parisi F"

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    On the use of mechanics-informed models to structural engineering systems: Application of graph neural networks for structural analysis
    (Elsevier B.V., 2023-12-14) Parisi F; Ruggieri S; Lovreglio R; Fanti MP; Uva G
    This paper investigates the application of mechanics-informed artificial intelligence to civil structural systems. Structural analysis is a traditional practice that involves engineers to solve different real-life problems. Several approaches can be used for this task, going from “by hand” computation to the recent advanced finite element method. However, when structures become complex, the success of the analysis can be complicated, often requiring high computational efforts and time. To tackle this challenge, traditional high-demanding methods can be supported by new technologies, such as machine-learning tools. This new paradigm aims to solve structural problems by defining the desired output after directly elaborating input data. One of the current limitations is that often the physics behind the problem is ignored. To solve this issue, resolution models can combine empirical data and available mechanics prior knowledge to improve the predictive performance involving physical mechanisms. In this paper, a method to develop a Mechanics-Informed Surrogate Model (MISM) on structural systems is proposed, for which input structured data are used to enrich the informative content of mechanics systems. Then, Graph Neural Networks (GNNs) are explored, as a method capable of properly representing and embedding knowledge about a structural system, such as truss structures. The main advantage of the proposed approach is to provide an alternative way to the usual black-box machine-learning-based models. In fact, in the proposed MISM, the mechanics of the structural system plays the key role in the surrogate model definition, in order to obtain physically based outputs for the investigated problem. For the case at hand, MISMs are developed and employed to learn the deformations map of the system, starting from the knowledge of the structural features. The proposed approach is applied to bi-dimensional and tri-dimensional truss structures and the results indicate that the proposed solution performs better than standard surrogate models.
  • Thumbnail Image
    Item
    Virtual reality for safety training: A systematic literature review and meta-analysis
    (Elsevier B.V., 2023-11-18) Scorgie D; Feng Z; Paes D; Parisi F; Yiu TW; Lovreglio R
    Unsafe behaviour in the workplace and disaster events can lead to serious harm and damage. Safety training has been a widely studied topic over the past two decades. Its primary aim is to save lives and minimise damage but requires regular refreshers. New digital technologies are helping in the process of enhancing safety training for better knowledge acquisition and retention. Among them, Virtual Reality (VR) can provide an engaging and exciting training experience, and there is a need to evaluate its application and effectiveness in safety training. This study aims to investigate VR safety training solutions applied to various industries (excluding medical and military applications), such as construction, fire, aviation, and mining. This was achieved by systematically reviewing 52 articles published between 2013 and 2021 to answer nine research questions. Fourteen domains were examined, with construction and fire safety training being the most prevalent since 2018. Findings reveal that only a small percentage (9.6 %) of the studies explicitly adopted theories while developing and testing VR applications. Additionally, this review highlights a critical need for long-term retention measurements, as only 36 % of studies provided such data. Finally, the two meta-analyses proposed in this work demonstrate that VR safety training outperforms traditional training in terms of knowledge acquisition and retention.
  • Thumbnail Image
    Item
    What do we head for while exiting a room? a novel parametric distance map for pedestrian dynamic simulations
    (Elsevier B.V., 2023-09-21) Parisi F; Feliciani C; Lovreglio R
    Identifying effective strategies describing crowd dynamics is crucial to enhance simulations of pedestrians for crowded event planning and management. Various modelling solutions have been proposed to describe how people try to exit from a built environment in normal and emergency. Several of these solutions rely on the use of distance maps or floor fields to account for the positions of existing goals and the location of obstacles to avoid. To date, distance maps are assumed to be static (they do not vary over time) and that pedestrians aim at the actual central coordinate of a door. In this work, we challenge the static goal assumption by proposing a novel parametric distance map which is variable depending on the polar coordinates defining the position of a pedestrian having the centre of an exit as the origin (i.e., the distance of the pedestrian and an angle between its direction and the perpendicular to the exit). In this work, we investigate what pedestrians head for while trying to reach an exit. Different parametric solutions are proposed and calibrated using likelihood-based optimisation methods with over 9000 trajectories of individual pedestrians who navigated through an indoor university atrium building to reach several exits. The results highlight good performance for this modelling approach: pedestrians head for targets in front of an exit when they are away from it, and their targets shift behind the exit as they get closer to it, (i.e., distance impact) while their angle does not have impact on this process. The proposed dynamic goal-based distance map can be applied for future pedestrian simulations for crowded event planning and management.