Fire Resistance Analysis Through Synthetic Fire Tests
No Thumbnail Available
Date
2023
Authors
Journal Title
Journal ISSN
Volume Title
Abstract
Fire resistance analysis is a complex procedure. In this pursuit, engineers design experiments. However, fire tests are expensive and complex and require specialized equipment that is not accessible to many engineers. This further constrains the ability to test and advance fire research. In order to overcome the above challenges, this paper adopts novel machine learning to generate synthetic fire test data via Generative Adversarial Networks (GANs) from real fire tests to expand our knowledge database. Thus, with the addition of new tests, engineers will have access to a much larger pool of data that can help us to better analyze and design structures for fire. In addition, the availability of more data allows us to seriously integrate machine learning into the fire domain. Thus, this paper presents a new approach to expanding fire test data and applying regression and classification machine learning to predict the fire response of reinforced concrete columns. GANs provide an efficient way to generate synthetic data from real fire tests. Moreover, new data additions contribute to improving predictions of classification-based machine learning in comparison with regression-based machine learning. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.
Description
Keywords
Complex networks , Engineers , Fire resistance , Fires , Regression analysis , Reinforced concrete , Complex procedure , Concrete , Engineers design , Fire resistance analysis , Fire test datum , Fire tests , Machine-learning , Real fire , Regression , Synthetic data , Generative adversarial networks