Prediction of Blast-Induced Structural Response and Associated Damage Using Machine Learning

Terrorist bombing-induced casualties are not only related to immediate fatalities but also to structural deterioration, damage, or even collapse that might occur and may lead to tremendous loss of life. Efficient assessment of blast-induced structural damage following explosion events is becoming a growing problem in modern societies. An attempt based on machine learning is made in this study to anticipate structures’ responses and the associated structural damage to reinforced concrete (RC) buildings exposed to extremely short_duration explosive loads. A program is developed to generate a set of analytically derived data for nonlinear building models subjected to explosive loads. Common machine learning models and Python libraries were utilized during the development of our program implementation to learn from a dataset. The latter has different features or input parameters, such as the amount of explosive charge, the distance from the building, fundamental period, and the building’s mass and rigidity, as well as the soil type. Our database is thus used, along with our regression-and-classification based implementations, to generate an output index that estimates and categorizes the state of damage based on the several most-important parameters of the explosion exposure. In the input database, the state of damage, based on the values of captured damage indices, is classified into one of four cases. Our code efficiently predicts those cases using a model that learns from the database. The prediction rates of the presented model reach an overall high accuracy. Therefore, the proposed model provides an accurate prediction of the level of structural damage by using the computed damage indices.

Cette oeuvre a été publiée sous la license Creative Commons Attribution 4.0 (CC-BY 4.0). Il est autorisé de partager et adapter l'oeuvre tant que l'auteur est crédité et la license est indiquée (avec le lien ci-dessus). Vous devez aussi indiquer si des changements on été fait vis-à-vis de l'original.