Vibrodiagnostics of rod structures using neural network methods of machine learning

Abstract

Vibrodiagnostics of rod structures using neural network methods of machine learning

Kadomtsev M.I., Razveeva I.F

Incoming article date: 14.01.2020

This article explores the use of machine learning methods in vibrodiagnostics. To assess the effectiveness of the use of a neural network, a hypothesis was put forward on the possibility of revealing hidden patterns indicating a defect for structures of the same type with different parameters. The selection of these features by visual analysis of the graphs is difficult due to the large amount of data, which indicates the relevance of solving the classification problem using a neural network. As a result of the study, a fully-connected two-layer neural network was obtained, showing sufficient accuracy of predictions, which confirms the fundamental possibility of using machine learning methods to monitor the state of standard structures.

Keywords: vibration diagnostics, building construction, building, structure, defect, damage, machine learning, neural network