Estimation of Vibration Frequency of Structural Floors Using Combined Artificial Intelligence and Finite Element Simulation

Abstract

Floor vibration due to human activities (walking, running, etc.) and operating machines generally makes inconveniences for residents. The natural vibration frequency of beams is determined as the main source and also the controlling parameter of such the phenomenon. Many studies have been conducted on determining the natural frequency of beams in recent years; however, the proposed formulations in many of them are not very practical for vibration control of tall building floors. In this paper, the finite element method (FEM), nonlinear (NL) dynamic analyses, and artificial intelligence (AI) techniques were adopted to constitute the simple frequency equations of the fixed ends and cantilever steel beams for controlling floor vibration. The input data required for training the AI based model are simulated in a NL-FE platform considering various cases of the steel moment connections. The analysis outcome of several hundred beams with different properties indicated that the calculated vibration frequency values of the fixed ends and cantilever beams were respectively 2.07 and 0.33 times larger than the frequency value of the simply supported beams with similar conditions. To this, the implemented soft computing technique was determined as an effective approach to improve the computational efficiency of the NL-FE simulations.