Methodology for analysis of traffic-induced building vibrations

Sammanfattning: Urban population growth leads to denser cities where buildings are being constructed closer to sources of vibration like motorways, railways and tramways. The risk of disturbing vibrations is thereby increasing which could affect residents and sensitive equipment, for instance equipment used in hospitals. In the Master’s thesis, the effect of structural design on traffic-induced building vibrations has been studied. A methodology was developed and numerical simulations were carried out using the finite element (FE) method. The methodology holds two parts, one part where a large FE model of the ground is reduced and one part where a parameter study is conducted. In the reduction part, the FE ground model is established and a dynamic condensation is performed which results in a reduced ground model. In the parameter study part of the developed methodology, a FE building model is created and a dynamic stiffness matrix for each studied frequency is determined. The dynamic stiffness matrix of the building model is assembled with the dynamic stiffness matrix of the reduced FE ground model and analyses are conducted. This allows for changes to be made in the building model and new analyses to be performed without the need to implement the large non-reduced ground model. The computational time to perform the parameter study was thereby decreased by 99.7% compared to using the non-reduced FE ground model. The parameter study was performed by steady-state analyses in the frequencies interval of 5–50 Hz with 1 Hz steps. A unit load was applied at a distance of 20 m from a reference building. The results from the parameter study showed that structural design can influence the response of the building significantly. The eigenfrequencies of the building are of importance and a coinciding frequency between the load and the eigenfrequency of the building resulted in a peak response of the vibration level inside the building. The work made in the thesis contributes in the ambition towards enabling predictions of vibrations by use of numerical models. As the developed methodology in the thesis can make numerical simulations more efficient in a way that less computational time and less computational resources are needed.