ABSTRACT
The dynamics of railway bridges is one of the most complex areas of theoretical analyses and experimental studies – (Frýba 1996, Yang et al. 2004). The results of dynamic analysis of railway structures are considerably uncertain as compared with static analyses, and as a rule require experimental verification. The basic element of these analyses is the determination of the dependency of the maximum amplitudes of displacements or accelerations on the train’s speed. It is important to analyse the structure’s behaviour during the train passages at the speed increasing from the quasi-static one to the maximum one permissible for the tested section.
The article presents examples showing how significant mistakes can be connected with the estimation of the extreme displacement amplitudes in case of measurement results extrapolation of train passages at speeds higher and lower than the registered ones.
The method of obtaining quasi static displacements history by means of filtering (Paultre et al. 1992) was tested.
The extrapolation of values under higher speed was realized with the use of polynomial extrapolation.
The first tested railway viaduct was designed as a steel plate girder with orthotropic deck. The theoretical length of span is 24.00 m. The viaduct is characterized with a high level of forced and free vibration. The damping was relatively weak.
The second tested railway viaduct was designed as a steel free-ends framework with parallel chords. The stiff bottom chords consist of two plate girders with composite reinforced concrete ballast pan. The theoretical length of span is 93.00 m. The railway track is the characteristic feature of the studied viaduct, since the truck is curved over the whole length of span with a radius of R = 2600 m.
The dynamic tests were conducted using a special train consisting of two locomotives and four railway cars placed between them.
Summing up the carried out analyses, we can state here that the estimation of the quasi-static value on the basis of the displacements registered during the train rides at the speeds close to the maximum ones instead of its measurement at the speed of approximately 10 km/h can result in significant mistakes and in most cases leads to significant undervaluation of the estimated dynamic amplification factor.
The extrapolation of the results of the measured displacements’ extreme values to higher (than obtained) speeds of the train rides can also lead to some mistakes. They are, however much smaller than the mistakes connected with the estimation of the quasi-static values on the basis of train rides at the speeds close to the maximum ones.
