دانلود رایگان مقاله انگلیسی ارزیابی قابلیت اطمینان مدل Eurocode برای سنجش خستگی پل های فولادی - نشریه الزویر

ABSTRACT

In European countries, the design of bridges is conducted following the specifications in the Eurocodes. For verification against fatigue of steel bridges, a simplified model is suggested based on a single vehicle load model together with k factors to estimate a representative stress range. Since the release of the Eurocodes the accuracy of this format has been discussed and questioned. In the current paper, a probabilistic model for fatigue assessment is suggested estimating the load effect from bridge weigh-in-motion (BWIM) measurements. The probabilistic model has been used to estimate the reliability reached with the existing verification format for road bridges. The result shows a large scatter depending foremost on the bridge geometry and the traffic volume. A tentative calibration of the verification format has been performed and new functions for two of the k factors have been derived. With these new functions a significant improvement in the consistency of the reliability level has been achieved. The study demonstrates the need for a more extensive calibration of the Eurocode model and indicates the parameters to focus on.

7. Conclusions

With the purpose to validate the Eurocode model for fatigue assessment of road bridges in steel, a probabilistic model has been suggested. It is based on vehicle loads determined by BWIM measurements. The model has been used to estimate the reliability achieved using the existing verification format by analyses of some simple bridge models. By an iterative process, the Eurocode verification format has been calibrated to render a more consistent reliability level. Considering the results, the following conclusions can be drawn:

• The estimated reliability indices for the existing verification format in the Eurocode EN 1993-2 show a large scatter for the investigated cases. The values range from b = 0.86 to b = 4.9. In general, higher reliability indices were reached for midspan sections in comparison to support sections.

• The scatter in the reliability indices is caused mainly by differences in the critical length of the influence lines and different values for Nobs, the number of heavy vehicles per slow lane per year.

• A tentative calibration of the verification format has been performed by deriving new functions for the factors k1 and k2 by statistical curve fittings.

• The calibration against a target reliability of b = 3.1, for a reference period of 100 years, rendered a significant improvement in the consistency of the reliability level, see Fig. 13.