Abstract: Because of uncertain thermal parameters, the stochastic temperature field of frozen soil roadbed can be correctly analyzed using the stochastic finite element method. Statistical characteristics of geotechnial thermal parameters were analyzed along the Qinghai-Tibet railway, and then random temperature fields of a typical frozen soil roadbed were calculated by the finite element method based on Monte-Carlo method. The results show that the distribution patterns of average temperature fields are similar to that of the standard deviation, and the bigger values are on the scope of top and slope surface of roadbed. The standard deviation, which is the biggest in the boundary, becomes smaller from the boundary to interior, and comes to more uniform at some depth. The standard deviations increase and its distribution scopes enlarge with the time.

Key words: stochastic temperature field, frozen soil roadbed, Monte-Carlo method, Qinghai-Tibet railway