چكيده به لاتين
The rail infrastructure manager just like any manager, has to efficiently make use of all the sources available for him, however has to deal with a complex situation in terms of high and difficult to control cost structures. It is necessary to find the optimal balance between reducing these costs on the one hand and to avoid break downs that may harm performance and safety on the other hand.
In order to do this, the infrastructure manager has to review current practices to find ways to keep and even enhance his performance level by setting new strategies mainly for track maintenance activities. If he can achieve "quality" for a new track and to maintain it with the optimum maintenance strategies he will be able to manage a track with low costs and a high efficiency.
Quality like any other engineering concepts should be quantitative to make it possible for the infrastructure manager to evaluate or compare in different conditions. The best way to describe the condition of the superstructure is using track quality indices. Although different directives recommend different indices, infrastructure manager has difficulties in choosing the best track quality index. Moreover, using different track quality indices result in different description of track quality and condition.
Despite the enormous influence that the current assessment of track quality has on sustainable development, no known quality index of easy construction and applicable to the majority of situations has been published. In addition, track quality is sometimes difficult to evaluate from a large number of sampling points.
This thesis propose a stochastic track quality index that takes into account the uncertainty about the quality classification still remaining after the data have been observed. In order to do that we set the problem in to a probability index by choosing Bayesian data analysis. So, Bayesian analysis is introduced and its statistical analysis has been developed.
Finally, the index proposed here is applied to data observed from Iran's Railway Network and its stochastic quality classification is obtained and compared with other quality indices.
Keywords: Track quality index, Track geometric parameters, Bayesian analysis, Probability distributions, Track recording car
