Page 272 - 360.revista de Alta Velocidad - Nº 5

Álvarez, Fernando. Balmaseda, Lucía. Gallego, Inmaculada. Rivas , Ana. Sánchez-Cambronero, San

With all these aspects, the models under study have been elaborated according to the following
criteria:

1. All models will maintain the constitutive model of elastic behaviour for the Rail, Bearing
plates and Sleepers due to their high stiffness.

2. The comparison between elastic and elastoplastic behaviour models will only refer to
granular materials of the railway section.
3. The sensitivity analysis will be performed by varying the value of the internal friction angle
corresponding to the Subballast and the Formation layer.

Table 2 shows, in summary, all the models or cases of analysis to be analysed, according to the
behaviour of the granular materials, the existence or not of confining ballast and the episodes
of load to be considered.

Table 2. Summary of characteristics of the models to be studied.

Constitutive
Model Confining ballast Load step
model-Infrastructure later
#1 Elastic Yes LS2
#2 Elastic No LS2
#3 Elastoplastic (Φ , Φ ) Yes LS1+LS2+LS3
1 1
#4 Elastoplastic (Φ ,Φ ) No LS1+LS2+LS3
1 1
#5 Elastoplastic (Φ , Φ ) Yes LS1+LS2+LS3
2 1
#6 Elastoplastic (Φ , Φ ) No LS1+LS2+LS3
2 1
#7 Elastoplastic (Φ , Φ ) No LS1+LS2+LS3
1 2
#8 Elastoplastic (Φ , Φ ) No LS1+LS2+LS3
2 2
4. Results

4.1 Adjustment model

During the resolution of the numerical model that considered the linear elastic behaviour of
all the materials of the platform, it was detected that the vertical compression stresses in the
ballast under the loaded sleeper reach values well below the usual ones of 100 to 120 kPa,
obtained in experimental observations performed for High-Speed sections in Spain (Gallego, et.
al., 2013). This could be because, considering the ballast as a linear elastic material, makes the
ballast elements located above the support plane of the crossbeam oppose the compressions
that occur in said plane, generating tensions that They decrease the value of the compressions
that are given in the sleeper-ballast interface. To solve the problem we defined the following
strategies to consider for the numerical model:

1. Eliminate the ballast that confines the sleepers in the longitudinal and transverse direction
of the track, see Fig. 2.
2. Assume in the ballast a law of elastoplastic behaviour, without cohesion, to avoid the
appearance of tractions in the same.

270 360.revista de alta velocidad