Instant ballast-condition analysis

Instant ballast-condition analysis

Visualization of measured raw data: the vertical force (a) and the penetration depth curves of the tamping tine (b) are shown. The penetration speed can be calculated at any time from the distance-time diagram (b).

Comparison of the ÖBB expert assessments with the condition classes of the ballast coefficient. The generally higher ratings of the condition classes can be explained by the reduction in the number of classes to four, in contrast to ÖBB’s five-level system. Sample 7 shows a deviation as an outlier, which is probably due to methodological differences in sampling.

Particle size distribution of ballast samples, shown for clean ballast (sample 2) and heavily contaminated ballast (sample 6). On the right: the weighted sum values are calculated by adding the screen passages for the three smallest opening widths. These total sum values vary greatly depending on the condition of the ballast. Both samples were correctly classified based on the ballast index: sample 2 in class I and sample 6 in class IV.

Precise condition assessment on first contact with the tamping tines

Selecting optimum tamping parameters is closely linked to the condition and the quality of the ballast. Thanks to a new system, it is now possible to determine its condition during the penetration process. This is achieved by measuring specific soil characteristics and subsequently calculation of a ballast coefficient. The application of this technology was analysed in detail as part of a collaborative research project.

The existing methods for determining soil characteristics emphasize the need for a value that can be quickly determined and continuously measured on site. Although laboratory tests provide a detailed picture of the ballast condition, they are labour-intensive,time-consuming and offer only selective data. In addition, they require skilled staff and the results can’t be used directly to determine operating parameters. The logical choice is therefore to use a tamping machine and, in particular, tamping tines as a measuring instrument. This enables seamless recording and measuring of values required immediately with clear spatial and temporatal classification. This information is highly relevant for both the machine operating company and the infrastructure manager. The newly developed, customized ballast coefficient provides an immediate overview of the current ballast condition using a single value and thus supports continuous monitoring and maintenance of the track system.

For this purpose, the penetration speed and vertical force are precisely recorded using a tried- and-tested sensor set-up on the tamping unit. The underlying evaluation algorithms are designed to analyse all relevant variables in real time during operation. As the vertical force and speed are measured during penetration, it is possible to determine the ballast condition at an early stage. These findings form the basis for automatically adjusting the filling and compaction parameters of each individual sleeper, which ensures a consistently high quality of the tamping process.

βBallast ‒ the ballast coefficient

A comprehensive data set was available from previous research projects with a Unimat 09-4x4/4S E³, whose tamping units were equipped with a additional sensors to measure interaction of the tamping uinit with the ballast bed. With aan abundance of measuring data available, the challenge was to extract the data that most accurately depicts the ballast condition and, at the same time, can be used efficiently. However, a detailed analysis of this data set showed that the machine settings had a significant impact on the measurement results. In particular, it was found that the penetration speed set by the operator has a decisive influence on the force measured during penetration. It was possible to establish a clear, direct correlation between the penetration speed and the vertical force, which led to the definition of the βBallast ballast coefficient. Additional tests confirmed that this coefficient shows a significant correlation with the actual condition of the ballast.

βBallast=Fmaxvmax

Maximum vertical force (F_max) during penetration in kilonewtons (kN) and maximum measured penetration speed (v_max) in metres per second (m/s).

This equation takes into account the fact that the lowering speed of the tamping units can be adjusted by the operator and therefore influences the vertical force. The penetration force thus varies with the speed set by the operator. Despite these variations, the ballast coefficientprovides reliable indicators of the ballast condition. . The above mentioned coefficient must therefore be calibrated in such a way that it relates the penetration force to the lowering speed. This aims to create a precise metric that considers both the penetration process conditions as well as the state of ballast. Long term practical experience show that ballast with a higher degree of contamination makes penetration of the tamping tines more difficult, which leads to an increase in the coefficient. This correlation between penetration force and lowering speed must be taken into account accordingly in the analysis metric of the ballast condition.

Ballast condition classes

Based on the statistical distribution of the data and for several practical reasons, the ballast coefficient values were quantitatively classified into four preliminary condition classes. This systematic classification was chosen due to its practicability and the possibility of standardizing the comparison process.

Ballast condition classβBallast in kNs/m
Class I (best condition)βBallast ≤ 35
Class II35 < βBallast ≤ 45
Class III45 < βBallast ≤ 60
Class IV (worst condition)βBallast > 60

How meaningful is the βBallast ballast coefficient?

To validate the condition classes, ÖBB took systematic ballast samples, the results from which were compared with the classifications determined from measuring data. The analysis showed that there was an excellent match between the expert-based assessments and the automated, data-enabled classification.

In order to be able to objectively compare the ballast coefficient with the actual condition of the ballast, a single parameter is also required, which on the one hand quantifies the ballast condition and, on the other hand, can be derived from the analysis data of the ballast samples taken.

The traditional indicators for assessing the ballast condition are proving to be inadequate for these purposes. Therefore, an alternative approach was developed: in the laboratory, the mass percentages of the screen passages for the opening widths 22.4 mm, 31.5 mm, and 40 mm were added together. This method corresponds to a weighted sum and favours smaller ballast grain sizes, as they are included in the results of the larger screen openings. The cumulative values of this summation vary as a function of the condition of the ballast and thus represent a meaningful metric.

When analysing the relationship between the ballast coefficient β and the weighted sum values of the particle size distribution, a clear correlation was found. This reinforces the assumption that the ballast coefficient is an effective tool for determining the condition of ballast.

Future research should aim to further improve the accuracy of the ballast coefficient and eliminate any remaining influences resulting from varying machine configurations.

Further information

Show all articles