**2. Methodology**

A Standard Element describes track behavior in its specific maintenance demands and service lives—based on different sets of parameters, different investment options, and its associated maintenance demand. The Standard Element approach was developed for cyclic track investment and maintenance but also can be used for permanent component exchange. Among the different options the most sustainable needs to be identified, finally in its economic efficiency.

One Standard Element describes the track behavior of one specific set of parameters in its investment, maintenance demand, and service life. All parameters

### *Assessing Average Maintenance Frequencies and Service Lives of Railway Tracks:… DOI: http://dx.doi.org/10.5772/intechopen.110488*

influencing track behavior in a relevant way, either the maintenance demand or the service life—or both, needs to be looked at. The relevant parameters must be identified. The entire network is clustered according to these parameters. The aim is to describe common situations precisely (Standard Element Approach). Individual cases are not pursued further within the framework of this method.

Every Standard Element is linked to a working cycle (**Figure 1**) consisting of the maintenance frequencies over the service life. Evaluating general track strategies demands analyzing different options of maintenance like minimized/reactive maintenance or preventive maintenance and their consequences on the service life. In the end, the working cycle linked to the Standard Element depicts the economically most sustainable option, a sufficient, mostly preventive maintenance, and the optimal service life.

The year zero shows the investment, followed by all planned maintenance actions including one line for small maintenance, which sums up all maintenance actions which are just reactive (e.g., rail breakages).

The core part of the Standard Element is the Working Cycle, describing maintenance demand and service life and thus track behavior for the given parameter set. This should be based on data and experience. It is generally carried out in working groups. Their members are decision-makers from the headquarter, technicians, and economists. Their knowledge is mainly based on a big amount of data. However, there is a second source of knowledge, the experiences of track engineers out at the track sections. Their knowledge is based on observation, does not fundamentally contradict data knowledge, but complements it in essential details. These two different sources lead to important discussions necessary for a proper definition of the working cycles.

The Standard Elements should cover the main part of the network. For verification, the number of kilometers of the various Standard Elements in the network must be calculated. Furthermore, the age of the sections is relevant. The verification of the working cycles is possible by comparing real investment and maintenance demand of the network with the demand described within the Standard Elements.

For identifying the most sustainable track solution a cost evaluation of all options is executed. Therefore, Standard Elements are transposed into time sequences of costs, starting in the year zero (year of investment) lasting until the next re-investment. The various options in general show different investment costs, different maintenance demand, and different service lives. Due to the different service lives the net present value cannot be used for ranking the options. The life cycle cost evaluation and thus a ranking of different options must be based on their average annual costs, including costs of fixing money. Within such an evaluation the cheapest option is the most sustainable one, as "cheap" is always taking all life cycle costs into account, from investment to next re-investment. Thus, the sustainable option requires an optimal balance of investment and maintenance and an optimal balance of maintenance and service life.

Based on verified Standard Elements, the most sustainable investment and maintenance options are identified for all relevant track situations in a network. This is also true for innovations, of course after a relatively short testing period.

**Figure 1.** *Standard element including working cycle.*

#### **2.1 Clustering the network**

#### *2.1.1 Parameters*

Within the Standard Element Approach "Parameters" are the boundary conditions that lead to certain track behavior and thus trigger the maintenance demands and service life. While some of those parameters are given, some are topics of track strategies.

The alignment of track is a major aspect when it comes to track maintenance demand. It is obvious that a curved track needs another maintenance regime or simply more maintenance than a straight track section. Consequently, service life is shorter on curved tracks compared to straight tracks.

The transport load is obviously triggering maintenance: the more trains operated, the higher the maintenance demand and the shorter service life.

Next to those two parameters hardly being influenced, the construction of the track itself defines another set of parameters.

The subsoil forms the foundation of a track and thus has a major impact on the amount of maintenance needed. Subsoil with sufficient bearing capability and a properly working water drainage system needs the least maintenance and delivers maximum service life.

The used track components are important parameters. Every single component goes along with either specific wear or damage phenomena or at least with different severances of those. The rail profile defines the durability of the rails. Smaller profiles lead to the necessity of through-going rail exchange in case of high transport volumes within the track's service life. Rail steel grade drives both (side) wear in sharp curves and rolling contact fatigue phenomena in larger curves. The sleeper type, mainly concrete and wooden sleepers, but also steel sleepers in some cases and newly concrete sleepers with under sleeper pads (USP), plays an important role in ballast maintenance: the different types lead to changes in the sleeper-ballast interface, decreasing or increasing the contact pressure and thus ballast bed deterioration and consequently impact tamping needs. Ballast and subsoil quality directly influence not only track service life but also tamping demand. The same is true for the quality of the dewatering system.

#### *2.1.2 Parameter values*

For the parameters concerning superstructure, the values are simply the components to be used, for example, wooden sleepers or concrete sleepers. To keep the number of combinations low, parameter values might be restricted to some combinations. An example would be "49E1 rails on low loaded lines only".

Values for the parameters "Subsoil" and "Drainage" are an issue as values are not available, especially for existing substructures. In case of executed subsoil rehabilitation, Evd values would be possible, but can also not cover the linked aspect of dewatering. Most infrastructure managers go for a "smart" characterization: "good" is derived from a situation in which superstructure quality and its maintenance are not influenced negatively by subsoil or dewatering topics, while "poor" would mean the other extreme.

Traffic load and alignment need clustering of discrete values. For higher traffic loads, a range of some 10 mio. Gross-tons per year (and track) is feasible looking at international experience with the Standard Element Approach. For lower traffic loads, smaller ranges are necessary though.

*Assessing Average Maintenance Frequencies and Service Lives of Railway Tracks:… DOI: http://dx.doi.org/10.5772/intechopen.110488*

The topic of alignment of track is dealt with between the two extrema "straight" (no side wear of rails, rail surface failures, tamping due to restoration of cant, or similar occurs) and the minimum radius for continuous welding of rails (jointed rails lead to different track behavior and maintenance actions anyhow).

The "number of tracks" parameter with its values "single-tracked" and "doubletracked" is not decisive for track behavior in the overwhelming part of effects, but is in case of track work costs due to different logistics and track closure times, respectively operational consequences of such closures.

We summarized the most typical parameter values for mixed traffic networks in **Table 1**.

#### *2.1.3 Parameter sets: The standard element*

A combination of exactly one value for every parameter delivers the description of one technical situation within the network, one Standard Element. On the lift top corner in **Figure 2**, we see the coding according to **Table 1**.

Theoretically, there is an enormous amount of possible Standard Elements combining all parameter values. Practical use shows that some 100 Standard Elements help to cover 90 percent of a network being sufficient for achieving valid and robust values for maintenance and renewal.


#### **Table 1.** *Parameters and parameter values.*

**Figure 2.** *Characteristic of a standard element.*

### **2.2 Working cycles: Input data**

In the next step, it is necessary to attach the average maintenance regime to the Standard Elements. This is done by depicting the frequency of different maintenance works between one re-investment and the next one. As this should be done based on existing data, some aspects need to be considered, which are discussed below.

It might be that


Besides different aspects evaluations of the set of Standard Elements may cover, it can be stated that one Standard Element (set of parameter values) can have different working cycles (maintenance regimes).

The process followed in the latter is based on the idea to depict a sustainable maintenance regime, so neither includes speed restrictions nor insufficient maintenance.

The most challenging task is forming the working cycles for different parameter sets. This is true for depicting the necessary maintenance frequency as well as for the service life. The latter can be approached by using track records. In most cases, infrastructure managers store the "year of track relaying" as status data in their data warehouses. In assessing track age parameter-set-wise, we come up with a survival analysis showing reached service lives so far. This works in case of evenly distributed track ages and/or parameter sets already used longer than one service life (e.g., wooden sleepers). It fails, if parameter sets are relatively new (e.g., new track components or higher transport volumes) as we only see the first part of the service life in the survival analysis. Alternatively (or additionally), we can analyze the reached service lives of tracks at the point in time of relaying. Doing statistics on these data also leads to robust values while delivering the deviations from the mean in addition. **Figure 3** shows such a frequency distribution analysis as an ex post evaluation for wooden sleepers. In this figure, 0 is no deviation from the depicted average service life, all values to the right stand for longer service lives, and all values to the left for shorter ones. **Figure 3** shows additionally the La Place distribution in blue that fits sufficiently good for wooden sleepers.

Generally, maintenance demand increases with increasing track age. When using maintenance records, we thus need to consider track age as an important variable. Taking sections with similar parameters from all over the network, we have executed maintenance frequencies for different time frames in the service live. Averaging these frequencies leads to a first approach of the working cycles. As maintenance is

*Assessing Average Maintenance Frequencies and Service Lives of Railway Tracks:… DOI: http://dx.doi.org/10.5772/intechopen.110488*

**Figure 3.** *Distribution of deviations from the average service life—Wooden sleepers.*

executed based on on-site track quality, these frequencies of course vary from section to section. Depending on the number of sections available, the averages of those frequencies might be more or less robust. Therefore, a consolidation process is necessary. This process is based on the experience of the track engineers. There are some rough plausibility checks that can be executed quite straightforwardly: Does maintenance increase over the service life? Does higher transport volume lead to tighter maintenance frequencies? Does poorer subsoil condition show intensified maintenance? With the knowledge and experience of the trained maintenance staff, we can derive consolidated maintenance frequencies.

In the end, the depicted maintenance frequencies in the working cycles need to deliver the executed maintenance amount of the analyzed network. The same is true for service lives: in calculating the average service life for all existing parameter sets and applying this to the total lengths of tracks, we can easily calculate the average amount of track renewal. If calculated values deviate from the existing maintenance and renewal works, the working cycles need to be adopted in an iterative way. Adding the unit costs for the single maintenance tasks, this plausibility check can also be executed for maintenance budgets.
