Data Analysis

Railway infrastructure assets are marked either as single (e.g. wheels or welds) or linear (e.g. rails, sleepers, ballast). Linear assets are characterised by their length and thus are more complicated to analyse than single assets (long assets can have different conditions and behaviour at different locations).

Therefore, to capture the local differences in the condition of linear assets, they must be analysed in shorter sections that have meaningful and manageable behaviour. Both dynamic segmentation capability and cross-correlation of single and linear assets are functions covered by RAMSYS that allow for this kind of analysis. Typical data-analyses in Ramsys consist of:

• measurement localisation and data verification
• correlation of several data types (e.g.asset inventory, condition data, maintenance and renewal works history, operational data, etc.)
• integrated data visualisation using GIS (Geographic Information System) thematic, linear-graphical, textual and visual modes
  
  
  
  
• generation of derived data such as defects or new parameters, and
• time-based condition data analysis.
  

Retrieving or analysing only one (e.g. the last) set of data such as measurements, allows only corrective maintenance. In other words, if a value is over the threshold, work is done, but if it is not, nothing is done. Having several consecutive measurements and their integrated analysis over time, allows detection of deterioration trends, which enables forecasting of asset behaviour, and therefore planning of actions before values reach respective thresholds, and before defects occur.

RAMSYS performs various analyses, deterioration modelling and optimisation, to reach optimal maintenance and renewal plans by allowing creation ofintelligent business rules and thresholds, which would then be run automatically.

This approach requires much less time and allows thorough analysis of all assets and their behaviour from all possible aspects, without any risks of human errors, whereas the manual methods, including that of visualization, are clearly slow, inefficient, and prone to human errors.