18.04.16
Understanding asset behaviour to maximise budget performance
Source: RTM Apr/May 16
Panayioti Yianni, Luis Neves, Dovile Rama and John Andrews from the University of Nottingham explore methods to manage large portfolio of assets.
A common theme in many sectors today is facing increased demand on assets whilst enduring budgetary restrictions. In this situation, understanding the intricate behaviour of the assets being managed is paramount. Working alongside Network Rail, the UK’s largest railway operator, allows investigations into assets which we have come to rely on, yet understand little about. The focus of this work was analysing the behaviour of Network Rail’s portfolio of railway bridges to understand their deterioration, the policies for inspection and maintenance and how they all interact.
The first stage of our work was to understand all the different deterioration mechanics of these structures and how this deterioration is captured in current inspection methods. Focusing on concrete bridges, the most common newly built bridge, all the different defects were analysed from decades of industry inspection data. Not only is it important to understand the severity of different defects, but the rate of their spread and the other defects that they can evolve into. Understanding the behaviour of these defects is key to taming their impact.
This led to the development of a stochastic Petri-Net model that simulates deterioration, inspection and maintenance, explicitly taking into account uncertainty in all these aspects. The main focus of the model was to define the deterioration process directly from the comments provided by visual inspections, using a bi-dimensional condition matrix, evaluating both the severity and extent of damage. This resulted in a model which is significantly more sympathetic to the observed deterioration, but also intuitive to the final user. The model was calibrated based on historical data of 30,000 bridges, inspected over the last two decades. The inspection policies of Network Rail, as well as current maintenance policies were included in the model. Expert judgment was used to take into account less than perfect inspection and maintenance occurrences.
Once a validated model was defined, it was possible to better understand the behaviour of the bridge stock as a whole, predict conditions of bridges over time, requirements for maintenance and expected maintenance costs. However, deterioration varies significantly across the bridge network, with seemingly similar bridges deteriorating at very different rates. This led onto the next phase of our study, comprising the analysis of local environmental effects. Through extensive collaboration with Network Rail we created a shortlist of possible factors that were affecting bridge deterioration from the underlying geology to quantity of rainfall.
New approach
The most straight forward place to start the analysis was to split the population of bridges into the two main types: overline and underline bridges. Existing methods to compare deterioration rates have a number of drawbacks so a new approach was developed to perform this analysis that is much more intuitive.
Different groups of bridges were compared in terms of the mean time between levels of damage. Statistical analysis was used to identify differences between groups of bridges (e.g. overline bridges vs underline bridges). Overline bridges carry the roadway over the train tracks and are subjected to different loading patters to underline bridges, which carry the heavier trains over the roadway. They are also exposed to different environmental stressors, including de-icing salts. The results showed that overline bridges don’t deteriorate nearly as quickly as underline bridges, demonstrating that different inspection and maintenance policies should be applied to each group.
Following this, a further analysis into underline bridges was performed, evaluating if annual tonnage played a role in deterioration. Results showed that there is no evidence to suggest a link between the gross tonnage and the deterioration rate. We suspect that deterioration is influenced by peak load, rather than annual tonnage and further work is necessary to define this relation.
Now that we have determined which bridges are at risk of rapid deterioration and which are not, the next phase of our work will be calculating the optimal way to manage these assets. By focusing on those that need more attention and relaxing with those that don’t, we can direct the budget to where it is needed, improving budgetary efficiency whilst maintaining the same level of health across the population of bridges.
Tell us what you think – have your say below or email [email protected]