Transport Infrastructure Ireland (TII) has a total network length of c.5300km, made up of roads of varying standards, from high speed, three-lane motorways to lightly trafficked, narrow, single-carriageway roads. The development process for many pavements has been evolutionary, rather than designed or planned.
Historically, TII pavement management practices included collection of pavement condition data, processing and presentation of that data and liaison with local authorities to identify the pavement, maintenance and rehabilitation needs in their areas. Technical advice on the design and construction of the resultant pavement-strengthening schemes was also provided by the authority’s engineering inspectorate.
TII’s pavement management activities have advanced in recent years. The procurement of a commercial, off-the-shelf pavement-management system, Deighton Total Infrastructure Management System (dTIMS), which has been configured to suit the Irish national road network, has allowed TII to obtain full value from the limited historical pavement data sets available along with the more detailed pavement data that is now available. dTIMS has a full range of condition analysis, performance modelling, programme optimisation and scenario-analysis capabilities.
[caption id="attachment_34468" align="alignright" width="300"]
Figure 1: Overview of PMS components [click to enlarge][/caption]As continued investment is essential in ensuring that maximum benefit is derived from the road network, the pavement management system (PMS) is an invaluable tool that uses reliable network (linear) referenced data to determine optimum strategies for evaluating and maintaining our pavements. The PMS is also where pavement inventory, condition, work history and wider network-related information is recorded.
The PMS system is shown schematically in Figure 1, both the PMS Data Repository and the Strategically Modelling Tool have been implemented in dTIMS, which is an open, fully flexible decision-support tool.
Pavement condition – data collection
Pavement condition data collection using high-speed machine-survey vehicles is carried out annually on the national road network. Most of the road condition data is collected using the road-surface profiler (RSP) machine.
In 2013, a laser-cracking measurement system (LCMS) has been used to collect cracking and ravelling data on the entire network. The skidding resistance data is collected using the sideways-force routine investigation machine (SCRIM). In addition, in 2013, a ground-penetrating radar (GPR) survey has been carried out on the entire network to improve the pavement construction and layer thickness data.
Datasets are collected for:
- Longitudinal profile (including international roughness index (IRI));
- Transverse profile (rut depth);
- Macrotexture (mean profile depth);
- Geometrics (crossfall, gradient and radius of curvature);
- Forward view/pavement-oriented digital video;
- DMI linear chainage co-ordinate system;
- GPS geo-referencing coordinate system;
- Crack analysis (alligator, longitudinal, transverse and ruts; patches, ravelling); and,
- SCRIM coefficient (SC).
The survey vehicles are equipped with global positioning system (GPS) and imaging technology is used to gather asset inventory information. All recorded data is referenced to 3-dimensional spatial co-ordinates. The GPS data must be differentially corrected in order to improve accuracy.
The TII corporate GIS, ArcGIS, is used to co-ordinate and cross-reference the data from a range of management systems including the PMS, bridge-management system, accident database, traffic-modelling database, routine maintenance management systems among others.
A new ArcGIS add-in has been developed for the TII to allow display and querying of the imagery collected on the annual surveys.The processed video is also available to the TII and its clients through a web browser system.
Network management – sub-networks
To manage the diversity in our network effectively, it was decided to define a series of sub-networks to group sections of the national road network with similar characteristics such that there is considerably less variation in pavement condition, traffic, construction and performance requirement.
This approach enabled, different service levels to be set for the five identified These different levels equalise the risk of failure and thereby enable fairer comparison of benefits accrued for lower volume roads.
Sub-network 0 comprises the high speed, high-volume pavement network, made up of the motorway and dual carriageway sections of the network.
Sub-network 1 comprises the remainder of the pavement network where significant geometric and pavement design has taken place in the construction and/or rehabilitation of the pavement sections.
Sub-networks 2, 3 and 4 are the legacy sub networks, typically constructed without formal geometric or pavement design over many years. The sub networks are distinguished from one another by traffic volume, with Sub-Network 4 carrying very low volumes of traffic of less than 2000 AADT.
| Subnetwork |
Classification |
Length (km) |
| 0 |
Motorway/Dual Cway |
1147 |
| 1 |
Engineered Pavements |
990 |
| 2 |
Legacy Pavements > 3500 AADT |
1129 |
| 3 |
Legacy Pavement, 2000 to 3500 AADT |
884 |
| 4 |
Legacy Pavement, < 2000 AADT |
1170 |
[caption id="attachment_34477" align="alignright" width="300"]
Figure 2: Cumulative frequency distribution of IRI across sub-networks[/caption]
Table 1: Sub-network classification and length
Figure 2 (right) shows the cumulative distribution frequency curves for the International Road Index (IRI). It can be seen that the five sub-networks are clearly distinguishable from one another. The levels of service definition for IRI on each of the sub-networks is given in Table 2 (below). Similar trends were observed in relation to the other performance parameters.
| Category |
Subnet 0 |
Subnet 1 |
Subnet 2 |
Subnet 3 |
Subnet 4 |
| Very Poor |
> 3 |
> 3.5 |
> 5 |
> 5 |
> 7 |
| Poor |
2.5 to 3 |
3 to 3.5 |
4 to 5 |
4 to 5 |
5 to 7 |
| Fair |
2 to 2.5 |
3 to 3.5 |
3.2 to 4 |
3.2 to 4 |
4 to 5 |
| Good |
1.5 to 2 |
2 to 2.5 |
2.7 to 3.2 |
2.7 to 3.2 |
3 to 4 |
| Very Good |
< 1.5 |
< 2 |
< 2.7 |
< 2.7 |
< 3 |
Table 2: IRI performance categories by sub-network
Pavement maintenance and renewal scheme ranking
Utilising condition prediction modelling in the PMS, the ranking of maintenance and renewal schemes is carried out using a ‘most-in-danger first’ approach. Schemes are ranked based on the percentage by which each of the three selection parameters (IRI, Rut, LPV3) in the section exceeds the relevant threshold values for the section (percentage above threshold or PAT) and is further refined by reference to predicted deterioration rates.
Using percentage values normalises the value for each parameter and allows direct comparison and ranking across the sub-networks.
[caption id="attachment_34478" align="alignright" width="300"]
![Figure 3: Performance prediction model IRI [click to enlarge]](http://www.engineersjournal.ie/wp-content/uploads/2017/02/Figure-3-300x180.png)
Figure 3: Performance prediction model IRI [click to enlarge][/caption]For each of the sample units (100m long) on a route section, the PAT for each parameter is calculated (capped at 150 per cent for all parameters). The sample unit characteristic PAT is the sum of the worst two parameter PATs. The value of the section is the average of the sample units.
This ranking scheme has been developed through a process of multiple refinements, with direct comparison of the ranking to rankings produced by experienced pavement engineers working within the TII as well as the results being further verified by visual inspection on the ground.
[caption id="attachment_34479" align="alignright" width="300"]
Figure 4: Performance prediction model RD [click to enlarge][/caption]The main objective of a future oriented pavement management process is to predict the effects and outcomes of maintenance activities from both technical and economic points of view. The methodology used in the TII PMS is advanced life-cycle-cost-analysis (LCCA), which enables prediction of future pavement condition by using performance models for various condition parameters (performance indicators) and the comparison of different maintenance treatment strategies under given preconditions (e.g. available budget).
To enable feasible performance prediction, the network must be sub-divided into homogeneous segments i.e. similar adjacent 100m dTIMS sectionsare aggregated into longer segments. Performance prediction models are then used to define the time-dependent change of road pavement characteristics or loading parameters during the analysis. Models are defined using so called analysis variables that can be applied to the input (actual condition) data.
For example, ‘Error! Reference source not found’ shows the deterministic performance prediction models for the performance indicators IRI and RD.
Treatment catalogue and assessment procedures
[caption id="attachment_34481" align="alignright" width="300"]
![Figure 5: Treatment options [click to enlarge]](http://www.engineersjournal.ie/wp-content/uploads/2017/02/Figure-5-300x218.jpg)
Figure 5: Treatment options [click to enlarge][/caption]The treatment catalogue is a list of representative heavy (rather than routine) maintenance treatments. It includes information about costs, triggers, reset values (technical and financial/economic effects) as well as possible subsequent treatments. Figure 5 gives an overview of this catalogue.
These procedures include the basics for the technical assessment of pavement condition (calculation of a total condition index as the maximum of condition classes of single parameters) and the cost-benefit-analysis on technical level
The system is configured so as to include the basics for a macro-economic assessment of pavement condition and the cost-benefit-analysis using external costs due to improved pavement condition and due to maintenance treatments (vehicle operation costs), time costs, environmental costs and accident costs).
For the optimisation, different analysis sets (standard for ten years, and long-term for 20 years) with different budget scenarios are implemented into TII PMS. During the optimisation, the system maximises the benefit (technical and macro-economic) of all possible maintenance treatment strategies (see Figure 5) over all analysis sections under the given budgetary constraints.
Beside standard budget scenarios, a do-nothing-scenario and a technical-optimum-scenario (unlimited budget) will be calculated automatically for comparison purposes.
Overview of results
The results of the analysis can be used in the different business processes of the TII on both strategic and planning (object) levels and can be used to define policies or general strategies for maintaining the TII road-network.
The results form a basis for reporting on the different levels, and can be subdivided into:
- PMS-sectioning results: output of the PAT- and PMS-sectioning process;
- Object level results: results of generating maintenance treatment strategies on each single road section and optimisation;
- Network level results (technical): summary of object level results, representing technical indicators;
- Network level results (economic): summary of object level results, representing monetary (economic) indicators;
- Also allows for proposed pavement schemes to be viewed relative to other network activity e.g. structure rehabs etc ensuring enhanced asset-management operations; and,
- Provides a high level view (bar charts/pie charts etc) of network performance over a given time period or using a specified budget. This quickly allows management/government to decide the most appropriate level of targeted investment in the network, optimising overall investment in the asset.
[caption id="attachment_34482" align="alignright" width="300"]
Figure 6: Section-based results of PMS-analysis (dTIMS CTTM) [click to enlarge][/caption]In Figure 6, the maintenance priority for each section across the national road network identified as needing treatment is shown.
The network level results summarise the section related maintenance treatment recommendations of the different scenarios to be analysed. The results are the output of the first analysis, which will be used to assess the pre-selected models.
Any adjustment of the model will also change the results to different extent. Hence this process will evolve over a number of years use on the system as the segmentation of homogenized reflects a higher confidence level in the structural behavior of the various pavement types.
Although the scenario ‘technical optimum’ uses an unlimited budget, the trigger for the maintenance treatments (as a part of the treatment catalogue) are still valid, so that the treatments will be applied when a road section reaches a poor condition only and not at every stage of condition.
[caption id="attachment_34485" align="alignright" width="300"]
Figure 7: Total Network Condition Index “Technical Optimum” [click to enlarge][/caption]This option effectively allows the system to ‘choose’ whatever budget it needs in any given year in order to optimise the traffic-weighted condition of the network, while still being conscious of delivering an economically efficient solution. Figure 7 shows the outcome from this largely theoretical exercise.
The unconstrained budget set out to ensure that no section of the network is in Poor and Very Poor after Year 1, and ensures that this continues throughout the full 20-year analysis period. Most of the sections are maintained in Good rather than Very Good condition over the lifetime of the analysis. Essentially it set out to bring the entire network to an optimum level immediately and maintain that condition.
This is usually the lowest whole life cost for an asset as the deterioration rates are lower in early life. Hence if over the lifetime of the asset there is a programme to renew the asset regularly by a minor intervention, this will have a lower WLC than allowing it to deteriorate to a lower operational level and intervening less frequently but with a more intensive treatment.
To achieve this condition would require an average annual spend of about €185m, however it is not evenly distributed. Most of the overall costs are to be expended in Year 1 to bring the network to optimal condition. In addition as we have a ‘live’ network there would be significant disruption and possible temporary closure to enable the works to be carried out.
Resources significantly beyond those deployed to meet the entire interurban programme would also be necessary. Whereas it may be a technically optimal solution, it is not a very practicable one.
The model’s predictions are continually being recalibrated, based on current network condition. To demonstrate how the system operates, a number predictions and graphs as to the likely network condition and spend level have been generated. They are
NOT working outputs but an exercise to demonstrate the capabilities of the system. In the following tables and graphs, the condition distribution of the total condition index is shown for each individual sub-network based on a budget scenario of a maximum of €60 million per year.
With regard to the treatment catalogue the treatment cost and length distribution over the next 20 years is shown in form of graphs for the scenario max. €60M/year.
Optimising programme costs
[gallery columns="4" ids="34496,34497,34495,34498,34499,34501,34502,34503"]
It can be seen when comparing the technical optimal to the €60 million per annum budget, there is a need to balance whole-of-life costing with a practical application of resource and availability management. The objective essentially would be to find an expenditure that can be implemented with minimal disruption, but is still capable of maintaining optimal operational characteristics.
[caption id="attachment_34506" align="alignright" width="300"]
Figure 16: Scenario €140m p.a. - Total Network Condition[/caption]
Technical optimal has not just a lower whole-life cost to the agency; it ensures the road condition is such that user costs are also minimised. Smoother roads with lower rolling resistance lower the vehicle maintenance and operational costs. In looking at scenarios, a useful exercise is to plot both agency and user costs for each option and determine at what investment level they are at a minimum for the country as a whole, i.e. agency and user combined.
The outcome of this exercise determined that to be a programme that required circa €140 million of investment per annum. Using this as our budget, we can demonstrate the outcome from an agency perspective for the network as before over a 20-year period.
This budget would result in a significantly different works programme, as illustrated in Figure 17.
In addition to the analysis carried out using dTIMS, the Authority implements a proactive approach in the management of skid resistance on its network. The skidding resistance of the entire national road network is measured in the left-hand wheel path using SCRIM.
[caption id="attachment_34507" align="alignright" width="300"]
Figure 17: Scenario €140m p.a. – Total Network Treatments[/caption]
The objective of the standard is to manage the risk of skidding collisions in wet conditions, so that this risk is broadly equalised across the national road network. Event locations with greater potential for vehicle/vehicle or vehicle/pedestrian interactions (e.g. roundabouts, approaches to traffic lights) are assigned more onerous levels of required skid resistance to reflect the higher risk of collisions.
Over 16,000 separate event locations have been identified on the network and an appropriate investigatory level of skid resistance assigned to each one.
Conclusions
Asset management focuses on the best way to achieve the overall objectives of the organisation rather than increasing the efficiency of maintenance activities. The ability to demonstrate that infrastructure is being preserved and the consequences of not investing in asset management are critical.
By enabling an understanding our network, defining ways of measuring our delivery, identifying means of improving system performance, and then continually monitoring to see how we are achieving targets, the PMS is an invaluable tool. Pavement models are enhanced by map-based reporting and analysis and the systematic availability of associated information like traffic modelling.
For further information or observations on the system, please contact TII. Email tom.casey@tii.ie.