5.7.2 Network segmentation and data aggregation
Network segmentation and data aggregation are used to characterise the road network for the various possible forms of WOLCC analysis. The segmentation and aggregation processes can be a major task in program development and can have a significant impact on the accuracy and reliability of the investment predictions using a WOLCC analysis. The input data must be representative of the network and appropriate to the level of analysis being undertaken. The analysis may vary from a strategic level such as assessing the impacts of different funding scenarios on the overall network performance to the development of project level recommendations on individual road segments (Austroads 2005).
There are two key interrelated steps required with regard to the preparation of the input data, namely road network segmentation and data aggregation.
- Road network segmentation is the task of subdividing the road network into manageable and homogeneous representative lengths. A segment is defined here as the smallest length of a road that is considered in the analysis process (e.g. smallest length considered for maintenance projects). Segments are also called sections or blocks.
- The task is to determine road segment lengths that are appropriate for the type of analysis being undertaken. Generally, segments need to be long enough to form an on-site project and short enough so that the data describing the length accurately portrays the characteristics of the segment in the analysis. Approaches vary from having a whole network characterised by a small number of uniform road sections (e.g. network wide budget assessment) to networks analysed in 100 metre length segments (e.g. program development). Ideally, segments need to have reasonably homogeneous properties with respect to surface conditions, strength, climate and traffic.
- Data aggregation is the aggregation of the road inventory and condition information into the defined segments in order to adequately reflect the characteristics of the road segment.
These two processes are potentially inter-dependent, as the road segments may be defined by using aggregated data. Further data aggregation is then conducted by transforming the road inventory and condition data into uniform segments that were defined during segmentation.
For more information on the above, refer to Part 6, Section 5.3. There are a number of approaches used to segment road networks to characterise their salient features. These approaches range from the rigorous and repeatable to those based on either engineering judgement or the use of a fixed pre-determined length. In principle, segmentation of pavement lengths can be based on:
- homogeneity where most of the relevant condition (rutting, roughness and strength) and road use parameters (traffic load) are relatively constant
- statistically representative values of the relevant condition and road use parameters that do not have significant standard deviation from their mean value.
Neither of the above approaches provides a satisfactory solution on their own; however, combining the two approaches is a practical means to gain representative segments. Improved computing capacity has allowed the typical use of 100 metre long segments to road sub-networks and extended program analyses.