FME & Neural Networks for the Automated Utilities Clash Detection, Engineering Analysis and Treatment Assessment for Light Rail Projects
During the design phase of light rail projects, significant time is spent identifying and avoiding utility clashes. Utilising the latest in available technology Arup has developed an automated clash detection and asset protection platform. This platform utilises available spatial asset data, to form the basis of a project specific utilities records. These records are compared to the latest project design, any impacted assets proceed through a predefined Risk (Red, Amber and Green), high level treatment (Divert, Protect and Leave), and specific works treatment analysis. This analysis is based on hierarchical functional analysis (IF, OR, AND) along a neural network (machine learning) to apply engineering logic to identify remediation options specific to the asset information (utilising size, material, type) and applying network utility owners (NUO) preferred treatment method for assets interacting with design. This engineering logic was developed in conjunction with the NUO's asset management teams, exemptions or specialised treatments for particular assets, safety clearances buffering for asset types. Outputs from this analysis are able to be utilised per project or can be developed into a holistic asset protection platform. This platform greatly reduces the impact of construction works on NUO's assets, identifying possible assets susceptible to stray current corrosion, delays and subsequent costs associated with relocating assets. In addition to improving designer, contractor and NUO's understanding of the interaction assets and design.
FME International User Conference 2017