Attribute Propagation for Utilities.

Utility systems such as electric, fiber/telco, gas, and water require the realistic modeling of network attributes or values over distance. For example, consider hydraulic pressure in a pipe network; as water flows away from the reservoir or pump, pressure decreases due to pipe friction, leakage, consumption, etc. Attribute propagation is the process whereby network attributes that change over distance (e.g., maximum allowable operating pressure, phase, etc.) are calculated and maintained. This is important for improving safety as well as efficiency. However, attribute propagation is challenging due to the size of the data, which could have tens of millions of nodes and edges per utility, and billions of nodes and edges at the nationwide scale. Additionally, results may need to be calculated and available quickly for interactive analysis. Previous approaches require immediate updates to all nodes and edges downstream of a node/edge being edited (to account for changes in attribute values), which could be computationally intensive and result in a slow user experience for editing attribute values. This paper presents Propagators, which feature an in-memory approach to attribute propagation. Propagators leverage a network index as well as a heuristic based on colocated sources with similar attribute values to increase computational savings. We present experiments that demonstrate the scalability of Propagators, which have been implemented in ArcGIS Pro and ArcGIS Enterprise.