Technique for Calibrating Network Models

Abstract
In calibrating a water distribution system model, the model user usually adjusts pipe roughness (e.g., Hazen‐Williams C factor) or water use so that pressures and flows predicted by the model agree with values observed in the field. This paper presents formulas to assist the user in deciding whether to adjust C or water use and by how much. The key to using the formulas is to observe pressures in the system for at least two significantly different use rates. Such data are often collected during fire flow tests. A model is considered calibrated to the extent that it can predict the behavior of the water distribution system over a wide range of operating conditions and water use.