Premium
Model calibration data: the good, the bad, and the useless
Author(s) -
Walski Thomas M.
Publication year - 2000
Publication title -
journal ‐ american water works association
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.466
H-Index - 74
eISSN - 1551-8833
pISSN - 0003-150X
DOI - 10.1002/j.1551-8833.2000.tb08791.x
Subject(s) - calibration , head (geology) , computer science , field (mathematics) , data loss , variety (cybernetics) , statistics , mathematics , artificial intelligence , geology , geomorphology , pure mathematics , computer network
Comparisons between field data and model results need to be made with “good” field data. Water distribution models are used to simulate a variety of conditions, including extreme flow events such as maximum hour demands and fire flows. Some models are tested only under average demand conditions, in which head loss is commonly of the same order of magnitude as errors in head loss measurements. Data collected when head loss is small in comparison with errors may give the false impression that the model is properly calibrated. Substantial errors may exist in the model that will not become apparent until the model is used at a time when head loss and velocity are high. This article contains guidelines for collecting data so that the model calibration will be meaningful.