The availability of a good hydraulic model increases the reliability of the results of methodologies using it. Thus the calibration of the model is a previous step that has to be done. The most uncertain parameters of the model are demands due to their constant variability. However, calibrating these demands requires a high computational cost that can be reduced by redefining the unknown parameters from demands to demand patterns. Besides, the number and location of the used sensors is highly correlated with the definition of such patterns. This paper presents a methodology for parameterizing and selecting sensors using the information from the singular value decomposition of the water distribution network sensitivity matrix.
Sanz, Gerard and Perez, Ramon, "Parameterization And Sampling Design For Water Distribution Networks Demand Calibration Using The Singular Value Decomposition: Application To A Real Network" (2014). CUNY Academic Works.