作者: Ramon Perez , Josep Cuguero , Joaquim Blesa , Miquel A. Cuguero , Gerard Sanz
DOI: 10.1109/SYSTOL.2016.7739770
关键词: Uncertainty analysis 、 Measurement uncertainty 、 Residual 、 Engineering 、 Data mining 、 Sensitivity analysis 、 Data modeling 、 Leak 、 Monte Carlo method 、 Synthetic data
摘要: The leak localisation methodologies based on data and models are affected by both uncertainties in the model measurements. This uncertainty should be quantified so that its effect methods performance can estimated. In this paper, a model-based methodology is applied to real District Metered Area using synthetic data. generation process of data, demands taken into account. was estimated it justify observed consists, first, generating set possible measurements, obtained Monte Carlo Simulation under certain assumption considering uncertainty, second, falsifying sets nodes correlation with residual order signal leaky nodes. assessment done means confusion matrix approach.