Reconciling continuum and non-continuum data with industrial application |
| |
Authors: | Ruben GonzalezBiao Huang Fangwei XuAris Espejo Joseph AmalrajWilliam Lam |
| |
Affiliation: | a The Department of Chemical and Materials Engineering, University of Alberta, Edmonton, Alberta T6G 2G6, Canada b Syncrude Canada Ltd., Fort McMurray, Alberta T9H 3L1, Canada |
| |
Abstract: | In order to perform data reconciliation, it is important that noises in the data have well-defined distributions. The motivation behind this study was to enable the comparison between a discrete and continuous data set so that means can be compared for gross error over the short term; this required that local variables exhibit similar distributions.A case study was done on a system where non-continuum loads from a dump truck were to be reconciled with two downstream continuum weightometers. An algorithm was developed using the binomial distribution and time delay in order to simulate the effect of the dump pocket.Regression analysis based on principal components was used to evaluate the performance of the smoothing algorithm and to determine the most likely maximum hopper capacity that locates between the two weightometers. |
| |
Keywords: | Data smoothing Binomial-normal data comparison Principal component regression |
本文献已被 ScienceDirect 等数据库收录! |