A multi-step balancing procedure for quality assessment of substance flow data
2005 (English)Report (Other academic)
All estimates of substance flows are more or less uncertain, which implies that the collected data can violate mass balance constraints that should be valid. In this article, we introduce multi-stage balancing algorithms that can accommodate prior information about mass balance constraints and uncertainty of the collected data. In particular, we formulate the balancing task as an optimization problem for a given set of prior information. If it is suspected that some flows have been overlooked, the balancing is achieved by minimizing the total increase in flows that is required to satisfy the given mass balance constraints. If the major problem consists of errors or uncertainty in the raw data, the sum of squares of all adjustments needed is minimized. We present a software prototype in which the balancing is integrated with a variety of tools for quality assessment of collected data, and use data from a previously published study of nitrogen flows in Sweden to illustrate the steps involved in the proposed algorithms.
Place, publisher, year, edition, pages
LiTH-MAT-R, ISSN 0348-2960 ; 1
substance flows, mass balance, matrix operations, optimization
IdentifiersURN: urn:nbn:se:liu:diva-86478OAI: oai:DiVA.org:liu-86478DiVA: diva2:577854