Fuzzy data reconciliation in reacting and non-reacting process data for life cycle inventory analysis

Data uncertainty is a critical issue in life cycle inventory analysis (LCI). Recent work has demonstrated that fuzzy mathematics provides a computationally efficient alternative to probabilistic methods for representing data uncertainty. One specific problem is the utilization of different, and potentially conflicting, LCI data sources such as physical measurements, estimates or databases. A fundamental requirement of a valid LCI is that the data must not violate material and energy balance principles; however, data from diverse sources may result in inconsistencies. Normally such inconsistencies in LCI data can be addressed through the use of data reconciliation methods based on probability theory. This paper presents an alternative data reconciliation method based on fuzzy mathematical programming. Two LCI case studies are included to illustrate the methodology.