Treatment of uncertainties in green engineering

Abstract

Uncertainties occurring in engineering are usually dealt with by means of statistical validation. Most value assignments are of statistical nature and thus only valid in a specific statistical way. This fact is, in most cases, only implicitly apparent but nevertheless almost always present. In green engineering this is not true. The calculation e.g. of life cycle assessments (LCAs) is generally based on a vast number of values. These data either cannot be validated in a statistical sense at all or they are too insufficient to be usable for statistical analyses. This means that final results in this field cannot be declared as trustworthy in a way that is common in engineering. In this paper, the method of interval arithmetic is proposed as a way to circumvent this severe epistemological problem. The consequences of the suggested approach are illustrated utilizing two typical examples: (1) Different reliabilities concerning theories for greenhouse effect influences of various chemical emissions on the greenhouse effect and thus on the global warming potential and (2) Situations, where one specific production site differs heavily from the average of the dedicated industrial sector. Computational implementation of the considered concept is briefly discussed. It is shown that misleading results, which may be quite common in this field, can possibly be avoided in the proposed way.