Search

Uncertainty is commonly not taken into account in LCA studies, which downgrades their usability for decision support. One often stated reason is a lack of method. The aim of this paper is to develop a method for calculating the uncertainty propagation in LCAs in a fast and reliable manner. The method is developed in a model that reflects the calculation of an LCA. For calculating the uncertainty, the model combines approximation formulas and Monte Carlo Simulation. It is based on virtual data that distinguishes true values and random errors or uncertainty, and that hence allows one to compare the performance of error propagation formulas and simulation results. The model is developed for a linear chain of processes, but extensions for covering also branched and looped product systems are made and described. The paper proposes a combined use of approximation formulas and Monte Carlo simulation for calculating uncertainty in LCAs, developed primarily for the sequential approach. During the calculation, a parameter observation controls the performance of the approximation formulas. Quantitative threshold values are given in the paper. The combination thus transcends drawbacks of simulation and approximation. The uncertainty question is a true jigsaw puzzle for LCAs and the method presented in this paper may serve as one piece in solving it. It may thus foster a sound use of uncertainty assessment in LCAs. Analysing a proper management of the input uncertainty, taking into account suitable sampling and estimation techniques; using the approach for real case studies, implementing it in LCA software for automatically applying the proposed combined uncertainty model and, on the other hand, investigating about how people do decide, and should decide, when their decision relies on explicitly uncertain LCA outcomes-these all are neighbouring puzzle pieces inviting to further work.

Qualitative valuation methods carefully try to avoid an aggregation across impact categories. However, such an aggregation often helps in obtaining a clear result for the valuation (which product scores better?). This article presents a new valuation method that uses an iterative approach. The application is demonstrated by the help of a case study for electric motors in trains. The approach combines two existing, unique valuation methods described earlier in literature, which both are of a rather non-aggregating nature, in line with ISO requirements, and were designed to be performed by LCA experts. The method is implemented in a computer software. Besides constants used within the method, the software needs as input solely indicator values from the Impact Assessment. The iterative nature of these methods itself, and especially the combination of these methods, helps in achieving a valuation result for the LCA with not more subjective and aggregating elements than necessary. Subjective elements are clearly separated from others. The algorithm seems highly sensitive to changes in impact categories regarded as important ones. The implementation in software greatly eases the application of the method by transferring routine work from LCA experts to a machine. It ensures a reproducible result and prevents erroneous steps in a rather complicated valuation procedure. It further helps in hiding the complexity of the method from the user. The approach of combining valuation methods in LCAs seems a fruitful one, and shows benefits when implemented in computer software, in terms of usability, and in terms of a more reproducible application. Care has to be taken to make sure users know what they do when performing an automated valuation procedure. We see three ways for extending the approach, namely: (i) become part of a toolbox of different valuation procedures; (ii) explicitly cope with uncertainty, and (iii) include different values for normalisation, in different regions worldwide. The software will be made available also in a stand alone version.

Goal and Background Geographical and technological differences in Life Cycle Inventory data are an important source for uncertainty in the result of Life Cycle Assessments. Knowledge on their impact on the result of an LCA is scarce, and also knowledge on how to manage them in an LCA case study.

Diese Arbeit hat das Ziel, eine praktisch anwendbare Methode zur Berechnung der Fehler in Ökobilanzen zu entwickeln. Die Entwicklung der Methode geschieht in der Arbeit mit Hilfe eines Modells, das die verschiedenen, heute üblichen Rechenschritte der Ökobilanz abbildet. Das Modell erlaubt es, die wahren Werte für die Fehler, die tatsächlichen Fehler , an verschiedenen Stellen der Ökobilanz wie z.B. in Zwischenergebnissen und insbesondere auch im Endergebnis der Ökobilanz, zu berechnen. Innerhalb des Modells wurden sechs verschiedene Methoden der Fehlerrechnung implementiert und untersucht. Wesentliches Kriterium war dabei die Darstellungsgüte der Methoden. Die Darstellungsgüte ist ein Maß dafür, wie gut der von einer Methode berechnete Fehlerwert dem Wert des tatsächlichen Fehlers entspricht. Der tatsächliche Fehler konnte so für eine Validierung der berechneten Fehlerwerte genutzt werden. Die entwickelte Methode unterscheidet die Berechnung zufälliger und systematischer Fehler. Zunächst werden systematische Fehler, also Fehler, die reproduzierbar sind, in den Daten korrigiert und falls erforderlich die Höhe der systematischen Fehler im Endergebnis berechnet. Diese Berechnungen und Korrekturen können vollkommen exakt durchgeführt werden. Ein Rückgriff auf Approximationsrechnungen ist nicht erforderlich. Auf Basis der um systematische Fehler korrigierten Werte werden anschließend die zufälligen Fehler berechnet. Als wesentlicher Parameter für die Darstellungsgüte hat sich der relative Fehler in Inputwerten der Rechnung erwiesen. Je nach Höhe dieses relativen Fehlers und je nach Rechenschritt der Ökobilanz empfiehlt das entwickelte Konzept, bestimmte Fehlerrechnungsmethoden nicht zu verwenden. Bei höheren Werten für den relativen Fehler ist insbesondere die Gaußsche Methode der Fehlerfortpflanzungsrechnung nicht mehr an-wendbar. In den Modellrechnungen hat sich gezeigt, daß sie dann den Fehler deutlich zu niedrig berechnet. Eine ebenfalls untersuchte Formel nach Bader/Baccini erbringt auch für höhere relative Fehler im Input Ergebniswerte, die nah an den tatsächlichen Werten liegen. Bei weiter erhöhten Werten für den relativen Fehler ist von den untersuchten Methoden nur noch die Monte Carlo Simulation in der Lage, den Fehler im Ergebnis gut abzubilden. Diese Methode stellt jedoch sehr hohe Anforderungen an die Zeit- und Hardwareressourcen, was den Einsatz in der praktischen Anwendung erschwert. Durch systematisch durchgeführte Parameteränderungen in den Modellrechnungen konnten für die Rechenschritte der Ökobilanz jeweils konkrete Werte für den relativen Fehler ermittelt werden, die wie Grenzwerte den Einsatz der Fehlerrechnungsmethoden beschränken. Der relative Fehler läßt sich in der praktischen Anwendung aus Inputwerten der Rechnung bzw. aus dem Ergebnis des vorangegangenen Rechenschritts bestimmen. Dieser Wert für den relativen Fehler ist dann mit dem ermittelten Grenzwerten abzugleichen, und aus diesem Abgleich läßt sich schließlich entscheiden, ob die Verwendung der Fehlerrechnungsmethode für den einzelnen Rechenschritt zulässig ist. The aim of this work is to develop a method for calculating errors in Life Cycle Assessments (LCAs), that can be applied in practice. This method is developed by putting up a model that covers the different calculation steps of a Life Cycle Assessment, as they are commonly used today. The model allows the calculation of true values for errors at different stages within the LCA calculation, and also in the final result of the LCA. In the model, six different methods for calculating errors where implemented and analysed. Essential for the analysis was the goodness of fit of each method. The goodness of fit is the measure to what extent the error, as calculated by an approximation method, corresponds to the true value of the error ( true error ). Thus, the true error was used to validate the calculated error. The error calculation method developed in this work distinguishes between the calculation of systematic errors and of random errors. At first, systematic errors, being errors that are reproducible, are cleared from the input data of the LCA, and if necessary or desired, the systematic errors in the result of the LCA can be calculated. Both clearing and calculation can be done in an exact way, without the need to refer to approximation formulas. In a second step, the random errors are calculated. The relative error turned out to be the essential parameter for discovering the goodness of fit of each method. Depending on the value of the relative error and the calculation step in the LCA, the method recommends not to use certain approximation formulas. The Gaussian error propagation formula turned out to largely underestimate the error, if the relative error had higher values. A formula developed by Bader and Baccini performed better in these cases. With even higher relative errors, from the methods analysed, only the Monte Carlo simulation was able to calculate the errors correctly. A systematic change of the parameter values accessible in each calculation step revealed distinct limits for the relative error specific for each calculation step in the LCA, and for each approximation formula. These limits span intervals specific for each approximation formula and each calculation step. In the calculation of an LCA, the relative error can be obtained either from input data or from a preceding calculation step. A check whether the error lies in the appropriate interval indicates whether an approximation formula should be used for calculating the errors in that calculation step.

Modelling data uncertainty is not common practice in life cycle inventories (LCI), although different techniques are available for estimating and expressing uncertainties, and for propagating the uncertainties to the final model results. To clarify and stimulate the use of data uncertainty assessments in common LCI practice, the SETAC working group ‘Data Availability and Quality’ presents a framework for data uncertainty assessment in LCI. Data uncertainty is divided in two categories: (1) lack of data, further specified as complete lack of data (data gaps) and a lack of representative data, and (2) data inaccuracy. Filling data gaps can be done by input-output modelling, using information for similar products or the main ingredients of a product, and applying the law of mass conservation. Lack of temporal, geographical and further technological correlation between the data used and needed may be accounted for by applying uncertainty factors to the non-representative data. Stochastic modelling, which can be performed by Monte Carlo simulation, is a promising technique to deal with data inaccuracy in LCIs.

This paper is a two-fold introduction. For one, it introduces a new, open source, LCA software. Second, it is to establish a new section in Int J LCA named ‘LCA Software’. Herewith, the editors of the journal recognise the growing possibilities and the impact of software, meaning both databases and calculation as well as modelling software, for practical applications as well as for the scientific development in LCA. This section is designed to house a broad variety of papers to beLCA focused and related to ICT (Information and Communication Technology). In this sense, announcements (as this one), conference reports, but also peer-reviewed papers on methodology and case studies, are most welcome.

Conservation and management of freshwater resources has become a major challenge of this century. In mining, one should be able to control the quantity of water intake, as well as the volume and quality of effluents. When the water loops in mining industry are closed, new methods for water quality control and optimisation at each process step are required, which facilitates also the recovery of additional valuable elements. The tightly closed water cycles necessitate that the tailings are filtered and stacked dry. Geopolymerisation of remaining tailings has the potential to be used for water and oxygen tight covers on the deposited tailings and for mine back-fill use. Detection of ore types and their classification or sorting by their geo-metallurgical properties before and in mill feed have a significant effect on metals recovery and on the process water quality in flotation. The main objective of this paper is to review the progress made towards a new paradigm shift at mine sites to recycle water and valorise tailings for an improved environmental and economic result, and for increased social acceptance of mine operations.