Advances in application of machine learning to life cycle assessment: a literature review

Purpose

Life Cycle Assessment (LCA) is the process of systematically assessing impacts when there is an interaction between the environment and human activity. Machine learning (ML) with LCA methods can help contribute greatly to reducing impacts. The sheer number of input parameters and their uncertainties that contribute to the full life cycle make a broader application of ML complex and difficult to achieve. Hence a systems engineering approach should be taken to apply ML in isolation to aspects of the LCA. This study addresses the challenge of leveraging ML methods to deliver LCA solutions. The overarching hypothesis is that: LCA underpinned by ML methods and informed by dynamic data paves the way to more accurate LCA while supporting life cycle decision making.

Methods

In this study, previous research on ML for LCA were considered, and a literature review was undertaken.

Results

The results showed that ML can be a useful tool in certain aspects of the LCA. ML methods were shown to be applied efficiently in optimization scenarios in LCA. Finally, ML methods were integrated as part of existing inventory databases to streamline the LCA across many use cases.

Conclusions

The conclusions of this article summarise the characteristics of existing literature and provide suggestions for future work in limitations and gaps which were found in the literature.

     

Challenges of organizational LCA: lessons learned from road testing the guidance on organizational life cycle assessment

Purpose

The novelty of the O-LCA method and the existing differences with the established product LCA practice, as well as the unique structure each organization, pose a broad range of methodological and application challenges, in addition to the general methodological gaps in LCA. In order to provide practitioners with lessons learned for future applications and boost future method development efforts, the paper discusses those challenges.

Methods

The challenges included in this paper were mainly identified from a survey administered to the road testers and from experiences during the piloting process. These are complemented with case studies from literature. The focus of the paper is on challenges exclusive to the organizational approach, although some additional issues common to product LCA but intensified in organizational LCA are also included. Each issue is characterized and exemplified, recommendations of reference standards are analyzed, and possible solutions discussed.

Results and discussion

With the goal and scope of O-LCA, some challenging issues were to select part of an organization as the reporting organization, and the operability of the reporting flow. Regarding the system boundary, the challenges were which parts of the supply chain should be included in the study, problems when setting the system boundary for the service sector, how to include supporting activities, and how to prepare the right system boundary diagrams. Regarding the inventory stage, the discussion starts with alternatives to the categorization of the inventory into activities and the aggregation of those activities into groups. It includes an equivalence table for an easier transfer from other organizational frameworks (ISO 14069 and the GHG Protocol). Some challenges during impact assessment and interpretation were the assessment of local impacts, scoping performance tracking, and the use of O-LCA results for an organization’s strategy.

Conclusions

The review of challenges is not meant as a complete overview of all possible challenges—new challenges may arise in future case studies. Further application testing is needed, along with research to support a future revision of the O-LCA Guidance, in line with the issues highlighted in this paper and new challenges may arise in future case studies. O-LCA has the potential to contribute in the future implementation of the life cycle concept in environmental management systems, in the development of organizational footprint metrics for region-specific impacts, and in the social dimension of life cycle assessment.

     

Ways to get work done: a review and systematisation of simplification practices in the LCA literature

Purpose

Within the field of life cycle assessment (LCA), simplifications are a response to the practical restrictions in the context of a study. In the 1990s, simplifications were part of a debate on streamlining within LCA. Since then, many studies have been published on simplifying LCA but with little attention to systematise the approaches available. Also, despite being pervasive during the making of LCA studies, simplifications remain often invisible in the final results. This paper therefore reviews the literature on simplification in LCA in order to systematise the approaches found today.

Methods

A review of the LCA simplification literature was conducted. The systematic search and selection process led to a sample of 166 publications. During the review phase, the conceptual contributions to the simplification discourse were evaluated. A dataset of 163 entries was created, listing the conceptual contributions to the simplification debate. An empirically grounded analysis led to the generative development of a systematisation of simplifications according to their underlying simplifying logic.

Results and discussion

Five simplifying logics were identified: exclusion, inventory data substitution, qualitative expert judgment, standardisation and automation. Together, these simplifying logics inform 13 simplification strategies. The identified logics represent approaches to handle the complexities of product systems and expectations of the users of LCA results with the resources available to the analyst. Each simplification strategy is discussed with regard to its main applications and challenges.

Conclusions

This paper provides a first systematisation of the different simplification logics frequently applied in LCA since the original streamlining discussion. The presented terminology can help making communication about simplification more explicit and transparent, thus important for the credibility of LCA. Despite the pervasiveness of simplification in LCA, there is a relative lack of research on simplification per se, making further research describing simplification as a practice and analysing simplifications methodologically desirable.

     

Energy simulation and LCA for macro-scale analysis of eco-innovations in the housing stock

Purpose

Energy consumption of buildings is one of the major drivers of environmental impacts. Life cycle assessment (LCA) may support the assessment of burdens and benefits associated to eco-innovations aiming at reducing these environmental impacts. Energy efficiency policies however typically focus on the meso- or macro-scale, while interventions are typically taken at the micro-scale. This paper presents an approach that bridges this gap by using the results of energy simulations and LCA studies at the building level to estimate the effect of micro-scale eco-innovations on the macro-scale, i.e. the housing stock in Europe.

Methods

LCA and dynamic energy simulations are integrated to accurately assess the life cycle environmental burdens and benefits of eco-innovation measures at the building level. This allows quantitatively assessing the effectiveness of these measures to lower the energy use and environmental impact of buildings. The analysis at this micro-scale focuses on 24 representative residential buildings within the EU. For the upscaling to the EU housing stock, a hybrid approach is used. The results of the micro-scale analysis are upscaled to the EU housing stock scale by adopting the eco-innovation measures to (part of) the EU building stock (bottom–up approach) and extrapolating the relative impact reduction obtained for the reference buildings to the baseline stock model. The reference buildings in the baseline stock model have been developed by European Commission-Joint Research Centre based on a statistical analysis (top–down approach) of the European housing stock. The method is used to evaluate five scenarios covering various aspects: building components (building envelope insulation), technical installations (renewable energy), user behaviour (night setback of the setpoint temperature), and a combined scenario.

Results and discussion

Results show that the proposed combination of bottom–up and top–down approaches allow accurately assessing the impact of eco-innovation measures at the macro-scale. The results indicate that a combination of policy measures is necessary to lower the environmental impacts of the building stock to a significative extent.

Conclusions

Interventions addressing energy efficiency at building level may lead to the need of a trade-off between resource efficiency and environmental impacts. LCA integrated with dynamic energy simulation may help unveiling the potential improvements and burdens associated to eco-innovations.

     

Benchmarking environmental performance of electric insulator supply chain in India using life cycle assessment

Purpose

This study aims at finding the environmental impacts generated by an electric disk insulator supply chain, used for the distribution of electricity by an open wire system, through a case study. This study also aims at benchmarking the environmental impacts of an electric insulator manufacturing process by taking ideal condition of zero waste as reference.

Methods

Cradle-to-grave life cycle assessment (LCA) has been carried out by following the guidelines provided in ISO 14040 series standards and using Umberto NXT software. ReCiPe endpoint and ReCiPe midpoint impact assessment methodologies have been used to calculate environmental impacts under various categories. The primary data has been collected from a medium-scale manufacturer of electric disk insulators located at Bikaner in north-west India. The secondary data has been taken from ecoinvent 3.0 database and literature. The environmental impacts using endpoint assessment (ecosystem quality, human health, and resources) and midpoint assessment (climate change, fossil depletion, human toxicity, metal depletion, ozone depletion, terrestrial acidification, and water depletion) categories have been computed. Finally, the results are compared and benchmarked against the ideal zero waste condition using three different production scenarios. The limitation of this study is that the data has been collected only from one manufacturer and its supply chain.

Results and discussion

It has been found that the use of steel, electricity, and fuel; transportation of product; and disposal of water generate high environmental impacts in the supply chain. It has also been found that in the electric disk insulator supply chain, the raw material extraction phase has the highest environmental impacts followed by manufacturing, disposal, transportation, and installation phases. This study has also found that benchmark scenario “B” (zero waste condition) is environmentally more efficient in comparison to scenario “A” (actual recycling condition) and scenario “C” (maximum waste condition).

Conclusions

This study has identified that raw materials, resources, and processes in the supply chain of an electric disk insulator manufacturing unit are responsible for the environmental damage. The various manufacturing processes and installation of the electric disk insulators are similar for all manufacturers except the machinery efficiency and the generated waste. This study provides environmental impacts associated with an electric disk insulator manufacturing process under zero waste or ideal conditions (scenario B). These results are used as a benchmark to compare environmental performance of electric disk insulator supply chain operating under actual conditions.

     

Reviewing the potential for including habitat fragmentation to improve life cycle impact assessments for land use impacts on biodiversity

Purpose

The biosphere is progressively subjected to a variety of pressures resulting from anthropogenic activities. Habitat conversion, resulting from anthropogenic land use, is considered the dominant factor driving terrestrial biodiversity loss. Hence, adequate modelling of land use impacts on biodiversity in decision-support tools, like life cycle assessment (LCA), is a priority. State-of-the-art life cycle impact assessment (LCIA) characterisation models for land use impacts on biodiversity translate natural habitat transformation and occupation into biodiversity impacts. However, the currently available models predominantly focus on total habitat loss and ignore the spatial configuration of the landscape. That is, habitat fragmentation effects are ignored in current LCIAs with the exception of one recently developed method.

Methods

Here, we review how habitat fragmentation may affect biodiversity. In addition, we investigate how land use impacts on biodiversity are currently modelled in LCIA and how missing fragmentation impacts can influence the LCIA model results. Finally, we discuss fragmentation literature to evaluate possible methods to include habitat fragmentation into advanced characterisation models.

Results and discussion

We found support in available ecological literature for the notion that habitat fragmentation is a relevant factor when assessing biodiversity loss. Moreover, there are models that capture fragmentation effects on biodiversity that have the potential to be incorporated into current LCIA characterisation models.

Conclusions and recommendations

To enhance the credibility of LCA biodiversity assessments, we suggest that available fragmentation models are adapted, expanded and subsequently incorporated into advanced LCIA characterisation models and promote further efforts to capture the remaining fragmentation effects in LCIA characterisation models.

     

Performance and environmental accounting of nutrient cycling models to estimate nitrogen emissions in agriculture and their sensitivity in life cycle assessment

Purpose

Several models are available in the literature to estimate agricultural emissions. From life cycle assessment (LCA) perspective, there is no standardized procedure for estimating emissions of nitrogen or other nutrients. This article aims to compare four agricultural models (PEF, SALCA, Daisy and Animo) with different complexity levels and test their suitability and sensitivity in LCA.

Methods

Required input data, obtained outputs, and main characteristics of the models are presented. Then, the performance of the models was evaluated according to their potential feasibility to be used in estimating nitrogen emissions in LCA using an adapted version of the criteria proposed by the United Nations Framework Convention on Climate Change (UNFCCC), and other relevant studies, to judge their suitability in LCA. Finally, nitrogen emissions from a case study of irrigated maize in Spain were estimated using the selected models and were tested in a full LCA to characterize the impacts.

Results and discussion

According to the set of criteria, the models scored, from best to worst: Daisy (77%), SALCA (74%), Animo (72%) and PEF (70%), being Daisy the most suitable model to LCA framework. Regarding the case study, the estimated emissions agreed to literature data for the irrigated corn crop in Spain and the Mediterranean, except N₂O emissions. The impact characterization showed differences of up to 56% for the most relevant impact categories when considering nitrogen emissions. Additionally, an overview of the models used to estimate nitrogen emissions in LCA studies showed that many models have been used, but not always in a suitable or justified manner.

Conclusions

Although mechanistic models are more laborious, mainly due to the amount of input data required, this study shows that Daisy could be a suitable model to estimate emissions when fertilizer application is relevant for the environmental study. In addition, and due to LCA urgently needing a solid methodology to estimate nitrogen emissions, mechanistic models such as Daisy could be used to estimate default values for different archetype scenarios.

     

The environmental impact of packaging in food supply chains—does life cycle assessment of food provide the full picture?

Purpose

Due to the urgency and the magnitude of the environmental problems caused by food supply chains, it is important that the recommendations for packaging improvements given in life cycle assessment (LCA) studies of food rest on a balanced consideration of all relevant environmental impacts of packaging. The purpose of this article is to analyse the extent to which food LCAs include the indirect environmental impact of packaging in parallel to its direct impact. While the direct environmental impact of food packaging is the impact caused by packaging materials’ production and end-of-life, its indirect environmental impact is caused by its influence on the food product’s life cycle, e.g. by its influence on food waste and on logistical efficiency.

Methods

The article presents a review of 32 food LCAs published in peer-reviewed scientific journals over the last decade. The steps of the food product’s life cycle that contribute to the direct and indirect environmental impacts of packaging provide the overall structure of the analytical framework used for the review. Three aspects in the selected food LCAs were analysed: (1) the defined scope of the LCAs, (2) the sensitivity and/or scenario analyses and (3) the conclusions and recommendations.

Results and discussion

While in packaging LCA literature, there is a trend towards a more systematic consideration of the indirect environmental impact of packaging, it is unclear how food LCAs handle this aspect. The results of the review show that the choices regarding scope and sensitivities/scenarios made in food LCAs and their conclusions about packaging focus on the direct environmental impact of packaging. While it is clear that not all food LCAs need to analyse packaging in detail, this article identifies opportunities to increase the validity of packaging-related conclusions in food LCAs and provides specific recommendations for packaging-related food LCA methodology.

Conclusions

Overall, we conclude that the indirect environmental impact of packaging is insufficiently considered in current food LCA practice. Based on these results, this article calls for a more systematic consideration of the indirect environmental impact of packaging in future food LCAs. In addition, it identifies a need for more packaging research that can provide the empirical data that many food LCA practitioners currently lack. In particular, LCA practitioners would benefit if there were more knowledge and data available about the influence of certain packaging characteristics (e.g. shape, weight and type of material) on consumer behaviour.

     

Comparative life cycle assessment of car disc brake systems—case study results and method discussion about comparative LCAs

Purpose

Two life cycle assessment (LCA) studies comparing a new low-particulate-matter-emission disc brake and a reference disc brake were presented. The purpose was to identify the difference in potential environmental impacts due to a material change in the new disc brake parts. Additionally, the validity was investigated for the simplification method of omitting identical parts in comparative LCA. This was done by comparing the results between the simplified and the full LCA model.

Methods

The two disc brakes, new disc brake and reference disc brake, were assessed according to the LCA ISO standards. The ReCiPe 2016 Midpoint (hierarchist) impact assessment method was chosen. Simplifying a comparative LCA is possible, all identical parts can be omitted, and only the ones that differ need to be assessed. In this paper, this simplification was called comparative LCA with an omission of identical parts.

Results and discussion

The comparative impacts were analysed over seventeen impact categories. The new disc brake alternative used more resources during the manufacture of one disc compared to the reference disc brake alternative. The shorter life length of the reference disc demanded a higher number of spare part discs to fulfil the same functional unit, but this impact was reduced due to material recycling. The new disc brake impacts were connected primarily to the coating and secondly to the pad manufacture and materials. The validity of the simplification method was investigated by comparing the results of the two LCA models. The impact differences were identical independent of the LCA model, and the same significant impact categories could be identified. Hence, the purpose of the study could be fulfilled, and the simplification was valid.

Conclusions

Both LCA models, simplified and full, revealed that the new disc brake had limited environmental advantages. The omission of identical parts made it more challenging to determine if an impact was significant or insignificant. The simplification seemed to be reasonable.

     

An effect factor approach for quantifying the impact of plastic additives on aquatic biota in life cycle assessment

Purpose

Plastic pervades now almost every aspect of our daily lives, but this prosperity has led to an increasing amount of plastic debris, which is now widespread in the oceans and represents a serious threat to biota. However, there is a general lack of consideration regarding marine plastic impacts in life cycle assessment (LCA). This paper presents a preliminary approach to facilitate the characterization of chemical impacts related to marine plastic within the LCA framework.

Methods

A literature review was carried out first to summarize the current state of research on the impact assessment of marine plastic. In recent years, efforts have been made to develop LCA-compliant indicators and models that address the impact of marine littering, entanglement, and ingestion. The toxicity of plastic additives to marine biota is currently a less understood impact pathway and also the focus of this study. Relevant ecotoxicity data were collected from scientific literature for a subsequent additive-specific effect factor (EF) development, which was conducted based on the USEtox approach. Extrapolation factors used for the data conversion were also extracted from reliable sources.

Results and discussion

EFs were calculated for six commonly used additives to quantify their toxicity impacts on aquatic species. Triclosan shows an extremely high level of toxicity, while bisphenol A and bisphenol F are considered less toxic according to the results. Apart from additive-specific EFs, a generic EF was also generated, along with the species sensitivity distribution (SSD) illustrating the gathered data used to calculate this EF. Further ecotoxicity data are expected to expand the coverage of additives and species for deriving more robust EFs. In addition, a better understanding of the interactive effect between polymers and additives needs to be developed.

Conclusions

This preliminary work provides a first step towards including the impact of plastic-associated chemicals in LCA. Although the toxicity of different additives to aquatic biota may vary significantly, it is recommended to consider additives within the impact assessment of marine plastic. The generic EF can be used, together with a future EF for adsorbed environmental pollutants, to fill a gap in the characterization of plastic-related impacts in LCA.

     

Life cycle environmental and economic assessment of industrial symbiosis networks: a review of the past decade of models and computational methods through a multi-level analysis lens

Purpose

Industrial symbiosis network (ISN) facilitation tools seek to holistically evaluate the environmental and economic performance of ISNs through life cycle assessment (LCA) and life cycle costing (LCC). ISNs have many stakeholders with diverse interests in the LCA and LCC results thus requiring multi-level analysis. The objective of this review was to examine the state-of-the-art methodologies used in LCAs and LCCs of ISNs and understand how multi-level analysis can be conducted.

Methods

The systematic literature review methodology was applied to develop a corpus of peer-reviewed LCA and LCC studies of ISNs published between 2010 and 2019 without any geographic boundary. Abstracts were reviewed to shortlist studies that conducted an LCA or LCC of an ISN with numerical results. LCA and LCC methodologies used in the shortlisted studies were collected and categorized. Each methodology was examined to understand how the foreground and background systems are represented, how waste-to-resource exchanges are analyzed, and how the results can be computed at the network, entity, and flow levels.

Results and discussion

The review yielded 42 LCA studies and 11 LCC studies of ISNs that used eight different methodologies. Process-based LCA was used in 71% of the LCA studies, whereas tiered hybrid LCA was used in 14% of the studies. Waste-to-resource exchanges in ISN scenarios were represented either through process analysis or as a black box. Fewer LCC studies that evaluate the economic performance of ISNs exist compared with LCA studies. Economic studies often evaluated financial feasibility, net present value, profitability, or payback period of specific waste-to-resource exchanges or the network overall.

Conclusions

The insights derived from this review chart future areas of research in multi-level modeling and analysis of the life cycle environmental and economic performance of ISNs. To improve the model construction and analysis process, research should be explored in developing a methodology for constructing a single model that represents multiple entities linked together by waste-to-resource exchanges and can provide LCA and LCC results for different stakeholder perspectives. The lack of LCC studies of ISNs merits the need for more research in this area at both the network and entity levels to quantify potential economic trade-offs between stakeholders. Developing a methodology for unified LCA and LCC modeling and analysis of ISNs can help ISN facilitation tool developers conduct simultaneous life cycle environmental and economic analysis of the potential symbiosis connections identified and how they contribute to the overall network.

     

A meta-analysis of the life cycle greenhouse gas balances of microalgae biodiesel

Purpose

Microalgae biodiesel has attracted considerable attention as a potential substitute for fossil fuels and biodiesel from food crops. Nevertheless, its reported climate impacts in the scientific literature vary significantly. This article describes and synthesizes the range of results found in the life cycle assessment (LCA) literature regarding microalgae biodiesel studies to investigate whether particular parameters, e.g. technologies, were associated with higher or lower greenhouse gas (GHG) emissions so that a best practice can be inferred from currently available LCA data and thereby recommended.

Methods

A systematic literature review and meta-regression analysis (MRA) of 36 LCA studies that report on the GHG emissions of microalgae biodiesel was conducted. An assessment of key aspects, including modelling choices and technologies, was performed. Furthermore, MRA models were formulated considering several variables of interest describing both technical and modelling choices to identify the main causes for the variability in GHG emissions per MJ of biodiesel. Variables chosen include: microalgae species; culture medium; cultivation system; source of CO₂; extraction technology; conversion technology; system boundary; geographical scope; inclusion or exclusion of capital goods; and how multifunctionality was handled.

Results and discussion

The reviewed studies altogether reported 308 results ranging from ?0.7 to 3.8 kg CO₂ eq. MJ^?1_biodiesel, portraying 19 different system configurations. Despite the comprehensive range of variables assessed, the models generated could not plausibly explain that the variability in GHG emissions depends either on the technologies considered or on the methodological choices adopted. However, the following relationships could be observed: location in Europe and high oil productivity were associated with lower emissions, whilst dry extraction should be avoided for leading to higher GHG emissions, on average.

Conclusions

There is a large degree of variability within the technologies considered, as well as the methodological choices adopted, so that no robust conclusions could be drawn from the MRA. Notwithstanding, average GHG emissions reported were more than twice as high as fossil diesel and, while there are some studies showing large benefits, none of the various algae technologies performed consistently better than fossil diesel, questioning the climate-mitigation potential of microalgae biodiesel.

     

Environmental impacts of European trade: interpreting results of process-based LCA and environmentally extended input–output analysis towards hotspot identification

Purpose

Trade is increasingly considered a significant contributor to environmental impacts. The assessment of the impacts of trade is usually performed via environmentally extended input–output analysis (EEIOA). However, process-based life cycle assessment (LCA) applied to traded goods allows increasing the granularity of the analysis and may be essential to unveil specific impacts due to traded products.

Methods

This study assesses the environmental impacts of the European trade, considering two modelling approaches: respectively EEIOA, using EXIOBASE 3 as supporting database, and process-based LCA. The interpretation of the results is pivotal to improve the robustness of the assessment and the identification of hotspots. The hotspot identification focuses on temporal trends and on the contribution of products and substances to the overall impacts. The inventories of elementary flows associated with EU trade, for the period 2000–2010, have been characterized considering 14 impact categories according to the Environmental Footprint (EF2017) Life Cycle Impact Assessment method.

Results and discussion

The two modelling approaches converge in highlighting that in the period 2000–2010: (i) EU was a net importer of environmental impacts; (ii) impacts of EU trade and EU trade balance (impacts of imports minus impacts of exports) were increasing over time, regarding most impact categories under study; and (iii) similar manufactured products were the main contributors to the impacts of exports from EU, regarding most impact categories. However, some results are discrepant: (i) larger impacts are obtained from IO analysis than from process-based LCA, regarding most impact categories, (ii) a different set of most contributing products is identified by the two approaches in the case of imports, and (iii) large differences in the contributions of substances are observed regarding resource use, toxicity, and ecotoxicity indicators.

Conclusions

The interpretation step is crucial to unveil the main hotspots, encompassing a comparison of the differences between the two methodologies, the assumptions, the data coverage and sources, the completeness of inventory as basis for impact assessment. The main driver for the observed divergences is identified to be the differences in the impact intensities of goods, both induced by inherent properties of the IO and life cycle inventory databases and by some of this study’s modelling choices. The combination of IO analysis and process-based LCA in a hybrid framework, as performed in other studies but generally not at the macro-scale of the full trade of a country or region, appears a potential important perspective to refine such an assessment in the future.

     

The environmental consequences of a change in Australian cotton lint production

Purpose

Changes in the production of Australian cotton lint are expected to have a direct environmental impact, as well as indirect impacts related to co-product substitution and induced changes in crop production. The environmental consequences of a 50% expansion or contraction in production were compared to Australian cotton production’s current environmental footprint. Both were then assessed to investigate whether current impacts are suitable for predicting the environmental impact of a change in demand for cotton lint.

Methods

A consequential life cycle assessment (LCA) model of Australian cotton lint production (cradle-to-gin gate) was developed using plausible scenarios regarding domestic regions and technologies affected by changes in supply, with both expansion (additional cotton) and contraction (less cotton) being modelled. Modelling accounted for direct impacts from cotton production and indirect impacts associated with changes to cotton production, including co-product substitution and changes to related crops at regional and global scales. Impact categories assessed included climate change, fossil energy demand, freshwater consumption, water stress, marine and freshwater eutrophication, land occupation and land-use change.

Results and discussion

For both the expansion and contraction scenarios, the changes to climate change impacts (including iLUC) and water impacts were less than would be assumed from current production as determined using attributional LCA. However, the opposite was true for all other impact categories, indicating trade-offs across the impact categories. Climate change impacts under both scenarios were relatively minor because these were largely offset by iLUC. Similarly, under the contraction scenario, water impacts were dominated by indirect impacts associated with regional crops. A sensitivity analysis showed that the results were sufficiently robust to indicate the quantum of changes that could be expected.

Conclusions

A complex array of changes in technologies, production regions and related crops were required to model the environmental impacts of a gross change in cotton production. Australian cotton lint production provides an example of legislation constraining the direct water impacts of production, leading to a contrast between impacts estimated by attributional and consequential LCA. This model demonstrated that indirect products and processes are important contributors to the environmental impacts of Australian cotton lint.

     

Functional unit influence on building life cycle assessment

Purpose

The building sector is one of the most relevant sectors in terms of environmental impact. Different functional units (FUs) can be used in life cycle assessment (LCA) studies for a variety of purposes. This paper aimed to present different FUs used in the LCA of buildings and evaluate the influence of FU choice and setting in comparative studies.

Methods

As an example, we compared the “cradle to grave” environmental performance of four typical Brazilian residential buildings with different construction typologies, i.e., multi-dwelling and single dwelling, each with high and basic standards. We chose three types of FU for comparison: a dwelling with defined lifetime and occupancy parameters, an area of 1 m² of dwelling over a year period, and the accommodation of an occupant person of the dwelling over a day.

Results and discussion

The FU choice was found to bias the results considerably. As expected, the largest global warming indicator (GWi) values per dwelling unit and occupant were identified for the high standard dwellings. However, when measured per square meter, lower standard dwellings presented the largest GWi values. This was caused by the greater concentration of people per square meter in smaller area dwellings, resulting in larger water and energy consumption per square meter. The sensitivity analysis of FU variables such as lifetime and occupancy showed the GWi contribution of the infrastructure more relevant compared with the operation in high and basic standard dwellings. The definition of lifetime and occupancy parameters is key to avoid bias and to reduce uncertainty of the results when performing a comparison of dwelling environmental performances.

Conclusions

This paper highlights the need for adequate choice and setting of FU to support intended decision-making in LCA studies of the building sector. The use of at least two FUs presented a broader picture of building performance, helping to guide effective environmental optimization efforts from different approaches and levels of analysis. Information regarding space, time, and service dimensions should be either included in the FU setting or provided in the building LCA study to allow adjustment of the results for subsequent comparison.

     

On the number of Monte Carlo runs in comparative probabilistic LCA

Introduction

The Monte Carlo technique is widely used and recommended for including uncertainties LCA. Typically, 1000 or 10,000 runs are done, but a clear argument for that number is not available, and with the growing size of LCA databases, an excessively high number of runs may be a time-consuming thing. We therefore investigate if a large number of runs are useful, or if it might be unnecessary or even harmful.

Probability theory

We review the standard theory or probability distributions for describing stochastic variables, including the combination of different stochastic variables into a calculation. We also review the standard theory of inferential statistics for estimating a probability distribution, given a sample of values. For estimating the distribution of a function of probability distributions, two major techniques are available, analytical, applying probability theory and numerical, using Monte Carlo simulation. Because the analytical technique is often unavailable, the obvious way-out is Monte Carlo. However, we demonstrate and illustrate that it leads to overly precise conclusions on the values of estimated parameters, and to incorrect hypothesis tests.

Numerical illustration

We demonstrate the effect for two simple cases: one system in a stand-alone analysis and a comparative analysis of two alternative systems. Both cases illustrate that statistical hypotheses that should not be rejected in fact are rejected in a highly convincing way, thus pointing out a fundamental flaw.

Discussion and conclusions

Apart form the obvious recommendation to use larger samples for estimating input distributions, we suggest to restrict the number of Monte Carlo runs to a number not greater than the sample sizes used for the input parameters. As a final note, when the input parameters are not estimated using samples, but through a procedure, such as the popular pedigree approach, the Monte Carlo approach should not be used at all.

     

Model uncertainty analysis using data analytics for life-cycle assessment (LCA) applications

Purpose

Objective uncertainty quantification (UQ) of a product life-cycle assessment (LCA) is a critical step for decision-making. Environmental impacts can be measured directly or by using models. Underlying mathematical functions describe a model that approximate the environmental impacts during various LCA stages. In this study, three possible uncertainty sources of a mathematical model, i.e., input variability, model parameter (differentiate from input in this study), and model-form uncertainties, were investigated. A simple and easy to implement method is proposed to quantify each source.

Methods

Various data analytics methods were used to conduct a thorough model uncertainty analysis; (1) Interval analysis was used for input uncertainty quantification. A direct sampling using Monte Carlo (MC) simulation was used for interval analysis, and results were compared to that of indirect nonlinear optimization as an alternative approach. A machine learning surrogate model was developed to perform direct MC sampling as well as indirect nonlinear optimization. (2) A Bayesian inference was adopted to quantify parameter uncertainty. (3) A recently introduced model correction method based on orthogonal polynomial basis functions was used to evaluate the model-form uncertainty. The methods are applied to a pavement LCA to propagate uncertainties throughout an energy and global warming potential (GWP) estimation model; a case of a pavement section in Chicago metropolitan area was used.

Results and discussion

Results indicate that each uncertainty source contributes to the overall energy and GWP output of the LCA. Input uncertainty was shown to have significant impact on overall GWP output; for the example case study, GWP interval was around 50%. Parameter uncertainty results showed that an assumption of ±?10% uniform variation in the model parameter priors resulted in 28% variation in the GWP output. Model-form uncertainty had the lowest impact (less than 10% variation in the GWP). This is because the original energy model is relatively accurate in estimating the energy. However, sensitivity of the model-form uncertainty showed that even up to 180% variation in the results can be achieved due to lower original model accuracies.

Conclusions

Investigating each uncertainty source of the model indicated the importance of the accurate characterization, propagation, and quantification of uncertainty. The outcome of this study proposed independent and relatively easy to implement methods that provide robust grounds for objective model uncertainty analysis for LCA applications. Assumptions on inputs, parameter distributions, and model form need to be justified. Input uncertainty plays a key role in overall pavement LCA output. The proposed model correction method as well as interval analysis were relatively easy to implement. Research is still needed to develop a more generic and simplified MCMC simulation procedure that is fast to implement.

     

Life cycle assessment of electrical and thermal energy systems for commercial buildings

Background, Aim and Scope The objective of this life cycle assessment (LCA) study is to develop LCA models for energy systems in order to assess the potential environmental impacts that might result from meeting energy demands in buildings. The scope of the study includes LCA models of the average electricity generation mix in the USA, a natural gas combined cycle (NGCC) power plant, a solid oxide fuel cell (SOFC) cogeneration system; a microturbine (MT) cogeneration system; an internal combustion engine (ICE) cogeneration system; and a gas boiler. Methods LCA is used to model energy systems and obtain the life cycle environmental indicators that might result when these systems are used to generate a unit energy output. The intended use of the LCA analysis is to investigate the operational characteristics of these systems while considering their potential environmental impacts to improve building design using a mixed integer linear programming (MILP) optimization model. Results The environmental impact categories chosen to assess the performance of the energy systems are global warming potential (GWP), acidification potential (AP), tropospheric ozone precursor potential (TOPP), and primary energy consumption (PE). These factors are obtained for the average electricity generation mix, the NGCC, the gas boiler, as well as for the cogeneration systems at different part load operation. The contribution of the major emissions to the emission factors is discussed. Discussion The analysis of the life cycle impact categories indicates that the electrical to thermal energy production ratio has a direct influence on the value of the life cycle PE consumption factors. Energy systems with high electrical to thermal ratios (such as the SOFC cogeneration systems and the NGCC power plant) have low PE consumption factors, whereas those with low electrical to thermal ratios (such as the MT cogeneration system) have high PE consumption factors. In the case of GWP, the values of the life cycle GWP obtained from the energy systems do not only depend on the efficiencies of the systems but also on the origins of emissions contributing to GWP. When evaluating the life cycle AP and TOPP, the types of fuel as well as the combustion characteristics of the energy systems are the main factors that influence the values of AP and TOPP. Conclusions An LCA study is performed to eraluate the life cycle emission factors of energy systems that can be used to meet the energy demand of buildings. Cogeneration systems produce utilizable thermal energy when used to meet a certain electrical demand which can make them an attractive alternative to conventional systems. The life cycle GWP, AP, TOPP and PE consumption factors are obtained for utility systems as well as cogeneration systems at different part load operation levels for the production of one kWh of energy output. Recommendations and Perspectives Although the emission factors vary for the different energy systems, they are not the only factors that influence the selection of the optimal system for building operations. The total efficiencies of the system play a significant part in the selection of the desirable technology. Other factors, such as the demand characteristics of a particular building, influence the selection of energy systems. The emission factors obtained from this LCA study are used as coefficients of decision variables in the formulation of an MILP to optimize the selection of energy systems based on environmental criteria by taking into consideration the system efficiencies, emission characteristics, part load operation, and building energy demands. Therefore, the emission factors should not be regarded as the only criteria for choosing the technology that could result in lower environmental impacts, but rather one of several factors that determine the selection of the optimum energy system. ESS-Submission Editor: Arpad Horvath (horvath@ce.berkeley.edu)