Scholarlink

Purpose <p>Interoperability among Life Cycle Inventory (LCI) formats remains a persistent challenge, even for widely used standards such as EcoSpold 2 and ILCD, limiting data exchange and reuse. Converters are commonly employed to bridge these gaps, yet existing tools are mostly commercial, tied to specific software ecosystems and lack robust open-source alternatives. To address this limitation, we introduce Lavoisier, an open-source conversion framework designed to enable reliable transformation of LCI datasets across different formats.</p> Methods <p>Lavoisier is designed with three modules: the input, the mapping of fields (including conversion logic) and the output, all of which accept new entries for different formats in a modular way. As a case study, we implemented in Lavoisier a bidirectional conversion flow between EcoSpold 2 and ILCD. This was achieved by leveraging existing documentation and gathering understanding of the structural and semantic characteristics of both formats. The primary goal in the case study was to ensure minimal data loss while respecting the constraints and modelling principles of each format. Conversions were assessed and classified into subclasses based on type (direct, indirect, incomplete, structural and not converted) and functional relevance of data (computational data, interpretative metadata, complementary metadata and not used).</p> Results and discussion <p>The results show that Lavoisier successfully converts the majority of computational data and interpretative metadata between EcoSpold 2 and ILCD, with conversion results and limitations logged. Direct and indirect conversions dominate across both pathways, with most losses concentrated in elements inherently unsupported by the target format (such as EcoSpold 2’s pedigree coefficients or ILCD’s richer source and contact structures). Tests using 115 EcoSpold 2 and 116 ILCD datasets showed high accuracy in unit handling, equation translation, uncertainty processing and intermediate flow reconstruction. The largest limitations stem from format-level constraints, particularly rigid reference lists for elementary flows, units and properties, which prevented full conversion. Overall, results indicate strong conversion performance constrained only by structural differences external to Lavoisier.</p> Conclusion <p>Lavoisier enables a translation of LCI datasets between EcoSpold 2 and ILCD that preserves nearly all commonly used quantitative and metadata information. Additionally, logs communicate limitations in a transparent way. However, full interoperability remains limited by intrinsic format (and methodology) differences such as incompatible reference lists, missing provider fields and non-extensible metadata structures. This is a barrier that converters alone cannot overcome. Even so, Lavoisier substantially reduces manual rework and has the potential to increase data reusability.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Bridging the interoperability gap in life cycle assessment: the lavoisier conversion framework and implementation for EcoSpold 2 and ILCD formats

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月16日

查看全文
  • José Paulo Pereira das Dores Pilz Savioli,
  • Tiago Emmanuel Nunes Braga,
  • Thiago Oliveira Rodrigues,
  • Silvana Aparecida Borsetti Gregorio Vidotti,
  • Cássia Maria Lie Ugaya

摘要

Purpose

Interoperability among Life Cycle Inventory (LCI) formats remains a persistent challenge, even for widely used standards such as EcoSpold 2 and ILCD, limiting data exchange and reuse. Converters are commonly employed to bridge these gaps, yet existing tools are mostly commercial, tied to specific software ecosystems and lack robust open-source alternatives. To address this limitation, we introduce Lavoisier, an open-source conversion framework designed to enable reliable transformation of LCI datasets across different formats.

Methods

Lavoisier is designed with three modules: the input, the mapping of fields (including conversion logic) and the output, all of which accept new entries for different formats in a modular way. As a case study, we implemented in Lavoisier a bidirectional conversion flow between EcoSpold 2 and ILCD. This was achieved by leveraging existing documentation and gathering understanding of the structural and semantic characteristics of both formats. The primary goal in the case study was to ensure minimal data loss while respecting the constraints and modelling principles of each format. Conversions were assessed and classified into subclasses based on type (direct, indirect, incomplete, structural and not converted) and functional relevance of data (computational data, interpretative metadata, complementary metadata and not used).

Results and discussion

The results show that Lavoisier successfully converts the majority of computational data and interpretative metadata between EcoSpold 2 and ILCD, with conversion results and limitations logged. Direct and indirect conversions dominate across both pathways, with most losses concentrated in elements inherently unsupported by the target format (such as EcoSpold 2’s pedigree coefficients or ILCD’s richer source and contact structures). Tests using 115 EcoSpold 2 and 116 ILCD datasets showed high accuracy in unit handling, equation translation, uncertainty processing and intermediate flow reconstruction. The largest limitations stem from format-level constraints, particularly rigid reference lists for elementary flows, units and properties, which prevented full conversion. Overall, results indicate strong conversion performance constrained only by structural differences external to Lavoisier.

Conclusion

Lavoisier enables a translation of LCI datasets between EcoSpold 2 and ILCD that preserves nearly all commonly used quantitative and metadata information. Additionally, logs communicate limitations in a transparent way. However, full interoperability remains limited by intrinsic format (and methodology) differences such as incompatible reference lists, missing provider fields and non-extensible metadata structures. This is a barrier that converters alone cannot overcome. Even so, Lavoisier substantially reduces manual rework and has the potential to increase data reusability.