i-Dave: Data-Knowledge Integration to improve the reliability of LCA projets in the enterprise of the future

Context 

Despite the growing interest in life cycle assessment (LCA) and product carbon footprint (PCF) methods and tools, their application faces three challenges:

• data collection
• the selection of relevant cost centers
• the robustness of reference databases.

These challenges are very difficult because experts mastering the industrial process do not necessarily have the LCA/PCF culture. In addition, validating and aggregating all factors is a very complex activity because the data is collected from very heterogeneous sources, in varied business contexts and life phases. Current collection methods remain mostly manual, based on questionnaires, and dedicated tools are often disconnected from the company's overall digital chain. These issues are more critical in the case of long-life systems. Once the studies are completed, another challenge concerns the exploitation of the results for the prediction of carbon trajectories. 

Project Objectives

The objective of the i-DAVE project is to propose a knowledge-based and AI-based framework, interoperable in a PLM approach, for the reliability of LCA/ECP studies. This is a dual decision-making aid: 1) Upstream of LCA studies to make the input data robust and configure the LCA/PCF study parameters. 2) Downstream for the exploitation of study results in the definition of the best action plans to reduce the PCF footprint.

Expected Scientific Advancement

A response to the above challenges is expected through three complementary solutions:

• A knowledge base defined in the form of an ontology of the LCA/PCF domain and coupled with an inference engine, implementing business rules to meet specific assistance needs such as the choice of scope and environmental cost centers, etc. It also involves supporting process traceability and the characterization of systems in terms of LCA/PCF factors.
• Intelligent connectors to ensure the interoperability of LCA/PCF tools with the different modules of the company's digital chain for extracting data from heterogeneous sources. The Product Lifecycle Management (PLM) approach will be used to ensure the cross-functional integration of all types of information systems and databases useful for LCA/PCF studies.
• A tool for managing low-carbon strategies in companies coupled with innovative dashboards. This tool contains in particular algorithms for predicting future trajectories based on history, and classification algorithms to help choose the best operational alternatives generating the minimum carbon emission impact. This module is based on a behavioral model containing useful performance indicators and cause-effect relationships between process parameters and environmental impacts.