Abstract:

Objective In the context of the 2030 carbon peak target, the green transformation of the textile industry is the only way to achieve the carbon neutrality. However, the flow and distribution of carbon footprints are extremely complex, making green carbon tracing and management for the whole life cycle very difficult. This study aims to address the complex issue of tracing and managing the lifecycle carbon footprint of wool blended fabrics in the green transformation of the textile industry. The specific goal is to develop a carbon emission knowledge graph (simplified ″carbon knowledge graph″) model for the production process of wool blended fabrics, quantify carbon emissions, and apply this model to provide guidance and decision-making support for the industry's carbon reduction efforts.

Method The study employs industrial intelligent technologies such as process mining and knowledge graph. Initially, a carbon emission quantification model is developed considering four dimensions: energy flow, material flow, personnel flow, and carbon flow throughout the various stages of wool blended fabric production, facilitating the modeling of local-level carbon emission management. Subsequently, a carbon emission knowledge graph modeling approach is introduced on the foundation of the quantification model, which allows for the calculation of carbon emissions in fabric production processes. In the end, directly-follows graphs are produced to model the global-level carbon emission distribution throughout the fabric production process, accomplishing bidirectional modeling of both global-level and local-level carbon emissions.

Results Using real data as an example, this paper constructs a carbon knowledge graph model for the production process of wool blended fabrics. The model comprises 8 045 nodes, 15 717 edges, 264 order nodes, and 779 process steps. The results show that the total carbon emissions for weaving in the current 264 orders amount to 385 010.707 kg CO₂, and the total carbon emissions for dyeing amount to 2 385 362.262 kg CO₂. Therefore, analyzing the energy consumption, material usage, and personnel carbon emissions of the dyeing process in the production of wool blended fabrics is crucial for understanding its carbon footprint. Furthermore, a specific order case involving the dyeing process of a fabric blend of 80% wool, 12% nylon, and 8% other fibers was selected for analysis. The experiment shows that the operation of production equipment is the main source of electrical energy carbon emissions in the actual production process, accounting for 93.92% of the total, or 1 634.263 kg CO₂. Moreover, the carbon emissions from the air compressor and lighting equipment were found to be 31.720 kg CO₂ and 24.494 kg CO₂, respectively. The results indicate that optimizing the carbon emissions of operating production equipment is key to reducing overall carbon emissions in the production process. In brief, the method proposed in this study has shown high effectiveness and feasibility in managing carbon emissions throughout the product lifecycle, providing an important reference for the field of low-carbon industrial manufacturing.

Conclusion This study successfully proposes a method for carbon knowledge graph modeling, carbon emission quantification, and application for the production process of wool blended fabrics. The effectiveness of the model has been verified through real-world cases, and it provides guidance and decision-making support for its application in industry carbon emission reduction. The research approach is expected to be applied to other textile production processes, aiding in the green transformation of China's textile industry.

Key words: wool blended fabric production, carbon emission knowledge graph, carbon knowledge graph, industrial intelligence