Transformer Based Prediction Model for Carbon Emissions from Cement Clinker Burning

The cement industry, as one of the main sources of carbon emissions in China, accounts for approximately 9.6% of the country's total carbon emissions. Among them, the process of burning cement clinker is the most carbon intensive and concentrated link in the entire cement production process. Therefore, accurately predicting the carbon emissions during the clinker burning process is of great significance for achieving the national dual carbon strategy goals. In terms of time series data processing during the clinker burning process, traditional carbon emission prediction methods cannot effectively capture complex dependency relationships and nonlinear interactions of multidimensional features, and have poor adaptability to different production environments. This article proposes a carbon emission prediction model for cement clinker burning based on Transformer architecture. Firstly, a Transformer model was designed to capture long-term dependencies on multidimensional production data during the clinker burning process. Then, the self attention mechanism is utilized to extract the relationships between various features at different time steps. Finally, verify through actual production data from the cement plant. The experimental results show that the model proposed in this paper has good performance and high practicality.