Abstract: To achieve accurate measurement of carbon emissions throughout the transformer life cycle and address the issue of missing data, a carbon emission measurement model is constructed. The sparrow search algorithm is introduced to optimize the long short-term memory model(SSA-LSTM) and compensate for missing data in key stages. A typical transformer is used to validate the model, and the results show that the model can accurately calculate carbon emissions across each stage. Among them, the raw material stage has the highest emissions (185.81 tCOe), followed by the operation stage (11.075 tCOe), while the retirement and recycling stage can achieve a carbon reduction of approximately 45 tCOe. The SSA-LSTM model used for imputing missing data outperforms traditional methods, achieving a mean absolute error(MAE) of 0.44 tCOe, a root mean square error(RMSE) of 0.53 tCOe, and a mean absolute percentage error(MAPE) of 2.35%. This study provides a methodological basis for the accurate accounting of carbon emissions across the transformer supply chain life cycle and the imputation of missing data in key links.

Key words: life cycle assessment, accurate carbon emission measurement, carbon emission factor method, sparrow search algorithm, long short-term memory network