Deep Learning‐Based Carbon Emission Monitoring and Prediction Study for Bridge Reconstruction and Expansion Projects

Highway reconstruction and expansion projects involve complex processes, including the demolition of old structures and the construction of new ones, with personnel and machinery contributing significantly to carbon emissions. Given the complexity of monitoring emissions, particularly for mixed old–new structures like bridges and interchanges, accurate emissions prediction is challenging. This study leverages a self‐encoding neural network model to introduce an artificial intelligence (AI)‐driven framework for carbon emission prediction in construction, applied to a bridge reconstruction project. Autoencoder network exhibits significant advantages in terms of small prediction error and short training time, making it suitable for carbon emission prediction tasks in the scenario of bridge renovation and expansion. This approach enables a clearer understanding of carbon emission sources and process‐specific contributions in bridge expansion projects, supporting effective carbon reduction strategies.