Damage assessment of modular integrated construction during transport and assembly using a hybrid CNN–gated recurrent unit model

Abstract

Modular integrated construction (MiC) offers improved sustainability and automation. Nevertheless, its performance is impeded by extensive logistics operations, including multimode transportation, recurring loading-unloading, stacking, and assembly. Such rigorous operations may cause inadvertent underlying damage to module structure, leading to supply chain disruptions, safety hazards and structural deterioration. A robust real-time damage prediction can mitigate such issues. Thus, this paper develops a hybrid deep learning model for MiC module damage prediction, integrating convolutional and sequential neural networks. The developed hybrid CNN-GRU model establishes correlations between module motion during logistic operations and corresponding structural variations. The multivariate training and testing data of MiC operations is collected using a multi-sensing IoT system. The model is validated for damage scenarios to assess damage level and location, demonstrating a 96 % (R2) accuracy. The model provides practical considerations through a robust, automated damage prediction to enhance the safety, productivity and proactive maintenance of MiC modules.