<![CDATA[The increasing complexity of urban infrastructure necessitates more efficient and proactive maintenance strategies. Traditional maintenance approaches often rely on reactive measures, leading to increased costs, unplanned downtime, and potential structural failures. The emergence of Artificial Intelligence (AI)-driven Digital Twin technology offers a promising solution by enabling predictive maintenance through real-time monitoring and advanced analytics. This study aimed to evaluate the effectiveness of AI-driven Digital Twin systems in enhancing predictive maintenance for urban infrastructure. A qualitative case study methodology was employed, analyzing multiple infrastructure projects that integrated Digital Twin technology. Data were collected from project reports, real-time sensor outputs, and expert interviews. The predictive capabilities of machine learning models, including Decision Trees, Support Vector Machines (SVM), and Deep Learning networks, were assessed based on their precision, recall, and F1-score. The results demonstrated that Deep Learning models achieved the highest fault detection accuracy, with an F1-score of 92.5%, outperforming other models. The adoption of Digital Twin systems resulted in a 30% reduction in maintenance costs and a 40% decrease in infrastructure downtime. Additionally, AI-driven predictive maintenance improved fault detection efficiency, reducing the average detection time from 15 days to 3 days. These findings highlight the potential of AI-enhanced Digital Twins in optimizing urban infrastructure resilience, cost efficiency, and sustainability. This study underscores the importance of integrating AI and Digital Twin technologies in predictive maintenance strategies. Future research should focus on addressing implementation challenges, including data security, interoperability, and computational costs, to facilitate broader adoption in smart city development]]>
