Application of Digital Twin Technology for Real-Time Monitoring and Predictive Maintenance of Coastal Bridges under Climate Change Scenarios

Abstract

Coastal bridges are increasingly vulnerable to structural degradation due to extreme environmental conditions and the accelerating impacts of climate change. Traditional inspection methods are no longer sufficient to provide early detection or preventive maintenance. This study proposes developing a digital twin system that integrates real-time sensor data and machine learning algorithms to monitor and predict the structural conditions of coastal bridges. Data from environmental (humidity, temperature, salinity) and structural (strain, displacement) sensors were collected at a case-study bridge in Semarang, Indonesia. The system employed a Long Short-Term Memory (LSTM) model to forecast stress anomalies and potential damage, achieving high accuracy with a Mean Absolute Percentage Error (MAPE) below 5%. The digital twin interface visualizes risk zones and enables early warning with an average lead time of 5–7 days. Field validations showed a 97.5% match between system predictions and manual inspections. This research contributes a replicable, scalable framework for adaptive infrastructure management that enhances maintenance decision-making and resilience to climate-induced deterioration in coastal regions.