Smart-building structural health monitoring (SHM) requires a unified digital representation capable of integrating heterogeneous sensing devices, continuous structural states, and burst-oriented post-event assessment without embedding device-specific logic throughout the software stack. This study proposes a semantic digital twin architecture in which SensorType, DeviceProfile, and site metadata form a semantic single source of truth and generate W3C Web of Things Thing Descriptions at runtime. The resulting WoT-driven contract governs field mapping, schema-on-write persistence, generic API access, state visualization, and engineering-threshold evaluation. To accommodate heterogeneous temporal behavior, event-driven seismic assessment and state-driven construction tilt monitoring are orchestrated as distinct workflows that share persistence, notification, and observability services while retaining separate timing contracts. Controlled extension experiments required no manual data-layer, backend, ingestion, or frontend modification, with a runtime source-hash difference of zero. Under a ten-building seismic-event burst, continuous write-lag p95 changed by â20 ms from a 969 ms baseline while all event jobs completed without restart or out-of-memory conditions. The ingestion path further sustained 71,040 points/s at 300 sensors with no dropped points. These results demonstrate that WoT-driven semantic interoperability and eventâstate workflow orchestration can provide an extensible integration foundation for smart-building SHM within a clearly defined configuration boundary.
Abstract Heritage buildings are highly vulnerable to structural degradation due to aging materials, environmental exposure, and natural disasters, necessitating intelligent and realâtime monitoring solutions. The current study proposes a dew computingâenabled digital twin framework integrated with Explainable Artificial Intelligence (XAI) for structural risk evaluation and health prediction of heritage infrastructure. The framework combines Internet of Thingsâbased sensing, dewâfogâcloud computing architecture, blockchainâbased data security, and a hybrid deep learning model to enable efficient, low latency, and reliable monitoring. Temporal structural data are processed using a Convolutional Neural NetworkâGated Recurrent Unit (GRU) model for feature extraction and timeâseries prediction of the Structural Health Index, while a Random Forest (RF) classifier categorizes structural risk into safe, degraded, and critical states. Shapley Additive explanationsâbased XAI is incorporated to enhance interpretability and support expert decisionâmaking. Experimental evaluation on a simulated dataset of 45,212 instances demonstrates the effectiveness of the proposed approach. The GRUâbased model achieves high prediction performance with an accuracy of 94.42%, sensitivity of 94.85%, specificity of 97.01%, and F1âscore of 94.43%. Regression analysis shows low prediction errors (Mean Absolute Error: 0.0158, Root Mean Squared Error: 0.0198) and a high coefficient of determination (), indicating strong agreement between predicted and actual structural states. The RF classifier further achieves 94.64% accuracy in structural risk classification. The framework exhibits low latency (~0.000195 s per sample), high reliability under noisy conditions (up to 99%), and strong scalability across increasing dataset sizes. Overall, the proposed system provides a robust, scalable, and interpretable solution for proactive Structural Health Monitoring and riskâaware maintenance of heritage buildings, significantly improving realâtime decisionâmaking and longâterm conservation strategies.
Load monitoring and damage identification are important tasks in the field of Structural Health Monitoring. Reconstructing unknown force inputs or system parameters usually involves the solution of an inverse problem which is mostly ill-posed. In the last decades a lot of effort has been spent in solving these problems separately. However, unknown loads and damage both have influence on the structural vibration pattern. The difficulty of simultaneous identification of external forces and structural damage is to distinguish these influences by using only the measured vibration effects. The use of prior knowledge of the unknown quantities is advisable for solving the combined inverse problem and to obtain meaningful solutions. In this contribution a sparsity-based reconstruction method in time domain is developed for identifying the unknown structural force excitation and damage parameter simultaneously by using output-only acceleration data. Sparsity means the majority of the solution vector is zero and only a very few elements are nonzero. Here damage is interpreted as additional load on the damaged structural element (virtual distortion). A numerical proof-of-concept experiment of a quadratic aluminum plate is presented. It shows that the proposed reconstruction method is able to identify the unknown external force and damage parameter by using a significant lower number of accelerometers than present methods.
An internal model based method is used to estimate the structural displacements under ambient excitation using only acceleration measurements. Strain measurements are incorporated to expand the method to single span concrete bridges subjected to moving vehicle loads. The structural response is assumed to remain the linear range for the duration of the loading. The excitation is assumed to be with zero mean and relatively broad bandwidth such that at least one of the fundamental modes of the structure is excited and dominates in the response. Using the structural modal parameters and partial knowledge of the load, their respective internal models can be established. These internal models can then be used to form an autonomous state-space representation of the system. It is shown that structural displacements, velocities, and accelerations are the states of such a system, and it is fully observable when the measured output contains structural accelerations only. Reliable estimates of structural displacements are obtained using the standard Kalman filtering technique. These displacement estimates can be used to determine the moment demand and provide insight into whether this demand is exceeding the capacity of the bridge. The effectiveness and robustness of the proposed method has been demonstrated and evaluated via numerical simulations of an eight-story lumped mass model along with a simply supported single span concrete bridge subjected to a moving traffic load. Experimental data of a three-story frame excited by ground accelerations from an actual earthquake record is also used. Lastly, field data from an inverted arch concrete bridge is analyzed as proof of concept for deployment of a structural health monitoring system for the purpose of displacement estimations.