Numerical benchmark for damage identification in Structural Health Monitoring

2026-03-12

AI Summary

This study addresses the critical lack of publicly available, high-quality benchmark datasets in structural health monitoring (SHM), a gap primarily due to data ownership restrictions, security concerns, and insufficient metadata. To overcome this limitation, the authors present a high-fidelity synthetic SHM dataset generated from a fixed-fixed steel beam model grounded in single-degree-of-freedom dynamics and Euler–Bernoulli beam theory. The dataset comprehensively incorporates realistic perturbations, including environmental and operational variability, multiple damage scenarios, measurement noise, and sensor faults. Leveraging parallel computing, the data are efficiently synthesized and released as the first open-source SHM benchmark of its kind, accompanied by complete metadata and fully reproducible code. This resource establishes a standardized platform for evaluating data-driven and hybrid physics-informed methodologies, significantly enhancing the comparability of algorithmic performance and the reproducibility of SHM research.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Search and Optimization: Sampling/Simulation-based Search
• Application Domains: Humanities & Computational Social Science

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Web Mining and Content Analysis: Web data provenance, reliability, and authenticity
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

The availability of a dataset for validation and verification purposes of novel data-driven strategies and/or hybrid physics-data approaches is currently one of the most pressing challenges in the engineering field. Data ownership, security, access and metadata handiness are currently hindering advances across many fields, particularly in Structural Health Monitoring (SHM) applications. This paper presents a simulated SHM dataset, comprised of dynamic and static measurements (i.e., acceleration and displacement), and includes the conceptual framework designed to generate it. The simulated measurements were generated to incorporate the effects of Environmental and Operational Variations (EOVs), different types of damage, measurement noise and sensor faults and malfunctions, in order to account for scenarios that may occur during real acquisitions. A fixed-fixed steel beam structure was chosen as reference for the numerical benchmark. The simulated monitoring was operated under the assumptions of a Single Degree of Freedom (SDOF) for generating acceleration records and of the Euler-Bernoulli beam for the simulated displacement measurements. The generation process involved the use of parallel computation, which is detailed within the provided open-source code. The generated data is also available open-source, thus ensuring reproducibility, repeatability and accessibility for further research. The comprehensive description of data types, formats, and collection methodologies makes this dataset a valuable resource for researchers aiming to develop or refine SHM techniques, fostering advancements in the field through accessible, high-quality synthetic data.

Problem

• Structural Health Monitoring
• benchmark dataset
• data accessibility
• Environmental and Operational Variations
• damage identification

Innovation

• Structural Health Monitoring
• synthetic dataset
• Environmental and Operational Variations
• open-source benchmark
• hybrid physics-data approaches