Addressing data scarcity in structural health monitoring through generative augmentation

TL;DR

This work tackles data scarcity and class imbalance in structural health monitoring by introducing STFTSynth, a GAN-based framework that generates realistic single-channel STFT spectrograms of bridge-related events. Built on a WGAN-GP foundation, STFTSynth combines dense residual blocks for spectral detail with Bidirectional GRUs to enforce temporal coherence, trained in an event-specific yet label-agnostic fashion. Across comparisons with DCGAN, WGAN-GP, and LSGAN, STFTSynth delivers superior SSIM and PSNR and generally lower FID, validated through both quantitative metrics and qualitative visual inspection. The approach offers a scalable, cost-effective augmentation strategy for rare SHM events (e.g., prestressing wire breakage), with potential to enhance downstream damage-detection pipelines in real-world bridge monitoring.

Abstract

Structural Health Monitoring plays a crucial role in ensuring the safety, reliability, and longevity of bridge infrastructures through early damage detection. Although recent advances in deep learning-based models have enabled automated event detection, their performance is often limited by data scarcity, environmental noise, and class imbalance. To address these challenges, this study introduces a customized Generative Adversarial Network model, STFTSynth, designed particularly for generating short-time Fourier transform spectrograms derived from acoustic event signals. In contrast to augmentation techniques such as MixUp, generative adversarial networks can synthesize high-quality spectrograms that mimic real-world events, enhancing dataset diversity and robustness. The proposed model integrates dense residual blocks for spatial consistency with bidirectional gated recurrent units for temporal dependency modeling. Model performance is evaluated against three baseline generative models using qualitative inspection and quantitative metrics, including Structural Similarity Index Measure, Peak Signal-to-Noise Ratio, and Fréchet Inception Distance. Results show that STFTSynth outperforms baseline models, producing high-resolution, temporally consistent spectrograms that align closely with real-world data. These findings indicate the potential of generative-based data augmentation as a scalable and cost-effective solution for bridge monitoring scenarios where rare events, such as prestressing wire breakage, suffer from data scarcity.

Figures (11)

• Figure 1: Architecture of a Gated Recurrent Unit (GRU)
• Figure 2: Overview of GAN workflow.
• Figure 3: Overview of the proposed GAN-based data augmentation workflow using STFT spectrograms.
• Figure 4: Cerqueta bridge along the A24 highway selected for controlled prestressing tendon breakage tests: (a) plan view; (b) lateral view from East.
• Figure 5: Controlled wire breakage test on the Cerqueta bridge: (a) operational window opened in the beam's web; (b) wire cutting using an electric trimmer; (c) strands with twisted wires after breakage