DATA ANOMALY DETECTION IN STRUCTURAL HEALTH MONITORING USING A MODIFIED TRANSFORMER ENCODER WITH 1D-CNN LAYERS
Article Sidebar
Main Article Content
Abstract
Structural health monitoring (SHM) systems produce continuous, high-volume sensor streams in which environmental effects, transmission faults and sensor malfunctions create data anomalies that corrupt downstream structural assessment. We present a lightweight classifier that assigns each one-hour acceleration record to one of seven patterns: normal, missing, minor, outlier, square, trend and drift. Each record is split into 30-s windows, and four statistical and 20 frequency-domain features are extracted per window, giving a 120 × 24 feature sequence. The sequence is classified by a Transformer encoder in which the position-wise feed-forward network is replaced by two 1 × 1 one-dimensional convolutional (1D-CNN) layers and positional encoding is removed. On data from 38 accelerometers on a cable-stayed bridge, the model reaches 97.1% test accuracy with about 51K parameters, against 95.8% for a standard encoder with 446K parameters, and exceeds the accuracies reported for image-based DNN/CNN and shapelet-based methods on data from the same bridge.
Downloads
Article Details
Section
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
How to Cite
References
Y. Bao, Z. Tang, H. Li, and Y. Zhang, “Computer vision and deep learning–based data anomaly detection method for structural health monitoring,” Struct. Health Monit., vol. 18, no. 2, pp. 401–421, 2019.
Z. Tang, Z. Chen, Y. Bao, and H. Li, “Convolutional neural networkbased data anomaly detection method using multiple information for structural health monitoring,” Struct. Control Health Monit., vol. 26, no. 1, e2296, 2019.
M. Arul and A. Kareem, “Data anomaly detection for structural health monitoring of bridges using shapelet transform,” arXiv:2009.00470, 2020.
A. Vaswani et al., “Attention is all you need,” in Advances in Neural Information Processing Systems, vol. 30, 2017.
S. Nuriev et al., “Data anomaly detection in structural health monitoring using modified transformer encoders with 1D-CNN layers,” Smart Struct. Syst., vol. 35, no. 6, p. 337, 2025.