Data for LambNet-T: A lightweight path-conditional transformer autoencoder for temperature-aware baseline learning in Lamb-wave SHM

Title of the Paper
LambNet-T: A lightweight path-conditional transformer autoencoder for temperature-aware baseline learning in Lamb-wave SHM

Authors
J. Horňas (hornas@VZLU.cz), O. Vích, L. Šedková, I. Mlch, B. Cabrnoch, M. Král

Affiliation
VZLU AEROSPACE, a.s. 
Beranových 130, 19900 Prague, Czech Republic

DOI
https://doi.org/10.1016/j.ultras.2026.107973

Description
The root directory of the supplementary data contains a README.pdf file describing the data observation, along with Comma-Separated Values
(CSV) files corresponding for each Baseline and Diagnostic measurement at 13 different temperatures in [-10, 50] °C range with 5 °C step across
all transducer paths (e.g., 1-7, 1-9, ..., 7-11) for impact I1 and impact I2 at 40 and 50 kHz of excitation frequency. The directory structure is
as follows:

     /
    ├─ README.pdf
    ├─ 01_Baseline_40kHz_I1.csv
    ├─ 01_Diagnostic_40kHz_I1.csv
    ├─ 02_Baseline_50kHz_I1.csv
    ├─ 02_Diagnostic_50kHz_I1.csv
    ├─ 03_Baseline_40kHz_I2.csv
    ├─ 03_Diagnostic_40kHz_I2.csv
    ├─ 04_Baseline_50kHz_I2.csv
    └─ 04_Diagnostic_50kHz_I2.csv*
            ├─ row id 1 (column-wise notation):        P01-07_T-10C,  ..., P07-11_T50C
            ├─ row id 2 (1st time sample):                164E-15,            ..., 1.80E-14
            ├─ row id 3 (2nd time sample):                -7.4727,            ..., -5.5183
            ├─ . . .
            └─ row id 6001 (6000th time sample):     -70.667,            ..., -57.333

* Column-wise structure of signal for specific sensor path (P) and temperature (T) in °C for 04_Diagnostic_50kHz_I2.csv as an example. Each
 time sample represents the amplitude value in mV.

-------- IF ANY DATASET IS USED, PLEASE CITE RELATED REFERENCE LISTED BELOW --------
[1] Jan Horňas et al. “LambNet-T: A lightweight path-conditional transformer autoencoder
for temperature-aware baseline learning in Lamb-wave SHM”. In: Ultrasonics (2026).
ISSN: 0041-624X. DOI: https://doi.org/10.1016/j.ultras.2026.107973.