Bench-Scale Study of Magnetically Influenced Dynamic Response in a Sloshing Tank

Overview Load‐cell–based CG_y measurements for a rectangular acrylic tank (600 × 60 × 300 mm; water depth h = 40 mm) under near‐resonant horizontal excitation (f ≈ 0.55 Hz).Two paired experimental campaigns were performed: Split‐wall (15 + 15) — symmetric, weak magnetic field; baseline verification (no measurable change). Single‐wall high‐field (Mag‐30@L vs Dummy‐30@L) — all 30 magnets concentrated on one wall, compared to mass‐matched nonmagnetic dummies; field ≈ tens of mT. The first 5 s of each run were discarded; the subsequent 20 s steady segment was analyzed.Sampling rate ≈ 191 Hz. A 2 mm ferromagnetic plate was mounted outside the tank bottom and kept identical in all runs. Contents raw/ — two ZIP archives (20 CSV files total; 10 × Mag, 10 × Dummy).Columns: Time (s), CG_y (mm) (SI units). figures/ — final publication figures code/ — magnet_paired_analysis.ipynb reproduces all analyses and figures. Reproducibility Unzip the raw CSVs, open the notebook, and run all cells.The script automatically: Computes P2P and RMS for each paired run. Performs paired t-tests and Wilcoxon signed‐rank checks (Mag – Dummy). Generates FFT spectra (Hann window) over [0.2–20 Hz], highlighting 0–5 Hz and 10–20 Hz bands. Produces mean ± SEM plots consistent with the published figures. All analyses are fully automated and self‐contained in the notebook. License Data: CC BY 4.0 Code: MIT (included under code/) Version History (Changelog) v3 (this): Added corrected Mag‐30@L / Dummy‐30@L archives Extended FFT to 20 Hz range Added ΔCG_y difference plot and SEM overlays Updated notebook and figure set v2 (Oct 2025): Added single‐wall high‐field dataset (Mag‐30@L vs Dummy‐30@L); 10 paired runs; improved statistics. v1 (Sep 2025): Initial release (split‐wall 15 + 15 configuration). Keywords sloshing; center of gravity; CG_y; static magnets; Mag‐30@L; Dummy‐30@L; FFT; bench‐scale; dataset; amplitude attenuation; paired design