Structural Nonlinearities for a Single Degree of Freedom System-MATLAB II

This repository contains part of a comprehensive collection of synthetic displacement time-series datasets developed to support the study “Data-Driven Classification of Structural Nonlinearities Using Interpretable Deep Learning on Time Series.” The datasets model the dynamic response of single-degree-of-freedom (SDOF) systems that exhibit a range of structural nonlinearities, including cubic stiffness (hardening/softening), Coulomb friction, clearance, and quadratic damping. Each sample consists of a 100,000-point univariate time series, generated by numerically solving the governing nonlinear equations of motion under logarithmic sweep-sine excitation. Solver environments include MATLAB (ode45, RK4/5) and Julia (Vern7), with systematic variations in sweep rate, forcing amplitude, stiffness/damping parameters, and additive Gaussian noise.The benchmark suite includes both binary-class datasets and more challenging six-class datasets, enabling controlled evaluation of classification models and post-hoc interpretability methods. Reason for Multiple Zenodo Deposits The complete dataset collection is too large to fit within Zenodo’s upload limits for a single record.To ensure full accessibility, the benchmark suite has been divided into three coordinated Zenodo uploads, with each part containing a distinct subset of the datasets used in the paper.Together, these deposits constitute the full dataset release. Dataset Overview This deposit contains one subset of the following datasets (full list shown for reference): LNM1, LNJ1: Binary linear vs. nonlinear classification (2 classes) DM1, DJ1: Six-class Duffing-type nonlinearities DJ19: Six-class dataset with 19 forcing amplitudes ONM19, OWM19, ONJ19, OWJ19: Reproductions and extensions of the Ondra et al. datasets, with and without added noise These datasets differ in complexity, solver type, and noise levels, allowing users to evaluate model robustness and interpretability across a wide range of nonlinear dynamic behaviors. File Formats CSV (.csv) — displacement time series stored as rows or columns MATLAB (.mat) — MATLAB-compatible structured arrays for efficient loading Every sample includes: 100,000-point displacement signal Integer-encoded class label Metadata: solver type, forcing parameters, SNR, dataset identifier How to Cite Please cite the associated research: Abusalameh, B., Wei, J., & Mengaldo, G. (2025). Data-Driven Classification of Structural Nonlinearities Using Interpretable Deep Learning on Time Series. and cite this Zenodo record using its DOI.