A Monte Carlo-based Sallen-Key benchmark dataset for data-driven predictive maintenance and few-shot fault diagnosis 核心 · 已核验

doi:10.6084/m9.figshare.32054736.v2

This dataset provides a high-fidelity, Monte Carlo-based frequency response benchmark for analog circuit predictive maintenance and few-shot fault diagnosis. Based on the classic Sallen-Key second-order active low-pass filter, the dataset incorporates 5% and 10% manufacturing tolerances (assuming a Gaussian distribution) to accurately simulate real-world physical component variations.It contains 9,000 AC sweep frequency traces across 9 distinct circuit states (1 normal state, 8 soft and hard fault modes). Both raw data (Raw_Trace_Data) and preprocessed data (Processed_Trace_Data with 2% AWGN and Min-Max normalization) are provided in CSV and NPY formats. This dataset accurately reflects the physical overlap between normal tolerance bands and minor fault features, making it an ideal benchmark for evaluating machine learning, metric learning, and robust AI algorithms in noisy, low-sample environments.

落地页

https://figshare.com/articles/dataset/A_Monte_Carlo-based_Sallen-Key_benchmark_dataset_for_data-driven_predictive_maintenance_and_few-shot_fault_diagnosis/32054736/2

许可证

CC-BY-4.0 (判读置信:inferred)

国内可访问性

国内直连:可达 (2026-07-11 检测) 代理通道:可达 (2026-07-11 检测)
检测口径:lychee 双通道单轮探测;「直连超时」表示检测窗口内未完成,系慢或不稳定证据,不构成封锁证据。

发布年份

2026

发布方

figshare

PHM 任务

fault_diagnosis

故障工况