Paderborn Dataset Processed for Domain Adaptation Scenarios

The dataset contains sensor readings of bearing machines under 4 different operating conditions, with each having 3 different classes, i.e., healthy, inner-bearing damage, and outer-bearing damage. Each operating condition refers to different operating parameters, including rotational speed, load torque, and radial force [. In our experiments, each operating condition is considered as one domain. Eventually, we can perform 12 cross-condition scenarios for domain adaptation. To construct the data samples for each domain, we adopted a sliding window to segment the data into small segments. We set the window size of 5120 and shifting size of 4096

搜集汇总

背景与挑战

背景概述

该数据集是Paderborn数据集的预处理版本，专为域适应场景设计，包含轴承机器在4种不同运行条件下的传感器读数，每种条件分为健康、内轴承损伤和外轴承损伤3类。通过滑动窗口方法（窗口大小5120，移动步长4096）将数据分割为样本，支持12个跨条件域适应实验，用于机器状态监测和故障诊断研究。