Abstract Effective incipient fault detection requires a method that can separate fault signatures under constant and time-varying operating conditions. In this work, we focus on improving informative frequency band identification methods for applications under constant and time-varying operating conditions. Many automatic band selection techniques exist and have proven effective under constant speed and load conditions. However, it has been shown that these techniques occasionally identify frequency bands that contain non-damage related information, especially under fluctuating operating conditions and at low damage levels. With this research, a new methodology is proposed which makes use of popular informative frequency band selection techniques, such as the Fast Kurtogram, to effectively identify damage under constant and fluctuating speed and load conditions. A key step in this methodology, the NICogram, requires healthy historical data, which is used to identify frequency bands that contain novel information in unclassified signals. The methodology uses multiple signals to identify whether a component is damaged or not through a probabilistic approach. It is shown that the method performs much better than the conventional informative frequency band identification methods on synthetic and experimental data.

中文翻译：

---

NIC 方法：一种概率方法，通过在时变操作条件下利用可用的健康历史数据来改进信息频带识别。

摘要 有效的初期故障检测需要一种能够在恒定和时变操作条件下分离故障特征的方法。在这项工作中，我们专注于改进恒定和时变操作条件下应用的信息频带识别方法。存在许多自动频带选择技术，并且已经证明在恒定速度和负载条件下是有效的。然而，已经表明，这些技术偶尔会识别包含非损坏相关信息的频段，尤其是在波动的操作条件和低损坏级别下。通过这项研究，提出了一种新方法，该方法利用流行的信息频带选择技术，例如快速库图，在恒定和波动的速度和负载条件下有效识别损坏。该方法的关键步骤 NICogram 需要健康的历史数据，用于识别包含未分类信号中新信息的频段。该方法使用多个信号通过概率方法来识别组件是否损坏。结果表明，该方法在合成和实验数据上的性能比传统的信息频带识别方法要好得多。