School of Mechatronical Engineering, Beijing Institute of Technology, Beijing 100081, China.
Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing 100190, China.
Sensors (Basel). 2019 Feb 21;19(4):910. doi: 10.3390/s19040910.
Fault diagnostic software is required to respond to faults as early as possible in time-critical applications. However, the existing methods based on early diagnosis are not adequate. First, there is no common standard to quantify the response time of a fault diagnostic software to the fault. Second, none of these methods take into account how the objective to improve the response time may affect the accuracy of the designed fault diagnostic software. In this work, a measure of the response time is provided, which was formulated using the time complexity of the algorithm and the signal acquisition time. Model optimization was built into the designed method. Its objective was to minimize the response time. The constraint of the method is to guarantee diagnostic accuracy to no less than the required accuracy. An improved feature selection method was used to solve the optimization modeling. After that, the design parameter of the optimal quick diagnostic software was obtained. Finally, the parametric design method was evaluated with two sets of experiments based on real-world bearing vibration data. The results demonstrated that optimal quick diagnostic software with a pre-defined accuracy could be obtained through the parametric design method.
故障诊断软件需要在时间关键型应用中尽早及时响应故障。然而，现有的基于早期诊断的方法并不充分。首先，没有通用的标准来量化故障诊断软件对故障的响应时间。其次，这些方法都没有考虑到提高响应时间的目标可能会如何影响设计的故障诊断软件的准确性。在这项工作中，提供了一种响应时间的度量方法，该方法使用算法的时间复杂度和信号采集时间来进行公式化。模型优化被构建到设计方法中。其目标是最小化响应时间。该方法的约束条件是保证诊断准确性不低于所需的准确性。改进的特征选择方法被用于解决优化建模。之后，获得了最优快速诊断软件的设计参数。最后，基于实际的轴承振动数据进行了两组实验来评估参数设计方法。结果表明，通过参数设计方法可以获得具有预定义准确性的最优快速诊断软件。
