It is of great significance for intelligent manufacturing to study diagnosis methods to realize the diagnosis of mechanical equipment faults. Multiscale weighted permutation entropy is an effective method recently proposed to measure the complexity and dynamic changes of dynamic systems. To solve the shortcoming of multiscale weighted permutation entropy that does not consider high-frequency components, this article proposes hierarchical weighted permutation entropy, which can comprehensively and accurately reflect the low-frequency and high-frequency information of the time series. The simulation signal verifies the effectiveness and superiority of hierarchical weighted permutation entropy. Then, a novel intelligent fault diagnosis method for common rail injectors based on hierarchical weighted permutation entropy and pair-wise feature proximity is proposed. Finally, the proposed method is applied to the common rail injector fault data, and the results verify the effectiveness of the proposed method. Compared with other methods, this method has a higher fault recognition rate and stronger robustness.

研究诊断方法对机械设备故障的诊断对智能制造具有重要意义。多尺度加权置换熵是最近提出的一种用于测量动态系统的复杂性和动态变化的有效方法。为了解决不考虑高频成分的多尺度加权置换熵的缺点，提出了一种分层加权置换熵，可以全面，准确地反映时间序列的低频和高频信息。仿真信号验证了分层加权置换熵的有效性和优越性。然后，提出了一种基于分层加权置换熵和成对特征接近度的共轨喷油器智能故障诊断新方法。最后，将该方法应用于共轨喷油器故障数据，结果验证了该方法的有效性。与其他方法相比，该方法具有较高的故障识别率和较强的鲁棒性。