Monitoring data contain the important status information of the monitored object, and are the basis for following data mining and analysis. However, the monitoring data usually suffer the pollution of the outliers, leading to negative effect on the subsequent data processing. To address the problem, this paper proposed an outlier detection method based on stacked autoencoder (SAE). SAE has a powerful capability of feature extraction and greatly preserves the original information of the data. The trained SAE by normal data can learn the characteristics of normal data. When a set of data with outliers are inputted to the trained network, there are larger reconstruction errors at the outliers between the original input data and the reconstructed data obtained by using the encoding parameters and the decoding parameter mapping, which provides a basis for locating outliers. Meanwhile, this paper introduced the Grubbs criterion and the PauTa criterion to identify the reconstruction errors corresponding to the outliers based on the traditional threshold method. The method can quickly isolate the abnormal data from the normal data according to the reconstruction error and the identification criterion. The effectiveness and superiority of the proposed method have been validated by experiment on real data and comparisons with traditional outlier detection algorithms.

Diagnosing the bearing fault, especially incipient fault is important for equipment health management while is still a challenge in which high-fidelity extraction of the fault signature is expected. A method termed flexible analytical wavelet transform (FAWT)-enhanced sparse representation with nonconvex regularization is proposed in this research. FAWT enjoys flexible covering along both the frequency and time axis as well as tunable oscillation property and is adopted to well match the fault impulses after parameters optimization. In the fabricated FAWT-enhanced sparse model with generalized minimax-concave regularization, an index termed harmonic-to-noise energy ratio of envelope spectrum (ES-HNER) is proposed which is found effective and robust for quantitative assessment of the richness of fault signature and could be automatically evaluated from the envelope spectrum, based on which the parameters for constructing the FAWT basis and threshold are optimized via maximizing the ES-HNER in the candidate parameters space. The sparse decomposition signals are further obtained via solving the FAWT-enhanced sparse model, upon which the bearing fault signature is expected to be exhibited on the envelope spectrum. The performance of the proposed method has been validated via analysis of both simulation and experiment signals as well as comparison with other methods.