A reliable data-driven tool condition monitoring system is more and more promising for cutting down on machine downtime and economic losses. However, traditional methods are not able to address machining big data because of low model generalizability and laborious feature extraction by hand. In this paper, a novel deep learning model, named multi-frequency-band deep convolution neural network (MFB-DCNN), is proposed to handle machining big data and to monitor tool condition. First, samples are enlarged and a three-layer wavelet package decomposition is applied to obtain wavelet coefficients in different frequency bands. Then, the multi-frequency-band feature extraction structure based on a deep convolution neural network structure is introduced and utilized for sensitive feature extraction from these coefficients. The extracted features are fed into full connection layers to predict tool wear conditions. After this, milling experiments are conducted for signal acquisition and model construction. A series of hyperparameter selection experiments is designed for optimization of the proposed MFB-DCNN model. Finally, the prediction performance of typical models is evaluated and compared with that of the proposed model. The results show that the proposed model has outstanding generalizability and higher prediction performance, and a well designed structure can remedy the absence of complicated feature engineering.

Bearing fault features are presented as repetitive transient impulses in vibration signals. Narrowband demodulation methods have been widely used to extract the repetitive transients in bearing fault diagnosis, for which the key factor is to accurately locate the optimal band. A multitude of criteria have been constructed to determine the optimal frequency band for demodulation. However, these criteria can only describe the strength of transient impulses, and cannot differentiate fault-related impulses and interference impulses that are cyclically generated in the signals. Additionally, these criteria are easily affected by the independent transitions and background noise in industrial locales. Therefore, the large values of the criteria may not appear in the optimal frequency band. To overcome these problems, a new method, referred to as multiband envelope spectra extraction (MESE), is proposed in this paper, which can extract all repetitive transient features in the signals. The novelty of MESE lies in the following aspects: (1) it can fuse envelope spectra at multiple narrow bands. The potential bands are selected based on Jarque-Bera statistics of narrowband envelope spectra; and (2) fast independent component analysis (fastICA) is introduced to extract the independent source spectra, which are fault- or interference-related. The multi-band strategy will preserve all impulse features and make the method more robust. Meanwhile, as a blind source separation technique, the fastICA can suppress some in-band noise and make the diagnosis more accurate. Several simulated and experimental signals are used to validate the efficiency of the proposed method. The results show that MESE is effective for enhanced fault diagnosis of rolling element bearings. Bearing faults can be detected even in a harsh environment.

Narrowband amplitude demodulation is an effective tool for extracting characteristic features in the fault diagnosis of rolling element bearings. The quality of demodulation largely depends on the frequency band selected for the demodulation. Numerous criteria have been constructed to determine the optimal frequency band. However, independent frequency interferences and in-band noises in narrowband signals can greatly affect the values of the criteria, which may lead to an inaccurate result in locating the optimal band, in demodulating fault features, and finally in the fault detection. Inspired by the nonlocal means (NL-means) denoising method that has been widely used in image processing, this paper proposes a narrowband envelope spectra fusion (NESF) method to enhance fault features and suppress in-band noises before criterion calculation. The method suppresses in-band noises by averaging envelope spectra at neighboring narrow bands. Meanwhile, some minor improvements are made to the conventional narrowband envelope spectrum calculation method to enhance the similarity of the narrowband envelope spectra containing fault features, and finally optimize the fusion process. Then, sparsity values of these denoised envelope spectra, which can lower the impact of independent frequency interferences, are utilized to determine the optimal band and select the optimal envelope spectrum. Frequency signatures of the extracted envelope spectrum can be utilized to indicate the status and fault types of rolling element bearings. A simulated bearing fault signal and three real bearing fault signals are used to validate the effectiveness of the proposed method through comparison studies with protrugram and sparsogram. The results show that the proposed method can effectively extract fault characteristics, even in a harsh environment.

This paper presents a supervised feature extraction method called weighted kernel entropy component analysis (WKECA) for fault diagnosis of rolling bearings. The method is developed based on kernel entropy component analysis (KECA) which attempts to preserve the Renyi entropy of the data set after dimension reduction. It makes full use of the labeled information and introduces a weight strategy in the feature extraction. The class-related weights are introduced to denote differences among the samples from different patterns, and genetic algorithm (GA) is implemented to seek out appropriate weights for optimizing the classification results. The features based on wavelet packet decomposition are derived from the original signals. Then the intrinsic geometric features extracted by WKECA are fed into the support vector machine (SVM) classifier to recognize different operating conditions of bearings, and we obtain the overall accuracy (97%) for the experimental samples. The experimental results demonstrated the feasibility and effectiveness of the proposed method.