Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Data Mining Practical Machine Learning Tools and Techniques

A new Ensemble Empirical Mode Decomposition (EEMD) is presented. This new approach consists of sifting an ensemble of white noise-added signal (data) and treats the mean as the final true result. Finite, not infinitesimal, amplitude white noise is necessary to force the ensemble to exhaust all possible solutions in the sifting process, thus making the different scale signals to collate in the proper intrinsic mode functions (IMF) dictated by the dyadic filter banks. As EEMD is a time–space analysis method, the added white noise is averaged out with sufficient number of trials; the only persistent part that survives the averaging process is the component of the signal (original data), which is then treated as the true and more physical meaningful answer. The effect of the added white noise is to provide a uniform reference frame in the time–frequency space; therefore, the added noise collates the portion of the signal of comparable scale in one IMF. With this ensemble mean, one can separate scales naturall...

/pdf/ensemble-empirical-mode-decomposition-a-noise-assisted-data-4htvj7tcuy.pdf

Ensemble empirical mode decomposition: a noise-assisted data analysis method

Deep learning and its applications to machine health monitoring

A review on empirical mode decomposition in fault diagnosis of rotating machinery

A new nonparametric degradation modeling method for truncated degradation signals by axis rotation

The cleanliness defect is commonly-occurred during the manufacturing process of rolling element bearings, which reduces the defect detection accuracy and further deteriorates the noise vibration harshness of bearings. To effectively detect the cleanliness defect of ex-factory bearings, this paper presents a method based on the sliding window kurtosis indicator. It first calculates the distribution of kurtosis and the threshold-overshoot sum of the sliding window kurtosis series to assess the degree of cleanliness. After that, an indicator-sensitive parameters and the judgment threshold are adaptively selected by a Bayesian optimization algorithm. The proposed method is verified with data from the bearing manufacturing processing. The results show that the proposed method has high accuracy for identifying cleanliness of bearings.

A Diagnosis Method of Cleanliness Defect in Rolling Element Bearing Manufacturing

Rotate vector (RV) reducers have advantages of high torque ratio, and extremely reliable functioning under dynamic load conditions, and so forth, and they are extensively used in precision transmissions, such as industrial robots. The torsional stiffness of RV reducers is a main parameter affecting the meshing vibration and transmission performance. In addition, the variable loads and tooth modifications have significant influence on the torsional stiffness of RV reducers. In this paper, a new method of calculating torsional stiffness for RV reducers considering variable loads and tooth modifications is presented. The dynamic stress and deformation of meshing teeth are calculated based on Hertz formulation, and the number of meshing teeth is determined by analyzing the dynamic stress of meshing teeth. Then, the torsional stiffness of RV reducers is obtained. Finite element method is applied to calculate the deformation of cycloidal-pin gear in the maximum force position δmax. The influence of variable applied loads and different tooth modifications of cycloidal-pin gear transmission on torsional stiffness are also studied. The results show that variable applied loads and different tooth modifications influence the torsional stiffness of RV reducers by changing teeth deformation and clearance, respectively. 

A new method for calculating time‐varying torsional stiffness of RV reducers with variable loads and tooth modifications


 Joint optimization of maintenance operations and spare parts inventory optimization is promising for ensuring the reliable and efficient operation of industrial systems. Extensive studies have been conducted on joint optimization emphasizing various aspects. However, those studies still suffer two limitations. 1) The considered system structures are restricted to particular cases, and thus are unable to fully address the joint optimization needs of actual industrial systems possessing various structures, and 2) imperfect maintenance, as has been commonly adopted in industries, is not addressed in existing works, which compromises the economic efficiency of the joint policy. Targeted at these limitations, this paper investigates the joint optimization for multi-unit systems with a generalized structure, i.e., the systems consist of multiple identical units, and each unit possesses a complex subsystem structure. First, a total cost rate model is formulated, in which maintenance actions include both perfect replacement and imperfect maintenance, and the spare parts inventory policy consists of an (s, S) policy and the option to place emergency orders. Based on the formulated model, a simulation optimization method is proposed. The simulation module evaluates the total cost rate iteratively, and the optimization module utilizes a variable neighborhood search-based algorithm to obtain a high-quality joint policy with tractable computation times. Two numerical examples are provided for demonstration, and sensitivity analyses are conducted to study the effects of parameters.

Joint optimization of maintenance and spare parts inventory for multi-unit systems with a generalized structure

Recently, a large number of deep learning-based prognostics methods for machinery remaining useful life (RUL) have been proposed. And massive monitoring data is the basis of deep learning-based RUL prediction methods. However, most existing methods usually assume that the monitoring data acquired from different sensors contain similar degradation information, and they lack consideration on effectively identifying the multi-sensor degradation information, which affects the prediction performance of deep networks. To overcome the weakness, this paper proposes a multi-channel attention bidirectional long short-term memory network (MCA-BiLSTM) for RUL prediction of machinery. In MCA-BiLSTM, sensitive features are firstly extracted and selected from the monitoring data of each sensor individually. Then, multi-channel bidirectional long and short-term memory (LSTM) network is constructed with the sensitive features as the inputs. Simultaneously, time attention and channel attention are utilized to realize the adaptive information fusion. Finally, the fusion representations are input into fully connected layers to predict RUL. The effectiveness of MCA-BiLSTM is verified by multisensor monitoring data from life testing of milling cutters.

Yaguo Lei

Papers

A new nonparametric degradation modeling method for truncated degradation signals by axis rotation

A Diagnosis Method of Cleanliness Defect in Rolling Element Bearing Manufacturing

A new method for calculating time‐varying torsional stiffness of RV reducers with variable loads and tooth modifications

Joint optimization of maintenance and spare parts inventory for multi-unit systems with a generalized structure

Remaining Useful Life Prediction Based on Multi-channel Attention Bidirectional Long Short-term Memory Network