Intelligent Diagnosis of Rolling Bearing Fault Based on Improved Convolutional Neural Network and LightGBM

Abstract

Aiming at the problems of weak generalization ability and long training time in most fault diagnosis models based on deep learning, such as support vector machines and random forest algorithms, one intelligent diagnosis method of rolling bearing fault based on the improved convolution neural network and light gradient boosting machine is proposed. At first, the convolution layer is used to extract the features of the original signal. Second, the generalization ability of the model is improved by replacing the full connection layer with the global average pooling layer. Then, the extracted features are classified by a light gradient boosting machine. Finally, the verification experiment is carried out, and the experimental result shows that the average training and diagnosis time of the model is only 39.73 s and 0.09 s, respectively, and the average classification accuracy of the model is 99.72% and 95.62%, respectively, on the same and variable load test sets, which indicates that the diagnostic efficiency and classification accuracy of the proposed model are better than those of other comparison models.

Classification of Breast Tumors Based on Histopathology Images Using Deep Features and Ensemble of Gradient Boosting Methods

“…We decided to use GBDT based methods as classifier because of their high prediction accuracy, and adaptability to non-linear characteristics and good training effect [31,49]. XGBoost, CatBoost and LightGBM were chosen as classifier candidates.…”

“…XGBoost, CatBoost and LightGBM were chosen as classifier candidates. XGBoost was chosen because of its new regularization method that resists overfitting, therefore making the model more robust [31,50]. CatBoosts improved generalization as well as its ability to capture high-order dependencies makes it a viable candidate as well [51].…”

“…Last but not least LightGBM is an attractive method to use as a classifier because of its capability in handling large-scale data. In addition, CatBoost's leaf-wise growth approach makes it more accurate than other level-wise GBDT methods [31,52]. As all three of our classifier candidates performed relatively well on this task, we decided to incorporate a soft majority voting method in the proposed method to enhance its predictive performance and improve the model's robust-ness by lowering its variance.…”

“…XGBoost, CatBoost and LightGBM were used alone in three of the seven models while the other four [32,33]. Light-GBM can quickly process massive data by instigating leafwise tree growth (see Figure 2) with depth limitation and gradient-based one-side sampling as well as exclusive feature bundling approaches [31,34]. LightGBM hyperparameters were tuned by trial and error.…”

Research on Vibration Fatigue Damage Locations of Offshore Oil and Gas Pipelines Based on the GA-Improved BP Neural Network

“…Tian et al [11] proposed an improved genetic algorithm (IGA) combined with the global optimization characteristics of the genetic algorithm (GA) and the local optimal solution of the simulated annealing (SA) algorithm in this paper, which adopts SA in the process of selecting subpopulations. Some scholars have found that the neural network has good performance in feature extraction and damage recognition [12,13].…”