过度拟合
计算机科学
辍学(神经网络)
人工智能
一般化
卷积神经网络
断层(地质)
机制(生物学)
抓住
方位(导航)
人工神经网络
机器学习
深度学习
数据挖掘
模式识别(心理学)
数学
数学分析
哲学
认识论
地震学
程序设计语言
地质学
作者
Xianghui Liao,Yuhan Mao
摘要
For various bearing states, the traditonal model has insufficient ability to extract important information, low accuracy and poor generalization ability of the model in complex environment.Therefore, a bearing fault diagnosis method based on CNN-BiLSTM-Attention is proposed. This method combines convolutional neural network (CNN) and bidirectional long and short term memory network (BiLSTM), adds the Dropout mechanism to the BiLSTM network to suppress the overfitting problem, then introduces the attention mechanism to automatically assign different weights to the BiLSTM network to improve the sensitivity and the ability to grasp different fault information. In order to verify the effect of this model, the Case Western Reserve University(CWRU) bearing fault dataset was used for experimental verification, the results show that the proposed model has higher accuracy and stronger generalization ability than other models in the multitask problem under complex working conditions.
科研通智能强力驱动
Strongly Powered by AbleSci AI