Data driven CAN node reliability assessment for manufacturing system

详细信息    查看全文

• 作者：Leiming Zhang ; Yong Yuan ; Yong Lei
• 关键词：bus ; off ; monitoring ; Controller Area Network ; industrial automation
• 刊名：Chinese Journal of Mechanical Engineering
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：30
• 期：1
• 页码：190-199
• 全文大小：
• 刊物主题：Mechanical Engineering; Theoretical and Applied Mechanics; Manufacturing, Machines, Tools; Engineering Thermodynamics, Heat and Mass Transfer; Power Electronics, Electrical Machines and Networks; Elec
• 出版者：Chinese Mechanical Engineering Society
• ISSN：2192-8258
• 卷排序：30

文摘

The reliability of the Controller Area Network(CAN) is critical to the performance and safety of the system. However, direct bus-off time assessment tools are lacking in practice due to inaccessibility of the node information and the complexity of the node interactions upon errors. In order to measure the mean time to bus-off(MTTB) of all the nodes, a novel data driven node bus-off time assessment method for CAN network is proposed by directly using network error information. First, the corresponding network error event sequence for each node is constructed using multiple-layer network error information. Then, the generalized zero inflated Poisson process(GZIP) model is established for each node based on the error event sequence. Finally, the stochastic model is constructed to predict the MTTB of the node. The accelerated case studies with different error injection rates are conducted on a laboratory network to demonstrate the proposed method, where the network errors are generated by a computer controlled error injection system. Experiment results show that the MTTB of nodes predicted by the proposed method agree well with observations in the case studies. The proposed data driven node time to bus-off assessment method for CAN networks can successfully predict the MTTB of nodes by directly using network error event data.