Du Yangkai, Huang Jiayuan, Tao Shuting, three students of Class of 2021 in computer engineering published a full conference paper on the International Conference on e-Business Engineering (ICEBE), and their paper , “Knowledge Graph Construction and Reasoning for Intelligent Maintenance of Power Plants ” was also selected for Springer-Verlag and will be published on “Advances in E-business Engineering for Ubiquitous Computing” of “Lecture Notes in Data Engineering and Communication Technologies (LNDECT)” . This paper was written during the last summer vacation, and Prof. Wang Hongwei advised this team. It is worth mentioning that only 13 papers were selected and recommended as long paper to publish from nearly 100 papers, according to the results of peer review. This paper was highly praised by the reviewing experts. The ICEBE is a prestigious conference in the IEEE computer society, and a flagship event co-sponsored by IEEE Technical Committee on Business Informatics and Systems and the National Engineering Laboratory for E-Commerce Technologies (NELECT). Since 2003, ICEBE has been provided as a high-quality international forum for researchers, engineers and business specialists to exchange the cutting-edge ideas, findings and experiences of e-business. In October this year, the 16th conference will be held at Fudan University and SAP China. According to incomplete statistics, it is extremely rare that sophomore undergraduates can publish a long paper in such a high-level conference in its history.
As one of the results of the project “Fault data analysis and mining for intelligent maintenance”, which was selected as Zhejiang University Student Research Training Program, according to the previous experience and feedback in power plant maintenance, they put forward a Bi-LSTM-CRF model and the attention-based Bi-LSTM model, proposes a novel process of an automatic construction and reasoning of knowledge graphs to support the intelligent maintenance of complex power equipment and a new generation of intelligent maintenance system platform.
This paper was reviewed and evaluated by three professional reviewers, all of the reviewers gave very positive feedbacks. One of them commented, “This is a well-written and thought-provoking paper which addresses the highly desirable goal of converting textual reports into machine-accessible data, in order to compile information about the reliability of individual physical components in a complex plant or system. While this paper focuses mainly on the components in a power station, the same problem exists in a wide range of other engineering contexts where reliability is critical, for instance in the aviation industry.”。
Project research is a very important part of ZJUI information system and data sciences program, guided by Prof. Wang Hongwei, the team of this program cooperate with China National Nuclear Power Co., Ltd., China Huadian Corporation Ltd. and other giant power corporations, focus on the knowledge driven intelligent design, manufacturing and maintenance of large engineering equipment. In the past year, this team published almost 20 papers in calculating methods in complex systems, knowledge extraction and modeling, fault diagnosis, the manufacturing system optimization etc.. This time, these undergraduates published a long paper in high-quality international conference, fully demonstrated the high level of ZJUI undergraduate scientific research training, highlights the internationalization and interdisciplinary education of ZJUI.
Writer/Photo | Tao Shuting