Under the “carbon peaking and carbon neutrality” strategic goal

high voltage direct current (HVDC) transmission systems

as an important part of the new power system

face many challenges in fault diagnosis. Sequence of events recording (SER) data is the key for monitoring and evaluating the operating status of HVDC transmission systems. To address the issues of weak adaptability

low automation

and poor scalability in existing SER data analysis methods

this article proposes the introduction of event knowledge graph technology to construct an abnormal event knowledge graph for the HVDC transmission systems

bridging the information gap between projects

faults

and abnormal events. Based on the characteristics of SER data and expert experience in fault diagnosis

a sequence link prediction (SLP) task is proposed and defined. A fault diagnosis dataset for the HVDC transmission systems is constructed

and benchmark models for SLP of SER events are established. Furthermore

a fault diagnosis method for HVDC transmission systems is proposed

leveraging SLP of SER events. The abnormal event knowledge graph and fault diagnosis dataset for the HVDC transmission systems are publicly released under the CC BY-NC-SA 4.0 license. Experimental tests and comparative analysis show that the models are validated to autonomously and accurately diagnose 53 typical faults in the HVDC transmission systems

such as busbar grounding and valve group short circuit. The method enhances the efficiency and accuracy of SER abnormal event analysis.