fault diagnosis event knowledge graph

Citation Author(s):
ChangHao
Men
Cheng-Geng
Huang
Yu
Han
Submitted by:
changhao Men
Last updated:
Tue, 10/08/2024 - 21:35
DOI:
10.21227/7y2m-ap98
Data Format:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

We organized and collected two years' worth of complete fault work orders from a wind farm, and structured these work orders into a fault diagnosis event knowledge graph using the proposed algorithm. This graph includes fault modes, fault impacts, fault symptoms, inspection schemes, root cause identification, and maintenance strategies, covering all potential fault information and handling methods for wind turbines. This dataset records the head entity-relation-tail entity information in the form of triples using JSON format. The head and tail entities correspond to fault-related event information, such as fault symptoms at the time of occurrence and fault-handling schemes, while the relation information encompasses the logical relationships between all events, covering the entire process from the onset of the fault to its resolution.

Instructions: 

The dataset is stored using the Neo4j graph database, specifically version 4.0.12. The final output format is a JSON file downloaded from the Neo4j graph database. The process of using and analyzing the data involves converting the JSON format back into Neo4j via Python conversion scripts, enabling re-visualization and further experimental analysis, including database queries and the large model-based retrieval-augmented generation processes supported by this dataset.

Comments

  

Submitted by LUIS ROBERTO CRUZ on Sun, 09/15/2024 - 02:00

A good dataset for reliability analysis.

Submitted by Qazi Jamal on Fri, 09/27/2024 - 01:57