Fault analysis method, system, device and storage medium

By dynamically adjusting the execution plan and data access method of query tasks using a large language model, the resource waste problem caused by static DAG is solved, and efficient execution and resource optimization of fault analysis are achieved.

CN122240371APending Publication Date: 2026-06-19WANBANG DIGITAL ENERGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
WANBANG DIGITAL ENERGY CO LTD
Filing Date
2026-03-17
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

Existing fault analysis systems based on large language models suffer from unnecessary task execution due to static DAGs, wasting computational resources and time, and making it difficult to dynamically adjust the execution of query tasks and data access methods according to actual fault problems.

Method used

The first major language model is used to determine the execution plan of the query task, and the second major language model is used to dynamically adjust the data access method. Combined with fault description and data source characteristics, the task status and data access are dynamically adjusted.

Benefits of technology

It improved the efficiency of query task execution, reduced resource waste, optimized the fault analysis process, and improved the overall performance of the system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122240371A_ABST
    Figure CN122240371A_ABST
Patent Text Reader

Abstract

This invention relates to the field of data processing technology, and provides a fault analysis method, system, device, and storage medium. The method includes: S1, acquiring a fault description and a preset query task set; S2, based on the fault description, obtaining an execution plan for each query task in the preset query task set based on a first major language model; S3, determining the data access method for the data required by the query tasks in the enabled state based on the execution plan of the query tasks and the characteristics of the data source; S4, executing the query tasks in the enabled state according to the execution plan of the query tasks and the selected data access method. This invention can perform semantic understanding and intent analysis based on the fault description, and then dynamically adjust the execution plan and data access method of the query tasks, avoiding meaningless activation of query tasks, thereby improving the execution efficiency of query tasks and reducing resource waste.
Need to check novelty before this filing date? Find Prior Art