Tao, Laifa (Hangzhou International Innovation Institute of Beihang Universit), Li, Shangyu (Beihang University), Huang, Qixuan (Beihang University), Zhao, Zhengduo (Beihang University), Su, Xuanyuan (Beihang University), Jin, Kaixin (Beihang University)

Research of Preventive Maintenance Plans Based on Maintenance Knowledge Fusion Large Model

Scheduled for presentation during the Invited session "Estimation and Prediction" (FrB1), Friday, November 1, 2024, 11:30−11:50, Room T1

7th IFAC Conference on Engine and Powertrain Control, Simulation and Modeling, Oct 30 - Nov 1, 2024, Dalian, China

This information is tentative and subject to change. Compiled on January 2, 2025

Abstract

As an important energy equipment, wind power equipment have a wide range of applications worldwide. But its high equipment maintenance costs seriously affect the profits of wind power generation enterprises. Fault prediction and health management technology, as key technologies for optimizing maintenance methods and reducing maintenance costs, are of great significance for reduce the failure rate of wind power equipment, reduce maintenance costs, and promote the rapid development of the clean energy industry to generate excellent preventive maintenance plans for wind power equipment. However, the current development of wind power equipment maintenance plans heavily relies on expert experience and lacks reliable explanatory support. In this case, we propose a preventive maintenance plan generation method for wind power equipment based on maintenance knowledge fusion large model. Our solution generation process no longer relies on expert experience, but relies on the reasoning ability of the latest artificial intelligence technology large language model. We fine tune the pretrained base large model using data and knowledge from wind power equipment fault manuals and maintenance manuals, and design reasonable question and answer prompts to achieve intelligent generation of wind power equipment preventive maintenance plans. Finally, the effectiveness of the above method was verified through the manual materials of UP77 and UP82 fan equipment.