摘要
随着电力系统复杂程度的提高,电力设备故障诊断面临严峻挑战。本文设计了一种基于多模态深度学习的电力设备故障诊断系统,提出多模态数据融合策略,利用门控循环单元、卷积神经网络和注意力机制实现多源数据特征提取与动态融合,并结合变分自编码器与极限学习机完成故障类型识别与程度评估。系统通过主动学习与增量学习提升鲁棒性,功能性测试实验验证了系统的高准确性与实时性。
With the increasing complexity of the power system,fault diagnosis of power equipment is facing severe challenges.This article designs a power equipment fault diagnosis system based on multimodal deep learning,proposes a multimodal data fusion strategy,uses gate controlled recurrent units,convolutional neural networks,and attention mechanisms to achieve multi-source data feature extraction and dynamic fusion,and combines variational autoencoder and extreme learning machine to complete fault type recognition and degree evaluation.The system improves robustness through active learning and incremental learning,and functional testing experiments have verified the high accuracy and real-time performance of the system.
作者
周元峰
ZHOU Yuanfeng(China Electronic Systems Technology Co.,Ltd.,Beijing 100141)
出处
《软件》
2024年第12期171-173,共3页
Software
关键词
多模态深度学习
诊断系统
电力设备
multimodal deep learning
fault diagnosis
electrical equipment
作者简介
周元峰(1988-),男,内蒙古赤峰人,本科,高级工程师,研究方向:电力行业数智化。
