摘要
针对传统关联规则算法在电力设备故障诊断中因忽视噪声消除导致的低准确率问题,提出了一种提出基于人工智能和深度学习的电力设备故障诊断方法.通过捕捉设备正常运行状态下的信号变化,收集频域故障数据,并利用度量空间距离映射为综合灰度级故障图像,在此基础上,运用过拟合函数消除图像噪声,得到纯净的故障数据,采用人工智能算法融合这些数据,形成单一特征的设备故障矢量集合,通过深度学习构建故障诊断模型,输入矢量数据,输出故障类型,实现精准诊断.仿真实验结果表明,该方法具有更高的诊断准确率,具有实际应用价值.
Abstract
A power equipment fault diagnosis method based on artificial intelligence and deep learning is proposed to address the issue of low accuracy caused by neglecting noise elimination in traditional association rule algorithms.By capturing signal changes during normal operation of the equipment,collecting frequency domain fault data,and mapping it into a comprehensive grayscale fault image using metric spatial distance,overfitting functions are used to eliminate image noise and obtain pure fault data.Artificial intelligence algorithms are used to fuse these data to form a single feature set of equipment fault vectors.A fault diagnosis model is constructed through deep learning,and vector data is input,Output fault types to achieve accurate diagnosis.The simulation experiment results show that this method has higher diagnostic accuracy and practical application value.
国家科技期刊平台
NSTL
万方数据