基于自组织竞争神经网络的抽油井故障诊断系统

Oil field fault diagnosis system based on self-organizing competitive neural network

本课题是针对油田边远井、孤立井能够更方便的进行故障诊断,创新的采用手持终端与自组织竞争神经网络相结合的方法进行油井故障诊断。以抽油机作为研究对象,从油井的示功图入手,利用自组织竞争神经网络对抽油井示功图进行智能识别分类,实现油井故障的自动诊断。实验表明,基于神经网络的故障诊断系统在手持android终端能够成功实现,并且诊断正确率在97.3%以上。

This article puts forward a new method based on hand-held terminal and self-organizing competitive neural network for the remote oil field wells to fault diagnosis. Pumping as the research object, indicator diagram is used to acquire to the fault characteristic, using the self-organizing competitive network to achieve intelligent classification and fault diagnosis automatically. The experiments show that the fault diagnosis system which is based on neural network in the android enddevice has been successfully achieved and the diagnostic accuracy rate is more than 97.3%.