神经网络在光纤故障预警系统中的应用仿真  

Application of Neural Network Technology in Optical Fiber Fault Early Warning System Simulation

作　　者：楼树美[1,2] 李淑玉[2,3] 

机构地区：[1]南京理工大学计算机学院,江苏南京210094 [2]信阳职业技术学院数学与计算机科学学院,河南信阳464000 [3]武汉理工大学计算机学院,湖北武汉430000

出　　处：《计算机仿真》2012年第12期255-258,共4页Computer Simulation

摘　　要：研究光纤故障预警系统的预警及时性问题。由于光纤承载业务量的不断增加,光纤线路产生大量的光功率告警信息,给预警系统带来巨大数据负担而影响预警及时性,传统的数据库方法对每一个光功率告警信息逐一进行预警分析,无法快速获取预警因子,而造成预警及时性不高。为解决上述问题,提出神经网络技术应用在光纤故障预警系统中。通过神经网络技术构建数据挖掘模型,将大量光功率告警信息输入到模型中进行数据分析,从中挖掘出具有最大决策价值的数据并快速提取其预警因子,避免逐一分析造成的预警因子提取效率不高的问题,利用预警因子快速完成预警。仿真结果表明,改进方法能够从大量光功率信息中快速获取预警因子,及时完成光纤故障的预警。Research warning timeliness problem of optical fiber fault early warning system. Due to the increase of fiber carrying portfolio, fiber optic lines have a lot of light power alarm information, which brings huge warning system data burden and influencing warning timeliness. Traditional database method for each optical power alarm information to process warning analysis one by one, can not quickly get warning factor, causing the problem of low warning timeliness. In order to solve this problem, this paper put forward a neural network technology application in optical fiber fault early warning system. Through the neural network technology a data mining model was built. A large mount of light power information were inputted to the alarm model to process data analysis, digging out the biggest decision value data and rapidly extract the warning factor, avoiding the low extraction efficiency problem of one by one to the early warning factor analysis. Finally, the warning factors were used to quickly finish early warning. Simulation results show that this method can quickly acquire warning factor from a large number of light power information, and timely complete the early warning of optical fiber fault.

关 键 词：神经网络 光纤故障 预警因子 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]