基于神经网络技术主通风机首次故障时间分析.pdfVIP

30 6 V o.l 30 N o. 6 2005 12 JOURNA L OF CH INA COA L SOC IETY D ec 2005 : 0253- 9993 2005) 06- 074 1- 05 阎善郁, 王洪德 , 116028) : 针对矿用主通风机系统的可维修特性, 将神经网络技术与系统可靠性工程理论相结 , 提出基于 BP 网络的首次故障时间模型识别方法和基于自适应神经网络的可靠性参数估计方法, 旨在使矿井通风系统可靠性工程与人工智能等学科理论的交叉渗透. 将机器学习引入矿用主通风 机可靠性研究过程中, 以提高系统可靠性分布类型识别和参数估计的自动化和智能化水平. : 主通风机首次故障时间; BP神经网络; 模型识别; 参数估计 : TD441 : A Analysis on main fan first failure time based on neural network technique YAN Shanyu, WANG H ongde (C ol lege of E nvironm en t S cience Eng ineering, D a lian J ia otong Universi ty, D a lian 116028, Ch ina ) Abstract: A mi ed at the reparable characterist ic o f m a in vent ilated m ach ine system u sed in m ine, com b ined neu ron network techn ique w ith system reliab ility eng ineering theo ries, and put forw ard the m ethod abou t the system first failure t mi e ident ify m odel based on the BP neural newt ork and the m ethod about system re liab ility param eters esti m at ion based on the adaptive neutron newt ork. The purpose are to m ake the m ine ventilated system reliab ility eng i neer ing in filtrated acrossw ith the artificial intelligence sub ject theory etc, and inducted the m achine learn ing to the course of the study on m ain ven tilated m ach ine reliab ility u sing in m ine, and advanced autom at izat ion and inte lli gentization leve l to identify system re liability d istribut ion m ode l and to estmi ate the param eters. K ey words: m a in fun first failure t mi e; BP neura l newt ork; m odel recognize; param eter estmi at ion . , . ,