基于RS-ANN的通风系统可靠性预警系统

系统可靠性预警是度量系统运行状态偏离可靠性指标界线的强弱程度,确定预警等级和做出决策警示的过程。笔者在对目前国内外有关系统预警方法的分析比较基础上,针对矿井通风系统可靠性运行的实际状况,应用了粗糙集(RS)理论和神经网络(ANN)技术,提出了一种基于粗糙集神经网络(RSANN)的矿井通风系统可靠性预警方法:首先,建立了一套适合于矿井通风系统可靠性的预警指标体系;然后,利用人工神经网络与粗糙集理论的优势互补,以粗糙集作为前置处理系统优化指标结构,构建了基于RSANN的通风系统可靠性预警仿真模型,并应用该模型进行了实例验证。其结果表明,该模型的仿真结论与基于ANN的结论十分吻合,训练效率提高了667倍。

Early-warning System of Ventilation System Reliability Based on RS-ANN

Early-warning of system reliability is a process for measuring the deviation of the system running state from reliability index margin line, determining early-warning rank and making alerting decision. Referring to the analysis and comparison of relevant system early-warning methods at home and overseas, and aiming at the reliability running fact of mine ventilation system, combining rough set (RS) theory and artificial neuron network (ANN) technique, an early-warning system based on rough set & artificial neuron network (RS-ANN) is proposed.In this system, a series of reliability early-warning indexes adaptive to mine ventilation system are set; and then take the advantage of complementary superiority of ANN and RS, taking RS as former disposal system of optimized index structure, a ventilation system reliability early-warning model based on RS-ANN is constructed and exemplified with real case for validation. The results indicate that the simulated conclusion of this model fully corresponds to the ANN's, and its training efficiency increases 667 times.