基于BP神经网络安全阀的失效评价  

Evaluation of Safety Valve Failure based on BP Neural Network

作　　者：苏清苗[1] 何世权[1] 权辉[2] 饶文武[1] 孟静华[3] 

机构地区：[1]兰州理工大学石油化工学院,甘肃兰州730050 [2]兰州理工大学能源与动力工程学院,甘肃兰州730050 [3]重庆大学城市科技学院土木工程学院,重庆401331

出　　处：《新技术新工艺》2012年第9期97-101,共5页New Technology & New Process

基　　金：国家自然科学基金项目(51079066)

摘　　要：在实际使用过程中,安全阀的失效涉及诸多因素,且易受随机性因素的影响,具有模糊性和不确定性的特点,这些特点决定了整个系统是非线性动力学的。将安全阀失效的相关影响因素作为因素集,应用BP人工神经网络模型,利用MATLAB神经网络工具箱GUI,对样本数据进行仿真,然后采用训练好的网络对安全阀的现状进行评价。BP神经网络强大的记忆功能使得能从安全阀以往失效情况中总结出一般的失效规律,快速、准确地对安全阀现状做出判断。将BP神经网络用于安全阀失效评价,具有良好的评价效果和重要的实用价值。The failure of safety valves is affected by many factors in real applications, and even easily affected by many random factors, so there are characters of fuzziness and uncertainty. Which determine that the changes of system state do not according to specific rules or functions, it is non-linear dynamics. The factors affecting the failure of safety valve were taken as the factor sets, BP artificial neural network model of failure valuation was established. Then the sample data were emulated by using neural network toolbox GUI in the MATLAB software. Lastly safety valve was evaluated by using the trained networks, its results are same as the practical results. BP neural network has powerful memory, summed up the general failure of the law from the safety valve failure in the past, quickly and accurately judge the status of the safety valves. So application of ANN technique in safety evaluation has feasibility and strong adaptability.

关 键 词：安全阀 BP神经网络 失效 安全评价 MATLAB 

分 类 号：U261.154[机械工程—车辆工程]