摘要
管道运输油气具有损失少、耗时少等特点,但是管道安全事故也时有发生,对管道寿命的有效预测是解决管道安全问题的重要手段.针对油气管道寿命预测现状,提出了一种基于疲劳寿命影响因素的BP神经网络疲劳寿命估值方法,并建立BP算法模型;分析各特征要素,确定输入层节点数;用收集到的样本数据,训练并测试BP算法的稳定性及精确度.结果表明,精确度达90%以上,从工程应用角度来说可进入实用.
Pipelines transportation has loss-reducing and time-saving traits, but has the problem of pipelines safety accidents. Therefore, effective forecasting of pipelines life is important means to solve this problem. The paper, based on the status quo of oil and gas pipelines life prediction, aims to put forward a method of BP neural network regarding fatigue life impact factors, to establish BP algorithm model, to analyzed attribute factors, to decide node number of input floor, and to test stability and precision of BP algorithm through sample data collected. Results prove that the precision is greater than 90% and can be put into practice from point of view of engineering application.
出处
《温州大学学报（自然科学版）》
2008年第4期14-18,共5页
Journal of Wenzhou University(Natural Science Edition)
关键词
BP算法
油气管道
疲劳寿命
估值
BP algorithm
Oil and gas pipelines
Fatigue life
Estimation
作者简介
万年红（1977-），男，江西新建人，讲师，硕士研究生，研究方向：神经网络，面向对象
