基于人工神经网络的长周期储存式压力容器安全评价研究

Study of Storage Pressure Vessel Safety Evaluation for Long Period Based on Artificial Neural Network

本文基于人工神经网络,设计了一个长周期储存式压力容器安全分析评价系统。按压力容器自身的特点进行建模,选择三种类型缺陷为模拟对象,利用有限元应力分析进行应力计算,获取各种状态的应力数据作为训练的样本数据,并选用带二次动量项的BP算法对样本数据进行学习,进而建立长周期压力容器安全评价智能系统软件。最后将该评价智能系统软件计算的结果与GB/T 19624—2004《在用含缺陷压力容器安全评定》计算结果进行比对,系统准确性高。利用该软件使压力容器安全评价变得方便、快捷、简单。

Ba, ed on artificial neural network,a pressure vessel safety analysis and evaluation systemsfor long period is designed. In this paper, based on the tank modeling with characteristics of the container, stress of pressure vessel with threq, kinds of defects is calculated by finite element stress analysis. The stress data in different condition is obtained as ~ training sample data, and the BP algorithm with momentum for the second sample data is chosen to learn the samp e data, the intelligent pressure vessel safety assessment system software for long period is built. The final results arc certified with high accuracy comparing to the national standard GB/T 19624. In this paper, The software makes the o ＂essure vessel safety assessment become more convenient fa^ter and ~n^ie.r