基于神经网络的高温下(后)预应力损失分析

Analysis on Prestress Loss at High Temperatures by Neural Network

提出了基于神经网络的高温下(后)预应力损失的评价方法.针对影响高温下(后)预应力损失的复杂多变性和相当强的不确定性,建立了基于神经网络的高温下(后)预应力损失模型.运用神经网络强大的自适应、自组织、自学习的能力以及高度的非线性映射性、泛化性和容错性的特点,通过对训练样本的神经网络学习,建立了预应力损失与温度、初始有效预应力以及高温条件的网络关系,最终实现了对测试样本的预应力损失预报.

A method that estimates prestress loss at high temperatures is proposed based on neural net- work. The model of prestress loss based on the neural network is established aiming at the polytropy and uncertainty that affect the prestress loss at high temperatures. The neural network has the characteristics of self adapting, self organization, self learning, nonlinear mapping, generalization and fault tolerance. By training, the neural network is established to predict the prestress loss. The predicting results fit the actual results well.