基于粒子群优化神经网络的橡胶弹性管道最大应力的算法预测模型

Algorithm Prediction Model of Maximum Stress of Rubber Elastic Pipe Based on Particle Swarm Optimization Neural Network

使用基于粒子群优化神经网络完成橡胶弹性管道最大应力预测模型是在管道广泛应用与发展的前提下,在追求更便利、更实用的背景下产生的。管道崩裂事件频繁发生,预测管道崩裂的属性和情况可有效防止及避免事故的发生,对财产、人身安全起到一定的保护作用。以4种不同材质的橡胶弹性管道为研究对象,通过仿真实验分别计算上管壁压力在1000N、1500N、2000N,下管壁压力在200N、500N,管内温度30℃、40℃,环境温度30℃、20℃等条件下的最大应力值,共计算了120组数据,然后整理为粒子群优化神经网络算法的数据集,再通过调整此算法的参数,获得一种对橡胶弹性管道最大应力值预测准确率为96%的算法预测模型。

The use of particle swarm optimization based neural networks to predict the maximum stress of rubber elastic pipelines has emerged in the context of the widespread application and development of pipelines,in pursuit of greater convenience and practicality.Pipeline cracking incidents occur frequently,and predicting the nature and situation of pipeline cracking can effectively prevent and avoid accidents,playing a certain protective role in property and personal safety.Taking four different materials of rubber elastic pipes as the research object,simulation experiments were conducted to calculate the maximum stress values under conditions such as upper pipe wall pressure of 1000 N,1500 N,2000 N,lower pipe wall pressure of 200 N,500 N,inner pipe temperature of 30℃,40℃,and ambient temperature of 30℃,20℃,etc.A total of 120 sets of data were calculated,and then organized into a dataset of particle swarm optimization neural network algorithm.By adjusting the parameters of this algorithm,an algorithm prediction model with an accuracy of 96%for predicting the maximum stress value of rubber elastic pipes was obtained.