基于粗糙集和BP神经网络的石化管道外腐蚀程度预测应用

Application on Petrochemical Pipeline Outside Corrosion Prediction Based on RS and BPNN

研究管道外腐蚀程度预测问题;传统预测方法有专家评分系统等,而造成管道腐蚀的原因众多,专家对各个因素的偏好不一样,因此这类方法个人主观因素较强,预测的结果不够客观、准确;为了克服个人主观影响,提高预测精度,提出粗糙集-BP神经网络预测模型;该预测模型首先利用粗糙集理论消除管道腐蚀影响因素中的冗余因素,然后利用BP神经网络依据处理后的数据进行学习建模,并测试;仿真结果表明了该模型简洁性、快速性和有效性。

This article studies the outer pipe corrosion prediction.Traditional forecasting methods is an expert rating system.The pipeline corrosion caused by many reasons,Experts on various factors preferences are not the same,so a strong subjective factors of such methods,the predicted results were not objective and accurate.In order to overcome the subjective influence to improve forecasting accuracy,we put forward the Petrochemical Pipeline Outside Corrosion model based on rough set theory combined with BP neural network.The prediction model using rough set theory to eliminate redundancy factors from pipe corrosion factors first.,and then use processed data modeling and testing based on BP neural network.Simulation results show that the model is simple,fast and effective.