RS-PSO-ELM下腐蚀管道失效压力预测

为提高腐蚀管道失效压力的预测精度并简化其计算过程,提出基于粗糙集(RS)和粒子群算法(PSO)融合极限学习机(ELM)的腐蚀管道失效压力预测模型。通过属性约简提取影响失效压力的关键因素,选用PSO优化ELM的输入权值和隐含层偏差,将归一化的核心指标数据代入计算。结果表明:该模型预测结果与实际值基本一致,与单一ELM模型相比,预测结果的均方差(MSE)降至0.255;与其他蚀管道失效压力评价模型相比,该模型预测结果的绝对误差平均值降至0.32。

Prediction of failure pressure of corrosion pipelines based on RS-PSO-ELM

In order to improve prediction accuracy of corrosion pipelines’ failure pressure and simplify its calculation process,a prediction model based on RS,PSO and ELM was proposed. Key factors that affected failure pressure were extracted in a way of attribute reduction,PSO was selected to optimize input weight and hidden layer deviation of ELM,and normalized core index data were computed in calculation.The results show that prediction of the model is basically consistent with actual values,its mean square error(MSE) is reduced to 0. 255 compared with single ELM model,and absolute mean error is reduced to 0. 32 compared with other assessment models of failure pressure.