基于PCA-SVM的油气管道腐蚀速率预测技术研究

为了对油气管道的腐蚀速率进行有效预测,对影响油气管道腐蚀速率的原因进行了梳理和总结。应用PCA主成分分析法对10大影响因素进行降维处理,选取累计贡献率大于98%的前8项影响因素代替原所有影响因素,忽略流速和压力的影响,利用LS算法对惩罚因子C、核参数σ和不敏感参数ε的取值进行了寻优。对PCA-SVM、PCA-BP、PCA-GRNN和PCA-WNN四种腐蚀速率预测模型进行对比,其中PCA-SVM模型的预测效果最好,平均绝对误差为1.04%,均方根误差为0.01339,但训练时间比其他模型长,今后应集中进行算法优化。

Prediction Technology Study of Oil and Gas Pipeline Corrosion Rate Based on PCA-SVM

In order to effectively predict the corrosion rate of oil and gas pipelines,reasons affecting the corrosion rate of oil and gas pipeline are reviewed and summarized.The principal component analysis(PCA)method is used to reduce the dimension of the 10 major influencing factors.The first 8 influencing factors with a cumulative contribution rate not less than 98%are selected to replace all the original influencing factors,ignoring the influence of flow rate and pressure.By LS algorithm,the values of penalty factor C,kernel parameter σ and insensitivity parameter ε are optimized.Finally,four prediction models of PCA-SVM,PCA-BP,PCA-GRNN,and PCA-WNN are compared.Among them,the PCA-SVM model has the best prediction effect,with an average absolute error of 1.04%and a root mean square error of 0.01339.However,its training time is longer than other models,so algorithm optimization should be concentrated in the future.