基于PSO-Hybrid的不锈钢应力腐蚀开裂敏感性预测模型  被引量:2

Prediction Model of Stress Corrosion Susceptibility of Stainless Steel Based on PSO-Hybrid

在线阅读下载全文

作  者:蔡起衡 李光海[2] 王强 曹逻炜 CAI Qiheng;LI Guanghai;WANG Qiang;CAO Luowei(College of Quality and Safety Engineering,China Jiliang University,Hangzhou 310018,China;China Special Equipment Inspection and Research Institute,Beijing 100029,China)

机构地区:[1]中国计量大学质量与安全工程学院,杭州310018 [2]中国特种设备检测研究院,北京100029

出  处:《腐蚀与防护》2023年第6期96-102,共7页Corrosion & Protection

基  金:国家重点研发计划课题(2018YFC0809004)。

摘  要:为了提高不锈钢应力腐蚀开裂(SCC)敏感性预测的准确性与科学性,通过主成分分析(PCA)提取出不锈钢SCC行为的主要影响因素作为后续模型的输入,随后将机器学习不同流派的代表算法混合成Hybrid模型,并用粒子群优化(PSO)算法进行优化,提出不锈钢SCC敏感性的预测模型PSO-Hybrid。以某奥氏体不锈钢实测数据为例,对比预测值与实际值,以验证模型的可靠性与优劣性。结果表明:Hybrid思想有一定的可行性和科学性,且经PSO优化后,Hybrid模型的平均准确度与马修斯相关系数各提高了3.3%与8.3%,PSO-Hybrid模型的预测准确度高、稳定性好。In order to improve the accuracy and scientificity of the prediction of stress corrosion cracking(SCC)sensitivity of stainless steel,the main influencing factors of SCC behavior of stainless steel were extracted by principal component analysis(PCA)as the input of the subsequent model,and then the representative algorithms of different schools of machine learning were mixed into Hybrid model,and optimized by particle swarm optimization(PSO)algorithm,and the prediction model PSO-Hybrid of SCC sensitivity of stainless steel was proposed.Taking the measured data of an austenitic stainless steel as an example,the predicted value and the actual value were compared to verify the reliability and superiority of the model.The results showed that the Hybrid idea was feasible and scientific,and after PSO optimization,the average accuracy of the Hybrid model and the Matthews correlation coefficient were increased by 3.3%and 8.3%respectively.The PSO-Hybrid model had high prediction accuracy and good stability.

关 键 词:不锈钢 应力腐蚀开裂 主成分分析 粒子群优化算法 敏感性预测模型 

分 类 号:TG174[金属学及工艺—金属表面处理]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象