Prediction of long-term creep life of 9Cr–1Mo–V–Nb steel using artificial neural network. (January 2020)

• Record Type: Journal Article
• Authors: Liang, Tuo; Liu, Xinbao; Fan, Ping; Zhu, Lin; Bi, Yao; Zhang, Yagang
• Is Part Of: International journal of pressure vessels and piping. Volume 179(2020)
• Publication Date: 2020-01
• Subjects: 9Cr–1Mo–V–Nb steel -- Creep life -- Time-temperature parameter -- Artificial neural network; Pressure vessels -- Periodicals; Pipe -- Periodicals; Récipients sous pression -- Périodiques; Tuyaux -- Périodiques; Pipe; Pressure vessels; Periodicals; 681.76041
• DOI: 10.1016/j.ijpvp.2019.104014
• Languages: English
• ISSNs: 0308-0161

Abstract

Abstract: In this work, the method of artificial neural network was employed to predict the long-term creep rupture time of 9Cr–1Mo–V–Nb steel using the NIMS datasheet. In order to verify the performance of this method, the long-term creep rupture times of 23 000–41000 h were predicted using the data lower than 17 000 h. Meanwhile, the detailed analyses were carried out by comparison with the traditional time-temperature parametric (TTP) methods, such as Larson-Miller, Manson-Harferd, and Orr-Sherby-Dorn method. The results showed that by the artificial neural network method, the predicted creep rupture times above had an average relative error of 17%, which was significantly lower than those of TTP methods. It further demonstrated that the artificial neural network offers a convenient tool to predict the accurate creep rupture time of 9Cr–1Mo–V–Nb steel due to its robust ability in law learning and extrapolation generalization. Highlights: TTP methods show poor results of long-term creep life prediction. Results of neural network are better in short-term prediction. Results of neural network are accurate enough in long-term prediction.