Rapid identification of erosion thinning and scaling thickening of inner wall of circular tube based on inverse heat conduction problem method

THERMAL SCIENCE AND ENGINEERING PROGRESS(2024)

引用 0|浏览0
暂无评分
摘要
Due to the erosion of fluid and the deposition of fouling in pipes, the internal defects of pipelines are widespread and seriously threaten the safety of industrial production. In this paper, an indirect inversion model based on modified one-dimensional correction method (MODCM) is proposed to identify the erosion thinning and fouling deposition of the inner wall of a two-dimensional circular tube. By constructing the relationship between the equivalent thermal conductivity and the geometric parameters of defect, the inversion of defects is transformed into the inversion of the equivalent thermal conductivity, which overcomes the difficulty of re-gridding caused by the change of the geometric domain during the iterative process, and effectively improves the calculation efficiency. In order to explore the validity and timeliness of the constructed models, the numerical experiments are carried out. The influence of measurement errors, thermal conductivity of pipe and thermal conductivity of fouling on the identification results are systematically analyzed. The research results show that the thermal conductivities of pipes and fouling have little effect on the inversion results, and the inversion accuracy decreases with the increase of measurement errors. When the measurement error is sigma = 0.5 celcius, the identification error remains about 0.2 %, which shows that the method has good robustness. In addition, compared with the typical direct identification models of Levenberg-Marquardt method (LMM) and conjugate gradient method (CGM), the proposed indirect inversion model is superior in computational efficiency and tracking ability of discontinuous boundaries.
更多
查看译文
关键词
Inverse heat conduction problem,Defect identification,Indirect inversion model,Equivalent thermal conductivity
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要