基于机器视觉的热障涂层内部裂纹在线检测方法及传热特性研究

Study of Internal Crack Detection Method based on Machine Vision and Heat Transfer Characteristics of Thermal Barrier Coatings

燃气轮机服役过程中,热障涂层(以下简称涂层)内部裂纹萌生和扩展是导致涂层失效的主要原因。通过数值重构方法获得了含不同长度裂纹的热障涂层(TBCs)微结构,基于耦合双分布格子波尔兹曼方法(DDF-LBM)建立了热障涂层与冷却气膜流动传热模型,研究了热障涂层内部和表面温度分布特性。结果表明:出现裂纹会极大地改变涂层的温度分布情况,增加涂层温度不均匀性,造成局部烧结,进一步产生应力集中,极易导致涂层分层断裂,从而影响其耐久性。同时,基于耦合检测算法(GEMSS)通过大量机器学习训练,提出了热障涂层内裂纹定位和长度估算的在线检测评估方法。该方法能有效确定裂纹位置,高精度估算裂纹长度,为高温叶片在线健康度评估和寿命预测提供理论基础和技术支撑。

Crack initiation and propagation inside the thermal barrier coatings are decisive factors that dominates the failure under services.The numerical reconstruction models of TBCs with different crack lengths were obtained.The temperature distribution characteristics of the surface and internal thermal barrier coatings were studied based on the coupled flow and heat transfer model of thermal barrier coatings and cooling film by using LBM-DDF method.The results show that the existence of cracks will greatly change the temperature distribution of the thermal barnier coating, resulting in increased temperature inhomogeneity and local sintering of the coating.This will lead to stress concentration, which can easily cause early layer fracture of the coating and affect the durability of the thermal barrier coatings.At the same time, an on-line orientation and length evaluation method for crack in thermal barrier coatings was proposed through a lot of trainings by machine learning based on GEMSS coupled detection method.This method can effectively determine the crack location and estimate the crack length with high accuracy.The results can provide theoretical basis and technical support for online health evaluation and life prediction of high temperature turbine blades.