MQENet: a mesh quality evaluation neural network based on dynamic graph attention

Purpose The goal of this paper is to solve the shortcomings of subjectivity and singleness of traditional mesh quality metrics. By combining graph learning technology, an intelligent mesh quality evaluation model based on graph dynamic attention is proposed to achieve a more comprehensive and objective evaluation process. Design/methodology/approach The authors propose MQENet, a structured mesh quality evaluation neural network based on dynamic graph attention. MQENet treats the mesh evaluation task as a graph classification task for classifying the quality of the input structured mesh. To make graphs generated from structured meshes more informative, the authors also introduce two novel structured mesh preprocessing algorithms. Findings We successfully applied graph neural networks to the task of mesh quality evaluation. At the same time, our proposed graph neural network overcomes the shortcomings of traditional mesh quality evaluation metrics. Originality/value Two novel structured mesh preprocessing algorithms can improve the conversion efficiency of structured mesh data. Experimental results on the benchmark structured mesh dataset NACA-Market show the effectiveness of MQENet in the mesh quality evaluation task.