The crossbeam plays a vital role in computer numerical controlled milling machines, especially in machines with a gantry structure, as it directly influences the machining precision. In this study, a machine tool crossbeam was designed, and the modal frequency of the crossbeam was analyzed using the finite element model (FEM) analysis. In the improved structure obtained through FEM analysis, the X-type structure of the internal unit of the crossbeam was replaced by an O-type structure. The specific structure dimensions were further optimized using a neural-network algorithm and a nondominated sorting genetic algorithm. Finally, we calculated the effect of each crossbeam dimension on the mass, deformation, and frequency in a sensitivity analysis. After optimizing the crossbeam dimensions with respect to deformation, modal frequency, and mass, the structural characteristics of the original and optimized crossbeams were compared. After optimization, the mass and deformation were reduced by 7.45% and 3.08%, respectively, and the modal frequency was increased by 0.42%. These results confirm that the optimization improved the performance of the crossbeam structure.

横梁在计算机数控铣削机床中，尤其是在具有龙门结构的机床中，起着至关重要的作用，因为它直接影响加工精度。在这项研究中，设计了一种机床横梁，并使用有限元模型（FEM）分析来分析横梁的模态频率。在通过有限元分析获得的改进结构中，横梁内部单元的X型结构被O型结构替代。使用神经网络算法和非支配排序遗传算法进一步优化了特定的结构尺寸。最后，我们在灵敏度分析中计算了每个横梁尺寸对质量，变形和频率的影响。在针对变形，模态频率和质量优化横梁尺寸之后，比较了原始横梁和优化横梁的结构特征。优化后，质量和变形分别减少了7.45％和3.08％，模态频率增加了0.42％。这些结果证实了该优化改善了横梁结构的性能。