【摘要】 准确的构架有限元模型可以为转向架前期设计和后期优化提供保障，从而避免激励与构架模态共振发生事故，但构架有限元建模产生误差是不可避免的。为获得准确的有限元模型，该文章推导了灵敏度矩阵加权算法，基于该算法，以构架模态试验数据为基准，以焊缝区等效弹性模量和主要板材厚度为修正参数，以模态频率和振型MAC(modal assurance criterion)值误差最小化为修正目标，对转向架构架有限元模型进行修正，修正结果显示与频率试验值误差控制在1%以内。另外，研究发现构架MAC值修正效果不明显，仅考虑频率修正项时依然可以获得准确模型。更多还原

【Abstract】 Correct finite element model of frame can provide guarantee for early design and later optimization of bogie to avoid accidents caused by resonance between excitation and modes of frame. However, errors caused by finite element modeling of frame are inevitable. Here, to obtain an accurate finite element model, a sensitivity matrix weighting algorithm was derived. Based on this algorithm, the finite element model of bogie frame was modified taking the frame modal test data as the benchmark, equivalent elastic modulus of weld zone and thickness of main plates as correction parameters, and error minimization for modal frequency and MAC(modal assurance criterion) value of vibration mode as the correction target. The correction results showed that the error between the modal frequency calculated with the modified finite element model and frequency test value is controlled within 1%; the MAC value correction effect of bogie frame is not obvious, and the correct model can still be obtained when only the frequency correction term is considered.更多还原