为研究采煤机螺旋滚筒在多种赋存条件下的振动特性，以MG2×55/250–BWD型薄煤层采煤机为工程对象，优化煤岩接触模型，建立与实际赋存条件相似的多种不同截割工况下煤壁离散元模型。结合DEM–MFBD(Discrete Element Method-Multi Flexible Body Dynamics)双向耦合数值模拟方法搭建采煤机截割部刚柔耦合虚拟样机模型与煤壁离散元模型的双向耦合试验平台，通过仿真试验得到不同煤岩工况下螺旋滚筒的截割过程，并分别对其振动特性的变化规律展开分析。研究结果表明：螺旋滚筒在截割过程中，三向均出现不同程度的振动，其中截割阻力方向振动加速度最大，牵引阻力方向振动加速度次之，侧向力方向振动加速度最小。随着模型中夹矸硬度以及层数比例的增加，截割过程中螺旋滚筒的振动强度不断加剧，最大振动加速度有效值的差值达到4 403.149 mm/s²。利用短时傅里叶变换将一维振动信号转化为二维时频谱图像，得到不同煤岩工况下振动信息变化特征在时频域中完成较好保留，其时频谱图像的特征样本效果优于各工况的时域一维信号曲线，主频能量位置、范围大小、特征团形状等信息具有明显区别，即使遇到夹矸层数不同，夹矸坚固性系数也存在差异的复杂工况，其时频谱图像中能量特征的分布形式也具有显著差别。通过振动模态分析发现，随着煤壁中含有夹矸硬度的增加，各部位的变形量均发生变化，其中截齿部位变化最为强烈。基于相似理论搭建采煤机振动信号测试试验平台，对不同煤岩工况条件下螺旋滚筒截割过程进行了测试研究，通过追踪螺旋滚筒的振动状态，发现其振动变化规律与双向耦合数值模拟一致。试验测试得到DEM–MFBD数值模拟方法获取的螺旋滚筒振动加速度有效值与依据相似比反推的试验数据之间的误差小于DEM离散元数值模拟方法与实验数据之间的误差，验证了DEM–MFBD数值模拟方法的准确性。研究结果对于提升螺旋滚筒工作可靠性具有重要意义，同时也为采煤机智能化开采的煤岩截割状态识别系统搭建过程中数据信息的获取提供了一种新的方法。

In order to study the vibration characteristics of shearer’s spiral drum under various occurrence conditions, the coal-rock contact model was optimized by taking the MG2×55/250–BWD thin seam shearer as the engineering object, and discrete element model of the coal wall under various cutting conditions similar to the actual occurrence conditions were established. Combined with DEM–MFBD (Discrete Element Method-Multi Flexible Body Dynamics) two-way coupling numerical simulation method, the two-way coupling simulation platform of the rigid-flexible coupling virtual prototype model of the shearer cutting section and the discrete element model of the coal wall was built. Through simulation, the cutting process of the spiral drum under different coal wall working conditions was obtained, and the variation law of the vibration characteristics of the spiral drum under different occurrence conditions was analyzed. The results show that: during the cutting process, the spiral drum vibrates to different degrees in three directions, of which the vibration acceleration in the cutting resistance direction is the largest, the vibration acceleration in the traction resistance direction is the second, and the vibration acceleration in the lateral force direction is the smallest. With the increase of the hardness of the gangue and the proportion of the number of layers in the model, the vibration intensity of the spiral drum during the cutting process continues to increase, and the difference between the effective values of the maximum vibration acceleration reaches 4 403.149 mm/s². The short-time Fourier transform was used to convert one-dimensional vibration signal into two-dimensional time-frequency spectrum image, and the characteristics of vibration information change under different working conditions were well preserved in time-frequency domain. The feature sample effect of time-frequency spectrum image is better than that of time-domain one-dimensional signal curve under various working conditions. The information such as the location, scope and shape of feature cluster of dominant frequency energy have obvious differences. Even if the number of rock parting is different, the distribution form of energy characteristics in the time-frequency spectrum image is also significantly different under complex working conditions where there are differences in the firmness coefficient of the rock parting. Through vibration modal analysis, it is found that with the increase of the hardness of the gangue in the coal wall, the deformation of each part changes, and the change of the pick part is the strongest. Building a vibration signal testing experimental platform for shearer based on similarity theory. The cutting process of spiral drum under different working conditions was tested experimentally. By tracking the vibration state of the spiral drum, it is found that the vibration change law is consistent with the results of the two-way coupling numerical simulation. The error between the effective value of the vibration acceleration of the spiral drum obtained by the DEM–MFBD numerical simulation method and the experimental data inferred based on similarity ratio is smaller than the error between the DEM discrete element numerical simulation method and the experimental data, which verifies the accuracy of the DEM–MFBD numerical simulation method. The research results are of great significance for the working reliability of the lifting spiral drum, and also provide a new method for obtaining data signals during the construction of the coal and rock cutting state recognition system for "unmanned" intelligent mining.