轧辊表面形态的振动激励机理及识别方法研究

Due to contact stress and high alternating loads, crack, spalling and other kinds of local defect can be easily found on the roll surface in the rolling process; besides, the surface morphology of the roll may be affected by the grinding process; these two problems lead to the surface quality of the steel strip hard to be guaranteed. Therefore, it is of significance to detect the abnormal vibration characteristics caused by local defects and morphology on roll surface in time for high-quality steel strip production and economic efficiency improvement. This project is aiming to investigate the underlying vibration mechanism and then develop some specific fault diagnosis methods for local defects and morphology with different shapes and levels. The main work includes: (1) establishing the contact process dynamic models between rolls and between roll and steel strip, with different kinds of local defect and morphology types on roll surface are taken into consideration, and then the underlying vibration mechanism and characteristics will be studied based on the dynamic models; (2) The influence of roll surface defects and morphology with different shapes and levels on the oil film lubrication state as well as the lubrication state will be discussed, further the regular pattern of the vibration characteristics to local defects and morphology, moreover, the relationship between the vibration characteristics and surface quality will be derived; (3) In order to effectively identify the local defects and morphology on roll surface, new guidelines for adaptive wavelet construction are proposed. Above all, the previous research will provide both theoretical and technical support for real-time local defects and morphology detection for rolling mill.

轧制过程中，轧辊在接触应力与高交变载荷的作用下，不可避免的会在轧辊表面产生裂纹、剥落等缺陷；此外，在磨削过程中，轧辊表面形态也会受到不同程度的影响，那么带钢表面质量就难以得到保障。因此，及时识别轧辊表面形态诱发的异常振动特征，对提高板带产品质量及企业经济效益，具有重要意义。本课题旨在对轧辊表面形态诱发轧机振动的激励机理及识别方法进行研究，在解明振动激励机理的基础上，研究其振动特征及识别方法，主要的研究内容包括：（1）构建轧辊表面缺陷时轧辊之间，轧辊与带钢之间接触过程动力学模型，解明其激励机理与边频产生机理；（2）建立油膜润滑情况下不同轧辊表面形态的动力学模型，查明其振动特征的变化规律，搭建轧机振动与振纹产生机理之间的桥梁；（3）提出构造自适应小波的新准则，准确定位轧辊表面局部缺陷尺寸大小。本课题将为实现对轧辊表面形态和早期故障特征的实时监测及识别，提供详实的理论分析依据与试验数据支撑。

我国作为钢铁生产大国，不锈钢产量约占全球不锈钢总产量的48.3%。随着国家供给侧的深入改革，产能已不是钢铁行业发展的主格调，只有拥有高品质、高技术产品的生产能力，才能更好地维护行业发展。在交变载荷及接触应力作用下，轧辊表面不可避免的会产生局部缺陷，严重影响带钢表面质量，因此及时识别轧辊表面形态诱发的异常振动特征，采取相应控制措施才能有效保证产品质量。但轧机轧制工况多变，多辊系接触，振源多，若不开展从机理到表征的系统性研究，难以实现异常振动特征的提取和保证带钢表面质量。本课题围绕轧辊表面形态诱发轧机振动的激励机理及识别方法开展了如下研究：（1）构建了不同轧辊表面形态时轧辊接触的时变刚度模型，考虑缺陷边缘接触效应，结合10自由度辊系动力学模型，从动力学角度解明了边频产生的机理。（2）建立了不同缺陷尺寸时变油膜刚度模型，研究了缺陷尺寸对油膜刚度变化的影响规律，并结合弹流润滑理论，提出了抑制带钢表面振纹的方法；同时研究了不同缺陷类型时变刚度变化规律及振动特性变化规律，为消除工况变化带来的影响，利用无量纲参数实现了振动特征的有效提取。（3）引入相似性判断理论，判断构造自适应小波函数的整周期性，根据小波理论构造自适应小波函数，利用构造的自适应小波函数，实现了在不停机状态下对轧辊表面缺陷尺寸大小的实时识别，精度达到90%以上。