Controlled Compensation of Defects in the Rings Geometry in Radial-Axial Ring-Rolling by Using Image Processing Methods

径向-轴向环件轧制工艺已成为生产直径范围为200 mm至13 m的热轧无缝环件的主要工艺。尽管环件径轴向轧制技术在过去得到了广泛的研究和发展，但在环件径轴向轧制过程中仍然存在几何误差和缺陷。如果环的几何形状在其制造过程完成后仍有缺陷，则将产生缺陷。为了避免因超过制造公差而导致的拒收，材料余量是必要的，因为环的几何形状可能会出现缺陷。特别是对于生产具有大尺寸的环，该材料余量可能是相当大的。今天，有很多方法被设计来避免这些错误。然而，这些方法中的大多数必须由操作员手动控制或作为预防措施激活。目前还不可能在此过程中自动检测环几何形状中的错误，以便采取适当的对策。这种可能性已经在DFG资助的一个名为Grundlagen zur Steuerung des Radial-Axial Ringwalterdurch den Einertium von Bildverarbeitungssystemen的研究项目中得到了发展。该项目介绍了一个图像处理系统的集成，该系统能够检测环件几何形状在径向-轴向环件轧制过程中的误差。基于这些结果，本研究项目旨在开发控制策略，以自动避免环的几何形状的错误。这些控制策略的输入量是由图像处理系统采集的关于环件几何形状的信息和传统的径向-轴向环轧机的环件测量系统的数据。为了实现这种方法，示例性地选择了圆度误差和爬坡误差作为应用领域。一个子目标是编目文献中描述的方法，并量化这些方法的控制变量，以避免不必要的环的几何形状的错误。基于这些方法的控制策略可以随后被开发并付诸实践。作为一个独立的软件实现和集成到环轧机的控制也是这个项目的一个子目标。沿着策略的制定和最终的结果评估，将进行滚动实验。首先，控制变量对环几何形状中的特定误差的外观的影响是这些实验的重点。通过改变这些控制变量，可以观察和记录影响。在进一步的实验中，可以实现控制的参数化和所开发的方法的功能能力的验证。应始终考虑将开发的方法扩展到环几何形状的其他缺陷的可能性。

The process of radial-axial ring-rolling has established itself as the main process for producing hot rolled seamless rings within a diameter range of 200 mm up to 13 m. Even though the field of radial-axial ring-rolling has been intensively explored and developed in the past, there are still errors and defects appearing in the rings geometry during the process. Reject will be produced if a rings geometry is still defective after its manufacturing process has been finished. To avoid this reject by exceeding the manufacturing tolerances, material allowance is necessary as defects may occur in the rings geometry. Especially for producing rings with large dimensions, this material allowance may be substantial. Today there is a high number of methods which have been designed to avoid many of these errors. However, most of these methods must be controlled manually by the operator or are activated as a preventative measure. There has not yet been the possibility to detect errors in the rings geometry automatically during the process in order to initiate adequate countermeasures. This possibility has been developed in a DFG-funded research project named Grundlagen zur Steuerung des Radial-Axial Ringwalzens durch den Einsatz von Bildverarbeitungssystemen. This project has introduced the integration of an image processing system which is able to detect errors in the rings geometry within the process of radial-axial ring-rolling. Based on these results, this research project aims at the development of control strategies to automatically avoid the errors of the rings geometry. Input quantities of these control strategies are the information about the rings geometry gathered by the image processing system and the data of the conventional ring measuring system of the radial-axial ring-rolling machine. To realize this approach, the errors of circularity and climbing are exemplarily chosen as the application field. A sub-goal is to catalogue the methods described in literature and to quantify the control variables of these methods to avoid unwanted errors of the rings geometry. Control strategies, based on these methods, can subsequently be developed and put into practice. The implementation as a standalone software and the integration into the control of the ring-rolling machine is also a sub-goal of this project. Along with the development of the strategies and finally to evaluate the results, rolling experiments will be executed. First, the influence of the control variables on the appearance of the particular errors in the rings geometry is in the focus of these experiments. By varying these control variables, the influence can be observed and documented. In further experiments, a parameterization of the control and a verification of the functional capability of the developed approach can be achieved. The possibility of an extension of the developed approach on other defects of the rings geometry shall always be considered.