KontROLL - Contact stress–controlled deep rolling process

The maximum stress that a component can withstand under cyclic loading is generally significantly lower than the maximum stress it can endure under static loading. The reason for this is the formation of cracks that gradually propagate under cyclic stress, eventually leading to fatigue fracture and, consequently, to component failure. However, suitable methods can counteract crack growth and thus fatigue. Surface treatments such as deep rolling are an effective method. Deep rolling tools are equipped with a roller that is pressed against the workpiece surface by springs. The rolling element plastically deforms the surface layer of the workpiece, thereby introducing compressive residual stresses into it. The degree of plastic deformation—and thus the potential increase in service life—depends significantly on the selected rolling force and the geometry of the workpiece. To achieve consistent process results and, in particular, uniform residual stresses in the workpiece, the rolling force must be adapted to the geometry of the component.

Objectives

The aim of the project is therefore to reduce unwanted variations in the resulting residual stresses during deep rolling by controlling the process. This should ensure that the desired service life of the machined components is achieved.

Benefits

Controlling the deep rolling process not only improves the process result. At the same time, it significantly simplifies the application and setup of the process in practice, so that the procedure can be used even with little prior knowledge.

Approach

To achieve the project goal, first a process model that predicts the necessary rolling forces depending on the workpiece geometry will be developed. For this purpose, a neural network will be trained. For the training, FE simulations provide the data basis. Subsequently, a control system will be implemented to ensure compliance with the predicted rolling force. The control system is based on the measurement of the rolling force during the process by a sensory tool that is already known from the predecessor project “ProMeFe” (ZF4070523LP9).

Are you also interested in a cooperation project?

Contact Jan Berlin via email at berlin@ifw.uni.hannover.de or by phone at +49 511 762 18095.