In order to shorten the process chain in manufacturing and thus to optimize the manufacturing time, current researches investigate the possibility of combining two process steps, turning and hardening.During cryogenic turning of metastable austenitic steel, deformation induced hardening in the surface layer of the workpiece can be achieved by applying high passive forces onto the workpiece [1]. This leads to an increase in the wear resistance [2] as well as the fatigue strength [3]. For the employed austenite-martensite phase transformation it is necessary to maintain low process temperatures, typically below room temperature. Thus, cryogenic cooling has to be applied to counteract the heat, generated during machining [4].For a better understanding of the influence of different cutting and cooling parameters on the temperature field during cryogenic turning, and thus martensite formation, knowledge of the exact temperature distribution in the workpiece and the workpiece surface temperature in the contact zone is essential. Since in situ measurements of the latter are hardly possible, an inverse determination via transient finite element simulation is performed [5].In order to model cryogenic turning, material properties, thermal loads and heat transfer coefficients defining convective heat transfer, have to be determined first. These model parameters are investigated independently in stand-alone experiments, applying only one thermal load at a time. The magnitudes of these values are obtained by iteratively modifying the initial values in the model until correspondence with measured temperature data from the experiment is achieved.The present finite element approach only takes thermal loads into account and is performed in the finite element program FEAP (Finite Element Analysis Program) with an Eulerian mesh, which requires special consideration of the rigid body rotation of the workpiece. The Eulerian treatment results in an unsymmetrical system matrix, and a special stabilisation is required to avoid numerical oscillations in the time integration scheme [5].
ModelThe model which is used for the finite element simulations represents the cylindrical geometry of the investigated workpiece with a diameter of d = 14.4 mm in the machined area, see e.g. [2], [5]. Whereas the material properties are assumed to be constant and known, see Table 2, process-dependent parameters like the heat transfer coefficients between the workpiece and the surrounding gas, cooling the workpiece by natural or forced convection,and the heat input by the toolq in have to be determined first. These parameters describe the unknown parameters of an inverse boundary value problem, which has to be solved. This is done by comparing the simulated temperature data with measured data from the experiment by iteratively incrementing the value of the investigated parameter in the model. Experimental data is provided by three thermocouples, measuring the workpiece's inner temperatures.For simplifying the determination of the unknown parameter values...