Determination of Coefficient of Friction and Anisotropy in Metalworking

Sammanfattning: The work material in metal working operations always showssome kind of anisotropy. To simplify the theoretical analysis,anisotropy is usually neglected and the material is assumed tobe isotropic, especially in determination of coefficients offriction. For predicting the material flow during plasticdeformation and for predicting the final material properties ofthe product, adequate descriptions of both coefficients offriction and flow stress curves have to be developed.In the present work, a new pattern describing the influenceof friction and anisotropy on material flow has been identifiedand verified by means of 3D-FEM analyses and experiments inring compression tests. Results reveal that the influence ofanisotropy on the ring deformation is quite similar to thatobtained by changing the frictional condition. If the influenceof anisotropy on material flow is incorrectly attributed tofriction, the possible error of the coefficient of friction canbe as high as 80% for a pronounced anisotropic material.Similar results have also been obtained from 2D-FEM analysesof upsetting, extrusion and rolling processes becauseanisotropy can have a significant effect on the shear strain inthe contact surface layer. In a rolling process, results showthat shear strain can be reduced by changing rolling parametersand the shear strain value can be an indicator to evaluate thedegree of optimization of the process.To determine the coefficient of friction and flow stress foranisotropic material, a modified two-specimen method (MTSM) hasbeen established using an inverse method. The method has beenapplied to cylinder upsetting, Rastegaev specimen upsetting,plane strain compression tests and rolling processes. Resultsshow that MTSM is able to evaluate the validity of mathematicalmodels when both the coefficient of friction and the flowstress are unknown and can also determine the coefficient offriction and flow stress provided that the selectedmathematical model is adequate. Regarding rolling, it is shownthat the Ekelund model combined with the Hitchcock equationmodified by Roberts gives the most accurate value of frictioncoefficient and flow stress in comparison with other forcemodels and equations.A number of experimental methods (yield loci, strain ratiosand multi-directional flow stress-curves) for determining theanisotropy have been compared. Results show that the yield locimeasurements are weak in predicting anisotropy when thematerial strain hardening differs in different directions. Itis also concluded that strain ratio (R-value) measurements areunreliable for describing anisotropy. The most reliable anduseful results were derived from multi-direction determinationsof the flow stresses.Key words:Coefficient of friction, anisotropy, flowstress, modified two-specimen method (MTSM) and FE-analysis