Tool Wear Evaluation Based on Design Space Exploration Coupled with Explicit Cutting Simulations and Cutting Forces Excitation Signature

Abstract

Cutting operations are still one of the main methods used in the industry for surface generation in mass production. The dynamics of these processes are complex and having a good insight into the interdependencies of the nonlinear physical phenomena can be translated into better cutting performance. One of the main cutting tool geometrical parameter, usually associated with wear, is the cutting edge radius. Continuous direct evaluation of this parameter exhibits some important limitations offline, if this evaluation needs to be online the whole process becomes very complicated due to the measurement limitations that might appear. For online cases, the best approach is to determine the amplitude of this parameter indirectly using various side effects that can be correlated. One possible option is to monitor the vibrations generated by the resulting cutting forces. In the latest years, the usage of explicit finite element methods (FEM) to simulate the cutting processes has grown exponentially with the continuous increase of computation efficiency. With the help of Computer-Aided Engineering (CAE in short) solutions and using the latest advances in design space exploration (DSE) solutions, it is possible to create models able to parametrically explore a design space (DS), having precise targets, and also get the important correlations for all the important, quantifiable, cutting parameters. This paper presents an innovative method to create tool wear evaluation models focused mainly on the cutting edge radius indirect evaluation. The proposed method uses guided simulation loops able to generate a variety of dynamic signatures which are further post-processed to get a complex correlative model. The model can be applied in real cutting cases using the reciprocity property and can determine online the state of the cutting edge radius for further tool wear evaluation. The paper concludes with an analysis of the obtained model and the applicability of the data for the intended purpose.