Characterization of coated abrasive wear by statistical invariants.

Statistical invariants are used to characterize the surfaces of coated abrasive belt samples at three conditions of their wear; namely, for a new belt, when was used to its half life and when it was used to its full life. An auto - regressive model of order four, A(4), was fitted to the surface profile data collected in five different directions using the DDS methodology. The model parameters are used to compute profile and surface spectral moments which are then combined to yield surface invariants and characterization parameters. The results indicate that the change in surface invariants and characterization parameters clearly differentiate coated abrasive surfaces as they wear. (A)