Model and algorithm based on accurate realization of dwell time in magnetorheological finishing.

Classically, a dwell-time map is created with a method such as deconvolution or numerical optimization, with the input being a surface error map and influence function. This dwell-time map is the numerical optimum for minimizing residual form error, but it takes no account of machine dynamics limitations. The map is then reinterpreted as machine speeds and accelerations or decelerations in a separate operation. In this paper we consider combining the two methods in a single optimization by the use of a constrained nonlinear optimization model, which regards both the two-norm of the surface residual error and the dwell-time gradient as an objective function. This enables machine dynamic limitations to be properly considered within the scope of the optimization, reducing both residual surface error and polishing times. Further simulations are introduced to demonstrate the feasibility of the model, and the velocity map is reinterpreted from the dwell time, meeting the requirement of velocity and the limitations of accelerations or decelerations. Indeed, the model and algorithm can also apply to other computer-controlled subaperture methods.