Automatic optimal control of Field Assisted Sintering Technology

In this paper, we introduce a novel model-free optimal control algorithm that finds the near optimal energy saving control trajectory from the observations of the process data. Conventional optimal control algorithms have not been able to find the optimal solution without the dynamic models of the system. Therefore a computationally expensive system identification process has been indispensable to control a black box dynamical system. The algorithm we provide in this paper eliminates the need for the system identification and quickly finds a near optimal solution by calculating the minimum dynamic information from the observed data. Since the algorithm runs with only time series type data, it is indeed model-free and especially versatile for a black-box dynamical system. We show the framework of this new optimizing method for a recent hot manufacturing process, Field Assisted Sintering Technology (FAST), whose dynamic mechanism is not revealed yet.