Data-Driven Nonlinear Robust Optimization for Gold Cyanidation Leaching Process

In order to operate the gold cyanidation leaching process under certainty, a data-driven nonlinear robust optimization method is proposed. By using the T
²
statistic limit based on PCA model, the correlations between the uncertain variables can be efficiently captured and a tighter uncertainty set can be defined. The first-order Taylor approximation is employed to linearize the inequality constraints. The equality constraints with state variables are also considered by using implicit function theorem, and the robust counterpart of the nonlinear robust optimization problem is formulated. The efficiency of the proposed data-driven nonlinear robust optimization method is verified through a simulated industrial GCLP. Compared with the nonlinear robust optimization methods with box and ellipsoid uncertainty set, the proposed method can effectively reduce the economic cost.