Nonlinear model predictive control using automatic differentiation

Although nonlinear model predictive control (NMPC) might be the best choice for a nonlinear plant, it is still not widely used. This is mainly due to the computational burden associated with solving a set of nonlinear differential equations and a nonlinear dynamic optimization problem. In this work, a new NMPC algorithm based on nonlinear least square optimization is proposed. In the new algorithm, the residual Jacobian matrix is efficiently calculated from the model sensitivity functions without extra integrations. Recently developed automatic differentiation techniques are applied to get the sensitivity functions accurately and efficiently. The new algorithm has been applied to an evaporation process with satisfactory results to cope with large setpoint changes, measured and unmeasured severe disturbances and process-model mismatches.