Optimal Operation of Heat Exchanger Networks with Changing Active Constraint Regions

Abstract

Abstract In this paper, we study the optimal operation of heat exchanger networks with stream splits. In particular, we extend previous approaches on the unconstrained optimization of the system to the constrained case, with temperature constraints on each flow branch, and with changing disturbances so that the set of optimally active constraints changes during operation. The simplest way to achieve optimal operation when some of the constraints are active, is to control the constraints to their limiting value, known as active constraint control. For the remaining unconstrained degrees of freedom, we propose to control linear combinations of the gradient as self-optimizing controlled variables. To automatically switch between the different active constraint regions, we use classical advanced control elements such as selectors, thereby achieving optimal operation using only the temperature measurements as feedback in different active constraint regions. The performance of the proposed feedback optimizing control structure for the heat exchange network is compared with the traditional model-based real-time optimization using simulations. In the presence of structural plant-model mistmach, we show that our proposed approach performs optimally for all disturbances, while traditional real-time optimization fails to converge for some cases, as the optimization problem becomes infeasible depending on the estimated disturbances.