VARIABLE TARGET MASS-EXCHANGE NETWORK SYNTHESIS THROUGH LINEAR-PROGRAMMING

A mass-exchange network (MEN) synthesis problem is considered with str eams whose target compositions are allowed to vary between upper and l ower bounds. The design task is to determine the minimum mass separati ng agent (utility) cost needed for the transfer of a single component from the rich to the lean streams. The mathematical formulation of thi s synthesis problem leads to a mixed integer nonlinear program. In thi s work, we propose a novel formulation of the problem that leads to a linear program. Stream decomposition is employed in attaining this nov el linear programming formulation and rigorous proofs are presented wh ich establish that the two formulations have the same solution. The li near programming formulation reduces the complexity of the considered MEN synthesis problem, thus making feasible its solution even for larg e-scale problems. Two examples, illustrating the procedure, are presen ted. Both demonstrate that significant utility cost savings can be ach ieved over the fixed composition MEN synthesis problem.