Optimal design of eco-efficient chemical processes

Technico-economic considerations have for long been taken into account as decisional criteria in the conceptual phase for new and retrofitted chemical processes. While the emphasis on economic aspect remains strong, another priority in evaluating chemical processes is the environment. Such problems, leading to multiple and most often conflicting goals, must be solved within the framework of complex multiobjective optimization. This study addresses the problem of analyzing the various objectives involved in eco-efficient processes, meaning that ecological and economic considerations are taken into account simultaneously at the preliminary design phase of chemical processes. The multiobjective methodology is performed by genetic algorithms implemented in the so-called MULTIGEN library, particularly well-suited to multiobjective optimization of mixed integer nonlinear programming problems. The trade-off between economic and environmental objectives is illustrated through the generation of Pareto curves. The methodology will be illustrated by the classical example of Williams and Otto Chemical Plant, which is often considered as a test bench for representing complex nonlinear programming problems incorporating the main features of a chemical processing plant, in the dedicated literature of process design. The original William and Otto Chemical Plant problem will be revisited here in a multiobjective mode. A key point will be the treatment of equality constraints involved in the material balances, which are often considered as one of the most critical phases in genetic algorithm implementation. This step was carried out by solving the set of nonlinear equations by a classical Newton-Rapshon method implemented within the Matlab solver. We will highlight the insight the design engineer can gain using the multiobjective synthesis procedure and demonstrate the computational efficiency achieved by tackling simultaneously environmental and economic issues.