"Multi­objective Optimization of Monolith Catalytic Reactors for the Reduction of NOx Emission

MIGUEL C. MUSSATI; ZORAN GOMZI; ANA M. ARIAS; VESNA TOMASIC; NICOLÁS J. SCENNA; SERGIO F. MUSSATI

San Sebastián

Conferencia; 5th-international-conference-structured-catalysts-and-reactors; 2016

Many real-world chemical engineering problems require the simultaneous optimization of two or more competing objectives. In multi-objective optimization approach, there is no solution of the problem that improves all the objectives simultaneously, i.e. an objective cannot be improved withoutdeteriorating at least one of the rest objectives. The main goal in multi-objective optimization is to obtain a set of optimal solutions among the conflicting objectives, which helps better understand the problem structure and helps the decision-maker in choosing the best compromise solution for theconsidered problem. The set of high-quality designs obtained in multi-objective optimization are usually represented by the Pareto?s curve. The Pareto?s curve includes the best design for each objective (i.e. the extremes of the curve) as well as many compromise designs representing various trade-offs among objectives. Thus, a given point in the Pareto?s curve represents the optimal design for a different quantification of relative preferences among objectives [1]. Different techniques dealing with multi-objective optimization problems have been developed and successfully applied in different applications. A recent review on the applications of multiobjective optimization in chemical engineering, and specially on the multi-objective optimization of chemical reactors and processes, can be found in [2, 3]. This paper addresses the multi-objective optimization of monolith catalytic reactors for the reduction of NOx emission. A steady-state, isothermal, first principle-based reactor model is considered. It consists of a set of partial differential equations (PDEs) system in axial and radial domains, which are discretized using the method of line methods (centered finite difference method CFDM). The resulting non linearmathematical programming problem (NLP) is implemented in GAMS environment (General Algebraic modeling system) and solved using CONOPT 3.0. Then, the main goal is to determine the optimal reactor sizes and the operating conditions (e.g. stream flowrate and composition) considering the following two competitive objectives: maximal NOx conversion versus minimal total annual cost (investment and operating cost) of the DeNOx process.The technique called the ε-constraint method for multi-objective optimization is applied for obtaining the Pareto?s curve, which provides quantitative and qualitative results useful not only to identify all the trade-offs existing among the process variables, but also to assist in the decisionmakingprocess in optimizing monolith catalytic reactors for the reduction of NOx emission.