Combined Use of Simulation Packages, Multi-Objective Optimization and Statistical Tools for the Environmentally Conscious Design of Thermodynamic Cycles

Wednesday, October 19, 2011
Exhibit Hall B (Minneapolis Convention Center)
Robert Brunet1, Daniel Cortés1, Dieter Boer2, Gonzalo Guillén-Gosálbez1 and Laureano Jiménez1, (1)Departament d'Enginyeria Química, Universitat Rovira i Virgili, Tarragona, Spain, (2)Departament d'Enginyeria Mecánica, Universitat Rovira i Virgili, Tarragona, Spain

Energy consumption is one of the main factors affecting global warming. In the last years the installation of air conditioning systems has increased worldwide (Balaras et al. 2007, Henning 2007), leading to higher electricity consumption. Energy is usually produced in power plants, using mainly Rankine cycles, which consume fossil fuels. -On the other hand, conventional air conditioning systems are based on compression cycles requiring mechanical energy for its operation. Considering all these issues, there is a clear need to develop systematic tools to assist in their design according to environmental and economic concerns.

This work introduces a novel framework for the optimal design of sustainable thermodynamic cycles that integrates process simulation packages (i.e. Aspen Hysys and Aspen Plus), multi-objective optimization, life cycle assessment (LCA) and principal component analysis (PCA). The optimization problems are formulated using multi-objective mixed-integer linear programming (MINLP) techniques in which the LCA metrics are described through standard algebraic equations. A solution strategy is proposed to tackle these problems based on decomposing them into two levels: a nonlinear programming (NLP) subproblem, where the binary variables are fixed, and a master mixed-integer linear programming (MILP) problem that predicts new values for the binaries. The NLP subproblems are solved by combining an external NLP solver with process simulation packages (i.e., Aspen Plus), while the master problem is solved with standard MILP solvers (i.e., CPLEX). PCA is used in this context to identify redundant LCA metrics that can be omitted, thereby reducing the complexity of the model.

Two different thermodynamic cycles are used to highlight the capabilities of the approach presented: a 10 MW Rankine power cycle, and an ammonia-water absorption cooling cycle with 100 kW of cooling demand operating at cooling and refrigeration conditions. Our approach provides valuable insight into the design problem, and a set of alternatives for reducing the environmental impact in energy systems.

Extended Abstract: File Not Uploaded
See more of this Session: Poster Session
See more of this Group/Topical: Computing and Systems Technology Division