Sustainable Development Based Methodology for the Optimized Integrated Design and Operation of Energy Conversion Processes and Networks

A design methodology allowing engineers to consider simultaneously and from the earliest phases of design some of the key factors of sustainable development (pollution, life cycle data) has been developed and successfully demonstrated on two examples of integrated energy systems. The principal idea is to take diverse factors related to thermodynamics, economics and the environment and integrate them in a way that the most optimal syntheses and designs (and operation) can be found. The solution space resulting from such a complete formulation, which can include tens of independant variables, is highly non linear and includes locally continuous but generally non contiguous subspaces. Solution procedures based on specially adapted genetic algorithms, have been implemented and shown to allow the user to tracks the most effective solution paths in a robust way. Pollution penalty factors accounting for both emissions and immission levels have been proposed and applied in dedicated sensitivity analyses. Using a simplified quasi-stationary formulation application examples include the predesign of combined cycle power plant with or without cogeneration and with or without CO2 separation as well as the predesign of district heating systems considering the use of both centralized and/or decentralized heat pumps as well as cogeneration units. Both exergetic and economic optimum solutions are compared. As FN did not finance the entire proposal, additional contributions from FOGA and Alliance for Global Sustainability allowed the completion of this project. Although the present results open exciting prospects and are already being extended in the framework of four different international initiatives, further work is still required. It includes, among others, a further refinement of the link between entropy and/or exergy and pollution, an extension to a broader range of pollutants including those having synergistic effects, a better account for the time dependancy and uncertainties, the coupling with top down approaches including neural net based simplified submodels and parallelisation and an extension of the models of technologies considered to other integrated systems such as direct electrochemical conversion or renewables based devices (new proposal in preparation)