The development of a new decision making model for material selection of polymer electrolyte fuel cell

Simulations and modeling polymer electrolyte fuel cell systems can be considered a powerful diagnostic tool in their system design for optimized component performance. Available mathematical models have been obtained for mass transfer, fuel storage, fuel reforming and processing, air delivery systems, as well as thermal and water management in very restricted and idealized situations, and do not consider material selection, engineering characteristics and design configurations of PEFC. In this thesis, a new approach has been carried out on the use of a Multiple Attribute Decision Making (MADM) model in material selection of fuel cell components. This approach provides solutions to material selection problems involving conflicting as well as multiple objectives.;For the second step, after introducing the theoretical background on Multiple Criteria Decision Making models, a case study is presented for the material selection of a bipolar plate in a PEFC. An analytical solution is considered for examining and evaluating the criteria and their related performance indices in the studied case. Then, the Compensatory TOPSIS MADM model is evaluated for the given material selection problem. This category of compensatory models selected the alternative with the highest score; the problem, then, consists of how to assess the appropriate multi-attributes utility function for the relevant decision situation. It is shown that such complex decision cases for material selection of fuel cell components, with multiple qualitative data, can be resolved by applying the same model in different stages. Finally, the trade-off between the cost factor and other predefined criteria in arriving at efficient, yet practical, candidate materials is discussed by comparing outcomes of different decision scenarios.;For the third step, a new non-compensatory approach is introduced for the material selection of bipolar plates in Polymer Electrolyte Fuel Cells (PEFC), using the original ELECTRE (Elimination Et (and) Choice Translating Reality) decision-making method. Inversely, with the compensatory TOPSIS models in which a single index is usually assigned to each multi-dimensional characterization representing an alternative, the original ELECTRE model gathers a set of preferences, ranking them according to how much each satisfies a given concordance. For the fourth step, the non-compensatory ELECTRE III decision model under pseudocriteria and fuzzy outranking relations is used for material selection of the bipolar plate in PEFC. Given a set of pre-defined attributes (material performance indices derived in the second step), ELECTRE III can arrange alternatives (bipolar plate materials) into equivalence classes that are completely or partially sorted. By introducing a decision matrix, the revised Simos method is used to define a set of weighting factors and to perform the ranking stability analysis. It is shown that the model may be useful not only to elicit the best performing materials, but also to recognize incomparable and/or indifferent alternatives. To get a more reliable solution, data uncertainties, often resulting from experimental tests, are incorporated into the model via definition of criteria thresholds.;For the fifth step, ELECTRE IV, using embedded outranking relations, has been applied to determine the best compromised possible candidate material for the bipolar plate. This step also investigates the effect of replacing components of the selection parameters (i.e. design parameters) with performance indices to solve the same problem. Through the introduction of different approaches to the solution algorithm, the effect of criterion of cost on material selection, using the ELECTRE IV, is studied. A simple multi-axial candidate material is also recommended from which safer engineering decisions may be attained.;For the first step, a comprehensive study was performed to provide the fundamental base on modeling of fuel cell components and stacks. Available mathematical and numerical models are classified according to the phenomena investigated in the polymer electrolyte fuel cell and are compared with each other in view of the assumptions and governing equations used.;Finally, for a given case study, advantages, disadvantages, similarities and differences observed between the results of the proposed methods are discussed.