科研之友
微信
新浪微博
Facebook
分享链接
摘要
Solar energy is one of the best renewable energy and could be a solution to the increasing energy demand and environment alarms. But its intermittence and uneven distribution make it hard to use directly. One promising way to overcome these obstacles is to store and transport solar energy in the form of chemical fuels like hydrogen or syngas, and the process of splitting water or carbon dioxide using metal oxides as recycled reagents is catching more and more attention. In these processes, a metal oxide is thermal reduced in the first step, and then reacted with water or carbon dioxide to produce hydrogen and carbon monoxide, in which the metal is recycled to the first step. More than 400 redox pairs have been proposed for solar thermochemical cycles. It is a fundamental question that how to select a proper redox pair for further scientific research. The redox pairs can be divided into two categories: one has been fully investigated and has a lot of criteria for evaluation, the other contains new candidates whose evaluation criteria are fuzzy and incomplete. For the first one, the multidimensional space composed by evaluation criteria is projected to the one dimension space (as a scalar) in the principle of concentrating the points into clusters while separating the clusters at the same time. Genetic algorithm is used to seek the best projection index during the process. The database from Sandia National Laboratories is used as sources, which contains more than 350 thermochemical cycles and each one has 16 criteria. The best projection index is a*=(0.3288, 0.2859, 0.1676, 0.3779, 0.2114, 0.2530, 0.1732, 0.2346, 0.2090, 0.0937, 0.1507, 0.1674, 0.2902, 0.2225, 0.1835, 0.4184), and iron oxide, zinc oxide, zinc-manganese ferrite are selected to compared with the redox pairs in the other category. For the second category, analytic hierarchy process (AHP) integrated with specialist marking method is applied for screening. Due to the features of emerging candidates are fuzzy and incomplete, some universal criteria are selected and scored based on experts'opinion. Then hierarchy structure are built based on the criteria and candidate redox pairs. Pairwise comparison matrices for relative importance judgment are generated based on 1-9 scaling method. If they satisfy consistency check, the normalized eigenvectors of the matrix are the weight value of the corresponding candidate. Ceria and perovskites, two emerging types plus three candidates from the above, are scored for five criteria including reduction temperature, separation complexity, element abundance, energy efficiency and test scale. Through the AHP methods, we get the ranking and final scores of the five candidates: perovskites (0.2810), zinc oxide (0.2558), zinc-manganese ferrite (0.1488), iron oxide (0.0957), and ceria (0.0464). The number in the bracket is the corresponding score of the candidate, which reflects the ranking position. The screening method, projection pursuit integrated with AHP, helps the decision makers find candidates from multiple criteria without much subjective decisions, as well as find one best suit their goal from the fuzzy and incomplete issues. It gives theoretical reference for the further research of certain redox pairs.
- 出版日期2017
- 单位中国科学院大学
