论文部分内容阅读
含能材料合成过程复杂,如果能从混合物光谱数据(中红外光谱)中分离出参与反应的中间体与过渡态的纯光谱信号和相对浓度信息,并使其尽可能独立,就能够准确地推断出未知物,从而得到含能材料合成反应机理。本文提出了一种基于主成份分析的快速定点算法与特征矩阵联合对角化算法混合的新算法,用于中红外光谱数据的分离研究。实验证明该混合算法具有可行性与有效性,同时相比经典算法可获得更多的分离物质。
The synthesis of energetic materials is complex and can be accurately inferred if the pure spectral signals and relative concentration information of intermediates and transition states involved in the reaction are separated from the mixture spectral data (mid-infrared spectrum) as much as possible Unknowns, resulting in a synthetic reaction mechanism of energetic materials. In this paper, we propose a new hybrid algorithm based on principal component analysis for fast fixed point algorithm and eigen-matrix combined diagonalization algorithm, which is used for the separation of mid-infrared spectral data. Experiments show that the hybrid algorithm is feasible and effective, and more isolated materials can be obtained than classical algorithms.