摘要
本文发展了一种将相图和其它相关的热力学性质放在一起关联的综合方法并实现了计算机程序化。该方法基于广义最小二乘原理,可以同时考虑多种不同来源、不同性质的原始数据,于同一拟合过程中得到所有的未知参数。结果,能够更加充分合理地利用原始数据,提取更多的信息,并且保证拟合结果在体系内彼此相容。对于线性拟合,文中给出一种由焓项、熵项和其他项三个加权部分构成的通用拟合关系式,以用于常见各类热力学性质和相平衡数据的变换。文中还给出回归Marquart二步参数优化法的结构框图和若干应用实例。
An extended least square scheme has been developed for fitting of phase diagrams and the correlated thermodynamic properties of a given system. A general fitting relationship, consisting of three weighted parts: enthalpy terms, entropy ones and the other items of thermodynamic quantities to be considered, is proposed for the linear regression approach. By using the scheme, all experimental data concerned with the thermodynamic uantities of interest, which may be available from various origins and thus of different accuracy, can be considered and handled synthetically and simultaneously. As a results, a complete series of empirical formulas for all corresponding quantities can be given through only a single regression process and a thermodynamic consistency can be well kept in the whole. The scheme has implemented in a Vax 11/780 computer. Two successful examples for LiF-BaF_2 and CsCl-BaCl_2, as well as the computer flow chart are presented.
出处
《计算机与应用化学》
CAS
CSCD
1992年第1期18-22,共5页
Computers and Applied Chemistry
关键词
相图
热力学
最小二乘原理
Phase diagram Thermodynamics Lease square Parameter estlmation
