Analysis of Samples for the ICTAC Lifetime-Prediction Round-Robin Exercise

Burnham, A K

Lawrence Livermore National Laboratory

关键词: Calorimetry;    Forecasting;    Probability;    37 Inorganic, Organic, Physical And Analytical Chemistry;    Kinetics;   

DOI  :  10.2172/899384
RP-ID  :  UCRL-TR-221563
RP-ID  :  W-7405-ENG-48
RP-ID  :  899384

美国|英语

来源: UNT Digital Library

【 摘 要 】

Derivation of chemical kinetic models for prediction of material and component lifetimes is of broad interest and value. This work analyzes data that was distributed to me, among others, by the International Confederation for Thermal Analysis and Calorimetry (ICTAC) as part of a blind study of kinetic analysis. The results from this report will be combined with results from other parties to create a broader comparison of kinetic analysis methods. In addition to the eight ICTAC data sets, which appear to contain one set of simulated data, presumably for ground truth comparison, I created an additional simulated data set to compare the reliability of isoconversional and model-fitting approaches. It is usually possible to fit the data well with both isoconversional and model fitting approaches, although the isoconversional method is usually faster and provides better fits to the data, particularly for complex reaction profiles. The two methods often, but not always, give similar predictions. Predictions of the isoconversional model will fail to the extent that the reaction contains competitive or crossing-concurrent reaction characteristics. Model fitting will either do better or worse depending on how well the derived model includes the appropriate characteristics, and the probability of deriving a good model depends both on the sophistication of the modeling software and the skill of the analyst.