Determination of reactivity ratios for acrylic acid and its dimer from classical parameter estimation and Bayesian approach. (19th December 2021)

• Record Type: Journal Article
• Authors: Hungenberg, Klaus‐Dieter; Schwede, Christian; Wulkow, Michael; Wulkow, Niklas
• Is Part Of: Canadian journal of chemical engineering. Volume 100:Number 4(2022)
• Page Start: 780
• Page End: 789
• Publication Date: 2021-12-19
• Subjects: Bayesian analysis -- parameter estimation -- probability distribution -- reactivity ratios; Chemical engineering -- Periodicals; Technology -- Periodicals; 660.05
• DOI: 10.1002/cjce.24330
• Languages: English
• ISSNs: 0008-4034

Abstract

Abstract: The reactivity ratios of acrylic acid (AA, M1 ) and its dimer β‐acroyloxypropionic acid (diAA, M2 ) are determined from cumulative copolymerization data by two different methods: classical parameter estimation (PE) by minimizing the objective function and a Bayesian analysis. Classical PE gives r 1 = 0.74 and r 2 = 1.23 at the minimum of the residual. From the Bayesian analysis, the probability distribution of the parameter sets is obtained, revealing the existence of parameter sets with rather the same probability. The influence of the number of data and the size of the measurement error are discussed. Abstract : Probability density over parameter space (in the back), parameter‐wise densities (red and blue), and ensuing estimate of potential forward simulations (right).