Estimation of kinetic coefficients in micro-reactors for biodiesel synthesis: Bayesian inference with reduced mass transfer model
Estimation of kinetic coefficients in micro-reactors for biodiesel synthesis: Bayesian inference with reduced mass transfer model
复制标题
DOI：
10.1016/j.cherd.2018.11.023
复制
发表时间：
2019-01
影响因子：
3.9
通讯作者：
J. M. Costa;C. Naveira-Cotta
中科院分区：
文献类型：
--
作者：
J. M. Costa;C. Naveira-Cotta
通过微反应器技术，可以在非常短的停留时间内获得生物柴油生产的高产率。微反应器中生物柴油合成反应机理的理论研究是实现生物柴油产量最大化的适当设计条件的关键。这种物理化学现象涉及复杂的液-液流动和传质过程，除了酯交换反应，和动力学系数的确定是必不可少的，在使用的数学模型的微反应器的设计。在这项工作中，一阶常微分方程的非线性耦合的数学模型已经开发，使用耦合积分方程方法（CIEA）的模型降阶，从一个扩散-对流-反应的三维数学模型，描述了在生物柴油合成中所涉及的局部物种浓度。使用Mathematica平台的NDSolve例程对ODE系统进行数值求解。马尔可夫链蒙特卡罗方法（MCMC）是在解决反问题，估计的动力学系数，使用合成的实验数据与低转化率，最大限度地存在的中间物种，并增加所需的参数的问题的敏感性。结果表明，CIEA与MCMC统计推断相结合，产生一个有效的和强大的组合，这样的反应传质问题的正-逆分析。
High yields in the production of biodiesel at very short residence times can be obtained by means of micro-reactors technology. The theoretical study of the reaction mechanisms involved in biodiesel synthesis in micro-reactors is crucial in achieving adequate design conditions towards maximizing biodiesel production. Such a physico-chemical phenomenon involves complex liquid–liquid flow and mass transfer processes, besides the transesterification reaction, and the determination of the kinetic coefficients is essential in using the mathematical models for the design of the micro-reactors. In this work, a nonlinear coupled mathematical model of first-order ordinary differential equations has been developed, using the Coupled Integral Equations Approach (CIEA) for model reduction, starting from a diffusive-convective-reactive three-dimensional mathematical model that describes the local species concentrations involved in the biodiesel synthesis. The ODEs system is numerically solved using the NDSolve routine of the Mathematica platform. The Markov Chain Monte Carlo method (MCMC) is employed in solving the inverse problem to estimate the kinetic coefficients, using synthetic experimental data with low conversion rates, which maximize the presence of intermediate species and increase the sensitivity of the problem to the desired parameters. The results presented indicate that the CIEA in combination with the MCMC statistical inference yield an efficient and robust combination for the direct - inverse analysis of such reactive mass transfer problems.