Methodology for a precise model-based characterization of liquid-phase adsorption processes

液相吸附是一种高效、高选择性的清洗和产品回收工艺，尤其适用于低热阻产品。为了设计和改进这些吸附过程，必须正确识别吸附平衡、基本吸附和解吸动力学以及吸附动力学。然而，当涉及到液相表征时，有几个因素是令人望而却步的，例如，未知或被忽视的过程和测量不确定性以及有限数量的吸附剂或吸附剂，特别是在研究和开发的早期阶段。此外，由于吸附物与吸附剂之间复杂的相互作用受各种因素的影响，对液相吸附的精确分析具有挑战性。为了获得可信的系统洞察，本研究项目旨在建立一种结合动态实验、创新系统识别概念和基于模型的实验设计的方法。具体而言，本项目侧重于考虑测量和过程不确定性的沸石BEA上单糖液相吸附的精确模型表征。为此，Scholl工作组在静态和动态操作条件下对吸附过程进行了实验研究，分析了工艺设计参数对测量不确定度的影响，并对实验装置的实验费用和材料需求进行了评价。与此同时，工作组Schenkendorf利用系统理论概念开发了用于模型和参数识别的代数识别程序，并将这些代数识别程序与基于模型的实验设计相结合。具有代表性的灵敏度分析确保了代数识别程序与常规参数识别方法的最佳组合。最后，对模型和参数的不确定性进行量化，并直接纳入稳健实验设计。因此，最优实验设计提供了有效和高效的实验，即以最小的材料投入提供高信息量的数据。总之，该项目有助于更好地理解吸附物、吸附剂和工艺设计参数之间的复杂相互作用，并对吸附动力学和吸附平衡进行可靠的不确定度量化，我们也将其应用于多组分吸附过程。根据这些结果，可以设计和运行更可靠的吸附工艺。

Liquid-phase adsorption is an efficient and highly-selective cleaning and product recovery process especially for products with a low thermal resistance. The adsorption equilibrium, the basic adsorption and desorption kinetics as well as the adsorption dynamics have to be identified properly to design and to improve these adsorption processes. However, several factors are prohibitive when it comes to liquid-phase characterization, e.g., unknown or neglected process and measurement uncertainties and a limited amount of adsorbent or adsorptive especially in the early stage of research and development. Moreover, a precise analysis of the liquid-phase adsorption is challenging because of the complex interactions between adsorbate and adsorbent which is affected by various factors. To gain credible system insights, this research project is aiming for a methodology which combines dynamic experiments, innovative system identification concepts, and model-based experimental design. In detail, this project focuses on a precise model-based characterization of the liquid-phase adsorption of monosaccharides on zeolite BEA considering measurement and process uncertainties. To this end, the working group Scholl studies the adsorption process experimentally under static and dynamic operating conditions, analyzes the impact of process design parameters on the measurement uncertainties, and evaluates the experimental setups regarding their experimental expense and material demand. In parallel, the working group Schenkendorf develops algebraic identification routines for model and parameter identification using systems theory concepts and combines these algebraic identification routines with model-based experimental design. A representative sensitivity analysis ensures an optimal combination of the algebraic identification routines with conventional methods for parameter identification. Finally, model and parameter uncertainties are quantified and directly incorporated in the robust experimental design. Thus, the optimal experimental design provides effective and efficient experiments, i.e., highly informative data with minimum material input. In summary, this project leads to a better understanding of the complex interaction of adsorbate, adsorbent, and process design parameters and in a credible uncertainty quantification of adsorption kinetics and adsorption equilibrium - which we also apply for multi-component adsorption processes. Based on these results, adsorption processes can be designed and operated more reliably.