Incremental Parameter Estimation under Rank-Deficient Measurement Conditions

Abstract

The computation and modeling of extents has been proposed to handle the complexity of large-scale model identification tasks. Unfortunately, the existing extent-based framework only applies when certain conditions apply. Most typically, it is required that a unique value for each extent can be computed. This severely limits the applicability of this approach. In this work, we propose a novel procedure for parameter estimation inspired by the existing extent-based framework. A key difference with prior work is that the proposed procedure combines structural observability labeling, matrix factorization, and graph-based system partitioning to split the original model parameter estimation problem into parameter estimation problems with the least number of parameters. The value of the proposed method is demonstrated with an extensive simulation study and a study based on a historical data set collected to characterize the isomerization of α -pinene. Most importantly, the obtained results indicate that an important barrier to the application of extent-based frameworks for process modeling and monitoring tasks has been lifted.

Active learning of chemical reaction networksviaprobabilistic graphical models and Boolean reaction circuits

“…[26][27][28][29] This approach effectively attributes measured rate data to individual reactions, enabling their independent determination; however, the authors acknowledge their incremental approach is vulnerable to inaccurate RNI. [30][31][32][33][34] DOE specifically for RNI is needed.…”

Special Issue on “Process Modelling and Simulation”

“…Also related to this goal, two contributions brought heat and power systems into the scope: [7] proposed a grey-box model of limited complexity that couples the production process with the plant's combined heat and power system in order to reduce operation costs, whereas [8] modeled the hydraulic dynamics in a nuclear reactor cooling pump with respect to different vane structures to ensure safe operation in case of power failures.Models for decision support must be tailored to the actual process, or the underlying equations should allow the transfer of the lab-scale data to any desired scale. In this sense, [3,4] proposed iterative methods for parameter estimation to progressively improve the plant-model match under realistic conditions, and [5] considered uncertainty in the estimation via robust optimization. Furthermore, a methodology for obtaining physically coherent grey-box models (or plant surrogate ones) from fundamental principles and plant data was proposed in [10], while [1] presented a quantitative validation method based on partial least squares to devise the suitable modelling depth according to the quality of the available experimental data.Once reliable prediction models are available, they can be used in numerical simulations to analyze the main features of the process or to evaluate the influence of the operating conditions as well as of the external disturbances.…”

“…Models for decision support must be tailored to the actual process, or the underlying equations should allow the transfer of the lab-scale data to any desired scale. In this sense, [3,4] proposed iterative methods for parameter estimation to progressively improve the plant-model match under realistic conditions, and [5] considered uncertainty in the estimation via robust optimization. Furthermore, a methodology for obtaining physically coherent grey-box models (or plant surrogate ones) from fundamental principles and plant data was proposed in [10], while [1] presented a quantitative validation method based on partial least squares to devise the suitable modelling depth according to the quality of the available experimental data.…”