(25c) Enhancing Chemical Process Efficiency Using Hybrid Information Quality

The onset of Industry 4.0 has resulted in an abundance of data from chemical processes,
opening new possibilities for better comprehension, monitoring, control, and optimization of
product quality and operational efficiency. Nonetheless, a review of the literature by Sansana
et al. (2021) highlights that while data-driven models frequently augment first-principles
models, the reverse—enhancing data-driven models through mechanistic elements—has not
been extensively pursued. This indicates that valuable data insights may be underutilized when
first-principles knowledge could complement them.

In this study, we aim to create a systematic framework designed to assist practitioners in
constructing hybrid solutions flexibly. By implementing a structured problem-solving method
based on the well-established concept of Information Quality (InfoQ) (Kenett & Shmueli,
2014; Reis & Kenett, 2018), and utilizing insights gained from applying industrial hybrid
modeling techniques to non-stationary systems, we introduce a new approach termed Hybrid
Information Quality (H-InfoQ).

The H-InfoQ implementation requires a thorough assessment of four key dimensions: the
chosen hybrid analysis method, fH, the available process information, XH, the defined analysis
goal, g, and the appropriate utility measure, U. Despite its comprehensive nature, this approach
raises concerns about reproducibility and operationalization, which poses challenges for
industrial practitioners attempting autonomous implementation. The final goal is to maximize
the utility of applying fH to XH, in the scope of the goal g: Max H-InfoQ = U {fH (XH)|g}. To
address these issues, a finer eight-dimensional assessment process is proposed. The eight
dimensions used to assess H-InfoQ are: data granularity, structure, integration, temporal
relevance, chronology of data and goals, generalizability, operationalization, and
communication. Instead of directly assessing the quality of the four components, H-InfoQ can
be indirectly analyzed using these dimensions. The goal is to construct an effective hybrid
model, defined as a mathematical abstraction of a system that addresses specific questions,
thereby maximizing utility for the user. The H-InfoQ framework is designed to guide
practitioners through a systematic thought process, enabling them to approach problems in a
structured manner. By adhering to this framework, users can identify and implement the most
suitable and effective technology available, aligned with the specific goals of the problem at
hand.

To illustrate the application and efficacy of the H-InfoQ framework, two industrial case studies
are assessed using this methodology. These instances were selected to offer a practical
showcase of how the framework can be employed in actual situations.

References

Sansana J, Joswiak MN, Castillo I, Wang Z, Rendall R, Chiang LH, Reis MS. Recent trends
on hybrid modeling for Industry 4.0. Computers and Chemical Engineering. 2021;151.

Kenett RS, Shmueli G. On Information Quality. Journal of the Royal Statistical Society Series
A: Statistics in Society. 2014;177

Reis MS, Kenett RS. Assessing the value of information of data-centric activities in the
chemical processing industry 4.0. AIChE Journal. 2018;64.