The forefront of chemical engineering research

Nature Chemical Engineering, Год журнала: 2024, Номер 1(1), С. 18 - 27

Опубликована: Янв. 11, 2024

Язык: Английский

---

Supercritical water gasification thermodynamic study and hybrid modeling of machine learning with the ideal gas model: Application to gasification of microalgae biomass

Energy, Год журнала: 2024, Номер 291, С. 130287 - 130287

Опубликована: Янв. 21, 2024

Язык: Английский

---

The enabling technologies for digitalization in the chemical process industry

Digital Chemical Engineering, Год журнала: 2024, Номер 12, С. 100161 - 100161

Опубликована: Июнь 5, 2024

In this paper, we provide an overview of the technologies that enable digitalization in chemical process industry and describe their applications to solve problems industrial settings. This is done through identification categorization these technologies, thereby providing structure otherwise loosely connected basket casting a spotlight on state-of-the-art offer great potential but are still underutilized applications. Furthermore, identify problem domains characterize connect them development aspects lend themselves digital solutions. For each connections, select most essential bridging gap between solution. allows practitioners better understand relevancy provides starting point for further investigation The connections substantiated by reference successful applications, highlighting previous works have been published field. They verified experts brainstorm sessions, interviews, workshop.

Язык: Английский

---

Efficient data-driven predictive control of nonlinear systems: A review and perspectives

Digital Chemical Engineering, Год журнала: 2025, Номер unknown, С. 100219 - 100219

Опубликована: Янв. 1, 2025

Язык: Английский

---

Machine learning in modeling, analysis and control of electrochemical reactors: A tutorial review

Digital Chemical Engineering, Год журнала: 2025, Номер unknown, С. 100237 - 100237

Опубликована: Апрель 1, 2025

Язык: Английский

---

Electrochemical and Chemometric Authentication of Panax notoginseng from Different Geographical Origins Using Graphene-Modified Screen-Printed Electrodes

International Journal of Electrochemical Science, Год журнала: 2025, Номер unknown, С. 101050 - 101050

Опубликована: Апрель 1, 2025

Язык: Английский

---

Feedback control of an experimental electrically-heated steam methane reformer

Process Safety and Environmental Protection, Год журнала: 2024, Номер 206, С. 469 - 488

Опубликована: Май 18, 2024

Язык: Английский

---

A tutorial review of machine learning-based model predictive control methods

Reviews in Chemical Engineering, Год журнала: 2024, Номер unknown

Опубликована: Дек. 10, 2024

Abstract This tutorial review provides a comprehensive overview of machine learning (ML)-based model predictive control (MPC) methods, covering both theoretical and practical aspects. It analysis closed-loop stability based on the generalization error ML models addresses challenges such as data scarcity, quality, curse dimensionality, uncertainty, computational efficiency, safety from modeling perspectives. The application these methods is demonstrated using nonlinear chemical process example, with open-source code available GitHub. paper concludes discussion future research directions in ML-based MPC.

Язык: Английский

---

Model predictive control of an electrically-heated steam methane reformer

Digital Chemical Engineering, Год журнала: 2023, Номер 10, С. 100138 - 100138

Опубликована: Дек. 27, 2023

Steam methane reforming (SMR) is one of the most widely used hydrogen (H2) production processes. In addition to its extensive utilization in industrial sectors, expanding it share as a clean energy carrier, and more sustainable efficient H2 methods are continuously being explored developed. One method replaces conventional fossil fuel-based heating with electrical through flow electrons across reformer. At UCLA, an experimental setup was built electrically heated steam process. This paper describes system components, explains digitalization introduces for building first-principles-based dynamic process model using parameters estimated via data-driven from data. The modeling approach uses lumped parameter approximation employs algebraic equations solve gas-phase variables. reaction calculated steady-state data, temperature change modeled respect electric current first-order model. overall then computational predictive control (MPC) scheme drive new set-point under unperturbed flowrate disturbance cases. performance robustness proposed MPC compared ones classical proportional-integral (PI) controller demonstrated be superior terms closed-loop response, robustness, constraint handling.

Язык: Английский

---

Improved fault detection and diagnosis using graph auto encoder and attention-based graph convolution networks

Digital Chemical Engineering, Год журнала: 2024, Номер 11, С. 100158 - 100158

Опубликована: Май 24, 2024

A powerful fault detection and diagnosis (FDD) system plays a pivotal role in achieving operational excellence by maximizing performance, optimizing maintenance strategies, ensuring the longevity resilience of process plants. In context FDD for multivariate sensor data, this study presents an improved approach using graph-based neural networks. This graph network uses adjacency matrix developed extracting expert domain knowledge topological information multi-sensor system. additional representation is incorporated along with data to capture spatial temporal networks efficiently. regard, we propose evaluate: 1) Graph Auto Encoder (GAE) based strategy 2) An Attention-based Spatial Temporal Convolution Network (ASTGCN) methodology. By leveraging form graphs, GAE captures complex relationships dependencies among sensors, enabling effective anomaly detection, which identifies abnormal patterns deviations from normal behavior, thus indicating potential faults The ASTGCN incorporates attention mechanisms selectively focus on relevant nodes their diagnosis. effectiveness proposed demonstrated benchmark Tennessee Eastman Process (TEP) problem. results show that approaches outperform traditional methods highlight importance systems.

Язык: Английский

---