Modeling and design of a combined electrified steam methane reforming-pressure swing adsorption process

Process Safety and Environmental Protection, Journal Year: 2024, Volume and Issue: 209, P. 111 - 131

Published: July 29, 2024

Language: Английский

Machine learning-based predictive control of an electrically-heated steam methane reforming process

Digital Chemical Engineering, Journal Year: 2024, Volume and Issue: 12, P. 100173 - 100173

Published: July 23, 2024

Hydrogen plays a crucial role in improving sustainability and offering clean efficient energy carrier that significantly reduces greenhouse gas emissions. However, the primary method of industrial hydrogen production, steam methane reforming (SMR), relies on combustion hydrocarbons as heating source for reactions, resulting significant carbon To address this issue, an experimental setup electrically-heated reformer (e-SMR) has been constructed at UCLA, lumped first-principle dynamic process model was built based parameters estimated from data previous study. Subsequently, implemented into computational predictive control (MPC) scheme, successfully driving production rate to desired setpoint. While these works are important pave way developing MPC large-scale e-SMR processes, may not accurately reflect actual behavior, particularly behavior changes with time. Therefore, development establishment adaptive data-driven approach implementing is necessary. need, present work investigates construction recurrent neural network (RNN) models in-depth, utilizing experimentally-validated model. Specifically, long short-term memory (LSTM) layer utilized RNN effectively capture complex correlations long-term sequential data. LSTM-based employed design MPC, its performance evaluated through comparison proportional–integral (PI) control. potential disturbances variability typical process, three distinct approaches were developed: integrator, real-time online retraining (transfer learning), offset-free MPC. These eliminated offset caused by disturbances. Overall, study underscores effectiveness dynamics process. It also outlines strategies employing RNN-based multiple general processes partially infrequent delayed measurement feedback. This valuable scenarios where new be challenging.

Language: Английский

Modeling and control of a protonic membrane steam methane reformer

Process Safety and Environmental Protection, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 1, 2024

Language: Английский