Zhihong Yuan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Zhihong Yuan has a total of 64 co-authors. Here are the top co-authors by number of shared publications:
Haoyang Hu
State Key Laboratory of Chemical Engineering and Low-carbon Technology, Department of Chemical Engineering
8 shared publicationsLifeng Zhang
Department of Chemical Engineering, The Sargent Centre for Process Systems Engineering, Imperial College London London UK
8 shared publicationsYue Li
Department of Chemical Engineering Beijing University of Technology Beijing China
6 shared publicationsBingzhen Chen
The State Key Laboratory of Chemical Engineering, Department of Chemical Engineering, Tsinghua University, Beijing, China
5 shared publicationsCongqin Ge
State Key Laboratory of Chemical Engineering and Low‐Carbon Technology, Department of Chemical Engineering Tsinghua University Beijing China
5 shared publications- Haoyang Hu
The State Key Laboratory of Chemical Engineering and Low-carbon Technology, Department of Chemical Engineering, Tsinghua University, Beijing, China
4 shared publications
Latest Publications
- Open Access
- 24 October 2026
- Published by Elsevier in Chemical Engineering Science
- Vol. 339
A unified optimization framework for integrated spatial–temporal berth allocation and refinery scheduling
- 24 October 2026
- Published by Elsevier in Chemical Engineering Science
- Vol. 339
Open AccessA unified optimization framework for integrated spatial–temporal berth allocation and refinery scheduling
- 5 May 2026
- Published by Wiley in AIChE Journal
- Vol. 72 (8) , e70436
Two‐stage distributionally robust optimization for the design of hybrid energy‐driven multi‐network processes
The present work focuses on optimizing integrated multi‐network systems that combine hybrid energy supplies, industrial processes, water resources, and carbon reductions. Given the inherent variability of renewable resources and volatility of market pricing and demand, a two‐stage distributionally robust optimization (TSDRO) framework with Wasserstein distance‐based ambiguity sets is proposed. By explicitly considering uncertainties in electricity demand and wind speed, the TSDRO framework simultaneously optimizes long‐term capacity investments and short‐term operational recourse. Case studies reveal that the framework identifies a robust system that prioritizes wind power as the dominant source, leverages photovoltaic and biomass for flexible regulation, and achieves a phased elimination of coal‐fired heating through substitution with biomass technologies. The optimal design not only reduces carbon emissions but also ensures a reliable energy supply while minimizing total cost. These outcomes underscore the distinct advantages of the proposed TSDRO method in balancing economic competitiveness, environmental performance, and operational resilience under uncertainty.
- 5 May 2026
- Published by Wiley in AIChE Journal
- Vol. 72 (8) , e70436
Two‐stage distributionally robust optimization for the design of hybrid energy‐driven multi‐network processes
The present work focuses on optimizing integrated multi‐network systems that combine hybrid energy supplies, industrial processes, water resources, and carbon reductions. Given the inherent variability of renewable resources and volatility of market pricing and demand, a two‐stage distributionally robust optimization (TSDRO) framework with Wasserstein distance‐based ambiguity sets is proposed. By explicitly considering uncertainties in electricity demand and wind speed, the TSDRO framework simultaneously optimizes long‐term capacity investments and short‐term operational recourse. Case studies reveal that the framework identifies a robust system that prioritizes wind power as the dominant source, leverages photovoltaic and biomass for flexible regulation, and achieves a phased elimination of coal‐fired heating through substitution with biomass technologies. The optimal design not only reduces carbon emissions but also ensures a reliable energy supply while minimizing total cost. These outcomes underscore the distinct advantages of the proposed TSDRO method in balancing economic competitiveness, environmental performance, and operational resilience under uncertainty.
- Open Access
- 7 March 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Accelerated discovery of highly active enzyme nanohybrids with parallelized Bayesian optimization in hybrid space
Artificial intelligence (AI) has significantly advanced protein engineering, enabling rapid enzyme evolution for diverse applications. However, the fragile nature of biomacromolecules requires enzyme immobilization to preserve catalytic activity under harsh industrial conditions, which often restricts substrate diffusion and reduces enzymatic activity. This challenge demands extensive trial-and-error experiments to optimize immobilized carriers with high activity for different enzymes. Here we show a machine-learning-guided workflow along with an algorithm named parallelized hybrid-space Bayesian optimization (PHBO) to accelerate the discovery of nanocarriers for specific enzymes and reactions. Leveraging prior knowledge, machine learning and iterative feedback, within limited number of experiments, this workflow explores the reaction space of over 10 7 experiments and achieves activity recovery of 100%, 90%, and 79% for glucose oxidase, catalase, and Candida Antarctica lipase B, respectively. These results demonstrate that data-efficient optimization can substantially accelerate the discovery of enzyme nanohybrids with high catalytic activity across diverse enzymatic systems.
- 7 March 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Open AccessAccelerated discovery of highly active enzyme nanohybrids with parallelized Bayesian optimization in hybrid space
Artificial intelligence (AI) has significantly advanced protein engineering, enabling rapid enzyme evolution for diverse applications. However, the fragile nature of biomacromolecules requires enzyme immobilization to preserve catalytic activity under harsh industrial conditions, which often restricts substrate diffusion and reduces enzymatic activity. This challenge demands extensive trial-and-error experiments to optimize immobilized carriers with high activity for different enzymes. Here we show a machine-learning-guided workflow along with an algorithm named parallelized hybrid-space Bayesian optimization (PHBO) to accelerate the discovery of nanocarriers for specific enzymes and reactions. Leveraging prior knowledge, machine learning and iterative feedback, within limited number of experiments, this workflow explores the reaction space of over 10 7 experiments and achieves activity recovery of 100%, 90%, and 79% for glucose oxidase, catalase, and Candida Antarctica lipase B, respectively. These results demonstrate that data-efficient optimization can substantially accelerate the discovery of enzyme nanohybrids with high catalytic activity across diverse enzymatic systems.
- Open Access
- 5 March 2026
- Published by Elsevier in Chemical Engineering Science
- Vol. 328
A unified continuous-time model for integrated crude scheduling and refinery planning optimization
- 5 March 2026
- Published by Elsevier in Chemical Engineering Science
- Vol. 328
Open AccessA unified continuous-time model for integrated crude scheduling and refinery planning optimization
- Open Access
- 5 August 2025
- Published by Wiley in AIChE Journal
- Vol. 71 (11)
Robust design of flexible supply chains with mobile and modular facilities
Modular manufacturing is increasingly being adopted by the chemical industry in applications where resource availability and product demand are highly distributed and uncertain. Modular plants are often built such that they can be easily relocated. Recent works have investigated how such mobility can increase flexibility at the supply chain level; however, we are lacking rigorous yet intuitive metrics for quantifying that flexibility. In this work, we develop a framework for the design of supply chains with mobile and modular facilities, where the uncertainty is described by an uncertainty set whose size can be adjusted using two flexibility indices that capture the spatial dissimilarity and temporal shift in demand. The resulting two‐stage robust optimization problem with mixed‐integer recourse is solved using an enhanced nested column‐and‐constraint generation algorithm. Through computational experiments, we examine the impact of the required level of flexibility on the supply chain design and the value of module mobility.
- 5 August 2025
- Published by Wiley in AIChE Journal
- Vol. 71 (11)
Open AccessRobust design of flexible supply chains with mobile and modular facilities
Modular manufacturing is increasingly being adopted by the chemical industry in applications where resource availability and product demand are highly distributed and uncertain. Modular plants are often built such that they can be easily relocated. Recent works have investigated how such mobility can increase flexibility at the supply chain level; however, we are lacking rigorous yet intuitive metrics for quantifying that flexibility. In this work, we develop a framework for the design of supply chains with mobile and modular facilities, where the uncertainty is described by an uncertainty set whose size can be adjusted using two flexibility indices that capture the spatial dissimilarity and temporal shift in demand. The resulting two‐stage robust optimization problem with mixed‐integer recourse is solved using an enhanced nested column‐and‐constraint generation algorithm. Through computational experiments, we examine the impact of the required level of flexibility on the supply chain design and the value of module mobility.
Autonomous De Novo Lead Halide CsPbBr 3 Perovskite Quantum Dots Synthesis Platform With Transfer Learning Accelerated Bayesian Optimization
Compared to traditional expert‐driven pattern, the emerging concept of self‐driving labs enabled by flow chemistry and artificial intelligence provides a new autonomous paradigm to significantly improve the R&D efficiency of functional material synthesis. In this work, focusing on on‐demand de novo synthesis of CsPbBr 3 quantum dots (QDs) by Ligand‐Assisted RePrecipitation (LARP) method, a micro TRansfer learning accelerated Bayesian Optimization driven reaction System (µTRBOS) is developed. Without any human supervision, µTRBOS can autonomously synthesize different types of high‐quality QDs, achieving user‐specified single‐peak fluorescent emission wavelengths between 455 and 505 nm with an error margin of less than 2 nm. These QDs exhibit particle sizes ranging from 2.5 to 7.4 nm. With knowledge transferred from existing experimental data, fewer than six experiments on average are required to optimize the synthetic conditions for each QD size. Additionally, the optimal synthetic conditions reveal the complex impact of temperature on LARP method.
Autonomous De Novo Lead Halide CsPbBr 3 Perovskite Quantum Dots Synthesis Platform With Transfer Learning Accelerated Bayesian Optimization
Compared to traditional expert‐driven pattern, the emerging concept of self‐driving labs enabled by flow chemistry and artificial intelligence provides a new autonomous paradigm to significantly improve the R&D efficiency of functional material synthesis. In this work, focusing on on‐demand de novo synthesis of CsPbBr 3 quantum dots (QDs) by Ligand‐Assisted RePrecipitation (LARP) method, a micro TRansfer learning accelerated Bayesian Optimization driven reaction System (µTRBOS) is developed. Without any human supervision, µTRBOS can autonomously synthesize different types of high‐quality QDs, achieving user‐specified single‐peak fluorescent emission wavelengths between 455 and 505 nm with an error margin of less than 2 nm. These QDs exhibit particle sizes ranging from 2.5 to 7.4 nm. With knowledge transferred from existing experimental data, fewer than six experiments on average are required to optimize the synthetic conditions for each QD size. Additionally, the optimal synthetic conditions reveal the complex impact of temperature on LARP method.
- Open Access
- 1 July 2025
- Published by The CACHE Foundation, Inc. in Systems and Control Transactions
- Vol. 4, 2081-2087
Integrating Renewable Energy and CO2 Utilization for Sustainable Chemical Production: A Superstructure Optimization Approach
Climate change, primarily caused by the excessive emission of greenhouse gases, particularly carbon dioxide (CO2), has intensified global efforts toward achieving carbon neutrality. In this context, renewable energy and CO2 utilization technologies have emerged as key strategies for reducing reliance on fossil fuels and mitigating environmental impacts. In this work, a superstructure model is developed to integrate renewable energy network and chemical production processes. The energy network integrates wind, solar, and biomass energy, complemented by storage systems to enhance reliability and reduce reliance on external power sources. The reaction network features various pathways that utilize CO2 as a raw material to produce high value-added chemicals such as polyglycolic acid (PGA), ethylene-vinyl acetate (EVA), and dimethyl carbonate (DMC), allowing for efficient conversion and resource utilization. A mixed-integer linear programming (MILP) model is formulated to minimize production costs while optimizing energy and reaction pathways. This research supports the green transition of the chemical industry by optimizing the integration of renewable energy and CO2-based chemical processes, promoting more sustainable production practices.
- 1 July 2025
- Published by The CACHE Foundation, Inc. in Systems and Control Transactions
- Vol. 4, 2081-2087
Open AccessIntegrating Renewable Energy and CO2 Utilization for Sustainable Chemical Production: A Superstructure Optimization Approach
Climate change, primarily caused by the excessive emission of greenhouse gases, particularly carbon dioxide (CO2), has intensified global efforts toward achieving carbon neutrality. In this context, renewable energy and CO2 utilization technologies have emerged as key strategies for reducing reliance on fossil fuels and mitigating environmental impacts. In this work, a superstructure model is developed to integrate renewable energy network and chemical production processes. The energy network integrates wind, solar, and biomass energy, complemented by storage systems to enhance reliability and reduce reliance on external power sources. The reaction network features various pathways that utilize CO2 as a raw material to produce high value-added chemicals such as polyglycolic acid (PGA), ethylene-vinyl acetate (EVA), and dimethyl carbonate (DMC), allowing for efficient conversion and resource utilization. A mixed-integer linear programming (MILP) model is formulated to minimize production costs while optimizing energy and reaction pathways. This research supports the green transition of the chemical industry by optimizing the integration of renewable energy and CO2-based chemical processes, promoting more sustainable production practices.
- 17 June 2025
- Published by American Chemical Society (ACS) in Industrial & Engineering Chemistry Research
- Vol. 64 (26) , 13258-13276
Optimal Design of Hybrid Energy-Driven Multiproduct Chemical Processes under Uncertainty
- 17 June 2025
- Published by American Chemical Society (ACS) in Industrial & Engineering Chemistry Research
- Vol. 64 (26) , 13258-13276
Optimal Design of Hybrid Energy-Driven Multiproduct Chemical Processes under Uncertainty
- 22 April 2025
- Published by Springer Nature in Journal of Global Optimization
- Vol. 92 (3) , 809-835
A strategy to tighten the relaxation of bilinear terms towards petrochemical scheduling problem
- 22 April 2025
- Published by Springer Nature in Journal of Global Optimization
- Vol. 92 (3) , 809-835
A strategy to tighten the relaxation of bilinear terms towards petrochemical scheduling problem
- 25 March 2025
- Published by Wiley in AIChE Journal
- Vol. 71 (7)
Adjustable robust superstructure optimization with mixed‐integer recourse for the continuous rufinamide manufacturing
To accelerate the applied research pace for the continuous synthesis of high value‐added rufinamide aiming at industrial applications, this work proposes a two‐stage adjustable robust optimization framework with mixed‐integer recourse to identify the optimal one from more than 500 possible alternative continuous synthetic routes. The overall rufinamide manufacturing process can be divided into three major processing steps, namely the halogenation of the precursor, azidation, and cycloaddition. A mixed‐integer nonlinear programming (MINLP) model is formulated under multiple uncertainties. Considering the numbering‐up of micro‐reactors for adjusting the production capacity, the mixed‐integer recourse leads to an intractable optimization problem. Hence, a tailored solution strategy based on the nested column‐and‐constraint generation (C&CG) is established to efficiently solve the resulting adjustable robust counterpart. Compared to the deterministic model, the benefits of implementing adjustable robust optimization (ARO) are fully demonstrated by evaluating the objective among sampled scenarios, where a cost reduction of up to 3% can be achieved.
- 25 March 2025
- Published by Wiley in AIChE Journal
- Vol. 71 (7)
Adjustable robust superstructure optimization with mixed‐integer recourse for the continuous rufinamide manufacturing
To accelerate the applied research pace for the continuous synthesis of high value‐added rufinamide aiming at industrial applications, this work proposes a two‐stage adjustable robust optimization framework with mixed‐integer recourse to identify the optimal one from more than 500 possible alternative continuous synthetic routes. The overall rufinamide manufacturing process can be divided into three major processing steps, namely the halogenation of the precursor, azidation, and cycloaddition. A mixed‐integer nonlinear programming (MINLP) model is formulated under multiple uncertainties. Considering the numbering‐up of micro‐reactors for adjusting the production capacity, the mixed‐integer recourse leads to an intractable optimization problem. Hence, a tailored solution strategy based on the nested column‐and‐constraint generation (C&CG) is established to efficiently solve the resulting adjustable robust counterpart. Compared to the deterministic model, the benefits of implementing adjustable robust optimization (ARO) are fully demonstrated by evaluating the objective among sampled scenarios, where a cost reduction of up to 3% can be achieved.