Machine learning in process systems engineering: Challenges and opportunities

Daoutidis, Prodromos; Lee, Jay H.; Rangarajan, Srinivas; Chiang, Leo; Gopaluni, Bhushan; Schweidtmann, Artur M.; Harjunkoski, Iiro; Mercangöz, Mehmet; Mesbah, Ali; Boukouvala, Fani; Lima, Fernando V.; del Rio Chanona, Antonio; Georgakis, Christos

doi:10.1016/j.compchemeng.2023.108523

Cited by 13 publications

(1 citation statement)

References 70 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, an intelligent model with high fidelity for time-series forecasting is needed to solve the multiple working condition problem during production . Subsequently, the system control, optimization, and scheduling of the power systems can be established based on the forecasting models. − Comprehensively considering the nonlinear, strong coupling, and time delay characteristics of the chemical process, artificial intelligence modeling approaches are suitable for handling complicated implicit correlations of multivariate variables in the long-time dimension compared with sequential modular or simultaneous equation approaches …”

Section: Introductionmentioning

confidence: 99%

Time-Series Forecasting of Large-Scale Green Ammonia Production: An Intelligent Dynamic Modeling Case in the Inner Mongolia Area of China

Zhang,

Li,

Zhang

et al. 2024

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

Renewable energy sources have been viewed as an important approach to solve the energy crisis and are carbon neutral. Compared with other storage mediums, green ammonia has the highest total energy efficiency and is an important raw material in the chemical industry. The fluctuation of renewable energy is a huge challenge for green ammonia production; thus, a high-precision prediction model is required to enable the building of an advanced control system for green ammonia production. In this study, a transformer-based multivariable multistep time-series prediction model that includes the temporal scales feature of renewable energy, called MSPTST, is proposed as the first step to solve the problem of unstable green ammonia production. The model outperformed other state-of-theart models, and its R 2 values were 0.986, 0.9998, and 0.990 under a high load condition, low renewable energy condition, and natural condition, respectively. Stable operation and swift load transitions are facilitated through a model predictive control based on the MSPTST framework. The impact of energy fluctuations on green ammonia synthesis is addressed by further integrating control theory and algorithms.

show abstract

Section: Introductionmentioning

confidence: 99%

Time-Series Forecasting of Large-Scale Green Ammonia Production: An Intelligent Dynamic Modeling Case in the Inner Mongolia Area of China

Zhang,

Li,

Zhang

et al. 2024

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

show abstract

Unveiling the interaction mechanisms of cold atmospheric plasma and amino acids by machine learning

Chai,

Wang,

Yusupov

et al. 2024

Plasma Processes & Polymers

View full text Add to dashboard Cite

Plasma medicine has attracted tremendous interest in a variety of medical conditions, ranging from wound healing to antimicrobial applications, even in cancer treatment, through the interactions of cold atmospheric plasma (CAP) and various biological tissues directly or indirectly. The underlying mechanisms of CAP treatment are still poorly understood although the oxidative effects of CAP with amino acids, peptides, and proteins have been explored experimentally. In this study, machine learning (ML) technology is introduced to efficiently unveil the interaction mechanisms of amino acids and reactive oxygen species (ROS) in seconds based on the data obtained from the reactive molecular dynamics (MD) simulations, which are performed to probe the interaction of five types of amino acids with various ROS on the timescale of hundreds of picoseconds but with the huge computational load of several days. The oxidative reactions typically start with H‐abstraction, and the details of the breaking and formation of chemical bonds are revealed; the modification types, such as nitrosylation, hydroxylation, and carbonylation, can be observed. The dose effects of ROS are also investigated by varying the number of ROS in the simulation box, indicating agreement with the experimental observation. To overcome the limits of timescales and the size of molecular systems in reactive MD simulations, a deep neural network (DNN) with five hidden layers is constructed according to the reaction data and employed to predict the type of oxidative modification and the probability of occurrence only in seconds as the dose of ROS varies. The well‐trained DNN can effectively and accurately predict the oxidative processes and productions, which greatly improves the computational efficiency by almost ten orders of magnitude compared with the reactive MD simulation. This study shows the great potential of ML technology to efficiently unveil the underpinning mechanisms in plasma medicine based on the data from reactive MD simulations or experimental measurements.

show abstract

Online Control and Optimization for Conventional and Emerging Carbon Capture Systems

Patrón,

Ricardez-Sandoval

2024

Reference Module in Materials Science and Materials Engineering

View full text Add to dashboard Cite

Machine learning in process systems engineering: Challenges and opportunities

Cited by 13 publications

References 70 publications

Time-Series Forecasting of Large-Scale Green Ammonia Production: An Intelligent Dynamic Modeling Case in the Inner Mongolia Area of China

Time-Series Forecasting of Large-Scale Green Ammonia Production: An Intelligent Dynamic Modeling Case in the Inner Mongolia Area of China

Unveiling the interaction mechanisms of cold atmospheric plasma and amino acids by machine learning

Online Control and Optimization for Conventional and Emerging Carbon Capture Systems

Contact Info

Product

Resources

About