Article
Environmental Sciences
Adrienne M. Marshall, Emily Grubert
Summary: Hydroelectric power has unique technical characteristics that can become more valuable with the growth of variable generation renewables. However, its use for electricity generation is constrained by complex physical, safety, and socioenvironmental considerations. Simplified optimization models can benefit from the use of empirical parameter values, and the study combines a large dataset with generation data to provide insights on the empirical constraints of hydroelectricity in the United States.
Article
Environmental Sciences
Ricardo E. Palma-Goyes, Fabiola S. Sosa-Rodriguez, Fernando F. Rivera, Jorge Vazquez-Arenas
Summary: This study proposes a continuous physicochemical model for the active chlorine production used to degrade recalcitrant sulfamethoxazole in an electrochemical flow reactor. Experimental validation shows that increasing the adsorption of electro-generated chlorine species on the anode surface can significantly enhance the oxidation potential of the system.
ENVIRONMENTAL SCIENCE AND POLLUTION RESEARCH
(2022)
Article
Computer Science, Artificial Intelligence
Nuno Basurto, Carlos Cambra, Alvaro Herrero
Summary: The paper proposes different mechanisms to deal with data irregularities in order to increase anomaly detection rates in robots, including strategies to overcome missing values and class imbalance. The evaluation of these strategies shows their positive effect on improving one-class classification results.
Article
Engineering, Environmental
Bart Klumpers, Tim Luijten, Stijn Gerritse, Emiel Hensen, Ivo Filot
Summary: Microkinetic modelling is a crucial tool in studying heterogeneous catalysis but transport phenomena in chemical reactors affect performance. To overcome computational cost, artificial neural networks are used to substitute microkinetic models and improve efficiency. A reactor model for Fischer-Tropsch synthesis reveals the undesired promotion of CO2 formation due to re-adsorption of in-situ generated water.
CHEMICAL ENGINEERING JOURNAL
(2023)
Article
Construction & Building Technology
Alexander Kuempel, Jens Teichmann, Paul Mathis, Dirk Mueller
Summary: This paper presents modular models for hydronic subsystems in air-handling units, which are validated using measured data and shown to have high accuracy for controller testing and tuning.
JOURNAL OF BUILDING ENGINEERING
(2022)
Article
Engineering, Environmental
Jun Yin, Jiali Li, Iftekhar A. Karimi, Xiaonan Wang
Summary: Modeling is crucial for reactor or process design, control, and optimization, but developing high-fidelity mechanistic models for complex reactor systems is time-consuming. Machine learning models using data-driven methods can fill the gap between complex systems and limited knowledge. This research presents a machine learning model architecture specifically for dynamic modeling of general flow reactors, achieving higher model accuracy, data efficiency, and model interpretability than commonly used models.
CHEMICAL ENGINEERING JOURNAL
(2023)
Article
Engineering, Electrical & Electronic
Ying Xu, Zhikang Rao, Zhe Liu, Weiqiang Zheng, Yunhong Zhou, Ning Lu, Yuting Yang
Summary: In electrochemical sensing, the low selectivity caused by similar redox potentials of analyte and interference species is a limiting factor in the detection of broad-spectrum antibiotic chloramphenicol (CAP) against metronidazole (MNZ). In this study, a machine learning-assisted detection method for CAP was developed in the presence of high-concentration MNZ interference. By quantifying features in electrochemical profiles and training an artificial neural network model using correlated features, the CAP concentration could be accurately predicted. The results demonstrate that this machine learning-assisted electrochemical sensing scheme can minimize the interference of MNZ and accurately determine CAP concentration in various samples.
IEEE SENSORS JOURNAL
(2023)
Article
Education & Educational Research
Hayat Sahlaoui, El Arbi Abdellaoui Alaoui, Said Agoujil, Anand Nayyar
Summary: This study compares different variations of the Synthetic Minority Oversampling Technique (SMOTE) combined with classification algorithms to create prediction models. The results show that SMOTE with Edited Nearest Neighbors is superior, and the balanced random forest classifier performs better when using SMOTE-ENN, achieving 96% accuracy, precision, and F-value. Smote also has faster execution time. For model interpretability, combining Local Interpretable Model-agnostic Explanations (LIME) and SHapley Additive exPlanations (SHAP) provides deeper insights.
EDUCATION AND INFORMATION TECHNOLOGIES
(2023)
Article
Electrochemistry
Michiel Vranckaert, Hannes P. L. Gemoets, Ruben Dangreau, Koen Van Aken, Tom Breugelmans, Jonas Hereijgers
Summary: Synthetic organic electrochemistry has gained attention due to its environmental impact and energy demand advantages. Researchers propose a novel electrochemical reactor concept that improves mass transfer through oscillatory flow regime and conductive pillar field electrodes.
ELECTROCHIMICA ACTA
(2022)
Article
Computer Science, Artificial Intelligence
Wenjun Yang, Chaoqun Li
Summary: This paper introduces a novel label integration method based on game theory, called Stackelberg label inference (SLI). It addresses the issue of low label quality and avoids poor results caused by the involvement of multiple noisy label sets. SLI demonstrates superior performance in both label quality and model quality when the number of labelers is relatively small.
INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS
(2021)
Article
Environmental Sciences
R. S. Nerem, T. Frederikse, B. D. Hamlington
Summary: Using satellite altimetry data from 1993-2020, we estimated a quadratic model to analyze the climate-driven global mean sea level change. We also accounted for errors in the quadratic coefficients and projected the model 30 years into the future to 2050, with a 90% confidence interval. Our findings suggest that the global mean sea level in 2050 will be 16.4 cm higher than in 2020, with an uncertainty range of 11.3-21.4 cm. This prediction aligns with the sea level projections of IPCC SROCC and AR6, and the hindcast extrapolation prior to 1993 matches well with tide gauge records. We believe this highlights the value of short-term observationally driven extrapolations as an additional tool for predicting future sea level change.
Article
Environmental Sciences
Mateusz Norel, Krzysztof Krawiec, Zbigniew W. Kundzewicz
Summary: This study investigated the spatially organized links between river runoff time series and climate variability indices, concluding that ENSO is the primary determinant. Machine learning approach, particularly convolution neural network, was found to model river runoff better than traditional baseline methods.
Article
Environmental Sciences
Hojjatollah Mahboobi, Alireza Shakiba, Babak Mirbagheri
Summary: Groundwater is a crucial water resource that is increasingly contaminated. This study aims to improve the spatial accuracy of predicting groundwater nitrate concentration through integrating machine learning models using a local approach. The findings demonstrate that the ensemble of ML models using geographically weighted regression achieves the highest predictive performance.
JOURNAL OF ENVIRONMENTAL MANAGEMENT
(2023)
Article
Mathematics, Interdisciplinary Applications
Xiaolu Chen, Tongfeng Weng, Chunzi Li, Huijie Yang
Summary: Recent advances have shown the effectiveness of machine learning models in predicting chaotic systems. This study focused on three commonly used models and found that they have almost identical long-term statistical properties as learned chaotic systems. Additionally, synchronization among machine learning models was achieved through signal sharing.
CHAOS SOLITONS & FRACTALS
(2022)
Article
Acoustics
Joachim Dominique, Julien Christophe, Christophe Schram, Richard D. Sandberg
Summary: This paper introduces a new data-driven approach for establishing empirical models describing turbulent boundary layer wall-pressure spectra. The approach directly builds models from a dataset using Gene Expression Programming, and modifications of the GEP algorithm are proposed to address specific issues in modeling wall pressure spectra. The method demonstrates consistency and better data matching, suggesting new ways to predict the influence of moderate pressure gradient.
JOURNAL OF SOUND AND VIBRATION
(2021)
Article
Engineering, Chemical
Zhe Wu, Junwei Luo, David Rincon, Panagiotis D. Christofides
Summary: This study introduces a method using dropout and co-teaching learning algorithm to develop LSTM neural networks for capturing the true process dynamics from noisy data. The performance and robustness of the modeling approaches are evaluated on an industrial chemical reactor example using data generated from a large-scale process simulator.
CHEMICAL ENGINEERING RESEARCH & DESIGN
(2021)
Article
Engineering, Chemical
Joonbaek Jang, Martina Ruscher, Maximilian Winzely, Carlos G. Morales-Guio
Summary: This article reports the development of a gastight rotating cylinder electrode cell with well-defined mass transport characteristics that can experimentally decouple mass transfer effects from intrinsic kinetics in electrocatalytic systems. The gastight rotating cylinder electrode cell enables the dimensionless analysis of electrocatalytic systems and should facilitate the rigorous research and development of electrocatalytic technologies.
Article
Engineering, Chemical
Sungil Yun, Matthew Tom, Junwei Luo, Gerassimos Orkoulas, Panagiotis D. Christofides
Summary: With increasing demands for microchips and the nano-scale semiconductor manufacturing industry, atomic layer etching (ALE) has become a critical etching process. This work develops microscopic models to characterize the thermal ALE process of aluminum oxide thin films with two precursors, setting the foundation for real-time, model-based operational parameter calculations.
CHEMICAL ENGINEERING RESEARCH & DESIGN
(2022)
Article
Engineering, Chemical
Mohammed S. Alhajeri, Junwei Luo, Zhe Wu, Fahad Albalawi, Panagiotis D. Christofides
Summary: Recurrent neural networks have shown their remarkable accuracy in approximating the dynamic evolution of complex, nonlinear chemical processes. By incorporating physical knowledge, the structure of recurrent neural network models can be further improved to achieve better accuracy and computational efficiency. This study investigates the performance of model predictive control based on two different recurrent neural network structures and demonstrates the improvements in model accuracy and control performance using an example of a complex chemical process.
CHEMICAL ENGINEERING RESEARCH & DESIGN
(2022)
Review
Computer Science, Interdisciplinary Applications
Yi Ming Ren, Mohammed S. Alhajeri, Junwei Luo, Scarlett Chen, Fahim Abdullah, Zhe Wu, Panagiotis D. Christofides
Summary: This article presents an overview of recent developments in time-series neural network modeling and its use in model predictive control (MPC). A tutorial on constructing a neural network-based model is provided, along with discussion on key implementation issues. A nonlinear process example is introduced to demonstrate different neural network-based modeling approaches and evaluate their performance. Finally, future research directions on neural network modeling and its integration with MPC are briefly discussed.
COMPUTERS & CHEMICAL ENGINEERING
(2022)
Article
Engineering, Chemical
Berkay Citmaci, Junwei Luo, Joon Baek Jang, Vito Canuso, Derek Richard, Yi Ming Ren, Carlos G. Morales-Guio, Panagiotis D. Christofides
Summary: With the increasing supply of clean energy, electrochemical reduction of carbon dioxide (CO2) has become a significant alternative source of carbon-based fuels. However, the complexity of the reaction mechanism and the lack of efficient concentration measurement sensors pose challenges for modeling, optimizing, and controlling the CO2 reduction process. Machine learning techniques offer a potential solution by capturing the dynamic behavior of the process from data.
CHEMICAL ENGINEERING RESEARCH & DESIGN
(2022)
Article
Electrochemistry
Derek Richard, Matthew Tom, Joonbaek Jang, Sungil Yun, Panagiotis D. Christofides, Carlos G. Morales-Guio
Summary: Understanding the complexity of mass, momentum, charge, and heat transport and their impact on reaction kinetics at the electrode/electrolyte interface is a major challenge in the field of energy and catalysis. Developing multi-physics models that accurately capture the complexity of real-world devices is crucial for scaling up electrocatalytic systems. Combining experimental electrocatalysis with computational fluid dynamics (CFD) modeling can provide insights into the hydrodynamics of gastight rotating cylinder electrode (RCE) reactors.
ELECTROCHIMICA ACTA
(2023)
Article
Engineering, Chemical
Berkay Citmaci, Junwei Luo, Joon Baek Jang, Carlos G. Morales-Guio, Panagiotis D. Christofides
Summary: The electrochemical reduction of CO2 gas is a new technique for mitigating the global climate crisis and storing energy from renewable sources. However, there is a lack of explicit models for CO2 reduction and limited effort in developing process modeling and control of CO2 electrochemical reactors. This study focuses on developing a control scheme for a rotating cylinder electrode (RCE) reactor using artificial and recurrent neural network modeling, nonlinear optimization, and process controller design. The experimental results demonstrate the effectiveness of the control system in regulating the production rates of ethylene and carbon monoxide.
CHEMICAL ENGINEERING RESEARCH & DESIGN
(2023)
Article
Chemistry, Multidisciplinary
Kangze Shen, Simran Kumari, Yu-Chao Huang, Joonbaek Jang, Philippe Sautet, Carlos G. Morales-Guio
Summary: Electrochemical partial oxidation of methane to methanol using transition metal (oxy)hydroxides as catalysts is investigated. CoOx, NiOx, MnOx, and CuOx are found to be active for this reaction. Systematic studies are carried out to evaluate the effect of catalyst film thickness, overpotential, temperature, and hydrodynamics on activity and methanol selectivity. It is shown that high-valence transition metal oxides are inherently active for methane activation and oxidation to methanol, and electrocatalytic oxidation enables thermodynamically favorable production of methanol.
JOURNAL OF THE AMERICAN CHEMICAL SOCIETY
(2023)
Proceedings Paper
Automation & Control Systems
Zhe Wu, David Rincon, Junwei Luo, Panagiotis D. Christofides
Summary: This research focuses on using LSTM to model and predict control nonlinear processes, comparing the performance of standard LSTM on datasets with Gaussian noise and noisy industrial datasets, and proposing a dropout LSTM method using Monte Carlo dropout to train LSTM more efficiently with noisy data.
2021 AMERICAN CONTROL CONFERENCE (ACC)
(2021)
Proceedings Paper
Automation & Control Systems
Zhe Wu, David Rincon, Junwei Luo, Panagiotis D. Christofides
Summary: This work introduces a co-teaching method to improve model accuracy by using both noisy and noise-free data, and demonstrates its application in a chemical process through a case study.
2021 AMERICAN CONTROL CONFERENCE (ACC)
(2021)
Proceedings Paper
Automation & Control Systems
Zhe Wu, Junwei Luo, David Rincon, Panagiotis D. Christofides
Summary: A co-teaching learning algorithm is proposed in this study to capture the ground truth of chemical processes using LSTM networks from noisy data. Experimental results demonstrate that the co-teaching LSTM model is more accurate in predicting process dynamics and achieves better closed-loop performance under model predictive control compared to the standard training process.
