A Light Attention-Mixed-Base Deep Learning Architecture toward Process Multivariable Modeling and Knowledge Discovery

Li, Yue; Hu, Lijuan; Li, Ning; Shen, Weifeng

doi:10.1016/j.compchemeng.2023.108259

Cited by 7 publications

(2 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These features reflect the inherent properties and patterns within the data and thus partially reveal the intrinsic mechanism of the mechanism model and construct three neural network surrogate models. Subsequently, the surrogate models undergo hyperparameter optimization by embedding a hyperparameter optimization framework named Optuna, 57 followed by retraining, evaluation, and screening. Ultimately, a deep learning hybrid framework is established, offering robust assistance for subsequent optimization decisions.…”

Section: Methodsmentioning

confidence: 99%

A Deep Learning Hybrid Framework Combining an Efficient Evolutionary Algorithm for Complex Many-Objective Optimization of Sustainable Triple CO₂ Feed Methanol Production

Cao,

Li,

Chang

et al. 2024

ACS Sustainable Chem. Eng.

Self Cite

View full text Add to dashboard Cite

Current mainstream technologies have exhibited limits in integrating global many-objective optimization methods with chemical production systems, resulting in subpar outcomes in terms of energy efficiency and environmental issues for methanol production systems. In this study, a novel deep learning hybrid framework is proposed, which involves the construction of a mechanism model with the ability to elucidate the underlying principles and interrelationships of chemistry on a macroscopic scale and a data-driven model to enhance the accuracy and dependability of predictions from available data. The efficiency and global search capability of the proposed framework are further improved through the integration of an advanced evolutionary algorithm, which incorporates many-criteria decision-making technology to provide a comprehensive set of trade-offs for the optimal solution sets. The results demonstrate that all four objective functions of carbon dioxide emissions, methane conversion rate, methanol production, and energy consumption in the triple CO 2 feed methanol production system are rapidly optimized, in which carbon dioxide emissions and energy consumption are reduced by 18.50% and 3.15%, respectively. Consequently, this considerably enhances the environment. This proposed framework holds significant potential in facilitating the efficient optimization and sustainable production of complex systems within process engineering.

show abstract

Section: Methodsmentioning

confidence: 99%

A Deep Learning Hybrid Framework Combining an Efficient Evolutionary Algorithm for Complex Many-Objective Optimization of Sustainable Triple CO₂ Feed Methanol Production

Cao,

Li,

Chang

et al. 2024

ACS Sustainable Chem. Eng.

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, many factors comprehensively affecting the H–Q curve are challenging to characterize in the curve expression, leading to the difficulty in accurately predicting the dynamic pump pressure lift under the influences of complex factors. The backpropagation neural network has a superior ability to establish the model of complicated nonlinear systems under the effects of many factors. − Quite a few prediction models of pump characteristics have been developed by this approach. Fazll and Cem created a flow rate prediction model of a pump based on the data of pump speed, temperature, and outlet pressure under experimental conditions.…”

Section: Introductionmentioning

confidence: 99%

Intelligent Dynamic Prediction Model of Pump Pressure Lift for a Petroleum Pipeline System with Nonisothermal Batch Transportation

Jiang,

Yuan,

Chen

et al. 2023

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

Accurate prediction of pump characteristics is pivotal for optimizing transportation schemes and device operations in long-distance petroleum pipelines. In response to the variable nature of pump characteristics due to production circumstances, a novel dynamic prediction model of pump pressure lift is proposed by coupling the neural network, the intelligent evolutionary strategy, and an adaptive method of a neural network coefficient matrix. This model enables parameter adjustments to align with optimal directions through intelligent optimization strategies in which the differential evolution performs better than the genetic algorithm and particle swarm optimization. The proposed model exhibits superior accuracy and computational efficiency compared with classical static and dynamic prediction models. Notably, the model achieves a prediction error reduction of 1.34%, surpassing existing dynamic models while offering an 87.93% computation time reduction when contrasted with the static model

show abstract

Development of hybrid surrogate model structures for design and optimization of CO2 capture processes: Part I. Vacuum pressure swing adsorption in a confined space

Du,

Cao,

et al. 2024

Chemical Engineering Science

View full text Add to dashboard Cite

A Light Attention-Mixed-Base Deep Learning Architecture toward Process Multivariable Modeling and Knowledge Discovery

Cited by 7 publications

References 47 publications

A Deep Learning Hybrid Framework Combining an Efficient Evolutionary Algorithm for Complex Many-Objective Optimization of Sustainable Triple CO₂ Feed Methanol Production

A Deep Learning Hybrid Framework Combining an Efficient Evolutionary Algorithm for Complex Many-Objective Optimization of Sustainable Triple CO₂ Feed Methanol Production

Intelligent Dynamic Prediction Model of Pump Pressure Lift for a Petroleum Pipeline System with Nonisothermal Batch Transportation

Development of hybrid surrogate model structures for design and optimization of CO2 capture processes: Part I. Vacuum pressure swing adsorption in a confined space

Contact Info

Product

Resources

About

A Light Attention-Mixed-Base Deep Learning Architecture toward Process Multivariable Modeling and Knowledge Discovery

Cited by 7 publications

References 47 publications

A Deep Learning Hybrid Framework Combining an Efficient Evolutionary Algorithm for Complex Many-Objective Optimization of Sustainable Triple CO2 Feed Methanol Production

A Deep Learning Hybrid Framework Combining an Efficient Evolutionary Algorithm for Complex Many-Objective Optimization of Sustainable Triple CO2 Feed Methanol Production

Intelligent Dynamic Prediction Model of Pump Pressure Lift for a Petroleum Pipeline System with Nonisothermal Batch Transportation

Development of hybrid surrogate model structures for design and optimization of CO2 capture processes: Part I. Vacuum pressure swing adsorption in a confined space

Contact Info

Product

Resources

About

A Deep Learning Hybrid Framework Combining an Efficient Evolutionary Algorithm for Complex Many-Objective Optimization of Sustainable Triple CO₂ Feed Methanol Production

A Deep Learning Hybrid Framework Combining an Efficient Evolutionary Algorithm for Complex Many-Objective Optimization of Sustainable Triple CO₂ Feed Methanol Production