Prediction of instantaneous flow characteristics of hydrocyclone with long short-term memory network based on computational fluid dynamics data

Powder Technology, Год журнала: 2024, Номер 439, С. 119668 - 119668

Опубликована: Март 16, 2024

Язык: Английский

---

Numerical simulation study of the phase transition heat transfer of nanoparticle-enhanced heat storage tubes

Applied Thermal Engineering, Год журнала: 2023, Номер 231, С. 121010 - 121010

Опубликована: Июнь 17, 2023

Язык: Английский

---

A hybrid safety monitoring framework for industrial FCC disengager coking rate based on FPM, CFD, and ML

Process Safety and Environmental Protection, Год журнала: 2023, Номер 175, С. 17 - 33

Опубликована: Май 10, 2023

Язык: Английский

---

Liquid–Liquid Dispersion Performance Prediction and Uncertainty Quantification Using Recurrent Neural Networks

Industrial & Engineering Chemistry Research, Год журнала: 2024, Номер 63(17), С. 7853 - 7875

Опубликована: Апрель 22, 2024

We demonstrate the application of a recurrent neural network (RNN) to perform multistep and multivariate time-series performance predictions for stirred static mixers as exemplars complex multiphase systems. employ two architectures in this study, fitted with either long short-term memory gated unit cells, which are trained on high-fidelity, three-dimensional, computational fluid dynamics simulations mixer performance, presence absence surfactants, terms drop size distributions interfacial areas function system parameters; these include physicochemical properties, geometry, operating conditions. Our results that while it is possible train RNNs single fully connected layer more efficiently than an encoder–decoder structure, latter shown be capable learning long-term underlying dispersion metrics. Details methodology presented, data preprocessing, RNN model exploration, methods visualization; ensemble-based procedure also introduced provide measure uncertainty. The workflow designed generic can deployed make other industrial applications similar data.

Язык: Английский

---

Machine learning analysis of pressure fluctuations in a gas-solid fluidized bed

Powder Technology, Год журнала: 2024, Номер 444, С. 120065 - 120065

Опубликована: Авг. 1, 2024

Язык: Английский

---

Study of atypical particle flow and free surface evolution behaviour in stirred tanks

Applied Thermal Engineering, Год журнала: 2023, Номер 234, С. 121320 - 121320

Опубликована: Авг. 10, 2023

Язык: Английский

---

Machine learning based prediction of biomass pyrolysis with detailed reaction kinetics for thermally-thick particles: from 1D to 0D

Chemical Engineering Science, Год журнала: 2023, Номер 280, С. 119060 - 119060

Опубликована: Июль 1, 2023

Язык: Английский

---

Numerical investigation of the gravity effect on two-phase flow and heat transfer of neon condensation inside horizontal tubes

Applied Thermal Engineering, Год журнала: 2023, Номер 233, С. 121162 - 121162

Опубликована: Июль 14, 2023

Язык: Английский

---

Predicting Spatiotemporal Distributions in a Bubbling Fluidized Bed for Biomass Fast Pyrolysis Using Convolutional Neural Networks

Industrial & Engineering Chemistry Research, Год журнала: 2024, Номер 63(8), С. 3744 - 3754

Опубликована: Фев. 15, 2024

Bubbling fluidized-bed biomass fast pyrolysis is a crucial technology for carbon neutrality and sustainability, computational fluid dynamics (CFD) one of the promising approaches to investigate optimize bubbling pyrolysis. However, traditional CFD still computationally costly pyrolysis, especially spatiotemporal transport-reaction behaviors, which are critical clarifying intrinsic characteristics optimizing operations. To address this issue, deep learning (DL) model centered on convolutional neural networks was developed based results efficiently predict distributions quantities each phase in fluidized bed Input DL sequence distributions, only an initial input required generate continuous outputs. The optimized by adjusting four typical parameters, i.e., length sequence, number neurons, rate, prediction step size. Accuracy short-term (10 frames) stability long-term (1000 were analyzed as well relationship between time-averaged length. It found that with satisfactory accuracy, several orders magnitude increase computation efficiency can be realized. Thus, paves way low-cost high-accuracy simulations

Язык: Английский

---

Optimization Research on the Performance of the RC-DTH Air Hammer Based on Computational Fluid Dynamics

Applied Sciences, Год журнала: 2025, Номер 15(2), С. 740 - 740

Опубликована: Янв. 13, 2025

To optimize the performance of RC-DTH air hammer, a mathematical model detailing each phase piston’s movement has been constructed in present work. Simultaneously, novel piston structure hammer (Type B) with diverse internal flow proposed. The impact structurally modified is analyzed using Computational Fluid Dynamics (CFD). Additionally, an energy testing system for developed to confirm validity numerical simulation results. Research results have shown that enhancing both intake stroke upper chamber (F1) and outlet lower (R2) can effectively improve performance. Conversely, increasing inlet (R1) (F2) tends diminish Moreover, quality influences its striking frequency while having minimal on single-impact energy. As increases, power diminishes. Once valve parameters are optimized, enhanced by 20.32%. Compared GQ89 Type B exhibits 84% increase 74% power.

Язык: Английский

---