Using an encoder-decoder convolutional neural network to predict the solid holdup patterns in a pseudo-2d fluidized bed

Bazai, H.; Kargar, Eshagh; Mehrabi, Mehdi

doi:10.1016/j.ces.2021.116886

Cited by 30 publications

(9 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The estimated range is the same as our previous work, 0.001−0.009. The number of neurons (16,32,64,128) and layers (1−4) in the LSTM module were examined to obtain the best value. In terms of the CLSTM model built on CNN and LSTM network (cases 5 and 6), two extra variables, convolutional kernel size, and stride size in the CNN layer were considered to examine the influence on predictive ability of instantaneous mass flow rates.…”

Section: Parameters Settingmentioning

confidence: 99%

“…Their results pointed out that the CNN-LSTM method is able to forecast more accurately than the LSTM model alone. Bazai et al 16 also revealed that an encoder-decoder CNN built on time series of CFD data is capable of correctly predicting particle volume fraction contours in a pseudo2d fluidized bed. In addition, an attention mechanism has also been considered as a possible way to improve model accuracy for ML.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Development and Evaluation of Deep Learning Models for Predicting Instantaneous Mass Flow Rates of Biomass Fast Pyrolysis in Bubbling Fluidized Beds

Zhong

Wei

et al. 2023

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

Computational fluid dynamics (CFD) has evolved into a vital tool for advancing bubbling fluidized-bed reactors for biomass fast pyrolysis. However, due to the enormous computational burden of CFD simulations, optimizing working parameters over a broad range or simulating large/industrial units is still extremely time-consuming. Because deep learning (DL) is a promising method to attain both precision and speed, two new DL models, which added an attention mechanism or a convolutional neuron network (CNN) layer in the basic long short-term memory (LSTM) model, were established to predict instantaneous mass flow rates of major species for biomass fast pyrolysis in a bubbling fluidized bed. Historical mass flow rates from a multifluid model (MFM) simulation were considered as the time series of data for the model training process. Influencing factors, including sequence length, learning rate, convolutional kernel and stride sizes in the CNN layer, and number of neurons and layers in LSTM module, were examined to improve forecasting ability. The results demonstrated that the hybrid model including both CNN and LSTM outperforms other models in predicting instantaneous mass flow rates of biomass fast pyrolysis in bubbling fluidized beds.

show abstract

Section: Parameters Settingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Development and Evaluation of Deep Learning Models for Predicting Instantaneous Mass Flow Rates of Biomass Fast Pyrolysis in Bubbling Fluidized Beds

Zhong

Wei

et al. 2023

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

show abstract

“…It was found that SVR performs best in the position reconstruction accuracy for all cases while RVR reconstruction speed outperforms SVR considerably due to the sparser nature of RVR. Except for these ML methods, researchers applied an encoder-decoder CNN to estimate the next-frame solid holdup pattern based on learning the first several frames as an input in a gas-particle fluidized bed . However, a possible challenge for this encoder-decoder DL is that the encoder must compress all input information into a fixed-length vector, and then pass it to the decoder.…”

Section: Current Status and Challengesmentioning

confidence: 99%

Review of Machine Learning for Hydrodynamics, Transport, and Reactions in Multiphase Flows and Reactors

Zhu

Chen

Ouyang

et al. 2022

Ind. Eng. Chem. Res.

120

View full text Add to dashboard Cite

Artificial intelligence (AI), machine learning (ML), and data science are leading to a promising transformative paradigm. ML, especially deep learning and physics-informed ML, is a valuable toolkit that complements incomplete domain-specific knowledge in conventional experimental and computational methods. ML can provide flexible techniques to facilitate the conceptual development of new robust predictive models for multiphase flows and reactors by finding hidden pattern/information/mechanism in a data set. Due to such emergence, we thereby comprehensively survey, explore, analyze, and discuss key advancements of recent ML applications to hydrodynamics, heat and mass transfer, and reactions in single-phase and multiphase flow systems from different aspects: (1) development of multiphase closure models of drag force, turbulence stresses and heat/mass transfer to improve the accuracy and efficiency of typical CFD simulations; (2) image reconstruction, regime identification, key parameter predictions, and optimization of multiphase flow and transport fields; (3) reaction kinetics modeling (e.g., predictions of reaction networks, kinetic parameters, and species production) and reaction condition optimization. These sections also discuss and analyze the key advantages and weakness of ML for solving the problems in the domain of multiphase flows and reactors. Finally, we summarize the under-solving challenges and opportunities in order to identify future directions that would be useful for the research community. Future development and study of multiphase flows and reactors are envisaged to be accelerated by ML and data science.

show abstract

“…This seems to be particularly true for machine eyesight systems that use deep learning. Finally, advancements in deep learning models techniques and applications, as well as experiments with more complicated designs, are due to two reasons: vast volumes of information and better processing performance [13] Implementation on true microcontrollers that processor speed and storage capacity, on the other hand, necessitates a different strategy [14]. Because the values, also known as design variables, play such a large role in deep learning model execution, the remedy for a deep learning model appropriate to different systems is a small model with fewer data to convey.…”

Section: Literature Surveymentioning

confidence: 99%

Evaluation of image quality based on visual perception using antagonistic networks in autonomous vehicles

Perumal¹,

Kulshrestha

Shastri

et al. 2022

ijhs

View full text Add to dashboard Cite

A method for just a point-to-point deep learning model for automated vehicles is described in this research. Our major goal was to develop an automated vehicle using a lightweight deep learning model that could be deployed on integrated modern vehicles. There is various point to point deep learning model used for automated vehicles, with camera pictures as input to the machine learning techniques and guiding direction projection as output, however, these deep learning methods are substantially more sophisticated than the cloud infrastructure we suggest. The proposed program's infrastructure, high computational, and summative assessment while automated vehicles are compared with different previous machine learning algorithms that We actually through order to achieve our goal, an accurate assessment. The proposed program's predictive model is 4 sets lower than PilotNet's and around 250 times less than AlexNet's. Although the innovative platform's intricacy and size are decreased in contrast to all other designs, resulting in reduced delay and greater refresh rate throughout reasoning, the model preserved its efficiency, accomplishing successful automated vehicles at the comparable economy as two additional designs. Furthermore, the proposed deep learning model decreased the processing capability, price, and storage requirements for true interpretation devices.

show abstract

Using an encoder-decoder convolutional neural network to predict the solid holdup patterns in a pseudo-2d fluidized bed

Cited by 30 publications

References 28 publications

Development and Evaluation of Deep Learning Models for Predicting Instantaneous Mass Flow Rates of Biomass Fast Pyrolysis in Bubbling Fluidized Beds

Development and Evaluation of Deep Learning Models for Predicting Instantaneous Mass Flow Rates of Biomass Fast Pyrolysis in Bubbling Fluidized Beds

Review of Machine Learning for Hydrodynamics, Transport, and Reactions in Multiphase Flows and Reactors

Evaluation of image quality based on visual perception using antagonistic networks in autonomous vehicles

Contact Info

Product

Resources

About