Pure hydrogen gas production in a coal supercritical water gasification system with CO2 as transporting medium

Applied Thermal Engineering, Год журнала: 2023, Номер 237, С. 121529 - 121529

Опубликована: Сен. 6, 2023

Язык: Английский

---

Monthly climate prediction using deep convolutional neural network and long short-term memory

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Июль 31, 2024

Climate change affects plant growth, food production, ecosystems, sustainable socio-economic development, and human health. The different artificial intelligence models are proposed to simulate climate parameters of Jinan city in China, include neural network (ANN), recurrent NN (RNN), long short-term memory (LSTM), deep convolutional (CNN), CNN-LSTM. These used forecast six climatic factors on a monthly ahead. data for 72 years (1 January 1951–31 December 2022) this study average atmospheric temperature, extreme minimum maximum precipitation, relative humidity, sunlight hours. time series 12 month delayed as input signals the models. efficiency examined utilizing diverse evaluation criteria namely mean absolute error, root square error (RMSE), correlation coefficient (R). modeling result inherits that hybrid CNN-LSTM model achieves greater accuracy than other compared significantly reduces forecasting one step For instance, RMSE values ANN, RNN, LSTM, CNN, temperature stage 2.0669, 1.4416, 1.3482, 0.8015 0.6292 °C, respectively. findings simulations shows potential improve forecasting. prediction will contribute meteorological disaster prevention reduction, well flood control drought resistance.

Язык: Английский

---

Machine Learning Model for CFD Simulations of Fluidized Bed Reactors

Industrial & Engineering Chemistry Research, Год журнала: 2025, Номер unknown

Опубликована: Янв. 3, 2025

Using detailed chemical kinetic models in CFD simulations of multiphase reactors is challenging. Detailed include radical species that span a wide range time scales, making the resulting system ODEs stiff. Solving large, stiff puts severe constraint on step, such impractical even for lab-scale reactors. Moreover, are difficult to converge. For this reason, most reactor rely global kinetics, when scheme available. This work targets problem, considering biomass thermochemical conversion at 1073–1273 K fluidized bed as an application. To end, gated recurrent unit (GRU) based neural network (RNN) model developed predict reactants and product evolution along length. Biomass devolatilization gas-phase chemistries represented by schemes comprising 20 with 24 reactions 39 118 reactions, respectively. A consisting ideal used generate training data. comprehensive compositions operating conditions used, ensuring applicability. The machine learning assessed against unseen test data CFD-DEM reactor. computational cost reduced 10 orders magnitude using GRU-based RNN model.

Язык: Английский

---

Microwave-assisted fluidized bed reactor pyrolysis of polypropylene plastic for pyrolysis gas production towards a sustainable development

Applied Energy, Год журнала: 2023, Номер 342, С. 121099 - 121099

Опубликована: Апрель 26, 2023

Язык: Английский

---

Machine learning applications in biomass pyrolysis: From biorefinery to end-of-life product management

Digital Chemical Engineering, Год журнала: 2023, Номер 8, С. 100103 - 100103

Опубликована: Май 16, 2023

The thermochemical conversion of biomass is a promising technology due to its cost-effectiveness and feedstock flexibility, with pyrolysis being particularly noteworthy method for diverse product range. Despite the potential pyrolysis, commercialization remains elusive, there growing need fully understand dynamics facilitate process scaling up. However, waste complex, time-consuming, capital-intensive. Machine Learning (ML) has emerged as possible means supporting accelerating research despite these challenges. This study provides comprehensive overview use ML in from biorefinery end-of-life management. In addition, success optimization control, predicting yield, real-time monitoring, life-cycle assessment (LCA), techno-economic analysis (TEA) during highlighted. Several methods have been utilized bid pyrolysis; potentiality artificial neural networks (ANNs) learn extremely non-linear input-output correlations led widespread adoption networks. Furthermore, current knowledge gaps future recommendations application are identified. Finally, this demonstrates development well scalability biomass.

Язык: Английский

---

Predicting biomass composition and operating conditions in fluidized bed biomass gasifiers: An automated machine learning approach combined with cooperative game theory

Energy, Год журнала: 2023, Номер 280, С. 128138 - 128138

Опубликована: Июнь 24, 2023

Язык: Английский

---

CFD-DPM data-driven GWO-SVR for fast prediction of nitrate decomposition in blast furnaces with nozzle arrangement optimization

Process Safety and Environmental Protection, Год журнала: 2023, Номер 176, С. 438 - 449

Опубликована: Июнь 15, 2023

Язык: Английский

---

Assessing bioenergy prospects of algal biomass and yard waste using an integrated hydrothermal carbonization and pyrolysis (HTC–PY): A detailed emission–to–ash characterization via diverse hyphenated analytical techniques and modelling strategies

Chemical Engineering Journal, Год журнала: 2024, Номер 492, С. 152335 - 152335

Опубликована: Май 18, 2024

Язык: Английский

---

Machine learning prediction of bio-oil production from the pyrolysis of lignocellulosic biomass: Recent advances and future perspectives

Journal of Analytical and Applied Pyrolysis, Год журнала: 2024, Номер 179, С. 106486 - 106486

Опубликована: Март 30, 2024

Bio-oil produced through pyrolysis of lignocellulosic biomass has recently received significant attention due to its possible uses as a second-generation biofuel. The yield and characteristics bio-oil are affected by reaction conditions the type feedstock that is used. Recently, machine learning (ML) techniques have been widely employed forecast performance bi-oil. In this study, comprehensive review ML research on carried out. Regression methods were most frequently build prediction models top five for random forest, artificial neural network, gradient boosting, support vector regression, linear regression. results developed quite consistent with experiment results. However, studies data had limitations such used restricted data, extraction features using their own knowledge, limited algorithms. We highlighted challenges potential cutting-edge in production.

Язык: Английский

---

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

---