Search

• Deeply investigation of pre-training models using different tokenization algorithms. • The BERT model for capturing advanced features of molecular is developed. • A novel deep learning framework is proposed for predicting complex properties. • Application of fine-tuning in correlating molecular structure and properties. With the rapid development of deep learning, research on quantitative structure–property relationships based on deep learning has received widespread attention. The deep learning architecture combining Bidirectional Encoder Representation from Transformers (BERT) and Feedforward Neural Networks (FNN) is proposed to compare the performance of different tokenization algorithms. And t-distributed stochastic neighbor embedding reveals valuable information about the mechanism of structure–property relationships. Additionally, a deep learning framework, BERT-Convolutional Neural Network (CNN)-FNN, is developed based on the optimal tokenization algorithm to accurately predict the σ -profile and V COSMO. The molecular structures are vectorized with the BERT model capturing local and global features of the entire molecule. And the CNN model enhances the latent representation associated with molecular properties, while the FNN model establishes the correlation. The deep learning frameworks predict σ -profile and V COSMO properties with R2 greater than 0.9703, making it a promising intelligent tool for guiding solvent design and screening. [ABSTRACT FROM AUTHOR]

• The efficient LTTMs were screened for the extraction of ethanol and ethyl propionate. • The relationship between the LLE behaviors and the weak interactions was investigated. • Molecular distributions of ethanol/ethyl propionate/LTTMs were unraveled. • Effect of LTTMs on the LLE of ethanol/ethyl propionate was experimentally explored. Low-transition temperature mixtures (LTTMs) have been proposed as green solvents for the separation of ethanol (EtOH)/ethyl propionate (EtPa) azeotropes. COSMO-SAC model was used to screen LTTMs, and ammonium acetate, ethylammonium chloride, and choline chloride were selected as hydrogen bond acceptors to form LTTMs with ethylene glycol. Liquid-liquid phase equilibrium experiments were carried out for the screened LTTMs and the separation system to confirm the validity of the selected solvents. And the distribution coefficients of EtOH in the three LTTMs were all greater than 1.2, and the selectivity for EtOH ranged from 6 to 50. The separation mechanism at the molecular level was revealed through quantum chemical calculations and wave function analysis. The thermodynamic behavior of the EtOH/EtPa/LTTMs system and the role of hydrogen bonding interactions were systematically investigated. The results indicated that electrostatic forces are the main molecular interactions, and the stronger interaction of LTTMs with ethanol is the basis for achieving the azeotropic separation. [ABSTRACT FROM AUTHOR]

[Display omitted] • The effecient DESs were screened for the separation of diethoxymethane and ethanol. • The key point of separating diethoxymethane and ethanol is identified. • The effect of DESs structure on LLE behavior was elucidated. • Microscopic mechanism on LLE behavior at the molecular level was revealed. The selection of efficient and environmental extractants is essential for the sustainable development of extraction processes. Based on the conductor-like screening model for the segment activity coefficient (COSMO-SAC), three deep eutectic solvents (DESs) were screened from 1000 DESs for the separation of diethoxymethane (DEM) and ethanol (EA). Subsequently, the liquid–liquid equilibrium behavior of the DEM + EA + DES system at different temperatures was measured. The results show that benzyltrimethylammonium chloride/glycerol (BTACl/GLY) has the highest selectivity and benzyltrimethylammonium chloride/ethylene glycol (BTACl/EG) and tetraethylammonium chloride/ethylene glycol (TEACl/EG) have similar extraction performance. More importantly, the extraction performance of DESs reported in this work is stronger than that of the solvents already reported. Additionally, analysis of the separation mechanism by various quantum chemical methods revealed that the purification of DEM and EA mixtures is mainly ascribed to the difference in electrostatic forces between the solvent and DEM/EA. The result confirmed that solvents with small molecular structures and strong nucleophilic sites are more competitive in separating the mixtures. [ABSTRACT FROM AUTHOR]

• The appropriate mixed solvent was selected for the desulphurization process. • Effect of the solvents on the LLE of thiophene/heptane was experimentally explored. • The relationship between LLE behaviors and microscopic interactions was unraveled. To meet the stringent requirements for sulfur content in gasoline, fuel oil desulfurization has become an indispensable process in petrochemical production. This study aimed to extract thiophene from heptane using a novel mixed solvent. In this work, the performance of single solvent N, N-dimethylformamide (DMF), dimethyl sulfoxide (DMSO) and sulfolane (SULF) was investigated for the separation of thiophene and heptane at 298.15 K or 308.15 K, respectively. Ternary liquid-liquid equilibrium (LLE) data (heptane + thiophene + DMF, heptane + thiophene + DMSO and heptane + thiophene + SULF) shows that DMF has the highest distribution coefficient and DMSO has the highest selectivity. Subsequently, compared to a single solvent, the mixed solvent composed of the molar ratio of 1:7 tetrabutylammonium bromide-SULF (DES1) and DMSO has the highest selectivity and applicable distribution coefficient. The conclusion can be obtained from the quaternary LLE data (heptane + thiophene + DMF/DMSO, heptane + thiophene + DES1/DMSO and heptane + thiophene + DES1/DMF). And the NRTL and UNIQUAC models were used to correlate the LLE data. The mechanism of thiophene extraction from thiophene/heptane was elucidated by molecular dynamics (MD) simulations. The analysis of non-bonded interaction energies shows that the electrostatic force plays a major role in the separation of thiophene/heptane for the mixed solvent, and radial distribution functions demonstrates that the oxygen atom of DMSO, nitrogen atom atom of DMF and bromide ion of DES1 have strong interaction with hydrogen atom of thiophene. Therefore, this study provides theoretical guidance for the application of mixed solvents in the desulfurization of fuel oil process. [Display omitted] [ABSTRACT FROM AUTHOR]

[Display omitted] • The new processes for separating benzene/isopropanol/water is proposed. • Multi-objective genetic algorithm is applied to optimize the proposed processes. • The proposed hybrid reactive-extractive distillation processes have better economic and environmental performances. The recycling of wastewater in chemical industry is of great significance because it can recover valuable organic solvents and prevent environmental pollution. In this work, benzene/isopropanol (IPA)/water ternary mixtures were separated by single and double reactive-extractive distillation column processes respectively. We conducted thermodynamic analysis on separation processes of ternary mixture to determine the feasible separation sequence. Furthermore, the processes are optimized by multi-objective genetic algorithm (MOGA) to obtain suitable operating conditions. And the processes were evaluated by the total annual cost (TAC), CO 2 emissions (E CO2), extraction efficiency (E ext) and thermodynamic efficiency. The results showed that compared with the traditional triple column extractive distillation (TCED) process, the TAC and E CO2 of the double column reactive-extractive distillation (DCRED) process were dramatically reduced by 52.9% and 49.5%, respectively. Similarly, TAC and E CO2 decreased by 49.5% and 53.5% in reactive-extractive dividing wall column (REDWC) process compared with the traditional process. With application of intensive processes, the final proposed process is more economically and environmentally sustainable than the separation system proposed in currently available literatures. This study can provide reference for the efficient separation of industrial wastewater containing benzene and IPA. [ABSTRACT FROM AUTHOR]

[Display omitted] • The reactive-extractive dividing wall column configuration is proposed for the separation of ethyl acetate/ethanol/water. • The waste heat of reactive-extractive dividing wall column process is recovered via organic Rankine cycles. • A multi-objective optimization approach considering safety is investigated. • The organic Rankine cycle assisted reactive-extractive dividing wall column processes have better economic performance. It is challenging and inevitable for the separation of ternary azeotropic mixture with multiple azeotropes in wastewater. Previous research proposed a hybrid double column reactive-extractive distillation (DCRED) process to separate ternary azeotropes in wastewater, which inspires the investigation into the reactive-extractive dividing wall column (REDWC) in this study. However, the design and waste heat recovery of REDWC have not been fully studied. To this end, two hybrid distillation processes, the REDWC process and the REDWC with feed preheating (REDWC-FP) process, are developed to separate ethyl acetate/ethanol/water ternary azeotropic mixture. The design parameters of the two hybrid distillation processes are optimized based upon the maximum total annual cost (TAC), CO 2 emissions and minimum process route index (PRI) through the multi-objective genetic algorithm (MOGA) considering safety. In view of massive waste heat in the REDWC process, several new configurations combining the REDWC and Organic Rankine cycles (ORCs) are further proposed. The optimal ORC system is determined via MOGA with the annual net profit, ORC thermodynamic efficiency and PRI as objective functions. Compared with the existing DCRED process, the TAC and CO 2 emissions of the REDWC process are reduced by 8.15% and 4.49%, respectively. The TACs of the ORC1-REDWC process with working fluid cyclohexane and the ORC2-REDWC process with working fluid benzene are reduced by 27.50% and 26.32%, respectively. [ABSTRACT FROM AUTHOR]

The gasoline fractionation scheme used in catalytic cracking has a significant impact on the blended gasoline's octane number. This study proposes an improved three-cut fractionation scheme for the fluid catalytic cracking gasoline fractionator and optimizes the process with respect to evaluation metrics such as total annual cost (TAC), light gasoline recovery, and CO 2 emissions (E CO2). To achieve energy efficiency and reduce emissions in the process, the dividing wall column is employed for process intensification. In comparison to the cascaded binary columns process, the middle cracked naphtha recovery of the FCC-DWC-M has increased by 1.96%, the TAC has decreased by 22.15%, and the E CO2 has decreased by 26.18%. Finally, the study investigates the dynamic characteristics of the FCC-DWC-M process, establishes a control scheme, and examines its performance while experiencing disturbances in feed flow rates of ±10% and ±20%. The findings indicate exceptional control performance of the FCC-DWC-M process. [Display omitted] • Propose two methods for fluid catalytic cracking gasoline fractionation. • The dividing wall column process is both cost-effective and eco-friendly. • The enhanced configuration's control scheme demonstrates strong robustness. [ABSTRACT FROM AUTHOR]

• The waste heat of extraction - extractive distillation is recovered by heat integration and organic Rankine cycles. • Compared with the heat-integrated process, organic Rankine cycle assisted process save 8.88% total annual cost. • The control scheme of organic Rankine cycle assisted extraction - extractive distillation process shows good robustness. The high-purity isopropanol butanol ethanol (IBE) mixture is a highly efficient gasoline additive that can be directly used without further separation. The product of isopropanol-butanol-ethanol fermentation contains a large amount of water, and a large amount of energy will be consumed in the purification process. In this paper, the separation IBE process is systematically studied from the aspects of process optimization, energy recovery and process control. The operation parameters of extraction combined with extractive distillation (E-ED) are optimized by a multi-objective genetic algorithm. Then, the pinch point technique is used to analyze the energy usage of the optimized process, and the maximum heat recovery heat integration scheme is obtained. Compared with the optimized process, the low pressure steam consumption of the heat integration process is reduced by 100%, the cooling water consumption is saved by 58.80%, and the TAC is reduced by 35.95%. In view of massive waste heat in the heat integration of the extraction combined with extractive distillation (E-ED-HI) process, a new configuration combining the E-ED-HI and organic Rankine cycle (E-ED-HI-ORC) is further proposed. The organic Rankine cycle using R600a can further reduce TAC by 8.88% based on the HI-E-ED process. Finally, the control scheme of E-ED process is proposed. The performance of the control scheme is examined with ±10% disturbance of the feed flow rate and composition change, and the results show that the process of extractive distillation and separation of IBE has good controllability. [ABSTRACT FROM AUTHOR]

• LTTMs were screened for ethyl acetate-isopropanol separation using COSMO-SAC. • VLE data LTTMs-ethyl acetate-isopropanol are measured. • The intermolecular mechanism was studied by using QC calculations. Ethyl acetate and isopropanol, as essential chemical ingredients and organic solvents, play an important role in the industry. At atmospheric pressure, ethyl acetate and isopropanol form an azeotrope, which can be separated by extractive distillation. The choice of extractant is critical in extractive distillation. Low transition temperature mixtures (LTTMs), which are emerging as efficient green solvents, are used as extractants in this work. The effect of type and structure on separation performance were studied, and three LTTMs were selected for the separation of the ethyl acetate-isopropanol system. The vapor–liquid equilibrium (VLE) data of ethyl acetate-isopropanol-LTTMs was measured by experiments, and the results showed that when the mole fraction of LTTMs was 0.1, all three LTTMs could break the azeotrope of ethyl acetate and isopropanol, especially LTTM1 (tetrabutylammonium bromide: ethanolamine = 1:2) having the best selectivity. To acquire the binary interaction parameters, the NRTL model was used to fit the experimental data, and the regression data was in good agreement with the experimental data. In addition, a quantum chemical method was used to investigate the interactions between molecules to gain insights into the microscopic mechanism. Results showed that the weak hydrogen bonding and van der Waals forces are formed between ethyl acetate and other systems, while moderate or weak hydrogen bonding are formed between isopropanol and other systems. [ABSTRACT FROM AUTHOR]

We investigated the effects of low dietary protein in isocaloric diets on sex reversal of Monopterus albus by evaluating the oestradiol (E2) and testosterone (T) concentrations, gonadosomatic index (GSI), sex ratio and gonad structure at the histological level. Fish (9.50 ± 1.50 g average initial weight; n=3 per group) were fed with five practical diets containing 100, 150, 200, 250 or 400 g kg−1 crude protein to apparent satiation for 15 months. Serum E2 and T concentrations were determined by radioimmunoassays. E2 concentrations and GSI significantly increased while T concentrations decreased as the dietary protein level was raised. Fish fed 400 g kg−1 of dietary protein had significantly higher E2 concentrations and GSI than those fed with lower dietary protein levels. The T concentrations of fish fed 100 g kg−1 of dietary protein was significantly higher than that of fish fed higher dietary protein levels. The shift of sex ratio towards more male and intersex fish was observed with decreasing dietary protein levels. Therefore, low dietary protein level may promote sex change from female to male in M. albus. This study provides important information for successful reproductive management and may be exploited for aquaculture of this species. [ABSTRACT FROM AUTHOR]

• The economic implications, controllability, and ecological benefits are selected as objective functions. • Complex network topologies are established on the real-time data. • Three control schemes are developed based on the principles of plant-wide control strategy. 1,3-Dioxolane (DOL) is an important chemical raw material with wide application, but the traditional production process of DOL has many problems that affect operational efficiency and quality, such as low conversion rate, high separation energy consumption, and complicated production process. Recently, a new process, which adopts reactive distillation and pressure swing distillation techniques, has been proposed for producing 1,3-dioxolane from aqueous formaldehyde solution and ethylene glycol. However, the researches on the process optimization and plant-wide control strategy for this new flowsheet are still in the original state. In this paper, multi-objective genetic algorithm (MOGA) is adopted to obtain the optimal process design parameters, and the economic implications, controllability, and ecological benefits of the process are selected as objective functions. Then, complex network topologies are established on the dynamic simulation data, and the key controlled variable is found by calculating central characteristics. Based on the principles of plant-wide control strategy and the complex network analysis results, three control schemes are developed in the dynamic simulation program, which can help to examine the robust stability and controllability of the system under different feed disturbances. The dynamic responses indicate that the CS3 scheme has better anti-interference ability, and the proposed analysis method can assist the development of the control system. [ABSTRACT FROM AUTHOR]