Distillation Process Research Articles

Applicability exploration and sustainable assessment of heat integration and vapor recompression heat pump to side-stream extractive distillation processes for separating ternary azeotropic system

The side-stream extractive distillation (SSED) process is attracting extensive attention due to its high energy efficiency for separating ternary azeotropic mixture. However, the correlation between the bottom temperature level of distillation columns depending on the flowrates and compositions of side streams and the introduction of heat integration (HI) and vapor recompression heat pump (VRHP) technology is ignored. In the work, a range of initialization extractive distillation processes including one conventional extractive distillation (CED) process and five SSED processes considering different flowrates, compositions, and feed locations of side streams are proposed for separating acetonitrile/isopropanol/water ternary azeotropic system. The optimization results with the minimum the total annual costs (TAC) as the optimization objective show that not all SSED schemes are superior to the CED separation process, and the SSED separation schemes can achieve up to a 2.72 % reduction in the TAC. Following that, the HI and VRHP assisted extractive distillation schemes are proposed based on the temperature levels between the top and bottom of different distillation columns and compressor restraints for the initialization extractive distillation schemes mentioned after optimization. Moreover, sustainable evaluations covering economy, energy consumption, environment, and energy utilization are implemented for the extractive distillation schemes proposed. The evaluation results indicate that the CED-HI scheme overcomes all SSED-HI and SSED-VRHP schemes in terms of economic costs, but the optimal SSED-HI processes can reduce 8.77 % of energy consumption and 4.73 % of CO2 emissions, while the energy utilization raised from 1.70 % to 1.78 % compared to the CED-HI process.

Comparative analysis of strategies of knowledge distillation on BERT for text matching

Large language model is a highly effective and promising language model technology that can improve the performance and robustness of natural language processing tasks. As a representative work, Bidirectional Encoder Representations from Transformers (BERT) has excellent performance in various natural language processing tasks. This model is pre-trained on large-scale language dataset and has gained attention from all walks of life since its introduction. However, its huge number of participants and scale make its performance in mobile very limited. As an effective technique to compress neural network, knowledge distillation can obtain a lightweight model with smaller parameters without losing too much model performance. Therefore, distillation of BERT models has been started, aiming at obtaining lightweight BERT models. In this paper, we will introduce several common BERT distillation models and analyse their model architecture, distillation process, and finally the compression efficiency and model effectiveness. It is concluded that the process of increasing recompression efficiency is often accompanied by decreasing model effectiveness.

Novel Study on Cryogenic Distillation Process and Application by Using CHEMCAD Simulation.

The purpose of this study was to examine the cryogenic separation process used in air separation plants with the CHEMCAD simulation program. A preliminary analysis was carried out with the literature about the operation of the process. In light of the technical information, the simulation of the cryogenic separation process was utilized with the program. While no changes were analyzed in the basic parts of the process, the filtration of the air and separation from moisture and carbon dioxide were not included in the simulation. The simulation showed the three basic components of air, nitrogen, oxygen, and argon obtained as a product of the desired purity rather than a one-to-one demonstration of the applications in the industry. In addition, the low-temperature separation process of an air separation unit was studied to obtain high-purity products (99.49-99.99%) achieving the expected separation efficiency. As a result, the production of all three substances in desired purity was achieved successfully.

Conceptual methods for synthesis of reactive distillation processes: recent developments and perspectives

AbstractReactive distillation (RD) is a process intensification technique that offers major advantages over conventional technologies by enabling the integration of reaction and separation into a single apparatus. This integration introduces and exploits complex interaction between mass transfer, chemical reaction and hydrodynamics within an RD column; however, these complexities can hinder the adoption of RD by industry. Many approaches have been developed by the scientific community to advance understanding and to expedite the initialization of RD design at the conceptual level. This paper critically discusses recent developments in conceptual methods for synthesizing RD processes. This review paper is the first to consider the range of available approaches for assessing the technical feasibility, controllability, economic viability, and sustainability of RD units by taking into account various configurations in which RD is treated as a new unit operation, and as part of process synthesis in a different way of designing processes based on functions. The review also addresses complex configurations, such as advanced RD technologies. Special attention is paid to process modeling and simulation as well as to education. Knowledge gaps to be filled by further research are indicated. © 2024 The Authors. Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI).

TIKP: Text-to-Image Knowledge Preservation for Continual Semantic Segmentation

Continual Semantic Segmentation (CSS) is an emerging trend, where catastrophic forgetting has been a perplexing problem. In this paper, we propose a Text-to-Image Knowledge Preservation (TIKP) framework to address this issue. TIKP applies Text-to-Image techniques to CSS by automatically generating prompts and content adaptation. It extracts associations between the labels of seen data and constructs text-level prompts based on these associations, which are preserved and maintained at each incremental step. During training, these prompts generate correlated images to mitigate the catastrophic forgetting. Particularly, as the generated images may have different distributions from the original data, TIKP transfers the knowledge by a content adaption loss, which determines the role played by the generated images in incremental training based on the similarity. In addition, for the classifier, we use the previous model from a different perspective: misclassifying new classes into old objects instead of the background. We propose a knowledge distillation loss based on wrong labels, enabling us to attribute varying weights to individual objects during the distillation process. Extensive experiments conducted in the same setting show that TIKP outperforms state-of-the-art methods by a large margin on benchmark datasets.

Distilling Reliable Knowledge for Instance-Dependent Partial Label Learning

Partial label learning (PLL) refers to the classification task where each training instance is ambiguously annotated with a set of candidate labels. Despite substantial advancements in tackling this challenge, limited attention has been devoted to a more specific and realistic setting, denoted as instance-dependent partial label learning (IDPLL). Within this contex, the assignment of partial labels depends on the distinct features of individual instances, rather than being random. In this paper, we initiate an exploration into a self-distillation framework for this problem, driven by the proven effectiveness and stability of this framework. Nonetheless, a crucial shortfall is identified: the foundational assumption central to IDPLL, involving what we term as partial label knowledge stipulating that candidate labels should exhibit superior confidence compared to non-candidates, is not fully upheld within the distillation process. To address this challenge, we introduce DIRK, a novel distillation approach that leverages a rectification process to DIstill Reliable Knowledge, while concurrently preserves informative fine-grained label confidence. In addition, to harness the rectified confidence to its fullest potential, we propose a knowledge-based representation refinement module, seamlessly integrated into the DIRK framework. This module effectively transmits the essence of similarity knowledge from the label space to the feature space, thereby amplifying representation learning and subsequently engendering marked improvements in model performance. Experiments and analysis on multiple datasets validate the rationality and superiority of our proposed approach.

DI-V2X: Learning Domain-Invariant Representation for Vehicle-Infrastructure Collaborative 3D Object Detection

Vehicle-to-Everything (V2X) collaborative perception has recently gained significant attention due to its capability to enhance scene understanding by integrating information from various agents, e.g., vehicles, and infrastructure. However, current works often treat the information from each agent equally, ignoring the inherent domain gap caused by the utilization of different LiDAR sensors of each agent, thus leading to suboptimal performance. In this paper, we propose DI-V2X, that aims to learn Domain-Invariant representations through a new distillation framework to mitigate the domain discrepancy in the context of V2X 3D object detection. DI-V2X comprises three essential components: a domain-mixing instance augmentation (DMA) module, a progressive domain-invariant distillation (PDD) module, and a domain-adaptive fusion (DAF) module. Specifically, DMA builds a domain-mixing 3D instance bank for the teacher and student models during training, resulting in aligned data representation. Next, PDD encourages the student models from different domains to gradually learn a domain-invariant feature representation towards the teacher, where the overlapping regions between agents are employed as guidance to facilitate the distillation process. Furthermore, DAF closes the domain gap between the students by incorporating calibration-aware domain-adaptive attention. Extensive experiments on the challenging DAIR-V2X and V2XSet benchmark datasets demonstrate DI-V2X achieves remarkable performance, outperforming all the previous V2X models. Code is available at https://github.com/Serenos/DI-V2X.

Performance assessment and optimization of adsorption-assisted air gap membrane distillation process for remediation of chromium-contaminated groundwater

Chromium contamination in groundwater is becoming a significant concern due to its potential contribution to the global cancer burden. Membrane distillation (MD) is a promising method known for producing high-quality water, yet issues namely temperature and concentration polarizations should be addressed. To overcome these issues, adsorption (ADS) is proposed to reduce Chromium concentration in the concentrated feed and enhance water production and resource recovery. This study assesses the performance optimization considered parameter effects of the process individually and subsequently executes it into a combined process. Slow sand filtration (SSF) is performed to eliminate particles and thus mitigate fouling. A novel Fe3O4@C-SiO2 is introduced in the ADS process. The technical and economic performances of single MD, hybrid systems (ADS-MD and SSF-MD), and an integrated SSF-ADS-MD system are compared. The results showed the integrated process outperformed various scenarios, achieving significant water recovery almost a 2-fold increase in a single process, and notable reductions in Chromium mass (485.3 gr) and abatement (31.4 mg/h). The economic analysis indicated that the benefit-cost ratio of the integrated process is slightly higher compared to hybrid ADS-MD primarily because of the lower costs in SSF. Ultimately, the results underscored the process of SSF and ADS in boosting MD performance, particularly in water production and Chromium reduction, with prudent economic considerations.

Dynamic control strategy of the ternary azeotrope separation process assisted by reactive-extractive distillation for ethyl acetate/isopropanol/water

The double-column reactive-extractive distillation (DCRED) process is a promisingly intensified compactness configuration for separating ternary water-containing azeotrope(s) by introducing the external reactions, such as the typical common adopted epoxy ethane (EO) hydration reaction. However, limited investigation has focused on the dynamic control analysis for the highly strong coupled DCRED process. In this paper, the efficient and robust control structure is proposed to achieve the dynamic operation of the DCRED process for the system of ethyl acetate/isopropanol/water. The composition controller is applied to precisely control the feed ratio of EO and water, thereby ensuring the qualified products when facing ± 10 % feed flowrate disturbances and ± 5 % feed composition disturbances. This work provides an alternative control strategy for the DCRED process, which is beneficial for its practical industrial applications.

MSSD: multi-scale self-distillation for object detection

Knowledge distillation techniques have been widely used in the field of deep learning, usually by extracting valid information from a neural network with a large number of parameters and a high learning capacity (the teacher model) to a neural network with a small number of parameters and a low learning capacity (the student model). However, there are inefficiencies in the transfer of knowledge between teacher and student. The student model does not fully learn all the knowledge of the teacher model. Therefore, we aim to achieve knowledge distillation of our network layer by a single model, i.e., self-distillation. We also apply the idea of self-distillation to the object detection task and propose a multi-scale self-distillation approach, where we argue that knowledge distillation of the information contained in feature maps at different scales can help the model better detect small targets. In addition, we propose a Gaussian mask based on the target region as an auxiliary detection method to improve the accuracy of target position detection in the distillation process. We then validate our approach on the KITTI dataset using a single-stage detector YOLO. The results demonstrate a 2.8% improvement in accuracy over the baseline model without the use of a teacher model.

Simultaneous production of electricity and potable water underwater by integrating concentrating photovoltaic with air gap membrane distillation

This paper combines underwater concentrating photovoltaic and air gap membrane distillation, enabling the simultaneous production of electricity and potable water in underwater environments. Given its submerged nature and absence of land occupation, it proves particularly advantageous for land-deficient regions with limited access to electricity and potable water, such as islands and Coastal cities. It primarily comprises innovatively designed underwater solar concentrators featuring extensive light interception angles, photovoltaic modules, and membrane distillation units. It utilizes exhausted heat from photovoltaic modules to separate clean water from seawater via membrane distillation process. Furthermore, it is engineered to prevent any contact between metal materials and seawater, thereby mitigating the risk of corrosion. An experimental setup with photovoltaic cells’ areas of 0.006 m2 is developed and tested in the laboratory and under real weather conditions to assess its performance. The steady-state experimental findings indicate that as solar irradiation increases from 500 W/m2 to 900 W/m2, its electricity generation efficiency ranges from 9.0% to 9.7%, while the water yield efficiency ranges from 32.8% to 35.1%. Besides, outdoor experiments demonstrate that its power output during normal incidence ranges from 38.89 W/m2-85.56 W/m2, with an average electrical efficiency of 10.5%. Its average potable water yield reaches 0.537 kg/m2/h and average water production efficiency reaches 48.3%.

Class similarity weighted knowledge distillation for few shot incremental learning

Few-shot class incremental learning illustrates the challenges of learning new concepts, where the learner can access only a small sample per concept. The standard incremental learning techniques cannot be applied directly because of the small number of samples for training. Moreover, catastrophic forgetting is the propensity of an Artificial Neural Network to fully and abruptly forget previously learned knowledge upon learning new knowledge. This problem happens due to a lack of supervision in older classes or an imbalance between the old and new classes. In this work, we propose a new distillation structure to tackle the forgetting and overfitting issues. Particularly, we suggest a dual distillation module that adaptably draws knowledge from two different but complementary teachers. The first teacher is the base model, which has been trained on large class data, and the second teacher is the updated model from the previous K-1 session, which contains the modified knowledge of previously observed new classes. Thus, the first teacher can reduce overfitting issues by transferring the knowledge obtained from the base classes to the new classes. While the second teacher can reduce knowledge forgetting by distilling knowledge from the previous model. Additionally, we use semantic information as word embedding to facilitate the distillation process. To align visual and semantic vectors, we used the attention mechanism of the embedding of visual data. With extensive experiments on different data sets such as Mini-ImageNet, CIFAR100, and CUB200, our model shows state-of-the-art performance compared to the existing few shot incremental learning methods.

Hydrophobicity and membrane distillation performance of glass fiber membranes modified by dip coating of pure PDMS

With growing concern for environmental issues associated with the use of chemicals and nanomaterials to achieve membrane hydrophobicity, polydimethylsiloxane has emerged as a promising and eco-friendly alternative. However, few studies have explored the application of pure PDMS solutions through dip coating to achieve hydrophobicity in membranes designed for membrane distillation; a process which demands the membranes possess specific porosity and pore size characteristics. This study explored how dip coating with different PDMS concentrations affected the hydrophobicity and physical properties of glass fiber membranes. Increasing the concentration of polydimethylsiloxane up to 15% resulted in a stepwise rise in hydrophobicity, but the functionality as a membrane was compromised when the concentration reached 20%. Hydrophobicity on the membrane surface was higher with 15% PDMS treatment than on the chemically grafted membrane, but the latter demonstrated superior performance in the membrane distillation process. Nevertheless, the study confirms the potential of the pure PDMS-coated membrane as a viable option for membrane distillation, with the possibility of enhancing performance through improvements in pore size and surface morphology.

Enhancing water recovery through integrated graphene oxide-modified forward osmosis and membrane distillation for real textile wastewater treatment

The textile industry is a highly water-intensive sector, emphasizing the increasing importance of reclaiming clean water from textile wastewater and advancing water recovery towards zero liquid discharge. This study proposes and assesses an integrated forward osmosis and membrane distillation (FO-MD) process for treating real textile wastewater, aiming for enhanced water recovery and ensure robustness towards zero liquid discharge. Firstly, a newly developed graphene oxide (GO)-enhanced FO membrane (GFO) was fabricated by adding GO into the polyamide layer during the interfacial polymerization process. Furthermore, the real textile wastewater treatment performance of the FO-MD process based on the GFO membrane was compared to that of a commercial FO membrane (CFO). The results indicated the GFO-based FO-MD process outperformed due to enhanced antifouling properties and stable water flux. Remarkably, the GFO membrane-based FO-MD system achieved over 80% water recovery and nearly 100% rejections for COD, Na+, K+, Cl-, and S2- during continuous 24-hour filtration. The findings highlighted FO-MD as a viable solution, surpassing the individual FO and MD processes by overcoming the limitations of FO draw solution dilution and MD membrane wetting challenges. Moreover, a preliminary techno-economic study for a 200 m3 day−1 textile wastewater estimated water production costs of $6.48 m−3 and $5.73 m−3 for CFO-MD and GFO-MD systems. The analysis highlighted maintaining the water flux above 10 L m−2 h−1 is crucial for economic viability. This work demonstrates the potential of the FO-MD system as a sustainable solution to address the textile wastewater treatment challenges, providing insights for FO-MD system design and optimization.

Study on the separation of methyl-ethyl-ketone / isopropanol / water ternary mixture by energy-efficient extractive distillation processes

The separation of ternary mixtures of methyl-ethyl-ketone (MEK)/isopropanol (IPA)/water, which are produced through chemical processes, is an imperative task aimed at reducing energy expenditures and mitigating environmental contamination. In the industry, Ethylene glycol (EG) and dimethyl sulfoxide (DMSO) are commonly employed as extractants. The selection of EG as the extractant was based on an investigation into the impact of relative volatility. Three distinct processes, namely conventional extractive distillation (CED), side-stream extractive distillation (SED), and heat pump extractive distillation (HPED), were evaluated by total annual cost (TAC), total energy consumption (TEC) and carbon dioxide emissions (CO2 emissions). The results demonstrate that both the enhanced SED and HPED processes effectively decrease energy consumption. Specifically, the TEC for the SED process exhibits a reduction of 15.36% compared to the CED process, resulting in a TAC saving of 9.06% and a decrease in CO2 emissions by 12.37%. Similarly, the HPED process surpasses the CED process by achieving a TAC saving of 15.41% and a CO2 emissions reduction of 22.61%. Consequently, this research presents a novel approach for the separation of MEK/IPA/water, offering significant practical implications.

Complete separation of benzene-cyclohexene-cyclohexane mixtures via temperature-dependent molecular sieving by a flexible chain-like coordination polymer

The separation and purification of C6 cyclic hydrocarbons (benzene, cyclohexene, cyclohexane) represent a critically important but energy intensive process. Developing adsorptive separation technique to replace thermally driven distillation processes holds great promise to significantly reduce energy consumption. Here we report a flexible one-dimensional coordination polymer as an efficient adsorbent to discriminate ternary C6 cyclic hydrocarbons via an ideal molecular sieving mechanism. The compound undergoes fully reversible structural transformation associated with removal/re-coordination of water molecules and between activated and hydrocarbon-loaded forms. It exhibits distinct temperature- and adsorbate-dependent adsorption behavior which facilitates the complete separation of benzene, cyclohexene and cyclohexane from their binary and ternary mixtures, with the record-high uptake ratios for C6H6/C6H12 and C6H10/C6H12 in vapor phase and highest binary and ternary selectivities in liquid phase. In situ infrared spectroscopic analysis and ab initio calculations provide insight into the host-guest interactions and their effect on the preferential adsorption and structural transformation.

Comparison of the composition and cost estimation of steam distillation and supercritical fluid extraction from leaves of Lippia graveolens

Lippia graveolens, mostly known as Mexican oregano, is an aromatic plant with great economic im-portance. However, very little is known about the reasons that could explain the variation observed with respect to the yield and composition of the essential oil obtained. The effect of steam distillation and supercritical fluid extraction processes on composition, yield and economic viability were evaluated. The essential oil was obtained by steam distillation, with a laboratory scale equipment and supercritical extraction equipment with a 1L capacity cell. The extracts were analyzed by GC–MS. where Carvacrol was the majority compound in both processes. The essential oil is obtained by steam distillation, with laboratory scale equipment and supercritical extraction equipment with a 1L capacity cell. In the case of steam distillation, a maximum yield of 4.1% on a dry weight basis was obtained, while the supercritical fluid extraction obtained provided a maximum yield of 2.8% on a dry weight basis at the same bed porosity (0.91). The costs that were obtained in a period of one year through. The process of extraction by steam distillation showed good results to obtained essential oil from L. graveolens. where it is observed that steam distillation is still the best option.

Experimental activities in view of hydrogen- deuterium equilibrium reaction in dielectric barrier discharge reactor: Effects of plasma treatment parameters and reactor design

Cryogenic distillation (CD) is a crucial process for separating and concentrating tritium in the tritium fuel cycle of fusion reactors. The equilibrator plays a vital role in decomposing HD, DT, and HT to improve the separation efficiency of hydrogen isotopes in CD. As the equilibrium reaction temperature decreases, a larger proportion of HD, HT, and DT decomposes into H2, D2, and T2, which enhances the performance of the equilibrator. While Pt/Al2O3 catalyst exhibits high activity at 303 K in the H2-D2 equilibrium reaction, it is almost inactive at 77 K. Non-thermal plasma (NTP) technology offers a promising and attractive alternative to conventional catalytic technology. In this study, H2-D2 equilibrium reaction was carried out without catalysts in a dielectric barrier discharge (DBD) reactor, achieving efficient conversion of HD even at 77 K. In the single-electrode DBD plasma reactor (S-P-Reactor), with an input power of 50 W, the conversion efficiency of HD was found to be 99.5% at 303 K, 99.5% at 273 K, and 99.2% at 77 K. To further optimize the energy efficiency of the S-P-Reactor, the reactor structure was enhanced, leading to the development of a multi-high-voltage DBD reactor (M-P-Reactor). The input power of the M-P-Reactor was reduced by approximately 10 W compared to the S-P-Reactor, while maintaining a conversion efficiency of over 95% for HD. The energy density of the reactor was calculated using the Q-V Lissajous graphic method, demonstrating that the M-P-Reactor exhibited superior energy conversion efficiency compared to the S-P-Reactor. Furthermore, the yield of HD was solely determined by the energy density and temperature. At a specific energy density, the yield of HD remained consistent for the same temperature. However, generating HD at lower temperatures required more energy due to the lower reactant activity. These results illustrate the potential application of NTP technology in improving hydrogen isotope separation performance in the cryogenic distillation process for fusion applications.

Energy reduction and efficiency separation method of associated gas in oilfield based on EOR-CO2

Cryogenic distillation is considered to be the most promising technology for the treatment of CO2-EOR gases, but it is an energy-intensive process. Considerable scientific efforts are being made to reduce energy consumption in cryogenic distillation and to improve process efficiency. Therefore, this study proposes a unique low-temperature distillation process. The novelty of the designed distillation process lies in the fact that using a small amount of refluxed liquid CO2 mixed with untreated oilfield-associated gas into the distillation column reduces the process energy consumption and improves the rectification effect. The sensitivity analysis of different design key parameters of the system is carried out using ASPEN HYSYS, and the energy consumption is regarded as the key function of this cycle optimization. The simulation results show that the optimized low-temperature distillation process has lower energy consumption and better distillation results. Compared with the traditional process, the product purity of the optimized process increased from 94.01% to 94.82%, the product recovery increased from 95.87% to 98.46%, The energy consumption of the reboiler and condenser of the distillation column was reduced, and the total energy consumption was reduced from 1.147MJ/kg to 1.001 MJ/kg.

Parametric Optimization of a Reactive Distillation Column for the Ethyl Tert-Butyl Ether Production

Combined processes in chemical industry are of current increasing interest, since they present a feasible solution for capital and operating expenses optimization. Multiplicities, which are specific to combined processes, cause certain control and design difficulties, thus, parametric sensitivity analysis and defining the range of existing multiplicities are the necessary part of the sustainable design and development of the combined process. Parametric and design optimization is the way of decreasing capital and operational costs. The paper discusses optimization of the design and operating parameters of an ethyl-tert-butyl ether synthesis reactive distillation column. The research aims to perform a parametric sensitivity analysis and explore existent multiple steady states of the process. The areas of multiple steady states are established, the optimal column design is developed, the optimal operation modes are determined. The procedure suggested in the study is universally applicable for design and optimization of other reactive distillation processes.