Chemical Engineering Research & Design最新文献

A deep learning method based on the Convolutional Neural Network (CNN) U-Net architecture has been explored as an image analysis tool for the detection and automated quantification of foam generated in a stirred tank. Multiple datasets of different sizes and with different types of foam were obtained experimentally and processed with data augmentation techniques in order to train the deep learning networks. The impact of adding attention and residual gates to the U-Net model to improve its performance, as well as the size of the dataset used to train the model have been assessed. The main factor influencing the accuracy of the models to detect the foam in the stirred tank is the size and quality of the dataset use to train the U-Net models; it must be sufficiently large and varied in terms of foam type/structure. The addition of attention gates and residual blocks to the U-Net model slightly improves the accuracy of foam detection, particularly for images that contain additional objects or obstructions, however the training time is 30 % longer than the standard U-Net model. In all cases, this image analysis method based on the U-Net model largely out performs conventional image analysis, which was not able to automatically differentiate between foam, additional objects in the tank and bubbles in the liquid. Two methods for the measurement of foam quantity (maximum foam height and foam volume) have also been developed based on the foam regions detected by the models. This is important for the application of the tool, which is to be used to understand the impact of operating conditions on foam formation in lab-/pilot-scale stirred tanks and then develop scale-up guidelines for foam control in industrial tanks.

Combustion with coal is an effective sludge disposal method, but it may weaken the combustion performance of coal. Mixing with biomass can improve the combustion characteristics of the coal-sludge mixture, yet the co-combustion kinetics remain to be studied. The combustion characteristics of sludge, wood chips (WC), and bituminous coal (BC) and their mixtures were investigated by TGA. Mixing 20 wt% sludge reduced the comprehensive combustion characteristic index (S) of BC. However, introducing WC offsets this weakening. The ignition temperature of the mixtures decreased from 430.5 to 274.0 °C, and the S indexes increased, indicating that WC enhances the combustibility of the mixtures. The optimum ratio of WC, sludge and BC blended for combustion was determined experimentally, resulting in the best ignition performance and the largest S index. The kinetic model of BC was represented by the Johnson-Mehl-Avrami (JMA) model. By adding 2–4 wt% WC, the kinetic models of mixtures shifted from JMA to the unimolecular decay law (F1) model and then to the phase boundary-controlled reaction (R3) model at 6 wt% WC content.

Established as a sub-discipline of Chemical Engineering in the 1960s by the late Professor R.W.H. Sargent at Imperial College London, Process Systems Engineering (PSE) has played a significant role in advancing the field, positioning it as a leading engineering discipline in the contemporary technological landscape. Rooted in Applied Mathematics and Computing, PSE aligns with the key components driving advancements in our modern, information-centric era. Sargent’s visionary foresight anticipated the evolution of early computational tools into fundamental elements for future technological and scientific breakthroughs, all while maintaining a central focus on Chemical Engineering. This paper aims to present concise and concrete ideas for propelling PSE into a new era of progress. The objective is twofold: to preserve PSE’s extensive and diverse knowledge base and to reposition it more prominently within modern Chemical Engineering, while also establishing robust connections with other data-driven engineering and applied science domains that play important roles in industrial and technological advancements. Rather than merely reacting to contemporary challenges, this article seeks to proactively create opportunities to lead the future of Chemical Engineering across its vital contributions in education, research, technology transfer, and business creation, fully leveraging its inherent multidisciplinarity and versatile character.

Woven screens are commonly used in filtration processes that involve molten polymer. Predicting the initial pressure drop through these screens is crucial for selecting an appropriate filtration system, as it directly affects the efficiency and cost-effectiveness in industrial processes. Existing computational models for predicting pressure drop through woven screens have not been developed for multi-layered configurations. This study develops an accurate and reliable methodology for using computational fluid dynamics (CFD) to model the pressure drop through a multi-layered screen. The computational model is validated through an experimental set-up.

Selective catalytic oxidation (SCO) of humins is a promising strategy to valorize undesired side streams of biomass conversion processes. Keggin-type polyoxometalates are efficient catalysts for the SCO of humins giving platform chemicals like formic acid, acetic acid or their respective esters up to combined yields of 51 %. Moreover, one of the main challenges for establishing continuous processes is the efficient catalyst recycling and product separation. Herein, we combined an optimization study for the process parameters in SCO with an integrated product separation carried out by using nanofiltration membranes. For this purpose, an enhanced reaction system consisting of 5 vol% methanol addition for CO₂ suppression in combination with 1.5 mmol pTSA as solubility promotor was applied using H₅[PV₂Mo₁₀O₄₀] (HPA-2) as a catalyst achieving a maximum humin conversion of 90 % resulting in 57 % carboxylic ester selectivity. In subsequent studies, an appropriate XN 45 nanofiltration membrane was identified allowing for >99 % catalyst and >90 % additive retention efficiently separating the acidic reaction products in the permeate. Furthermore, long-time stability of the membrane and thus of the separation process could be confirmed up to 168 h time on stream. Based on these results, the developed process represents an important milestone in the valorization of humins combined with efficient separation of the molecular polyoxometalate catalyst and could therefore be the basis for future developments.

Municipal sludge particles with size range between 0.045 and 0.2 mm were difficult to be precisely separated due to fine size distribution and high content of organic matter, resulting in low sand recycling efficiency. In view of the above problems, this paper proposed a flushing water device which was settled at the cone section of conventional hydrocyclone, aiming to utilize the supplementary water to strengthen the shear force field. Comparative test results showed that the organic matter content in the underflow product was significantly reduced after the addition of flushing water. Using the organic matter content and the sand recovery rate as the response index, the prediction models were established by using response surface optimization test. The influence order of different parameters on the response index was apex diameter, flushing device position and flow rate. Finally, an underflow product with organic matter content of 3.96 % and sand recovery rate of 48.03 % was obtained under the conditions of optimized parameter combinations. The research results can contribute to the development of a new method of residual sludge disposal that improves the sand washing efficiency and the cost-effectiveness of sludge treatment.

A novel diglycolamide resin with 2-ethylhexyl terminal alkyl chain functionalised to silica gel was synthesised and characterised by ATR-FTIR, SEM, TG/DSC, & ¹³C (CP/MAS) solid state NMR techniques. Effective functionalization of the organic ligand upon silica gel was confirmed from the two carbonyl stretching frequencies at 1735 cm⁻¹ and 1619 cm⁻¹ in the FTIR spectrum and carbonyl carbon atom signal at 170 ppm in ¹³C (CP/MAS) solid state NMR spectrum. Thermo gravimetric measurement showed 17.3 ± 0.2 % (W/W) weight loss corresponding to the ligand loading value of 425 µmol/g. The resin showed very high uptake of trivalent lanthanide and actinide ions from 4.0 M HNO₃ medium and very low distribution of Cs & Sr; K_d ∼ 1.1×10³ mL/g for Am(III), ∼ 1.41×10³ mL/g for Eu(III), ∼2.5 mL/g for Sr and 1.53 mL/g for Cs. FTIR spectroscopic study of the Eu³⁺ adsorbed resin showed single carbonyl stretching frequency at 1654 cm⁻¹, instead of the two carbonyl stretching frequencies in fresh silica-DGA(2EH) resin indicating complexation through carbonyl oxygen. Very low K_d of ∼1.2–5.1 mL/g for lanthanides/actinides from 4.0 M HNO₃ containing water soluble DGA namely N, N, N, N, Tetraethyldiglycolamide (TEDGA) in the concentration range 0.3–0.1 M indicates 99 % back adsorption feasibility.

Distillation is an energy-intensive separation technique. Semicontinuous distillation has gained attention as a cost-effective alternative to conventional multi-column distillation of multi-component mixtures. However, the cyclic behavior of semicontinuous distillation poses operational challenges considering cost of energy and need for maintaining consistent product purity. The challenge is addressed in this work by proposing a dual-objective Model Predictive Controller (MPC) that handles both product purity and energy consumption through a combined tracking-economic objective function. The proposed MPC features Extended Kalman Filter for state estimation and the successive linearization technique for building the prediction model. The nonlinear plant model is implemented in OpenModelica, which is linked to Matlab for MPC implementation. The effectiveness of the proposed MPC is demonstrated on semicontinuous distillation of Benzene, Toluene, and o-Xylene. The dual-objective MPC is shown to yield an energy saving of 8% per feed processed compared with conventional tracking MPC, while also performing well under process disturbances. It is also shown that the feed processed during a certain time period is 8% higher in the dual-objective MPC than the tracking MPC. The considerable economic improvement is gained without degrading the product purities, indicating that dual-objective MPC is an effective approach to energy-efficient operation of semicontinuous distillation.

This work introduces an online learning-based model predictive control (MPC) approach for the modeling and control of switched nonlinear systems with scheduled mode transitions. Initially, recurrent neural network (RNN) models are constructed offline, utilizing sufficient historical operational data to capture the nominal system dynamics for each mode. Subsequently, we employ real-time process data to develop online learning RNN models, aiming to approximate the dynamics of switched nonlinear systems in the presence of of bounded disturbances. In cases where the initial RNN model is unavailable for a specific switching mode due to very limited historical data, we use real-time data from closed-loop operations under a proportional–integral (PI) controller to build online learning RNN models. To evaluate the predictive performance of online learning RNNs, a theoretical analysis on their generalization error bound is developed using statistical machine learning theory. Additionally, considering the presence or absence of initial RNN models, two MPC schemes are developed. These schemes employ RNNs as prediction models to stabilize switched nonlinear systems, ensuring closed-loop stability by accounting for the generalization error bound derived for online learning RNNs. Finally, the effectiveness of the proposed MPC schemes is demonstrated through a nonlinear process example with two switching modes.

This study explores the utilization of coal fly ash combined with natural clay as a precursor for creating ceramic membranes through the uniaxial compaction method, aiming at lignin separation. These membranes were sintered at varying temperatures from 600 to 1000 °C. Thermal Gravimetric Analysis (TGA) revealed the robust thermal stability of the membranes. Fourier Transform Infrared (FTIR) analysis affirmed the presence of silica and alumina within the membranes. X-ray Diffraction (XRD) peaks indicated the crystalline structure and the presence of metal oxides originating from the fly ash and clay components. Scanning Electron Microscopy (SEM) and Atomic Force Microscopy (AFM) images illustrated the porous nature and rough surface of the ceramic membranes. The average pore radius of the fabricated membrane expanded from 43.12 to 7.91 nm with increasing temperature. Concurrently, membrane flux decreased from 4.13 to 0.96 ×10⁻⁶ m³/m² s with higher temperatures. Lower contact angles indicated the hydrophilic properties of membranes. All membranes exhibited superior corrosion resistance and negative zeta potential, owing to the significant silica content in the fly ash. Lignin separation efficiency increased from 65 % to 81 % with increasing temperatures. The optimal sintering temperature was determined to be 800 °C, achieving a lignin recovery greater than 76 % and a membrane permeability of about 7×10⁻⁹ m³/m² s kPa. Thus, the fly ash-clay-based ultrafiltration membranes synthesized in this study offer the potential for lignin biomass recovery from biorefinery effluent.