Batch Process Research Articles

Practical exercises of computer-aided process synthesis for chemical engineering undergraduates

The study presents ten different exercises covering various computational tools. These exercises are practical applications presented to improve the understanding and skills of students in important concepts of chemical-aided process synthesis. A few exercises aim to build a foundation in computational techniques for chemical engineering undergraduates. The exercises are based on a spreadsheet that covers the design of regression analysis to find the optimum Antoine constants, array calculation for multicomponent distillation material balance, and the generation of a Gantt chart to plan and study the activities of batch processes. The other exercises included an introduction to process simulation, simulation, and reactor rating, and a simulation of multicomponent shortcut distillation. These exercises provide students with hands-on experience in utilizing process simulation software essential for analysing and optimizing chemical processes in real-world scenarios. The exercises also included the design of a heat exchanger network and solving a linear programming problem. An anonymous survey was collected from the cohort that had undergone the exercises, and the practical grades were compared with the batch that did not study the proposed exercises. Additionally, student feedback on practical exercises was collected. Based on the experience of the course coordinator and the collected feedback from participants, it was clear that the exercises helped students to inculcate critical thinking and self-learning abilities. An article essentially sheds light on the computer-aided practical exercises that enable chemical engineering graduates to engage in lifelong learning.

Dynamic controlled variables based dynamic self-optimizing control

Self-optimizing control is a strategy for selecting controlled variables, where the economic objective guides the selection and design of controlled variables, with the expectation that maintaining the controlled variables at constant values can achieve optimization effects, translating the process optimization problem into a process control problem. Currently, self-optimizing control is widely applied to steady-state optimization problems. However, the development of process systems exhibits a trend towards refinement, highlighting the importance of optimizing dynamic processes such as batch processes and grade transitions. This paper formally introduces the self-optimizing control problem for dynamic optimization, termed the dynamic self-optimizing control problem, extending the original definition of self-optimizing control. A novel concept, ”dynamic controlled variables” (DCVs), is proposed, and an implicit control policy is presented based on this concept. The paper theoretically analyzes the advantages and generality of DCVs compared to explicit control strategies and elucidates the relationship between DCVs and traditional controllers. Moreover, this paper puts forth a data-driven approach to designing self-optimizing DCVs, which considers DCV design as a mapping identification problem and employs deep neural networks to parameterize the variables. Three case studies validate the efficacy and superiority of DCVs in approximating multi-valued and discontinuous functions, as well as their application to dynamic optimization problems with non-fixed horizons, which traditional self-optimizing control methods are unable to address.

Application of smart overpressure protection devices to batch and semi-batch processes: The smart relief mode

Smart overpressure protection devices (SmOPs) combine complex mathematical models implemented to safety-related programmable logic controllers (sPLC) with adaptable mechanical safety devices. The devices dynamically adjust set points and relief areas based on real-time hazard potential assessments and are being used to protect chemical reactors against exothermal runaway reactions. Providing two modes of operation, the "zero-emission mode" prevents reactions from exceeding reactor limits, while the "smartrelief mode" initiates controlled fluid discharge to avoid exceeding the permissible plant limits. In this article, the “smart relief mode” of smart overpressure protection devices as an innovative safety solution for chemical reactors is introduced. It is applied to two different types of reaction and shows three different typical industrial use cases.

A decentralized control strategy for improved batch tablet coating process operation

This paper proposes a control strategy for batch tablet film-coating processes using readily available process instrumentation. Outlet air conditions (temperature and relative humidity), which impact coating quality, are controlled by the manipulation of inlet air temperature and coating spray rate. Controller design, tuning and validation rely on the use of a dynamic process simulator, calibrated using data collected from a pilot plant coating unit. The final 2 × 2 decentralized PI controller demonstrated good performance for both setpoint tracking and disturbance rejection in simulation. Additionally, controlled variable setpoints were determined to maximize plant profits considering batch duration and energy consumption as optimization criteria. For the case study considered, the total batch processing time can potentially be reduced by 40 % while simultaneously reducing energy usage by 30 %, in comparison with the more conservative fixed recipe approach typically adopted in production.

A Selective Migration-Based Improved GPR Modeling Method for Batch Process

A Selective Migration-Based Improved GPR Modeling Method for Batch Process

Quantitative Analysis of CGRP-IR Afferent Axons in the Mouse Stomach Using Zeiss Arivis Vision4D for Automated Tracing

Gastric nociceptive afferents detect and respond to noxious stimuli by transmitting signals from the stomach to the brain. Calcitonin gene-related peptide (CGRP) is a marker for nociceptive afferents and is crucial to gastric functions, including gastric acid secretion, gut motility, and inflammation. However, the topographical quantitative analysis of CGRP-immunoreactive (CGRP-IR) axons in the whole stomach is challenging due to the large amount of axon innervation in various targets within the tissue. Manual tracing of axons is time-consuming, labor-intensive, and extremely tedious, while currently existing automatic neuron tracing programs often need help with large file sizes or tracing inaccuracies. In this study, we used Zeiss Arivis Vision4D, a novel and sophisticated software with a user-guided machine learning capability, for automated tracing of CGRP-IR axons in flat-mounts of the whole mouse stomach muscular layers and thick serial sections of the antrum-pylorus region (C57BL/6J, male, 2-3 months old, n=7). We utilized a combination of techniques, including the Zeiss M2 Imager for automated scanning, Zen 3.3 (blue edition) for image processing, and Vision4D for stitching the entire stomach flat-mounts and sections, 3D fiber tracing, and density analysis. Our results demonstrated that: 1) Vision4D detected and traced almost all CGRP-IR axons of different gastric targets (blood vessels, muscles, myenteric neurons) in different regions (fundus, corpus, antrum) of the stomach. 2) The x-y-z tracing allowed the visualization of axons in space, portraying its continuous trajectory across layers in the muscular flat mounts (longitudinal muscle, myenteric plexus, circular muscle) and cross sections (muscles, submucosa, mucosa). 3) Vision4D also enabled quantitative morphological and laminar analysis based on target and layer separations. The axon density heatmap generated from the tracing data showed high CGRP-IR axon innervation in the blood vessels. After the removal of blood vessels, dense innervation was seen in the upper fundus and antrum-pylorus followed by corpus, indicating the abundance of CGRP in these regions. 4) Vision4D is a powerful tool in handling large data sizes (hundreds of gigabytes of optical sections in more than 300 image frames per stomach) and outstanding in its capability to track nerve fibers in a highly accurate and time effcient manner. This has paved the way for data batch processing to be accessible, as well as shown that multi-marker tracing of the same tissue sample is attainable. 5) The tracing data will be integrated into a stomach scaffold for a 3D map of CGRP-IR afferent axons. This study represents, for the first time, the automatically traced CGRP-IR axons using whole stomach preparations and highlights the potential of Vision4D in creating axon tracing models in other organs and species. This study is supported by NIH 1 U01 NS113867-01, NIH R15 1R15HL137143-01A1 and Carl ZEISS Company. This is the full abstract presented at the American Physiology Summit 2024 meeting and is only available in HTML format. There are no additional versions or additional content available for this abstract. Physiology was not involved in the peer review process.

GENTI: GPU-Powered Walk-Based Subgraph Extraction for Scalable Representation Learning on Dynamic Graphs

Graph representation learning is an emerging task for effectively embedding graph-structured data with learned features. Among them, Subgraph-based GRL (SGRL) methods have demonstrated better scalability and expressiveness for large-scale tasks. The core challenge of applying SGRL to dynamic graphs lies in accommodating the extraction of subgraphs to evolving data with efficient computation. To address the efficiency bottleneck, we propose GENTI, a GPU-oriented SGRL algorithm for dynamic graphs. Our approach mainly improves the critical subgraph extraction stage by disentangling it into two phases, namely neighbor sampling and subgraph gathering, which are respectively performed on CPU and GPU in an asynchronous fashion. The design favorably eliminates the dependence of feature learning on subgraph extraction, and is capable of exploiting the GPU's batch processing ability to remarkably boost computations throughout the pipeline. Dedicated data structures are designed for efficiently managing the dynamic graph storage and conforming efficient subgraph operations. Extensive empirical results on various real-world dynamic graphs show that GENTI achieves up to 30× faster in subgraph extraction time than the state-of-the-art walk-based methods and up to 26× acceleration in overall learning time, while maintaining comparable prediction performance. In particular, it is able to complete learning on the largest available graph of 1.3 billion edges within 24 hours, while all other baselines exhibit prohibitive overhead.

Meta-learning of feature distribution alignment for enhanced feature sharing

Meta-learning aims to learn common feature representations between tasks to quickly adapt and perform well when facing new tasks. However, current meta-learning algorithms ignore the differences in feature distributions among tasks. The significant differences in task features may lead to an over-fitting of the model to specific features of certain tasks during the learning process, thereby causing a decrease in adaptability to other tasks. In this paper, we propose a novel method for feature distribution alignment. Specifically, we employ knowledge distillation (KD) to minimize the Kullback–Leibler (KL) divergence between task features, thus promoting greater similarity in the extracted feature distributions. This incentivizes the model to prioritize learning shared parameters and general features across tasks during the training process, rather than task-specific features. Furthermore, we also use an adaptive dropout strategy, which adjusts the training process for the next batch based on the feature distribution of the current task. This method reduces the impact of feature distribution differences during the training iterations. Extensive experiments and analyses demonstrate that our method improves the performance of various meta-learning models in few-shot classification, few-shot fine-grained classification, cross-domain few-shot classification, and non-mutually exclusive few-shot classification.

An adaptive strategy for time‐varying batch process fault prediction based on stochastic configuration network

AbstractFault prediction ensures safe and stable production, and cuts maintenance costs. Due to the changing operating conditions that lead to the changes in the characteristics of industrial processes, there is a need to monitor the fault state of batch processes in real‐time and to accurately predict fault trends. An adaptive slow feature analysis‐neighborhood preserving embedding‐improved stochastic configuration network (SFA‐NPE‐ISCN) algorithm for batch process fault prediction is proposed. Firstly, SFA is used to extract the time‐varying features of process data and establish the update index of the NPE model. Then, to extract local nearest‐neighbor features and reconstruct them by the NPE model with adaptive update capability, square prediction error (SPE) statistics are constructed as fault state features based on the reconstructed error. Further, the hunter‐prey optimization (HPO) algorithm optimizes the weights and biases in the stochastic configuration network, and the singular value decomposition (SVD) and QR decomposition of column rotation are introduced to solve the ill‐posed problem of SCN and obtain the prediction model of ISCN. Finally, the obtained statistics SPE is formed into a time series, and the ISCN model is used to predict the process state trend. The effectiveness of the proposed algorithm is verified by case studies of industrial‐scale penicillin fermentation processes and the Hot strip mill process.

Continuous flow synthesis of hierarchical low silica X zeolite

Zeolites, renowned for their versatile applications in catalysis, adsorption, and ion exchange, have long been synthesized using conventional batch processes. However, the inherent limitations of these methods, such as resource-intensive conditions and inconsistent product quality, underscore the need for a sustainable and efficient approach. In this study, a continuous flow synthesis process was established for the synthesis of industrially important low silica X (LSX) zeolite using a tubular reactor. The synthesis gel was subjected to aging for 5 days at room temperature to facilitate nucleation and crystal growth combined with the fast-heating rate in a tubular reactor at 363 K & 1.1 atm., which in turn produces LSX after 40 min. The synthesized product was confirmed by the XRD, FE-SEM, EDS, XRF, TEM, and N2 adsorption-desorption; the data was compared with the LSX sample synthesized by batch process. The result implies that LSX prepared by continuous flow has a pure phase of LSX with the hierarchical structure, which provides better adsorption capacity of CO2 at 298 K up to 20 bar. Due to continuous flow synthesis, the crystallization time was reduced and faster kinetics which may be helpful for scale-up the process for LSX synthesis.

Real-Time Monitoring of Multistep Batch and Flow Synthesis Processes of a Key Intermediate of Lifitegrast by a Combined Spectrometer System with Both Near-Infrared and Raman Spectroscopies Coupled to Partial Least-Squares.

Process analytical technology (PAT) has been successfully applied in numerous chemical synthesis cases and is an important tool in pharmaceutical process research and development. PAT brings new methods and opportunities for the real-time monitoring of chemical processes. In multistep synthesis, real-time monitoring of the complex reaction mixtures is a significant challenge but provides an opportunity to enhance reaction understanding and control. In this study, a combined multichannel spectrometer system with both near-infrared and Raman spectroscopy was built, and calibration models were developed to quantify the desired products, intermediates, and impurities in real-time at multiple points along the synthetic pathway. The capabilities of the system have been demonstrated by operating dynamic experiments in both batch and continuous-flow processes. It represents a significant step forward in data-driven, multistep pharmaceutical ingredient synthesis.

The Performance of Iron Nanocomposites as Persulfate Activators in Permeable Reactive Barrier Technology for Treatment of Landfill Leachate

Landfill leachate contains a large organic load and a wide spectrum of pollutants, and it needs appropriate treatment to reduce the severity of these contaminants. Therefore, the present study aimed to remove chemical oxygen demand (COD) and ammonia (NH3 ) as target contaminants from the landfill leachate. To this end, the zero-valent iron nanofibers/reduced ultra-large graphene oxide (ZVINFs/rULGO) and converter sludge (CS) were used as persulfate (PS) activators in permeable reactive barrier (PRB) technology for removing COD and NH3 . The Box–Behnken design (BBD) was employed to determine the possibility of using each activator and optimize three parameters: activator, pore volume (PV), and flow (mL/min) using target contaminants. In the Batch process, the effects of pH, activator concentration, and PS/COD mass ratio on COD removal were examined. The BBD experiments provided a satisfactory predictive model, indicating that ZVINFs/rULGO is superior to CS when used as activators of PS in PRB technology. The biochemical oxygen demand (BOD5 )/COD ratio increased from 0.24 in raw leachate to 0.67 in treated leachate, and leachate toxicity was reduced by more than 85%. Additionally, the ZVINFs/rULGO and CS contributed to the activation of PS.

Novel Multi-flow Multi-scale Convolutional Neural Network Developed for Quality Prediction of Batch Processes to Fuse Data With Different Sampling Frequencies

Novel Multi-flow Multi-scale Convolutional Neural Network Developed for Quality Prediction of Batch Processes to Fuse Data With Different Sampling Frequencies

From Batch to Continuous Small-Scale Production of Particles: Mixer Design Methodology for Robust Operation

Layered double hydroxides (LDHs) are a vital tool in many different areas, such as drug delivery, catalysis, anion exchange (materials), polymer processing, etc. Conventionally, LDHs are synthesized in a batch process that consists of particle generation and ripening, where product properties are manipulated for stability and the optimal uptake of genetic material. Continuous processing and intensive mixing holds high promise for improved particle generation and characteristic control. In this contribution, an iterative method, using the mentioned particle generation as a use case, was applied to quickly generate a continuous process optimization platform for continuous, plugging-free particle generation with the required characteristics. Assisted by rapid prototyping and additive manufacturing, a vortex mixer was produced that delivers satisfactory long-term results.

Robust predictive minmax fault‐tolerant control for multi‐phase batch processes under partial actuator faults and asynchronous switching

AbstractIn view of the uncertainties, partial actuator faults, asynchronous switching as well as unknown external interference in multi‐phase batch processes, a robust predictive minmax fault‐tolerant control approach is presented. First, considering that a controller and a system state cannot switch synchronously during switching, a match‐and‐mismatch switching model is constructed. Then, an output tracking error is introduced to establish an equivalent extended switching model. In this extended model, the robust control issue is transformed into the minmax optimization issue by establishing the minmax performance index. On this basis, the control laws with control input and interference input are designed. Second, sufficient conditions are given to assure the system separately has asymptotic and exponential stabilities in each phase and batch. By calculating the given sufficient conditions online, the gains of the control laws, and the minimum and maximum operating time under match and mismatch cases are obtained. With the maximum operating time, the controller receives the switching signal beforehand to eliminate the mismatch case of switching. Additionally, a robust criterion is supplied to confirm the system's robustness. Finally, a simulation study of an injection molding process shows that the approach is valid and feasible.

Apparent metabolizable energy and ileal amino acid digestibility of commercial soybean meals of different origins in broilers

We studied the chemical composition and the in vivo AMEn content and apparent (AID) and standardized (SID) ileal digestibility of CP and amino acids (AA) of 27 samples of soybean meals (SBM) from Argentina (ARG), Brazil (BRA), and USA, collected in Spain. On 88% DM basis, the BRA meals had more CP (46.9 vs. 46.0 and 45.9%; P < 0.05) and less sucrose (5.21 vs. 6.28 and 6.47%; P < 0.001) and stachyose (4.20 vs. 4.66 and 4.78%; P < 0.05) than the USA and ARG meals. Urease activity, protein dispersibility index, KOH protein solubility, and trypsin inhibitor activity values, were higher for the USA meals than for the South American meals (P < 0.05). In the in vivo trial, broilers received a common crumble diet from 0 to 16 d of age and then, their respective experimental diets (53% of a N-free diet and 47% of each of the 27 SBM tested) in mash form, for 5 d. The AMEn (2,334 vs. 2,282 and 2,277 kcal/kg; P = 0.062) and the AID (87.3 vs. 86.7 and 86.4%; P = 0.054) and SID (91.9 vs. 91.2 and 90.8%; P < 0.05) of the protein, were greater for the USA meals than for the ARG and BRA meals. In fact, the SID of Lys (93.5 vs. 93.0 and 92.1%; P < 0.001) and of the sum of Lys, Met, Thr, Trp, and Cys (91.4 vs. 91.0 and 90.2%; P < 0.05) were greater for the USA meals than for the ARG and BRA meals. In summary, the chemical composition, protein quality indicators, AMEn content, and ileal digestibility of the CP and the AA of the SBM, varied with the country of origin of the soybeans. In order to increase the accuracy of the feed formulation process, the composition and nutrient content of commercial batches of SBM, by country of origin should be controlled and periodically updated.

An integrated tube robust iterative learning model predictive control strategy based on dynamic partial least squares algorithm for batch processes

AbstractIn this paper, an integrated tube robust iterative learning model predictive control (Tube‐RILMPC) strategy based on the dynamic partial least squares (DyPLS) identified algorithm is proposed. The problems of large amount of online data calculation and input and output variable dimensions disaster in the original variable space are solved. This integrated Tube‐RILMPC strategy enhances the tracking performance and robustness of two‐dimensional (2D) control system. The output trajectories of system are located in the tube the nominal trajectory, which reduces the modelling error caused by the uncertain model. Based on the worst‐case performance index of ellipsoidal uncertainty and polytopic uncertainty, a robust iterative learning control (ILC) strategy is designed. Finally, the superiority of the proposed control algorithm is verified by comparative simulation.

Improving performance of aerosol jet printing using machine learning‐driven optimization

AbstractAerosol jet printing is a promising technology for printing functional materials on a variety of substrates with high precision and resolution. This technology has the potential to revolutionize the manufacturing industry by providing a low‐cost, high‐resolution printing technique that can be used to produce additively printed electronics, sensors, and energy devices. However, the optimization of this process has traditionally relied on time‐consuming trial‐and‐error methods, hampering its efficiency and scalability. Machine learning (ML) models have the potential to overcome these challenges and improve the quality, speed, and efficiency of the printing process. In this paper, we propose an approach that leverages ML algorithms to streamline and enhance the aerosol jet printing optimization process. Our methodology involves data collection through systematic experimentation with various parameter settings. This data set serves as the foundation for training different ML model capable of predicting printed line characteristics and optimal printing process parameter. We validate our approach by performing experiments on different inks, and we compare the results of our ML‐based optimization approach to those obtained using traditional trial‐and‐error methods. The results demonstrate that our approach offers significantly higher accuracy and efficiency. To enhance our approach's accessibility and ease of use, we incorporate AutoML techniques which automates the process of selecting the most suitable ML algorithms and hyperparameters, reducing the burden of manual configuration. Furthermore, we introduce a user‐friendly web‐based interface that facilitates the entire ML pipeline, from data preprocessing to prediction and batch processing. This interface empowers users to efficiently manage and manipulate their data, select appropriate ML algorithms, and execute predictions, ultimately improving accuracy and model performance.

Kinetics, equilibrium thermodynamics and machine learning modeling of the adsorption of acid dyes by corncob derived magnetic modified biomass

The removal of Acid dyes (Acid Orange 52, AO-52 and Acid Red 87, AR-87) from aqueous solutions in a batch process using Magnetic modified corncob biomass (Corncob@Fe3O4) is reported. The effects of factors such as dyes concentrations, contact time, adsorbent dosage and temperature were investigated. The kinetics and equilibrium data were modeled appropriately while Response Surface Methodology (RSM), Artificial Neural Network (ANN) as well as multiple linear regression (MLR) based on Box–Behnkem design of experiment were used to model the combined effect of the factors. The thermodynamics parameters were obtained from van’t Hoff expression. Although all the kinetic models fitted the data, first order kinetic model best fitted the data with highest R2 and lowest Reduced Chi square (χ2). The monolayer adsorption capacities from Langmuir model are 47.85 and 22.48 mg/g for AO-52 and AR-87 respectively. The thermodynamic parameters obtained showed that the process is spontaneously feasible and entropy driven. Statistical evaluation showed that the process is well modeled in the order of ANN > RSM > MLR for AO-52 and MLR > ANN > RSM for AO-87. The study showed that magnetic modified corncob biomass is effective and competed favorably with other adsorbents for the adsorption of AO-52 and AO-87 dyes from aqueous solution.

A Study on the Valorization of Rice Straw into Different Value-Added Products and Biofuels

This work depicts that rice straw (RS), which is one of the major lignocellulosic wastes all over the world and causing many environmental problems, has considerable amounts of protein, ash, macronutrients, and micronutrients of approximately 11.38%, 16.77%, 2.27 mg/kg, and 771.9 mg/kg, respectively; besides, a C/N ratio of 15.18, a total N, P2O5, and K2O content of 1.85%, and a considerably low concentration of undesirable heavy metals and silica of approximately 77.69 mg/kg and 109 mg/kg are also present, which recommends its applicability as a precursor feedstock for the production of organic fertilizer and animal fodder. The batch solid-state fermentation (SSF) of RS by Trichoderma longibrachiatum DSMZ 16517 produced considerable amount of total reducing sugars (TRS) of approximately 339.2 mg TRS/g RS under the optimum operatic conditions of 20% (w:v) substrate concentration, pH 7, 1% inoculum size, a 9-day incubation period, and 30°C incubation temperature. The readily available and cost-effective agroindustrial waste, sugarcane molasses, proved to enhance the fungal biomass growth and (hemi) cellulolytic enzymes activities. The inoculated RS-SSF batch process with T. longibrachiatum precultured on 10% molasses enhanced the (hemi) cellulolytic enzymatic activities and TRS production rate by approximately 5.82 and 3.8 folds, respectively, relative to that inoculated by T. longibrachiatum precultured in the conventional potato dextrose broth medium. The separate hydrolysis and fermentation processes by different yeast strains Candida tropicalis DSM 70156, C. shehatae ATCC 58779, and Saccharomyces cerevisiae ATCC 64712 revealed an efficient bioethanol yield and productivity that ranged between 0.36 and 0.38 g/g sugars and 0.22 and 0.23 g/L/h, respectively, with concomitant competent fermentation efficiencies that ranged between 48.35% and 51.25%. The proximate analysis of rice straw before and after fungal hydrolysis proved calorific values of approximately 15.8 MJ/kg and 16.05 MJ/kg, respectively, recommending their applicability as primary and secondary solid biofuels. Thus, this study proved the waste prosperity of RS for environmental opulence and sustainability.