S.A. GokulaKrishnan , S. Srithar , R. Surya Prakash , A. Infant Ebinesh , T. Raguram , B. Sasikumar , Muhammed Afnas , G. Arthanareeswaran , Mangalaraja Ramalinga Viswanathan
{"title":"Influence of covalent organic framework integrated hybrid PEI membrane on removal of divalent salts-fulvic acid residues","authors":"S.A. GokulaKrishnan , S. Srithar , R. Surya Prakash , A. Infant Ebinesh , T. Raguram , B. Sasikumar , Muhammed Afnas , G. Arthanareeswaran , Mangalaraja Ramalinga Viswanathan","doi":"10.1016/j.jtice.2025.106416","DOIUrl":"10.1016/j.jtice.2025.106416","url":null,"abstract":"<div><h3>Background</h3><div>Natural organic matter (NOMs) can form complexes with divalent salts, increasing their bioaccumulation in aquatic systems and posing health risks to aquatic organisms. This study explores the use of triazine-structured Covalent Organic Frameworks (COFs), specifically NENP-1, incorporated into polyetherimide (PEI) membranes to enhance the removal of fulvic acid (FA) with Mg²⁺ and Ca²⁺ from saline wastewater.</div></div><div><h3>Methods</h3><div>Flat sheet membranes (M2-M5) with varying COF concentrations (0.25 %, 0.5 %, 0.75 %, and 1 %) were fabricated by the phase inversion method. The presence of characteristics peaks at 3486 cm<sup>-1</sup> (secondary amine), 1495 - 1349cm<sup>-1</sup> (triazine structure) functional group in the was confirmed, which confirmed the NENP-1 successful embedding in the matrix of PEI polymer. FESEM analysis revealed that the 0.25 wt.% membrane had uniform nanoparticle dispersion and a thin selective layer, while higher COF content (1 wt.%) caused aggregation and dense layer formation.</div></div><div><h3>Significant Findings</h3><div>Among the membranes (M1-M5), the M2 (0.25 wt.%) COF membrane showed the best performance, achieving solute fluxes of 180 and 190 L·m⁻²·h⁻¹ for FA+MgCl₂ and FA+CaCl₂, with 72 % and 79 % rejection, respectively. Further, the M2 (0.25 wt.%) membrane also demonstrated good antifouling properties, with a flux recovery ratio of 78.5 %. Overall, NENP-1/PEI membranes present a viable solution for removing NOMs and divalent salts from contaminated water sources.</div></div>","PeriodicalId":381,"journal":{"name":"Journal of the Taiwan Institute of Chemical Engineers","volume":"179 ","pages":"Article 106416"},"PeriodicalIF":6.3,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145218107","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Rui Huang, Yuhan Sun, Xiaoqing Xu, Renjie Ji, Xiaoyu Yan, Jiale Xu, Chunyi Li
{"title":"Extending reactor lining service life through thiophene treatment of corroded 304 stainless steel","authors":"Rui Huang, Yuhan Sun, Xiaoqing Xu, Renjie Ji, Xiaoyu Yan, Jiale Xu, Chunyi Li","doi":"10.1016/j.jtice.2025.106430","DOIUrl":"10.1016/j.jtice.2025.106430","url":null,"abstract":"<div><h3>Background</h3><div>The lining was a critical component of chemical processing equipment, whose structural stability directly determined production efficiency and product quality. The lining structural damage not only significantly increased the maintenance costs of the equipment but also caused serious influence on the enterprise’s economic benefits. Therefore, research on lining failure mechanisms and protective technology development had the important value of engineering application.</div></div><div><h3>Methods</h3><div>The lining precursor was prepared using a corundum-aluminum dihydrogen phosphate system, with embedded 304 stainless steel that was calcined to obtain the composite lining. The stainless steel which was extracted from the prepared lining evaluated in a C<sub>3</sub>H<sub>6</sub>/C<sub>3</sub>H<sub>8</sub>/H<sub>2</sub> mixed atmosphere. After gas-phase thiophene treatment of the prepared lining, the extracted stainless steel sample was tested under identical conditions for comparison. A systematic analysis on corroded zones of the stainless steel surface was conducted using XRD, SEM, XPS, and electrochemical measurements.</div></div><div><h3>Significant Findings</h3><div>The passive film of stainless steel was corroded during lining fabrication, leading to Fe<sub>3</sub>O<sub>4</sub> formation on the stainless steel surface, which could catalyze hydrocarbon gases converted to filamentous carbon at high temperatures. The introduction of thiophene could poison Fe<sub>3</sub>O<sub>4</sub>, reduce its catalytic activity, prevent the formation of filamentous carbon, and thereby extend the service life of the lining.</div></div>","PeriodicalId":381,"journal":{"name":"Journal of the Taiwan Institute of Chemical Engineers","volume":"179 ","pages":"Article 106430"},"PeriodicalIF":6.3,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145218110","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Divya Dagar , Yu-Fon Chen , Bhim P. Kafle , Shanmuganathan Venkatesan , Yuh-Lang Lee
{"title":"Mechanistic elucidation of novel carbon quantum dots with high selectivity and sensitivity for picric acid detection","authors":"Divya Dagar , Yu-Fon Chen , Bhim P. Kafle , Shanmuganathan Venkatesan , Yuh-Lang Lee","doi":"10.1016/j.jtice.2025.106432","DOIUrl":"10.1016/j.jtice.2025.106432","url":null,"abstract":"<div><h3>Background</h3><div>Carbon quantum dots (CQDs) have evolved as promising alternatives to traditional semiconductor quantum dots due to their high photoluminescence, stability, tunable emission wavelengths, and diverse applications.</div></div><div><h3>Methods</h3><div>In this study, a novel precursor system composed of terephthalic acid and m-phenylenediamine was developed to synthesize CQDs. The as-prepared CQDs were characterized and applied for picric acid (PA) detection in various environmental samples. Moreover, the underlying sensing mechanism was thoroughly elucidated.</div></div><div><h3>Significant findings</h3><div>The CQDs exhibited excitation-independent green emission at 509 nm with a high quantum yield of 56.3 %. Characterizations confirmed that the green-emissive CQDs (G-CQDs) possessed a graphitic carbon framework and abundant surface functional groups, which contributed to excellent colloidal stability and robust photoluminescence across a wide range of pH values, ionic strengths, storage durations, and under prolonged UV irradiation. For PA sensing, the G-CQDs demonstrated high selectivity and sensitivity, achieving a low detection limit of 17 nM. This sensing capability was not affected by the presence of various cations and anions, and was successfully validated in real samples including sea, river, and tap waters. Mechanistic investigations revealed that the sensing process was primarily governed by the Inner Filter Effect (IFE), rather than by dynamic or static quenching pathways.</div></div>","PeriodicalId":381,"journal":{"name":"Journal of the Taiwan Institute of Chemical Engineers","volume":"179 ","pages":"Article 106432"},"PeriodicalIF":6.3,"publicationDate":"2025-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145218148","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Rajapriya Govindaraju, Jose Paul, Jihyeon Kim, Jongsung Kim
{"title":"Iron and nitrogen co-doped MXene quantum dots: A novel dual-mode electrochemical and fluorescence sensor for dopamine detection","authors":"Rajapriya Govindaraju, Jose Paul, Jihyeon Kim, Jongsung Kim","doi":"10.1016/j.jtice.2025.106421","DOIUrl":"10.1016/j.jtice.2025.106421","url":null,"abstract":"<div><h3>Background</h3><div>MXene quantum dots (MQDs) possess excellent conductivity, abundant active sites, and unique optical properties, yet their metal/heteroatom-modified sensing applications remain underexplored. Here, we report Fe- and N-co-doped MXene quantum dots (FeOMQDs) for the detection of dopamine (DA).</div></div><div><h3>Methods</h3><div>FeOMQDs were synthesized via a one-step hydrothermal process using MXene sheets, FeCl₃, and o-phenylenediamine, and characterized by FTIR, TEM, and XPS. Their fluorescence and electrochemical properties were exploited for DA sensing.</div></div><div><h3>Significant Findings</h3><div>The fluorescence sensing mechanism is governed by DA-induced quenching of FeOMQDs photoluminescence through metal–analyte complexation. The platform exhibited a linear range of 1–10 nM with a detection limit of 0.56 nM. For electrochemical analysis, FeOMQDs were immobilized on glassy carbon electrodes (GCE) and screen-printed electrodes (SPE), followed by evaluation via cyclic voltammetry (CV) and differential pulse voltammetry (DPV). Enhanced electrocatalytic activity toward DA oxidation was observed, attributed to the synergistic interactions between Fe centers and the MXene matrix. The FeOMQDs–GCE demonstrated a linear range of 2–20 nM with detection limits of 1.8 nM (CV) and 1.09 nM (DPV), along with an extended linear range of 40–1000 nM in CV studies. The probes showed excellent reproducibility, stability, and selectivity, with reliable recovery in human serum, confirming their promise for practical bioanalytical applications.</div></div>","PeriodicalId":381,"journal":{"name":"Journal of the Taiwan Institute of Chemical Engineers","volume":"179 ","pages":"Article 106421"},"PeriodicalIF":6.3,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145156787","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Sifei An , Qiang Wang , Haini Yang , Daquan Zhang , N.N. Andreev
{"title":"Preparation of Schiff base derivates from leucine and its inhibition effect on the atmospheric corrosion of mild steel","authors":"Sifei An , Qiang Wang , Haini Yang , Daquan Zhang , N.N. Andreev","doi":"10.1016/j.jtice.2025.106429","DOIUrl":"10.1016/j.jtice.2025.106429","url":null,"abstract":"<div><h3>Background</h3><div>Schiff bases are attractive as eco-friendly corrosion inhibitors for the protection of metals. In this study, a novel volatile corrosion inhibitor (VCI) based on Schiff base (LEU-T) is synthesized from leucine and salicylaldehyde. This study evaluates the volatile corrosion inhibition performance of LEU-T on mild steel and examines its anti-corrosion effect in a simulated atmospheric corrosion solution.</div></div><div><h3>Methods</h3><div>The corrosion inhibitor is characterized by Fourier transform infrared (FT-IR) absorption spectra. The corrosion behavior, inhibition efficiency, and anti-corrosion mechanism of the inhibitor on mild steel are evaluated using volatile inhibition ability (VIA) tests, electrochemical measurements (including potentiodynamic polarization (PDP) and electrochemical impedance spectroscopy (EIS)), and theoretical calculations. Scanning electron microscopy (SEM) and atomic force microscope (AFM) confirm the formation of the inhibitive films on mild steel.</div></div><div><h3>Significant Findings</h3><div>The results demonstrate that LEU-T exhibits outstanding inhibition efficiency of 82.18 % in VIA tests. Electrochemical measurements indicate that the formed inhibitive films on the steel substrate lead to an increase in the charge transfer resistance. The corrosion inhibition efficiencies of leucine and LEU-T, as determined by electrochemical impedance spectroscopy, are 67.84 % and 83.90 %, respectively. SEM and AFM observations indicate that the inhibitor forms a protective VCI coating, effectively retarding the steel corrosion. The LEU-T has larger adsorption energy (-136.13 kJ/mol) compared to leucine (-88.48 kJ/mol) on mild steel.</div></div>","PeriodicalId":381,"journal":{"name":"Journal of the Taiwan Institute of Chemical Engineers","volume":"179 ","pages":"Article 106429"},"PeriodicalIF":6.3,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145156789","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yeshuang Guo, Hai Wang, Long Li, Xinlong Sun, Shaojun Li
{"title":"Real-time optimization of continuous catalytic reforming using transfer learning and reinforcement learning for feedstock variability","authors":"Yeshuang Guo, Hai Wang, Long Li, Xinlong Sun, Shaojun Li","doi":"10.1016/j.jtice.2025.106419","DOIUrl":"10.1016/j.jtice.2025.106419","url":null,"abstract":"<div><h3>Background</h3><div>Real-time optimization (RTO) of continuous catalytic reforming (CCR) is hindered by model mismatch caused by feedstock property variations, which degrade the performance of existing optimization strategies.</div></div><div><h3>Methods</h3><div>To address this challenge, this study proposes an RTO method integrating transfer learning (TL) and reinforcement learning (RL). A high-precision surrogate model of the CCR process is constructed as the agent's interactive environment. A TD3 agent is designed based on optimization objectives, incorporating dropout layers during training to enhance robustness. The trained critic network serves as a monitor, detecting feedstock changes via absolute temporal difference error.</div></div><div><h3>Findings</h3><div>TL enables the agent to adapt efficiently to varying feedstock properties. Experimental results demonstrate that the proposed method rapidly identifies optimal operating conditions, with the monitor effectively detecting feedstock changes. Across multiple test cases, the TL strategy reduces average training time to one-tenth of training from scratch, without compromising performance, demonstrating its effectiveness and potential for RTO applications.</div></div>","PeriodicalId":381,"journal":{"name":"Journal of the Taiwan Institute of Chemical Engineers","volume":"179 ","pages":"Article 106419"},"PeriodicalIF":6.3,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145156788","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yu Sha , Yong Wang , Jinming Zhang , Tianpeng Chen , Qingguo Liu , Pengpeng Yang , Hanjie Ying , Yong Chen , Wei Zhuang , Wenjun Sun
{"title":"Facile co-deposition of folic acid and polydopamine: Regulation of surface roughness on yeast adhesion in the continuous fermentation","authors":"Yu Sha , Yong Wang , Jinming Zhang , Tianpeng Chen , Qingguo Liu , Pengpeng Yang , Hanjie Ying , Yong Chen , Wei Zhuang , Wenjun Sun","doi":"10.1016/j.jtice.2025.106425","DOIUrl":"10.1016/j.jtice.2025.106425","url":null,"abstract":"<div><h3>Background</h3><div>Adhesion is the basis of the cell immobilization, with the surface roughness being a critical factor in promoting the cell initial adhesion and biofilm formation. This study presented a versatile co-deposition approach that combined folic acid (FA) and dopamine (DA) for generating the rough surface.</div></div><div><h3>Methods</h3><div>Inspired by DA self-polymerization, a rough hierarchical structure incorporated with FA was constructed on the surface of cotton fibers (CF) through π-π interactions and hydrogen bonds (DA+FA-CF). Various modification parameters were investigated to explore the effect of the surface roughness on cell adhesion, biofilm formation, and continuous fermentation of <em>Saccharomyces cerevisiae</em>.</div></div><div><h3>Significant Findings</h3><div>The optimal FA concentration, modification time and deposition time (0.5 g/L, 6 h and 8 h) yielded the highest surface roughness, which significantly increased the immobilization efficiency from 54.92 % to 74.91 %, and the biomass from 5.95 g/L to 8.70 g/L, respectively. Moreover, DA+FA-CF reduced the fermentation period by 66.67 % and achieved an ethanol yield of 39.67 %, with a peak production of 23.75 g/L in the fifth-batch of continuous fermentation. This strategy offered a facile and biocompatible approach to improving continuous fermentation performance through surface roughness-regulated cell immobilization.</div></div>","PeriodicalId":381,"journal":{"name":"Journal of the Taiwan Institute of Chemical Engineers","volume":"179 ","pages":"Article 106425"},"PeriodicalIF":6.3,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145156790","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Cost-effective large-scale production of Chlorella vulgaris-based biodiesel and bioethanol: A comparative assessment","authors":"Bavatharny Thevarajah , P.H.V. Nimarshana , Jo-Shu Chang , Thilini U. Ariyadasa","doi":"10.1016/j.jtice.2025.106395","DOIUrl":"10.1016/j.jtice.2025.106395","url":null,"abstract":"<div><h3>Background</h3><div>Microalgae-based biofuel production is a promising alternative to conventional renewable energy sources, despite its economic viability at industrial scale remains unrealized. The present study is the first to present a novel evaluation of large-scale biodiesel and bioethanol production from <em>Chlorella vulgaris</em>, integrating the cost analysis of chemicals and energy demands, to identify the most viable production route for real-world implementation.</div></div><div><h3>Methods</h3><div>Three biofuel production routes, namely, biodiesel via two-step transesterification and bioethanol from whole and defatted biomass through acid hydrolysis, were systematically assessed, with a functional unit of 1000 kg (dry biomass) of <em>C. vulgaris</em>.</div></div><div><h3>Significant findings</h3><div>Total costs associated with chemicals and energy demands were $ 2.30/L, $ 95.52/L, and $ 0.73/L to produce biodiesel, bioethanol, and defatted biomass-based bioethanol, respectively. Consequently, the most profitable biofuel production using defatted <em>C. vulgaris</em> biomass has a profitability ratio of 1.01 (>1). Moreover, biodiesel production demonstrates the most renewable system with a net energy ratio of 6.28 and a net energy balance of +29.93. Nonetheless, the production viability of biodiesel and bioethanol (whole biomass) could be enhanced by incorporating cost- and energy-effective strategies. Additionally, the valorization of defatted biomass for bioethanol production would be a promising strategy, facilitating a circular bioeconomy and enhanced sustainability.</div></div>","PeriodicalId":381,"journal":{"name":"Journal of the Taiwan Institute of Chemical Engineers","volume":"179 ","pages":"Article 106395"},"PeriodicalIF":6.3,"publicationDate":"2025-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145109658","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Sara Kishan Roodbari, Vahid Mohebbi, Reza Mosayyebi Behbahani
{"title":"Gas hydrate dissociation of CO2-CH4 mixtures in porous media with sodium chloride – A fundamental basis for prediction of carbon capture in gas hydrates reservoirs","authors":"Sara Kishan Roodbari, Vahid Mohebbi, Reza Mosayyebi Behbahani","doi":"10.1016/j.jtice.2025.106423","DOIUrl":"10.1016/j.jtice.2025.106423","url":null,"abstract":"<div><div>Background: The rising atmospheric CO<sub>2</sub> levels, driven primarily by growing global energy demands, pose urgent climate challenges. In support of the International Energy Agency's ambitious carbon reduction goals, this study investigates CO<sub>2</sub>/ CH<sub>4</sub> hydrate systems as a potential carbon capture solution. We specifically examine hydrate behavior under conditions mimicking deep-sea environments, with particular attention to sodium chloride's influence.</div><div>Methods: Laboratory experiments employed a high-pressure system to analyze CH<sub>4</sub> and CO<sub>2</sub> hydrate equilibrium conditions in reservoir rock samples featuring 20 nm pores. The isochoric pressure-search method was applied to perform tests in saltwater solutions (3.5% NaCl). Key parameters—including concentration, pressure, and temperature—were rigorously controlled to accurately simulate hydrate formation and dissociation processes.</div><div>Significant Findings: It was observed that increasing CO<sub>2</sub> levels reduce the equilibrium pressure of CH<sub>4</sub> + CO<sub>2</sub> mixtures. The thermodynamic model used exhibited a strong predictive ability with an average absolute deviation of 3.38%. It was found that in nanopores (<20 nm), capillary forces govern hydrate formation while salinity effects become marginal, as evidenced by comparative water/salt system analyses. The study highlights the role of pore size, water salinity, and capillary pressures on hydrate stability, which are critical for improving CO<sub>2</sub> sequestration methods and advancing natural gas recovery techniques.</div></div>","PeriodicalId":381,"journal":{"name":"Journal of the Taiwan Institute of Chemical Engineers","volume":"179 ","pages":"Article 106423"},"PeriodicalIF":6.3,"publicationDate":"2025-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145119789","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Online improving NOX concentration prediction at SCR system outlet through continual learning","authors":"Peng Chen , Baochang Xu , Wei He , Hongtao Hu","doi":"10.1016/j.jtice.2025.106417","DOIUrl":"10.1016/j.jtice.2025.106417","url":null,"abstract":"<div><h3>Background</h3><div>Accurately predicting NO<sub>X</sub> concentration at SCR system outlet is crucial for optimizing process parameters and reducing NO<sub>X</sub> emissions. Existing prediction models often struggle to maintain accurate long-term online predictions when operating conditions change and new data arrive. Therefore, frequent model updates are required in practical applications. However, new models often exhibit catastrophic forgetting of learned patterns, leading to a deterioration in NO<sub>X</sub> concentration prediction accuracy.</div></div><div><h3>Methods</h3><div>A novel continual learning algorithm, termed TDIR, is proposed. This algorithm dynamically identifies historical temporal samples that are most susceptible to interference from new data and prioritizes their replay. The TDIR algorithm is integrated with the iTransformer architecture to establish TDIRformer, an online self-updating model for NO<sub>X</sub> emission prediction. The model utilizes variate tokens to capture cross-feature correlations, effectively addressing prediction inaccuracies caused by multivariate coupling. It also employs the TDIR algorithm for online updates, which mitigates catastrophic forgetting and improves the prediction accuracy of NO<sub>X</sub> concentrations.</div></div><div><h3>Significant findings</h3><div>Experimental results show that TDIRformer significantly outperforms LSTM, Informer, PatchTST, and iTransformer, while TDIR also surpasses the continual learning methods EWC and MAS. Additionally, the TDIR algorithm demonstrates strong generalization capability, achieving RMSE reductions of 10.7 %, 9.3 %, and 11.9 % when applied to LSTM, Informer, and PatchTST models, respectively.</div></div>","PeriodicalId":381,"journal":{"name":"Journal of the Taiwan Institute of Chemical Engineers","volume":"179 ","pages":"Article 106417"},"PeriodicalIF":6.3,"publicationDate":"2025-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145109659","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}