Scientific African最新文献

{"title":"Long-term spatiotemporal trends and population-weighted exposure to PM2.5 in Greater Cairo (2000–2020)","authors":"Gamil Gamal ,&nbsp;Naglaa Zanaty ,&nbsp;Pavol Nejedlik","doi":"10.1016/j.sciaf.2025.e03150","DOIUrl":"10.1016/j.sciaf.2025.e03150","url":null,"abstract":"<div><div>Fine particulate matter (PM₂.₅) pollution poses significant environmental and health risks in urban areas, such as Greater Cairo (GC). To provide the first high-resolution, two-decade assessment (2000–2020) of chronic exposure, this study integrates 1 km satellite-derived PM₂.₅ concentrations with comprehensive population distribution dynamics (LandScan Global). Trends were assessed using the Mann–Kendall test, and Sen’s slope was also calculated. We utilized advanced exposure metrics: the Population-Weighted Mean Concentration (PWMC) and Total Exposure (TE). A multi-level classification based on the five World Health Organization (WHO) Interim Targets and Guideline Value (IT-1, IT-2, IT-3, IT-4, and the Guideline of 5 µg/m³) was used to quantify the health significance of exposure evolution rigorously. PWMC consistently violated strict international health standards, remaining 7 to 10 times higher than the WHO Guideline (5 µg/m³) over the entire period, indicating a 100% population exposure risk. Concentrations exhibited a non-linear temporal pattern, peaking around 2010 (domain-averaged peak at 45.6 µg/m³) before showing a decline through 2020. Crucially, analysis revealed that high population growth increased the overall health burden (TE), actively counteracting localized improvements in pollution intensity (PWMC). Robust K-means clustering confirmed that the highest exposure burdens persisted in dense urban cores and industrial corridors. These findings provide quantitative evidence essential for targeted air quality management, demonstrating the urgent need for integrated urban planning and localized policies to reduce the cumulative health burden in this megacity.</div></div>","PeriodicalId":21690,"journal":{"name":"Scientific African","volume":"31 ","pages":"Article e03150"},"PeriodicalIF":3.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145924991","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An integrated computational approach combining QSAR modeling, molecular docking, and ADME profiling for the discovery of selective CYP11B1 inhibitors","authors":"Mohamed El Yaqoubi ,&nbsp;Mouad Lahyaoui ,&nbsp;Ahmed Mazzah ,&nbsp;Hafsa El-idrissi ,&nbsp;Yousra Seqqat ,&nbsp;Amal Haoudi ,&nbsp;Riham Sghyar ,&nbsp;Taoufiq Saffaj ,&nbsp;Bouchaib Ihssane ,&nbsp;Fouad Ouazzani Chahdi ,&nbsp;Youssef Kandri Rodi","doi":"10.1016/j.sciaf.2025.e03176","DOIUrl":"10.1016/j.sciaf.2025.e03176","url":null,"abstract":"<div><div>CYP11B1 inhibitors play a critical role in controlling cortisol biosynthesis and represent promising therapeutic candidates for disorders such as Cushing’s syndrome and hypertension. In this study, a series of novel analogues were designed and evaluated using an integrated approach combining quantitative structure–activity relationship (QSAR) modeling, molecular docking, and ADME predictions. Multiple linear regression (MLR), partial least squares (PLS), and principal component regression (PCR) models were constructed to establish robust predictive relationships between molecular descriptors and inhibitory activity against CYP11B1. The models were rigorously validated through external test-set prediction, Y-randomization, and applicability-domain (AD) analysis, all satisfying OECD criteria (R² = 0.725–0.772, Q² = 0.701–0.752, RMSE = 0.242–0.310).</div><div>Docking simulations revealed that compound D3 exhibited the most favorable binding affinity (−7.45 kcal/mol) and formed stable π–H and π–cation interactions with key residues Arg404 and Leu113, suggesting selective inhibition of CYP11B1. ADME and drug-likeness evaluation indicated predicted favorable pharmacokinetic properties, including high gastrointestinal absorption, absence of blood–brain barrier penetration, and good solubility, with D3 also demonstrating the lowest synthetic-accessibility score (SA = 3.09).</div><div>Overall, this integrated computational approach successfully identified D3 as a potent and synthetically feasible CYP11B1 inhibitor candidate. These findings provide a validated framework for the rational design and optimization of new inhibitors with improved pharmacological and metabolic profiles.</div></div>","PeriodicalId":21690,"journal":{"name":"Scientific African","volume":"31 ","pages":"Article e03176"},"PeriodicalIF":3.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145924984","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Impacts of zootechnical factors on Salmonella contamination in swab samples using real-time PCR at the Yaounde slaughterhouse","authors":"Chelea Matchawe ,&nbsp;Célestin Godwe ,&nbsp;Clarisse Engowei Mbah ,&nbsp;Tata B. Ndakoh ,&nbsp;Mélanie F.K. Gondam ,&nbsp;Henriette A. Essomba ,&nbsp;Fadimatou Ahmadou ,&nbsp;Manuela A. Baomog ,&nbsp;Mélissa Wangue ,&nbsp;Séverin Loul ,&nbsp;Marie-Chantal Ngondé ,&nbsp;Bonglaisin J. Nsawir ,&nbsp;Lucy M. Ndip ,&nbsp;Marco Galeotti ,&nbsp;Edi Piasentier","doi":"10.1016/j.sciaf.2026.e03183","DOIUrl":"10.1016/j.sciaf.2026.e03183","url":null,"abstract":"<div><div><em>Salmonella</em> contamination of beef carcasses remains a major public health concern, particularly in low- and middle-income countries where abattoir hygiene and traceability systems are often inadequate. This study aimed to generate context-specific data on Salmonella contamination along the cattle slaughter chain at the Yaoundé abattoir using real-time PCR, and to evaluate the influence of zootechnical factors of slaughtered cattle within a One Health framework. A total of 705 swab samples were collected from live cattle (<em>n</em> = 145), carcasses (<em>n</em> = 310), butchers’ hands (<em>n</em> = 145), and meat contact surfaces (<em>n</em> = 105). <em>Salmonella</em> detection was performed using TaqMan probe-based real-time PCR. Overall, 14.9% (95% CI: 12.4%–17.6%) of samples were positive for <em>Salmonella</em>, with prevalence rates of 5.20% (95% CI: 2.4%–10.6%) in live cattle, 17.73% (95% CI: 13.89%–22.38%) in carcasses, 1.9% (95% CI: 0.52% – 6.68%) in contact surfaces, and 5.52% (95% CI: 2.8% – 10.5%) in butchers’ hand swabs. <em>Salmonella</em> occurrence differed significantly among sample categories (<em>p</em> &lt; 0.05). However, multivariate logistic regression revealed that none of the assessed zootechnical factors (sex, age, breed, transport, origin, cleanliness, body condition, and production system) were independently associated with contamination (<em>p</em> &gt; 0.05). These findings suggest that <em>Salmonella</em> contamination in slaughtered cattle is driven by systemic hygiene and biosecurity shortcomings rather than individual animal-related factors. Overall, the moderate prevalence observed reflects gaps in slaughter hygiene and biosecurity. Strengthening sanitation practices, enforcing Hazard Analysis and Critical Control Point (HACCP) measures, and adopting molecular surveillance tools such as real-time PCR are essential to reduce contamination risks and protect public health.</div></div>","PeriodicalId":21690,"journal":{"name":"Scientific African","volume":"31 ","pages":"Article e03183"},"PeriodicalIF":3.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146077844","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integrating GOF tests and cross validation for copula model selection","authors":"Kevin Otieno ,&nbsp;Linda Chaba ,&nbsp;Evans Omondi ,&nbsp;Collins Odhiambo ,&nbsp;Bernard Omolo","doi":"10.1016/j.sciaf.2026.e03203","DOIUrl":"10.1016/j.sciaf.2026.e03203","url":null,"abstract":"<div><div>In dependence modeling, choosing the right copula is crucial, as different copula models can yield distinct interpretations of the relationship between variables. However, real-world applications are often constrained by the limitations of existing copula selection methods, which lack consistency and robustness across datasets. The selection methods in the literature that includes goodness-of-fit (GoF) tests and selection criteria, often yield conflicting results, thereby misrepresenting the dependence structure and leading to misleading conclusions. This study developed an integrated copula selection framework that combines GOF tests with cross-validation techniques. We integrated block-based cross-validation with GoF tests, where data was partitioned into blocks of different sizes <span><math><mrow><mo>(</mo><mi>K</mi><mo>)</mo></mrow></math></span>. A copula was fitted on the training set, and its performance was validated on the test set using GoF measures. The selection process was repeated across multiple folds, and an aggregation method was applied to determine the most suitable copula. The approach was tested through Monte Carlo simulations and an empirical study was tested on weather variables in Kenya. The findings show that Kendall-based Kolmogorov–Smirnov (KendallKS) and Cramér–von Mises (KendallCvM) test statistics integrated with stratified cross-validation, when <span><math><mrow><mi>K</mi><mo>=</mo><mn>5</mn></mrow></math></span>, perform better when the Benjamini–Hochberg (BH) procedure was used for aggregation. The proposed tests successfully identified the true copula and consistently rejected incorrect alternatives, with performance improving as sample size and dependence level increased. The empirical application further demonstrates the method’s robustness in real-world settings. These findings demonstrate that the proposed approach enhances the reliability and stability of copula selection.</div></div>","PeriodicalId":21690,"journal":{"name":"Scientific African","volume":"31 ","pages":"Article e03203"},"PeriodicalIF":3.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146077787","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Structure-based design and computational evaluation of new acrylamide derivatives as potent inhibitors of human dihydroorotate dehydrogenase for the treatment of rheumatoid arthritis","authors":"Rachid Haloui ,&nbsp;Amine Ballari ,&nbsp;Khadija Khaddam Allah ,&nbsp;Ayoub El-Mrabet ,&nbsp;Abdelmoula El Abbouchi ,&nbsp;Samir Chtita ,&nbsp;Ahmed Mazzah ,&nbsp;Amal Haoudi ,&nbsp;Souad Elkhattabi","doi":"10.1016/j.sciaf.2026.e03190","DOIUrl":"10.1016/j.sciaf.2026.e03190","url":null,"abstract":"<div><div>Human dihydroorotate dehydrogenase (DHODH) is a therapeutic target for the treatment of rheumatoid arthritis (RA). The development of new DHODH inhibitors could facilitate the discovery of a drug for RA therapy and contribute to sustainable health by promoting the design of safer and more efficient treatments. Using 3D-QSAR modeling techniques, we established a quantitative relationship between the DHODH inhibitory activity of 35 acrylamide derivatives (M1-M35) and their molecular fields. This model guided the design of 22 new acrylamide-based molecules (P1-P22) predicted to exhibit higher activity than M35, the most active in the initial database. The designed molecules were then docked into the DHODH active site, and the resulting complexes were evaluated using MM-GBSA free binding energy calculations. These analyses identified P4, P11, and P18 as the most DHODH inhibitors with docking scores and binding free energies superior to those of M35. The ADMET properties of the P4, P11, and P18 molecules were also predicted. The results show that they have good pharmacokinetic properties and are non-toxic. The molecular dynamics simulations of both free DHODH and its complexes with the best candidates confirmed the stability and validity of the obtained results. Finally, molecules P4, P11, and P18 show excellent capacity to inhibit the DHODH protein for the treatment of RA.</div></div>","PeriodicalId":21690,"journal":{"name":"Scientific African","volume":"31 ","pages":"Article e03190"},"PeriodicalIF":3.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146077789","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A sustainable hybrid cryptographic framework for energy optimization and scalability in blockchain-based e-voting","authors":"Adil Marouan ,&nbsp;Morad Badrani ,&nbsp;Abderrahim Zannou ,&nbsp;Nabil Kannouf ,&nbsp;Abdelaziz Chetouani","doi":"10.1016/j.sciaf.2026.e03191","DOIUrl":"10.1016/j.sciaf.2026.e03191","url":null,"abstract":"<div><div>Electronic voting (e-voting) is increasingly recognized as a viable alternative to traditional paper-based elections, offering automation, transparency, and improved accessibility. However, blockchain-based e-voting frameworks still face critical challenges, notably excessive energy consumption, scalability bottlenecks, and trade-offs between performance and security. This paper proposes a sustainable hybrid cryptographic framework that integrates the Elliptic Curve Digital Signature Algorithm (ECDSA), the Edwards-Curve Digital Signature Algorithm (EdDSA), and Boneh–Lynn–Shacham (BLS) signatures. The proposed design reduces computational and energy overhead through lightweight ballot signing and aggregated verification, while ensuring interoperability with widely deployed blockchain infrastructures. Moreover, the framework incorporates Layer-2 batching within a Proof-of-Stake (PoS) consensus mechanism to minimize transaction costs and latency. Experimental results show that the proposed system reduces per-transaction energy consumption by nearly 50%, doubles transaction throughput from 120 to 250 TPS, and enhances scalability compared to conventional blockchain voting solutions. These findings confirm that hybrid cryptography represents a sustainable and practical pathway towards secure, transparent, and energy-efficient blockchain-based e-voting systems applicable to both institutional and national contexts.</div></div>","PeriodicalId":21690,"journal":{"name":"Scientific African","volume":"31 ","pages":"Article e03191"},"PeriodicalIF":3.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146077791","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Digitization and the energy paradox: Does broadband adoption increase the electricity intensity of African businesses?","authors":"Youssef Bouazizi,&nbsp;Mohammed Ghazouani","doi":"10.1016/j.sciaf.2025.e03130","DOIUrl":"10.1016/j.sciaf.2025.e03130","url":null,"abstract":"<div><div>The digitization of African firms reveals a persistent energy paradox: while digital tools can improve efficiency, they may also intensify electricity demand in contexts marked by infrastructural fragility. This study provides the first multi-country, firm-level assessment of how broadband adoption shapes electricity intensity in Africa, using 10,755 observations from the post COVID World Bank Enterprise Surveys. To address endogeneity, the analysis relies on a causal identification strategy based on Instrumental Variables with the Two Stage Least Squares estimator and the Limited Information Maximum Likelihood estimator. The results show that broadband access increases electricity intensity in the short run, and that conventional estimations underestimate this effect due to measurement error and self-selection. At the same time, broadband adoption significantly promotes product and process innovation, stimulates investment in research and development, and reduces financial obstacles, revealing indirect channels that may support efficiency gains over time. A comprehensive set of robustness and sensitivity tests confirms the stability of these findings across alternative specifications, subsamples, and placebo exercises. By uncovering both the direct and indirect effects of digital integration, this study provides new evidence on the conditions under which connectivity can support Africa’s productive modernization. The results underscore the need for complementary reforms, including reliable electricity infrastructure, strengthened digital and energy capabilities, and expanded access to green and digital finance, in order to transform digitalization into a driver of sustainable energy transitions.</div></div>","PeriodicalId":21690,"journal":{"name":"Scientific African","volume":"31 ","pages":"Article e03130"},"PeriodicalIF":3.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145737943","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A survey of blockchain-based solutions for secure social networks","authors":"Fatima Anter ,&nbsp;Fatna El Mendili ,&nbsp;Nabil Benamar ,&nbsp;Nabil Mrani ,&nbsp;Mohammed Fattah","doi":"10.1016/j.sciaf.2026.e03233","DOIUrl":"10.1016/j.sciaf.2026.e03233","url":null,"abstract":"<div><div>Online social networks have become essential platforms for communication and information sharing, yet their centralized architectures expose users to serious privacy, security, and integrity risks. Blockchain technology has emerged as a promising paradigm for addressing these challenges by enabling decentralization, transparency, immutability, and user-centric control over data. This survey aims to systematically examine blockchain-based solutions designed to enhance security and privacy in social networks. Following the PRISMA methodology, we review and analyze peer-reviewed studies published between 2016 and 2025, focusing on blockchain architectures, consensus mechanisms, smart contracts, and their integration with machine learning and deep learning techniques for malicious behavior detection. The survey categorizes existing blockchain-based social networks, compares centralized and decentralized models, and evaluates key trade-offs such as scalability versus security and privacy versus transparency. Additionally, regulatory and governance implications are discussed. The findings indicate that blockchain can significantly improve privacy, security, data integrity, and user empowerment in social networks; however, challenges related to scalability, usability, interoperability, and legal compliance remain. This survey provides a comprehensive overview for researchers and practitioners and identifies open research directions for the secure evolution of social networks.</div></div>","PeriodicalId":21690,"journal":{"name":"Scientific African","volume":"31 ","pages":"Article e03233"},"PeriodicalIF":3.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146188167","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Sequential electrocoagulation–Fungal biotreatment as a pretreatment approach for olive mill wastewater (Fez, Morocco)","authors":"M.S. Fahmi ,&nbsp;H. Bellouk ,&nbsp;H. El Harchli ,&nbsp;A. El Mahjoub ,&nbsp;F. El Khammar ,&nbsp;A. Chaouch ,&nbsp;F. Khalil","doi":"10.1016/j.sciaf.2026.e03241","DOIUrl":"10.1016/j.sciaf.2026.e03241","url":null,"abstract":"<div><div>The present study aims to evaluate a preliminary sequential treatment approach combining electrocoagulation (EC) and fungal bioremediation for the purification of olive mill wastewater (OMWW) diluted at a 1:10 ratio. EC, conducted with aluminum electrodes at a current density of 38 A·m⁻² for 4 h, achieved an initial reduction in chemical oxygen demand (COD) from 9520 to 4096 mg·L⁻¹ (56 %) and in phenolic compounds (PC) from 0.55 to 0.28 g·L⁻¹ (49 %). Subsequently, biological treatment with <em>Aspergillus niger</em> for 72 h further decreased COD to 2856 mg·L⁻¹ and PC to 0.137 g·L⁻¹, corresponding to cumulative removals of 70 % and 75 %, respectively. The treated effluent exhibited marked physicochemical improvements, including a decrease in turbidity to 15 NTU, a near-neutral pH (7.1), and a conductivity of 10.0 mS·cm⁻¹. The operational cost of the integrated EC–fungal process was estimated at 1.75 USD·m⁻³, with an associated carbon footprint of 3.4 kg CO₂·m⁻³. These values represent indicative laboratory-scale estimates, requiring validation at the pilot scale with a full accounting of electrode consumption, sludge management, and fungal biomass handling. Overall, the sequential EC–biotreatment constitutes a promising and sustainable pretreatment strategy for high-strength OMWW, facilitating subsequent polishing stages rather than serving as a stand-alone treatment solution.</div></div>","PeriodicalId":21690,"journal":{"name":"Scientific African","volume":"31 ","pages":"Article e03241"},"PeriodicalIF":3.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146188164","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"State of health estimation for microgrid storage batteries using CNN-BiLSTM-attention-transfer learning","authors":"Khalid Errakkas,&nbsp;Mohammed Kissaoui,&nbsp;Rachid Lajouad,&nbsp;Abdelmounime El Magri,&nbsp;Ahmed Khayat,&nbsp;Youness Atifi","doi":"10.1016/j.sciaf.2026.e03244","DOIUrl":"10.1016/j.sciaf.2026.e03244","url":null,"abstract":"<div><div>This study investigates data-driven state-of-health forecasting for second-life lithium-ion battery modules operated under long-term residential microgrid duty cycles. While CNN–recurrent–attention models are widely used in laboratory SOH studies, reliable field deployment is limited by distribution shift and by evaluation leakage when overlapped time windows are used across train and test. We address these issues using a two-stage domain-adaptation framework: a CNN–BiLSTM multi-head attention model is first pre-trained on NASA laboratory aging data to learn general degradation representations, then selectively adapted to microgrid data by freezing the CNN feature extractor and the lower BiLSTM while fine-tuning the upper temporal and attention/regression layers. For microgrid scenarios, we enforce a strict chronological split using date-labelled quarter folders and construct sliding windows only after splitting, preventing raw-sample overlap between train/validation/test. Experiments show low error on NASA cells (average RMSE ≈ 1.24% and MAE ≈ 0.98%) and robust transfer performance on three microgrid scenarios with RMSE below 3.5% (PV: 2.76%, PV-TOU: 2.34%, TOU: 3.12%). These results indicate that selective freezing reduces target-data requirements while maintaining generalization under real duty-cycle variability, supporting practical SOH monitoring and maintenance planning for second-life battery energy storage systems.</div></div>","PeriodicalId":21690,"journal":{"name":"Scientific African","volume":"31 ","pages":"Article e03244"},"PeriodicalIF":3.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146188189","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}