Advancing Open Science

Microbial lipid production from renewable carbon sources, particularly lignocellulosic hydrolysates, is a promising alternative to plant-derived oils and fats for food applications, as it can minimize the land use by utilizing agricultural wastes and byproducts from food production. In this context, a standard approach to prevent oxygen limitation at reduced air gassing rates during long-term aerobic microbial processes is to operate bioreactors at increased pressure for elevating the gas solubility in the fermentation broth. This study investigates the effect of absolute pressures of up to 2.5 bar on the conversion of the carbon sources (glucose, xylose, and acetate), growth, and lipid biosynthesis by Cutaneotrichosporon oleaginosus converting a synthetic nutrient-rich lignocellulosic hydrolysate at low air gassing rates of 0.1 vessel volume per minute (vvm). Increasing pressure delayed xylose uptake, reduced acetic acid consumption, and reduced biomass formation. Lipid accumulation decreased with increasing pressure, except for fermentations at 1.5 bar, which achieved a maximum lipid content of 83.6% (±1.6, w/w) (weight per weight in %). At an absolute pressure of 1.5 bar, a lipid yield from glucose, xylose, and acetic acid of 38% (w/w) was reached after 6 days of fermentation. The pressure sensitivity of C. oleaginosus may pose challenges on an industrial scale due to the dynamic changes in pressure when the yeast cells pass through the bioreactor. Increasing liquid heights in full-scale bioreactors will result in increased hydrostatic pressures at the bottom, substantially reducing lipid yields, e.g., to only 23% (w/w) at 2.0–2.5 bar, as shown in this study. However, further scale-up studies with dynamic pressure regimes (1–2.5 bar) may help to evaluate scale-up feasibility.

Intelligent recommender systems are essential for identifying at-risk students and personalizing learning through tailored resources. Accurate prediction of student performance enables these systems to deliver timely interventions and data-driven support. This paper presents the application of machine learning models to predict final exam grades in a university-level programming course, leveraging multi-modal student data to improve prediction accuracy. In particular, a recent raw dataset of students enrolled in a programming course across 36 class sections from the Fall 2024 and Winter 2025 terms was initially processed. The data was collected up to one month before the final exam. From this data, a comprehensive set of features was engineered, including the student’s background, assessment grades and completion times, digital learning interactions, and engagement metrics. Building on this feature set, six machine learning prediction models were initially developed using data from the Fall 2024 term. Both training and testing were conducted on this dataset using cross-validation combined with hyperparameter tuning. The XGBoost model demonstrated strong performance, achieving an accuracy exceeding 91%. To assess the generalizability of the considered models, all models were retrained on the complete Fall 2024 dataset. They were then evaluated on an independent dataset from Winter 2025, with XGBoost achieving the highest accuracy, exceeding 84%. Feature importance analysis has revealed that the midterm grade and the average completion duration of lab assessments are the most influential predictors. This data-driven approach empowers instructors to proactively identify and support at-risk students, enabling adaptive learning environments that deliver personalized learning and timely interventions.

A reconfigurable analog beamformer for the use case of multiband Global Navigation Satellite System (GNSS) multiantenna receiver systems is designed and tested. The beamformer board operates in all existing GNSS frequency bands. In this paper, the two commonly used GNSS bands, the E1/L1 and E5a/L5 GNSS bands at 1.575 GHz and 1.176 GHz, respectively, are studied. An analog weighting of the complex excitation of up to 14 individual channels is realized using attenuators and phase shifters, digitally controlled by proprietary PC software. We present an analysis of the relative errors between the channels and a simple calibration of constant errors which is applied and validated. The beamformer is then demonstrated in an exemplary test case, to generate an ad hoc pattern from an array of antennas.

To address issues such as significant scale differences, complex pose variations, strong background interference, and similar category characteristics of pests in the images obtained from field traps, this study proposes a pest recognition method based on a two-stage “segmentation–detection” approach to improve the accuracy of field pest situation monitoring. In the first stage, an improved segmentation model, BAE-UNet (Background-Aware and Edge-Enhanced U-Net), is adopted. Based on the classic U-Net framework, a Background-Aware Contextual Module (BACM), a Spatial-Channel Refinement and Attention Module (SCRA), and a Multi-Scale Edge-Aware Spatial Attention Module (MESA) are introduced. These modules respectively optimize multi-scale feature extraction, background suppression, and boundary refinement, effectively removing complex background information and accurately extracting pest body regions. In the second stage, the segmented pest body images are input into the YOLOv8 model to achieve precise pest detection and classification. Experimental results show that BAE-UNet performs excellently in the segmentation task, achieving an mIoU of 0.930, a Dice coefficient of 0.951, and a Boundary F1 of 0.943, significantly outperforming both the baseline U-Net and mainstream models such as DeepLabV3+. After segmentation preprocessing, the detection performance of YOLOv8 is also significantly improved. The precision, recall, mAP50, and mAP50–95 increase from 0.748, 0.796, 0.818, and 0.525 to 0.958, 0.971, 0.977, and 0.882, respectively. The results verify that the proposed two-stage recognition method can effectively suppress background interference, enhance the stability and generalization ability of the model in complex natural scenes, and provide an efficient and feasible technical approach for intelligent pest trap image recognition and pest situation monitoring.

Predicting the thermo-physiological comfort of technical clothing requires an understanding of how microscopic textile structures influence macroscopic properties such as air, heat, and moisture permeability. This work represents the first step towards a multi-scale predictive tool capable of estimating key comfort-related properties from the geometrical features of woven fabrics. Focusing on air permeability, the effect of structural and design parameters was investigated while keeping the fibre material (cotton) constant. A computational framework that combines validated Computational Fluid Dynamics (CFD) simulations with a Fully Connected Neural Network (FCNN) was developed, enabling fast and accurate predictions before production. The CFD model accounts for both intra- and inter-yarn porosity, ensuring reliability across a wide range of fabric configurations. The FCNN, trained on simulation and literature data, achieved a mean absolute relative error of 2.01% and a maximum error of 7.72%, demonstrating excellent agreement with experimental results. The analysis highlights how weave type and yarn density govern airflow resistance, offering an efficient tool for the design and optimisation of breathable technical textiles.

As the automation and intelligence of low-voltage distribution networks continue to advance, the inter-layer coupling between medium- and low-voltage distribution networks is increasingly strengthened, making traditional fixed-point iteration methods inadequate for distributed power flow calculation in such a collaborative framework. To address this issue, this paper proposes a distributed power flow calculation method for medium- and low-voltage distribution networks based on edge intelligence. First, a cooperative operational framework for medium- and low-voltage distribution networks is designed by integrating edge intelligence technology. Then, a distributed power flow calculation model is established, and its fixed-point iterative characteristics are analyzed. A convergence index calculation method based on small perturbations is proposed, followed by an iterative algorithm based on continuous intersection estimation. Finally, simulation case studies validate the proposed method in terms of accuracy, convergence, and computational efficiency, demonstrating its capability to meet the modeling and analytical needs of power flow calculation in medium- and low-voltage distribution networks, providing methodological support for the development of distributed intelligent power grids.

Considering the escalating international geopolitical tensions and the ensuing great power maneuvers, China’s oil supply faced unprecedented threats. To safeguard against these risks and harness domestic resources more effectively, addressing the stability of refined oil supply had become an urgent imperative. The complex network theory is integrated into oil product delivery logistics, accounting for transportation volumes, distances, and node importance. Through simulation, we evaluated each scheme’s efficacy using a case study from a province in northwest China. The results demonstrate notable improvements in network robustness across all four strategies. The key node focuses on protection measures emerged as the most effective, followed by the oil depot resource optimization strategy and the network topology optimization strategy, in descending order. By mitigating the risks stemming from international uncertainties, our strategies ensured the timely supply of refined oil products, thereby upholding the stable functioning of the national economy.

Point cloud semantic segmentation is a key technology for applications such as autonomous driving, robotics, and virtual reality. Current approaches are heavily reliant on local relative coordinates and simplistic attention mechanisms to aggregate neighborhood information. This often leads to an ineffective joint representation of geometric perturbations and feature variations, coupled with a lack of adaptive selection for salient features during context fusion. On this basis, we propose LSSCC-Net, a novel segmentation framework based on LACV-Net. First, the spatial-feature dynamic aggregation module is designed to fuse offset information by symmetric interaction between spatial positions and feature channels, thus supplementing local structural information. Second, a dual-dimensional attention mechanism (spatial and channel) is introduced to symmetrically deploy attention modules in both the encoder and decoder, prioritizing salient information extraction. Finally, Lovász-Softmax Loss is used as an auxiliary loss to optimize the training objective. The proposed method is evaluated on two public benchmark datasets. The mIoU on the Toronto3D and S3DIS datasets is 83.6% and 65.2%, respectively. Compared with the baseline LACV-Net, LSSCC-Net showed notable improvements in challenging categories: the IoU for “road mark” and “fence” on Toronto3D increased by 3.6% and 8.1%, respectively. These results indicate that LSSCC-Net more accurately characterizes complex boundaries and fine-grained structures, enhancing segmentation capabilities for small-scale targets and category boundaries.