KITopen
Not a member yet
63711 research outputs found
Sort by
Human-centered evaluation of statistical parametric mapping and explainable machine learning for outlier detection in plantar pressure data
Get PDFPlantar pressure mapping is essential in clinical diagnostics and sports science, yet large heterogeneous datasets often contain outliers from technical errors or procedural inconsistencies. Statistical Parametric Mapping (SPM) provides interpretable analyses but is sensitive to alignment and its capacity for robust outlier detection remains unclear. This study compares an SPM approach with an explainable machine learning (ML) approach to establish transparent quality-control pipelines for plantar pressure datasets. Data from multiple centers were annotated by expert consensus and enriched with synthetic outliers resulting in 798 valid samples and 2000 outliers. We evaluated (i) a non-parametric, registration-dependent SPM approach and (ii) a convolutional neural network (CNN), explained using SHapley Additive exPlanations (SHAP). Performance was assessed via nested crossvalidation; explanation quality via a semantic differential survey with domain experts. The ML model reached high accuracy and outperformed SPM, which misclassified clinically meaningful variations and missed true outliers (Matthews Correlation Coefficient: ML = 0.96 ± 0.01; SPM = 0.78 ± 0.02). Experts perceived both SPM and SHAP explanations as clear, useful, and trustworthy, though SPM was assessed less complex. These findings highlight the complementary potential of SPM and explainable ML as approaches for automated outlier detection in plantar pressure data, and underscore the importance of explainability in translating complex model outputs into interpretable insights that can effectively inform decision-making
Between vision and barriers: a qualitative study on school leadership and ESD implementation in Germany
Get PDFThis study explores how school leaders in Baden-Württemberg, Germany, perceive and enact their role in the implementation of Education for Sustainable Development (ESD) within the framework of the Whole School Approach (WSA). Using a qualitative research design, we conducted semi-structured interviews with eleven school leaders from various school types. Thematic analysis identified five core dimensions shaping ESD implementation—vision, leadership, curriculum, resources, and communication—derived from the WSA and strongly reflected in participants’ narratives. In addition, an inductively identified cross-cutting theme—leaders’ understanding of ESD—emerged as central to how they engaged with all other dimensions. A notable contribution of this study is its nuanced documentation of leadership as a balancing act between aspiration, delegation, and constraint navigation. While school leaders expressed openness to sustainability, their engagement was often fragmented and environmentally focused. The study highlights the gap between policy ambition and school-level practice and underscores the need for leadership development, institutional support, and structural flexibility to realize the holistic aims of the WSA
Results of the SPIZWURZ bundle test
Get PDFThe 21-rod SPIZWURZ test bundle with three types of unirradiated claddings (opt. ZIRLO, Zry-4, DX-D4) was used to conduct a long-term integral experiment, approximately simulating dry storage conditions. A variety of parameters typical of an integrated test allowed for the acquisition of a large amount of experimental data necessary for verifying the corresponding computer codes. The claddings were preliminarily hydrogenated to concentrations of 100 and 300 wppm in a specially designed HOKI tubular furnace, distributed as uniformly as possible over a 1.3 m length. Hydrogenation was carried out at 450 °C by sequentially feeding fixed masses of hydrogen through a specially treated inner surface of the claddings. It was noted that the rate of hydrogenation of the opt. ZIRLO claddings is 1.5 times lower than for the Zry-4 claddings. After the hydrogen loading of the cladding tubes, the axial distribution of hydrogen was determined by laser scanning profilometry calibrated by hot gas extraction.
During the experiment, two values of internal rod pressure were used: 106 and 146 bar, which were maintained constant throughout the experiment (250 days). The peak cladding temperature decreased in steps of ≈15 K from 400 to 165 °C (average cooling rate ≈0.9 K/day). The maximal cooling rate during each temperature step was 6 K/h, step duration was about 10 h.
The post-test laser scanner measurements of the outer cladding diameter showed significant creep: radial deformation values are between 0.2 and 3.3% (diameter increase and the corresponding wall thinning). The largest creep of 3.3% was measured for opt. ZIRLO claddings hydrogenated to 300 wppm. The corresponding maximum creep value was 0.93% for Zry-4 and 1% for DX-D4. A clearly visible dependence of the degree of creep on the hydrogen concentration is observed for the opt. ZIRLO claddings: the creep of claddings hydrogenated to 300 wppm is 1.2-1.5 times higher than that of claddings hydrogenated to 100 wppm. A number of claddings show radially asymmetric wall thinning, which can be associated with the radial shift of the pellets from the central axis of the rod and the corresponding asymmetric heat supply along the circumference of the cladding.
The metallographic investigations revealed a uniform distribution of hydrides throughout the entire cladding circumference for all three cladding types used. In the DX-D4 claddings, hydrogen primarily diffused toward the outer liner. The degree of hydride reorientation was significantly higher in the Zry-4 claddings compared to the opt. ZIRLO claddings.
The difference in the behavior of the Zry-4 and opt. ZIRLO may be due, in part, to their different grain microstructures. Opt. ZIRLO claddings have a finer grain size than Zry-4. Moreover, although the temperatures during hydrogenation (450 °C) and the experiment itself (max. 400 °C) were relatively low, EBSD measurements showed grain growth from approx. 6 μm for the initial state to post-test 16 μm for Zry-4 (12 μm after hydrogenation), and from initial 3 μm to post-test 4.3 μm for opt. ZIRLO
Highly-cyclable Na-ion battery exploiting a nanostructured tin-carbon anode, layered-oxide P3/P2 cathode and a glyme-based electrolyte
Get PDFAlternative materials to (purely) carbon-based anodes could enhance the energy density of sodium-ion batteries, and thus favor their complementarity to lithium-ion batteries. This work provides a viable setup of Na-ion cells
combining a P3/P2 sodium-deficient layered cathode and a tin-carbon Na-alloying anode with a glyme-based electrolyte. Galvanostatic cycling in sodium half-cells of the water-processed alloying anode with sodium car-boxymethyl cellulose (CMC) binder shows a maximum capacity of ~260 mAh g, a capacity retention exceeding 70 % after 150 cycles, and an average Coulombic efficiency over 99 %. The multi-metal cathode evidences a great cycling stability over 100 cycles, with average Coulombic efficiency between 99.5 and 99.6 % as favored by the presence of Al ions in its structure. Full Na-ion batteries exploiting ad hoc chemically-sodiated tin-based anode and sodium-deficient layered cathode operate with average working voltage of 3 V, and maximum capacity of 120 mAh g retained for 95 % over 100 cycles in the best experimental setup. The rationally designed full-cell reaches theoretical energy density between 310 and 250 Wh kg as referred to the cathode weight
Strengthening Resilience for Critical Supply Chain Networks: Strategic Optimization and Decision Support
No full textAs global supply chains become more interconnected and complex, strengthening their ability to anticipate, manage, and recover from disruptions presents a growing challenge, particularly in critical sectors such as food and pharmaceuticals, where disruptions can have severe consequences. Resilience in these contexts involves strategic planning approaches that strengthen supply networks, improve transparency across operations, and support targeted responses to disruptions. These capabilities enable both public authorities and companies to manage operations more effectively, monitor performance, and maintain the continuity of supply. While interest in supply chain resilience is rising, there is still considerable potential for deeper evaluations of network strategies, datadriven decisions, and insights from real-world case studies. Given these challenges, this dissertation presents six research studies that propose strategic optimization and decision support methodologies, contributing to resilience in critical supply chain networks.
Study A focuses on optimizing facility placement and the distribution of critical goods, accounting for disruptions in warehouse availability and operability, route failures, coverage limitations, and demand variability. Study B addresses network design under varying impact scenarios, integrating location-allocation decisions to minimize shortages and leverage economies of scale in storage. The study employs stochastic modeling to assess crisis intensity, facility size, capacity limits, and cost-weighting strategies.
The models in Studies A and B are applied to Germany’s national food stockpiling system, demonstrating their applicability and practical relevance. Study C analyzes national food stockpiling strategies through a comparative analysis of different countries, highlighting their benefits and limitations. Study D introduces a routing model that optimizes fleet composition by balancing the strengths of different vehicle types, such as trucks and drones, and considers the specific requirements of humanitarian logistics. A case study based on the COVID-19 pandemic assesses the model’s relevance and sensitivity. Study E develops a Knowledge Graph to systematically structure, link, and integrate heterogeneous data sources, enabling a comprehensive analysis of drug shortages and supporting complex queries to uncover contributing factors and enhance strategic planning. Study F deepens the understanding of supply chain resilience in the pharmaceutical sector. It emphasizes the strategic importance of resilience-building capabilities and proposes a structured approach to link them with sector-specific vulnerabilities.
Beyond the findings of the studies, this dissertation offers overarching implications across key areas, including network design, disruption management, data-driven decisions, and the development of strategic insights for diverse stakeholders. Collectively, these outcomes provide valuable contributions to strengthening the resilience of critical supply chain networks
Activity Boost for Cobalt–Selectivity Shift and Activation During CO 2 Methanation by Carbon Nitride‐Modified Al 2 O 3 Support
No full textImmobilizing single atom on high-entropy oxides as separator regulators for catalyzing low-temperature lithium-sulfur battery
No full text
