Zhejiang Ocean University (ZJOU; simplified Chinese: 浙江海洋学院; traditional Chinese: 浙江海洋學院; pinyin: Zhèjiāng hǎiyáng xuéyuàn) is a comprehensive public university based in Zhoushan city, Zhejiang province, China.
Reliability-constrained jacket member sizing requires repeated response and probability evaluations under uncertain wind–wave loading. This study develops QAE–ABC for fixed-base OC4 jacket sizing by integrating multi-fidelity surrogate modeling, quantum amplitude estimation (QAE), and artificial bee colony (ABC) search. The selected NA-QAE-ABC design weighs 537.13 t, a 20.29% reduction from the 673.88 t baseline, and satisfies the screening criterion with an out-of-time system-response exceedance probability of 1.331×10−3 and an equivalent reliability index of βout=3.004. Historical North Sea records and multi-seed 600 s coupled OpenFAST simulations support response surrogates and a five-response reliability metamodel. Probability intervals rank feasibility, while adaptive queries target candidates near the probability boundary. Leave-one-design–environment-cell-out validation yields R2=0.9446 and 0.9677 for tower-base bending moment and tower-top displacement, without false-safe classifications at diagnostic upper-response thresholds. Controlled ablation against direct metamodel and Monte Carlo baselines isolates QAE’s algorithmic contribution under simulated amplitude counts. After site-specific recalibration, all 90 runs at three additional sites satisfied the out-of-time criterion, whereas equivalent-spring analysis showed that the fixed-base optimum depended on the support model. The workflow supports uncertainty-aware, probability-constrained member sizing within the prescribed geometry, support representation, and five-response screening scope.
Chestnut tannin, a bark-derived renewable polyphenolic product from the forest industry, was converted into porous carbon through hydrothermal carbonization and potassium-based activation. The effects of activator type and potassium hydroxide/carbon microspheres (KOH/CMS) mass ratio on pore evolution, surface chemistry, and electrochemical behavior were investigated to clarify the relationship between activation-regulated structure and charge storage. At a fixed activator/CMS mass ratio of 1:1, KOH produced a more electrochemically favorable porous framework with uniformly developed micropores, a higher defect level, and improved ion-accessible pathways, although K2CO3 and KHCO3 also induced visible pore development and surface roughening. Subsequent optimization of the KOH/CMS ratio identified CMS-KOH-3 as the best-performing sample, with a high specific surface area of 3069.9 m(2) g(-1) and a total pore volume of 1.989 cm(3) g(-1). Benefiting from its balanced micro/mesoporous architecture and reduced ion/electron transport resistance, CMS-KOH-3 delivered a specific capacitance of 243.2 F g(-1) at 0.5 A g(-1) and retained 76.9% of its capacitance at 10 A g(-1) in 6 M KOH electrolyte. A symmetric aqueous supercapacitor assembled with CMS-KOH-3 electrodes achieved an energy density of 16.04 Wh kg(-1) at 375 W kg(-1) and retained 91.3% of its initial capacitance after 18000 cycles. This study clarifies how potassium-salt activation regulates the pore evolution and surface chemistry of chestnut tannin-derived hydrochar, providing a thermochemical route to biomass-derived porous carbon electrodes for aqueous supercapacitors.
High-dimensional gene expression classification faces inherent challenges including limited sample sizes, pervasive feature redundancy, and the difficulty of building interpretable models without introducing excessive structural complexity. While conventional feature selection methods can identify discriminative genes, they fail to provide explicit guidance for organizing screened features to support localized fuzzy inference. To address these limitations, this paper proposes a causal-prior-guided hierarchical neuro-fuzzy system that integrates interaction-aware feature screening, structural-prior-guided subspace construction, and parallel fuzzy inference. Specifically, redundancy-controlled joint mutual information (RJMI) is adopted to preserve complementary and task-relevant genes. Restricted feature-level structural learning further yields sparse conditional dependence priors, which are fused with target dependency information to establish a Soft Markov Blanket-inspired (Soft-MB) hierarchical structure consisting of core, extended, and auxiliary feature layers. Guided by this hierarchy, multiple compact feature subspaces are constructed and separately modeled via parallel first-order Takagi-Sugeno-Kang (TSK) fuzzy subsystems with adaptive aggregation weights. Experimental results across twelve high-dimensional biomedical datasets demonstrate that the proposed method achieves the highest Macro F1 scores on ten datasets and the best average Macro F1 rank of 1.25 among seven evaluated methods. Friedman tests and Holm-corrected Wilcoxon signed-rank tests support the overall comparative advantage of the proposed method under cross-validation protocols. Ablation and sensitivity analyses verify the complementary functionality of each embedded component, whose effectiveness varies across datasets. Furthermore, a TCGA-STAD case study provides model-level interpretability through learned subspace weights and category-specific fuzzy rules. Overall, the proposed framework delivers competitive classification performance and transparent local fuzzy modeling, while avoiding overinterpretation of the learned dependency structure as a rigorous causal graph.
Abstract Coal gangue (CG) is a solid waste produced during coal mining that exhibits pozzolanic activity after calcination. Utilizing calcined coal gangue (CCG) to replace the calcined clay in limestone calcined clay cement ( LC 3 ) not only facilitates the resource utilization of industrial solid waste but also provides significant economic and environmental benefits. This study focuses on CCG–based LC 3 with 30% clinker content, using methods such as compressive strength testing, isothermal calorimetry, x-ray diffraction, and microstructural characterization to systematically investigate the effects of kaolinite content (20%, 40%, and 60%) and the addition of CH on hydration reactions and mechanical properties. The results indicate that CH significantly enhances the pozzolanic reactivity of CCG, particularly in the LC 3 system with high kaolinite content (60%), where the 28 day compressive strength increased from 24.0 to 36.9 MPa. Microstructural analyses revealed that CH addition promotes the formation of calcium (aluminate) silicate hydrate gels (C─(A)─S─H) and carboaluminate phases, while markedly refining the pore structure and reducing porosity. When the mass ratio of CCG:LS:CH was 2 ∶ 1 ∶ 1 , the system exhibited optimal overall performance. This study confirms the critical role of CH in improving the performance of low-CCG-based LC 3 cement and provides theoretical and practical guidance for the efficient utilization of CG with varying qualities.
To address challenges such as inadequate market participant motivation within the Credit Pool Management (CPM) mechanism under the Dual Credit Policy (DCP), we propose a New Credit Pool Management (NCPM) mechanism designed to flexibly regulate credit supply and demand in the market. This mechanism enables the government to inject credits at the ceiling price or to withdraw credits at the floor price once market prices reach predefined thresholds, subject to intervention caps. We then introduce a set of credit market clearing conditions capturing the interaction between the credit market and Automobile Manufacturers’ (AMs) decision-making, which endogenously incorporate the credit ceiling and floor prices. Building on this, we formulate a bi-level model that characterizes the interaction between the government’s credit pool parameter optimization at the upper level and the AMs’ decision optimization coupled with credit market equilibrium conditions at the lower level. Finally, we conduct a comparative analysis using numerical examples to evaluate the performance of the DCP under the unregulated market scenario, the CPM mechanism, and the NCPM mechanism. We found that, compared to scenarios with the unregulated market and the CPM, the NCPM mechanism can, under specific conditions, effectively regulate credit supply and demand while maintaining price stability at a limited fiscal outlay. Furthermore, it enhances profits for new energy AMs, increases consumer surplus in the New Energy Vehicle (NEV) market, and boosts the share of NEVs.