This study introduces a hybrid calculation-estimation framework to optimize flow field designs for vanadium redox flow batteries (VRFBs), prioritizing directly calculable geometric and physical parameters over empirically fitted coefficients to enhance model fidelity and predictive accuracy. A fully validated three-dimensional multi-physics model, coupling fluid dynamics with electrochemical kinetics, is developed to systematically evaluate three distinct flow field architectures: a conventional serpentine design, a nature-inspired biomimetic leaf-venation network, and a modified serpentine channel featuring embedded micro-pillar perturbators. Comparative analysis reveals that biomimetic design achieves the most favorable trade-off between hydraulic and electrochemical performance. Its low-resistance, hierarchically branched architecture facilitates uniform electrolyte distribution across the porous electrode, resulting in a 35% reduction in pressure drop and a corresponding 3.2% increase in net system efficiency relative to the conventional baseline. In contrast, while the perturbator-enhanced design achieves the highest limiting current density (190 mA cm⁻²) by inducing localized vortex mixing to enhance mass transport, this gain is offset by a significant increase in pumping losses. The findings underscore that directly calculated parameters such as branching geometry and flow path length are critical drivers of performance. This work provides a principled modeling strategy and offers generalizable design guidelines, demonstrating that nature-inspired engineering is a key pathway toward developing next-generation, high-efficiency VRFB systems.

Existing fire emergency decision-making models often struggle with accurately handling multi-granularity uncertain linguistic information, loss aversion, and a lack of adaptability to dynamic fire evolution. To address these gaps, this study adopts two-tuple linguistic representation (2TLR) for quantifying multi-granularity linguistic information and combines the Analytic Hierarchy Process (AHP) with the entropy weight method (EWM) to determine the ability weights of the experts. Furthermore, a six-dimensional dynamic reference point is generated via the random forest algorithm, and the integration of prospect theory (PT) with a sequential decision-making framework (SDF) is implemented for the dynamic optimization of response plans. Validation through real-world cases demonstrates that the proposed Multi-stage Prospect Selection (M-PS) model outperforms both the TOPSIS method and the single PT model, compared with these two methods, the proposed M-PS model can effectively prioritize the avoidance of high-risk scenarios, accurately reflect decision-makers’ loss aversion tendency, and realize dynamic decision-making through updating the decision plan sequentially, thereby providing reliable support for urban fire emergency management. At the same time, in this study we conduct a comparative analysis of core metrics between existing methods and the proposed M-PS model. The evaluation across five dimensions demonstrates that the proposed M-PS model delivers superior performance.

This work develops a control-oriented, lumped-parameter model for vanadium redox flow battery (VRFB) stacks. The framework integrates mass, charge, energy, and momentum transport with electrochemical kinetics via a coupled system of ordinary differential equations (ODEs) and algebraic constraints, bridging system dynamics and electrochemical engineering. A key methodological advancement is the application of a hydraulic-electrical network analogy, utilizing Kirchhoff's laws to simulate electrolyte flow and shunt current pathways across a 20-cell stack, thereby transforming complex three-dimensional physics into a tractable, control-oriented formulation. The model directly links physical fidelity to actionable performance insights. Simulations identify that non-uniform flow distribution induces significant local state-of-charge gradients, exacerbating shunt currents. This parasitic effect can reduce effective charging current by up to 2.1% and increase discharge overpotentials. Through analysis of these coupled interactions, the study demonstrates that optimized flow management and thermal control can mitigate losses. Specifically, regulating stack temperature below 40 °C via a novel targeted tank-cooling strategy rather than full-system cooling prevents vanadium precipitation while improving round-trip efficiency, achieving a 27.2% reduction in cooling energy consumption. Furthermore, the model reveals that tank-based heat rejection dominates convective heat transfer (85.8%), enabling a transformative redesign where thermal management is consolidated at the tanks. This permits a more compact stack enclosure and reduces balance-of-plant complexity. The work establishes a validated mathematical framework that advances the fundamental understanding of coupled transport in VRFBs and provides a direct pathway to designing more efficient, compact, and cost-effective systems.

This study proposes a framework for optimizing cyclic discharge-reload operations in successive batch bioreactors under integral input constraints (e.g., total substrate consumption, cumulative energy use). Prior research has shown that candidates for optimal fed-batch and periodic on/off operation, formulated in an input-affine system, can be derived from bang-bang form. However, the practical application of these results to successive batch configurations requires a prior knowledge of operational periodicity, which limits flexibility and adaptability to variability. To address this limitation, we reformulate the switching strategy into an algebraic form using the Fliess series expansion. This expansion explicitly embeds periodicity constraints, enabling direct computation of switching parameters from predefined boundary conditions. The overall performance, defined as the over-yield relative to steady-state operations, is quantified via integrated integrals. For practical implementation, we leverage self-cycling fermentation as a foundational framework to determine required periodicity, and design optimal trajectories. Analytical results are derived to clarify the quantitative relationship between boundary conditions and switching parameters for any arbitrary anchor point (i.e., reference state for cycle initialization). The proposed framework is validated using an experimentally calibrated ethanol fermentation model. Simulation results demonstrate that the optimized cyclic strategy achieve a 25.58% relative increase in ethanol yield compared to traditional batch operations. This method enables the seamless integration of on/off operation with optimal periodicity, addressing critical gaps between theoretical periodic switching strategy and industrial bioprocess implementation.

Digital currency markets exhibit extreme volatility, heavy tails, and bursty order-flow that complicate surveillance and heighten manipulation risk. This systematic review synthesizes the state of the art in stochastic differential equation (SDE)–based approaches for detection and monitoring of abusive practices in cryptocurrencies. Following PRISMA 2020 and PICOS, an Elsevier-indexed search (2015–2025) yielded 273 records; after screening and eligibility assessment, 20 primary studies (2018–2025) were included. The literature clusters into six methodological families: (i) stochastic volatility and jump–diffusion models for heavy-tailed returns and volatility smiles; (ii) Hawkes and related point-process models for clustered order arrivals, spoofing, pump-and-dump, and wash trading; (iii) Markov and regime-switching diffusions that delineate latent “fair” versus manipulated regimes; (iv) hybrid SDE–machine learning frameworks for high-frequency prediction; (v) change-point and sequential detection methods grounded in likelihood ratios and optimal stopping; and (vi) meta-studies consolidating performance trends. Across studies, Hawkes-type intensities consistently outperform Poisson and threshold baselines in event detection; regime-switching models align with known market breaks; and hybrid neural–SDE systems achieve the strongest forecasting but at reduced interpretability. We formalize a taxonomy linking model structure to surveillance objective, and we delineate a practical trade-off between transparency (classical SDEs) and predictive accuracy (hybrid models). The review highlights open needs in explainable hybrid designs, reproducible datasets, and real-time deploy ability for exchanges and regulators. By connecting applied probability, financial engineering, and control, the paper clarifies how SDE frameworks can underpin robust, auditable market-integrity tools for digital assets.

Medical images contain highly sensitive diagnostic and personal information that requires robust protection during storage and transmission. To address this, we propose a region-of-interest (ROI)-based hybrid encryption algorithm that combines pixel-level and bit-level permutation with bit-wise diffusion driven by an improved Lorenz chaotic system. The scheme first employs a robust ROI perception mechanism to accurately identify diagnostically important areas while avoiding unnecessary processing of non-critical regions, thereby enhancing computational efficiency and security. Image-dependent SHA-256 hashing is integrated to generate keystreams tightly bound to image content, improving key sensitivity and resisting plaintext attacks. Dual-layer chaotic scrambling ensures both global confusion and local diffusion, while a dedicated bit-wise diffusion stage further randomizes the ciphertext, strengthening resistance against differential, statistical, and noise-based attacks. Experimental evaluations demonstrate that the proposed method achieves high security and robustness: the average information entropy of encrypted images reaches 7.9992, and NPCR and UACI values are 99.63% and 33.47%, respectively. Compared with existing encryption techniques, the proposed algorithm exhibits higher randomness, stronger differential attack resistance, and better protection of sensitive medical data, without embedding ROI location metadata into non-interest regions. The results indicate that this approach provides an efficient and secure framework for safeguarding medical images in telemedicine, healthcare information systems, and other critical applications.