安永

In conventional air pollution research, natural dispersion, industrial emissions, and ecological absorption are typically regarded as the dominant mechanisms shaping PM2.5 exposure. However, under conditions of intensified human intervention in the Earth’s surface, this logic is undergoing a paradigmatic shift. Drawing on multi-regional panel data from China spanning 2015–2022, this study develops a machine learning model with PM2.5 concentrations as the outcome variable. By incorporating the Human Footprint Index (HF), alongside a suite of ecological, meteorological, and socio-economic variables, the analysis seeks to identify the associated mechanisms underlying pollution anomalies. The results indicate that: (1) at the global scale, the Human Footprint Index (HF) surpasses all natural and socio-economic variables, emerging as the primary determinant of PM2.5 exposure; (2) the effect of HF exhibits pronounced regional heterogeneity, with a strong positive structure observed in Central and Northeastern China, while in the Eastern region the effect tends towards neutrality due to more favourable dispersion conditions and intensified governance, and although the overall contribution in the Western region remains relatively low; (3) the marginal pollution effect of HF demonstrates a nonlinear threshold pattern, appearing buffered or insensitive at low-intensity levels, but rising sharply in pollution risk once a critical threshold is exceeded. The findings suggest that the explanatory logic of air pollution is shifting towards a new paradigm centred on the Human Footprint. Accordingly, this study advocates the development of a pollution early-warning and governance framework that is sensitive to the intensity of human activity and grounded in the identification of spatial threshold effects. The analysis further demonstrates the theoretical and practical potential of interpretable machine learning for uncovering pollution-related mechanisms and informing regionally differentiated policy design.

Permanent magnet synchronous motors (PMSMs) are widely used in electric vehicle (EV) applications due to their high efficiency and excellent torque characteristics. Field-oriented control (FOC) enables precise decoupled regulation of torque and flux in these systems, with proportional-integral (PI) controllers remaining the practical standard for current and speed loops. However, tuning these controllers is challenging because the closed-loop dynamics depend on the motor, inverter, and the vehicle's dynamic load, which results in time-varying inertia and operating conditions. This paper presents a comprehensive PI controller design methodology for PMSM FOC that explicitly incorporates vehicle dynamics into the control framework. The approach accounts for inverter delay, sampling effects, and load-dependent inertia, enabling systematic gain selection for both current and speed loops. The tuned controllers are first evaluated through detailed simulations, including representative EV drive cycle, to assess tracking accuracy, transient response, and robustness. Experimental validation on a real-time controller platform further confirms the practical applicability of the proposed method. Simulation results show that the proposed approach achieves drivetrain efficiencies of 92.81% and 91.84% for the Japanese and combined urban and extra-urban driving cycle (CUEDC) respectively, compared with 59.89% and 85.05% obtained using conventional tuning without considering vehicle dynamics, corresponding to efficiency improvements of approximately 7.98% and 55.13%. These findings demonstrate that integrating vehicle dynamics into the PI tuning process significantly enhances efficiency, improves stability margins, and ensures consistent performance across a wide range of operating conditions.

The evaluation of output statistics in systems with high-dimensional uncertain parameters is important for real-time decision-making in large-scale systems. In this paper, we develop an MPCM-Taguchi method that combines the Multivariate Probabilistic Collocation Method (MPCM) with the Taguchi method to accurately estimate system output statistics while significantly reducing the number of simulations. The estimation algorithm is developed, and its theoretical analysis is provided. Numerical studies and the application to power buffers in Direct Current (DC) Microgrids with uncertain loads validate the MPCM-Taguchi uncertainty evaluation method.

Introduction Brain disorders, including neurological conditions and mental disorders, pose a considerable economic burden in Denmark. However, the costs of illness are not limited to the patients themselves, as the consequences of living with a brain disorder may also impact close relatives. We aimed to describe any societal costs related to healthcare utilisation and income loss for the closest family relatives of patients with brain disorders. Methods This population-based cohort study included the closest family relatives (proxies for caregivers) of patients with prevalent (by 1 January 2021) or incident (during 2016–2021) brain disorders, using registry data. They were compared with corresponding relatives of matched population comparisons. Patients were categorised into three age strata: children and young people (0–24 years), adults (25–64 years) and older adults (65+years). We specified criteria for identifying the closest family relative for each age stratum. Using data from national registries, we estimated attributable healthcare costs and, for working-age relatives (18–65 years), income loss. Results In 2021, close relatives of patients with brain disorders included 125 495 fathers and 96 154 mothers of children and young people, 880 661 relatives of adults and 378 826 relatives of older adults. The pooled attributable costs of brain disorders incurred by closest family relatives were €2407 million for prevalent disease in 2021 and €794 million for incident disease the first year following incidence. The dominating cost component was income loss for working-age relatives. Conclusions The higher healthcare costs and especially the lower income for relatives of patients with brain disorders could be associated with the societal economic burden of brain disorders in Denmark. Informal care provided by relatives should be considered in discussions of healthcare prioritisation.

Business trips and traditional secondments remain among the best-known and most common forms of cross-border employee assignments. Since the end of the COVID-19 pandemic, new forms of flexible working have been increasing. Variations of ‘remote work’ include not only working from home, but also virtual assignments, ‘workations’ and hybrid business trips (mixing work and leisure time). The Posted Workers Directive and the rules on social security coordination remain cornerstones in governing such new situations in Europe, but were not designed for them. Modern EU Court case law as well as proposed measures may provide renewed fair solutions balancing Member State interests, employer compliance obligations and employee protection needs.