Altay Arbak

Prismatic radial slots machined into the aluminum rotor disk of an eddy current brake (ECB) increase the convective cooling area, but their effect on steady-state temperature has not been quantified systematically. This paper presents a two-dimensional polar finite difference thermal model, coupled with a one-dimensional pin-fin shaft model through an iterative energy balance loop, to evaluate 12 slotted disks for an axial-flux ECB with double-sided excitation, eight poles, and a rated speed of 3000 min⁻¹. Four slot counts (N = 12, 24, 36, 48) and three bridge subdivisions (single, double, triple) are compared with a solid disk under the same 400 W Joule heat load, which the excitation current holds constant for every disk; the effect of the slots on the loss distribution is not modeled. The model predicts a solid disk peak temperature of 165.5 °C, which the best slotted configuration (48 slots, triple subdivision) reduces to 130.8 °C, a 34.7 °C (21.0%) reduction. Slot count is the dominant design variable: increasing N from 12 to 48 lowers the peak temperature by 21.9 °C (14.2%), while bridge subdivisions contribute only 0.6–1.4 °C. At N = 48, slot-wall convection carries 37.0% of the heat. These are numerical predictions; the ranking survives variations in the face convection, slot width and heat load, and of the slot-wall convection down to one third of its estimate.

This study presents a comprehensive 3E (Energy, Exergy, Environmental) analysis and multi-criteria selection of nine low global warming potential (GWP) refrigerants as alternatives to R404A in commercial refrigeration systems. Thermodynamic calculations were performed using the REFPROP 10.0 database for evaporation temperatures between -20 °C and 0 °C and condensation temperatures between 35 °C and 45 °C. Environmental effects were also evaluated for the same temperatures. The TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method was applied with base-case criteria weights of Total Equivalent Warming Impact (TEWI) (40%), Exergy Efficiency (35%), and COP (25%). R1234ze(E) achieved Rank 1 in 13 of the 15 operating conditions; R290 led at the remaining two, where it marginally outperforms R1234ze(E) on all three criteria simultaneously. Sensitivity analysis across five weight scenarios confirmed the robustness of these rankings under all tested weight distributions. The choice between R1234ze(E) and R290 represents an operational sustainability trade-off: R1234ze(E) offers practical safety advantages (A2L) for retrofit applications, while R290 provides superior environmental credentials (GWP = 3, PFAS-free) for new installations. All top-ranked refrigerants comply with the GWP thresholds of the Kigali Amendment and the EU F-Gas Regulation (2024), confirming their long-term regulatory viability.

Flux-switching permanent-magnet (FSPM) generators are gaining prominence in renewable energy applications due to their robust rotor structure and high-power density. However, thermal management remains a critical challenge, as excessive heat can degrade insulation and permanent magnet (PM) properties. This study presents a comprehensive thermal analysis of a 12-slot/10-pole FSPM generator using 3D Computational Fluid Dynamics (CFD) via ANSYS Icepak. Electromagnetic losses, including copper, core, and eddy-current losses, were integrated into the finite-volume formulation to solve for conduction, convection, and radiation effects. Simulation results identify the primary hotspots in the central axial region of the armature windings, reaching a peak steady-state temperature of 78 °C. The rotor and PMs exhibit strong thermal coupling, stabilizing within a range of 69 °C to 78 °C. Transient analysis confirms that all components reach thermal stability after approximately 90 minutes. The findings demonstrate that peak temperatures remain well below the critical limits for NdFeB magnets and Class F insulation, validating the thermal reliability of the proposed design for long-term operation.

Net-zero energy buildings (NZEBs) achieve their annual balance through photovoltaic (PV) generation, which can mask substantial thermal loads (heating, cooling, and dehumidification) in aggregate energy metrics. This paper presents a two-year measured case study of the NIST Net-Zero Energy Residential Test Facility (NZERTF), a 252 m² detached house in a mixed-humid climate, to characterise the thermal load drivers of daily net energy consumption (NC). Two feature selection methods, mutual information and LassoCV, are applied to 14 candidate variables; cross-validated Lasso selects eight predictors, including degree-days and indoor relative humidity (RH_in). Three ensemble models (RF, XGBoost, and LightGBM) are trained on Year 1 and evaluated on a 362-day Year 2 test set. XGBoost achieves the best out-of-sample accuracy (R² = 0.830), with winter posing the greatest prediction challenge. SHAP (SHapley Additive exPlanations) attribution identifies solar irradiance as the dominant driver (15.33 kWh/day) acting through dual PV-generation and cooling-demand channels. RH_in ranks fifth globally, with winter importance 1.8× higher than summer. The case study demonstrates that SHAP-based attribution of monitored building data can expose thermally distinct load drivers hidden within an otherwise balanced annual energy account.