A Data-Driven Validation Methodology for a Multi-City Solar Chimney Dataset in Smart-City Applications

Abstract

This article presents a methodology for a validated multi-city parametric dataset for data-driven assessment of single-channel solar chimneys as energy-efficient components of smart-city buildings. The dataset covers Mexico City, Hermosillo, Mérida, Monterrey, Toluca, and Villahermosa across tropical, dry, and temperate Köppen–García climates. The proposed dataset combines nine absorber-wall materials, four cardinal orientations, and 24 representative days (the warmest and coldest day of each of the twelve calendar months), generating 5184 CSV files (6×24×4×9=5184). A transient global energy balance model calculates heat conduction, natural convection, radiative exchange, and buoyancy-driven ventilation under real climatic conditions included in the dataset. The GEB model is experimentally validated against an aluminum-absorber prototype. The energy balance model discretizes massive and phase-change absorber walls into 21 layers and represents latent thermal storage through an effective specific heat formulation. Consequently, the ∼133.64 GB dataset records time-dependent temperatures, airflow rates, air changes per hour, instantaneous efficiency, and layer-resolved liquid fractions. A proposed open Python 3.12.7toolkit also organizes the heterogeneous file hierarchy and enables reproducible, scalable retrieval of specific configurations. The resource supports big-data analytics, surrogate modeling, machine-learning training, predictive performance assessment, and data-driven decision-making for passive ventilation. Furthermore, the dataset supports future integration with IoT-based building monitoring systems for model calibration, operational prediction, and automated energy management. Overall, the proposed resource advances scalable, climate-responsive solutions for resilient, energy-efficient, and sustainable smart cities.