Heritage-Informed Design: AI-Enhanced Vernacular Serambi Strategies for Regenerative Urban Mosques

This study investigates how vernacular architectural intelligence, specifically the serambi, can inform sustainable and regenerative mosque design in rapidly urbanizing Malaysian contexts. Modern mosques increasingly rely on energy-intensive cooling systems despite the longstanding climatic efficacy of vernacular Malay architecture. Addressing this gap, the research integrates field measurements, Computational Fluid Dynamics (CFD), and Artificial Intelligence (AI)–assisted prediction to evaluate the thermal performance of varying Serambi Opening-to-Wall Ratios (OWR). Empirical data were collected across five daily prayer periods at Ara Damansara Mosque over seven days and subsequently used to calibrate IESVE-based CFD simulations. A supervised Artificial Neural Network (ANN) was then trained to predict thermal comfort indicators across multiple façade configurations. Results demonstrate that an OWR of 60–70% optimally balances ventilation, temperature reduction, humidity control, and daylighting, reducing indoor air temperature by 3–4 °C and achieving comfortable air velocities between 1.0–1.5 m/s. The AI-CFD framework demonstrates strong predictive accuracy (R2 > 0.94), underscoring its utility for generative design and early-stage decision-making. This research positions vernacular knowledge not as nostalgic heritage but as a regenerative, culturally embedded design intelligence that supports sustainable urban development. The findings contribute a validated, heritage-informed design model that can strengthen contemporary mosque performance while advancing the discourse on decolonial and climate-responsive urban regeneration.