Yeralti Mi̇mari̇si̇nde Ti̇poloji̇k Siniflandirma Ve Yapay Zekâ Destekli̇ Mekânsal Temsi̇l: Deri̇nkuyu Yeralti Şehri̇ Örneği̇
Year: 2026
Type: article
Abstract: This study examines the Derinkuyu Underground City—one of the largest urban-scale underground settlements in the world—through the lenses of architectural typology and spatial representation. The aim is to render the complex spatial organization of an organically developed, rock-cut settlement with indistinct floor boundaries legible by means of a systematic typological classification and an interpretive virtual-reconstruction approach. The research relies on a multi-stage, qualitative method: spaces were documented through field observation and measured surveys; a typological classification system was developed on the basis of functional, formal, volumetric and relational criteria; and selected representative spaces were reconstructed through an AI-assisted, scenario-based and probabilistic digital workflow. The findings show that spatial organization in Derinkuyu constitutes a network-based, relational and flexible system defined by accessibility, circulation intensity and control mechanisms rather than by a floor-based linear order. Six main typological groups were identified, and it is demonstrated that threshold spaces—narrowed passages, abrupt level changes and rolling-stone doors—operate not merely as physical connectors but as an active control device regulating privacy and security. In the representation dimension, two use scenarios were constructed for each of three representative typologies (church, kitchen, living–resting space); these representations reveal that the same physical volume can produce distinct spatial identities depending on use practices, and that digital reconstruction can serve as an analytical, knowledge-producing method rather than a mere visualization tool.
Source: DergiPark
Authors: Nas, Sezin, Aytıs, Saadet
Language: Turkish
Subfield: Urban Studies
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7169690727
