Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

The Breathing Skeleton, Algorithmic Service Routing as a Tectonic Generator

Rishi, Bhavya LU (2026) ASEM01 20261
Department of Architecture and Built Environment
Abstract
The design process of buildings has historically treated
mechanical, electrical, plumbing (MEP) Systems,
particularly Heating, Ventilation and Air Conditioning
(HVAC) as an afterthought which leads to compromised
spatial qualities inefficient routing systems. This
systemic marginalisation of environmental engineering
withing architectural discourse has traditionally
prioritised the visual form of the structural envelope
over the thermodynamic performance of its interior.
Consequently, modern architectural practices frequently
encounter severe downstream conflicts during the
construction phase, necessitating costly revisions and
destructive structural compromises. (Abdelhameed &
Saputra, 2020)
This thesis explores the... (More)
The design process of buildings has historically treated
mechanical, electrical, plumbing (MEP) Systems,
particularly Heating, Ventilation and Air Conditioning
(HVAC) as an afterthought which leads to compromised
spatial qualities inefficient routing systems. This
systemic marginalisation of environmental engineering
withing architectural discourse has traditionally
prioritised the visual form of the structural envelope
over the thermodynamic performance of its interior.
Consequently, modern architectural practices frequently
encounter severe downstream conflicts during the
construction phase, necessitating costly revisions and
destructive structural compromises. (Abdelhameed &
Saputra, 2020)
This thesis explores the integration of HVAC services
during the conceptual architectural design stage using
generative routing algorithms. Recognizing ventilation as
the most space occupying building service, this project
shifts away from traditional manual routing and explores
computational path finding. Initial experiments with
Agent-based Modelling (ABM) proved too chaotic and
unstandardized, leading to the formulation of a custom
rule-based logic utilising ‘Minimum Spanning Trees’ and
‘Steiner Minimal Trees’. These algorithmic methodologies
operate within the parametric environment of
Grasshopper and Python, establishing rigorous
mathematical constraints that prioritize orthogonal
efficiency, structural clash avoidance, and optimal
thermodynamic distribution metrics.
These algorithmic workflows are applied to a real
world renovation case study, Academic Block 3 at the
Indian Institute of Technology (IIT) Delhi. By developing
Manhattan MST for the efficient rectilinear routing on
the ground floor, and expressive Steiner trees on the
first floor, the research culminates in a newly proposed
lightweight top floor. Here, a 3D Steiner network
dictates both the HVAC ducting and the structural
framework simultaneously, featuring custom 3D joints
that integrate air distribution and load transfer. The
findings demonstrate that generative routing can elevate
building services from hidden necessities to the primary
architectural design drivers, bridging the intellectual and
physical gap between the structural skeleton and the
mechanical respiratory systems of a building. (Less)
Please use this url to cite or link to this publication:
author
Rishi, Bhavya LU
supervisor
organization
course
ASEM01 20261
year
type
H2 - Master's Degree (Two Years)
subject
keywords
Algorithmic, Grasshopper, Rhino, Building Services, HVAC
language
English
id
9233321
date added to LUP
2026-06-09 08:58:19
date last changed
2026-06-09 08:58:19
@misc{9233321,
  abstract     = {{The design process of buildings has historically treated
mechanical, electrical, plumbing (MEP) Systems,
particularly Heating, Ventilation and Air Conditioning
(HVAC) as an afterthought which leads to compromised
spatial qualities inefficient routing systems. This
systemic marginalisation of environmental engineering
withing architectural discourse has traditionally
prioritised the visual form of the structural envelope
over the thermodynamic performance of its interior.
Consequently, modern architectural practices frequently
encounter severe downstream conflicts during the
construction phase, necessitating costly revisions and
destructive structural compromises. (Abdelhameed &
Saputra, 2020)
This thesis explores the integration of HVAC services
during the conceptual architectural design stage using
generative routing algorithms. Recognizing ventilation as
the most space occupying building service, this project
shifts away from traditional manual routing and explores
computational path finding. Initial experiments with
Agent-based Modelling (ABM) proved too chaotic and
unstandardized, leading to the formulation of a custom
rule-based logic utilising ‘Minimum Spanning Trees’ and
‘Steiner Minimal Trees’. These algorithmic methodologies
operate within the parametric environment of
Grasshopper and Python, establishing rigorous
mathematical constraints that prioritize orthogonal
efficiency, structural clash avoidance, and optimal
thermodynamic distribution metrics.
These algorithmic workflows are applied to a real
world renovation case study, Academic Block 3 at the
Indian Institute of Technology (IIT) Delhi. By developing
Manhattan MST for the efficient rectilinear routing on
the ground floor, and expressive Steiner trees on the
first floor, the research culminates in a newly proposed
lightweight top floor. Here, a 3D Steiner network
dictates both the HVAC ducting and the structural
framework simultaneously, featuring custom 3D joints
that integrate air distribution and load transfer. The
findings demonstrate that generative routing can elevate
building services from hidden necessities to the primary
architectural design drivers, bridging the intellectual and
physical gap between the structural skeleton and the
mechanical respiratory systems of a building.}},
  author       = {{Rishi, Bhavya}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{The Breathing Skeleton, Algorithmic Service Routing as a Tectonic Generator}},
  year         = {{2026}},
}