deakin research onlinedro.deakin.edu.au/eserv/du:30033114/roetzel-optimisedfacade-2009.pdf · do we...

Deakin Research Online This is the published version of the Power Point presentation: Tsangrassoulis, A. and Roetzel, A. 2009, Optimized façade design using pattern searching, in Proceedings of the International Sustainability and Building SMART conference, Oslo, Norway, September 24th 2009, The Conference, [Oslo, Norway]. Available from Deakin Research Online: http://hdl.handle.net/10536/DRO/DU:30033114 Every reasonable effort has been made to ensure that permission has been obtained for items included in Deakin Research Online. If you believe that your rights have been infringed by this repository, please contact [email protected] Copyright : 2009, BuildingSMART

http://hdl.handle.net/10536/DRO/DU:30033114

mailto:[email protected]

• Tsangrassoulis1 & A. Roetzel21Dept. of Architecture, Univ. of Thessaly, Volos, Greece

2Dept. of Architecture, Hafencity University, Hamburg, Germany

Optimized façade design using pattern searching

Building’s environmental impact

40% of primary energy

72% of electricity

39% of CO2 emmissions

13.6% of potable water

www.freedigitalphotos.net

http://www.freedigitalphotos.net/

Increase in building design complexity + tight energy legislation

Location

Building morphology

Materials-Components

HVAC systems

Control of systems

Building Operation

Other design needs

Energy Performance Assessment

E-man Studio, Athens

Optimization process intime consuming since many antagonistic issues have to be resolved

Shading

Lighting

Problem: Optimize energy consumption-comfort

Solution: Explore large regions of possible solutions fast (global optimization methods)

“Land” of solutions

We must approachminimum during design phase

Ener

gy c

onsu

mpt

ion

Parameter #1

Parameter #2

Do we have to go through all possible solutions ?

Requirements when a building design optimization problem is approached:

• Existence of non-analytic objective function• Non-linear inequality constrain• Time consuming simulation• Use of continouus and dicrete design variables

Genetic algorithms

• Fitness function (I.e. energy consumption)

• Definition of parameters i.e. window width varying from 0.1 to 3.1 m with 0.1 m steps.(3.1-0.1*10 which is < 25. Thus 5 bits to describe this parameter.

• 0110010011000111001 This is a chromosome

window width window height Other parameter

Genetic algorithms

01100100110001110011100101101011100011011000011000000000100000011101110001101001111111001111000………………………….………………………….………………………….………………………….…………………………..

Generation #1

Simulation to estimate fitness function

01100100110001110011100101101011100011110010110101110001100000011101110001100000001110111000110………………………….………………………….………………………….………………………….…………………………..

011001 11010111000111100100011000111001110010 1110111000110000000 1101011100011………………………….………………………….………………………….………………………….………………………….…………………………..

Generation #2

SELECTION CROSSOVER&

MUTATION

....

.

..

. . ..

.

..

..

........

..... . ..

Genetic algorithms

Generation #1 Generation #n

Hooke-Jeeves pattern search

0 1 2 3 4 5 6 7 8 90

1

2

3

4

5

6

7

x

x

Starting pointExploratory moves

Pattern moves

What happens if starting point is here ?

MIT Design Advisor COMFEN

Focus on the facade

80% of building’s cost have been determined during the conceptual design phase.

You have to design first and then energy consumption is calculated

New tools have been presented recently

Case study using Genopt+EnergyPlus

5.4

m

3.5 m

Plan

EnergyPlus used to evaluatePrimary Energy Consumption

GENOPT “drives” the optimizationprocedure


θ

BlindsSouth orientation


Closed blinds Horizontal blinds


Optimization procedure took into account :• Four cardinal Orientations• Window size• Three glazing types (clear, Spectrally selective, reflective)• Two external shading systems (blinds, overhang)• Dimming

South orientation


Average processing time with Core 2 Quad 2.66 was ~15 min


Spectrally selectiveglazing

Clearglazing

Reflectiveglazing

.


#1

#2

#8

Alternative design scenarios

No Window sill


Spectrally selectiveglazing

Clearglazing

Reflectiveglazing

.


•Optimization is extremely useful during the initial design stage

•For simple façade design problems setting up Genopt is relatively fast

•Split up the problem to different alternative scenarios.

• Using a lighting control system (with or without shading) have a major impact on opening size and position.

• Changing objective function can drive to different design

How is defined an objective function thattakes into account both energy consumption and comfort?

• Unfortunately, without knowing the mathematical limitations results can be problematic

Thank you for your attention

deakin research onlinedro.deakin.edu.au/eserv/du:30033114/roetzel-optimisedfacade-2009.pdf · do we...

Documents