[ARCH 689] Project 2 - Jae Yeon

[ARCH 689] Parametric Modeling in Design

Instructor - Dr. Wei Yan

Final Project - Algorithms and Scripting for Parametric Design

Melbourne Recital Centre

1. Introduction - Voronoi Algorithm

Voronoi algorithm was defined by Gerogy F. Voronoyi when he studied the general n-dimensional case in 1908.  Voronoi algorithm is a method to divide space into a number of reference points also known as seeds, sites, or generators.  Voronoi cells generate from each seed to each corresponding region as shown in Figure 1. [1] 

Fig. 1.  Demonstration of Voronoi Cell Generation [2]

2. Project Description

The Voronoi algorithm is picked up by me for a parametric facade design.  However, the Voronoi algorithm generates in a certain boundary such as a square, a circle, or a sphere. Therefore, the original facade design is changed from a unique boundary to a square boundary for a parametric as shown in Figure 2.  

Fig. 2.  Boundary Change

Each reference point will be placed in the square boundary randomly through the 'Populate 2D' node as shown in Figure 3.  Plus, the number of seeds will be controlled by the number slider for seeds. 

Fig. 3. Populate 2D Node

 3. Voronoi Algorithm for Parametric Design

1) Design for Parametric Voronoi 2D 

The Voronoi 2D algorithm needs the reference points to generate the Voronoi cell. However, reference points are picked up by the user manually.  If the user wants to show the different Voronoi designs, the user needs to pick up another reference points again and again until the expectation that the user makes a decision.  However, this is a time-consuming work and difficult to satisfy with the many opinions of participants at the same time. But the Voronoi 2D node can provide the various design options when it goes with the Populate 2D node to generate the random facade designs as shown in Figure 4.

Fig. 4.  Parametric Voronoi 2D Algorithm with Random Generated Reference Points [3]

2) Count Variable for Reference Points

The number of Voronoi cell is changed by the count slider.  The parametric design will give the various design options to determine the best fit of a facade design as shown in Figure 5. 

Fig. 5. Parametric Design with a Count Variable

3) Seed Variable for Reference Points 

The reference points are located randomly by the seed slider as shown in Figure 6.  The design options are diversified by the seed slider as a parametric variable when this variable is combine with a count variable.

Fig. 6. Parametric Design with a Seed Variable

4) Voronoi Facade Thickness

The thickness of Voronoi cell can be changed by the Z - dimensional factor slider.  It helps to figure out the proper thickness of facade.  One of possible final designs is shown in Figure 7.

Fig. 7. Final Facade Design

5) Design for Parametric Voronoi 3D

The Voronoi 3D algorithm can provide another perspective for the design options through the algorithm as shown in Figure 8.  The Populate 3D node used to generate the reference points in the 3D-space.

Fig. 8.  Parametric Voronoi 3D Algorithm with Random Generated Reference Points [3]

The reference points are located up and down by the number slider in the 3-dimensional space.  Random generated random points are in the 3D space with the wall contained Voronoi cells as shown in Figure 9. This Voronoi 3D node can be the another parametric design option because this Voronoi 3D node dose not contain the Voronoi cell thickness. This is almost portrays the original facade design of Melbourne Recital Centre. 

Fig. 9. Parametric Design with Voronoi 3D Node

4. Project Movie

References

1. Voronoi diagram, Retrieved from http://en.wikipedia.org/wiki/Voronoi_diagram, Wikipedia, (10 April 2014).
2. Edclidean Voronoi Diagram File, Retrieved from http://en.wikipedia.org/wiki/File:Euclidean_Voronoi_Diagram.png, (06 November 2013).
3. Voronoi, Retrieved from http://www.grasshopper3d.com/forum/topics/3d-voronoi-4 (2014)

[ARCH 689] Project 1 - Jae Yeon

[ARCH 689] Parametric Modeling in Design

Instructor - Dr. Wei Yan

Midterm Project - Parametric Modeling and Physically-based Form Finding

Melbourne Recital Centre

1. Introduction

Melbourne Recital Centre is located in Melbourne, Australia.  The building was designed by Ashton Raggatt McDougall and opened as a concert and recital hall as shown in Figure 1.  It is the second largest auditorium for the classical music in Melbourne. [1]

Fig. 1.  Melbourne Recital Centre [2]

2. Project Scope

The facade of the targeted project (Melbourne Recital Centre) is proper to be studied as a parametric modeling.  Since the entire modeling is improper to meet the purpose of this class (ARCH 689), my project is concentrated on the facade of  Melbourne Recital Centre.  Furthermore, the use of Rhino was minimized and the use of Grasshopper was maximized for a parametric modeling.  Last, I am devoted to the two-dimensional design to meet the project requirements. 

3. Parametric Form for the Curved Design

The facade of the project is modeled on the Grasshopper nodes as shown in Figure 2.  Each point of the facade model was taken by the Rhino model.  The shape of facade is built up by the NURBS node and the NURBS curve is extruded to the Y-direction.  Thus, the thickness of facade can be controlled by the slide node as shown in Figure 3.

Fig. 2.  Nodes for Building Contour 

Fig. 3.  Extruded Inside and Outside Facade Model

Since the project has the unique contours for the architectural aesthetics, the inside and outside facade points are needed to be defined through the designed boundary for the algorithmic model or the parametric model.  The inside points and outside points can be classified by the created nodes as shown in Figure 4 [3]. 

Fig. 4.  Nodes for the Reference Points of Facade

The determined points of the glass facade are shown in Figure 5 and 6.  The points will be the reference points for the parametric modeling with respect to the designer's intent. 

Fig. 5.  Inside Reference Points of Glass Part Facade

Fig. 6.  Outside Reference Points of Glass Part Facade

The determined points of the concrete structure part are shown in Figure 7 and 8.  The points will be the reference points for the parametric modeling with respect to the designer's intent as well.

Fig. 7.  Inside Reference Points of Concrete Structure Part

Fig. 8. Outside Reference Point of Concrete Structure Part

The pipeline node is added to change the exiting design as shown in Figure 9.  However, the shape of facade is not distorted by the reference points.   

Fig. 9. Node for the Own Design Intent to the Original Design

 This is the result of the parametric design are shown in Figure 10 and Figure 11.

Fig. 10.  Changed Design of Internal Facade 

Fig. 11. Changed Design of External Facade 

The thickness of the inside and outside facades are controlled by the designed nodes.  Also, the design of facade is changed by the designed nodes as shown in Figure 12.

Fig. 12. Changed Facade Design 

4. Parametric, Physically-Based Model

Since the project was focused on the two-dimensional facade, Unary Force in Kangaroo was used for the physically-based model as shown in Figure 13.  The initial condition of Unary Force is as shown in Figure 14 and the deflection is as shown in Figure 15.  The weight of the gravity and the numbers of knot are controlled by the parametric sliders.

Fig. 13.  Parametric, Physically-Based Model with Unary Force [4]

Fig. 14.  Initial Condition of Psysically-Based Model with Unary Force

Fig. 15.  Deflection of Psysically-Based Model with Unary Force

5. Analyses

My analysis approach is little bit different comparing with the description of the project requirement.  My analysis approach is seems like a case study to figure out the proper algorithmic design to apply the project which has a unique boundary design.

Case 1) Cellular Automata

To apply Cellular Automata for the algorithmic design of the targeted project as shown in Figure 16, the reference surface should be rectangular as shown in Figure 17.  Cellular Automata is improper for the targeted project which has the unique boundary design.

Fig. 16.  Cellular Automata Nodes for this Project [5]

Fig.  17.  Limitation of Cellular Automata

Case 2) Voronoi

Personally, I supposed that Voronoi was the proper algorithm for this project.  However, Voronoi has the limitation same as the Cellular Automata.  It should has the rectangular boundary to work it well as shown in Figure 18.    

Fig. 18.  Limitation of Voronoi [6]

Case 3) External Database for Parametric Design (gHowl)

The 'gHowl' engine is proper to apply the project which has a unique boundary design.  The user can make the own design with a unique boundary as shown in Figure 19.  The user punches in the numbers and picks the cells of Microsoft Excel sheet up where the user wants to make the model to run through the designed nodes as shown in Figure 20.  Also, if the user divides cells finely, the user can design the model in detail.  However, the user needs to enter the data manually and it takes a long time to run it.  But it is a good algorithm for the unique design with a free boundary condition as shown in Figure 21.

Fig. 19.  Input Data Type for gHowl Algorithm

Fig. 20.  Nodes for gHowl Algorithm [7]

Fig. 21.  Free Boundary Condition Design with gHowl Algorithm

6. Project Movie

References

1. Melbourne Recital Centre, Retrieved from http://en.wikipedia.org/wiki/Melbourne_Recital_Centre, Wikipedia, (17 March 2014).

2. John Madden, Retrieved from http://john-madden.com.au/index.php/portfolio-item/mrc-mtc/, Melbourne recital Centre + MTC, (2014).

3. Grasshopper Tutorial Boundary Inclusion, Retrieved from https://www.youtube.com/watch?v=-xhgakVFUoY, (25 May 2010).

4. Lance Walters, Retrieved from http://vimeo.com/36219411, Kangaroo Basic Catenary System, (2012)

5. Cellular Automata: Support for Excitable Media Cellular Automata, Retrieved from http://morphocode.com/rabbit/, Plug-in for Grasshopper Rabbit, Morphocode, (2014)

6. Voronoi, Retrieved from http://www.grasshopper3d.com/forum/topics/3d-voronoi-4 (2014)

7. Grasshopper gHowl, Retrieved from http://www.grasshopper3d.com/group/ghowl (2014)

[ARCH 653] Project 2 - Jae Yeon

[ARCH 653] Building Information Modeling in Architecture

Instructor - Dr. Wei Yan

Final Project - Accessing BIM Data for Creative Design and Applications

America's Cup Building

1. Introduction

Building Information Modeling (BIM) users encounter plenty of limitations to design the customized models or the repeatable models when they work for the projects.  To push the limitations, Application Programming Interface (API) and Visual Programming (Modified Dynamo) can be key for the tailored designs, time-consuming tasks and etc.  

Autodesk Revit has a .NET API with C# (Computer programming language).  One of advantages of C# is that it is easy-to-learn, easy-to-use and does not need to consider data capacity. [1]

The modified Dynamo install for the class project 2.  It is a user-friendly tool, especially, users who has no computer language experience.  This visual program provides each node as Window Forms Application instead of a long code as shown in Figure 1.  Also, users don't need to consider each relation of Main Method in Class of C# because users can consider each relation of Node through 'Pathing Parameter' in Dynamo.  Thus, everyone can use this program intuitively.  [2]

Figure 1.  Comparison between Microsoft Visual C# and Modified Dynamo
 

2. Parametric BIM through Application Programming Interface (API)

There are two cases.  The first case is for the panels are patterned panels that are loaded from Mass Family with Divide Surface.  The second case is for the panels are from Curtain System (by face) in the Project level of Revit.  For my project, I picked the first case up.  Every relative element ID of Mass Family was replaced by me.  Generated random colors are renamed as 'MyMaterial'.  For my project, I disabled the thickness of every Mass Family because random thickness was improper for my project.  The initial condition of facade is as shown in Figure 2. 

Figure 2.  Initial Condition of Facade

After ran code in Autodesk Revit 2014, every facade was changed by random generated color as shown in Figure 3 and 4.

Figure 3.  Applied Condition of Facade (View 1)

Figure 4.  Applied Condition of Facade (View 2)

3. Parametric BIM through Visual Programming (Modified Dynamo)

I was not able to consider the occlusion of each PV panel through Project 1.  So, I decided to improve the problem in accordance with the instructor's suggestions as shown in Figure 5. Automatically appropriated arrangement of PV panel is my goal of this part.

Figure 5.  Instructor's Comments for My 1st Project

The brief concept of occlusion problem is presented by Figure 6.  The short mushrooms of the blue shaded area cannot photosynthesize because sunshine are blocked by the tall mushrooms of the red shaded area.  

Figure 6.  Occlusion Problem Concept with Mushrooms [3]

The occlusion problem among PV panels is considered in Project 2 to get maximized solar energy.  PV panel will be arranged based on the Reference Grid Slider example of the modified Dynamo as shown in Figure 7.  Each important node is explained as shown below.   

Significant Nodes Explanation

1. Initial PV panels can be counted as many as the user wanted with x and y-axis.
2. Generated PV panels can be distributed regarding to x-axis.
3. Distributed PV panels can be arranged regarding to y-axis.
4. Developed Envelope Solar Panel can be picked by Select Family Node.

Figure 7.  Reference Grid Slider for PV Panel

The construction sequence of a sustainable energy plant with PV Panel and Wind Power Generator is shown in Figure 8.  (In here, Wind Power Generators are arranged by me manually in the spare space).

Figure 8. Construction Sequence of A Sustainable Energy Plant

4. Minor Changes

Toposurface of my first project, some models around the building, and etc. were redesigned by me to meet the Project 2 requirements as shown in Figure 9.

Figure 9.  Minor Changes

5. Video

References

1. Autodesk - My First Plug-in Training, Retrieved from http://usa.autodesk.com/adsk/servlet/index?siteID=123112&id=16777469
2. Dynamo Visual Programming for BIM, Retrieved from http://autodeskvasari.com/dynamo
3. Mush Occlusion Image, Global Illumination, Retrieved from http://renderman.pixar.com/resources/current/rms/rfmGlobalIllumination.html 

[ARCH 653] Project 1 - Jae Yeon

[ARCH 653] Building Information Modeling in Architecture

Instructor - Dr. Wei Yan

Midterm Project 1 - Parametric BIM

America's Cup Building

1. Introduction

America's Cup Building is located in Valencia, Spain and constructed from 2005 to 2006. Architect 'David Chipperfield' designed this building and he arrayed the unevenly designed floor white decks for the aesthetic aspect as shown in Figure 1. [1]   

Fig. 1. America's Cup Building

2. Building Information Modeling Diagram for America's Cup Building

(Here, the chosen project was renovated by 3 and 4 slightly to satisfy the requirements of project.)

Fig. 2. Conceptual Modeling Steps

3. Building Information Model Project

Building floors, roofs and interior walls of entire project was created as '.rvt' file type as shown in Figure 3. 

Fig. 3. An Entire Project Building Information Model

4. Parametric, Conceptual Mass Family

To control and move accurately the mass at the same time, parameters were assigned on each Mass.  Since this project was designed with unevenly designed floor deck, every deck should move simultaneously to respect the architect's design.  The applied parametric, conceptual mass family is shown as in Figure 4.

Fig. 4. Parametric, Conceptual Mass Family Model

5. Parametric Envelope (Facade) Family

A parametric envelope family was created to applied on the facade of the building. This model was redesigned by the "Rectangle Surface with Curved Surface" to load for the curtain panel family as shown in Figure 5.

Fig. 5. Parametric Envelop Family

6. Additional Parametric Envelope Family with Equation Description

Solar panel needs to move with respect to the angle of sunshine to collect the maximum energy.  The angle of sunshine is different by date, latitude of project, and hours as shown in Figure 6.  

Fig. 6. Solar Panel Model with Family Type

The solar panel inclination was changed by the following formula.[2] [3]

Each parametric solar panel was allocated on the parking lot roof as shown in Figure 7.  This prefabricated model was located on the columns. 

Fig. 7.  Prefabricated Solar Panels

However, there are some limitations.  In here, n is the day of the year.  But it is impossible to pick up the real date.  Thus, the number day from the New Years day is punched into the blank manually.  For instance, n is 31 for 31 January. Also, hours should be selected from 1 to 24.  Lastly, the latitude of the project location is 40 degree North and 4 degree West.  Theses limitation can be solved by coding through the Revit Application Program Interface.

7. Rendering Images

Fig 8. Exterior Rendering

Fig 9. Interior Rendering

8. Video

References

1. America's Cup building, Retireved from http://en.wikiarquitectura.com/index.php/America's_Cup_Building, en.wikiarquitectura, (22 August 2013)
2. Manual of Solar Collectors and Photovoltaic in Energy PRO, Report, (2009)
3. Bernard Bourges, European Simplified Methods for Active Solar System Design, July 1990)